Designing the European Model
9780230_547018_01_prexiv.indd i
3/19/2009 7:58:31 PM
Also by Seppo Honkapohja: (co-edi...
22 downloads
1064 Views
3MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Designing the European Model
9780230_547018_01_prexiv.indd i
3/19/2009 7:58:31 PM
Also by Seppo Honkapohja: (co-editor K. Arrow) FRONTIERS OF ECONOMICS (co-editor F. Försund) LIMITS AND PROBLEMS OF TAXATION (co-editors W. Bergström and J. Södersten) GROWTH AND DISTRIBUTION: INTERGENERATIONAL PROBLEMS INFORMATION AND INCENTIVES IN ORGANIZATIONS THE STATE OF MACROECONOMICS (co-editor M. Ingberg) MACROECONOMIC MODELLING AND POLICY IMPLICATIONS (co-editors T. Gylfason, T. Andersen, A. Jon Isachsen and J. Williamson) THE SWEDISH MODEL UNDER STRESS: A VIEW FROM THE STANDS (co-author G. W. Evans) LEARNING AND EXPECTATIONS IN MACROECONOMICS Also by Frank Westermann: (co-author A. Tornell) BOOM-BUST CYCLES AND FINANCIAL LIBERALIZATION (co-editor M. M. Hutchison) JAPAN’S GREAT STAGNATION, FINANCIAL AND MONETARY POLICY LESSONS FOR ADVANCED ECONOMIES
9780230_547018_01_prexiv.indd ii
3/19/2009 7:58:32 PM
Designing the European Model Edited By
Seppo Honkapohja Bank of Finland and Former Chairman of the EEAG
and
Frank Westermann University of Osnabrueck, Germany and Former Secretary General of the EEAG
9780230_547018_01_prexiv.indd iii
3/19/2009 7:58:32 PM
Selection and editorial matter © Seppo Honkapohja and Frank Westermann 2009 Individual chapters © contributors 2009 All rights reserved. No reproduction, copy or transmission of this publication may be made without written permission. No portion of this publication may be reproduced, copied or transmitted save with written permission or in accordance with the provisions of the Copyright, Designs and Patents Act 1988, or under the terms of any licence permitting limited copying issued by the Copyright Licensing Agency, Saffron House, 6-10 Kirby Street, London EC1N 8TS. Any person who does any unauthorized act in relation to this publication may be liable to criminal prosecution and civil claims for damages. The authors have asserted their rights to be identified as the authors of this work in accordance with the Copyright, Designs and Patents Act 1988. First published 2009 by PALGRAVE MACMILLAN Palgrave Macmillan in the UK is an imprint of Macmillan Publishers Limited, registered in England, company number 785998, of Houndmills, Basingstoke, Hampshire RG21 6XS. Palgrave Macmillan in the US is a division of St Martin's Press LLC, 175 Fifth Avenue, New York, NY 10010. Palgrave Macmillan is the global academic imprint of the above companies and has companies and representatives throughout the world. Palgrave® and Macmillan® are registered trademarks in the United States, the United Kingdom, Europe and other countries. ISBN-13: 978–0–230–54701–8 hardback ISBN-10: 0–230–54701–X hardback This book is printed on paper suitable for recycling and made from fully managed and sustained forest sources. Logging, pulping and manufacturing processes are expected to conform to the environmental regulations of the country of origin. A catalogue record for this book is available from the British Library. A catalog record for this book is available from the Library of Congress. 10 9 8 7 6 5 4 3 2 1 18 17 16 15 14 13 12 11 10 09 Printed and bound in Great Britain by CPI Antony Rowe, Chippenham and Eastbourne
9780230_547018_01_prexiv.indd iv
3/19/2009 7:58:32 PM
Contents List of Figures
vi
List of Tables
viii
List of Boxes
xi
Preface
xii
Members of the EEAG
xiii
Introduction
1 Part I Labour Markets
1 Welfare to Work (from the 2002 EEAG Report) 2 Labour Market Reform in Europe (from the 2004 EEAG Report) 3 Pay-setting Systems in Europe: Ongoing Developments and Possible Reforms (from the 2004 EEAG Report) 4 Outsourcing (from the 2005 EEAG Report) 5 Longer Working Hours – The Beginning of a New Trend? (from the 2005 EEAG Report) Part II 6
7 8 9
11 12 13 14
82 122 144
Macroeconomic Issues
Fiscal Policy and Macroeconomic Stabilisation in the Euro Area: Possible Reforms of the Stability and Growth Pact and National Decision-Making Processes (from the 2003 EEAG Report) Acceding Countries: The Road to the Euro (from the 2004 EEAG Report) Economic Growth in the European Union (from the 2006 EEAG Report) Global Imbalances (from the 2006 EEAG Report) Part III
10
35 61
177 229 259 296
Institutions
Rethinking Subsidiarity in the EU: Economic Principles (from the 2003 EEAG Report) Financial Architecture (from the 2003 EEAG Report) Pensions and Children (from the 2005 EEAG Report) Prospects for Education Policy in Europe (from the 2006 EEAG Report) Mergers and Competition Policy in Europe (from the 2006 EEAG Report)
Index
331 366 399 447 469 497
v
9780230_547018_01_prexiv.indd v
3/19/2009 7:58:32 PM
Figures 1.1 1.2 1.3 1.4 1.5 2.1 2.2 2.3 3.1
3.2 3.3 3.4 4.1 4.2 4.3 4.4 4.5 4.6 4.7 5.1 5.2 5.3 5.4 5.5 6.1 6.2 7.1 7.2 7.3 8.1a 8.1b
Social assistance and earnings dispersion in the OECD, 1997 Duration of unemployment insurance Wage replacement versus wage supplement US earned income tax credit for various household types Welfare caseloads, 1987–1999 The equilibrium rate of unemployment Effects of structural reforms on unemployment Unemployment rate in the United Kingdom The relationship between bargaining co-ordination (centralisation) of collective bargaining and the real wage (unemployment) Earnings dispersion Relative wage per employee and relative productivity per employee in Eastern versus Western Germany Unemployment rates in Eastern and Western Germany Share of value added in the manufacturing sector’s own production Share of manufacturing in GDP Employment in manufacturing sector Hourly labour cost in Europe, 2002 Change in net foreign direct investment (stocks) Share of manufacturing value added in total value added and manufacturing employment in total employment Bazaar economy Annual hours worked per capita, 2005 Decomposition of differences in hours worked per capita Decomposition of the trend growth in labour utilisation 1970–2005 Desired working time Real wages and employment in the long run Gross government debt/GDP, 1970–2003 The actual budget balance as a function of the output GAP Nominal effective exchange rates Share of foreign currency liabilities in the banking system in 1999 Ratio of tradables-to-non-tradables prices Index of real GDP per capita Index of real GDP per capita
40 43 46 50 55 67 68 73
94 98 99 100 124 126 127 129 130 138 139 145 146 148 156 165 180 191 233 240 254 261 262
vi
9780230_547018_01_prexiv.indd vi
3/19/2009 7:58:32 PM
Figures
8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9 9.1a 9.1b 9.2 9.3 9.4 10.1 12.1 13.1 13.2 13.3 13.4 14.1 14.2 14.3 14.4
Cumulative real growth per capita, 1995–2004 Difference between GDP and GNP growth from 1995 to 2004 Beta-convergence among EU25 countries Income dispersion (sigma convergence) Comparison of income distribution across the EU25 Information and communication technology expenditure, 2003 Personal computers per 1000 people, 2003 Expenditure on education in EU countries, 2002 Current account balances Net foreign assets Official reserves as a percentage of total foreign liabilities, developing countries US current account balance versus US stock of foreign assets and liabilities United States: Net portfolio debt and equity cum FDI positions Public expenditure of country, state and local governments in selected federalist nations in 1996 Estimated effective retirement age of older male workers in 2000 Cumulative spending per student vs. mathematic proficiency Students per teacher vs. mathematic proficiency Students per teacher vs. reading proficiency The Finnish education system Value of M&A deals by region of the sellers, 1981–2004 Number of M&A deals by region of the sellers 1981–2004 Value of cross-border M&A purchases Value share of hostile M&A sales by region of the seller
9780230_547018_01_prexiv.indd vii
vii
263 264 265 266 267 276 277 280 300 300 302 303 310 341 407 455 456 456 460 471 471 472 474
3/19/2009 7:58:32 PM
Tables 1.1 Net replacement rates by family type at the APW level, 1997 1.2 Unemployment rates by education for population 25 to 64 years of age, 1999 1.3 Underground economy as a percentage of GDP, 1998 1.4 Periods of benefit sanction following a voluntary quit and refusal of work or an ALMP placement 1.5 Welfare-to-work programmes 1.6 US federal earned income tax credit parameters in 2000 3.1 Coverage of collective agreements and unionisation 3.2 Bargaining levels 3.3 Unemployment rates under various bargaining regimes (ceteris–paribus differences to decentralised systems) in various studies 3.4 Unemployment rates under various rates of union density and coverage of collective agreements (ceteris–paribus differences to 15% union density or coverage in different studies) 3.5 The coverage of collective agreements in eastern and western Germany, 2002 3.6 Wage dispersion in Swedish hospital care run by regional authorities 4.1 International outsourcing – The share of foreign intermediate products in totalintermediate products 5.1 Real income, labour productivity, and labour input in EU-15 in percent of US levels 5.2 Standard working time for full-time workers according to collective agreements and/or legislation, 2003 5.3 Average actual and standard working time for full-time employees in EU-15, 2002 5.4 Major reductions in the standard work week in European economies, 1980–2004 6.1 Government spending, excluding interest payments, as a percentage of GDP in the EU countries 6.2 Gross government debt as a percentage of GDP in the EU countries, 1980–2003 6.3 General government actual fiscal balance (net lending) as a percentage of GDP in the EU countries 6.4 General government cyclically adjusted fiscal balance as a percentage of GDP in the EU countries
39 41 42 44 49 51 84 87
91
95 106 109 125 145 149 150 151 184 188 193 193
viii
9780230_547018_01_prexiv.indd viii
3/19/2009 7:58:32 PM
Tables ix
6.5 Projected increases in age-related public expenditures in the EU countries in percent of GDP, 2000–2040 6.6 Required primary surpluses in the EU countries to stabilise the debt ratio after 2020 under various assumptions (debt ratios in parenthesis) 6.7 Deficit levels consistent with a sdebt-to-GDP ratio 6.8 Possible ways of conditioning the deficit ceiling on the debt ratio 6.9 The fiscal position of accession countries in 2002 6.10 Different measures of government debt as percentages of GDP for the EU countries, 2001 6.11 Correlations between the government debt measures in 6.10 7.1 The evolution of exchange rate regimes in acceding countries 7.2 Outstanding euro-denominated bank-deposits 8.1 Growth accounting for the US 8.2a Growth accounting for Germany 8.2b Growth accounting for France 8.2c Growth accounting for Italy 8.3a Growth accounting for Ireland 8.3b Growth accounting for Finland 8.3c Growth accounting for Sweden 8.3d Growth accounting for UK 8.4a Growth accounting for Spain 8.4b Growth accounting for Greece 8.5 Growth rates of IT and non-IT capital services in EU-15 countries, 1995–2004 8.6 Average share of GDP imputed to IT and non-IT capital, 1995–2004 8.7 Changes in total annual hours worked, annual hours per employee and number of workers employed, 1995–2004 8.8 TFP growth in EU-15 countries, 1995–2004 8.9a Product market regulation in EU-15 countries and the US 8.9b Strictness of employment protection legislation in EU-15 countries and the US 8.10 Early-stage versus (expansion and replacement) venture capital investment in EU-15 countries and the US percentage of GDP 8.11 Expenditure on R&D in EU-15 countries and the US 8.12 Growth rates of total hours worked in Eastern Europe A8.1a Strictness of employment protection legislation, selected new EU members A8.1b Product market regulation, selected new EU members
9780230_547018_01_prexiv.indd ix
198
200 204 207 208 211 212 234 239 269 270 270 270 271 271 272 272 272 272 274 275 278 279 282 283
284 285 286 291 291
3/19/2009 7:58:32 PM
x
Tables
A8.2a Early-stage venture capital investment in selected new EU members A8.2b Expansion and replacement venture capital investment in selected new EU members A8.3 Expenditure on R&D in selected new EU members 9.1 Private and official financing of the US current account deficit 9.2 Official reserves in percent of total foreign liabilities for developing countries 10.1 EU expenditures by function 10.2 Central, state and local government expenditures by function in % of total expenditure – 1996 10.3 Eurobarometer-Policies: National or EU level decision-making – Results in % of EU average 10.4 Eurobarometer-Policies: National or EU level decision-making – Results in % by country 11.1 Merger and acquisition activity in the euro-area financial industry 11.2 Fiscal costs of select banking crisis 12.1 Dependency ratios in selected countries 12.2 The effects of ageing on per capita output 12.3 The relative performance of a pension fund invested in stocks, relative to a pension fund invested in bonds 13.1 PISA ranking in average math proficiency 13.2 PISA ranking in average reading proficiency 13.3 The soaring cost of education in the United States A13.1 The institutional characteristics of educational systems 14.1 M&As formally blocked by EU
9780230_547018_01_prexiv.indd x
291 291 292 301 302 340 342 345 347 371 372 403 406 415 449 451 457 465 486
3/19/2009 7:58:32 PM
Boxes 2.1 2.2 2.3 2.4 4.1 5.1 6.1 6.2 6.3 6.4 7.1 7.2 9.1 9.2 9.3 10.1 10.2 11.1 11.2 11.3 11.4 11.5 11.6 12.1 12.2 12.3 12.4 13.1 13.2 13.3 13.4 13.5 14.1
The equilibrium rate of unemployment Sweden’s deregulation in the 1990s Lags in the effects of British labour market reforms Labour Market Reform in Germany Competitiveness Some formulas for the short-run employment effects of longer working hours at unchanged pay per worker The cyclically adjusted budget balance Long-run government debt The golden rule Different measures of the government’s financial situation The nominal convergence criteria Financial fragility and the sustainability of hard pegs External imbalances, the current account balance and the balance of payments Real return differentials in favour of the US Useful exchange rate and international relative price definitions Heterogeneity and inefficiency of uniform decisions German unification and social dumping The rationale for financial regulation Financial stability facilities Regulation, prudential supervision and moral hazard Optimal regulatory design Lamfalussy diagnosis and priorities Basel II Why a social security system? Shifting to a fully funded system cannot improve everybody’s welfare: an example Pensions reforms as a way to circumvent inefficiencies coming from taxes and regulations An actuarial system reduces the tax wedge The cost disease in economics The French ZEP experience Main features of the Finnish education system The Milwaukee Voucher Programme Education reform in the UK Why mergers happen in waves
66 71 72 78 135 160 195 197 202 209 230 245 297 305 308 334 359 374 375 376 380 387 388 400 410 416 422 453 457 459 462 463 477
xi
9780230_547018_01_prexiv.indd xi
3/19/2009 7:58:32 PM
Preface The European Economic Advisory Group (EEAG) at CESifo was formed in 2001 with the aim to provide economic analysis and policy recommendations to important issues of contemporary economic policy in Europe. EEAG has met several times every year in Munich and other European locations. This work has resulted in an annual report in which the group takes a stance on what they believe are some of the most important topics to be addressed and the key policy conclusions that are needed to enhance economic growth and welfare in the European Union. The first EEAG Report on the European Economy was published in 2002 and five reports were published in the years through to 2006. The material in the reports continues to be used as input to policy analysis and discussions, and in teaching European economic policy issues. At the five-year anniversary of the group it seemed appropriate to republish a selection of the material in the 2002 to 2006 EEAG reports in book form. This book presents a collection of individual chapters from these reports, which – grouped by topic – form an analytical view on a number of issues, such as labour market problems, macroeconomic policy, and the design of European institutions, which continue to be topical in European debates. The book also has an extensive introduction by the editors. It contains the main policy conclusions of the group, with updates to any major changes that have taken place since the initial publication of the chapters. The chapters are reproduced in their original form and the members of EEAG are joint authors of them. The composition of the group has varied somewhat over the five-year period as indicated below. We hope that the republication of the material facilitates the use of the economic analysis and policy conclusions by the group in the wide variety of current European concerns. Cambridge and Osnabrueck, April 2007 Seppo Honkapohja and Frank Westermann
xii
9780230_547018_01_prexiv.indd xii
3/19/2009 7:58:32 PM
Members of the EEAG* Lars Calmfors, University of Stockholm (2003–2006) Giancarlo Corsetti, European University Institute, Florence John Flemming, Wadham College, Oxford (2002–2003) Seppo Honkapohja, Universities of Cambridge and Helsinki John Kay, St. John’s College, Oxford (2003–2006) Willi Leibfritz, OECD (2002–2005) Gilles Saint-Paul, Universitè des Sciences Sociales, Toulouse Hans-Werner Sinn, ifo Institut and Universität München Jan-Egbert Sturm, KOF, ETH Zürich (2006) Xavier Vives, IESE and INSEAD Frank Westermann served as the Secretary General of EEAG in 2002–2005. In the name of the group, we also wish to thank Doina Radulescu, Sascha O. Becker and Tobias Seidel for assistance to the group. Wolfgang Ochel, Wolfgang Nierhaus, and Wolfgang Meister (support on the content), Gebhard Flaig, Timo Wollmershäuser, Oliver Hülsewig and Oscar-Erich Kuntze (economic forecast), Siegfried Schönherr, Volker Rußig, Ludger Wößmann and Martin Werding (comments), Heidemarie C. Sherman and Paul Kremmel (editing), as well as Elsita Walter (statistics and graphics) and Elisabeth Will (typesetting and layout). Finally, we thank Katja Drechsel and Nammette Lindenberg for editorial support when finalising this book.
* The years in parentheses indicate membership years in the group when they have not served the whole five-year period. xiii
9780230_547018_01_prexiv.indd xiii
3/19/2009 7:58:32 PM
This page intentionally left blank
Introduction Seppo Honkapohja, Frank Westermann
Part I: Labour Markets Unemployment rates vary between 4.5 percent in countries like Ireland and Austria and close to 10 percent in Germany, France and Spain. In some of the Eastern European countries, the figures are even above 15 percent (For instance in Slovakia and Poland). Given these high unemployment rates, one of the most pressing issues in contemporary economics is how to deal with various problems in the labour markets. The economic and social consequences of labour market problems are tremendous, and since the beginning of the EEAG, the group has devoted at least one chapter to this topic every year.
1
Welfare to Work
In its first report, in 2002, the group presented a concrete model of how the welfare state could be reformed in order to tackle this important problem. It has been argued that traditional social programmes of the modern welfare state have concentrated too much on replacing incomes for those who do not have a job, thereby generating implicit minimum wage expectation that the private industry would have to overbid to employ the workers. This offers an incentive to those capable of earning only very low wages to qualify for (higher) benefits by declining jobs that, as a result, are also not offered. In this chapter an alternative is developed, which is already implemented in varying degrees in a number of countries, in which tax credits are used to supplement the wages available to low productivity workers. A detailed proposal is put forward on a basis which should allow the living standards of both the working and most of the non-working poor to rise at no net cost to governments, while raising employment output and growth. In essence, it implies requiring government work in exchange for existing welfare benefits, cutting welfare benefits for those who do not work although they are 1
9780230_547018_02_int.indd 1
3/19/2009 8:00:10 PM
2
Designing the European Model
classified as being able to, and paying a wage subsidy to those who take low paid jobs in the private sector. Since the appearance of this first report of the EEAG, this issue has entered the discussion in Germany and several other countries. Concrete policy changes, like the German “Agenda 2010”, the “Harz” Reforms, or the present discussion of the “Kombi-Lohn” (a special form of wage subsidies) have been inspired by this proposal as well as the parallel “activating social aid” proposal of the Ifo Institute. The chapter provides a coherent conceptual structure to think about welfare state reform of this kind and gives in detail a possible solution to current labour market problems.
2
Labour Market Reform in Europe
Because of political constraints, the margin of manoeuvre for governments to combat unemployment with institutional reforms is typically quite narrow. The second chapter of this book analyses alternative ways of labour market reform that can reduce unemployment, perhaps less efficiently, but which would encounter less opposition. Furthermore, it is argued that present circumstances in Europe could make more radical reforms possible in the near future, pointing out several reasons why governments have become more ambitious in this area. An important complement to labour market reform is product market liberalisation. Increasing competition in product markets can have a strong beneficial effect on the equilibrium rate of unemployment. This is because increased competition raises firms’ productivity and imposes discipline on pay setting, since it reduces monopoly rents that are available to workers. In Sweden, for example, aggressive product market deregulation in the 1990s may have contributed to the significant fall in unemployment. A second approach consists in eliminating inefficiencies in the welfare system without radical changes in its basic structure. Relevant measures include the suggestion that dismissal costs are made more efficient in many countries by replacing the current system of legal procedures with a simple “firing tax”, which would be paid to the worker as severance payment. Alternatively, as discussed above, incentives to work could be increased by replacing means-tested welfare payments with in-work benefits such as earned-income tax credits. Finally, search activity of the unemployed should be tightly monitored, with sanctions in the form of reduced benefits if search is not active. However, an open issue in this respect is how to make sure that the unemployment insurance administration engages in effective monitoring and applies sanctions. Third, there are specific institutions or reforms that may enhance convergence of the interests of employees (who do not profit from many reforms) and firms (who gain from reforms). An important example is provided by “profit sharing” and the promotion of stock ownership among workers.
9780230_547018_02_int.indd 2
3/19/2009 8:00:11 PM
Introduction 3
These can make policies of wage moderation, which boost profitability and job creation, more acceptable to workers. If adopted on a large scale, they can be an efficient means of reducing equilibrium unemployment. A second example is provided by recent reforms liberalising temporary employment contracts, which have increased labour market flexibility in many European countries. These reforms have met with much criticism. For instance, it has been pointed out that the reforms may increase the protection of permanent workers. Yet, there has been substantial mobility from temporary to permanent contracts in many countries, and firms value temporary contracts as a way to test the quality of new workers. The experiences of the United Kingdom and, to a lesser extent, Spain suggest that in situations perceived as “crisis”, government can be more ambitious in pushing reforms directly targeted at the labour market. In principle, a critical point has also been reached in a number of European countries, not so much because of overall macroeconomic performance, but because of budgetary problems and the feeling that “globalisation” is making the burden of labour rigidities unbearable. The chapter gives a list of further main factors that could promote, or delay, reforms in the next few years. One of these factors could, paradoxically, be the very effort made by governments to combat unemployment. Some of the policies might make it an even bigger problem, as they tend to increase social spending per unemployed worker. Active labour market policies in France and Sweden require considerable spending per recipient. But in France they have failed to produce a reduction in unemployment, and in Sweden their efficiency has been questioned. One view is that they may have succeeded in reducing registered unemployment, but at the cost of lowering regular employment. Similarly, when high persistent unemployment is erroneously fought using fiscal and monetary policies, this can lead to excess deficits and/or inflation, which in the end create the need for more drastic structural reforms. Furthermore, financial problems in other politically sensitive areas, such as pensions and health care, raise the social value of high employment rates. The recent drive for labour market reform in Germany stems from a more general crisis of the welfare state, with the recognition that the prereform state commitments in this area are unsustainable. Opposition to the removal of labour market rigidities weakened as it became clear that failing to reform the labour market would increase the need for reductions in pension and health insurance benefits that are more valuable than the gains from existing rigidities. Changes in the international economic environment may also raise the cost of labour market rigidities. Increasing openness reduces producers’ ability to transfer changes in their labour costs to prices. Since a faster pace of technical progress raises the need for labour turnover, it penalises those societies that impose a tax on turnover in the form of employment protection provisions. New technologies may increase the demand for skilled
9780230_547018_02_int.indd 3
3/19/2009 8:00:11 PM
4
Designing the European Model
workers and reduce the demand for unskilled workers, thus reducing the real wages and/or employment of the latter. The reform package called “Agenda 2010” recently adopted in Germany is a good illustration of these effects. The reform was pushed forward because it had become clear to the political elite of the country that the welfare state in Germany was no longer affordable and proved incompatible with the degree of wage flexibility required by globalisation. In addition to a large number of minor reforms, the package abolishes the second-tier unemployment benefit system, which used to pay roughly 60 percent of the worker’s last net wage until retirement age. While Agenda 2010 falls short of reforming the welfare system and the system of wage bargaining, where major reasons for the inflexibility of the German labour market can be sought, it goes much further than seemed possible only a short while ago. The German example is therefore a good illustration of how a severe crisis and a sense of urgency enhance reform possibilities and help to implement fundamental changes that are otherwise politically doomed.
3 Pay-setting Systems in Europe: Ongoing Developments and Possible Reforms The performance of the labour markets – as well as macroeconomic performance in general – is intimately associated with the functioning of pay setting. The third chapter of the book analyses different forms of organising pay-setting systems in Europe and comments on their effects on employment and welfare. It has become commonplace to relate the rise in Western European unemployment in the 1970s and 1980s to aggregate real wage rigidities. The reduction in unemployment in many of the smaller EU countries throughout the 1990s was associated with real wage moderation. But in some countries a badly functioning pay-setting system has contributed to persistent unemployment: in Germany, for instance, a crucial problem has been the compression of pay differentials between Eastern and Western Germany. Another important aspect of pay setting relates to the use of pay as an incentive to promote effort and labour productivity. Increased international competition and new forms of work organisation have substantially increased the importance of this last aspect of pay setting. A key feature of pay-setting arrangements that is common to most Western European countries is the high coverage of collective agreements, usually above 60 percent of employees. The main exception is the United Kingdom, where the overall coverage rate was estimated to be as low as 36 percent in 2001. In most Western European countries the main bargaining level is the sectoral one. But the general trend is toward enlarging the scope of local bargaining. Somewhat paradoxically, in some countries this development has occurred simultaneously with social pacts between top-level labour market organisations, sometimes also involving governments that have set norms
9780230_547018_02_int.indd 4
3/19/2009 8:00:11 PM
Introduction 5
for pay increases at the national level. In some EMU countries, such coordination efforts have been a means to restrain wage increases. The most far-reaching examples are Finland, Ireland and the Netherlands, but less ambitious schemes have also been adopted in Belgium, Greece and Spain. Most of the new EU member states in Eastern and Central Europe find themselves in an entirely different situation from what prevails in the EU-15 states. In most of the new member countries collective bargaining is of much less importance than in Western Europe. When collective bargaining does take place, the main level is the enterprise one. Estonia, the Czech Republic, Hungary, Latvia, Lithuania and Poland are all characterised by decentralised pay bargaining at the level of the enterprise, low unionisation and coverage of collective agreements. The forces behind the decentralisation of pay bargaining in the past decades are likely to continue operating also in the future: one should expect further decentralisation of pay setting in Western Europe. There are three possible scenarios. The first scenario is massive decentralisation, de-unionisation and reduction in the importance of collective bargaining, as has happened in the United Kingdom, New Zealand and Australia. In these countries, singleemployer bargaining has almost completely replaced multi-employer bargaining. The Anglo-Saxon model of pay setting has its advantages. The combination of enterprise-level bargaining and low unionisation/coverage of collective agreements is likely to produce aggregate real wage restraint. At the same time, it promotes relative wage flexibility and the use of incentive pay. But the radical change in pay-setting practices in the United Kingdom, New Zealand and Australia have only occurred following fundamental changes in the legal system. Without sweeping legal reforms, a development in this direction in continental Western European countries is unlikely. The second scenario is a slow and disorganised process of decentralisation, leading to a reduction in the importance of collective bargaining in Western Europe. This has occurred in Eastern Germany, where many firms have left the employers’ associations or violated sectoral collective agreements in order to reduce wage levels. However, because of the inertia of wage-bargaining institutions, a process of spontaneous, disorganised decentralisation is likely to be slow. In principle, decentralisation could promote real wage restraint, although the effects might not be large. It is also possible that a move to single-employer bargaining weakens the incentives for wage restraint. This risk is great if coverage of collective agreements and unionisation remain high. A third scenario is organised decentralisation, according to which higherlevel union confederations and employers’ associations choose voluntarily to leave more scope for local bargaining. This could involve larger freedom at the enterprise level to determine the margin for wage increases and its distribution among individual employees. Which pay-setting system should the
9780230_547018_02_int.indd 5
3/19/2009 8:00:11 PM
6
Designing the European Model
EU countries opt for? While there is no general answer to this question, the chapter on pay-setting systems gives are some basic guidelines for reform. First, it is pointed out that pay-setting systems are very slow to change: it takes a long time or extraordinary circumstances to achieve fundamental changes. Therefore, specific recommendations for individual countries would have to take into account the existing bargaining systems, which may differ considerably among countries for historical reasons. Furthermore, there are good reasons for keeping existing bargaining systems in those current EU member states that have achieved wage restraint through formal or informal co-ordination of bargaining at the multi-sector level. Such arrangements have worked well especially among smaller countries. But aggregate real wage moderation should be combined with relative wage flexibility. One way to achieve this could be to publicise a “corridor” for wage increases, rather than a single figure, when wage norms are formulated. Alternatively, a co-ordinated agreement on a “normal wage increase” could be reached with the understanding that “above-normal” wage raises should only be granted in areas with labour shortages. The chapter also recommends allowing more relative wage flexibility and applies this in particular to Germany, where wage bargaining mostly takes place at the industry level. In Germany, there is a strong need for diversity in wages between the western and the eastern regions, as well as for greater opportunities to set wages at the enterprise level that are lower than the norms agreed upon at the industry level. If the stagnation problem of Germany is to be solved, labour market reform must be extended to paysetting practices as well, which so far has not taken place. Finally, the chapter highlights that the acceding countries, with industrial relations of an Anglo-Saxon type, should keep their present systems. These countries are likely to face strong pressures from Western European trade unions, and possibly also from EU institutions, to change their industrial relations systems so as to conform better to “EU standards”. Such pressures should be resisted. The existing combination of enterprise-level bargaining and limited importance of collective agreements is likely to produce better macroeconomic outcomes than a move to industry-level bargaining of the Western European (German) type. What will be the outcome of the present trend towards greater importance of local-level pay setting in Western European countries? For political and economic reasons, it may turn out to be impossible to combine aggregate real wage restraint with relative-wage flexibility within the current bargaining systems. If this is the case, one cannot, in the long run, rule out a development towards an Anglo-Saxon system in which a move to singleemployer collective bargaining is accompanied by massive de-unionisation and a decline in the importance of collective bargaining in general. But such a development would probably occur only as a consequence of a long period of deep economic crisis, followed by radical reform of basic labour
9780230_547018_02_int.indd 6
3/19/2009 8:00:11 PM
Introduction 7
legislation. This point is underlined by the failure to include even modest changes of the pay-setting system in the recently adopted labour market reforms in Germany.
4
Outsourcing
Since the fall of the Iron Curtain and the integration of China into the world trading system, international trade in goods and services has increased significantly. This is a consequence of the large differences in factor endowments and hence in relative prices between the earlier OECD countries and the countries that have been opened up to international trade. Trade in intermediate products has developed particularly rapidly, due to outsourcing activities of firms that have tried to make use of the huge wage differences between the formerly separated parts of the world. Major improvements in information and communication technologies in the 1990s are a second reason for the increase in outsourcing. The new international production patterns have caused the domestic value added per unit of output, the socalled production depth, to decline in many sectors. This trend towards a reduction in domestic production depth has been particularly strong in the manufacturing sector. Its importance has declined significantly in recent years in most Western European countries, where it now usually accounts for about one quarter of total production and one fifth of aggregate employment. Furthermore, the share of manufacturing value added in GDP has been declining in recent years. A closer look at the input-output tables shows that this trend of de-industrialisation is related to outsourcing activities of domestic firms. Parts of the production process have been moved to low-wage countries. In particular, the new members of the European Union in Eastern Europe are at the receiving end of outsourcing done by Western European firms. Outsourcing is not limited only to manufacturing. In Germany, for example, it has been shown that the same phenomenon applies to the export sector as a whole. From 1991 to 2002, an additional unit of real exports induced, on average, a 55 percent increase in intermediate imports. Only 45 percent of the increase in exports implied additional value added in Germany, a phenomenon that has been caricatured as the “bazaar effect”. Nevertheless, export-induced value added grew relative to GDP, which is a natural implication of increased specialisation. In principle, outsourcing activities can lead to gains from trade for all countries involved. The low-wage countries of Eastern Europe and Asia find new and profitable employment activities for their large labour forces and are able to increase their wages. The high-wage countries of the West are able to withdraw part of their endowments of labour and capital from labour-intensive sectors and use them more productively in the service and high-tech sectors, where they may have comparative advantages.
9780230_547018_02_int.indd 7
3/19/2009 8:00:11 PM
8
Designing the European Model
Outsourcing is a special form of international trade that can be expected to boost world GDP and world welfare, because it allows countries to specialise on the basis of their comparative advantages. However, for the gains from trade to occur, it is essential that the domestic factor markets in the West are flexible enough to allow for the necessary factor migration between shrinking and expanding sectors. While capital markets in Europe do seem largely to meet this requirement, labour markets are rather rigid. For one thing, national job protection measures prevent workers from moving easily between sectors. For another, collective wage agreements and welfare state provisions that ensure high replacement payments for the non-employed prevent the necessary wage flexibility. Gains from trade go hand in hand with a tendency towards factor price equalisation. In particular, the specialisation in more capital-intensive production requires lower wages of the less-skilled workers relative to wages of skilled workers in order to prevent unemployment. If wages are rigid, this process cannot take place. The sectors where the West has a comparative disadvantage shrink too quickly, setting more labour free than is useful, and the growing sectors where there is a comparative advantage do not create enough additional jobs. The result is a growing level of unemployment. The EEAG sees strong signs for such a deficiency of the adjustment process in some European countries. Thus, the group advocates policies to make the labour market more flexible. The necessary measures range from more limited job protection policies via opening clauses for collective wage agreements toward policies of activating social aid that reduce labour costs and give further incentives to work, changing thereby the role of the welfare state from a competitor to a partner of private enterprises.
5
Longer Working Hours – The Beginning of a New Trend?
Both hours worked per capita and hours worked per employee are low in several continental European countries, such as Germany, France, Belgium and the Netherlands, as compared to the United States. This accounts for a large part of the income difference between Western Europe and the United States. The low working hours in Europe reflect to a large extent low standard working hours for full-time employees. Recent company-level deals in Germany on longer working time may represent a reversal of the earlier trend towards shorter working time that could spread also to other Western European countries with low working hours. It seems that especially the French discussion has been much affected by working time developments in Germany, but also employer demands in Belgium and the Netherlands have been influenced. The agreements in Germany have implied longer working hours without pay compensation. The deals can be seen as cost-cutting measures made necessary by both increased international competition in product markets and credible
9780230_547018_02_int.indd 8
3/19/2009 8:00:11 PM
Introduction 9
employer threats to outsource jobs abroad associated with ongoing globalisation in general and EU enlargement in particular. One way of thinking about the deals on longer working time is as a convenient way of reducing hourly wages without reducing the pay per employee. It is natural that such agreements take place at the decentralised level of firms exposed to harsh competitive pressures rather than at more centralised levels of bargaining. The deals on longer working time can also be viewed as a labour supply response in collective bargaining to a reduction in the hourly wage that is required as a response to the forces of globalisation. The desired working time on the part of trade unions may respond differently to a wage change than the desired working time on the part of employees when acting on their own. The EEAG argues that it is likely to be in the interest of unions (workers acting collectively) to accept increases in working time when real wages per unit of time have to fall. When analysing the effects of longer working hours, it is essential to distinguish between the short run and the long run. In the short run, longer working hours at unchanged pay, as in Germany, will by definition prevent job losses in firms where there is an acute risk of outsourcing production because costs are too high. Also in firms where this is not the case, such agreements will have positive employment effects, provided that the longer working time of employees makes possible to operate the capital stock for longer hours. Indeed, such an increase in capital utilisation is one of the major advantages of longer working hours. The capital utilisation effect makes an immediate output increase possible without the need to restructure the production process. In other firms where the operating time of capital does not increase with working time (for example, because shift work or overlapping working times of employees have detached the operating time of capital from working time of individual workers) or where it takes time to restructure the production process, the employment effects are likely to be negative in the short run. Lower wages per hour will, however, make it profitable even for these firms to expand output. In the long run, the hourly real wage level in an open economy must be such that capital earns the same rate of return as abroad. It follows that the long-run feasible hourly wage depends on the world market rate of return to capital. A critical factor for the long-run employment effects of an economywide lengthening of working hours is therefore how wage-setting incentives are affected. If longer working time creates stronger incentives for wage moderation, lower unemployment is needed in equilibrium to discourage wages from rising above the feasible level. Although neither theoretical nor empirical research gives unambiguous conclusions, there is a presumption that longer working hours would contribute to wage moderation. If so, one should expect positive employment effects in the long run from longer working time. This can add to the positive long-run output effects of an increase in working time that can arise already at an unchanged employment level.
9780230_547018_02_int.indd 9
3/19/2009 8:00:11 PM
10 Designing the European Model
Part II: Macroeconomic Issues The first chapter in every EEAG report provides an economic outlook on the macroeconomic environment of the member countries of the EU. This outlook is supported by a business survey of the Ifo Institute, which illustrates the development of the current economic conditions as well as the expectations about the future development. Aside from this survey, and the analysis of main macroeconomic indicators, most EEAG reports have dealt in further depth with topics related to specific macroeconomic analysis. Issues of growth and stabilisation policy have been reoccurring, but also special topics, like the option to introduce the Euro as a new currency in Eastern Europe, as well as pension reforms and the problem of global imbalances in international aggregate accounting, have been analysed. The second part of the book consists of chapters dealing with these types of macroeconomic problems.
6 Fiscal Policy and Macroeconomic Stabilisation in the Euro Area: Possible Reforms of the Stability and Growth Pact and National Decision-Making Processes The recent budgetary problems of some member states in the EU have focused attention on fiscal policy and the fiscal rules in the EU. A key issue is the need to combine long-run sustainability of fiscal policy with shortrun flexibility, because fiscal policy is the only remaining stabilisation instrument in the case of country-specific cyclical developments. In the first chapter of this section, it is argued that fiscal policy should play a larger role as a stabilisation tool than envisaged by the conventional wisdom that has prevailed in recent years. The group highlights that the problems of using fiscal policy in this way are not due to technical ineffectiveness but to problems of political economy. The chapter emphasises that there is a continued need for fiscal rules at the EU level to ensure fiscal discipline. In view of the future strains on government budgets arising from ageing populations, the present “close to balance or in surplus” budget targets for the medium term should not be relaxed, although the targets should be set explicitly in cyclically adjusted terms. It would be unwise to introduce a golden rule, according to which government investment can be financed through borrowing. The underlying rationale for a golden rule is that public projects are expected to generate a flow of tax revenues as high as the interest payment on the additional debt incurred to finance them. There is, however, no reason for this to be true: many public projects are desirable for reasons that are independent of tax revenue considerations. Moreover, the classification of expenditure among different categories is arbitrary. Allowing budget flexibility via a golden rule is likely to cause massive re-classification to take advantage of the rule.
9780230_547018_02_int.indd 10
3/19/2009 8:00:11 PM
Introduction 11
This is not to deny that there may be sound reasons to allow for larger deficit financing of public investment – such as efficiency of the tax regime or intergenerational fairness, as also future generations will benefit from public capital. But experience shows that these good reasons seldom are primary concerns in the actual budget processes. Recent amendments from the European Commission from July 2005 aim at increasing the flexibility of the EU fiscal rules through changes in the interpretation of the Stability and Growth Pact but without revisions of the Maastricht Treaty. The agreed changes involve more discretionary decisions on the fiscal goals. The idea is to allow temporary deviations from the medium-term budget objective of “close to balance or in surplus” on a caseby-case basis if they can be justified in terms of growth-enhancing expenditure increases or tax cuts, or as a consequence of structural reform.1 A loosening of medium-term budget objectives without doing anything about the maximum deficit ceiling of three percent of GDP increases the risk that this ceiling will be breached, which is likely to cause more conflicts among member states. Also, the more complicated the rules become and the more discretionary judgements are involved, the greater is the danger that the credibility of the fiscal rules is undermined. Instead, EEAG suggests that there is a strong case for more fundamental reforms of the fiscal rules involving Treaty. These changes should focus on the excessive deficit procedure and the deficit ceiling, as they form the backbone of the rules. A simple and transparent reform would be to let the deficit ceiling depend explicitly on the debt level of the country: countries with low debt (less than 55 percent of GDP according to the EEAG proposal) should be allowed to run larger budget deficits than three percent of GDP. The lower the debt–GDP rating the higher the maximum deficit for these countries should be. This would serve both to give low-debt countries greater scope for stabilisation policy in recessions and to enhance the incentives for long-run fiscal discipline, preventing pro-cyclical fiscal policies in booms. Changes in the fiscal rules must not, however, accommodate the current budgetary problems of some countries. This would ruin the future credibility of clear fiscal rules at the EU level. If France, Germany, Italy or Portugal were to breach the three percent deficit ceiling for more than a single year, sanctions must be imposed, as a natural consequence of earlier insufficient fiscal retrenchment, in the common interest of establishing credibility for the rules. The present fiscal policy framework at the EU level suffers from the fundamental problem that the ultimate decisions on excessive deficits are political. The threat of sanctions has low credibility, as governments are likely to try to avoid political conflicts with each other. This is an argument for transferring decisions on deposits and fines from the political level of the Council to the judicial level of the European Court of Justice.
9780230_547018_02_int.indd 11
3/19/2009 8:00:12 PM
12
Designing the European Model
Recent events have shown that there are limits to how much fiscal rules at the EU level can achieve on their own. It would be impossible to uphold these rules if governments repeatedly came into conflict with them. This consideration suggests that one should rely much more on national institutions that are conducive to both long-run fiscal discipline and effective short-run stabilisation policy. One possibility would be to require the member states to adopt national laws on fiscal policy that set well-defined longrun sustainability goals, but also outline clear principles for the use of fiscal policy as a stabilisation instrument. In this respect, economists have recently begun to discuss whether there are lessons for fiscal policy to be learnt from the recent development of theory and institutions of monetary policy. A parallel could be drawn between delegation of monetary policy to independent central banks, and delegation of decisions about fiscal stabilisation policy to an independent fiscal policy committee. The underlying idea is to separate decisions aimed at stabilisation from other aspects of fiscal policy concerning distribution and social efficiency. Such separation would reduce decision lags as well as politicoeconomic risks of pursuing pro-cyclical policies and deficit bias. At the same time, it could help the government to define more clearly the political goals of alternative policy measures. Such a development has taken place in other areas of economic policy-making in addition to monetary policy: examples include competition policy as well as market regulation and supervision. The idea of delegation of fiscal policy stabilisation decisions may be unfamiliar to many people, and is not on the current political agenda. There is, however, a case for starting to think about the possibility of such a reform, and exploring the extent to which it would be compatible with generally accepted principles of democratic governance. In line with the principle of subsidiarity, national delegation could be seen as an alternative to the recent proposals of the European Commission, according to which it should be given greater discretionary powers in assessing fiscal policies of member states. One idea would be for member states to establish an independent fiscal policy committee at the national level. A politically realistic way to move in this direction in the next few years is to set up independent fiscal policy committees at the national level that play an advisory role. 2 Governments could be required to seek the advice of these committees before making their budget decisions and to use the committees’ estimates of cyclical conditions, government expenditures and tax revenues as a basis for budget calculations. The task of these committees could be to propose how much the actual budget balance in a given year should deviate from the cyclicallyadjusted budget balance and to make recommendations on specific tax or expenditure changes with the aim of stabilising the business cycle. The general goal of such reform would be to lessen many of the problems that now hamper the use of fiscal policy as an effective stabilisation tool, such as
9780230_547018_02_int.indd 12
3/19/2009 8:00:12 PM
Introduction 13
long-decision lags, deficit bias, irreversibility of decisions, and confounding of objectives.
7 Acceding Countries: The Road to the Euro The ten countries that joined the EU in 2004 are supposed to also join EMU in the next few years – the third and last phase of accession. Slovenia entered EMU in the beginning of 2007, and several other countries are likely to submit their applications for EMU entry in the coming years. Participation in EMU is conditional on satisfying the convergence criteria established in the Maastricht Treaty. When should these countries adopt the new currency? What are the challenges to policymakers in the period between EU accession and the adoption of the euro? There are clear benefits from participating in EMU. The traditional argument is the credibility of low inflation, which applies to the new member countries as it did to southern European countries in the 1990s. A common currency eliminates currency risk and therefore drastically reduces interest rate differentials. In addition, a common currency is likely to increase trade with other EU countries. In this respect, adopting the euro is equivalent to a drop in transaction costs in cross-border exchanges of goods and services within the EU economic area. By reducing the stock of external debt denominated in foreign currency, adopting the euro will also substantially reduce vulnerability to currency and financial instability (although in principle EMU countries could still issue large stocks of dollar-denominated debt). The main and well-known disadvantage of participation in a monetary union is the loss of national monetary policy as an instrument of macroeconomic stabilisation and of the exchange rate as an adjustment mechanism. Whether and under what circumstances the adoption of the euro is a net economic benefit is the subject of an ongoing academic debate and political discussion in present EU countries that are not members of EMU. However, from a political point of view it appears certain that the accession countries will ultimately join EMU, so the relevant policy issue is one of timing. The road to EMU may be quite difficult. Policy choices on the timing of EMU participation directly impinge on the acceding countries’ ability to use monetary policy to stabilise their economies in the next few years and to build an economic environment that favours high rates of investment and growth, economic integration and financial stability. Fiscal and monetary authorities in the new member countries currently operate in a regime of high capital mobility. This is the result of a relatively rapid process of liberalisation and deregulation implemented in the last few years. It is too early to say whether or not the financial and legal systems of these countries can weather volatile capital movements. However, it would be naïve to hope for the better and envision years without large (global or region-specific) shocks.
9780230_547018_02_int.indd 13
3/19/2009 8:00:12 PM
14
Designing the European Model
The chapter points at the intrinsic financial and currency fragility of fastgrowing emerging markets in a world of liberalised capital flows. Based on the experience of the 1990s, emerging countries may be exposed to highly volatile capital flows, which can be a formidable challenge to macroeconomic stability. In boom periods, large inflows of short-term capital lead to domestic overheating and high rates of domestic credit expansion, causing excessive risk taking. Since debt is usually denominated in foreign currency, these inflows also expose domestic institutions and individuals to severe currency risk. In bust periods, currency devaluation worsens the balance sheets of banks and industrial firms. EU membership has increased short-term speculative capital inflows into the acceding countries. These flows may also respond to moral hazard distortions at both domestic and euro-area levels. Vulnerability to crises and contagion emphasises the need for building well-established mechanisms at the EU level to deal with such contingencies. Structural imbalances in the new member countries may cause acute problems. Deteriorating fiscal conditions could constrain the use of budget policies for stabilisation purposes. Stabilisation is likely to fall disproportionately on monetary and financial authorities, both from a macro perspective and from a financial stability perspective. In such an environment, mandatory adoption of a regime of limited exchange rate flexibility (ERM II) for two years before entering EMU is controversial. Overall, there is no single strategy that may be recommended to all acceding countries as regards macroeconomic stabilisation on the road to the euro. Arguments in favour of adopting the euro as early as possible include smaller financial risk due to the elimination of currency mismatch in the balance sheets of banks and firms (which implies the risk of a self-fulfilling run on the country’s debt), interest rate convergence (with the associated gains in terms of the interest bill for the government as well as investment financing by firms) and overall gains in monetary credibility. Arguments for a slower pace to the euro include the need to remove financial distortions creating moral hazard and therefore raising the country’s default risk, easier relative-price adjustment without the need of costly nominal wage and price adjustments, and the need to make fiscal and financial policy sustainable and compatible with a fixed exchange rate before participation in the EMU. Several recommendations can, however, be made. First, countries that are already able to sustain hard pegs should be helped to achieve a smooth and fast transition to the euro. In this set of countries, mainly small ones, priority should be given to institutional reforms and building a policy framework consistent with participation in the euro area without suffering from major macroeconomic imbalance. Second, delaying participation in ERM II is a realistic option for countries that are currently unable to sustain hard pegs and have large domestic
9780230_547018_02_int.indd 14
3/19/2009 8:00:12 PM
Introduction 15
imbalances. Here the policy priority is achieving a sustainable fiscal stance and stabilising inflation at the correct relative prices. This task requires both institutional and policy reform. Third, for both groups of countries, the convergence criteria in terms of inflation, interest rates, debt and deficits provide desirable targets to guide policy and should not be relaxed. Though they are not first-best targets, these convergence criteria should be judged relative to existing distortions that could derail the stabilisation efforts. Fourth, ERM II allows for large fluctuation bands around the exchange rate parity. Once in the ERM II, a country should be able to use the exchange rate flexibility implied by such an arrangement: exchange rate stability should not be mechanically assessed with reference to much narrower bands. Fluctuations in the exchange rate in response to domestic and foreign shocks are not necessarily an indicator of tension in the foreign exchange market, but can be part of an efficient adjustment process. If the dollar suffers further depreciation, it would be reasonable to expect exchange rate fluctuations within ERM II. To declare that candidate countries will be accepted in the euro area only if they can peg to the euro within narrow bands may raise the possibility of speculative attacks driven by self-fulfilling prophecies. During the transition to the euro, strict domestic stabilisation with some exchange rate flexibility is better than exchange rate based stabilisation with very limited flexibility.
8 Economic Growth in the European Union Economic growth has been sluggish in many EU countries. Up to the 1990s, levels of GDP per capita in Western European countries were catching up with that of the US, but these tendencies were dramatically reversed in the 1990s. In particular, France, Germany and Italy started to fall further behind the US. The European growth problems have led to major political discussions within the EU. The 2000 Lisbon strategy for growth and employment is an expression of political concerns about low growth. This chapter analyses the reasons behind the varying growth performance of Western EU countries. The first observation is that slow growth is not a universal phenomenon among the old EU countries. Some countries – notably Ireland, Finland, Greece, UK, Spain and Sweden – have performed well over the last decade. Furthermore, one is beginning to see “growth miracles” in several new EU member countries. Second, a process of convergence in per capita incomes in the EU is taking place. This process is largely driven by the convergence between the EU-15 and the new member countries, that is, living standards in the new EU countries appear to be catching up with the old EU members in a long-term perspective. Determinants of economic growth are analysed by decomposing GDP growth into the contributions from growth of labour input, IT capital
9780230_547018_02_int.indd 15
3/19/2009 8:00:12 PM
16
Designing the European Model
input, non-IT capital input and technological progress (total factor productivity). Growth accounting reveals that the unsuccessful countries, France, Germany and Italy, have been growing mostly through traditional capital accumulation and to a much smaller extent through general technological progress. Labour input often played a substantial negative role, particularly in Germany. In contrast, there have been different roads to prosperity in the successful countries. In one group, consisting of Ireland, Finland, the UK and Sweden, there has been a large increase in the contribution by IT capital growth, though all production factors have made a positive contribution in these countries, including labour input for most episodes. In addition, relatively rapid IT capital growth has been coupled with relatively high total factor productivity growth in these countries. The best performer, Ireland, has had rapid growth in all factor inputs. Spain and Greece make up a second group of success cases, which have primarily grown through conventional capital accumulation and labour input growth. There are large variations among countries in the determinants of growth in capital and labour inputs and in factors that influence technological progress. Finland, the UK and Sweden had higher shares of IT capital, relative to other capital, earlier, so the recent fast accumulation of IT capital has for this reason resulted in larger contributions to growth. These countries are also at the top in terms of indicators of IT diffusion. Determinants of technological progress are likely to have been quite diverse, as technological progress is influenced by a number of factors such as education and innovativeness of the economy. Finland and Sweden had the highest levels of education spending (relative to GDP) among EU countries, but there appear to be no systematic relationships between this factor and growth for EU countries. The amount of regulation is one determinant of the degree of competition among firms, which in turn influences innovativeness. In many, though not all, cases the successful countries have done well in terms of indicators of deregulation, venture financing, and R&D spending. The analysis put forward in this chapter leads to several policy conclusions. First, the EEAG recommends that the Lisbon strategy should be modified. The Lisbon strategy argues for the creation of a uniform model of a high-tech information society for the EU, whereas the European experiences suggest that there are different routes to success. Instead, the EU should allow for a flexible strategy for growth, in which there is scope for high-tech driven growth as well as growth based on more traditional means of capital accumulation, increased labour input and imitative adoption of new technologies from the leaders. One key element in growth policy is improvement of the educational systems. This should be done at both the national and EU levels. Education influences growth through the accumulation of human capital, and there are also important complementarities between education and the ease of
9780230_547018_02_int.indd 16
3/19/2009 8:00:12 PM
Introduction 17
adoption of innovations and new technologies. An important question concerns the level of education at which improvements should be focused. Countries that are close to the frontier should specifically focus on improving the tertiary education system, as high-technology innovations require more advanced skills than lower-level innovations. The latter are often process improvements and rely on imitative adoption of known technologies. While the US does not stand out in the quality of secondary education, it is well ahead of EU countries in university education, which is likely to matter the most for economic growth of the most advanced countries. The best universities in the US compete strongly with each other for the best graduate students and researchers. In European countries, the university system is largely not exposed to strong competition, though the UK with its national research and teaching quality audits is partly an exception. A third policy conclusion concerns the potential to increase labour input to enhance economic growth. In most EU countries, labour input has not grown much, and in some countries labour input growth was even negative for some periods. Labour input can be raised through labour market reforms such as reduced unemployment benefits, increased employment tax credits, and lower marginal tax rates on labour and pensions to provide incentives to a longer working life. Decentralised collective agreements that allow lengthening working hours (as in Germany) and reversals of earlier legislated working time reductions (for example in France) are other desirable measures. Another policy conclusion concerns the regulatory policies in the EU. Europe tends to have a relatively high level of regulations that limit competition by restricting entrepreneurial activities, entry and labour market adaptability, which in turn can suppress innovation and technological advancements. Growth effects of competition appear to depend on the distance of the industry from the technology frontier, so that increased competition yields the largest productivity gains in sectors that are far behind the frontier. Technology policy should focus on provision of opportunities for creation of new firms and industries and less on glorifying national champions. Improvements of venture capital financing and R&D continue to be important policy areas for the EU countries. There are big variations in the amount of venture capital investments in the EU, and Europe is lagging behind the US in this respect. Also, competition policies should focus more on facilitating entry of new firms to improve innovativeness of European economies. Reduction of trade barriers to competition and entry in the service sector is important, as exporters of services tend to be subjected to national regulations in both the country of origin and in the host country. Since the service sector makes up around 70 percent of both GDP and employment in the EU-15, lower trade barriers for services have potentially large growth effects. For this reason, it was important that the new EU Services Directive
9780230_547018_02_int.indd 17
3/19/2009 8:00:12 PM
18 Designing the European Model
that was agreed on in December 2006 was not watered down.3 A related issue is that the imposition of national pay conditions on posted workers from other EU member states prevents effective cross-border price competition. This limits the gains from trade in services to economies of scale, more effective organisation and greater product diversity. It also means that the old EU member states forsake the welfare gains that could come from allowing service providers from the new member states to compete effectively by compensating for lower productivity through lower wages. Such competition is a natural exploitation of different comparative advantages. It is not “unfair wage dumping”. Wage competition among countries in trade with services should be allowed in the same way as in trade with goods. Growth-enhancing policies for new EU member countries include facilitating technology transfer and improvement of productivity in industries that are mostly behind the high-technology frontier. Education policy and financing of new firms and innovations continue to be major items in the policy agenda for the new EU members.
9
Global Imbalances
The large and persistent current-account deficits run by the United States from the second half of the 1990s have generated widespread concerns about the sustainability of current macroeconomic imbalances at the global level and the risk of disorderly adjustment and slowdown in macroeconomic activity. This chapter reviews the current debate and discusses the implications of global adjustment for European macroeconomic developments and policy-making. Currently, large external deficits in the US are matched by large surpluses in Japan, Asian emerging markets, oil-producing countries and a few European countries. However, the euro area as a whole is close to external balance. The composition of external financing of the US deficit changed significantly after 2000 with a falling share of private capital inflows (accounting for 90 percent of total inflows in 1997–1999, but for only 40 percent in 2003–2004) and an increasing share of public inflows. A further dimension of current global imbalances concerns the high level of international reserves held in dollar assets. At the same time, there has been a strong expansion of cross-border holdings of financial instruments, which have doubled since 1990 from about 60 percent of world GDP to above 120 percent now. Though the US current account deficit is large in terms of US GDP, it is small relative to the stock of US foreign gross assets. The US typically borrows from international markets by issuing dollar-denominated assets but lends abroad mostly by acquiring equities and foreign-currency denominated bonds. Therefore, dollar depreciation leaves the dollar value of US liabilities unaffected but raises the dollar value of US assets and improves the US net foreign asset position.
9780230_547018_02_int.indd 18
3/19/2009 8:00:12 PM
Introduction 19
There are a number of views on the causes of current imbalances, with quite different implications for the need for corrective policy measures. A widespread view attributes the persistent US current account imbalances to low US national savings. Private savings in the US have been trending downward for some time and US public savings have also deteriorated markedly since 2000. Some studies suggest that the impact of fiscal consolidation in the US on external trade is limited in the short run, but greater fiscal discipline would certainly help reduce imbalances in a longer-term perspective. A second view of the US external deficits argues that they are essentially driven by expectation of high future growth. This view has two important policy implications. First, it is not appropriate to talk about “imbalances”, as trade flows are in fact balanced in an intertemporal perspective. Second, significant dollar depreciation in real terms may not be required for some time and should therefore not be expected. However, current expectations about high US growth in the future may be too optimistic. If and when expectations are revised downwards, restoring US external balance would then require a sharp correction of spending plans, possibly implying large movements in exchange rates and relative prices. A third view of US current account deficits argues that the deficits are a mirror image of a “saving glut” in the rest of the world. A variant of this view is that there is an “investment drought” outside the US. This view offers a potential explanation of the simultaneous occurrence of low real interest rates and low investment. According to this argument, one may expect interest rates to rise as soon as investment picks up again. A fourth view suggests that a desire for “exported growth” and a build-up of currency reserves in Asian emerging markets have substantially contributed to the current global imbalances. In particular, imbalances are due to China’s exchange rate policy and its strong influence on the policies of the other emerging markets in the region. China’s formal abandonment of the inflexible peg against the US dollar has not led to any significant appreciation of the renminbi so far. However, given the internal consequences of distorted relative prices, due to an artificially low exchange rate and the threat of protectionist measures by the US, one should expect some noticeable correction in the near future. Predictions of further sizeable depreciation of the dollar in real effective (multilateral) terms emphasise the need for a fall in the relative price of US non-tradables, which is tantamount to a reduction in US income relative to the rest of the world. According to some studies, the required real rate of depreciation of the dollar might be quite large, depending on several factors that ultimately affect the elasticity of substitution between traded and nontraded goods in the US and between US and foreign traded goods, as well as on the impact on the level of economic activity. Many studies suggest that adjustment could necessitate a protracted period of real dollar weakness.
9780230_547018_02_int.indd 19
3/19/2009 8:00:12 PM
20
Designing the European Model
According to the consensus view, the most important policy contribution to adjustment should come from a reduction in the US fiscal deficits. Without any fiscal rebalancing in the US, a reduction in Asian saving, possibly associated with a slowdown or reversal in reserve accumulation, increases the risks of financial strain in the global currency and asset markets. Looking at the adjustment of global imbalances from a “euro” viewpoint, there may or may not be further dollar depreciation vis-à-vis the euro. However, correcting the US current account deficit in any case requires an improvement in US net exports, and Europe is likely to experience a drop in external demand with negative effects on European growth. It is possible that the resolution of current imbalances will proceed relatively smoothly. However, it is also possible that the current build-up of imbalances will be followed by one of the “hard landing” scenarios. Suppose that there is a disorderly adjustment with strong relative price and exchange rate movements and financial turmoil across markets. In this scenario, it is highly plausible that European financial and non-financial firms would suffer from strong deterioration of their balance sheets and liquidity shortages. This scenario would call for European monetary and supervisory authorities to stress test their institutional framework. If the financial crisis is moderate, the euro system may be able to contain it. However, if the financial crisis is sufficiently severe, monetary authorities may face difficult tradeoffs between financial stability and price stability. Governments may then have to shoulder large fiscal costs to stave off a serious financial crisis. This would raise important issues regarding the distribution of fiscal costs across countries that policies to mitigate a serious crisis would entail. In this scenario, the relatively weak public finances in many European countries are an aggravating factor, as they would imply undue constraints on emergency financing in the case of a crisis. Increasing the fiscal room of manoeuvres in a possible future financial crisis adds a strong precautionary motive for stronger fiscal discipline now. Even if European monetary authorities were successful in fighting financial contagion and other undesired effects of liquidity shortages in the event of a worldwide financial crisis in the context of an unwinding of global imbalances, the euro area would still in such a situation face a severe aggregate demand problem. It would be difficult to deal with this problem under the current framework for monetary and fiscal policy. Perhaps the most important risk for Europe associated with global imbalances is to become exposed to a severe downturn without having access to effective policy instruments to stabilise the economy.
Part III: Institutions A third major part of the analyses by the EEAG is policy recommendations about possible changes in institutions, needed to enhance the efficiency
9780230_547018_02_int.indd 20
3/19/2009 8:00:12 PM
Introduction 21
and overall welfare in the European economy. In fact, Ifo Institute started with monitoring and assembling information about different institutional arrangements in the European Union already two years before the formation of the EEAG. This database, that has now accumulated over the past 8 years, is referred to as DICE – the Database for Institutional Comparisons in Europe – and has served as the origin, is complemented or extended by the chapters of the EEAG that we will present in this part of the book. Institutions subject to investigation are, for instance, those that govern the functioning of financial markets, education and competition among firms. We start by introducing a fundamental chapter that revisits the principle of subsidiarity, which is a central concept in the delineating the powers between the EU Commission and the national governments. It thus guides the determination the level of government responsible for taking action in the various different areas where policy actions are needed. After this chapter specific institutional issues are considered in the three other chapters in this part.
10 Rethinking Subsidiarity in the EU: Economic Principles A reconsideration of the EU policies and the concept of subsidiarity have been timely due to the recent enlargements and the proposed constitution for the enlarged EU. The challenges ahead require careful consideration of the division of responsibilities for decision making of public sector activities. Analysis of economic efficiency provides a useful guideline for assessing, which public sector tasks should be delegated to the competence of the EU and which tasks should be the responsibility of national governments of the member states. While there are reasons for using subsidiarity as the basic principle, in a number of tasks there are sound economic reasons for deviations from it. These exceptions must be analysed case by case. Maintenance and promotion of the single market is the most basic EU-level task. It involves not only the removal of obstacles to trade and economic integration but also activities, such as the design and implementation of an active competition policy, that facilitate the functioning of the single market. The EU involvement has both an internal and an external dimension. In fact, it should not be forgotten that regional free trade areas might lead to trade diversion rather than trade creation. To be consistent with its ultimate goal of promoting the welfare of European citizens, EU-level trade policy should be geared towards global free trade. A second reason for delegation of specific tasks to the EU-level of government arises from the existence of public goods, which have geographically widely dispersed benefits. Defence, foreign policy and internal security are public goods where common EU-level decision may be appropriate, though the forms of implementation could partly be national with the EU level having a co-ordinating capacity. Whether other public goods qualify for
9780230_547018_02_int.indd 21
3/19/2009 8:00:12 PM
22
Designing the European Model
centralised provision is controversial, as in most cases benefits tend to be more concentrated locally. A third reason for delegating public intervention to the EU level arises from the need to regulate economic activities that generate important spillovers or externalities across borders. This is the case for telecommunication networks, environmental concerns, aspects of standardisation and product quality, as well as the financial system. The significance of spillovers and externalities must be assessed case by case. If the externalities involve only a few neighbouring countries, the EU function could be limited to co-ordination. In the area of environmental and resource administration, the management of fishing rights can be an EU concern because it involves management of a common property resource. However, it is difficult to extend the same argument to agriculture as a whole. A country or region should decide on its own whether to subsidise agriculture for aesthetic or environmental reasons, and implement its policy at the local level. Reforms of the EU agricultural policy that relies significantly on national policies should stay clear of providing nationally administered subsidies to production or exports as a way to promote competitiveness of national producers. If agricultural support moves to national level, the EU has a potentially important role in ensuring a level playing field and in defining food safety standards. The current activities of the EU accord rather poorly with economic principles. Nearly half of the EU budget is devoted to agricultural subsidies and guarantees. Structural funds and operations are the second largest item in the EU budget. The remaining significant items in the EU budget consist of external action that is policies towards non-EU countries (for example, development aid and pre-accession strategy), international operations, research and technological development, and EU administration. While the EU budget is small in comparison to the budget of central government in federal states, the EU exerts great power through regulatory policies in different ways, including regulations, directives and decisions. The regulatory activity of the EU has grown significantly over the years. Agriculture and fishery stand out also in terms of the number of EU regulations: looking at five-year periods, about 40–50 percent of the total is in this area. In terms of EU regulations, matters concerning the single market and non-sectoral business relations (especially competition policy) are also significant. As discussed above, activities associated with agriculture are not natural EU-level tasks, with the possible exception of food safety. Agriculture and structural policies are largely redistributive in nature and as such they are not natural responsibilities of the EU-level government. Decentralisation according to subsidiarity is likely to lead to competition between national jurisdictions, which can be good or detrimental depending on the nature of the activity. In general, beneficial effects can be expected from a yardstick competition, as countries try to imitate successful
9780230_547018_02_int.indd 22
3/19/2009 8:00:12 PM
Introduction 23
neighbours. However, in the case of factors of production that are mobile across borders, tax competition is problematic because it tends to drive tax rates down to a level that equals the marginal cost of providing public infrastructure. So, with fiscal competition, in the long run taxes on mobile factors become similar to prices or user fees for public infrastructure. But this means that the tax base for generating revenue towards the general government budget is likely to erode with the passage of time. Note that the revenue from taxes on mobile factors may not even cover the cost of providing the infrastructure. This is because tax competition equates tax rates to the marginal costs of producing the infrastructure, but in the case of public goods marginal costs are typically below average production costs. In that case, tax competition would result in a race “below the bottom”, whereby infrastructure is under-priced and the immobile factors are forced to pay for the services enjoyed by the mobile ones. Unless the distortions from tax competition offset other distortions, such as the tendency of local and national government to spend and tax excessively for political-economy reasons, there are potentially large losses of welfare. To prevent such outcome, tax harmonisation on the EU level might be considered. However, mere tax rate harmonisation will create a strong incentive at the country level to compete with each other through the provision of infrastructure goods, possibly resulting in overprovision of such goods. This problem can be avoided if the EU ban on explicit subsidies is extended to indirect subsidies through the provision of under-priced infrastructure. In principle, the cost of infrastructure should be covered with taxes on the benefiting firms and agents alone. With deepened integration and increased mobility of capital and people, the welfare state will come under financial pressure. In a closed system redistributive taxation and the welfare state can be seen as insurance systems as they protect citizens who happen to experience unfavourable personal circumstances. With open borders, increasing factor mobility puts limits to this insurance activity since rich net-contributors to the welfare state of a country may be inclined to move to countries with a less-redistributive system, while poor people have the opposite incentive – to migrate to countries with a relatively more redistributive welfare state. This has and will continue to create problems: The migrants from Eastern and South-Eastern Europe who have come to Western Europe after the fall of the iron curtain, and will continue to come in the foreseeable future, exhibit a highly differential mobility among European countries. This differential mobility is likely to trigger off a sort of deterrence competition among these countries. One important source of difficulties is the adoption of the “residence principle” for migrant workers and employees in the EU, as regards the eligibility to social benefits and social security contributions. While people who migrate from one EU country to another for reasons other than work are excluded from the welfare system of the host country, people who migrate
9780230_547018_02_int.indd 23
3/19/2009 8:00:12 PM
24 Designing the European Model
in order to work are fully and immediately included. Full and immediate inclusion implies full participation in the national redistribution system. This creates an incentive to migrate above and beyond the economic incentive from wage and employment differences. Moving away from a “residence principle” towards a “home-country principle” to define benefits and responsibilities for the migrants can in principle reduce distortions. Partially delayed integration, in which migrants are immediately entitled to contribution-financed social benefits but are only gradually entitled to social benefits that are funded from general tax revenues, may provide a practical solution. Social standards in health, work, and elsewhere are another aspect of modern welfare state. The recent enlargements challenge these standards because of the differences across member states – especially between the EU-15 and the new EU members. Economic analysis suggests that rapid harmonisation of work-related social standards is detrimental, since it would enforce the same mix of pecuniary wages and social standards on virtually all countries, whereas a different mix may best suit local labour market conditions. Different countries are in very different stages of economic development and premature harmonisation of social standards would slow down the process of development. If instead countries are allowed to compete, these standards will rise in line with wages and living standards in the poorer EU countries. Instead of focusing on harmonisation, it will be important to provide free access to new markets to the accession countries. This is the best way to facilitate the development process. Redistribution among different EU countries raises difficult political issues and polarises opinions. Once again, it is important to take into account the major differences in the stages of economic development. These differences suggest that inter-jurisdictional competition could be beneficial, as in the case of social standards discussed above. Interpersonal and interregional redistribution is primarily a national responsibility. Deviating from this principle could involve huge welfare and efficiency losses in Europe. East Germany is a good example of the problems that may occur. The quick adoption of the West German welfare system in East Germany has had extremely adverse consequences, because the underdeveloped market economy of East Germany turned out to be unable to generate jobs that could compete with the generous replacement incomes provided by the welfare state. Mass unemployment and a very poor growth performance were the result with little improvement in sight.
11 Financial Architecture Alternative models for reforming financial architecture in Europe will have profound implications for the degree of financial market integration, competitiveness in the financial industry, and financial and monetary stability.
9780230_547018_02_int.indd 24
3/19/2009 8:00:13 PM
Introduction 25
Reform proposals should be assessed in terms of their contributions to the welfare of European citizens, including the price they will pay for financial and payment services, the range of opportunity for insurance and portfolio diversification, and the reliability and trust of the financial institutions in the area. The financial architecture in Europe is clearly in a process of deep change. In its present shape, there are at least three significant problems. First, there are areas in which the present financial architecture arrangements are not adequate for financial stability. For instance, in the event of a crisis, there is no clear chain of command among the institutions potentially involved in any intervention. How would the euro system react to the threat of a major disruption like the one ensuing from the possible bankruptcy of Long Term Capital Management (LTCM) in the US in 1998? Who in Europe would have the responsibility to organise a rescue of a large financial institution, as did the president of the Federal Reserve Bank of New York in the case of LTCM? A response based on improvised co-operation may not be enough – it may come too late. Moreover, there could be misaligned incentives for national supervisors dealing with transnational firms, leading to too little interventions, as they do not internalise cross-border spillovers from the crisis of such firms. Conversely, national authorities may have strong incentives to provide excessive help to national champions. This view is in contrast to the conclusions of Brouwer’s reports (of March 2000 and April 2001)4 in the EU, according to which all these potential issues can be satisfactorily addressed with just a little bit more co-operation among supervisors in the various member states. Second, the present arrangements hinder European financial market integration to a large extent. Legislation is slow, rigid, and lags behind market developments. Regulatory fragmentation prevents the emergence of liquid European markets (as arguably was the case in the failure of the London Stock Exchange and the Deutsche Börse to create iX). Protection of national champions and regulatory barriers avert the emergence of pan-European banks. Third, the present arrangements hinder the competitiveness of EU financial markets and institutions. There is considerable uncertainty about the normative and regulatory framework in Europe. Market fragmentation resulting from regulatory barriers slows down and distorts the emergence of cross-national firms that may be able to compete at international level. Until recently, the “official” view has been that this state of affairs is not worrisome because European banking and financial markets remain segmented.5 In a framework of segmented markets, all that is needed is more co-operation among different regulators and authorities. This view may clearly backfire, as it justifies a slow pace of reforms and policies that do not remove obstacles to integration. Ultimately, this may just be a way to endanger stability.
9780230_547018_02_int.indd 25
3/19/2009 8:00:13 PM
26 Designing the European Model
Many political-economy issues are at the heart of the problem, namely, the tension between economic integration and the lack of willingness to relinquish national political control. But while these political economy issues slow down the pace of regulatory and institutional innovations, there are important sources of systemic risk to which the European markets are exposed. The recent events have stressed the threat of terrorist action, and possible financial weakness associated with economic slow-downs. Some European banks are heavily exposed to emerging markets and to particular sectors, such as telecoms, which have recently experienced deep crises. The process of consolidation within countries has led to the creation of many “national champions”, which may create incentives for national authorities to provide excessive guarantees. At the same time, the expansion of crossborder activities may increase potential spillovers and externalities across countries, while creating incentives for under-provision of supervision and liquidity support by national authorities. The present approach to reforms is gradualist, based partially on the socalled “comitology”, consisting in delegation of powers to define rules to various committees. This approach has its limits, and may yield more costs than benefits in the long term. It may be preferable not to wait for a major crisis to strike in order to put the house in order. There is good reason to endorse in general the well-intentioned recommendations of the committees and groups seeking to remove the obstacles to European financial integration. Yet the question is whether a more ambitious approach would be more appropriate. In particular, what prevents the immediate setting of clear procedures for crisis lending and management with the European Central Bank at the centre? Why not put a crisis framework in place now, and confront the fiscal issues related to the possible costs of intervention? By the same token, a debate should be opened with a view towards evaluating the benefits of more centralised supervisory arrangements in banking, insurance and securities. In addition to the current decentralised regulatory competition framework, there are other long-run models that one could follow. In the first model, the ECB and ESCB might gain a larger role in supervision of banking, with the contemporaneous creation of separate specialised European-wide supervisors in securities and insurance. The second model consists of an integrated supervisor for banking, insurance and securities, a European Financial Supervision Authority (EFSA), whereas the ECB would have access to supervisory information in order to maintain systemic stability. Different models present different trade-offs between efficiency, accountability but also suitability to specific circumstances and features that may differentiate markets and financial institutions across regions. It may be important to note here that in neither of the two models above, supervision need be completely centralised at the European level. First, national supervisors will need to be involved in day-to-day operations.
9780230_547018_02_int.indd 26
3/19/2009 8:00:13 PM
Introduction 27
Second, national institutions could still have the supervision of entities that trade mostly within one national jurisdiction (under the home-country principle). The door should be left open in the Convention on the Future of Europe to the necessary institutional changes to implement more centralised regulation, perhaps along the lines of one of the models above. At the same time, the EU-wide competition policy in the banking sector should limit help to national champions (which are “too big to fail”), and remove obstacles to cross-border mergers. Domestic competition policy should also be reinforced, as to keep in check local market power. Reforms of the financial architecture are admittedly quite complicated, as technical aspects are strictly interwoven with legal and institutional aspects. Given the large interests at stake, the process of reform is the target of particularly strong lobbies, both private and public. It would be a great cost for society if the need to reconcile conflicting special interests resulted in a lower protection of European citizens against the many risks that an inefficient and vulnerable financial system entails.
12
Pensions and Children
Public pension systems in most countries are based on the pay-as-you-go principle, in which current contributions are used to pay the pensions of people in retirement. Under current conditions, most pay-as-you-go pension systems in Europe are not sustainable: the old-age dependency ratios are forecasted to grow from the current 0.2–0.3 range to as high as 0.4–0.68 pensioners per worker in 2050, which would eventually require a very large increase in tax rates, and/or a reduction in pensions. Reform is required and it should aim not only at fixing the budget problem but also at designing a more efficient pension system. A general slowdown in the growth of living standards associated with ageing is inevitable. Pension reform that entails a move to a partially funded system (in which workers make savings in personal accounts toward their future pensions) will not prevent the slowdown and cannot benefit all generations. However, such a move may help stimulate national savings and smooth the pension burden across generations. The pension crisis results from a lack of human capital. Partial funding means filling the human capital gap with real capital. It thus helps in mitigating the forthcoming provision crisis when the baby boomers begin to claim their pensions. Funded pension components may also increase the scope for individual flexibility by allowing people to choose their pension level and retirement age at an actuarially fair rate, and thus alleviate political conflicts associated with ageing. One should ensure that private pension funds have an appropriate risk structure. That includes limiting exposure to stock market fluctuations and
9780230_547018_02_int.indd 27
3/19/2009 8:00:13 PM
28
Designing the European Model
minimising the correlation between the financial risk of pension wealth and labour market risk. Thus, portfolios of pension funds should be adequately diversified, with a critical mass of risk-free assets and a very limited exposure to assets in the firm and sector in which the worker is employed. Simple legal rules should be designed to supervise and regulate fund management so as to minimise budget risks and social costs associated with financial instability and moral hazard in financial markets. These rules should also provide strong incentives to contain the managing costs of pension funds. Especially in the initial phase of a reform, these costs may levitate in a privatised system because of aggressive advertising by an excessive number of providers. The introduction of an individually based, privately managed, funded pillar of the pension system would allow for a great deal of individual flexibility, provided it satisfies these requirements. It would be a good idea for those European countries that have not already done so to complement the existing pay-as-you-go system with such a pillar. A number of other margins of manoeuvre also exist that would contribute to fixing the problem of sustainability of the pension system. To the extent that part of ageing is due to an increase in life expectancy and that people are healthier, it is perfectly natural to raise the retirement age, which has trended downwards for many years in most countries. Pre-retirement schemes that are meant to artificially reduce registered unemployment, while increasing the burden on pensions, should be avoided altogether. Structural reform in the labour market, although desirable in its own right, will also have a positive effect on pension finance by increasing employment, thus increasing the tax base for contributions. The fiscal system could be amended so as to reduce its distortionary impact on people’s decision to have children. When deciding on the number of children, people may ignore the fiscal benefits brought by children to society in the form of contributions to pensions and they may therefore have fewer children than is socially desirable. One could envisage reforms to address this issue. A partial indexation of pay-as-you-go pension claims on the number of children is one possibility. Additional self-financed mandatory funded pensions for those who have no or only few children could then supplement the pay-as-you-go pension for those with no or only few children. People who do not raise children have on average more funds to save for their old-age pension. Personal income taxation can also be differentiated according to the number of children and systems of child allowance be used to provide stronger incentives towards having children.
13
Prospects for Education Policy in Europe
Education is an important productive input into the wealth of a nation. It enhances individual productivity, which shows up in higher wages. The rate of secondary enrolment comes out as one of the significant determinants of
9780230_547018_02_int.indd 28
3/19/2009 8:00:13 PM
Introduction 29
differences in GDP per capita across countries. An educated workforce is also a valuable asset at times of rapid technological change, because educated workers are better at adopting new technologies. In most European countries, the public sector holds a quasi-monopoly on the provision of education. While government intervention may be justified on the grounds that education has social aspects and that parents’ decisions may not reflect their children’s best interest, it is not clear that direct provision is the adequate form of government intervention. One may consider a more decentralised approach that would contain costs and allow for greater diversity of individual choices. In many countries, primary and secondary educational systems are under pressure. On the one hand, the costs of education are soaring as both enrolment rates and the length of studies trend upward, while the cost per pupil grows as fast as GDP per capita. On the other hand, there is a perception that standards and achievements are going down. Some argue that in order to solve these issues, one should spend more resources on facilities, hire more teachers to reduce class size, and perhaps employ more staff to take care of discipline and other non-curricular aspects. Others insist that educational systems can be made a lot more efficient by relying on competition and free parental choice. What does the evidence say? The EEAG points out in this chapter that there are large disparities between countries in terms of achievements in reading, mathematics and science. These disparities occur among countries that are similar in economic and demographic terms. Therefore, the way schools are organised seems to matter a lot. Furthermore, the amount of resources devoted to education does not seem to have a large impact. In a cross section of countries, it only has a small impact on achievements; the US spends a large amount per student, but does worse than the Slovak Republic that spends only little. Econometric studies at the individual level suggest that traditional recipes based on increased spending fail. For example, there is hardly any evidence that reducing class size has any impact on achievement. These findings are confirmed by event studies such as those of the unsuccessful French “Zone d’Education Prioritaire” experience. On the other hand, a growing body of empirical studies that compare similar groups of pupils exposed to different policies suggests that enhancing competition between schools has positive effects on achievements. Competitive mechanisms re-allocate resources from the worse to the best schools by allowing parents to choose and by adjusting school resources so that the successful schools can grow to accommodate increased demand. These mechanisms can take different forms: they can rely on the private sector to different degrees and involve different compensation mechanisms in order to offset potential unwanted effects on the distribution of income. For example, vouchers of some amount can be given to attend private schools.
9780230_547018_02_int.indd 29
3/19/2009 8:00:13 PM
30 Designing the European Model
The amount of vouchers can be adjusted to reflect distributional concerns. It has been shown that such schemes also benefit pupils who continue to attend public schools, because these are disciplined by competition from private schools. Hence, even students that are too poor to attend a private school, despite the voucher, indirectly benefit from school competition. But one can also think of other mechanisms where parental choice is increased and management is decentralised to the school level, but where there is less reliance on monetary rewards and smaller distributional effects. The organisation of public schools has a large impact on achievements. Mere increases in spending, in particular in the form of smaller classes, seem to be an inefficient way of raising achievements. In contrast, substantial improvements can be obtained if one fosters competition, both among students to get into the good schools and among schools to attract the good students. The available evidence suggests that while raising performance, such policies would not be particularly “unfair” or “non-egalitarian” relative to current practices.
14 Mergers and Competition Policy in Europe Merger activity is gathering pace in Europe. 2005 has seen large-value mergers or acquisitions such as Italy’s Unicredito of Germany’s HVB in the banking industry and France’s Pernod Ricard of the UK’s Allied Domecq in the food and drink sector. The pace of activity in utilities has been especially hectic and France’s Suez has acquired Belgium’s Electrabel, France Telecom has bought Spain’s Amena and Telefónica (Spain) has launched a bid for O2 (UK). Within Spain, Gas Natural has also announced its intention to take over Endesa in the largest operation of the year. At the same time, private equity firms (mostly British and American) are buying up firms, in particular conglomerates, with a view to restructure them and sell them for a profit. Not so long ago, mergers were basically an Anglo-Saxon phenomenon. However, the end of the millennium merger wave was driven, at least in terms of cross-border operations that have been gaining weight in the total, by activity in the EU-15. This reflects the long-term effects of market integration in Europe. But broader trends in the world economy are also important – the revolution in information technology, the widening of markets, the strength of corporate profits and the availability of cheap credit. Globalisation, especially in the form of competition from emerging economies like China and India, has induced restructuring and redeployment to increase productivity, and mergers are an integral part of these processes. Mergers raise a host of public policy issues. It is not clear that mergers create value for shareholders and consumers. A consolidation wave poses a threat to competition, which is the main driver of economic efficiency and productivity growth. The preservation of
9780230_547018_02_int.indd 30
3/19/2009 8:00:13 PM
Introduction 31
competition in different markets is of utmost importance. Domestic mergers are in general more threatening to competition than cross-border ones. It may be agreed that globalisation lessens the need for merger control, but it is important to establish that European merger control is up to the task of ensuring that the merger wave is good not only for investment bankers but also for consumers. A related issue is that many European governments have a protectionist instinct and view with suspicion foreign takeovers of their national champions or of firms that are considered to be in strategic sectors. Banking and utilities are often viewed as examples of such sectors. France and Italy tend to protect their firms, as shown by, for example, the discussion over whether French Danone could be taken over by PepsiCo, and the obstacles put by the former governor of the Bank of Italy to the foreign takeover of Antonveneta and BNL. France has issued a list of strategic sectors where national interests are to be protected. Despite this, the trend towards cross-border mergers seems robust. The public policy question is whether ownership matters and whether Europe needs either national or European champions. Globalisation is associated with technological change, with decreases in trade and transport costs of goods, capital, people and information, and with liberalisation and market integration that simultaneously enlarge the market and increase competitive pressure. In many sectors, the number of firms will have to be reduced in an integrated or enlarged market to reap economies of scale. At the same time, a sufficient level of competition is needed for innovation, and the timely termination of bad projects drives productivity growth. Furthermore, domestic competition is a key to international success and competitiveness, whereas fostering national champions may be self-defeating. The policy challenge is how to allow the needed restructuring and potential increase in firm size in some sectors, while at the same time protecting competition. The EEAG’s first conclusion is that a vigorous competition policy is needed, but care must be taken not to try to enforce low concentration in natural oligopoly industries where only a limited number of firms can survive. Furthermore, merger control should take into account the need of a larger firm size in several industries and the potential dynamic efficiencies (for example innovation) generated by mergers. The second conclusion is that artificial obstacles to hostile and crossborder mergers should be removed in Europe. Hostile takeovers are a sign of health of the market for corporate control. Cross-border mergers should proceed without regulatory obstacles as they may keep in check the increase in domestic concentration. It is acknowledged in this chapter, that ownership is not neutral, in particular, in some industries like banking where relationships are important, but on balance this is insufficient justification for protectionism. European as well as national competition policy must play a major role in keeping markets open.
9780230_547018_02_int.indd 31
3/19/2009 8:00:13 PM
32 Designing the European Model
The third conclusion is that care must be taken in not promoting European champions that end up being effectively protected from closure. Can the independence of competition policy be maintained given the politics of the Commission? States can lobby Commissioners and other Directorates than the one for competition (like Industry or Energy) to further national policies. Such lobbying would be hard to resist if it is done simultaneously by more than one large EU member state. An independent institutional body might protect competition policy from these industrial policy pressures. The fourth and final conclusion is that the 2004 reform of the merger control procedure in the EU was a step in the right direction, increasing checks and balances for merging parties and the role of economic analysis. However, the guarantees for the parties and the quality of analysis and decision-making, as well as the protection against the lobbying pressures of national governments and firms, could still be improved. One example of an independent institutional body would be an administrative panel, which is located within the Commission and recommends or even decides on merger cases. Another possibility would be a European Competition Agency.
Notes 1. Council Regulation (EC) No 1055/2005 of 27 June 2005 can be downloaded at: http:// eur-lex.europa.eu/LexUriServ/site/en/oj/2005/l_174/l_17420050707en00010004. pdf The Commission’s review of the implementation of the revised Pact is given at: http://europa.eu/scadplus/leg/en/lvb/l25084.htm 2. The new government in Sweden that was elected into office in the fall of 2006 has recently introduced the idea of a fiscal policy committee. 3. The Directive 2006/123/EC of the European Parliament and of the Council, of 12 December 2006, can be downloaded at: http://www.europarl.europa.eu/ comparl/imco/services_directive/061227_oj_services_en.pdf 4. The ‚Brouwer’ reports are the reports of the Economic and Financial Committee that have been first discussed at the April 2000 Ecofin Council. The main result of these reports was that all in all, the current regulatory and supervisory arrangements in Europe are adequate to ensure financial stability, though their practical functioning could be enhanced. The second report of April 2001 contains a list of proposals that could improve financial crisis management in the EU. The first report can be downloaded at: http://www.consilium.europa.eu/ueDocs/cms_ Data/docs/pressdata/en/misc/ACF16BD.htm. The second one, from April 2001, is available at: http://www.consilium.europa.eu/uedocs/cmsUpload/Brouwer%20 Report%20on%20Financial%20Crisis%20Management.pdf 5. Ecofin has taken steps in the meantime to recognise the problem according to our analysis: On 14 May 2005, the EU (Ecofin) has agreed to a financial crisis plan between central banks, banking supervisors and finance ministries of the EU in financial crisis situations. However there is still no agreement on who would cover the cost in the case of crisis.
9780230_547018_02_int.indd 32
3/19/2009 8:00:13 PM
Part I Labour Markets
9780230_547018_03_cha01.indd 33
3/19/2009 4:50:19 PM
This page intentionally left blank
1 Welfare to Work
1.1
Replacing or supplementing labour market incomes?
Social policy evens out the distribution of income. It prevents social unrest, it satisfies the taxpayers’ sense of justice and it insures against random variations in people’s lifetime careers. Ideally, it insures risks that are not privately insurable, either because risk markets suffer from adverse selection or because private insurance comes too late in a person’s life, when the veil of ignorance has already been lifted. Welfare programmes “... serve to even out differences in life chances, to achieve greater equality between generations and to redress inequality by race, gender, or health status. More generally, these programmes are intended to help people reallocate income over the lifecycle, to insure against events which cause income loss, and to provide a sense of security to all citizens” (Atkinson 1999, pp. 5–6). This is the theory. In practice poverty is dominated by joblessness, and large quantities of money are spent on the condition that the beneficiaries do not work and do not earn. The replacement of labour income with public transfers becomes the dominant form of assistance. As employment is the most important source of income, the replacement of labour income is a plausible implication of the social-policy motives expressed by Atkinson. The problem, however, is that the fraction of people not having a job is not exogenously given but depends heavily on policy itself. Benefits under the condition of not working operate like a wage paid for idleness which the market wage has to exceed. Since no one is willing to work at a market wage below the social benefit attainable without working – with the exception of those who fear being stigmatised as “welfare scroungers” – this benefit is a lower bound on market wages. However, in a market economy, an upper bound on an individual’s market wage is given by his (or her) productivity, i.e. the value added he or she is capable of creating. Thus there is a fundamental problem with people whose productivity is below the benefit that the welfare state is willing to provide. 35
9780230_547018_03_cha01.indd 35
3/19/2009 4:50:20 PM
36 Designing the European Model
These people, in principle, cannot find a job in a market economy under traditional policies. The wage has to be above their benefit to make them offer their labour, and the wage has to be below productivity to make firms demand this labour. The two conditions are mutually exclusive. Although driven by good intentions, the wage replacement policy turns out to be a policy of increasing the reservation wage – the wage below which a worker will refuse a job – and of preventing the creation of jobs which otherwise would have been available. This problem used to be minor when benefits were low relative to average incomes. However, the gradual expansion of the welfare state (expressed as the proportion of gross domestic product being spent on unemployment compensation and social assistance) has increased the number of people who are affected and has therefore increased the number of unemployed, in particular among the less educated, whose productivity is low relative to the minimum income which the state provides them. Unfortunately, this situation seems unlikely to change in the future. The productivity effects of the New Economy are likely to stimulate aggregate income growth and with it the growth of social standards. However, the number of people who just cannot keep up with the New Economy and who are unable to cope with modern work requirements may be increasing. The digital divide may not only be a problem among nations but also among the people within a nation. European integration may increase the desire for harmonisation of social standards. If traditional welfare benefits are harmonised, many people in the less-developed regions of Europe may find themselves in a situation where their labour productivity is below common European benefit standards. In Europe, there are regions where labour productivity is only a quarter of that elsewhere. Harmonising social standards without changing the conditions under which social benefits are paid would undoubtedly create mass unemployment in many of the less-developed regions if the benefits were sufficient for the more productive regions. The problem of the Mezzogiorno would spread. Thus it is opportune to search for alternative ways of designing the welfare state, ways that make it possible to help the needy without driving many of them into unemployment. Basically, these ways involve redefining poverty and the conditions under which the welfare state delivers its benefits. To satisfy Atkinson’s definition of welfare programmes, it is not necessary to make benefits conditional on people being jobless. They could also be made conditional on people being employed and not earning enough. A new definition of poverty would capture that. It is not a person who does not work who is poor but someone who works to his physical and mental capacity and is nevertheless unable to earn a sufficient income. With this definition of poverty, the welfare state would not replace labour income but supplement it when it is inadequate. Supplementing income to reach a social target level has very different implications for the functioning
9780230_547018_03_cha01.indd 36
3/19/2009 4:50:20 PM
Welfare to Work 37
of the labour market than the current system because it circumvents the problem described above. Even people with very low labour productivity would be able to find jobs because social benefits would no longer establish a lower bound to wages. People would be willing to work at very low wages, because they know that this would make them eligible for social benefits, and, for the same welfare state expenditures, they could even have higher incomes than in the current system. A number of mainly Anglo-Saxon countries have followed this line of reform and have moved from a wage replacing to a work complementing welfare system. This chapter will report on the experience of welfare-towork programmes and develop a proposal for a useful reform along these lines, respecting European norms of social protection.
1.2 The traditional approach In the OECD countries, an important part of social protection against unemployment is unemployment insurance. In order to be eligible for compensation, claimants must have worked and contributed to the insurance fund for a given period of time, they must be involuntarily unemployed and they must be actively looking for work. Financial assistance for those no longer eligible for unemployment insurance takes two forms: unemployment assistance and social aid. Unemployment assistance is designed as a follow-up benefit to unemployment compensation, paying a lower benefit than unemployment insurance. Social aid is given to those who qualify for neither unemployment insurance nor unemployment assistance. The government acts as a provider of last resort to secure a minimum standard of living. Social assistance in the EU member countries normally has an unlimited duration (see European Commission, MISSOC 2000). 1.2.1
Replacement policy
The traditional social security systems of most OECD countries can be characterised as passive. Benefits are provided to secure a minimum standard of living, and recipients receive the benefits without a strong obligation to look for work. This is especially true for social aid, which is provided without any significant obligation imposed on the recipient. Such a social security system leads to welfare dependency. It encourages inactivity, does not provide sufficient incentives to look for work and increases the opportunity cost of working in the market economy. In short, by following a wage replacement policy, the traditional social security system pushes the reservation wage up and thus destroys part of the employment opportunities which otherwise would have been available. The extent to which the required wages are artificially pushed up is influenced by the level of unemployment benefits and social assistance, the
9780230_547018_03_cha01.indd 37
3/19/2009 4:50:20 PM
38
Designing the European Model
duration of entitlement, the coverage of the system and the strictness with which the system is operated – as well as social attitudes. The influence of the welfare system on reservation wages can be represented and quantified by the net replacement rate (NRR) defined as: Benefit income when unemployed – tax on benefit income NRR = Earned income + benefit income when employed – tax on earnings and benefits The net replacement rate is the fraction of current or potential income which the social system provides to a person if he or she does not work. It varies according to the type of household, employee, sector of industry, wage and salary group and the reasons for not working. Table 1.1 shows the net replacement rates for an average production worker receiving unemployment benefits (at the beginning of receipt of benefits) or social assistance (long-term benefit recipient). It demonstrates that the net replacement rate at the beginning of unemployment is relatively high for a couple with two children but lower for someone who is single. Hence, the breadwinner has little incentive to seek regular work. This is all the more true if the (participating) spouse is long-term unemployed. There are, of course, differences in the net replacement rate from one country to another. The net replacement rates for long-term benefit recipients are lowest in the United States and Spain and highest in the Scandinavian countries (except Norway), Switzerland and the Netherlands. The replacement rate can be explained by the intended insurance function. However, a replacement rate also defines a minimum reservation wage, below which no one is willing to accept a job. In fact, for most people the minimum reservation wage may be even higher than that because when they decide to work they not only require a compensation for the lost special benefits but also for the time lost for leisure and for working at home or even for the loss of black market income. The higher the replacement rate, the better is the insurance protection, but the lower is the number of jobs which employers are willing to provide, given the skill distribution of the unemployed. 1.2.2 High unemployment of low-skilled workers The destruction of jobs and output resulting from the traditional policy is particularly severe at the lower end of the income distribution. The information provided in Table 1.1 refers to an average production worker. Workers with an income below the average will have a higher replacement rate than that reported in the table, and what is more: the replacement rate would be above one for people who do not work because their productivity and potential wage is below the level of social aid. However, since they do not work,
9780230_547018_03_cha01.indd 38
3/19/2009 4:50:20 PM
Welfare to Work 39 Table 1.1
Net replacement rates by family type at the APW level,a 1997 On unemployment benefitsb
Austria Denmark Finland France Germany Ireland Italy Netherlands Norway Portugal Spain Sweden Switzerland United Kingdom Australia Canada United States
Single
Couple 2 children
57 62 60 71 60 33 36 75 66 79 76 72 73 50 37 63 60
73 77 84 74 74 62 54 85 73 77 74 84 84 64 74 69 61
On social assistancec Single 54 48 58 38 54 33 28 60 36 42 25 58 61 50 37 25 7
Couple 2 children 70 97 97 50 52 62 62 79 54 61 43 100 90 73 74 59 48
Note: a APW: Average production worker. bIn the first month of benefit receipt: after tax and including unemployment benefits, family, and housing benefits. cFor long-term benefit recipients (60 months): after tax and including social assistance, family, and housing benefits. Source: OECD (1999), pp. 34 and 37.
the wage at which they would find employment is not known. Thus, no statistical information is available on the replacement rates of this important group. Social aid (and to a lesser extent unemployment assistance) is particularly problematic for the functioning of the labour market because, unlike unemployment insurance, it is a lower bound on the feasible wage distribution very much like a legally prescribed minimum wage. This lower bound is of limited importance for average production workers, but it destroys jobs for the less well qualified whose labour productivity is below the social aid level or not sufficiently above it to compensate for the work effort. The consequence of social aid is that it compresses the wage distribution and concentrates unemployment on the lower qualification (or productivity) levels. As even less educated people are normally able to fulfil some useful functions in the economy, a wage is conceivable at which these people could find employment. The problem is that society considers this wage to be too low and it is therefore replaced by a higher level of social aid; but the good intentions turn out to have adverse employment consequences for those people who seemingly benefit.
9780230_547018_03_cha01.indd 39
3/19/2009 4:50:20 PM
40 Designing the European Model Earnings dispersion (D5/D1) (y) 2.5
Canada
United States
Ireland
2.0 Australia Japan
1.5
y = –0.0105x + 2.0745 (–2.93) (12.88) R2 = 0.3493
0.5
Netherlands Switzerland
Norway France Denmark Belgium Öster- Finland, Sweden reich
Italy 1.0
United Kingdom Portugal Germany
0.0 0
10
20
30
40 50 60 70 Social assistance/Median earnings in % (x)
Figure 1.1 Social assistance and earnings dispersion in the OECD, 1997 (Relationship between the relative level of social assistance for a single-person household and earnings dispersion in 18 OECD countries) Note: The net replacement rate is calculated for the level of social assistance of a single and the earning of an average production worker. OECD database on earnings deciles refers to gross earnings. The cranings dispersion is the relation between median gross income (D5) to the gross income of the lowest decile (Dl).
Source: OECD (1999), p. 37.
Figure 1.1 gives some indication of the compression of the wage distribution through the social system. It regresses the earnings dispersion amongst relatively low earners as measured by the ratio of the median decile to the lowest decile of the wage distribution with the ratio of social aid and the average wage income across the 18 OECD countries for which the data were available. There is a significant negative correlation between these variables indicating that in countries like the United States and Canada, which have low levels of ordinary welfare payments, the distribution is indeed much wider than in countries like Denmark, Austria, Finland and Sweden where welfare payments are rather high. Table 1.2 gives an overview of the employment situation among the EU countries. The majority of European countries are reporting high rates of unemployment. Many of the low-skilled unemployed are unemployed for over a year. The standardised unemployment rate for EU members is nearly 9 per cent. In every country, and for both sexes, the less skilled have the higher rate of unemployment. Average unemployment rates in excess of 10 per cent occur only in France, Italy and Spain, and even there they are experienced only by women. For the less skilled such rates prevail in Finland, France, Germany, Ireland, Italy (for women), Spain, the UK (for men) and Canada. The unweighted average unemployment rate of workers with a “below upper-secondary education” is roughly 50 per cent higher than the general unemployment rate of the 12 EU member countries examined in Table 1.1.
9780230_547018_03_cha01.indd 40
3/19/2009 4:50:20 PM
Welfare to Work 41 Table 1.2 Unemployment rates by education for population 25 to 64 years of age, 1999 Below upper secondary education Austriaa Denmark Finland France Germany Irelanda) Italy Netherlands Norwaya) Portugal Spain Sweden Switzerland United Kingdom Australia Canada United States
Men Women Men Women Men Women Men Women Men Women Men Women Men Women Men Women Men Women Men Women Men Women Men Women Men Women Men Women Men Women Men Women Men Women
8.0 6.0 6.8 7.2 12.0 14.4 14.1 16.7 17.7 14.1 11.7 11.4 7.8 16.6 3.6 6.7 3.4 2.4 3.9 4.6 10.5 22.8 8.5 9.7 4.1 5.7 12.7 7.3 9.2 7.6 10.7 10.3 7.0 8.8
All levels of education 3.9 4.3 3.6 5.0 8.1 9.3 9.0 12.3 8.4 9.5 7.4 6.5 6.7 13.0 2.1 4.1 2.2 2.1 3.8 4.5 9.2 20.1 6.5 5.8 2.2 3.1 5.5 4.1 6.1 5.4 6.4 6.0 3.5 3.5
Note: a1998. Source: (2001) p. 274.
Again, it should be noted that the statistical information is incomplete and misleading since those whose productivity is below social aid do not work enough to qualify for benefits and often do not look for work and, therefore, do not count as unemployed.
9780230_547018_03_cha01.indd 41
3/19/2009 4:50:21 PM
42
Designing the European Model
1.2.3 Black market activities The policy of providing social assistance through replacing labour income has not only destroyed jobs by increasing reservation wages; it has also worked as a policy of subsidising black market activities. It is true, of course, that this was not intended, but as informal labour is the natural alternative to formal labour and as the payment of benefits stops when formal labour income is obtained, it is clear how the incentives have worked. Table 1.3 reveals that the underground economy has reached a high level in most OECD countries. Black market activities are, of course, only partially brought about by the generous provision of social assistance. Other factors like high tax rates also play a significant role. The replacement policy, however, not only provides incentives to work in the informal market, but also to use informal labour for home improvements, especially where home ownership is widespread. The spread of do-it-yourself stores not only reflects a fashion but also a rational reaction to economic incentives which undermine the division of labour and prevent the productivity gains that it entails. 1.2.4 Attempts to keep disincentive effects under control The disincentive effects have not been overlooked by policy-makers, but in most countries policy reactions have not called into question the traditional policy as such. There have been a number of attempts, however, to keep the disincentive effects under control.
Table 1.3 Underground economy as a percentage of GDP, 1998 1998 Austria Denmark France Germany Ireland Italy Netherlands Norway Spain Sweden Switzerland United Kingdom Australia Canada United States
9.1 18.4 14.9 14.7 16.3 27.8 13.5 19.7 23.4 20.0 8.0 13.0 14.1 15.0 8.9
Source: F. Schneider (2000).
9780230_547018_03_cha01.indd 42
3/19/2009 4:50:21 PM
Welfare to Work 43
One obvious provision in this regard has been the limitation of the periods during which people are eligible for benefits. Figure 1.2 gives an overview for EU countries. In some countries, however, eligibility is of unlimited duration. Belgium is an extreme example, but even a country like Germany, which at first glance seems to have restrictive rules, pays unemployment assistance (Arbeitslosenhilfe) at a replacement rate of slightly more than 50% for an indefinite period of time.1 Entering the formal labour market as unemployed and then moonlighting is an extremely attractive option under these conditions. Another attempt to keep misuse under control is the imposition of sanctions if a job is left voluntarily or if a job offered an unemployed person is not accepted. Table 1.4 gives an overview of the situation in some European countries. At first glance, the table signals a rather harsh approach in some countries, which even exclude the unemployed from benefits if they refuse a job twice.
months unlimited
70 60 50 40 30 20 10
United Kingdom
Sweden
Spain(4)
Portugal (3)
Netherlands
Luxembourg
Italy
Ireland (1)
Greece
Germany (1)
France (1)
Finland (1)
Denmark
Belgium (2)
Austria (1)
0
depending on age
depending on employment record
depending on age and employment record
for persons difficult to place
Figure 1.2 Duration of unemployment insurance Notes: (1) Unlimited unemployment assistance (2) Degressive (3) Unemployment assistance: 12 to 30 months (depending on age) (4) Unemployment assistance: 6 to 18 months.
Source: EU Commission, MISSOC (as of 1 January 2000).
9780230_547018_03_cha01.indd 43
3/19/2009 4:50:21 PM
44 Designing the European Model Table 1.4 Periods of benefit sanction following a voluntary quit and refusal of work or an ALMP placement First voluntary Refusal of work or ALMP placement Refusal of Second refusal Subsequent work or ALMP First refusal refusals placement quit or dismissal Subsequent for fault Denmark
5 weeks
Finland
3 monthsb
France
4 monthsc
Germany Norway Spain Switzerland
12 weekse 8 weeks exclusiong 6–12 weeks
United Kingdom
1–26 weeks
Australia
4–5 weeks
1 week (job), exclusion (ALMP) a 2 monthsb (job), 0–2 months (ALMP) temporary or definitive exclusiond 12 weekse 8 weeks exclusion 6–12 weeks 1–26 weeks (job), 2 weeks (ALMP) 4–5 weeks
exclusion
2 months or exclusion
2 months or exclusion
temporary or definitive exclusiond exclusionf 12 weeks
Temporary or definitive exclusiond
6–12 weeks or exclusionh 1–26 weeks (job), 4 weeks (ALMP) 6 weeks
6–12 weeks or exclusionh 1–26 weeks (job), 4 weeks (ALMP) 8 weeks
26 weeks
Note: ALMP: active labour market policy. a A first refusal of an ALMP placement leads to exclusion only during the “active period” (after 12 months of unemployment). bReduced to one month if the job in question is for less than five days. cAdmission to benefit after 4 months of unemployment is conditional on providing active job search during these 4 months. dThe word »exclusion« in this table generally implies an indefinite benefit stop or definite loss of remaining benefit entitlement. Legislation also provides for temporary exclusions. When an attitude of refusal of work is observed, exclusion is in principle definitive. eReduced in some circumstances. f Exclusion follows when sanctions totalling 24 weeks have been pronounced. gExclusion in cases of a quit, but a 3-month waiting period in cases of dismissal for fault. hA second refusal of an ALMP place leads to exclusion, and a second or third refusal of a job might lead to exclusion. Source: OECD Employment Outlook 2000, p. 135.
The actual policies are, however, much milder than suggested by the table, since the jobs rejected must have been appropriate, and the definition of what is appropriate is always a matter of ambiguity. Also, of course, the provisions do not alter the fact that public money flows if people do not work
9780230_547018_03_cha01.indd 44
3/19/2009 4:50:21 PM
Welfare to Work 45
and stops flowing if they begin to work. This is a challenge to the ingenuity of beneficiaries to invent reasons why an appropriate job cannot be found or why one offered is not appropriate. The problem may not be avoidable with unemployment insurance because benefits have to be provided if someone does not work. However, the provision of social aid and social assistance to the long-term unemployed or people who have never entered the labour force is clearly another matter. Here, in particular, the traditional policy should be reconsidered.
1.3 Welfare-to-work policies: Wage replacement vs. wage supplement The alternative to the policy of wage replacement is a policy of wage supplementation. Benefits are not given on condition of staying away from formal employment but on condition of participating in it and nevertheless not earning enough. When the reason for the assistance is not a random or temporary loss in employment but a permanent handicap that results in labour productivity too low to permit earning sufficient income even with full-time work, the policy of supplementation may be a useful alternative which could at least partly replace traditional welfare programmes. 1.3.1 The basic argument We consider a stratified society consisting of groups of (potential) workers each characterised by a particular (potential) productivity level. Figure 1.3 illustrates the case of a relatively low productivity group. Given the total amount of capital assigned to this group and the conditions pertaining in other labour market segments, there is a well-defined demand curve for labour of this quality as a function of its (net) cost to an employer. The supply of labour is assumed to be inelastic. In the absence of intervention, the market-clearing wage would be w*, but this is below the socially acceptable minimum wage w. To prevent anyone from having to live on an income as low as w* two policies are possible. One is a wage replacement policy and the other is a policy of wage supplementation. The first offers a public benefit payment of the minimum socially acceptable income. This places a floor under the market wage and A*–A people, who are in excess of the market demand for such low productivity workers at wage w, will be unemployed. Between them they will receive public funds represented by the area CDHI. Under the alternative policy of supplementing low earnings from public resources there is no floor to the market wage which for everyone in the relevant skill class falls to w*. Their income is brought up to the social minimum w by an employment subsidy w-w*, costing the government an amount represented by the area BDFE.
9780230_547018_03_cha01.indd 45
3/19/2009 4:50:22 PM
46
Designing the European Model Employment potential
Wage rate
B D E
F Labour demand
G
H A
I A*
Employment
Figure 1.3 Wage replacement versus wage supplement Source: Sinn et al. (2001).
We now need to relax some of the assumptions underlying Figure 1.3. The demand curve was drawn on the assumption of a given stock of capital available to co-operate with the category of low productivity labour being considered. If the policy is changed only for this group, raising its employment, we would expect capital to be reallocated from co-operating with other groups, raising the productivity of the group in question and reducing the cost of the wage supplementation programme. The situation is slightly more complicated but qualitatively unchanged, if there are several distinct groups (or a continuum) of low-skilled workers whose equilibrium wage would fall below w in the absence of intervention. It is also important to acknowledge that the combined effect of raising the employment of the least skilled groups and diverting capital to co-operate with them will be to depress the market wage for some other groups of workers whose wage was, and remains, above w. These workers, and their representatives in organised trade unions, are therefore likely to resent and resist the policy change advocated here. The big advantage of the wage supplement policy is that it does not generate unemployment. It is possible to help the low-skilled workers without eliminating some of the jobs necessary to employ all of them. This is not only better, because it preserves the dignity of the people in need and gives them a chance of improving their qualifications on the job, but also
9780230_547018_03_cha01.indd 46
3/19/2009 4:50:22 PM
Welfare to Work 47
because it goes along with a higher level of GDP. In Figure 1.3, the additional employment results in additional value added produced which is equal to the area CFIH. The analysis shows that the frequent claim that more growth is needed to generate more employment may have to be turned around for, in fact, it is the additional employment that generates more output. We believe that a change over from the wage replacement policy to the wage supplement policy would be a key element in a programme that would generate more output in Europe. The policy switch will not necessarily increase budgetary costs. The replacement policy gives a large subsidy to a small number of people. The supplement policy gives a small subsidy to a large number of people. Which case will cost the government less cannot be determined a priori. In Figure 1.3, the answer depends on the relative sizes of the areas CDIH and BDFE. The size of these areas is a function of the labour demand elasticity in the low-wage sector.2 Information on labour demand elasticities is scarce. With a linear homogeneous production function the constantcapital wage elasticity of labour demand is the ratio of the elasticity of substitution and the non-labour income share in GDP. While the former has been estimated to be in the range of 0.6 to 0.7, the latter can be assumed to be about 1/3. Thus, the absolute labour demand elasticity with a given stock of capital (and endogenous output) lies in the range of 1.5–2.3 These estimates do not refer to the low-wage sector and do not consider the dynamics of labour demand resulting from the fact that lower wages will induce more capital investment. In the low-wage sector, the labour demand elasticity is higher than in the overall economy. If dynamic aspects are included, the elasticity increases further. Taking into account the special situation of the low-wage sector, long-term adjustments and the self-financing aspect of the promotion of higher employment, a wage supplement policy can be expected to cost the government less than the traditional policy and it would also generate more employment and a higher GDP. If the wage were raised from w* to the same level w (Figure 1.3) regardless of the workers’ productivity level, very low productivity persons would have no incentive to study or otherwise raise their productivity a little; unless they could qualify themselves for a job paying more than w, they would be no better off. On the other hand, to pay a significant subsidy to everyone who works would be impossibly expensive. In practice, therefore, wage supplements are made a function of wages paid (or more often of earnings). Typically, the supplement or tax credit rises with earnings over a certain range of lower earnings and is spaced out over a range of higher earnings. The first effect ensures some incentive to improve one’s skills and productivity – possibly through formal qualifications. The second has the disadvantage of adding to the effect of positive marginal income tax rates in tending to discourage both work and training. This disadvantage is hardly avoidable,
9780230_547018_03_cha01.indd 47
3/19/2009 4:50:22 PM
48
Designing the European Model
but it seems less problematic than leaving people in idleness. The advantage of a wage supplementation policy is that, unlike the current systems, people can actually be induced to work and that it involves lower budget costs for the government, while achieving the same social objective. 1.3.2 Real-world examples: A country comparison The wage supplement policy is no longer pure theory. Several OECD countries have introduced policies that provide incentives to low-skilled workers to participate in the labour market. These policies go beyond the well-known active labour market policy or the tightening of the eligibility requirements for social benefits. Instead, they constitute a fundamental change of welfare policy. Social benefits are no longer provided to inactive persons but to people who are employed or who make efforts to become employed. The welfare-to-work policies consist of a variety of measures: ●
●
●
●
Employment-conditional benefits, tax credits or wage subsidies and payroll-tax rebates given to employers. The obligation of benefit recipients to be active (participation in public employment or job training). Otherwise, they lose their entitlement to benefits. By taking part in these activities, the benefit recipient will increase his or her human capital and become accustomed to working. The shortening of the duration of benefits. New benefits are often made available only for a limited period of time. The promotion of intensive job search.
There are mainly eight countries that have actually introduced major welfare-to-work programmes: the United States, Canada, the United Kingdom, Ireland, Denmark, France, the Netherlands and Sweden. The programmes differ substantially. The major features of six of these programmes are summarised in Table 1.5. The US welfare-to-work programme consists essentially of an employmentconditioned tax credit (the Earned Income Tax Credit – EITC) and a workfare system. The UK’s approach comprises the “working families’ tax credit” and the “new deal programme.” Denmark is making those active who have been unemployed for a long time and has made the eligibility criteria for social benefits more demanding. Sweden is “activating” the long-term unemployed and is offering a two-year wage subsidy to employers who hire unemployed workers aged 57 or older. In France, firms implementing the 35-hour week are entitled to relief on the employer’s social security contributions; an employment-conditional tax credit was introduced in 2001. The Netherlands obliges the long-term unemployed to participate in public employment or training programmes and reduces employers’ social security contributions when hiring certain types of unemployed workers. In the following we shall look more closely at the experience of the different systems.
9780230_547018_03_cha01.indd 48
3/19/2009 4:50:23 PM
Welfare to Work 49 Table 1.5 Welfare-to-work programmes Elements Employmentconditional benefits and tax credits Wage subsidies Payroll tax rebates Obligation to work for welfare recipients Duration of benefit provision limited Support of job search Other: Tightening of eligibility requirements for social benefits Active labour market policy
US
UK
Denmark
Sweden
France
Netherlands
X
X
–
–
X
–
–
X
–
X
–
–
–
–
–
–
X
X
X
X
X
X
–
X
X
–
–
–
–
–
X
–
–
–
–
–
–
–
X
–
–
–
–
–
X
X
–
X
Source: CESifo DICE Database.
United States The US welfare-to-work programme provides employees in low-wage occupations with an earned income tax credit (EITC) whose goal is to create work incentives for low-wage earners and boost their incomes. The beneficiaries are subject to federal income tax. If the tax credit is higher than the income tax owed, the difference is paid out to the eligible families. Otherwise, it is deducted from the income tax. The earned income tax credit is administered by the Internal Revenue Service. Employment is required for eligibility and the programme is primarily aimed at working people with children. The amount of tax credit received is based on gross earnings. Figure 1.4 illustrates the three ranges of the earned income tax credit. In the first range the increase in the tax credit is proportional to income. In the second range the tax credit remains constant. In the third range it declines until a maximum income is reached. The amount of tax credit and the income limits differ according to household type.
9780230_547018_03_cha01.indd 49
3/19/2009 4:50:23 PM
50 Designing the European Model Tax credit 4000
3000
Families with one child
Families with two children
2000
1000 Families with no child 0 0
5000
10000
15000
20000
25000
30000
35000 40000 Gross income
Figure 1.4 US earned income tax credit for various household types Source: R. Zahradnik, I.J. Lav, A D.C. Earned Income Tax Credit could provide tax relief and reduce child poverty, Center on Budget and Policy Priorities, 9 February, 2000.
Distinctions are made in the case of families with two or more children, families with one child, and people without children. The highest credit is given to families with two or more children. The parameters of the earned income tax credit in the year 2000 are shown in Table 1.6. A family with two or more children and a yearly income of between $1 and $9,720, for example, receives a tax credit of 40 cents for every additional dollar earned. With a gross income of $9,720 the maximum credit of $3,888 is reached. This remains constant until gross income reaches $12,690. For every dollar earned above $12,690, the tax credit is then reduced by 21 cents. With gross income reaching $31,152, the tax credit is reduced to zero. In the third range in which the tax credit is reduced, the marginal charges on income are higher than the marginal rate of income tax. As a rule, in this range of tax credit reduction, the marginal effective tax rate capturing both the reduction of the earned income tax credit and the increase in ordinary taxes amounts to about 50%. In 1999, nearly 19 million workers took advantage of the tax credit. It amounted to an average of $1,632 (Economic Report of the President 2001, p. 200). In addition to the earned income tax credit, the US has pursued a workfare model since 1996. The temporary assistance to needy families (TANF) programme was introduced in order to overcome welfare dependency. The legal entitlement to welfare was eliminated, and willingness to accept work was made a condition for welfare assistance. If this work requirement is not fulfilled the claim to welfare lapses. The principle of reciprocity was firmly
9780230_547018_03_cha01.indd 50
3/19/2009 4:50:23 PM
Welfare to Work 51 Table 1.6
US federal earned income tax credit parameters in 2000
Household type
Phase I Credit Income percentage limitsa($)
Families with one child Families with two or more children Families with no children
Phase II
Phase III
Maximum Phase-out Income benefits ($) rate (%) limitsa($)
34.0
0–6,920
2,353
15.98
40.0
0–9,720
3,888
21.06
7.65
0–4,610
353
7.65
12,690– 27,413 12,690– 31,152 5,770– 10,380
Note: a Annual amounts for income or EITC assistance. Source: H. Johnson (1999).
established: the state is obliged to provide money and jobs and, in return, the welfare recipient is obliged to work. In addition, the TANF programme sets a limit of five years on welfare benefits during a person’s lifetime. Moreover, it gives priority to work over education and training. Finally, to a greater extent than in the past, the federal government gives the individual states authority to decide the (final) nature of their welfare programmes. Global subsidies to the states’ budgets have created an incentive for the implementation of welfare reform. United Kingdom Along with the United States, the United Kingdom has had a long tradition of assisting working people who have low incomes. In 1971 a family income supplement was introduced. This was replaced in 1988 by the Family Credit (FC). This in-work benefit was in turn replaced at the end of 1999 by the Working Families’ Tax Credit (WFTC). The credit is paid to all low-income families with children who have at least one adult working for at least 16 hours per week. The payments are set at a level that guarantees families a minimum income of £200 per week, with additional payments for larger families. Furthermore, 70 per cent of all childcare costs (up to £150) are covered. A small additional payment is made if at least one adult per family works for more than 30 hours per week. When earnings exceed £90, the credit is reduced by 55 per cent of any additional earnings. However, because this adjustment is only made every six months, temporary marginal increases in hours worked are not discouraged. This “taper rate” comes in addition to income tax and social insurance contributions. The credit is paid through wage packets every month. In 2000, it covered 1.1 million recipients (in a country with a total of around 20 million households and 27 million wage and salary earners), costing over £5 billion per year – about two-thirds of a percentage point of GDP.
9780230_547018_03_cha01.indd 51
3/19/2009 4:50:23 PM
52 Designing the European Model
Since 1998, the working families’ tax credit has been supplemented by a new arrangement, which aims at increasing peoples’ employability and at helping them find work. It consists of a number of different strands. The new deal’s target group are the young long-term unemployed, aged 18 to 24. After young people have been unemployed for 6 months, they enter a “gateway” of intensive counselling with a personal adviser. This can last for a maximum of 4 months, during which time they are expected either to be placed into a regular job or to have entered one of four subsidised programmes, lasting at least six months: ●
●
●
●
A subsidised job with a regular employer (secured by a 6-month subsidy of £75 a week) Work experience in the voluntary sector (while receiving benefits plus £15 a week) Work experience in an environmental project (while receiving benefits plus £15 a week) Full-time vocational education (while receiving benefits).
All of the programmes described above include at least one day a week of training. The new arrangement is well funded and has received general approval. Since April 1998, many young people have been provided with work, and long-term unemployment among young people has already declined by two-thirds (in a period of falling unemployment). In addition, a pilot programme was initiated for those aged 25 and over who have been unemployed for two years or more. It offers personal advisers, a subsidised job, full-time education, training and continuing guidance for finding work. Denmark Denmark’s welfare-to-work policy does not provide in-work benefits. Thus it does not really fit the listing made here. Nevertheless, the Danish approach is worth including because it involves exceptionally harsh punishment for people who do not accept job offers. Recent reforms included two elements. One is compulsory activation, introduced in 1996. Compulsory activation applies to the longer-term unemployed. It requires these unemployed to participate in full-time private and public on-the-job training in return for receiving unemployment benefits and social assistance. The length of unemployment before compulsory activation takes place is one year for those aged 25 years and older. For younger people, compulsory activation already applies after six months, and in addition the benefits are cut in half. One of the aims of this “right and duty” policy is to stimulate adequate job-search behaviour, thereby avoiding the need for compulsory activation.
9780230_547018_03_cha01.indd 52
3/19/2009 4:50:24 PM
Welfare to Work 53
The other element of the reforms is the tightening of the generous unemployment benefit system itself. The benefit period has now been reduced to four years, and activation does not start for a new benefit period. The eligibility criteria for unemployment benefits have been made more demanding. After the first job refusal, there is a one-week penalty and after the second refusal unemployment benefits are cut altogether. The duration of occupational protection allowing unemployed people to refuse a job offer that involves a change of occupation has been reduced to three months. At the same time surveillance of benefit eligibility has been tightened. Sweden In Sweden, the so-called activity guarantee came into force in 2000. It applies to people who receive either unemployment or other social benefits and who have not had ordinary (unsubsidised) work for 27 months. This group comprised 50,000 persons (1¼ per cent of the workforce) at the end of 2000 who are obliged to participate in full-time activation to maintain their benefit entitlement. The activity guarantee requires job seekers to participate in an active labour market programme or some other education or training in order to increase their human capital. Each activation period within the framework of the activity guarantee is planned for a maximum duration of six months, at the end of which an evaluation is made and a new six-month period may be planned. There is no fixed time limit to the total duration of the activity guarantee period; in principle it may be open-ended. In addition to this new form of activating the long-term unemployed, a generous two-year wage subsidy is offered to employers hiring unemployed workers aged 57 or above during the activity guarantee period. This subsidy is 75 per cent of the wages, although it is subject to a maximum of SEK 525 per day. France France’s welfare-to-work policy consists mainly of programmes intended to increase employment and to train specific groups having difficulty in finding work. Employers are exempt from social contributions or receive wage subsidies if they offer employment initiative contracts, on-the-job training, apprenticeships, etc. Since 1 January 2000, firms implementing the 35-hour week have been entitled to more generous relief on employers’ social security contributions. The scheme combines both flat-rate relief and declining-rate relief. The relief decreases from FRF 21,500 per year if the legal minimum wage (SMIC) is paid to FRF 4,000 if a wage 80% above the minimum wage is paid. Thereafter, a flat-rate exemption of FRF 4,000 is provided. In addition to the existing programmes, an employment-based tax credit was introduced in September 2001. Workers not earning more than 1.4 times the minimum wage (FRF 99,016 per year) will receive the tax credit. A
9780230_547018_03_cha01.indd 53
3/19/2009 4:50:24 PM
54
Designing the European Model
couple with two children will receive FRF 3,400 up to FRF 9,400. This threeyear programme is estimated to cost FRF 25 billion. Netherlands The Netherlands’ welfare-to-work programme consists of an extensive programme of subsidised jobs for long-term unemployed workers who normally receive social assistance benefits (called Melkert jobs after the former Minister of Social Affairs and Employment). The aim is to provide about 60,000 jobs for 32 hours a week, at either the legal minimum wage or slightly above. It includes four different programmes: Melkerts 1, 2, 3 and 4. The jobs are created in municipalities (maintenance of public areas, education, childcare, etc.) and in healthcare (hospitals, home care etc.). Recipients of social assistance who refuse these jobs suffer benefit sanctions. In order to increase labour demand, employers’ social security contributions for low-paid workers were cut (SPAK). Employers paying less than 115 per cent of the legal minimum wage are entitled to these cuts. SPAK reduces gross labour costs by 10 per cent. Moreover, firms hiring long-term unemployed workers can qualify for an additional reduction in social security contributions for a period of four years (VLW). The combination of SPAK and VLW can cut labour costs by up to 23 per cent. 1.3.3 Evaluation of the US experience Since the labour market conditions and the welfare-to-work programmes differ from country to country and since some of these programmes have only been introduced in recent years, a comparative assessment of their effects is not yet possible. However, there is now ample evidence on the results of US reforms. The US welfare-to-work programme, consisting of the earned income tax credit (EITC) and the temporary assistance to needy families (TANF), had the objectives of increasing labour supply and employment and supporting the poor. The latter objective has largely been achieved. Half of all payments go to families with income below the poverty line. With regard to increasing labour supply, one must distinguish between two effects: the encouragement of participation in the labour force and the encouragement of the supply of additional working hours by those already employed. There is no doubt that participation has increased. The rise in net incomes and the accompanying decrease in net replacement rates have created work incentives. The effect on the supply of working hours, however, unsurprisingly, is not as clear. As income increases due to the earned income tax credit, a household can afford to enjoy more leisure and reduce working hours (the income effect). On the other hand, there are substitution effects caused by changes in relative prices between leisure and working time (the substitution effect). The substitution effect depends on which of the three ranges applies (see Figure 1.4). In the initial range, where the state subsidises
9780230_547018_03_cha01.indd 54
3/19/2009 4:50:24 PM
Welfare to Work 55 Caseload (indexed) 1.4 1.2
United States
1.0 0.8 0.6 0.4
Wisconsin
0.2 0.0 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 Figure 1.5 Welfare caseloads, 1987–1999 (as a multiple of the 1987 caseload) Notes: The data refer to to ordinary welfare payments (AFDC) and temporary assistance to needy families (TANF). Wisconsin data: calendar year monthly average. US data: federal fiscal year monthly average.
Source: US Department of Health and Human Services, Administration for Children and Families; Wisconsin Department of Workforce Development, Economic Support Division.
each dollar earned with 40 cents, there is a clear substitution effect towards working more. In the flat intermediary range, there is no substitution effect. And in the third range where the tax credit is phased out the substitution effect is negative. Empirical studies confirm that there are, in fact, these effects, but they also show that the net overall effect on hours worked is positive. The people entering the labour force because of the tax credit and working longer in response to a higher net marginal wage are more important than those who work slightly less because the phasing out of the credit creates a highly effective marginal tax burden. The great achievement of the programme has been to integrate the unemployed into an orderly working life. Part of the prosperous growth and the employment miracle in the United States during the last two decades can, in our opinion, be attributed to this effect. The additional jobs brought about growth since the employed were productive and generated income which financed additional consumption. This was supply-side policy at its best. There is particularly good information on the results of the 1996 welfare reform carried out in Wisconsin. Basically these can be summarised as follows. ●
●
The number of TANF welfare recipients has been reduced considerably. This is illustrated in Figure 1.5. Aggregate welfare expenditure has fallen.
9780230_547018_03_cha01.indd 55
3/19/2009 4:50:24 PM
56 Designing the European Model ●
●
●
●
● ●
About 70 per cent of former welfare recipients were successful in finding jobs. They are now in a position to provide for themselves by working. Thirty per cent do not work after leaving welfare. Illegal work is becoming less attractive and, indeed, unattainable for those obliged to work in the formal sector. Income of former welfare recipients who find a job has increased. It has decreased for those who continue not to work. In-work benefits are lower than out-of-work benefits because they are only needed as a supplement to earned income. Close targeting of beneficiaries is a prerequisite for all measures. Work as a condition for social benefits activates a self-selection mechanism amongst welfare recipients which helps to limit abuses. In addition, welfare recipients are induced to be more active.
1.4 A proposal for Europe Helping the needy under the condition that they remain idle is a strange idea which has little in common with both the basic principles of justice and the positions of Europe’s traditional parties, whether conservative, labour oriented, social democratic, Christian or green. Replacing wage income of the unemployed sounds better than paying for idleness, but it results in similar policy measures. Unemployment largely follows from paying people under the condition that they not work. The wage replacement policy that characterises Europe’s welfare states increases the reservation wage, pushing it dangerously close to too many people’s productivity or even above it. The high and persistent level of unemployment in Europe makes it necessary to reconsider the European welfare state, the wage replacement policy and the definition of poverty. Unemployment benefits in the usual sense of the word have a useful insurance function against random, temporary job losses. They should therefore remain a crucial ingredient. If the benefits are given only for a limited period of time, if the replacement rate is moderate, if recipients face penalties in case of job rejection and if they are forced to actively seek employment, this type of insurance will only generate limited moral hazard effects relative to the income security it provides. Unemployment assistance that is paid for an unlimited period of time is much more problematic, since it provides a reasonable income to those who earned a relatively high wage and is all too often only the basis for additional income from moonlighting to make ends meet. Unemployment assistance of this sort should be abolished altogether and integrated with the payment of social aid. Social aid itself needs to be reformed, however. Since it is fixed in absolute terms, independently of the previous wage, it prevents all those people from participating in the labour market whose labour productivity is lower than,
9780230_547018_03_cha01.indd 56
3/19/2009 4:50:24 PM
Welfare to Work 57
or not sufficiently above, social aid. Labour productivity is an upper bound on wages, and social aid is a lower bound. The two bounds define an empty set of job opportunities for an increasing number of people. We believe that a modified earned income tax credit system of the American type, albeit with significantly higher benefit levels, is to be recommended. Instead of taking money away if someone decides to accept a job, he or she should be given money. And instead of defining poverty as not having a job, it should be defined in terms of earning too little when working. This principle was explained in the introduction. One major difference from the American system refers to the treatment of those who do not find a job despite the new policies or claim not to have found a job. To maintain their work incentives, they should receive only very low benefits during their search. The US benefit level satisfying this requirement may be too low, however, when judged by European social preferences. In fact, the minimum income definitions specified by social laws and supreme courts of justice preclude a simple translation of the American solution to Europe. To avoid this difficulty we include public jobs in the programme we propose. Starting from a system with a given level of social aid which satisfies a country’s minimum income requirement, we define four different categories of people and the welfare payments for which they are eligible. 1. People who cannot work for medical or social reasons to be defined by law. They receive the traditional type of social aid. 2. People who can work but do not, for whatever reason. These people receive only a benefit level of the American type, much below the current level of social aid in Western Europe, but much above American levels. 3. People who work in simple government jobs. They receive a wage income equal to the country’s minimum income requirement (the previous social aid). The government is obliged to provide the necessary number of jobs. 4. People who work in the private sector. They receive the earned income tax credit, properly adjusted to ensure that the sum of market wage and government subsidy exceeds the minimum income requirement (i.e. the income earned by category three). This system defines an income ladder which people will be able to climb. It reduces the lower bound on wages to the level of social aid received by the second category of people and creates the additional jobs that are needed if that level is sufficiently low. While it is true that, in the short run, the lower bound on wages could also be given by union wages, we believe that union wages will soon react to the new incentives and ensure that their members become eligible for the earned income tax credit.
9780230_547018_03_cha01.indd 57
3/19/2009 4:50:24 PM
58 Designing the European Model
In countries with statutory minimum wages it may be necessary, however, to reduce these wages to the level of social aid as given to people in category two above or to define minimum incomes such that they include the funds received from the government. This should not be a major problem. It is crucial for the principles of our proposal that the second category exists. It is necessary to make sure that people climb the income ladder by working more rather than less. There will not be many people belonging to this category, though, and those who do will not be a problem. People who cannot work are in category one and people who need more income but cannot find a job in the private sector will be in category three. The few who congregate in category two must have alternative incomes from undeclared work in the informal economy which makes it preferable not to spend their time on the jobs provided by the government. Compared to a traditional welfare system, the system we propose will shift the vast majority of the current welfare recipients into categories three and four. Most people will find a job in the private sector because low wages will fall. They will still be better off than before, because the sum of the earned income tax credit and the market wage will exceed the previous level of social aid. The others, who fail to find a job in the private sector, will work for the government where they receive an income that satisfies the legal or Supreme Court minimum income definitions. In the short run, after introducing the new system, many people may find themselves in category three. Instead of receiving social aid for free they will have to work for it. Over time, however, the free market wage for simple labour will decline and more and more jobs will be created, as the American example has shown. Thus more people will gradually be integrated into the private job market, and category three will run dry. In the end, unemployment among the less skilled will largely be eliminated, and the economy will be closer to the full employment low-wage situation depicted by point F in Figure 1.3. In sum, our proposal will create a better welfare state by improving its incentive structure and provide more income to the needy, given the overall expenditure which the government can afford. This new type of activating welfare state will better satisfy the goals defined at the outset than the current one, and it will bring about a higher activity level and more economic growth from which all will benefit.
Notes 1. The United States provides unemployment insurance for six months but no unemployment assistance. 2. The crucial variable for finding out which policy is cheaper is the labour demand elasticity. Express the unemployed A*–A resulting from the replacement policy as part of the total labour force A*, A* A A* ( HI)
9780230_547018_03_cha01.indd 58
3/19/2009 4:50:24 PM
Welfare to Work 59 and let the subsidy in the supplementing case be given by w w* dw ( DF). The ratio of the budgetary costs K in the two cases is then given by Kr ( A * A )w A* w Ks ( w w*) A* wA*
where the subscripts r and s stand for the replacement and supplementing policies. The ratio of γ and δ is the demand elasticity for labour. This becomes clear if the first two equations are rewritten as
A * A A*
A A*
and
w w *
w
w
.
w
According to these equations, γ is a relative change of employment and δ is a relative change of the wage rate. The ratio of these variables,
A/A* w/w
,
is the labour demand elasticity. Obviously, if, and only if, y/ 1 is the supplement policy the cheaper alternative since then Kr / Ks > 1. See Sinn (2000). 3. Burgess (1988) determined a value of 1.85 and Nickel and Symons (1990) estimated 1.92. Franz and König (1986) only found a value of one. Estimates of the output-constant demand elasticities are lower, usually in the range between – 0.3 and – 0.5, because they assume that capital input diminishes as employment goes up. See Fuchs, Krueger, Poterba (1998) and Hammermesh (1993). Of course, reference to the output-constant elasticity makes no sense in the present context, since capital input will increase rather than diminish when wages fall. The relevant elasticity for our purposes is one which incorporates all endogenous factor adjustments, and that elasticity is even higher than the capital-constant elasticity to which the text refers.
References Atkinson, A. B. (1999) The Economic Consequences of Rolling Back the Welfare State, The MIT Press, Cambridge, London. Blundell, R., A. Duncan, J. McCrae and C. Meghir (2000) “The Labour Market Impact of the Working Families’ Tax Credit,” Fiscal Studies 21(1), 75–104. Burgess, S. (1988) “Employment Adjustment in UK Manufacturing,” Economic Journal 98, 81–103. Economic Report to the President (2001) Washington DC.
9780230_547018_03_cha01.indd 59
3/19/2009 4:50:24 PM
60
Designing the European Model
European Commission (2000) “MISSOC 2000 Social Protection in the EU Member States and the European Economic Area,” Luxembourg, OPOCE. Franz, W. and H. König (1986) “The Nature and Causes of Unemployment in the Federal Republic of Germany since the 1970s: An Empirical Investigation,” Economica 53, 219–44. Fuchs, V., A. Krueger and J. Poterba (1999) “Economists’ Views about Parameters and Policies: Survey Results in Labor and Public Economics,” Journal of Economic Literature 36, 1387–425. Hammermesh, D. (1993) Labor Demand, Princeton University Press, Princeton, New Jersey. Johnson, H. (1999) A Hand Up: How State Earned Income Tax Credits Help Working Families Escape Poverty, Center on Budget and Policy Priorities, Washington DC (http://www.cbpp.org/11-12-99sfp.pdf). Nickell, S. and J. Symons (1990) “The Real Wage-Employment Relationship in the United States,” Journal of Labor Economics 8, 1–15. OECD (1999) Benefit Systems and Work Incentives, OECD Publishing, Paris. OECD (2000) Employment Outlook, OECD Publishing, Paris. OECD (2001) Education at a Glance, OECD Publishing, Paris. Schneider, F. (2000) “The Shadow Economies around the World: Sizes, Causes and Consequences,” Journal of Economic Literature 38(1), 77–144. Sinn, H.-W. (2000) “The Threat to the German Welfare State,” Atlantic Economic Journal 3, 279–94. Zahradnik, R. and I. J. Lav (2000) A D.C. Earned Income Tax Credit Could Provide Tax Relief and Reduce Child Poverty, Center on Budget and Policy Priorities, Washington DC (http://cbpp.org/2-9-00sfp.htm).
9780230_547018_03_cha01.indd 60
3/19/2009 4:50:25 PM
2 Labour Market Reform in Europe
2.1
Introduction
To solve the European unemployment problem, economists have usually advocated reforms that reduce labour market rigidities, which to increase the cost of labour: such reforms include less generous unemployment benefits, reductions in minimum wages and more flexible employment protection provisions. These recommendations are based on a cornerstone of modern economic theory: the notion of an “equilibrium rate of unemployment” to which the labour market converges in the absence of shocks, once all prices and wages have adjusted. This view holds that the equilibrium rate is entirely determined by real frictions at the microeconomic level, such as the workers’ bargaining power, information and incentive problems at the firm level, the efficiency of job search, etc. While these parameters themselves depend on the above-mentioned institutions, they do not depend on short-run fiscal and monetary policies, which only have a transitory effect on employment. However, in many countries structural reforms have typically been difficult if not impossible; they have suffered delays and have often not been designed in an optimal way from the point of view of economic efficiency. Hence, reducing the minimum wage for youths in France in 1994 encountered such violent opposition that the proposal was eventually withdrawn. Reducing firing costs in Spain in the 1980s was only possible by liberalising the use of temporary contracts for newly hired workers, while leaving employment protection unchanged for permanent workers. This created a dual labour market, which has been much criticised for a number of reasons: the inequality it creates in working conditions between workers with identical jobs; the burden it imposes on the unemployment benefit system by increasing the inflow of eligible unemployed workers; and the potential wage inflation it has induced by creating a cushion of temporary workers who shelter permanent ones from job loss. Nevertheless, in Italy, France and 61
9780230_547018_04_cha02.indd 61
3/19/2009 8:04:29 PM
62 Designing the European Model
Portugal, reductions in employment protection have taken the same form, and across-the-board reductions in employment protection are rarely seen. More recently, in France, a reform of the unemployment benefit system was implemented. It was originally intended to follow the recommendations of many economists, by making sure that the payment of benefits would be conditional on the workers actively seeking a job. In order to get the benefits, the worker now has to regularly report his job search activities to the administration. However, in the process of being negotiated between employers’ associations, unions and the government, the reform has been considerably watered down. Sanctions and tight monitoring have been replaced by a bona fide pledge of active job search on the part of the worker. Worse, as a compensation for the unions, the previous negative dependence of benefits on unemployment duration was abolished. As a result, it is far from clear that the unemployed search more actively because of the reform, especially given that monitoring rests on social workers who traditionally consider the unemployed as their customers. Similarly, there are a number of examples of active labour market policies, which most economists consider superior substitutes for passive unemployment compensation, that are ill-designed in the sense that they fail to boost the participants’ prospects for a regular job. For example, in 1997, France implemented the “emploi jeunes” programme, which directed young unemployed workers, who often had an appreciable educational background, to exclusively public jobs with a low skill content. Such a programme is unlikely to enhance the employability of participants in the medium term. Swedish labour market policies are also famous and many authors claim that they have played a key role in maintaining a low unemployment rate there. However, the huge empirical literature that has tried to evaluate them is generally cautious or negative. A recent survey by Calmfors et al. (2001), for example, concludes that active labour market policies in Sweden have “probably reduced open unemployment, but also reduced regular employment” and that they are not an efficient means of employment promotion when used on a large scale. Again, an important issue is that neither politicians nor bureaucrats have a strong incentive to design and manage them so as to improve the long-run efficiency of the labour market. Finally, more systematic evidence based on synthetic indicators suggests that overall, European labour markets have been substantially more rigid since the mid-1970s than before, while there was a divergence in the late 1980s and 1990s between those countries that deregulated their labour markets to some extent and apparently enjoyed some benefits in the form of lower unemployment rates, and those where rigidity continued to increase or remained high, where unemployment did not fall. The latter group includes the three largest continental European countries: Germany, France and Italy. Thus, according to Nickell (2003), relative to the 1980s,
9780230_547018_04_cha02.indd 62
3/19/2009 8:04:29 PM
Labour Market Reform in Europe
63
unemployment benefit replacement ratios had gone up in Austria, Finland, Italy, Norway, the Netherlands, Portugal, Sweden and Switzerland. They remained roughly constant in France, Germany and Denmark, and fell in Belgium, Ireland, Spain and the United Kingdom. Employment protection indices have decreased in many countries, with the exception of France, but not by much, a fact that should be taken with caution, given the qualitative nature of these indices. It is also known that such a fall is mostly due to marginal liberalisations, such as more scope for using temporary contracts. In any case, employment protection in continental European countries remains much stricter than in Anglo-Saxon ones. Finally, labour taxes remain high or have gone up in most countries. From the early 1980s to the late 1990s, the total average tax rate on labour, which is computed by adding VAT, income taxes and payroll taxes, has increased from 58 per cent to 66 per cent in Austria, from 46 per cent to 51 per cent in Belgium, from 65 per cent to 68 per cent in France, from 56 per cent to 64 per cent in Italy, and from 40 per cent to 45 per cent in Spain. Elsewhere in Europe, the total average tax rate has remained high and roughly constant, with four exceptions where it has fallen: Ireland (from 37 per cent to 33 per cent), the Netherlands (from 55 per cent to 43 per cent), Norway (from 65 per cent to 60 per cent) and the United Kingdom (from 51 per cent to 44 per cent).1 Interestingly, in Ireland, the Netherlands and the United Kingdom, unemployment declined a lot during that period. That same analysis implies that, when comparing the evolution of employment in a panel of countries and relating it to changes in labour market institutions, the traditional “neoclassical” reform strategies based on reducing labour costs do work, although it is not straightforward to disentangle which aspects are more important. The Dutch reforms, which have been successful in reducing unemployment, illustrate this point: they have been quite comprehensive2, encompassing nationwide agreements on wage moderation (see also Chapter 3) and reductions in firing costs, unemployment benefits and minimum wages as well as cuts in welfare spending and better monitoring of the unemployed. Therefore, while we do not know exactly what works and what does not, we are confident enough to say that comprehensive reforms do work. However, they have not been pursued, except in Anglo-Saxon and a few small countries. The overall picture suggests that 1. Those countries that managed to reduce unemployment all liberalised their labour markets along at least some dimensions; 2. None of those who did not liberalise (France, Germany, Italy) managed to reduce unemployment. 3. Small countries have less of a hard time reforming than large ones. 4. Many policies are put into place to fight unemployment, but a lot of them are ineffective, misguided or counter-productive.
9780230_547018_04_cha02.indd 63
3/19/2009 8:04:29 PM
64
Designing the European Model
One implication of the last point is that reforms often encounter fierce political opposition, despite the claim by many economists that they are worthwhile in order to reduce unemployment. Furthermore, those policies that actually are implemented in order to fight unemployment are so heavily shaped by political constraints that they are often counter-productive.
2.2
Political constraints faced by reformers
So, why is it that labour market reforms have been inadequate and/or rare in light of standard recommendations by economists? 2.2.1 Resistance by interest groups A typical explanation is that labour market rigidities benefit powerful interest groups because they create and protect rents for their members and that these groups are likely to oppose reforms or at least to distort them so as to protect themselves (see Saint-Paul 1993, 1996 and 2000). If the total gains from reform in terms of extra output and employment are large enough, one should in principle buy the support of workers whose rents are eliminated by the reform, by promising to transfer a share of the total gains to them. In practice, however, such a possibility rarely enters the political debate and sounds remotely abstract. This is true for a variety of reasons, from the difficulty to clearly identify the losers from a broad reform – thus generating incentives for everyone to claim being a loser – to constitutional provisions about equal treatment. 2.2.2 The role of ideology Resistance by interest groups is enhanced by ideological views that question the idea that only painful structural reforms may effectively reduce the long-run unemployment rate. Indeed, many analysts and journalists dispute that notion and hold the traditional Keynesian view that a fiscal and monetary expansion could bring unemployment back to its level of the 1960s. Consequently, they ascribe Europe’s high unemployment rate to macroeconomic mismanagement in the 1970s and 1980s and advocate more expansionary fiscal and monetary policies. This is compounded by the observation that in the United States the unemployment rate, while substantially lower than in Europe, is still not negligible (some 5 per cent vs. 8 per cent in the EU and goes up quite rapidly in recessions. Thus, one may ask, what is the use of harming some of the most disadvantaged workers and risking severe social disruptions to obtain a labour market which does not look that desirable after all? One factor that has been overlooked in the debate about European labour market reform is the role of ideologies and statements about the functioning of the economy in shaping our beliefs about what should be done. In particular, while in the United States there is a well-defined “mainstream” in
9780230_547018_04_cha02.indd 64
3/19/2009 8:04:29 PM
Labour Market Reform in Europe
65
economics, with other approaches still existing but being marginalised, in Europe one still typically is of the opinion that there are many competing schools that are equally worth considering. The combination of a given policymaker’s preferred “school” and some emotional discourse about “helping people” may lead to erroneous policies which will not cure unemployment. However, these policies are not uniformly erroneous and often benefit some interest groups. These interest groups then have an interest in promoting the underlying “school” or ideology regardless of whether or not it is correct. Some examples may illustrate our point. The view that all unemployment is Keynesian and there is no such thing as a long-run equilibrium rate of unemployment, or that such rate is zero or very low, implies that the bulk of unemployment would eventually be eliminated by traditional fiscal and monetary tools, and this underlines much of the short-run Keynesian stimulation policies. Instead of recognising the failure of such policies, the advocates of such ideologies argue that macroeconomic stimulus has not gone far enough. One should note that similar ideas have been vindicated within the mainstream of economics by the “hysteresis” view (see Box 2.1). The idea that an increase in wages will help reduce unemployment because it stimulates consumption is popular among union leaders, who often advocate wage increases in order to lift the economy out of a slump. While such an effect may exist, any positive effect on employment is bound to be short-lived, while the long-term effects are likely to be negative. In the long run the logic of equilibrium unemployment prevails, and unemployment depends on whether and to what extent wage aspirations are compatible with productivity, as discussed in Box 2.1. Higher wage pressure makes it necessary for unemployment to increase in order to bring actual wages back to what productivity levels allow. In the short run, an increase in wages may indeed increase employment: if workers consume more than capitalists, total aggregate demand may go up. But even that is not guaranteed: investment will fall because higher wages reduce profitability, and so will exports, as higher wages reduce the country’s competitiveness vis-à-vis the rest of the world. Despite this, trade union leaders tend to overemphasise the view that higher wages promote employment, because increasing wages benefits incumbent employees, provided their jobs are protected enough, regardless of the effect on demand and employment. The “lump-of-labour” fallacy, which states that the total amount of work is fixed and can only be shared among those who want to work, has led to many misguided policies, such as early retirement to “make room” for the young, or working time reduction. Both theory and evidence run counter to the “lump-of-labour” view. The long-run equilibrium rate of unemployment does not depend on the size of the labour force. The reason is that the size of the labour force does not affect the link between the unemployment
9780230_547018_04_cha02.indd 65
3/19/2009 8:04:29 PM
66
Designing the European Model
Box 2.1
The equilibrium rate of unemployment
According to standard macroeconomic theory, the unemployment rate converges to an “equilibrium rate” in the absence of structural reform. This equilibrium rate is the one which makes workers’ wage-setting behaviour compatible with firms’ labour demand and price-setting behaviour. The former depends on factors such as unions’ bargaining power, employees´ wage aspirations and labour market institutions, such as minimum wages, unemployment benefits and employment protection. The latter depends on productivity, taxes imposed on firms, and the degree of competition among them. An increase in unemployment has a moderating effect on wages by exerting discipline on workers’ wage demands, while at the same time increasing the wages that firms are able to pay. This is because at lower employment levels, everything else equal, productivity is higher, as the least productive jobs are shed first. Therefore, if initially the unemployment rate is lower than the equilibrium rate, workers will tend to demand wages that are higher than what firms are willing to pay for labour. Firms will stop hiring and start firing, and unemployment will tend to rise until it is again equal to the equilibrium rate. Economists usually summarise the determinants of equilibrium unemployment as in Figure 2.1 (see, for example, Nickell and Layard 1999 or Calmfors and Holmlund 2000 for a more detailed account). The WS curve shows how wage demands increase with employment. The LD curve represents the wage that firms can pay workers, which depends on technology (which determines productivity), as well as the degree of competition in product markets. For simplicity, we have only drawn the LD curve for the long run when the capital stock is adjustable, implying that the feasible wage does not depend on unemployment. In the long run, LD is horizontal, as depicted in the Figure. In the medium term, however, productivity falls when employment goes up, as it takes time for the capital stock to adjust upwards; consequently, the wage that firms are willing to pay is a declining function of employment. The mediumrun LD curve (not drawn here) would be downward sloping. The equilibrium level of unemployment is the one which makes wage aspirations compatible with the wages firms are willing to pay, that is it is determined by the intersection of the two curves. According to the equilibrium rate theory, only structural reforms which change the long-run determinants of firms’ and workers’ behaviour can have a lasting effect on unemployment. A reduction in unemployment benefits, for example, reduces wage demands by lowering workers’ incomes outside employment. Thus, as illustrated in Figure 2.2, the WS curve shifts down, and (once the capital stock has adjusted to the new increased profitability of the firms) equilibrium is restored at a new, lower rate of unemployment. Note that wages do not decline, because in the long run they are pinned down by the behaviour of the firms. However, they would have to fall in the short run, until enough capital is accumulated to absorb the increase in employment. Also, everything else equal, a reform that increases the wages that firms are willing to pay at any level of unemployment (for example because productivity has risen) will reduce the equilibrium rate of unemployment. This is because less discipline needs to be imposed on workers, who can now be paid more (the LD curve shifts up in this case).
9780230_547018_04_cha02.indd 66
3/19/2009 8:04:29 PM
Labour Market Reform in Europe
67
Other policies only have a transitory effect on unemployment, because they do not change the equilibrium rate. This is particularly true for fiscal and monetary policy, which stimulate the economy in the short run but not in the long run. While a macroeconomic expansion may speed recovery if unemployment is initially higher than the equilibrium rate, it would fail to reduce it permanently below the equilibrium rate. Such an attempt would simply result in greater inflationary pressure and/or mounting public debt. This view does not commend a total consensus, though, because there are mechanisms such that the transitory effects of fiscal and monetary policy can have persistent effects by changing the equilibrium rate of unemployment in the future (the so-called “hysteresis” effect pointed out by Blanchard and Summers 1986). This will happen, for example, if workers lose skills during spells of unemployment (see Pissarides 1992); a temporary contraction will then have long-lasting effects on equilibrium unemployment by reducing workers’ productivity. However, most specialists agree that while these mechanisms make the effects of macroeconomic policies more persistent, they are not strong enough to affect the equilibrium rate in a distant enough future (see Layard et al. 1991).
Wages WS
LD
Labour demand schedule
Wage-setting schedule Equilibrium rate
0 Figure 2.1
N
Employment
The equilibrium rate of unemployment
Source: EEAG.
rate and wage aspirations, nor does it affect the wages firms are willing to pay in the long run. A reduction in the labour force decreases unemployment initially. However, this leads to more wage pressure. Higher wages then lead to fewer hirings and also less investment, as profitability goes down. That contributes
9780230_547018_04_cha02.indd 67
3/19/2009 8:04:29 PM
68
Designing the European Model Wages
WS
LD
Labour demand schedule
New rate
Wage-setting schedule WS
Old rate
0 Figure 2.2
N
Employment
Effects of structural reforms on unemployment
Source: EEAG.
to bringing employment back to the equilibrium rate, which is what happens in the long run. Thus, a 10 per cent decrease in the workforce eventually leads to a 10 per cent decrease in employment and an unchanged unemployment rate. Working time reduction (WTR) is less straightforward to analyse, but there are good reasons to believe that it may lead to increases in wages per hour that offset any tendencies of the equilibrium unemployment rate to decrease. One mechanism is that a reduction in working time tends to reduce the total wage income of each employee: this creates strong incentives for unions to push up wages per hour, which will in turn reduce the total number of hours worked. Also, the tendency for total wage income to fall when working time is reduced makes work less attractive relative to non-work. That reduced attractiveness is another factor working in the direction of increasing wage pressure when working time is reduced (see, for example, Calmfors 1985 for a more detailed analysis). As far as evidence is concerned, one may simply point out that since the mid 1980s, thanks to their flexible labour market, the United States has been able to create millions of jobs so as to absorb a large number of immigrants. And recent empirical studies of working time reductions conclude that they have had an adverse impact on employment (Hunt 1999; Crépon and Kramarz 2002). Nevertheless, we may see such policies being advocated, and even implemented, because they again benefit some groups of workers by sheltering them from competition, although this remains to be further investigated.3
9780230_547018_04_cha02.indd 68
3/19/2009 8:04:30 PM
Labour Market Reform in Europe
69
Another ideology is the general scepticism, among many analysts and policymakers, about the allocative role of prices in general, and wages in particular. Dismissing the common sense view that less labour is demanded when its price goes up amounts to dismissing all policies that would lead to reductions in wages, or in the total cost of labour, in order to create jobs. Such a view may be supported by the difficulties one encounters when estimating such effects, especially with aggregate data: careful empirical work, however, has made a rather convincing case that increases in labour costs reduce employment (Laroque and Salanié 2000, 2002). The view that it does not, however, remains popular in some places, and again it benefits workers who are already employed and whose labour is a substitute for the jobs that would be created by such reductions in labour costs. Finally, reforms may also be blocked by “analytical myopia”: the general public tends to be more confident about the direct effects of policies than about their indirect ones, which shapes beliefs in a way rarely friendly to employment-enhancing reforms.4 Thus, the direct effect of a reduction in the minimum wage is to reduce the income of minimum wage earners; subsequent job creation only comes later. The direct effect of reducing employment protection is that some workers will lose their jobs; the benefits in terms of job creation come later and hinge on the firms’ rational calculations taking into account the reduced cost of having to dismiss a worker in the future. Reductions in the generosity of unemployment benefits impose “hardship” on the unemployed, but their beneficial effects on employment involve the complex process of wage bargaining and so on. 2.2.3 Labour market reform under political constraints: what does it look like? Following the preceding discussion, one may even ask why governments would want to implement a reform of the labour market at all. Most of the pressure for implementing structural labour market reforms comes from employers, who want lower labour costs and more freedom in managing the workforce, and from the financial unsustainability of the welfare system when unemployment gets too high. The actual policy being followed results from a complex game where these factors interact with the objectives of unions and representatives of incumbent workers to protect their rents and with the government’s need to satisfy its electorate while demonstrating its competence. This process results in a number of outcomes, none of which is fully satisfactory from the point of view of economic efficiency. The political economy approach to designing labour market reforms helps to understand the characteristics of the economic environment and of reform design that make reform more likely to be politically viable. Saint-Paul (2000) has discussed how it may be rational for groups that would otherwise block labour market deregulation to actually support structural reforms if properly
9780230_547018_04_cha02.indd 69
3/19/2009 8:04:30 PM
70 Designing the European Model
designed. Such rational design sometimes resembles reforms that are actually undertaken. For example: ●
●
●
Political support for reform is more likely in times of “crisis” when incumbent employees are more exposed to the risk of job loss. Indeed, Saint-Paul (1996) has found that most reduction in employment protection provisions have taken place at times of rising unemployment. Reform is more viable if it is designed in a two-tier fashion, leaving existing provisions unchanged for incumbent workers, but applying new rules to labour contracts signed after the reform. Reform is more viable if there is a lag between the date it is decided and the date it is implemented. This increases the probability that a voter who is in the group of losers at the time of the decision (say, employed in a high-wage, protected job) may have moved to a group that is benefiting from the reform at the time it is implemented (say, employed in a precarious job, or unemployed).
In other words, postponing reform induces people to vote as if under a veil of ignorance, taking into account the interests of groups to which they do not belong, because they might belong to them at the time reform is implemented. However, other policies may be difficult to rationalise and be driven by short-termism or concerns about politicians’ public image. In any case, none of the approaches undertaken so far are fully satisfactory; and, after almost 30 years of high unemployment, we are seeing their limits.
2.3 Some reform strategies: advantages, drawbacks and political problems The preceding discussion highlights the difficulty, but not the impossibility, of implementing “orthodox” labour market reforms. Thus it is natural to analyse alternative ways of reducing unemployment, that might perhaps be less effective but would encounter less political opposition. 2.3.1 Trying alternatives to labour market reforms Product market liberalisation On paper, and this may sound surprising to people who are not professional economists, deregulating product markets may have positive employment effects almost as large as labour market deregulation. This is because the equilibrium rate of unemployment is the one which makes workers’ wage aspirations compatible with the wages that firms are willing to pay, which in turn depend on productivity and on the degree of competition among firms. By increasing productivity and competition among firms, product market deregulation increases the wages that firms can pay to workers, thus
9780230_547018_04_cha02.indd 70
3/19/2009 8:04:30 PM
Labour Market Reform in Europe
71
allowing for a “tighter” labour market, that is lower unemployment, according to the mechanisms described in Box 2.1. Furthermore, product market liberalisation may have positive side-effects on wage formation, which yields extra dividends in terms of job creation. As discussed by, for example, Blanchard and Philippon (2003), greater competition in product markets reduces the monopoly rents that are appropriable by workers and makes labour demand more sensitive to wages, since it is more difficult for firms to pass higher wages on to their customers in the form of higher prices. This tends to impose greater discipline on workers in their wage demands, thus building in wage moderation and leading to additional reductions in the equilibrium rate of unemployment. One may thus conclude that instead of deregulating the labour market, which faces fierce opposition, one may achieve equivalent results by deregulating the product market. The experience of the Scandinavian countries suggests that this could be partly true: in Sweden, for example, product market deregulation was quite aggressive in the 1990s (see Box 2.2), and during that period, unemployment fell quite rapidly after the large increase at the beginning of the decade. However, we do not believe such a conclusion to be warranted. First, one can also observe political opposition to deregulation from the employees of regulated firms. Second, it is not totally clear whether the effects on employment are large enough, especially given that only a fraction of the economy is affected by deregulation and privatisation. Even if a large fraction is regulated, each sector has its specific regulations, and it takes time to change all of them.
Box 2.2 Sweden’s deregulation in the 1990s Sweden is one of the countries that went farthest during the 1990s in deregulating their product markets, substantially increasing competition and the number of companies in a number of sectors. In particular: • The number of companies owned by the state declined by more than 50 percent. Employment in state-owned enterprises also declined. • Privately managed social services have grown substantially. • The public monopoly on telecommunications was abolished, and the market opened to international players. The market share of the former state monopoly has been falling rapidly. • Similarly, the state monopoly on postal services was abolished. • Airlines have been deregulated to an extent similar to other EU countries. • Private operators have been allowed in railways. • Electric power has also been deregulated to a large extent, with the standard outcome of a state-owned monopoly for operating the network and a number of competing suppliers. A complete overview of Sweden’s deregulation is given in CEEP, 2000.
9780230_547018_04_cha02.indd 71
3/19/2009 8:04:30 PM
72 Designing the European Model
Keynesian policies A number of continental European countries have resorted to short-run Keynesian economic stimulus rather than structural reform. As we argued in Box 2.1, such policies cannot have long-lasting effects on unemployment. Indeed, while public debt increased in France in the 1990s, the unemployment rate has typically remained at 9 per cent, while in the United States, which pursued structural reforms in the 1980s, it has fallen to about 5 per cent, in accordance with the equilibrium rate theory. If they do not work, why do governments pursue such policies? Because they work well over the short-run, and the electoral cycle imposes a short horizon on politicians; and because their costs are diffuse and remote in the future, while structural reforms are politically painful, as discussed above. Furthermore, as the UK experience suggests (Box 2.3), the gains from structural reforms only show up after several years, a decade in the case of that country, which is beyond a government’s electoral horizon. Finally, it is also true that such policies benefit from uncertainties about the “true model” of the economy, which leaves room for ideology, as discussed below. We do not want to imply that such policies should be abandoned altogether. They can be quite efficient, since unemployment can be above the equilibrium rate for fairly long periods as a result of a recession or a fall in the equilibrium rate. This is illustrated by the Fed’s policy in the 1990s: the Fed deliberately allowed unemployment to stay below the earlier estimated equilibrium rate, an estimate which proved ex post to have been too high. However, when one observes high and persistent unemployment for a decade or more, that is a sign that the equilibrium rate is itself high, and Keynesian policies are bound to fail.
Box 2.3 Lags in the effects of British labour market reforms In the United Kingdom, major reforms took place in the 1980s, in particular reductions in the power of trade unions (which are discussed in greater detail in Chapter 3), and in the generosity of unemployment benefits. As described by Blanchflower and Freeman (1993), these reforms have been broad: “industrial relations laws that weakened union power; measures to enhance self-employment; privatisation of government-run or owned businesses; reduction in the value of unemployment benefits and other social receipts relative to wages; new training initiatives; tax breaks to increase use of private pensions; lower marginal taxes on individuals; and elimination of wage councils that set minimum wages.” However, it took a long time for theses changes to translate into durable reductions in equilibrium unemployment. As Figure 2.3 suggests, unemployment was still around 10 percent in 1994.
9780230_547018_04_cha02.indd 72
3/19/2009 8:04:30 PM
Labour Market Reform in Europe
73
Active labour market policies Active labour market policies have been popular in many countries as a tool to reduce unemployment. They are appealing to politicians because one has a sense of directly tackling the problem by acting upon the unemployed individuals. They also directly impact on unemployment statistics for sheer accounting reasons. Finally, in a world where social insurance is provided to the unemployed, relief jobs or training periods seem a more productive way of using the unemployed workers’ time and monitoring them than passive compensation. However, things may go wrong with active labour market policies, which may explain why they are so costly. Specifically, economic analysis shows that they can reduce the equilibrium rate of unemployment if they increase the search intensity of the long-term unemployed. However, it is precisely that sort of competitive pressure that unions and dominant interest groups want to combat when they resist orthodox reforms. If they have a say in the design of active labour market policies, they may well have an interest in designing them so as to prevent active search of a regular private sector job by programme participants (Saint-Paul 1998). A telling example is the above-mentioned French “emploi jeunes” programme, which was a programme of relief jobs aimed essentially at young workers who just finished their studies, consisting exclusively of jobs in public administrations or non-profit businesses. Clearly, this removes a good deal of competition for private sector jobs.
12
in% 12
10
10
8
8
6
6
4
4
2
2 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 Figure 2.3
Unemployment rate in the United Kingdom (seasonally adjusted data)
Source: European Commission: European Economy Autumn, 2003.
9780230_547018_04_cha02.indd 73
3/19/2009 8:04:30 PM
74 Designing the European Model
Such perverse incentives may explain why active labour market policy has failed to reduce unemployment in some countries, despite its substantial costs. 2.3.2 Eliminating inefficiencies in the welfare system Another approach is to try to eliminate inefficiencies in the welfare system without questioning its basic construct. In principle, that opens the door for changes that are not radical, but would face little opposition as most workers would benefit. Simplifying employment protection legislation One case in point is employment protection. In most countries, especially southern European ones, it is extremely difficult to shed labour without offering workers high compensation for dismissal. The reason is that most of the costs are in the form of highly uncertain legal procedures, which have even led, in some cases, to some layoffs being reversed years later, jeopardising the firm’s financial soundness and the job stability of all current employees. The existence of these legal procedures often lead to generous settlements, but this is by no means systematic, and such an outcome depends on the outcome of bargaining between firms and workers. Such a costly and inequitable system is in part due to misunderstandings about the role of employment protection. Employment termination is a normal component of a market economy, just like any other contract termination, and should not be criminalised. In addition, promoting an employment protection policy on the ground that it will reduce unemployment is misguided. The implied reduction in job destructions is offset by a reduction in job creation, and countries with the lowest unemployment rates (the United States and the United Kingdom) have little job protection. The role of employment protection legislation should be to give the right incentives to firms and to compensate workers for the cost of job loss. In that respect, we advocate the elimination of the current system of legal procedures in many countries, especially the southern European ones, and its replacement by a simple “firing tax” which would be paid to the worker as severance payment. Monitoring the search activity of the unemployed In principle, monitoring the search activity of the unemployed and imposing sanctions on them in the form of reduced benefits if search is not active enough may be a productive way to reconcile a high degree of social insurance with an efficient labour market. And this philosophy seems to work well in both Sweden and the Netherlands, which have achieved low unemployment rates. However, the recent French experience with a reform called PARE (Plan d’Aide au Retour à l’Emploi) highlights a number of difficulties with this approach. In particular, there is an agency problem in that
9780230_547018_04_cha02.indd 74
3/19/2009 8:04:30 PM
Labour Market Reform in Europe
75
the social workers in charge of monitoring the unemployed may well treat them, rather than the taxpayers, as their customers, and fail to report misdemeanour. In the French case, as compensation for the supposed monitoring of the unemployed, the duration and level of benefits was increased. If the monitoring turns out to exist only on paper, the reform will have been counterproductive. In-work benefits Replacing means-tested welfare payments to the poor by in-work benefits such as earned income tax credits or the French “prime à l’emploi” is also a non-controversial proposal on which a large majority of economists would agree. Indeed, this (as proposed in the 2002 EEAG report) is likely to have sizeable effects on employment by increasing work incentives. It is not clear to us whether or not there should be a smaller constituency to oppose such policies, relative to, say, reductions in employment protection or the extent of collective bargaining. After all, in all cases, competition between outsiders and insiders is enhanced. However, in-work benefits surely are less vulnerable to opposition based on cognitive issues than those discussed in Section 2.2. It seems harder, for example, to argue that they would “reduce demand”; and their impact effects are more favourable than for the traditional prescriptions, so that they are less vulnerable to “analytical myopia”. 2.3.3 Establishing convergence of interests between insiders and outsiders Finally, one may enhance the viability of labour market reforms by increasing convergence of interests between the incumbent, protected employees, who do not profit from many reforms and the unemployed or the firms who gain from many reforms. This can be built into the design of a specific reform, or one may develop institutions (such as stock ownership) that are not per se labour market reforms but may help enhance their viability in the future. Two-tier systems Some continental European countries have also adopted the two-tier reform strategy, as outlined above. This has been especially true with respect to reducing employment protection legislation, which in Spain, but also in France, Italy and Portugal, has taken the form of a liberalisation of the use of temporary contracts. This strategy has come under much criticism for various reasons. It has been accused of maintaining a dual labour market, with haves and have-nots, thus fuelling social tensions. Another argument has been that, by allowing firms to use temporary workers as a buffer against labour demand fluctuations, the strategy has increased the protection of permanent workers, thus creating excess wage pressure and eventually
9780230_547018_04_cha02.indd 75
3/19/2009 8:04:31 PM
76 Designing the European Model
reducing employment (Bentolila and Dolado 1994). Finally, by increasing turnover and at the same time the fraction of workers who are eligible for unemployment benefits, the two-tier strategy has put pressure on the financing of the unemployment insurance system. In our view, these criticisms are exaggerated, to say the least, because no other way of increasing labour market flexibility in continental Europe has been found. The idea that a two-tier system of employment protection leads to a segregated labour market, for example, is overstated. There is substantial mobility from temporary to permanent contracts, and firms value temporary contracts as a way to test the quality of newly hired workers. Furthermore, the two-tier system has generated political dynamics that are favourable for reform. In Spain, for example, the government has been able to trade reductions in the employment protection provisions associated with permanent contracts against further restrictions in the use of temporary contracts. Profit sharing The political economy approach suggests that the virtues of policies such as profit sharing or the promotion of stock ownership have been underestimated. These policies generated a lot of interest in the 1980s, following work by Martin Weitzman (1984), who argued that profit sharing enhances a country’s macroeconomic stability over the business cycle. Profit sharing also creates convergence between the interests of workers and those of capitalists by making each worker a little bit of a capitalist. Our discussion in Section 2 implies that such schemes would make policies of wage moderation, which boost profitability and job creation, more acceptable to incumbent workers. Profit sharing should therefore be seen as a way of reducing the rate of equilibrium unemployment.
2.4
The current situation: An opportunity for reform?
The preceding discussion implies that the margin of manoeuvre for governments to combat unemployment with institutional reforms is typically quite narrow. However, past experiences like that of the United Kingdom (or, to a lesser extent, that of Spain in the 1980s) suggest that in situations perceived as a “crisis,” one can be substantially more ambitious. To some extent, such a point has been reached in a number of European countries, not so much because of overall macroeconomic performance (the current slowdown is milder than the previous one), but because of budgetary problems and the feeling that “globalisation” is making the burden of labour rigidities unbearable. For one thing, paradoxically, the very efforts made by governments to combat unemployment tend to make it a bigger problem. Beyond their sheer psychological effect of making them a political test, these efforts tend to increase social spending per unemployed worker, making unemployment
9780230_547018_04_cha02.indd 76
3/19/2009 8:04:31 PM
Labour Market Reform in Europe
77
more of a financial problem. Hence, active labour market policies in France and Sweden represent considerable spending per recipient and, in the former case, have failed to bring unemployment back to its level of the 1960s. In the case of Sweden, the efficiency of the large-scale measures undertaken has been questioned (Calmfors et al. 2001). Similarly, when high persistent unemployment is erroneously fought using fiscal and monetary policies, this leads to excess deficits and/or inflation, which in the end create the need for more drastic structural reforms. Second, financial problems in other areas, such as pensions and health care, which are even more sensitive, make it more valuable to increase employment. The recent drive for labour market reforms in Germany comes from a more general crisis of the welfare state and a recognition that the commitments of the government are unsustainable. Labour market rigidities can be tackled because it is clear that the alternative is reducing benefits that are more valuable than these rigidities, such as pensions or health insurance benefits. Thus, Germany has recently started an ambitious and comprehensive reform package called “Agenda 2010”. It is too early at this stage to assess whether it is going to be implemented fully or only partly. But it includes a number of measures that are described in Box 2.4. The programme does little to reduce the rents of incumbent employees, as predicted by the political economy approach. For example, as will be discussed in more detail in Chapter 3, the reforms do not encompass the system of pay bargaining. On the other hand, the programme is more ambitious about reducing unemployment benefits and tightening eligibility requirements as well as enhancing measures to bring the long-term unemployed back to work. These measures are clearly more likely to be acceptable to incumbent employees than changes in the pay-setting system, which might reduce their rents. Furthermore, the design of the reforms confirms to some extent the principles discussed above in that there is a sense in which they affect marginal, non-core groups of workers more than others (reducing unemployment benefit duration for elderly people exerts less downward pressure on wages than reducing it for core workers, as the elderly unemployed compete less intensively with insiders). Their political viability is also enhanced by some delay in implementation (for example, the gradual phasing out of the previous system of non-contributory benefits avoids political opposition from those whose benefits would otherwise immediately fall, leading more workers to consider the reform under some “veil of ignorance”). However, similar changes have been opposed or neutralised by unions in other circumstances. The German example is therefore illustrative of how a severe crisis enhances reform possibilities. Third, changes in the international economic environment may increase the cost of labour market rigidities. Increased openness to international markets makes producers more sensitive to increases in labour costs. A
9780230_547018_04_cha02.indd 77
3/19/2009 8:04:31 PM
78
Designing the European Model
Box 2.4
Labour Market Reform in Germany
In the course of 2003, a first round of reforms of German labour market policy became effective that had been enacted in December 2002 (see EEAG, 2003, box on p. 31). Further amendments of labour market regulation were passed in the summer of 2003. In December 2003, decisions were taken regarding a second round of reforms, following intense negotiations between the two houses of parliament in which the Bundesrat opposition took an active role in pushing through more stringent measures than those planned by the Federal government. Core elements of recent changes are as follows. Unemployment benefits: The time limit for entitlement to unemployment insurance benefits (Arbeitslosengeld) for workers aged 55 and over will now be reduced from 32 to 18 months, while the standard limit of 12 months for other beneficiaries is left unchanged. The dual system of non-contributory benefits with unlimited duration − unemployment assistance (Arbeitslosenhilfe) and social assistance (Sozialhilfe) – is now integrated in one comprehensive scheme (Arbeitslosengeld II) that is basically modelled on the less generous scheme of the former social assistance. However, those whose contributory benefits have expired will be moved only gradually, over a period of two years in each individual case, to the lower level of benefits. Also, incentives to take up a new job are still limited by high transfer rates (that even exceed 100 per cent over a certain range of incomes in the case of family households). On the other hand, job offers now have to be accepted without any binding minimum level of wages, thus creating some leeway for expanding the low-wage sector, which might now offer job opportunities for the large number of unemployed people with low skills. Job Centres: Building on the 2002 reform, all kinds of services for individuals seeking employment (administration of benefits, counselling, job placement) are to be provided in Job Centres operated by the Federal Employment Services and act as “one-stop” agencies. In the future, Job Centres will also be responsible for job seekers who are currently on social assistance. Alternatively, municipalities, some of which have been rather successful in the past in re-integrating this particular subgroup of individuals in the labour market, can opt for becoming responsible for all the long-term unemployed (12 months or longer, i.e., living on Arbeitslosengeld II). In any case, municipalities are still responsible for welfare recipients who are not classified as being job seekers. As the new assignment of responsibilities has some inconsistencies and as the administrative changes associated with the introduction of Job Centres are still under way, it is as yet unclear whether the discontinuities involved in the re-organisation will harm the attempts to fight long-term unemployment. Protection against dismissal: Up until now, the strict rules regarding protection against dismissal that are in place in Germany are applicable to all firms with five or more employees. This threshold is now raised to ten or more employees, while the rules themselves remain unchanged. In addition, only new employees in firms that stay within the “5-to-10 employees” category will no longer be subject to these rules. As the existence of this threshold is usually considered an obstacle to further expansion of small-sized enterprises, extending it could lead to some job growth in firms that end up employing less than 10 individuals.
9780230_547018_04_cha02.indd 78
3/19/2009 8:04:31 PM
Labour Market Reform in Europe
79
The 2002 reform package was primarily meant to have an impact on the way public employment services are operated. Specifically, rules regarding active job search and availability of work for those receiving unemployment benefits were tightened. So far, the major effect appears to be that, over the year, unemployment growth related to the business cycle was lower than expected because some registered as unemployed decided to withdraw from the labour force rather than re-enter the labour market, while others were channelled into new “nonstandard” forms of employment (in particular, a subsidised form of self-employment for individuals formerly unemployed or employment with non-profit temp work agencies). Apart from a surge in so-called “mini jobs” (with less than 400 Euro of monthly pay, effectively a domain of second earners or individuals seeking a second source of income, not of people otherwise unemployed), employment continued to decrease over the year of 2003. Measures taken in the new 2003 package are potentially much more important for moving the unemployed back into work. Yet, when compared to a number of proposals made by leading German economists (see for instance, German Council of Economic Experts, 2002, No. 472−476, Joint Forecast of the Economic Research Institutes, Fall 2003, pp. 38–40, or Sinn et al. 2003), the changes introduced so far may still be insufficient to substantially reduce the current level of structural unemployment.
greater pace of technical progress increases the need for labour turnover and penalises those societies that impose a tax on turnover in the form of employment protection provisions. New technologies may increase the demand for skilled workers and reduce the demand for unskilled workers, thus reducing the real wages and/or employment of the latter. These arguments illustrate how a feeling of urgency helps to implement reforms that are otherwise politically doomed. In other words, crises and severe recessions are more conducive to reform than booms. This is somewhat unfortunate and paradoxical, as most economists tend to agree that the adverse effects of reforms on job destruction are more bearable in booms. However, the record suggests that political incentives to implement structural reforms in booms are quite weak. This is probably because governments’ popularity surges in booms and they do not want to jeopardise it with risky reforms.
Notes 1. Source: Nickell (2003). 2. See for example Barrell and Genre (1999). 3. See Marimon and Zilibotti (2000), for a model where “insiders” may prefer to take welfare gains in the form of reduced working time rather than higher wages. 4. See Gersbach and Schniewind (2001) for an analysis.
9780230_547018_04_cha02.indd 79
3/19/2009 8:04:31 PM
80
Designing the European Model
References Barrell, R. and V. Genre (1999) “Labour Market Reform in the UK, Denmark, New Zealand and the Netherlands,” National Institute of Economic and Social Research, http://www.niesr.ac.uk/staff/rbarrell/LMP1999.pdf. Bentolila, S. and J. Dolado (1994) “Labour Flexibility and Wages: Lessons from Spain,” Economic Policy, 18, 53–99. Blanchard, O. and T. Philippon (2003) “The Decline of Rents, and the Rise and Fall of European Unemployment,” mimeo, Massachusetts Institute of Technology. Blanchard, O. and L. Summers (1986) “Hysteresis and the European Unemployment problem,” NBER Macroeconomics Annual. Blanchflower, D. and R. Freeman (1993) “Did the Thatcher Reforms Change British Labour Market Performance?” CEP Discussion Papers no. 0168, Center for Economic Performance LSE. Calmfors, L. (1985) “Work Sharing, Employment and Wages,” European Economic Review 27, 293–310. Calmfors, L. and B. Holmlund (2000) “Unemployment and Economic Growth: A Partial Survey,” Swedish Economic Policy Review 7, 107–54. Calmfors, L., A. Forslund and M. Hemström (2001) “Does Active Labour Market Policy Work? Lessons from the Swedish experiences,” Swedish Economic Policy Review 5, 61–124. CEEP (2000) “Development of the Public Enterprises in Sweden,” The Swedish National Chapter to the CEEP Statistical Survey 2000. Crépon, B. and F. Kramarz (2002) “Employed 40 Hours or Not-Employed 39: Lessons from the 1982 Mandatory Reduction of the Workweek,” Journal of Political Economy 110(6), 1355–89. European Commission (2003) European Economy, Statistical Annex, Autumn 2003. European Economic Advisory Group at CESifo (2003) Report on the European Economy 2002, CESifo, Munich. Gersbach, H. and A. Schniewind (2001) “Awareness of General Equilibrium Effects and Unemployment,” IZA Working Paper no. 394. Hunt, J. (1999) “Has Work-Sharing Worked in Germany?” Quarterly Journal of Economics 114(1), 117–48. Joint Forecast of the Economic Research Institutes (Fall 2003) “Die Lage der Weltwirtschaft und der deutschen Wirtschaft im Herbst 2003,” ifo Schnelldienst 56(20), 3–41. Laroque, G. and B. Salanié (2000) “Une décomposition du non emploi en France,” Economie et statistique 331(1), 47–66 . Laroque, G. and B. Salanié (2002) “Labor Market Institutions and Employment in France,” Journal of Applied Econometrics 17, 25–48. Layard, R., S. Nickell and R. Jackman (1991) Unemployment – Macroeconomic Performance and the Labour Market, Oxford University Press, Oxford. Marimon, R. and F. Zilibotti (2000) “Employment and Distributional Effects of Restricting Working Time,” European Economic Review 44(7), 1291. Nickell, S. (2003) “Labour Market Institutions and Unemployment in OECD Countries,” CESIfo DICE Report 2, 2003. Nickell, S. and R. Layard (1999) “Labour Market Institutions and Economic Performance” in O. Ashenfelter and D. Card, eds, Handbook of Labor Economics, 3, Elsevier Science and Technology, North-Holland, Amsterdam, 3029–84. Pissarides, C. (1992) “Loss of Skill During Unemployment and the Persistence of Employment Shocks,” The Quarterly Journal of Economics 107(4), 1371–91.
9780230_547018_04_cha02.indd 80
3/19/2009 8:04:31 PM
Labour Market Reform in Europe
81
Saint-Paul, G. (1993) “On the Political Economy of Labor Market Flexibility,” NBER Macroeconomics Annual. Saint-Paul, G. (1996) “Exploring the Political Economy of Labour Market Institutions,” Economic Policy 23, 263–315. Saint-Paul, G. (1998) “A Framework for Analyzing the Political Support for Active Labor Market Policy,” Journal of Public Economics 67, 151–65. Saint-Paul, G. (2000) The Political Economy of Labour Market Institutions, Oxford University Press, Oxford. Sinn, H.-W. et al. (2003) “Welfare to Work in Germany: A Proposal on How to Promote Employment and Growth,” CESifo Research Report no. 1. Weitzman, M. (1984) The Share Economy: Conquering Stagflation, Harvard University Press, Cambridge, MA.
9780230_547018_04_cha02.indd 81
3/19/2009 8:04:31 PM
3 Pay-setting Systems in Europe: Ongoing Developments and Possible Reforms
During the last twenty years, Western Europe has been characterised by both high unemployment and low growth as compared to the United States. However, in the last decade, experiences have been diverse among European countries. Several smaller countries have managed to reduce unemployment substantially, whereas the large EU countries (France, Germany, Italy and Spain) have been less successful (see EEAG 2004, Chapters 1 and 2 of this book). As discussed at length in 2003 EEAG report (EEAG 2003), Germany is the EU country that has suffered the worst stagnation. Macroeconomic performance is intimately associated with the functioning of pay setting. It influences output and employment in a number of ways: ●
●
●
●
Aggregate real wage moderation, as was achieved in, for example, the Netherlands in the 1980s and the first half of the 1990s, is a precondition for high equilibrium levels of output and employment; that is for high average output and employment over the business cycle. Aggregate wage flexibility contributes to output and employment stabilisation in the case of macroeconomic shocks. Such flexibility has become even more important than before with the common currency, as changes in relative wage costs represent the only way of changing real exchange rates among countries in the euro area. Relative-wage flexibility is required to reduce labour market imbalances at sectoral, occupational and regional levels that otherwise raise equilibrium unemployment. Vivid illustrations of the importance of this factor are provided by Eastern Germany and Southern Italy, where a compression of the wage differentials to the other parts of the countries is a major cause of high unemployment. Pay differentials according to skills determine the returns to investment in education and training and thus the pace of human capital accumulation, which is a fundamental determinant of the rate of long-run growth. 82
9780230_547018_05_cha03.indd 82
3/19/2009 8:03:22 PM
Pay-setting Systems in Europe ●
83
The extent to which pay is linked to individual or group performance at the level of the enterprise and the workplace has an important influence on labour productivity.
As discussed in Chapter 2, pay setting and macroeconomic performance are affected by a number of basic “institutional” factors: the generosity of unemployment insurance, the scope and design of active labour market policy, the degree of employment protection, the extent of competition in product markets, and tax levels (see also Nickell and Layard 1999; Calmfors and Holmlund 2000). The way pay bargaining is conducted is also a fundamental determinant of macroeconomic performance. This chapter assesses the ongoing developments of pay-setting systems in the European countries, including the new EU member states, and gives recommendations on appropriate reforms. The chapter is structured in the following way. Section 1 reviews the present pay-setting practices in the European countries. Section 2 analyses the advantages and disadvantages of various systems. Section 3 discusses possible future developments. Section 4 offers some normative conclusions as to what pay-setting systems strive for.
3.1
A review of pay-setting systems in Europe
The most striking observation on pay-setting systems in Europe is probably their diversity and the number of country-specific features. There is no such thing as a uniform European model of wage setting. Still most countries show many similarities. To characterise pay setting, we focus below on two key aspects. The first is the importance of collective bargaining and trade unions. The second aspect concerns the level at which collective bargaining occurs (the degree of centralisation) and the extent of co-ordination among various bargaining units. 3.1.1 Unionisation and the coverage of collective agreements A key common feature in most Western European countries is the importance of collective agreements. As shown in Table 3.1, collective agreements cover over 60 per cent of all employees in most of the current EU countries, and in some of them (Austria, Belgium, Denmark, Finland, France, the Netherlands, Spain and Sweden) coverage is even above 75 per cent. In most cases, coverage is higher in the total economy than in the market sector. This reflects the fact that collective agreements usually cover a larger fraction of the labour force in the public than in the private sector. The coverage rates of collective agreements are much more similar among the current EU countries than unionisation rates. The latter vary from only 10 to 15 per cent in France and Spain to 70 to 90 per cent in Belgium and the Scandinavian countries. The similarity of coverage rates, despite the
9780230_547018_05_cha03.indd 83
3/19/2009 8:03:22 PM
84 Designing the European Model
large differences in union density, can be explained by various extension mechanisms: in some countries employers choose voluntarily to extend collective agreements to all employees. In others they are legally required to do so. Most present EU countries – though not the Scandinavian ones – also have legal provisions for extending sectoral collective agreements to nonunionised firms in the sector (Ebbinghaus and Visser 2000; Calmfors et al. 2001). The main outlier in Western Europe in terms of coverage of collective agreements is the United Kingdom, where the overall coverage rate in 2001 was estimated at 36 per cent. This reflects a process where coverage has fallen pari
Table 3.1
Coverage of collective agreements and unionisationa
Country
Old EU member states Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxemburg Netherlands Portugal Spain Sweden UK b New EU member states Cyprus Czech Republic Estonia Hungary Latvia Lithuania Malta Poland Slovakia Slovenia
Total economy (2001)
Market sector (mid 1990s)
Coverage
Unionisation
Coverage
Unionisation
98 100 85 90 90 67
40 69 88 79 9 30
97 82 52 67 75 80
34 44 68 65 <4 25
32
60 78 62 81 94 36
35 50 27 30 15 79 29
65–70 25–30 29 34 < 20 10–15 60–70 40 48 100
70 30 15 20 30 15 65 15 40 41
43 36 79 80 67 72 35
19 < 20 < 15 77 19
Continued
9780230_547018_05_cha03.indd 84
3/19/2009 8:03:22 PM
Pay-setting Systems in Europe
Table 3.1
85
Continued
Country
Total economy (2001) Coverage
Unionisation
22 (23)c 32
23 30g
Japan
21
22h
New Zealand
45d
22
70–77e 53f 15
Other countries Australia Canada
Norway Switzerland US
Market sector (mid 1990s) Coverage
Unionisation
21
24
55h 23h
62 50
44 22
14h
13
10
Notes: a Coverage refers to the percentage of employees covered by collective agreements and unionisation to the percentage of employees with union membership; b Figures do not include Northern Ireland; c The parenthesis refers to the coverage of wage awards (see Section 1.1) and to 2000; d 1997; e 2000–01; f 1994; g 2000; h 1996–98. Sources: Industrial Relations in the EU Member States and Candidate Countries (2002) for the total economy and Ebbinghaus and Visser (2000) for the market sector in most cases. The sources for the total-economy coverage data are Australian Workplace (2003) for Australia, http://www.gov. nf.ca/labour/unionization_rates (2002) for Canada, Industrial Relations in the EU, Japan and the US (2002) for Japan and the US, Bray and Walsh (1998) for New Zealand, http://www.eiro.eurofound. ie/2002/12/study/tn0212102s for Norway, and Ochel (2001) for Switzerland. The sources for totaleconomy unionisation data are Ebbinghaus and Visser (2000) for Japan, Norway, and the US, OECD for Canada and Switzerland, and ERA (2003) for New Zealand. Differences among countries should be taken only as broadly indicative, as data are not always exactly comparable.
passu with unionisation over the last twenty years. According to Brown et al. (2000), around half of all employees and around two-thirds of employees in the private sector now have their wages set unilaterally by employers in the United Kingdom. The development in New Zealand, and also in Australia, has been similar to that in the United Kingdom: both coverage of collective agreements and unionisation have fallen dramatically, although the development in these countries occurred mainly in the 1990s. For New Zealand, Bray and Walsh (1998) reported that in the mid-1990s around 50 per cent of all employment contracts were between an individual employee and an employer. The developments in the United Kingdom, New Zealand and Australia imply that the industrial relations systems there have approached the US system. In most Western European countries, except the United Kingdom, coverage rates have remained quite stable, even though there have been significant reductions in unionisation rates in most of them over the last two decades (the exceptions are the Scandinavian countries and Belgium; see Ebbinghaus and Visser 2000; Calmfors et al. 2001; Ochel 2001). The average union density in Western Europe declined from 44 per cent in 1979 to 32 per cent in 1998. It is evident from Table 3.1 that the new EU member states are very different from the present ones, in terms of both coverage of collective agreements
9780230_547018_05_cha03.indd 85
3/19/2009 8:03:22 PM
86
Designing the European Model
and unionisation. With the exception of Slovenia (with almost 100 per cent coverage), Cyprus and Malta, coverage is much lower than in Western Europe. Unionisation rates are also very low in some of the new member states. In Estonia, Lithuania, Poland and Hungary union density is around 20 per cent or lower. This reflects to a large extent the difficulties of the old trade unions in these countries, which did not play a wage-bargaining role under communism, to adapt and obtain legitimacy with employees as well as the difficulties of building up new unions (Flanagan 1998). 3.1.2 Bargaining levels Table 3.2 gives a summary picture of how collective bargaining is conducted in different countries. Again there are large differences among countries, to a large extent along the same lines as the differences in the coverage of collective agreements. In the present EU countries, the most important bargaining level is usually the sectoral one. This applies to Austria, Denmark, Germany, Greece, Italy, the Netherlands, Portugal, Spain and Sweden. In several of these countries (primarily Austria, Denmark, Germany and Sweden), there is a strong element of pattern bargaining, with the engineering sector usually acting as a wage leader, setting the pace for the whole economy. The standard situation in the present EU member states is one of multilevel (or at least two-tier) collective bargaining, where multi-employer bargaining at the sectoral level is complemented to a smaller or larger extent by bargaining at the enterprise level on the implementation of the sectoral agreements. The general trend over the last twenty years has been that the scope for local bargaining has increased (Traxler et al. 2001; Calmfors et al. 2001; Traxler 2003). This picture is confirmed by the development of various measures of bargaining co-ordination that seek to measure the importance of the different bargaining levels (see also Section 2.1). According to Visser (2000), co-ordination decreased from the mid-1970s to the mid-1990s in ten out of thirteen EU countries examined.1 A similar picture is given by Ochel (2000), who found that co-ordination decreased in eight out of thirteen EU countries between 1975–79 and 1995–99. 2 However, the only current EU country where single-employer bargaining at the local level has become completely dominating is the United Kingdom, where multi-employer bargaining at the sectoral level has almost ceased to exist (Brown et al. 2000; Nickell and Quintini 2002). Again, similar developments have occurred in New Zealand and Australia, where the earlier systems of letting government tribunals determine sector, and occupationbased national wage awards were replaced in the 1990s by a system of basically single-employer bargaining (Honeybone 1997; Bray and Walsh 1998). France is also a country where the enterprise is the most important bargaining level (in recent years stimulated by government financial incentives for firms to conclude local collective agreements on working time reductions), although bargaining at the enterprise level coexists with bargaining at the sectoral level (Jefferys 2000; Dufour 2003; Eironline France 2003).
9780230_547018_05_cha03.indd 86
3/19/2009 8:03:23 PM
Pay-setting Systems in Europe Table 3.2
87
Bargaining levels
Country
National guidelines
Inter- Sectoral Enterprise sectoral level level level
Old EU member states Austria Belgium
Denmark Finland France Germany Greece Ireland Italy
Luxemburg Netherlands
Portugal Spain Sweden
Pattern bargaining Centrally agreed guidelines for wage increases with the government 2003–04 Pattern bargaining Tripartite national pay agreement 2003–04 Pattern bargaining National general collective agreement 2002–03 Tripartite national pay agreement 2003–04 Social pacts with government 1993 and 1998 setting guidelines for the wagebargaining process
XXX
XXX X
X X
XX XXX
XX XX
X X
XX
X XXX XXX
XX X X
XXX
X
X
XX
X
Centrally agreed ceiling for wage increases with government 2003; tripartite national wage freeze 2004–05
XX
XX XXX
XX X
Centrally agreed guidelines for wage increases 2003
XX
XXX XXX
X X
XXX
XX
X
XXX
XXX X
X XXX
X
XXX
XX
XXX
Intersectoral agreements setting guidelines for the wagebargaining process; pattern bargaining
UK New EU member states Cyprus Czech Republic Estonia Hungary
Tripartite national agreements on minimum wages Tripartite national agreements on minimum wages National guidelines for wage increases agreed with government and tripartite national agreements on minimum wages
X
Continued
9780230_547018_05_cha03.indd 87
3/19/2009 8:03:23 PM
88 Designing the European Model
Table 3.2
Continued
Country
National guidelines
Latvia
Tripartite national agreements on minimum wages
Lithuania Malta Poland
Slovakia Slovenia Other countries Australia Japan New Zealand Norway
Inter- Sectoral Enterprise sectoral level level level X
X
XXX
X X
XXX XXX XXX
XX
X
XXX
XX
X
X
XX
XXX
XX
X XXX
XXX X
X
XX XXX
National guidelines for wage increases agreed with government and tripartite national agreements on minimum wages Tripartite national agreements on minimum wages Tripartite national pay bargains National wage awards for minimum wages Pattern bargaining Pattern bargaining; tripartite agreement on guidelines for wage increases 2003
Switzerland US
XXX
Notes: XXX = dominating level. XX = important, but not dominating level. X = existing level. Sources: Industrial Relations in the EU Member States and Candidate Countries (2002), Collective Bargaining Coverage and Extension Procedures (2002), individual Eironline country reports. For New Zealand: Bray and Walsh (1998).
The enterprise is the dominating bargaining level in most of the new EU member states (the Czech Republic, Estonia, Hungary, Malta, Poland, Latvia and Lithuania). The sectoral level is dominating in Cyprus and of substantial importance in the Slovak Republic and Slovenia. One important explanation of the limited role of sectoral bargaining in the former communist countries is that it has proved difficult to organise employers’ associations in a situation of rapid structural change, where many old (state) firms have been closed down or privatised and many new firms have been started (Flanagan 1998). Perhaps somewhat surprisingly, the trend towards greater importance of the local bargaining level has in many of the current EU countries occurred at the same time as there have been social pacts between the peak-level labour
9780230_547018_05_cha03.indd 88
3/19/2009 8:03:23 PM
Pay-setting Systems in Europe
89
market organisations, sometimes also involving the government as a third actor, setting national norms for pay increases to be negotiated at the sectoral level. Indeed, such agreements became more frequent in the pre-EMU period in the 1990s as a means to promote real wage restraint without resorting to inflation and exchange rate depreciations. Such income policies have also been used frequently in some EMU countries after 1999 as a means to restrain wage increases.3 Table 3.2 also gives an overview of the extent to which such attempts at co-ordination of wage increases have been made recently. The most comprehensive central agreements in force are probably those in Finland and Ireland, where the peak-level labour market organisations agreed low wage increases with the government for 2003–04 in exchange for tax cuts (Finland) or other changes in government policies (Ireland). Tripartite bargaining has also taken place in the Netherlands. For 2003, the peak-level labour market organisations agreed on a ceiling for wage increases in exchange for cuts in taxes and social insurance contributions (Eironline Netherlands 2002a) and for 2004–05 they accepted a pay freeze in exchange for a government commitment to refrain from reductions of certain expenditures. Another recent example of tripartite bargaining is Belgium, where a central “indicative norm” on wage increases was negotiated for 2003–04, at the same time as the government undertook to implement certain employment measures.4 National central agreements on wage increases have also been negotiated between peak-level employer organisations and union confederations in Greece and Spain. Italy and Sweden provide examples of weaker co-ordination efforts, where earlier social accords (Italy) or intersectoral agreements (Sweden) established guiding principles for wage setting and bargaining procedures.5 Germany provides an exception to the pattern described, as the attempts there of achieving co-ordinated wage restraint through social pacts (Bündnis für Arbeit) have been largely unsuccessful. Social pacts have also played a much smaller role in the new EU countries than in the old ones. Slovenia is the only new member country where tripartite national pay bargains have a dominating influence, whereas national guidelines agreed centrally with the government play some role in Hungary and Poland. In the other ex-socialist countries joining the EU, tripartite bargaining at the national level applies only to minimum wages.
3.2 Advantages and disadvantages of various pay-setting systems There is a large literature on the advantages and disadvantages of various pay-setting systems. This literature has emphasised the effects on: ● ●
the aggregate (equilibrium) wage and employment levels the flexibility of the aggregate wage level
9780230_547018_05_cha03.indd 89
3/19/2009 8:03:23 PM
90 Designing the European Model ● ●
relative wages and wage dispersion incentives for effort and productivity
3.2.1
The aggregate wage and employment levels
Much of the literature on pay-setting systems has focused on the determination of the aggregate wage level. This literature has usually tried to distinguish between, on the one hand, the effects of collective bargaining and unionisation per se, and, on the other hand, the effects of different degrees of co-ordination/centralisation of the collective bargaining that takes place. Theoretical modelling of trade union behaviour usually assumes that unions strive for real wages that trade off the benefits of a wage increase for employed members against the income (and utility) loss of those members who may become unemployed because of the wage rise. The bargaining process between unions and employers is modelled as providing a negotiated wage that balances unions’ wage objectives against employers’ interests in high profits (see, for example, Nickell and Layard 1999 or Calmfors and Holmlund 2000). A well-known argument based on such an analytical framework is that a high degree of co-ordination of collective wage bargaining (which may come about either because bargaining is formally conducted at a high level, such as the national one, or because separate bargaining units co-ordinate their actions at that level) promotes real wage restraint because it allows negative externalities of high wage levels for individual bargaining areas to be internalised. Wage setters will take into account that high wages for one group may have a negative impact on other groups. This could occur for several reasons (Calmfors 1993): ● ● ●
●
●
because the aggregate consumer price level is pushed up; because the prices of inputs to other production sectors are increased; because high wages in one sector cause job losses there that reduce employment opportunities for everyone in the economy; because such job losses raise costs of unemployment benefits and reduce the tax base; or because high wages for one group can give rise to pure envy effects.
For these reasons, one should expect collective bargaining at the sectoral level to result in higher real wage levels, and thus also higher unemployment, than co-ordinated multi-sector bargaining, as negative externalities will be internalised to a lesser extent. This hypothesis receives strong empirical support from attempts to relate unemployment differences among countries to differences in labour-market institutions, as is illustrated in Table 3.3 (see also Calmfors et al. 2001; Calmfors 2001).6
9780230_547018_05_cha03.indd 90
3/19/2009 8:03:23 PM
Pay-setting Systems in Europe
91
It has been more difficult to show in time series wage equations for individual countries that periods with highly co-ordinated collective bargaining have been associated with wage moderation (see, for example, Hartog 1999 on the Netherlands or Walsh 2002 on Ireland), although Koskela and Table 3.3 Unemployment rates under various bargaining regimes (ceteris–paribus differences to decentralised systems) in various studiesa Study
Intermediate co-ordination
High Measure of co-ordination bargaining structureb
A: Studies finding a hump-shaped relationship between bargaining co-ordination and unemployment 1 2
Zetterberg (1995)c Bleaney (1996)d
2.6 3.5
1.5 2.1
3 4
Scarpetta (1996)e Elmeskov et al. (1998)f Elmeskov et al. (1998)g Elmeskov et al. (1998)h Cukierman & Lippi (1999)i Daveri & Tabellini (2000)j Nicoletti et al. (2001)l Average
0.9 1.3
12.0 2.4
1.2
4.4
6.9
4.6
5.8
3.2
Centralisation
5.8
7.2
Geographicalk
3.6
2.2
Centralisation/ co-ordination
3.5
3.9
5 6 7 8 9
Centralisation Centralisation/ co-ordination Centralisation Centralisation Centralisation/ co-ordination Co-ordination
B: Studies finding a monotonic relationship between bargaining co-ordination and unemployment 1 2 3 4 5 6 7 8
Layard et al. (1991) Zetterberg (1995)m Scarpetta (1996)n Bleaney (1996)o Elmeskov et al. (1998)p Hall & Franzese (1998)q Iversen (1998)r Nickell & Layard (1999)s
4.7 0.4 6.2 2.0 0.8
10.4 2.4 12.3 3.9 5.7
Co-ordination Centralisation Co-ordination Co-ordination Co-ordination
2.6
5.1
Co-ordination
3.3 4.6
4.1 6.0
Centralisation Co-ordination
Continued
9780230_547018_05_cha03.indd 91
3/19/2009 8:03:23 PM
92
Designing the European Model
Table 3.3
Continued
Study
9 10 11 12
Blanchard & Wolfers (2000)t Belot & van Ours (2001)u Belot & van Ours (2001)v Nickell et al. (2003)x Average
Intermediate co-ordination
High co-ordination
Measure of bargaining structureb
4.4
8.9
Centralisation
2.6 (0)
5.2 (0)
Co-ordination
1.9
1.9
Co-ordination
7.2 3.4
14.4 6.7
Co-ordination
Notes: a The table shows how the unemployment rates under intermediate and high co-ordination/centralisation differ from that under decentralisation/low co-ordination when other factors are controlled for. High co-ordination can be interpreted as bargaining/co-ordination at the national level, intermediate co-ordination as bargaining/co-ordination at the sectoral level and low coordination/decentralisation as uncoordinated bargaining at the enterprise level. b Measures of centralisation capture the level at which actual bargaining takes place. Measures of co-ordination try to capture informal co-ordination among formally independent bargaining units as well. c Equation (5) in Table 4.14. We have classified the countries ranked 1–3 and 7–9 as centralised, the countries ranked 13–17 as intermediately centralised and the countries ranked 4–6 and 10–12 as decentralised. d Equation (4) in Table 5. Bleaney distinguishes between highly centralised systems, highly decentralised systems, moderately centralised systems with a high degree of corporatism and moderately centralised systems with a low degree of corporatism. In the table, the last two categories have been amalgamated to one. e Equation (8) in Table 1. The entry for intermediate centralisation refers to the country ranked 14 and the entry for co-ordination to the country ranked 1. The comparison is with the country ranked 17. f Equation (2) in Table 2. g Equation (4) in Table 2. h Equation (4) in Table 4. In the equation, taxes and bargaining co-ordination are interacted. The effects are evaluated at the average tax ratio for the sample period 1983–95. i Equation 4.5 in Table 4.2. In the regression, centralisation is interacted with central bank independence. The effect in the table is evaluated at the sample average for the central bank independence variable. At high levels of central bank independence the hump-shaped relationship between centralisation and unemployment turns into a positive (!) monotonic one (that is with higher centralisation being associated with higher unemployment). j Equation (5) in Table 9. In the equation, taxes and bargaining co-ordination are interacted. The effects are evaluated at the average tax ratio for 1983–95. k This study associates the Scandinavian countries with high co-ordination, the European continental countries with intermediate co-ordination, and the Anglo-Saxon countries with low co-ordination/decentralisation. l The dependent variable is non-employment and not unemployment. The study has not been included when computing the averages. m Equation (3) in Table 4.14. The countries ranked 1–5 are classified as highly co-ordinated, the countries ranked 6–10 as intermediately co-ordinated, and the countries ranked 11–17 as uncoordinated/decentralised. n Equation (2) in Table 1.
9780230_547018_05_cha03.indd 92
3/19/2009 8:03:23 PM
Pay-setting Systems in Europe
93
o
Equation (1) in Table 5. Equation (1) in Table 2. q Decade equation in Table 2. Countries have been assigned co-ordination values from 0 (lowest coordination) to 1 (highest co-ordination). In our calculation, high co-ordination is 0.875, intermediate co-ordination is 0.5, and low co-ordination 0.125. In the regression the co-ordination variable is interacted with a central bank independence variable ranging from 0 to 1. In the calculation in the table the central bank independence variable has been assigned the value 0.5. r Full model in Table 2. High co-ordination refers to the average score for the three countries with the highest co-ordination, low co-ordination to the three countries with the lowest coordination, and intermediate co-ordination to the halfway distance between the two groups. In the regression the co-ordination variable is interacted with a central bank independence variable ranging from 0 to 1. In the calculation in the table the central bank independence variable has been assigned the value 0.5. Note that the monotonic relationship does not hold at high degrees of central bank independence. Then the relationship is instead u-shaped. s The equation explains the log of the unemployment rate. In the calculation of the effect on the unemployment rate, it has been assumed that unemployment under decentralisation is equal to the average rate of unemployment among the countries studied during the estimation period. t Equation (1) in Table 1. In the equation, macroeconomic shocks and bargaining co-ordination are interacted. The entries show the differences in the increase of unemployment between the 1960–65 period and the post-1995 period. u Equation (1) in Table 3a. The equation has been estimated without fixed country and time effects. With such effects, there are no significant unemployment differences among various bargaining regimes as indicated by the zeros in the parentheses. v Equation (2) in Table 3b. This equation interacts various labour market institutions with each other (for example employment protection and co-ordination as well as union density and coordination). The effects in the table are evaluated at the average values of employment protection and union density in 1960–94. The equation has been estimated with fixed country and time effects. x Equation (1) in Table 5. This equation interacts various institutional variables with each other (for example union density and co-ordination as well as the employment tax rate and coordination); the effects in the table are evaluated at the average sample values of union density and tax rates. The effects in the table are steady-state effects. p
Source: Calculations by the EEAG.
Uusitalo (2003) – after controlling for other factors – report significantly lower wage increases in Finland in years with centralised wage agreements than in years when the main locus of bargaining has been the sectoral level. It is less clear, both theoretically and empirically, how decentralised single-employer collective bargaining at the enterprise level compares with multi-employer bargaining at the sectoral level. If various degrees of internalisation of negative externalities were the whole story, then enterprise-level bargaining would imply an even higher aggregate wage level than sector-level bargaining, since the degree of internalisation is lower. In this case there is thus a monotonic negative relationship between, on the one hand, the degree of co-ordination and, on the other hand, the real wage and unemployment (curve I in Figure 3.1). But enterprise-level bargaining may also have a wage-restraining effect because it increases competitive pressures to restrain wages. These pressures may be weak in sectoral bargaining, since all domestic competitors in a sector are then
9780230_547018_05_cha03.indd 93
3/19/2009 8:03:23 PM
Real wage (unemployment)
94 Designing the European Model
II
I
Low (firm level)
Intermediate (sectoral level)
High (national level)
Bargaining co-ordination (centralisation)
Figure 3.1 The relationship between bargaining co-ordination (centralisation) of collective bargaining and the real wage (unemployment)
exposed to similar wage increases, which makes it easy to shift them on to prices. If the competition effect due to loss of market power outweighs the reduction in the internalisation of negative externalities, enterprise bargaining leads to lower wage levels than sectoral bargaining. In that case, there is a hump-shaped relationship between, on the one hand, the degree of co-ordination of collective bargaining and, on the other hand, the levels of real wages and unemployment (curve II in Figure 3.1).7 The height of the hump depends on the extent of competition from abroad. Because sectoral bargaining only reduces competitive pressures from other domestic producers, the hump will be less pronounced the larger is the degree of competition from abroad (Danthine and Hunt 1994; Calmfors et al. 2001). The empirical research on the determination of aggregate unemployment in Table 3.3 does not give a clear verdict on how collective bargaining at the enterprise level compares with sectoral bargaining. Almost half the studies indicate that decentralisation to the enterprise level results in lower unemployment than sectoral bargaining, whereas the other half gives the reverse result. As can be seen, there is a tendency in the studies to find more support for the hump-shape hypothesis with “centralisation measures,” focusing on the actual level of bargaining, than with “co-ordination measures,” attempting to capture also informal co-ordination among separate bargaining units. But the studies finding a hump-shaped relationship also
9780230_547018_05_cha03.indd 94
3/19/2009 8:03:23 PM
Pay-setting Systems in Europe
95
Table 3.4 Unemployment rates under various rates of union density and coverage of collective agreements (ceteris–paribus differences to 15% union density or coverage) in different studiesa Study 1 2 3 4 5 6
Layard et al. (1991) Scarpetta (1996)b Elmeskov et al. (1998) Hall & Franzese (1998) Iversen (1998) Nickell & Layard (1999)c
7 8 9 10 11
Nickell & Layard (1999)e Nicoletti et al. (2001)e) Belot & van Ours (2001)f Belot & van Ours (2001)g Nickell et al. (2003)h
45%
75%
2.5 1.8 0 0 0 2.8 3.7 6.5d 2.4 2.1 1.8 (0) 4.7 0 (2.1)
4.9 3.6 0 0 0 6.5 9.0 15.5d 4.8 4.2 3.6 (0) 9.4 0 (4.2)
Explanatory variable Coverage Union density Union density Union density Union density Coverage Union density Total Union density Union density Union density Union density Union density
Notes: aThe table shows how much higher the unemployment rate is at 45% and 75% density or coverage rates compared to 15% density or coverage rates when other factors are controlled for. b Equation (2) in Table 2. c The equation explains the log of the unemployment rate. In the calculation of the effect on the unemployment rate, we have assumed that unemployment at 15% density and coverage rates is equal to the average rate of unemployment among the countries studied during the estimation period. d The sum of coverage/density effects. e The dependent variable is non-employment and not unemployment. f See footnote (u) to Table 3. g Equation (2) in Table 3B. The equation interacts various labour market institutions with each other (for example union density and co-ordination). The entries in the table refer to the effects of changes in union density under decentralisation. At higher levels of co-ordination there are no significant effects. h Equation (1) in Table 5. The figures not in parenthesis are long-run effects. The figures in parenthesis are impact effects. The regression equation interacts union density and co-ordination. The effects in the table are evaluated at the sample average of co-ordination. Source: Calculations by the EEAG.
agree that bargaining under high centralisation/co-ordination (at the multisector or national level) is associated with lower unemployment than bargaining under decentralisation/low co-ordination (at the enterprise level). There is, however, a consensus in the empirical literature that less importance of collective bargaining and unions per se – given the degree of coordination of the collective bargaining that takes place – is associated with lower wages or factors likely to be correlated with lower wages. Table 3.4 reports estimates of the effects on unemployment from cross-section and panel data sets of countries. There is also a large body of microeconomic evidence indicating that higher unionisation and coverage of collective agreements lead to higher wages and lower profitability of firms (see, for example,
9780230_547018_05_cha03.indd 95
3/19/2009 8:03:23 PM
96
Designing the European Model
Addison and Hirsch 1989; Belman et al. 1997; Kleiner 2001 or Askildsen and Nilsen 2002). This is, of course, to be expected, since the main raison d’être of trade unions is to push up wages and various employee benefits above the levels that would otherwise prevail. An important conclusion is that the chances that decentralisation of collective bargaining is associated with wage restraint increase if such a development is associated with a reduction in unionisation and the coverage of collective agreements. This is all the more so, as there is some evidence that low unionisation has a larger unemployment-reducing effect under decentralised bargaining than under intermediate or high centralisation/coordination: for example, Belot and van Ours (2001) find that variations in unionisation only affect unemployment under decentralised bargaining. EMU and different bargaining structures An issue that has received much interest in the recent research literature on bargaining co-ordination is how monetary unification in Europe affects wage outcomes. One strand of literature has focused on the interaction between the central bank and wage setters in an economy with monetary policy autonomy, that is a flexible-exchange rate economy outside the EMU (Coricelli et al. 2000; Soskice and Iversen 2000). The starting point is that, provided there is some co-ordination of collective bargaining, wage setters could be expected to act strategically and take into account the anticipated responses of the central bank to wage settlements. More precisely, central bank behaviour can act as a deterrent to high wages, because wage setters realise that wage increases that threaten the price stability goal of the central bank will trigger an interest rate rise that raises the cost of high wage increases in terms of employment losses (the effective elasticity of labour demand). For a country outside EMU, with a flexible exchange rate and an inflation-targeting central bank, like Sweden, an extra incentive for wage restraint is thus added under these conditions. But this incentive does not work for the EMU countries, as national wage setters will always be too small relative to the ECB to be able to trigger any monetary policy reaction. So, one could argue that monetary unification weakens the incentives for wage restraint when there is some co-ordination of collective bargaining. The argument has been made that the effect is larger with only intermediate co-ordination (bargaining at the sectoral level) than with high co-ordination. The reason is that in the latter case the incentives for wage moderation may be very strong anyway because of various internalisation effects, so the potential deterrence role of monetary policy may then not matter much anyway (Soskice and Iversen 2000; Holden 2001).8 But with weaker internalisation effects in the case of intermediate co-ordination, the loss of a deterring national monetary policy in EMU is more serious. This might be one explanation of why it may be difficult for Germany – with its heavy reliance
9780230_547018_05_cha03.indd 96
3/19/2009 8:03:24 PM
Pay-setting Systems in Europe
97
on sectoral bargaining – to achieve sufficient real wage moderation (see also Soskice and Iversen 1998; Hall and Franzese 1998). 3.2.2
Aggregate wage flexibility
Another aspect of pay setting concerns the degree of aggregate nominal wage flexibility as an adjustment mechanism in the case of macroeconomic shocks. This aspect has become more important than before in the euro zone, as there is no longer any nominal exchange rate channel for changing real exchange rates among the member countries. An old argument already dates back to Keynes (1936), who argued that the concern of employees over relative wages would make them oppose money wage reductions, unless all wages were cut simultaneously so as to preserve existing wage relativities. A more modern version of this co-ordination failure argument is that because of product demand interrelationships, the benefit of changing wages (and thus also prices) in individual firms depends on whether or not other firms do the same (Ball and Romer 1991). With small demand shocks, adjustment costs may make it unprofitable for each firm to change the wage independently of what happens in other firms. With very large shocks, it will always pay to adjust the wage, even if others do not. But for shocks of intermediate size, the individual wage setter may gain from adjusting only if others do the same. This could give rise to multiple equilibria: which one materialises will then depend on the expectations of what other wage setters will do. Co-ordination of pay setting is a way of removing such indeterminacy and ensuring that the economy ends up in a good equilibrium in which wages adjust. A similar argument can be made with respect to the length of wage contracts, which is an important determinant of nominal wage flexibility. In systems with decentralised and unsynchronised wage setting, contract length may be chosen in a socially inefficient way (Ball 1987). Most notably, there is an aggregate-demand externality: wage setters in an individual bargaining area do not take into account that a long-term wage contract on their part may contribute to aggregate demand fluctuations in the economy. The reason is that money wage stickiness in a part of the economy means lower flexibility of the aggregate price level in the case of nominal shocks. With co-ordination of wage setting, wage setters have an incentive to internalise this externality. This effect works towards shorter wage contracts and thus more nominal wage flexibility, when there is more co-ordination.9 At the same time, it should be recognised that decentralised wage setting is easier to combine with various systems of performance-related pay (see Section 2.4), including profit-sharing arrangements, which increase aggregate pay flexibility. Also, lower unionisation and less importance of collective agreements are likely to make labour markets conform more closely to the textbook competitive model, with demand and supply factors exerting a more direct influence. Overall, however, there has been little empirical
9780230_547018_05_cha03.indd 97
3/19/2009 8:03:24 PM
98 Designing the European Model
research on the importance of various bargaining systems for nominal wage flexibility. 3.2.3 Relative wages and wage dispersion The determination of relative wages is an equally important aspect of wage setting as the determination of the aggregate wage level. Aggregate real wage restraint is not enough for achieving low unemployment if there are large demand and supply imbalances between regions, sectors, occupations and skills. Trade unions typically regard a “fair distribution of wages” as a central goal. Sometimes the objective is formulated as “equal pay for equal work,” but many times the objective is the even more far-reaching one of general pay compression. This latter objective has been particularly important in the Scandinavian countries and Italy.10 Reducing pay compression is a rational objective for a trade union if the utility gain for an employer member is seen as larger the lower is the initial wage. Then a union will be more inclined to trade off wage increases against employment losses for low-wage than for high-wage members. Another factor working in the same direction is that the income loss in case the wage rise makes a member unemployed (the wage when employed minus unemployment compensation) is smaller the lower the initial wage is (provided that unemployment compensation is fixed and thus independent of the wage). 2.5 D9/D5 1.5
0.5
0.5 D5/D1
1.5
Figure 3.2
c
Au Sw stri ac itz er la N nd a et he r la nd sd G er m an yd Au st ra lia a Ja pa nb Fr a n N ce c ew Ze U ni al te an d db Ki ng do U ni ma te d St at es a
c
Ita ly
nl Fi
Sw ed e
n
c
an d
2.5
Earnings dispersion
Notes: D9, D5 and D1 refer to earnings deciles. a) 1998; b) 1997; c) 1996; d) 1995. Source: OECD database on earnings deciles; calculations by the EEAG.
9780230_547018_05_cha03.indd 98
3/19/2009 8:03:24 PM
Pay-setting Systems in Europe
99
There exists overwhelming evidence that high degrees of unionisation and coverage of collective agreements compress the wage structure in all dimensions. It is also well known that higher degrees of centralisation of collective bargaining reduce wage dispersion (see, for example, Rowthorn 1992 or Wallerstein 1999). This is illustrated in Figure 3.2, which shows that earnings dispersion is higher in New Zealand, the United Kingdom and the United States, where multi-employer bargaining is rare and rates of unionisation and coverage of collective agreements are lower than in other countries. Reductions in wage dispersion in centralised collective agreements seem to come about mainly through higher wages at the lower end of the wage distribution. For example, Blau and Kahn (1996) found that centralisation of bargaining reduces wage dispersion by narrowing the wage differential between the 50th and 10th per centile, but has no significant effect on the differential between the 90th and 50th per centile. It is also striking how the large increases in wage dispersion in the United Kingdom over the last twenty years and in New Zealand over the last ten years have coincided in time with de-unionisation, reduced coverage of collective agreements in general and virtual disappearance of multi-employer bargaining. A clear-cut example in Europe of how wage policies aimed at compressing wage differentials across regions can contribute to high unemployment is Germany. As has been reported by the German Council of Economic Experts (2002), the lowest union wages in West German industry rose much more than average wages over the period from 1970 through 1999. This was a direct implication of the unions’ policy of negotiating for lump sum rather
Relative wage, relative productivity in % 90 80 Relative wage
70 60
Relative productivity
50 40 30 20 10 0
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Figure 3.3 Relative wage per employee and relative productivity per employee in Eastern versus Western Germany Source: Federal Statistical Office, Germany.
9780230_547018_05_cha03.indd 99
3/19/2009 8:03:24 PM
100 Designing the European Model Unemployment rate in % 20 18 16 Eastern Germany
14 12 10 8 6
Western Germany
4 2 0
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 Figure 3.4
Unemployment rates in Eastern and Western Germany
Source: Federal Employment Services, Germany.
than proportional wage increases. While the unemployment rates among people with high and average skills have not risen much in Germany, the unemployment rate of those who have neither vocational training nor a university degree rose from about 6 to 22 per cent in the period from 1975 to 2001.11 The most adverse consequences of wage compression in Germany have occurred in Eastern Germany. Wages per hour increased from 7 per cent of wages in Western Germany in 1989, before unification, to about 72 per cent in 2002 (Sinn 2003). The relative wage per employee in 2002 amounted to 77 per cent, while aggregate productivity (GDP per employee) had only reached a level of 57 per cent of the level in Western Germany (see Figure 3.3). This is a major cause of the high unemployment in Eastern Germany (currently around 18 per cent, see Figure 3.4). 3.2.4 Performance-related pay During the last two decades, it has been a general trend in most countries for firms to increasingly use pay systems to stimulate both individual and group effort of employees. This trend towards more incentive pay conflicts with traditional union policies of pay compression, as described in Section 2.3. In most Western European countries, union policies have had a large influence on the possibilities of differentiating pay both among and within firms. Sectoral wage bargaining has limited the possibilities of differentiating pay depending on individual company performance. This has been quite a binding constraint, especially in countries like Germany and Austria, where local bargaining about the implementation of sectoral agreements has played only a limited role. Within firms, collective agreements
9780230_547018_05_cha03.indd 100
3/19/2009 8:03:24 PM
Pay-setting Systems in Europe
101
traditionally set standard pay rates for broadly defined groups or tasks based more on formal criteria than on individual performance. An important objective for each firm is to align the incentives of the employees with those of the firm to stimulate high performance. Profit maximisation requires that this is done in as cost-efficient a way as possible. One method is close monitoring of employees combined with threats of dismissal or missed career opportunities. But monitoring is efficient as an incentive device only if the probability of discovering “shirking” is high (Nalbantian and Schotter 1997) and may therefore be costly. It will be the more so the more complex the relationship is between employee work and firm performance. Performance-related pay is therefore likely to be part of any optimal incentive package. The optimal mix of individual pay incentives and group incentives, such as profit-related pay or employee stock ownership plans, depends on the character of production. The advantages for the employer of incentive pay linked to group rather than individual performance are greater when (i) capital intensity is high and the risks of misuse of capital equipment is large; (ii) when labour inputs are strongly complementary, so that the pay-off from encouraging co-operation among employees is high and the potential disruptive costs from conflicts over the use of differentiated pay are large; and (iii) when it is difficult to measure individual output (which explains why profit sharing and employee stock ownership plans are common in sectors with complex employee tasks and rapidly changing technology; Kruse 1996). Another advantage of profit-related pay is that it introduces more flexibility of the average pay level in the firm, as discussed in Section 2.2. One should therefore expect profit sharing to be common in firms with a high variability of performance and in new firms with uncertain prospects, which is an empirical prediction that seems to be borne out by the data (Kruse 1996). The benefits of incentive pay seem increasingly to have been realised by employers over the last twenty years. So-called Human Resource Management practices have been adopted by a large number of firms (Lazear 1999). These practices have been described as aiming to substitute co-operative employer-employee relationships, focused on “managing the contribution and commitment of individual employees,” for more collectivist and adversarial industrial relations (see, for example, Godard and Delaney 2000 or Roche 2001). Flexible work assignment, cross training, teamwork and formal employee participation are typical elements of human resource management policies. Performance-based pay is also seen as a key ingredient (see, for example Lazear 1999; Ichniowski et al. 1996; Godard and Delaney 2000 or Rayton 2003). Not very surprisingly, performance-based pay has been found to be associated with superior economic performance in empirical studies. For example, the profit-sharing literature has found significant links between firm performance and profit-related pay (Kruse 1993). A widely quoted study by
9780230_547018_05_cha03.indd 101
3/19/2009 8:03:24 PM
102
Designing the European Model
Ichniowski et al. (1997) of productivity in steel finishing lines concluded that “innovative work practices” (incentive pay, team structures and flexible job assignments) lead to higher performance than “traditional work practices” (narrow job specifications, close supervision, hourly pay and strict work rules). A very recent study by Rayton (2003) of US firms finds that average employee compensation is significantly positively related to firm performance (in terms of stock returns and bond ratings) in high-performance firms, whereas there is no significant relationship between average employment compensation and performance in low-performance firms. Decentralisation of pay setting to the level of the individual firm makes it easier in general to design incentive pay systems that are appropriate for the specific activities performed. Also average employee compensation can more easily be linked to firm performance in this way, because benefits in the form of profit sharing or employee stock ownership plans can best be traded off against standard wages at the local level. Indeed, profit sharing arrangements and employee stock ownership plans have often been adopted by firms in distress as an explicit part of concession bargaining at the firm level (see, for example, Bell and Neumark 1993).12 The standard view has been that unions are opposed to incentive pay systems because they are likely to increase wage dispersion. But it has also been claimed that higher unionisation may not necessarily make such pay systems less likely and that the key determinant of the probability that such systems are adopted is the general industrial-relations climate (Heywood et al. 1998). Interestingly, a recent study using panel data for US firms found that, although unionised firms were less likely than non-unionised firms to have profit sharing and employee stock ownership plans in 1975, they were equally likely to have adopted them subsequently (often as part of concessionary bargaining; Kruse 1996).
3.3
Alternative ways of decentralising pay setting
The development towards more decentralised pay setting in the current EU countries over the last two decades, which was described in Section 1.2, has taken place on the initiative of employers. Several driving factors have been pointed out: ●
Standardised tasks and hierarchical Tayloristic organisation within firms seem increasingly to have been replaced by more flexible tasks and flatter hierarchies, allowing greater decentralisation of decision-making. This makes an increased use of performance-related pay systems, which have to be decided on at the enterprise/workplace level, profitable for employers (Lindbeck and Snower 2000, 2001). This development can be seen as part of a more intensive search for efficiency-enhancing measures at the level of the firm driven by increased competition (due to product market deregulations and “globalisation”).
9780230_547018_05_cha03.indd 102
3/19/2009 8:03:24 PM
Pay-setting Systems in Europe ●
●
103
If most of the competition for a firm comes from other domestic producers, it may regard uniform wages across a sector as an advantage because they provide a “level playing field.” But increasing international competition renders this consideration less relevant. Instead, it becomes more important for each firm to be able to adjust its wage level to that of foreign competitors, which requires greater scope for wage bargaining at the level of the firm (Crouch 2000a, b; Calmfors et al. 2001). A desire to reduce the political power of unions by limiting their role as a national actor in general may be a third factor (Elvander 1999; Calmfors et al. 2001).
The development towards more decentralisation of pay bargaining can also be seen as an outcome of “meta bargaining” between unions and employers about the level at which wage negotiations should take place. To the extent that the relative bargaining strength of employers has increased, they may have been able to shift the locus of bargaining in their preferred direction. One possible explanation of such a shift in bargaining power is the increasing degree of international capital mobility, which gives the employer side a better “fall-back” position in the case of disagreement with unions. Another explanation of the shift in bargaining power in favour of employers is the fall in unionisation rates that has taken place in many Western European countries (see Figure 3.1 in Section 1.1). This development has been attributed to a number of structural changes in the economy (Ebbinghaus and Visser 2000; Calmfors et al. 2001). One such change is the shift in employment from the traditionally heavily unionised manufacturing sector to the private service sector with many small establishments that are costly for unions to organise. At the same time, employment in the public sector, where unionisation is in most countries higher than in the private sector (presumably because politicians for election reasons are more favourable to unions than private-sector management) has stagnated. Increased relative importance of part-time and temporary employment is usually also held as an important explanation of the trend towards lower unionisation, since employees on such contracts are likely to see smaller benefits of being unionised than permanently employed full-time employees. The forces behind a decentralisation of pay bargaining are likely to operate also in the future. Thus, one should expect further decentralisation to occur in continental, Western European countries. But this process can take place in different ways, which may have quite different implications. We shall distinguish between three possible scenarios: 1. Massive decentralisation and de-unionisation of the Anglo-Saxon type 2. Disorganised and gradual decentralisation/de-unionisation 3. Organised decentralisation where higher-level collective agreements allow more scope for local wage setting.
9780230_547018_05_cha03.indd 103
3/19/2009 8:03:24 PM
104
Designing the European Model
3.3.1 Massive decentralisation and de-unionisation The first scenario is massive decentralisation/de-unionisation of the type that has taken place primarily in the United Kingdom and New Zealand, but also in Australia. As discussed in Section 1, developments in these countries have not only meant a large reduction in unionisation, but also a dramatic reduction in the coverage of collective agreements and a radical shift from multi-employer to single-employer bargaining. From a theoretical point of view, the Anglo-Saxon model has its benefits. It allows an increased use of incentive pay and more relative wage flexibility. As we discussed in Section 2.1, it is an open question to which extent collective bargaining at the enterprise level in a system with high unionisation and high coverage of collective agreements creates incentives for aggregate wage moderation as compared with higher-level bargaining. But de-unionisation and low coverage of collective agreements represent strong wage-moderating forces. In both the United Kingdom and New Zealand, there has been much talk of trying to achieve larger ‘individual variation’ in employment contracts. It is, however, not clear to what extent this has happened in practice. A weaker role for collective agreements appears to have led to more legal regulation of ‘minimum employment conditions’ (Bray and Walsh 1998; Brown et al. 2000). Employers also often seem to have stuck to standardised employment contracts independently of the degree of trade union presence because of the cost savings implied (Bacon and Storey 2000). A development of the Anglo-Saxon type appears very unlikely in most Western European countries in the near future. Existing bargaining institutions in these countries seem to be changing only at a slow pace. One should not expect radical changes in the bargaining institutions unless there are massive reforms in the legal regulations. Indeed, it was such reforms that triggered off the dramatic changes in pay-setting systems in the United Kingdom, New Zealand and Australia. In the United Kingdom, there was a sustained series of legal changes during the 1980s, which gradually restricted union powers. These legal reforms involved: legislation on election of union officials and secret ballots before strikes; outlawing of secondary action; making unions liable for fines and civil damages if legal rules on industrial action were not observed; de facto abolition of closed shops; legal recognition of the right not to belong to a union; and rules making it more difficult for unions to be recognised by employers (Gregory 1998). Even despite these radical changes, the reduction in the importance of collective bargaining has been only a gradual process. The main explanation of the decreased scope of collective bargaining has been the difficulties of unions to get recognition at new workplaces rather than de-recognition at already existing ones (Machin 2000). The legal reforms seem to have led to a situation where traditional union wage premiums have largely disappeared, and where unionisation no longer appears
9780230_547018_05_cha03.indd 104
3/19/2009 8:03:24 PM
Pay-setting Systems in Europe
105
to cause negative productivity effects (Addison and Belfield 2001; Forth and Millward 2002). In New Zealand, the changes in wage-setting institutions were triggered by the Employment Contracts Act of 1991, which introduced completely new labour legislation. Here, unlike in the United Kingdom, no attempts were made to regulate the internal structure of unions. Instead, the guiding principle behind the new legislation was to view the employee and the employer as the primary parties to the bargaining process. Each employee must choose between bargaining on his/her own or being represented by a “bargaining agent,” who could be another person or an organisation. A trade union can be such an organisation, but it cannot bargain on behalf of an employee (not even a union member) unless he/she has given it explicit bargaining authority. Legislation allows explicitly for two types of employment contracts: individual and collective ones. Single-employer bargaining is encouraged, for example, by explicit provisions prohibiting industrial action to force an employer to become a party to a multi-employer collective agreement (Honeybone 1997; Bray and Walsh 1998). Changes in the legal regulations in Australia have been later and less radical. But the Workplace Relations Act of 1997 clearly favours collective agreements at the enterprise level and restricts the earlier use of compulsory arbitration (with “wage awards” now being confined mainly to minimum wages). The legislation also allows for individual employment contracts, although, unlike in New Zealand, these have not become the norm, but instead require an active “opt-out” (Bray and Walsh 1998). The upshot is that dramatic changes in wage-setting systems of the type that have occurred in Anglo-Saxon countries would seem to require massive legal interventions, abolishing, for example, legal restrictions on individual employment contracts that stipulate lower wages or other less favourable employment conditions than in collective agreements. Such favourability clauses exist, for example, in Germany, the Netherlands and France. In the United Kingdom and New Zealand, the fundamental changes in pay-setting institutions were made politically possible because of deep economic crises and widespread dissatisfaction with the workings of the traditional systems. But it is difficult to see that the political preconditions for such radical reforms exist in Western European countries today. This point is underlined by the failure to include even very modest changes of the pay-setting system in the recently adopted labour market reforms in Germany. It appears that political resistance to reforms of pay-setting practices is often much harder than to other types of labour market reform. The main reason is that changes in the pay-setting regime have a large, direct impact on employed insiders, as discussed in Chapter 2. Also, union officials, who see their role challenged, have strong incentives to try to organise political opposition to such reforms.
9780230_547018_05_cha03.indd 105
3/19/2009 8:03:24 PM
106
Designing the European Model
3.3.2
Disorganised decentralisation
An alternative scenario is that decentralisation and de-unionisation proceed in a spontaneous way without legal changes and with opposition from the union side. Such a disorganised decentralisation process seems to have been going on in Germany. It has, to a large extent, been concentrated in Eastern Germany, where many firms have left the employers’ associations in order to conclude enterprise agreements outside the sectoral collective agreements, and others have, in violation of existing legislation, paid wages below the minimum levels stipulated in the sectoral agreements. In Eastern Germany, Western German unions and employers’ organisations stepped in immediately after unification in 1990 to carry out wage negotiations. Before the privatisation programmes could begin, long-term wage contracts were concluded that foresaw full wage equalisation with Western Germany in only five years. After privatisation, many East German entrepreneurs tried to negotiate new contracts, but they only managed to postpone the time of wage equalisation by one year. As a consequence, many employers broke the union contracts with the silent consent of the unions, whereas others left the employers’ organisations. In 2002, no more than 20 per cent of all firms with 43 per cent of the work force were covered by sectoral agreements in Eastern Germany, whereas the corresponding shares were 44 per cent and 63 per cent, respectively, in Western Germany (see Table 3.5). Unions in Eastern Germany have been much weakened by this development. When the metalworkers’ union (IG Metall) called for a strike in the spring of 2003 to reduce the working week in Eastern Germany to 35 hours, Table 3.5 2002
The coverage of collective agreements in eastern and western Germany,
Percentage of firms Western Germany Sectoral collective wage agreement “Orientation” towards sectoral collective wage agreement Enterprise collective wage agreement No collective wage agreement
Eastern Germany
Percentage of employment Western Germany
Eastern Germany
44
20
63
43
22
34
16
23
2
4
7
12
32
42
14
22
Note: “Orientation” towards sectoral collective agreement means that firms/workers are not formally encompassed by such agreements, but that wage conditions have been “influenced” by them. Source: IAB Betriebspanel.
9780230_547018_05_cha03.indd 106
3/19/2009 8:03:25 PM
Pay-setting Systems in Europe
107
the weakness of unions there was clearly exposed. The strike was unpopular among employees who feared that the union demands would further undermine the competitiveness of their companies. After only a few weeks, the union had to give in without having achieved any of its goals. Spontaneous and disorganised decentralisation has also occurred in the public sector in Germany. The earlier joint bargaining association (Tarifgemeinschaft) for public sector employers, encompassing the federal government, the states and the municipal employers, collapsed in 2003 because bargaining goals turned out to be too diverse. To a large extent, this was related to tensions between Western and Eastern states (Eironline Germany 2003b).13 However, one would expect spontaneous decentralisation and de-unionisation processes to be slow, unless they occur in such extreme economic situations as in Eastern Germany. As pointed out in Section 3.1, it does appear that, in the absence of fundamental changes in the legal framework, bargaining institutions are very persistent. A possible explanation is that the design of labour market institutions to a large extent reflects deeprooted social attitudes. For example, Black (2001) finds that differences in the degree of centralisation of wage bargaining among countries depends on differences in societal values relating to such basic factors as “power distance” (the degree of inequality of power in organisations and institutions that the population considers normal), general attitudes towards “individualisation,” the cultural values attached to “masculinity” and attitudes towards risk. Another factor that counteracts tendencies toward spontaneous moves to single-employer bargaining is the demand for insurance on the part of employers against the costs of labour market conflicts. An important function of employers’ associations in some countries, such as the Scandinavian ones and Germany, is to provide this type of insurance through joint conflict funds. Indeed, in Sweden, for example, multi-employer bargaining initially emerged on the initiative of employers who sought to protect themselves against union action against individual firms through such cooperation (Skogh 1984). The operation of such a system of insurance creates incentives for hierarchical control in the employers’ associations to address moral hazard problems, so that individual employers do not choose too aggressive a stand against unions leading to an excessive number of labour market conflicts. Spontaneous moves to single-employer bargaining are not likely to proceed at a fast pace in most Western European countries. Nor are unionisation and the coverage of collective agreements likely to fall rapidly. The structural changes working in this direction, which we discussed in the introduction to Section 3, occur only slowly over time, and trade unions may, to some extent, develop strategies to cope with them (Calmfors et al. 2001). Experiences from the United States and Canada seem also to imply
9780230_547018_05_cha03.indd 107
3/19/2009 8:03:25 PM
108 Designing the European Model
that negative management attitudes to unionisation may matter much less for union recognition and the existence of collective agreements than the legal framework (Kleiner 2001). The available empirical research surveyed in Section 2.1 does not permit clear-cut conclusions on the macroeconomic effects of moving from a system of sectoral collective bargaining, such as in Germany, to a system of collective bargaining at the enterprise level. If the hump-shape hypothesis (see Section 2.1) is correct, such decentralisation will promote aggregate real wage restraint, although the effects may not be very large. But there is also the possibility that the incentives for wage restraint are weakened. This risk is most apparent if a move to single-employer bargaining would occur at the same time as coverage of collective agreements and unionisation remain fairly high. 3.3.3
Organised decentralisation
A third possibility is organised decentralisation (a term coined by Traxler 1995), according to which higher-level union confederations and employers’ associations choose voluntarily to leave more scope for local bargaining. At least theoretically, this scenario would seem to offer the greatest possibilities of combining the benefits of co-ordination in terms of internalisation of various effects of wage increases with the benefits of larger relative-wage flexibility. Several models exist for such organised decentralisation. Local determination of the distribution of wage increases One possibility is to retain the determination of the total margin for wage increases (the average wage increase in each firm/workplace) at the higher (usually sectoral) level, but increase the freedom of local bargainers to determine how this margin should be distributed among the individual employees in the firm/workplace. In many countries there has been a general trend towards such separation of the decisions at the sectoral level on the total margin applicable to each firm and the decisions at the enterprise level on how the agreed wage increase should be distributed among individuals. But the development does not seem to have gone very far in countries like Austria, Belgium, Germany and Italy. The development has gone much further in the Netherlands and the Scandinavian countries. In the Netherlands, sectorwide agreements have been interpreted as “increasingly adopting the character of framework agreements, which then need to be developed in detail at company level” (Eironline Netherlands 2002a). This can occur through negotiations between the employer and the local works council. Avtalsrörelsen and Lönebildningen (2002) give an overview of the differences in the designs of existing collective agreements in Sweden in 2002. According to their study, collective agreements with standard wage increases at the sectoral level that applied to all employees existed only in the private sector and did not encompass more than ten per cent of the employees
9780230_547018_05_cha03.indd 108
3/19/2009 8:03:25 PM
Pay-setting Systems in Europe Table 3.6
109
Wage dispersion in Swedish hospital care run by regional authorities 1988
1994
2002
90/10 90/50 50/10 90/10 90/50 50/10 90/10 90/50 50/10 Registered nurses (sjuksköterskor) Assistant nurses (undersköterskor)
1.18
1.05
1.12
1.19
1.07
1.11
1.29
1.15
1.12
1.12
1.02
1.10
1.08
1.03
1.04
1.13
1.06
1.07
Note: 90/10 is the wage ratio between the 90th and the 10th percentiles etc. Source: Calmfors and Richardson (2003).
there. All other collective agreements left some scope for local bargaining about the distribution of wage increases among individual employees. Usually local bargaining about the distribution of wage increases was conducted under the constraint of a minimum guaranteed wage increase for everyone (sometimes only as a guarantee when the collective agreement was evaluated ex post) or a fall-back clause stipulating that such a guarantee would apply if the local parties could not agree. Delegating the decisions on the distribution of wage increases among individuals to the local level makes it possible to use pay increases as an individual incentive mechanism. It is instructive to study how this has influenced wage setting in hospital care in Sweden, where employers (the regional authorities) have tried (in fact much more consistently than in large parts of the private sector) to introduce pay policies that aim explicitly at stimulating effort and on-the-job training (Calmfors and Richardson 2003). This has been achieved for registered nurses (sjuksköterskor in Swedish), where, since 1995, the union side has accepted that the distribution of nationally agreed pay increases are a subject for local negotiations only, without any influence from higher bargaining levels. Such a pay system has not been accepted by the trade union organising less educated assistant nurses (undersköterskor in Swedish), for whom collective agreements have contained guarantees of pay rises for the individual (more or less amounting to the standard pay increases negotiated centrally). As can be seen from Table 3.6, wage dispersion has increased substantially for registered nurses (in the upper half of the wage distribution), whereas it has stayed more or less constant for assistant nurses. Local determination of the total margin for wage increases Avtalsrörelsen and Lönebildningen (2002) also show that a substantial share of employees in Sweden is subject to higher-level, multi-year agreements that do not specify any figures for wage increases at all, but leave the determination of both, the total margin for wage increase (the average wage increase) and its distribution among individual employees in each firm/workplace,
9780230_547018_05_cha03.indd 109
3/19/2009 8:03:25 PM
110
Designing the European Model
entirely to annual local bargaining. Such contracts apply to 7 per cent of the employees in the private sector and to as much as 32 and 28 per cent in central government and local government (municipalities and regional authorities), respectively. These contracts apply mainly to white-collar employees with higher education. This represents another form of organised decentralisation, where higher-level collective agreements delegate also the determination of “total” wage increases to bargaining at the local level. The local bargaining is then conducted under a “peace obligation,” ruling out the use of industrial action. With such delegation of bargaining rights, higher organisational levels retain some co-ordination capacities through their influence on local negotiators as well as the capacity to “recall” the delegation in future collective agreements. This “recall capacity” can be seen as a constraint on the behaviour of local parties, at the same time as the delegation increases the scope for adjusting wage developments to local conditions. So far, collective agreements leaving the determination of the total margin for wage increases to the local level have mainly been used as a means to raise relative wages for groups with a favourable labour market position. But such agreements could also allow downward wage flexibility in situations of unfavourable labour market developments. Local incentive pay schemes The possibilities of using incentive pay mechanisms at the level of the individual firm/workplace can be enhanced by sectoral agreements on the design of performance-related pay systems. For example, the Confederation of Danish Industries and the Central Organisation of Industrial Employees in Denmark concluded a framework agreement on a new pay system (Plus Pay) in 2002, outlining the general principles for how basic pay (80 per cent of the total wage) and qualification-motivated supplements should be combined with task-related bonuses and performance-related pay elements, with the specific design in each firm to be negotiated in local bargaining (Eironline Denmark 2002). A similar framework agreement on more decentralised pay was also concluded in the public sector in Denmark in 2002 (Eironline Denmark 2003). Another example is the 2002 agreement in the German banking sector on variable pay, allowing variations of annual earnings of up to eight per cent) based on performance appraisals and achievements of agreed targets (Eironline Germany 2003a). We are not, however, aware of any sectoral or other higher-level collective agreements in EU countries on guiding principles for how local bargaining parties could trade off employee benefits in the form of profit sharing or employee stock ownership arrangements against standard pay rates in individual firms. In Finland, for example, unions have tried to negotiate such framework agreements, but employers have not been willing to do so (Eironline Finland 2003). On the other hand, unions have often been opposed to such compensation schemes, because they are likely to increase
9780230_547018_05_cha03.indd 110
3/19/2009 8:03:25 PM
Pay-setting Systems in Europe
111
pay dispersion. However, in a system where the main locus of pay bargaining is the sectoral level, such framework agreements may be necessary for a more widespread introduction of profit-related pay. Opening clauses A final possibility of increasing the flexibility of company pay levels in a system with sectorally negotiated pay increases is to widen the scope also for downward deviations from them. In Germany, for example, these possibilities are very restricted. The so-called favourability principle in the federal Collective Agreements Act (§ 4) states that “sites which are subject to collective agreements can only approve regulations which depart from the collective agreement if the changes are in the employees’ favour” (Bispinck 2003). Downward adjustments of the pay level to save jobs are not allowed under this stipulation, unless an explicit opening clause has been negotiated in the sectoral agreement. Such clauses are, however, usually quite restrictive and allow only limited and temporary deviations from the sectorally determined pay levels (Lange et al. 2001). A possible reform, proposed, for example, by the German Council of Economic Experts (2002) would be to amend the statutory regulations to establish a legal framework for decentralised pay bargaining. Each employer and local works council could be given complete freedom to deviate both upwards and downwards from the sectoral agreements if they agree on that. Alternatively, such a deviation could be allowed if a certain proportion of the employees agree to it. Introducing such “symmetric” flexibility at the local level would allow both, greater wage dispersion among firms in general and greater possibilities to adjust wages locally to adverse shocks. One would not, however, expect such a reform to radically affect wage outcomes, as the sectoral wage agreement would serve as a fall-back option in local bargaining in the case the parties cannot agree. Still, it did not prove possible to include such limited changes of the pay-setting system in the recent labour market reforms in Germany (see Chapter 2).
3.4
Conclusions
Each system of pay setting has its advantages and disadvantages. There is ample research evidence that highly co-ordinated collective bargaining at the multi-sector or national level, such as in many smaller European countries, promotes aggregate real wage restraint and low unemployment. Controlling for other factors, unemployment appears to be systematically lower with high co-ordination of collective bargaining than with intermediate co- ordination (where the locus of bargaining is the sectoral level). There are reasons to believe that EMU membership could increase the benefits of high co-ordination. The explanation is that monetary policy in EMU cannot react to wage hikes in individual countries, which weakens its
9780230_547018_05_cha03.indd 111
3/19/2009 8:03:25 PM
112
Designing the European Model
disciplining force. Under such circumstances, there may be a greater pay-off from wage-setting arrangements that in themselves foster wage moderation than if national monetary policy could be used as a direct means to influence domestic wage setting. However, pay compression and inflexibility of relative wages seem always to be serious disadvantages of centralised wage bargaining. Decentralisation of collective bargaining to the level of the firm facilitates the adjustment of wages to the labour market situation of different skill categories, occupations, regions and sectors. Such single-employer bargaining also facilitates the use of performance-related pay to boost labour effort and productivity. It appears, however, that decentralised collective bargaining is less conducive to aggregate real wage restraint and low unemployment than highly co-ordinated bargaining. But it is less clear how decentralised collective bargaining compares with intermediate centralisation (sectoral bargaining) in this respect. According to some empirical studies, decentralised collective bargaining leads, other things equal, to lower unemployment than sectoral bargaining. According to others it leads to higher unemployment. What is clear, however, is that low unionisation and low coverage of collective agreements promote real wage restraint and high employment. Therefore, the chances that single-employer bargaining results in wage moderation increase if a development in this direction goes hand in hand with a reduced importance of collective bargaining. As to the flexibility of the aggregate pay level, there are conflicting tendencies. On the one hand, an increased use of profit-related pay under single-employer collective bargaining can help raise aggregate pay flexibility. But, on the other hand, it can become more difficult to change wages in response to macroeconomic disturbances if bargaining is decentralised and unsynchronised, because the incentives for wage adjustments for a particular bargaining area may be weak unless other bargaining areas also adjust their wages. Against this background, which pay-setting systems should the EU countries opt for? One cannot give a generally valid answer to this question. A characteristic feature of pay-setting systems is their high degree of inertia: it usually takes a very long time or very special circumstances to achieve fundamental changes. So, the answer is likely to depend on the starting point, which for historical reasons differs a lot among countries. Several of the current EU member states, especially many of the smaller states (for example Belgium, Finland, Ireland and the Netherlands) have been able to achieve wage moderation through formal or informal coordination of bargaining at the multi-sector level, involving peak-level labour market organisations and sometimes also governments. Through such social pacts it has often been possible to establish generally accepted norms for wage increases, serving as benchmarks for subsequent negotiations at lower bargaining levels. In countries where such arrangements have
9780230_547018_05_cha03.indd 112
3/19/2009 8:03:25 PM
Pay-setting Systems in Europe
113
worked, there is little reason to abandon them now. But there are good reasons for treating central wage norms only as indicative, and not binding, in order to give more room for relative wage changes. The trick is to avoid such recommendations being regarded as “floors” for wage increases. It is important to get an understanding that a functioning market economy means that wage increases across firms, sectors, occupations and regions must deviate from the average, depending on the situation in the specific labour market. One way of signalling the need for this could be to publicise a corridor for wage increases rather than a single figure when guidelines are formulated. Alternatively, one might define a “normal” wage increase, which lies below the desired average, with the understanding that “abovenormal” wage increases should only be granted in bargaining areas with labour shortages. Both in countries with a social-pact tradition and in countries with more traditional bargaining at the sectoral level, such as Austria and Germany, one could also conceive of other reforms within the existing pay-setting system in order to increase relative pay flexibility among both firms and individuals. Such organised decentralisation could encompass reforms of the following type: ●
●
●
●
The distribution among individual employees of agreed “aggregate” wage increases at higher bargaining levels should be determined at the local level to a much larger extent than is currently the case. Higher-level agreements should therefore by and large abstain from providing individual guarantees of wage increases. The scope for opening clauses that give local wage bargainers also the right to deviate downwards from higher-level collective agreements should be increased. Such flexibility would, for example, be very helpful for adjusting wages in Eastern Germany and Southern Italy to the prevailing unemployment situations. Where there is bargaining at the sectoral level, framework agreements would facilitate the introduction of variable performance-related pay systems at the level of the individual firm. Such framework agreements should allow the possibility of wage adjustments at the company level in connection with the introduction of profit-sharing systems or employee stock ownership plans. One could also conceive of higher-level collective agreements that delegate the determination of the total margin for wage increases (the average wage increase in each firm/workplace) to the local level, but where the local negotiations are conducted under a “peace obligation,” ruling out industrial action, and the higher level retains the organisational capacity to “recall” the delegation in future collective agreements. Such “discretionary delegation” increases the flexibility of relative wages at the same
9780230_547018_05_cha03.indd 113
3/19/2009 8:03:25 PM
114
Designing the European Model
time as the higher-level collective agreements serve as a “stand-by option” if co-ordinated wage adjustments are needed in a macroeconomic crisis. How should one expect wage bargaining systems in continental Western European countries to develop? We have surveyed the tendencies to decentralisation that have taken place and concluded that this development is likely to continue. How should one evaluate this? Obviously, the conclusion depends on the starting point from which decentralisation occurs. For the countries with social pact co-ordination, the key question is whether or not it is possible to have both general wage norms determined in a co-ordinated way and relative wage changes that are decided in a decentralised manner. In our view the jury is still out. An optimistic scenario is that the process of organised decentralisation proceeds in the manner we have described so as to make greater diversity of wages compatible with continued aggregate real wage moderation. A pessimistic scenario is that disorganised decentralisation, increasing the number of independent pay-setting actors, ultimately makes co-ordinated wage restraint unfeasible. For a country such as Germany, where the main locus of collective bargaining is the sectoral level and the extent of co-ordination of collective bargaining is lower than in many of the smaller EU countries, the prospects are different. It is possible that a disorganised decentralisation process could lead to more aggregate wage restraint. But the difference to current outcomes may be small. And it is also possible that more single-employer collective bargaining could weaken the incentives for such wage restraint. A key requirement for overcoming the present stagnation in Germany seems to be that labour market reforms are extended to pay-setting practices in such a way that greater diversity of wages is allowed. If it turns out to be impossible to combine aggregate real wage restraint with relative-wage flexibility in the continental Western European economies, one could not in the long term rule out an Anglo-Saxon development, where decentralisation of collective bargaining is accompanied by massive de-unionisation and reduction in coverage of collective agreements. But such a development would probably only occur as the consequence of radical political reform of basic labour legislation after a deep economic crisis. An important reason why this scenario may not materialise is that trade unions may find it in their self-interest to acquiesce in both wage restraint and organised decentralisation precisely because they want to avoid such a development. Most of the new EU member states in Eastern and Central Europe find themselves in an entirely different situation from the present EU states in Western Europe. Estonia, the Czech Republic, Hungary, Latvia, Lithuania and Poland are all characterised by decentralised pay bargaining at the level of the enterprise and low unionisation and coverage of collective agreements. These countries are likely to face strong pressures from Western European trade unions, and possibly also from EU institutions, to change
9780230_547018_05_cha03.indd 114
3/19/2009 8:03:25 PM
Pay-setting Systems in Europe
115
their industrial relations systems so as to conform better to “EU standards.” Such pressures should be resisted. The existing “Anglo-Saxon” combination of enterprise-level bargaining and limited importance of collective agreements in these countries is likely to produce better macroeconomic outcomes than a move to collective bargaining at the sectoral level of the Western European type.
Notes 1. The ten countries are Austria, Belgium, Denmark, Finland, France, Germany, Italy, Spain, Sweden and the United Kingdom. According to Visser, co-ordination increased in Ireland, the Netherlands and Portugal. Outside the EU, co- ordination was found to have decreased in Switzerland but to have remained stable in Norway. 2. Concerning the EU countries, the difference between Visser and Ochel is that the latter found bargaining co-ordination to have decreased in Portugal, increased in Finland and Italy and remained unchanged in Germany. As to countries outside the EU, Ochel reports a decrease in co-ordination in Australia, Canada, New Zealand, Norway and Switzerland and unchanged co-ordination in Japan and the United States. 3. See, for example, Crouch (2000a, b), Calmfors et al. (2001), and Calmfors (2001) for more detailed discussions. 4. The Competition Act of 1996 stipulates formally that wage bargaining must be based on a pay norm set by the Central Economic Council (an advisory body to the government) according to which wages should not rise any faster than in Belgium’s three main neighbouring countries (France, German and the Netherlands) (Calmfors et al. 2001; Eironline Belgium 2003). 5. In Italy, social accords of 1993 and 1998 established a two-tier wage-bargaining process, according to which wage increases at the sectoral level should be linked to forecast inflation, whereas bargaining at the enterprise level should be based on productivity increases (Baccaro et al. 2001; Bertola and Garibaldi 2003). In Sweden, an agreement between a number of industry unions and employers’ associations in 1997 (the Industrial Agreement) has sought to establish a bargaining framework conducive to industrial peace and wage moderation with the aim of strengthening the role of the manufacturing sector as wage leader (Elvander 1999). Similar agreements on the bargaining framework have later been concluded in the public sector. 6. Some of the earlier studies in the table exploit only cross-section variation among countries, but later studies use panel data, thus exploiting also time-series variation. 7. This hypothesis was originally formulated by Calmfors and Driffill (1988). Note that the hump-shaped relationship presupposes that the degree of bargaining coordination is changed along the dimension firm-sector-nation. Changes in the degree of co-ordination along the dimension of various types of labour would have other effects. For example, decentralising collective bargaining so that different types of labour in each firm negotiate separately would raise the aggregate wage level if a wage increase for one group reduces labour demand for other groups (that is, if the various types of labour are complements in labour demand), since this negative externality cannot then be taken into account. 8. Iversen (1998) did indeed find empirical support for the hypothesis that a more non-accommodating central bank has a larger unemployment-reducing effect
9780230_547018_05_cha03.indd 115
3/19/2009 8:03:25 PM
116
9.
10.
11. 12.
13.
Designing the European Model at intermediate levels of co-ordination than at high or low levels. But Hall and Franzese (1998) and Cukierman and Lippi (1999) found only partial empirical support for the theoretical predictions. But, as analysed by Ball (1987), there is also an externality working in the opposite direction. It arises because wage setters in an individual bargaining area will not, under decentralisation, take into account that an increase in contract length there, and the consequent reduction in nominal wage variability, will reduce real wage variability in other bargaining areas. For example, the Trade Union Congress in Sweden sees “the possibility of achieving the basic elements of solidaristic wage policy with special wage increases for low-wage groups in order to attain equitable wage differentials” as a main advantage of co-ordinated bargaining (Riktlinjer för samordnade förhandlingar 2003, 19). IAB (2002a). There are, however, also many examples of such systems being adopted because firms want to raise employee compensation in periods of good performance without taking on an obligation to pay higher wages also in the future (Kruse 1996). Similarly, in Austria, the earlier centralisation of the (informal) bargaining in the public sector was abandoned in 2002 because of diversity in bargaining objectives (Eironline Austria 2002). Similar developments occurred much earlier in Sweden.
References Addison, J. and B. Hirsch (1989) “Union Effects on Productivity, Profits, and Growth: Has the Long Run Arrived?” Journal of Labor Economics 7, 72–105. Addison, J. and C. Belfield (2001) “Updating the Determinants of Firm Performance: Estimation Using the 1998 UK Workplace Employee Relations Survey,” British Journal of Industrial Relations 39, 341–66. Askildsen, J. and Ø. Nilsen (2002) “Union Membership and Wage Formation,” European Journal of Political Economy 18, 345–63. Avtalsrörelsen och lönebildningen 2002 (2003) Medlingsinstitutets årsrapport, Medlingsinstitutet (National Mediation Office), Stockholm. Baccaro, L., M. Carrieri and C. Damiano (2002) “The Resurgence of the Italian Confederal Unions: Will It Last?” Discussion Paper 144, International Institute for Labour Studies, Geneva. Bacon, N. and J. Storey (2000) “New Employee Relations Strategies in Britain: Towards Individualism or Partnership?” British Journal of Industrial Relations 38, 407–27. Ball, L. (1987) “Externalities from Contract Length,” American Economic Review 77, 615–29. Ball, L. and D. Romer (1991) “Sticky Prices As Coordination Failure,” American Economic Review 81, 539–52. Bell, L. and D. Neumark (1993) “Lump Sum Payments and Profit Sharing Plans in the Union Sector of the United States Economy,” Economic Journal 103, 602–19. Belman, D., J. Heywood and J. Lund (1997) “Public Sector Earnings and the Extent of Unionization,” Industrial and Labor Relations Review 50, 610–28. Belot, M. and J. C. van Ours (2001) “Unemployment and Labor Market Institutions: An Empirical Analysis,” OSA Working Paper 2001–10. Bertola, G. and P. Garibaldi (2003) “The Structure and History of Italian Unemployment,” Working Paper 907, CESifo, Munich.
9780230_547018_05_cha03.indd 116
3/19/2009 8:03:25 PM
Pay-setting Systems in Europe
117
Bispinck, R. (2003) “Country Report on Germany,” ETUI, Brussels. Black, B. (2001) “National Culture and Industrial Relations and Pay Structures,” Labour 15, 257–77. Blanchard, O. and J. Wolfers (2000) “The Role of Shocks and Institutions in the Rise of European Unemployment: The Aggregate Evidence,” Economic Journal 110, C1–C33. Blau, F. and L. Kahn (1996) “International Differences in Male Wage Inequality: Institutions versus Market Forces,” Journal of Political Economy 104, 791–837. Bleaney, M. (1996) “Central Bank Independence, Wage-Bargaining Structure, and Macroeconomic Performance in OECD Countries,” Oxford Economic Papers 48, 20–38. Bray, M. and P. Walsh (1998) “Different Paths to Neo-Liberalism? Comparing Australia and New Zealand,” Industrial Relations 37, 358–87. Brown, W., S. Deakin, D. Nash and S. Oxenbridge (2000) “The Employment Contract: From Collective Procedures to Individual Rights,” British Journal of Industrial Relations 38, 611–29. Calmfors, L. (1993) “Centralisation of Wage Bargaining and Macroeconomic Performance – A Survey,” OECD Economic Studies 21, 161–91. Calmfors, L. (2001) “Wages and Wage-Bargaining Institutions in the EMU – A Survey of the Issues,” Empirica 28, 325–51. Calmfors, L., A. Booth, M. Burda, D. Checchi, R. Naylor and J. Visser (2001) “The Future of Collective Bargaining in Europe,” in T. Boeri, A. Brugiavini and L. Calmfors, eds, The Role of Unions in the Twenty-First Century, Oxford University Press, Oxford, 1–155. Calmfors, L. and B. Holmlund (2000) “Unemployment and Economic Growth: A Partial Survey,” Swedish Economic Policy Review 7, 107–53. Calmfors, L. and J. Driffill (1988) “Bargaining Structure, Corporatism and Macroeconomic Performance,” Economic Policy 6, 14–61. Calmfors, L. and K. Richardson (2003) Marknadskrafterna och lönebildningen i landsting och regioner, Landstingsförbundet, Stockholm. Coricelli, Fabrizio, Alex Cukierman and Alberto Dalmazzo (2000) “Monetary Institution, Monopolistic Competition, Unionized Labor Markets and Economic Performance“, p. 44 Discussion Paper Series 2407, Centre for Economic Policy Research, London (CEPR), January 2000. Crouch, C. (2000a) “National Wage Determination and European Monetary Union,” in C. Crouch, ed., After the Euro, Shaping Institutions for Governance in the Wake of European Monetary Union, Oxford University Press, Oxford. Crouch, C. (2000b) “The Snakes and Ladders of Twenty-First-Century Trade Unionism,” Oxford Review of Economic Policy 16, 70–83. Cukierman, A. and F. Lippi (1999) “Central Bank Independence, Centralisation of Wage Bargaining, Inflation and Unemployment: Theory and Some Evidence,” European Economic Review 43, 1395–434. Danthine, J. P. and J. Hunt (1994) “Wage Bargaining Structure, Employment and Economic Integration,” Economic Journal 104, 528–41. Daveri, F. and G. Tabellini (2000) “Unemployment, Growth and Taxation in Industrial Countries,” Economic Policy 15, 48–104. Department of Employment, Workplace Relations and Small Business, “Agreement Making in Australia under the Workplace Relations Act 2000 and 2001,” Australian Workplace,www.workplace.gov.au/Workplace. Dufour, C. (2003) Collective Bargaining in France in 2002: A Sudden Change of Climate, Institut de Recherches Economiques et Sociales (IRES), Noisy-Le-Grand.
9780230_547018_05_cha03.indd 117
3/19/2009 8:03:25 PM
118
Designing the European Model
Ebbinghaus, B. and J. Visser (2000) Trade Unions in Western Europe since 1945, Macmillan, London. EEAG (2003) Report on the European Economy 2003, European Economic Advisory Group at CESifo, Munich. EEAG (2004) Report on the European Economy 2003, European Economic Advisory Group at CESifo, Munich. Eironline Austria (2002) “Changing Industrial Relations in Local Public Sector,” European Industrial Relations Observatory on-line, European Foundation for the Improvement of Living and Working Conditions, www.eiro.eurofound. eu.int/2002/04/feature/at0204202f.html. Eironline Belgium (2003) “2002 Annual Review for Belgium,” European Industrial Relations Observatory on-line, European Foundation for the Improvement of Living and Working Conditions, www.eiro.eurofound.eu.int/2003/01/feature/ be0301306f.html. Eironline Denmark (2002) “New Pay System Agreed for Industry Sector,” European Industrial Relations Observatory on-line, European Foundation for the Improvement of Living and Working Conditions, www.eiro.eurofound.eu.int/2002/11/inbrief/ dk0211102n.html. Eironline Denmark (2003) “2002 Annual Review for Denmark,” European Industrial Relations Observatory on-line, European Foundation for the Improvement of Living and Working Conditions, www.eiro.eurofound.eu.int/2003/01/feature/ dk0301106f.html. Eironline Finland (2003) “2002 Annual Review for Finland,” European Industrial Relations Observatory on-line, European Foundation for the Improvement of Living and Working Conditions, www.eiro.eurofound.eu.int/2003/01/feature/ fi0301106f.html. Eironline France (2003) “2002 Annual Review for France,” European Industrial Relations Observatory on-line, European Foundation for the Improvement of Living and Working Conditions, www.eiro.eurofound.eu.int/2003/01/feature/ fr0301108f.html. Eironline Germany (2003a) “2002 Collective Bargaining Round Examined,” European Industrial Relations Observatory on-line, European Foundation for the Improvement of Living and Working Conditions, www.eiro.eurofound. eu.int/2003/02/feature/de0302203f.html. Eironline Germany (2003b) “Public Sector Employers’ Bargaining Association Collapses,” European Industrial Relations Observatory on-line, European Foundation for the Improvement of Living and Working Conditions, www.eiro. eurofound.eu.int/2003/06/inbrief/de0306202n.html. Eironline Greece (2003) “2002 Annual Review for Greece,” European Industrial Relations Observatory on-line, European Foundation for the Improvement of Living and Working Conditions, www.eiro.eurofound.eu.int/2003/01/feature/ GR0301103F.html. Eironline Netherlands (2002a) “Debate over Viability of Collective Bargaining System,” European Industrial Relations Observatory on-line, European Foundation for the Improvement of Living and Working Conditions. www.eiro.eurofound. eu.int/2002/11/feature/nl0211104f.html. Elmeskov, J., J. Martin and S. Scarpetta (1998) “Key Lessons for Labour Market Reforms: Evidence from OECD Countries’ Experiences,” Swedish Economic Policy Review 5, 205–52. Elvander, N. (1999) The Industrial Agreement: An Analysis of Its Ideas and Performance, ALMEGAs förlag, Sandviken.
9780230_547018_05_cha03.indd 118
3/19/2009 8:03:25 PM
Pay-setting Systems in Europe
119
ERA (2003) The Report on Current Employment Relations in New Zealand, no. 12, Department of Labour, Wellington. Flanagan, R. (1998) “Institutional Reformation in Eastern Europe,” Industrial Relations 37, 337–57. Forth, J. and N. Millward (2002) “Union Effects on Pay Levels in Britain,” Labour Economics 9, 547–61. Godard, J. and J. Delaney (2000) “Reflections on the ‘High Performance’ Paradigm’s Implications for Industrial Relations as a Field,” Industrial and Labor Relations Review 55, 482–90. Gregory, M. (1998) “Reforming the Labour Market: An Assessment of the UK Policies of the Thatcher Era,” Australian Economic Review 31, 329–44. Hall, P. A. and R. J. Franzese (1998) “Mixed Signals: Central Bank Independence, Coordinated Wage-Bargaining, and European Monetary Union,” International Organization 52, 505–35. Hartog, J. (1999) “Wither Dutch Corporatism? Two Decades of Employment Policies and Welfare Reforms,” Scottish Journal of Political Economy 46, 458–87. Heywood, J., O. Hübler and U. Jirjahn (1998) “Variable Payment Schemes and Industrial Relations: Evidence from Germany,” Kyklos 51, 237–57. Holden, S. (2001) “Monetary Regimes and the Co-ordination of Wage Setting,” CESifo Working Paper no. 429, Munich. Honeybone, A. (1997) “Introducing Labour Flexibility: The Example of New Zealand,” International Labour Review 136, 493–507. IAB (2002) Die Entwicklung der Flächentarifbindung 1995–2001, Ergebnissse aus dem Betriebspanel. http://doku.iab.de/betriebspanel/ergebnisse/2002_11_01_05_ tarifbindung.pdf. Ichniowski, C., K. Shaw and G. Prennushi (1997) “The Effects of Human Resource Management Practices on Productivity: A Study of Steel Finishing Lines,” American Economic Review 87, 291–313. Ichniowski, C., T. A. Kochan, D. Levine, C. Olson, and G. Strauss (1996) “What Works at Work: Overview and Assessment,” Industrial Relations 35, 299–333. Iversen, T. (1998) “Wage Bargaining, Central Bank Independence, and the Real Effects of Money,” International Organization 52, 469–504. Jefferys, S. (2000) “A ‘Copernican Revolution’ in French Industrial Relations: Are the Times A’Changing?” British Journal of Industrial Relations 38, 241–60. Keynes, J. M. (1936) The General Theory of Employment, Interest, and Money, Macmillan, London. Kleiner, M. (2001) “Intensity of Management Resistance: Understanding the Decline of Unionization in the Private Sector,” Journal of Labor Research XXII, 519–40. Koskela, E. and R. Uusitalo (2003) “The Un-intended Convergence: How the Finnish Unemployment Reached the European Level,” CESifo Working Paper no 878, Munich. Kruse, D. (1993) Profit Sharing: Does It Make a Difference? Upjohn Institute for Employment Research, Kalamzoo, MI. Kruse, D. (1996) “Why Do Firms Adopt Profit-Sharing and Employee Ownership Plans?” British Journal of Industrial Relations 34, 515–38. Lange, T., G. Pugh and L. Funk (2001) “Reforms in the New German Labour Market: A Case for Institutional Corporatism?” International Journal of Manpower 22, 494–507. Layard R., S. Nickell and R. Jackman (1991) Unemployment, Macroeconomic Performance and the Labour Market, Oxford University Press, Oxford. Lazear, E. (1999) “Personnel Economies: Past Lessons and Future Directions,” Presidential Address to the Society of Labor Economists, Journal of Labor Economics 17, 199–236.
9780230_547018_05_cha03.indd 119
3/19/2009 8:03:25 PM
120
Designing the European Model
Lindbeck, L. and D. Snower (2000) “Multitask Learning and the Reorganization of Work: From Tayloristic to Holistic Organization,” Journal of Labor Economics 18, 353–76. Lindbeck, L. and D. Snower (2001) “Centralized Bargaining and Reorganized Work: Are They Compatible?” European Economic Review 45, 1851–75. Machin, S. (2000) “Union Decline in Britain,” British Journal of Industrial Relations 38, 631–45. Nalbantian, H. R. and A. Schotter (1997) “Productivity under Group Incentives: An Experimental Study,” American Economic Review 87, 314–41. Nickell, S. and G. Quintini (2002) “The Recent Performance of the UK Labour Market,” Oxford Review of Economic Policy 18, 202–20. Nickell, S. and R. Layard (1999) “Labor Market Institutions and Economic Performance,” in O. Ashenfelter and D. Card, eds, Handbook of Labor Economics, vol 3, Elsevier, Amsterdam, North Holland, 3029–84. Nickell, S., L. Nunziata and W. Ochel (2003) Unemployment in the OECD since the 1960s. What Do We Know? Bank of England. Nicoletti, G., R. Haffner, S. Nickell, S. Scarpetta and G. Zoega (2001) “European Integration, Liberalization and Labour Market Performance,” in G. Bertola, T. Boeri and G. Nicoletti, eds, Welfare and Employment in a United Europe, MIT Press,Cambridge, MA, 147–235. Ochel, W. (2000) Collective Bargaining (Centralization and Co-ordination), Ifo Institute, Munich. Ochel, W. (2001) “Collective Bargaining Coverage in the OECD from the 1960s to the 1990s,” CESifo Forum 4, 62–5. Rayton, B. (2003) “Firm Performance and Compensation Structure: Performance Elasticities of Average Employee Compensation,” Journal of Corporate Finance 9, 333–52. Riktlinjer för samordnade förhandlingar (2003) LO, Stockholm. Roche, W. (2001) “The Individualization of Irish Industrial Relations?” British Journal of Industrial Relations 39, 183–206. Rowthorn, R. E. (1992) “Centralization, Employment and Wage Dispersion,” Economic Journal 102, 506–23. Sachverständigenrat zur Begutachtung der gesamtwirtschaftlichen Entwicklung – German Council of Economic Experts (2002) Jahresgutachten 2002/2003: Zwanzig Punkte für Beschäftigung und Wachstum, Wiesbaden. Scarpetta, S. (1996) “Assessing the Role of Labour Market Policies and Institutional Settings on Unemployment: A Cross-Country Study,” OECD Economic Studies 26. Sinn, Hans-Werner (2003) Ist Deutschland noch zu retten? Econ Verlag: München. Skogh, G. (1984) “Employers’ Associations in Sweden,” in J. Windmuller and A. Gladstone, eds, Employers’ Associations and Industrial Relations: A Comparative Study, Clarendon Press, Oxford. Soskice, David and Torben Iversen (1998) “Multiple Wage-Bargaining Systems in the European Single Currency Area”, Oxford Review of Economic Policy, 14(3), pp. 110–24. Soskice, David and Torben Iversen (2000) “The Non-Neutrality of Monetary Policy with Large Wage and Price Setters”, Quarterly Journal of Economics (February, 2000), pp. 265–85. Traxler, F. (1995) “Farewell to Labour Market Associations? Organized versus Disorganized Decentralization as a Map for Industrial Relations,” in C. Crouch and F. Traxler, eds, Organized Industrial Relations in Europe: What Future? Avebury, Aldershot, 3–19.
9780230_547018_05_cha03.indd 120
3/19/2009 8:03:25 PM
Pay-setting Systems in Europe
121
Traxler, F. (2003) “Bargaining (De)centralization, Macroeconomic Performance and Control over the Employment Relationship,” British Journal of Industrial Relations 41, 1–27. Traxler, F., S. Blaschke and B. Kittel (2001) National Labour Relations in Internationalized Market: A Comparative Study of Institutions, Change, and Performance, Oxford University Press, Oxford. Visser, J. (2000) “A Combined Indicator of Union Centralisation and Coordination,” Working Paper no 00/03, Amsterdam Institute for Advanced Labour Studies. Wallerstein, M. (1999) “Wage-Setting Institutions and Pay Inequality in Advanced Industrial Societies,” American Journal of Political Science 43, 649–80. Walsh, B. (2002) “When Unemployment Disappears: Ireland in the 1990s,” CESifo Working Paper no. 856, Munich. Zetterberg, J. (1995) “Unemployment, Labour Market Policy and the Wage Bargaining System,” in J. Johannesson and E. Wadensjö, eds, Labour Market Policy at the Crossroads, EFA, Ministry of Labour, Stockholm, 57–115.
9780230_547018_05_cha03.indd 121
3/19/2009 8:03:25 PM
4 Outsourcing
4.1
Introduction
With the fall of the Iron Curtain in Europe and the opening of China, the percentage of the world population integrated with the Western trading system has increased dramatically. Regions with very different endowments of capital and labour as well as very different commodity price ratios have been merged, creating huge potentials for gains from trade. And international trade indeed has grown rapidly. While world GDP increased by 50 per cent in nominal terms from 1990 to 2002, the trading volume grew by 90 per cent. Among the increasing trading activities, international outsourcing and offshoring activities have received particular attention in the public debate in Western countries. Both relate to the so-called intra-industry trade of intermediate products. Outsourcing means that domestic firms give up parts of their intermediate production chains and instead buy parts from foreign suppliers (Feenstra and Hanson 2004). Offshoring means that domestic firms set up new factories abroad to produce the intermediary products themselves.1 As the economic implications of these two phenomena are the same, we treat the two equally in our analysis. Both phenomena involve a reduction in domestic production depth. Downstream activities close to consumers such as the final assembly of parts typically remain in the country, but labour-intensive upstream activities are often moved abroad in order to benefit from lower wages. Cars are good examples. Renault, Porsche and Audi are national brand names that require national content. But in fact, parts from Nissan or from own plants in Slovakia and Hungary account for substantial fractions of the value added. An extreme example is a car like the Porsche Cayenne. The car appears to be produced in Leipzig. However, in Leipzig only the steering, the gear box and a few other parts are added, while the main assembly line as such is in Bratislava. Leipzig accounts for only 12 per cent of the value of the Cayenne; 88 per cent of the value is already contained in what comes from Bratislava. 122
9780230_547018_06_cha04.indd 122
3/19/2009 3:48:37 PM
Outsourcing 123
Comprehensive empirical information on outsourcing and offshoring developments in the United States is provided by Feenstra (1996). Reviewing evidence of long-run changes in trade and production structures in manufacturing, he finds that there have been tendencies for more use of imported intermediate inputs in production. Related results have been obtained by Hild (2004) and Sinn (2004a) for Germany (see the appendix to this chapter). Outsourcing and offshoring activities not only apply to production for domestic use but also to production for exports. Increasing fractions of the value-added content of exports seem to be coming from abroad indicating that trade in intermediaries is growing even faster than trade on average. Evidence for this is given by Ng and Yeats (2002). They show that between 1984 and 1996 East Asian imports and exports of manufactured components grew annually two to three times as fast as imports and exports of traditional production. Related results have been found for Germany. Sinn called this the “bazaar economy effect”, a term which has triggered off a wide debate among German researchers.2 If a country specialises in “bazaar” activities, its factors of production move from other sectors towards activities in which their value added rises, but the export and import volumes rise even faster than this. International outsourcing and offshoring have gained importance in Europe since the mid-1990s, in particular after the ex-communist countries in Eastern Europe had overcome their transformation crises and EU membership came in sight. The public in Western EU countries has been alarmed by potential job losses and low-wage competition from the East. By contrast, business representatives have tended to play down these fears, pointing to the advantages the new trading opportunities are creating and to the possibility of cutting production cost, thus saving jobs in the West that otherwise would have been lost. This chapter tries to shed some more light on this issue.
4.2 The decline in the share of value added in production Outsourcing and offshoring activities, which reduce the domestic production depth, seem to be particularly important for the manufacturing sector. Outsourcing of services which also has received much attention in the media, does not (yet) seem to be an important empirical phenomenon.3 The reduction in domestic production depth is demonstrated quite clearly by the decline in the share of the manufacturing sector’s value added in its own production. As Figure 4.1 shows, this decline is a pronounced empirical trend in all other major EU countries. In Germany, for instance, the own value-added share in production declined from 40 per cent in 1970 to less than 34 per cent in 2003. In principle, the observed deviation of value added from production could be a result of domestic outsourcing to other sectors rather than
9780230_547018_06_cha04.indd 123
3/19/2009 3:48:37 PM
124
40
Designing the European Model %
% West Germany United Kingdom
38
40 38
36
36 Germany
EU15 34
34
32
32 Italy
30 28 26
30
France
28
Spain 1990
Figure 4.1
1992
1994
1996
1998
2000
2002
2004
26
Share of value added in the manufacturing sector’s own production
Sources: OECD, STAN database for Industrial Analysis, Vol. 2004 release 07; Germany: Federal Statistical Office, National Accounts: Special series 18, series S.21 and series 3.1; Ifo Institute calculations.
international outsourcing and offshoring. In fact, this has been the position taken by the Financial Times Deutschland in a series of articles countering the bazaar economy view.4 However, Table 4.1, which refers to all countries for which Eurostat data are available, shows that foreign outsourcing plays a more than proportionate role in explaining the development. In all major Western European countries, the share of foreign intermediate products in the total value of intermediate products increased from 1995 to 2000. This shows that the declining share of value added is indeed primarily due to the increased share of foreign, rather than domestic, outsourcing activities. As the reader can easily verify, the strength of this development differs among countries with Germany being more affected than others. The reason could be that Germany’s geographic and cultural proximity to the new EU member countries makes it particularly attractive for German firms to locate part of their production activities there.5 While the Eurostat information in Table 4.1 refers to the whole economy, the German Statistisches Bundesamt and the Council of Economic Advisors recently showed that the same tendency holds true for total exports as well as for exports of manufacturing goods.6 From 1991 to 2002, the fraction of foreign intermediary inputs in German exports, including the import of exported merchandise, increased from 26.7 per cent to 38.8 per cent, that is, by about 11 percentage points. More than eight of these 11 percentage points can be attributed to the period from 1995 to 2002. Similarly, the fraction of foreign intermediary inputs in the exports of the manufacturing sector
9780230_547018_06_cha04.indd 124
3/19/2009 3:48:37 PM
Outsourcing 125 Table 4.1 International outsourcing – The share of foreign intermediate products in totalintermediate products (in percent)
Italy Denmark Finland Netherlands Austria Sweden Germany
Year
Share
1995 2000 1995 2000 1995 2000 1995 2000 1995 2000 1995 2000 1995 2000
17 19 22 26 20 24 29 30 25 29 23 28 20 26
Note: The values reported above refer to the total economy. Source: Eurostat and own calculations. The underlying data are stored in CIRCA, the Intranet of Eurostat, and can be purchased at the Eurostat data shop.
increased from 26.7 per cent to 38.1 per cent in the period from 1990 to 2000. We are not aware that similar statistical information is available for other countries, but the tendency is strong enough to expect a phenomenon of wider importance for many European countries. Regressing real export-induced imports on real exports, where the export price index is used as a deflator, the results – published by the German Federal Statistical Office – imply a marginal intermediary import propensity of exports of 55 per cent. This means that 55 per cent of an additional real euro of German exports is value added coming from abroad and only 45 per cent is produced in Germany.7 Less value added per unit produced in the manufacturing sector does not necessarily imply that GDP is growing more slowly than it otherwise would. A development towards a bazaar economy could well be accompanied by an increasing prosperity of the exporting sector in general and the manufacturing sector in particular. In principle, at least some countries could specialise in engineering, final assembly and industrial sales activities with an ever increasing share of income earned this way. Figure 4.2 shows a mixed picture in this regard. While the downward trend of the manufacturing value-added share in GDP is a general phenomenon for the old EU, there are substantial differences among the various countries. Most countries, for instance Italy, Great Britain and Germany, display long-term reductions in their share of manufacturing value added in GDP. The overall decline since 1970 is very pronounced indeed. Only in France has the share stabilised, albeit at a very low level.
9780230_547018_06_cha04.indd 125
3/19/2009 3:48:38 PM
126
Designing the European Model %
%
34
34 Ireland
32
32 West Germany
30 28
30 28
Finland
26
26
24
24 Germany
Italy 22
22
20
France
18
20
EU12
18
United Kingdom
16
16 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04
Figure 4.2
Share of manufacturing in GDP
Sources: The World Bank Group, World Development Indicators Database 2003 and 2004. Germany: Federal Statistical Office, National Accounts: Special series 18, Series S.21 and Series 1.3; Ifo Institute calculations.
Exceptions to the general pattern appear to be Ireland and Finland. Ireland as one of the fastest growing economies in Europe has experienced a substantial increase in the share of manufacturing in total GDP. In Finland, the share fell up to 1991. However, thereafter it increased over much of the 1990s along with the “Finish miracle” boom that was particularly driven by the IT and other high-tech industries.8 The Finnish and Irish development may also, to some extent, have to do with the EU single market, which has removed the disadvantage of a limited domestic market size from which the smaller European countries had suffered before. Interestingly however, Germany, where the outsourcing and offshoring activities are particularly pronounced, seems to have succeeded in stabilising the share of manufacturing value added in GDP to some extent. However, Germany has had the slowest growth among all EU countries since 1995. Part of the stabilisation of the share may therefore simply reflect the weak performance of the rest of the economy. The value-added shares are only approximate indicators of employment shares, as firms in the old EU countries may react to the competitive pressures by increasing the capital intensity of their production and dismissing unskilled labour with relatively low value added per capita. Indeed, Figure 4.3 shows that the decline in manufacturing employment shares is more pronounced than the decline in value-added shares.
9780230_547018_06_cha04.indd 126
3/19/2009 3:48:38 PM
Outsourcing 127 %
% 25
25 West Germany
22
22 19
19
United Kingdom Germany
16
16
Italy
13
13
10
10 Spain 7
France 7
70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04
Figure 4.3 Employment in manufacturing sector (as a share of working age population) Sources: OECD, STAN database for Industrial Analysis, Vol. 2004 release 07; Working Age Population (between 15 and 64 years old): OECD Economic Outlook database; Germany: Federal Statistical Office, National Accounts: Special Series 18, series S.21 and Series 1.3; Ifo Institute calculations.
4.3 The general motives for outsourcing and offshoring activities Despite the media attention that outsourcing and offshoring is currently receiving, the academic literature on this topic is mostly recent, diverse and relatively scarce. Outsourcing as a reason for change in industrial structures, with outsourcing as a particular form of vertical disintegration, is emphasised by Grossman and Helpman (2002, 2003). They argue that outsourcing has been increasing in recent years due to improvements in communication technologies and the reduction of transaction costs between intermediate and final goods producers. In related work, McLaren (2000) argues that globalisation has increased the thickness of markets and that thicker markets imply less vertical integration.9 Another related idea is suggested by Casella and Rauch (2002) and Rauch and Casella (2003), who argue that outsourcing helps reduce informational barriers to trade and that foreign intermediaries have access to information that domestic firms would otherwise not have. According to these hypotheses, trading services are an important aspect of outsourcing. This is similar in spirit to Feenstra et al. (2002). These authors show that Chinese imports to the United States are to a large extent channelled through intermediaries
9780230_547018_06_cha04.indd 127
3/19/2009 3:48:38 PM
128
Designing the European Model
in Hong Kong and they view this as a form of outsourcing trading services that indicates benefits to firms from informational advantages of the Hong Kong intermediaries. A related question is that, given that a firm has decided to outsource and offshore part of the production process, where to outsource. Hanson et al. (2001) show that distance is among the most important factors in the decision to outsource in US data, next to tax differences and differences in labour cost. While not surprising, this finding is particularly important for Europe, as the proximity of Eastern Europe allows small and medium-sized firms to outsource parts of their production processes to the new members of the European Union, which are close-by neighbours.
4.4 Opening trade with the ex-communist countries: Low-wage competition at the extreme While these are important motives that certainly play a role in explaining outsourcing, it seems to us that the idiosyncratic shock to the world trading system that came about with the end of the communist regimes in the late 1980s and early 1990s was of particular importance. The more or less sudden opening of trade between devastated, poor ex-communist countries and the highly productive and rich Western countries was like opening the weirs between two lakes of different height. Much of what has happened to international trade since that time can be explained by this picture, including the rapid emergence of outsourcing and offshoring activities. In particular, outsourcing and offshoring is one way of taking advantage of the low labour cost in the Eastern European accession countries. Figure 4.4 gives an idea of how large the relocation incentives may be by comparing the wage costs for industrial workers across the EU countries. Figure 4.4 shows that the hourly labour cost in most Western European countries exceeds the labour cost in the new EU member countries by a factor of five to ten.10 Moving parts of the production process to Eastern Europe is therefore not a marginal saving for the firms but a large discrete jump in reducing total labour costs. Clearly this plays a major role in the location choice of newly established firms and is large enough in magnitude to justify a relocation of already existing plants if the share of labour cost in total production cost is large enough.11 Figure 4.5 indicates that the share of FDI in GDP that flows to the countries in net terms is negatively related to the hourly labour cost. In particular the Eastern European countries, which have the lowest wages within the EU, are the ones that have received the largest share of FDI. This is not identical but likely to be highly correlated with offshoring and outsourcing.12 A strong link between wages and direct investment flows has also been found by Marin et al. (2003) and in a most recent study of the Ifo Institute by Becker et al. (2004). The latter have exploited the voluminous direct
9780230_547018_06_cha04.indd 128
3/19/2009 3:48:38 PM
Outsourcing 129 West Germany Denmark Belgium Finland Netherlands Sweden Austria United Kingdom France Ireland Italy East Germany Spain Greece Cyprus Slovenia Portugal Czech Republic Hungary Poland Estonia Slovakia Lithuania Latvia
26.36 25.73 23.35 23.20 22.64 21.86 21.64 19.89 19.50 17.17 16.60 16.43 15.37 9.47 9.06 9.01 6.59
0
Figure 4.4
4.17 3.97 3.61 3.19 2.91 2.83 2.29 5
10
15
20
25
30
Hourly labour cost in Europe, 2002
Sources: Cologne Institute for Business Research: Deutschland in Zahlen 2004, Table 12.13; East Germany: Ifo Institute calculations.
investment data base of the Bundesbank that had not previously been accessible. They found that German and Eastern European workers are close substitutes rather than complements and that a wage increase in Germany has a significant and sizable positive impact on the number of jobs created by German firms in Eastern Europe.13
4.5
Impact on wages
While it is clear that outsourcing and offshoring activities are driven by wage differences, there is less evidence for the reverse causality. Theoretically, outsourcing reflects changes in international cost structures and productivity and thereby major shifts in comparative advantage. Such shocks will have the initial impact that some factors will earn lower income or lose employment if factor prices are rigid. Outsourcing is like technical progress, which may or may not hurt the different types of labour in the economy. The effects on the income distribution depend, among other things, on the relative capital intensities of the sectors that experience and do not experience outsourcing. In particular whether a factor will gain or lose depends on whether it is a complement or
9780230_547018_06_cha04.indd 129
3/19/2009 3:48:38 PM
130 Designing the European Model Change in net position, percentage points of GDP 1 80 Ireland 60 Czech Republic 40 Slovakia
Hungary
20 Sweden Netherlands Denmark
Poland 0 Portugal
Greece
Italy Spain
–20
Germany
Austria France
Finland United Kingdom
–40 0
5
10
15
20
25
30
Hourly labour cost, euros 2
Figure 4.5
Change in net foreign direct investment (stocks)
Notes: 1 Change of inward FDI stocks (as percentage of GDP) during the same period (1995–2002). 2 2002, euros. Source: OECD.
a substitute to the factor that has become more abundant by opening trade (see, for example, Jones 2003; Kohler 2004; Bhagwati et al. 2004). Clearly those wage earners who are substitutes to the low-wage earners who offer their labour abroad and who trigger off outsourcing and offshoring are hurt. They have to accept wage reductions to preserve their jobs. By contrast, those who offer labour that is complementary to the low-wage earners abroad will gain. Theoretically, it is even possible that the average wage rate rises. From an empirical perspective it seems clear that low-skilled wage earners, basically industrial workers, will lose, as there is an abundance of people in the ex-communist countries including China who are now offering their labour for low-skilled industrial jobs.14 And it may also seem plausible that highly skilled individuals whose services are complements to these large numbers of low-skilled workers will gain. However, it is unclear where the borderline between skilled and unskilled labour lies. According to a study of Geishecker and Görg (2004), even German workers with a normal school degree and accomplished vocational training programme belong to the group of losers, and only workers with a university degree can clearly be identified as winners. If this is true, then substantial fractions of wage earners in Germany are likely to belong to the group of losers. This view is implicitly confirmed by Marin (2004) who found that outsourcing to Eastern Europe does not only involve low-skilled labour, as is often assumed, but includes high-skilled labour, too. Labour markets in Eastern Europe are not comparable to labour markets in third-world countries. Rather
9780230_547018_06_cha04.indd 130
3/19/2009 3:48:39 PM
Outsourcing 131
a high degree of skill abundance seems to prevail. However, Marin finds only few negative effects on jobs of outsourcing from Germany and Austria. The empirical studies that try to estimate the impact of international outsourcing on wages face the difficulty of separating the effects of outsourcing from the technological changes in cost and production that often are the initial reason for outsourcing. These studies must also consider the closely related issue of skill-biased technical change and its role in the wage gap between skilled and unskilled workers that widened in the 1980s and 1990s.15 Feenstra and Hanson (2004) survey the empirical literature that has attempted to estimate the reasons behind the widened wage gap between skilled and unskilled workers in the US. This literature has employed several related methodologies, for example estimations of demand for skilled labour, price-cost conditions or economy-wide GDP functions to study the role of outsourcing. Feenstra and Hanson conclude that in addition to skillbiased technical change, “international trade is an important explanation of the increased wage gap”. However, the quantitative results in these studies are sensitive to the framework employed and also to the variables used to measure skill-biased technical progress.
4.6 Outsourcing, factor price equalisation and the welfare state: How to evaluate the process The question that arises is how to evaluate the patterns of industrial restructuring documented above. In principle, there is a strong presumption that the countries between which trade is opened up will experience gains from trade, because they can specialise in those sectors where they have comparative advantages and move away from others where they have disadvantages. As a rule, a country will gain more the further the price ratios of its products move away from the price ratios that prevailed, or would have prevailed, in autarchy, because then they can specialise more and buy imports at increasingly lower prices in terms of exported goods. Outsourcing and offshoring are important elements that contribute to gains from trade. If one sees the world as consisting of two kinds of countries, the highly industrialised countries of the West and the ex-communist countries that recently joined the world trading system, both groups of countries will unambiguously gain if they open up trade under free competition, since before trade was opened, the commodity price ratios differed widely. Both groups of countries will be able to import goods more cheaply than it would have been possible for them to produce these goods themselves. However, when the world is seen in a more disaggregated way and if various institutional constraints that limit the role of free competition are taken into account, the gains from trade are less obvious. Consider, for example, a third group of countries, say the resource-rich developing countries of the South. Then it may well be the case that the
9780230_547018_06_cha04.indd 131
3/19/2009 3:48:39 PM
132 Designing the European Model
terms of trade will not improve for the countries of the West if the excommunist countries join the trade between the West and the developing countries as a third partner. Suppose, for example, that the West has specialised in skill-intensive goods and that the ex-communist countries will do the same. Then the terms of trade for the West may worsen in the sense that the price of skill-intensive goods in terms of natural resources falls, moving closer to the level that would have prevailed in the West in autarchy. In this case the West’s gains from trade will decline. This is the fear that Samuelson (2004) expressed recently. The recent rise of the world oil price, which many have attributed to increasing demand from China, can be interpreted in the light of this argument. However, as has been shown by Dixit and Grossman (2004), the empirical conditions for this scenario to be relevant are unlikely to be satisfied. A more relevant concern for the countries of Europe seems to be that the condition that the markets in the countries among which trade is opened operate competitively is not satisfied. Gains from trade and specialisation require that the countries internally be organised such that the factors of production, capital and labour in particular, can freely move from the disadvantaged to the advantaged sectors. There are always disadvantaged sectors with dying firms and growing unemployment. The crucial test for gains from trade is how quickly a country is able to return to full employment after trade is opened and jobs in the import-competing industries are destroyed by creating new jobs in the rising export sectors. This test applies to outsourcing and offshoring activities in particular. The domestic job losses they incur must be compensated for by new jobs that are established in other parts of the economy. To be concrete, an efficient process of international outsourcing and offshoring would imply that the work time set free in the firms that dislocate their labour is used in other sectors such as services or construction whose output cannot easily be provided by the ex-communist low-wage countries or in downstream manufacturing sectors that specialise in bazaar activities. Some European countries seem to meet the test. The UK where unemployment declined despite substantial job losses in manufacturing is a good example. The UK seems to have succeeded well in managing the structural change towards a service-based economy. Its growth and employment performance in the last ten years has been outstanding. Other countries seem to be having more difficulties. Germany, which has been growing more slowly than any other EU country since 1995, is the most prominent example. However, the views on Germany differ widely. A number of German researchers have argued that Germany is able to master the current period of globalisation well and to capture gains from trade since the exporting sectors’ employment and value-added shares are growing relative to the rest of the economy. The Federal Statistical Office has been able to demonstrate that this is the case in Germany despite the
9780230_547018_06_cha04.indd 132
3/19/2009 3:48:39 PM
Outsourcing 133
pronounced tendency to reduce the value added per unit of export. The growth in the export volume overcompensated the declining production depth due to outsourcing and offshoring. The same researchers see it as a confirmation of Germany’s gains that the export surplus has increased in recent years.16 This line of reasoning escapes our understanding of how economies function. As explained, opening trade with a number of previously excluded countries means specialisation, and specialisation means increased exports and imports. Sectors where a country has a comparative advantage expand, export their products and absorb more factors of production such that their value added increases. Other sectors where a country has a disadvantage decline, giving way to imports. They lose the factors of production, and hence their value added falls behind. Showing that value added and employment in the exporting sectors have increased relative to the rest of the economy is close to a tautology. It implies that countries specialise, but it does not in itself imply gains from trade, let alone an optimal reaction to international trade. One reason why increasing exports cannot, by themselves, be interpreted as signs of gains from trade has been given by Brecher (1974) in a HeckscherOhlin trade model with a minimum wage constraint. If the minimum wage constraint applies in the high-wage, capital-abundant country, the labourintensive sector of that country has difficulties surviving the low-wage competition from abroad, setting free an excessive amount of capital and labour. The capital-intensive exporting sector then expands more rapidly, and the country overspecialises. The capital-intensive goods are used to buy more labour-intensive goods abroad, replacing the reduced national production of such goods. Both imports and exports grow beyond their optimal size. There is an export boom, but it is a sign of a pathologic reaction of the economy rather than of a successful reaction to international trade, because the capital-intensive sector is unable to absorb the workforce set free in the labour-intensive sector. This example shows that it is not always better to export more and to generate more export-induced value added.17 While it is debatable whether this case already applies to Europe, the argument makes it clear that a high level of exports in itself cannot possibly be interpreted as a sign for gains from trade if wages are sticky. Similarly, the mere fact that a country has an export surplus does not show gains from trade either. By definition, an export surplus is a capital export. While a capital export could be the sign of welfare-improving intertemporal trade, it could also be the consequence of bad national policies.18 When wages are fixed above the market clearing level and factor proportions cannot react to the forces of globalisation, a capital export can even be taken as an indicator of job exports. It is difficult to say to what extent this case applies to the European economies, but it is clear that a trade surplus in itself cannot be interpreted as a sign of gains from trade.
9780230_547018_06_cha04.indd 133
3/19/2009 3:48:39 PM
134 Designing the European Model
An indication that the case of inflexible labour markets is relevant for Europe is the increasing unemployment from which many European countries have suffered in recent years. Germany, whose unemployment has risen for more than three decades, is a characteristic example. In the period from 1995 to 2003, when the reduction in domestic production depth was particularly pronounced, 1.9 billion hours of work disappeared from the German manufacturing sector (industrial production without construction), which is a decline of 14 per cent. However, only 290 million additional hours of work were created in the rest of the economy. There was a net loss of 1.61 billion hours of work in the period under consideration. This undoubtedly was not a sound development that speaks of gains from trade. Outsourcing and offshoring per se do not cause unemployment. The problem is instead the inflexibility of the labour market. In order for the structural change accompanying specialisation to operate efficiently and bring about gains from trade, wages would have to give way as a reaction to increased low-wage competition from abroad. However, both trade unions and the relatively high replacement incomes of the welfare state have prevented wages from falling. Replacement incomes, such as unemployment benefits and social aid, act as lower bounds on the wage distribution. They were not a problem when growth was high and wages were rising, because they then followed wages at a sufficient distance. However, when trade is opened with huge low-wage areas like China and Eastern Europe, wages come under pressure while the replacement incomes stemming from better times prevent them from giving way. Thus, too much outsourcing and offshoring activity may have been induced, and too few jobs may have been created in the rest of the economy. This has added to the already existing unemployment problems caused by the interaction of macroeconomic shocks in the past and unfavourable institutional conditions. This is the general dilemma of market integration between high-wage and low-wage countries. Such integration will bring about gains from trade, but only in the sense that the winners gain more than the losers lose. The reason is that market integration tends to bring about factor price equalisation or at least a narrowing of the gap in factor prices. As argued above, at least wages for unskilled labour in the EU-15 come under pressure, and by the same mechanism wages in the new accession countries are lifted. Thus there are losers in the West, and in the current situation the losers may well comprise substantial fractions of the workforce of the Western European countries. It is understandable that unions and politicians try to stem the tide and prevent the distributional consequences.19 However, if they do that by fixing wages and social replacement incomes, they prevent the necessary adjustments of the economy and hence the gains from trade from occurring. There are no gains from trade if political constraints make the corresponding convergence of factor prices impossible.
9780230_547018_06_cha04.indd 134
3/19/2009 3:48:39 PM
Outsourcing 135
Box 4.1
Competitiveness
In the discussion on outsourcing, and the process of globalisation in general, the term “competitiveness” is often used in a misleading way. Often, it is argued that the lower cost of inputs from Eastern Europe makes Western European firms more competitive and therefore gives them an advantage in placing their products on the world markets. This undoubtedly is the case However, it is not the most interesting and relevant form of competitiveness. While the firms – and the owners of capital in general – including perhaps skilled workers remain competitive and even increase their competitiveness through outsourcing, it is the unskilled workers who are unable to compete. The main concern raised about competitiveness in this chapter is that workers in a highly regulated labour market may not be able to compete with the low-wage alternatives that are opening up for internationally mobile capital in Eastern Europe.
Nevertheless, the exporting sector of such an economy may flourish for the reasons given by Brecher (1974), as cited above. Also many firms in such an economy may stay competitive. However, unemployment shows that an increasing fraction of the workforce may at the same time have lost its competitiveness (see Box 4.1). In this sense, the judgment of international outsourcing and offshoring is not trivial. In principle, these are good and natural consequences of a fruitful integration of markets. However, the inflexibility of wages in the West may create more shifts in production processes and a more rapid reduction in domestic production depth than is optimal. The subtle truth is that a process that is good in principle may go too far because of the inflexibility of Western labour markets. The EEAG had previously argued that one of the most powerful tools in the fight against unemployment is a partial conversion of the welfare state from the payment of wage replacement incomes to a policy of permanently paying wage subsidies to the very low-skilled, and it had presented a system of activating social aid to accomplish this without increasing the cost of the welfare state. This strategy remains the appropriate medicine against an overly rapid process of outsourcing and offshoring, because it establishes the wage flexibility necessary to bring about the gains from trade while at the same time compensating the victims of factor price equalisation.20
4.7
Conclusions
Since the fall of the Iron Curtain and the integration of China into the world trading system, international trade in goods and services has increased significantly, given that the factor endowments and hence relative prices differ substantially. Trade in intermediate products has developed particularly rapidly due to outsourcing and offshoring activities of firms that have tried
9780230_547018_06_cha04.indd 135
3/19/2009 3:48:40 PM
136 Designing the European Model
to make use of the huge wage differences between the formerly separated parts of the world. This has caused the domestic value added per unit of output – the so-called production depth – to decline in many sectors. This trend towards a reduction in production depth has been particularly strong in the manufacturing sector. This trend of de-industrialisation is related to outsourcing and offshoring activities of domestic firms. Parts of the production process have been moved to low-wage countries. In particular the new members of the European Union in Eastern Europe are the focus of a new restructuring of the economy. Outsourcing and offshoring are not limited to manufacturing only. In Germany, for example, it has been shown that these phenomena apply to the export sector as a whole. From 1991 to 2002, the additional unit of real exports, on average, induced a 55 per cent increase in intermediate imports. Only 45 per cent of the increase in real exports implied additional value added in the exporting country, a phenomenon that has been caricatured as the “bazaar effect.” If a country specialises in bazaar activities, its factors of production move from other sectors towards such activities such that their value added rises, but the export and import volumes rise even faster than this. In principle, outsourcing activities may lead to gains from trade for all countries involved. The low-wage countries of Eastern Europe and Asia find new and profitable employment activities for their affluent labour forces and are able to increase their wages. And the high-wage countries of the West are able to withdraw part of their endowments of labour and capital from labour-intensive sectors to use them more productively in the service and high-tech sectors where they may have comparative advantages. Outsourcing and offshoring is just a special form of international trade that can be expected to boost world GDP and world welfare, because it allows the countries to specialise in their comparative advantages. However, for the gains from trade to occur it is essential that the domestic factor markets in the West are flexible enough to allow for the necessary factor migration between shrinking and expanding sectors. While capital markets in Europe do seem to meet this requirement, labour markets are quite rigid. For one thing, national job protection measures prevent workers from moving easily between sectors. For another, the repercussions of collective wage agreements and the welfare state that is based on wage replacement payments prevent the necessary wage flexibility. Gains from trade go hand in hand with a tendency towards factor price equalisation. In particular, the specialisation on more capital-intensive production requires lower wages so as to prevent unemployment. If wages are rigid, this process cannot take place. The sectors where the West has a comparative disadvantage shrink too quickly setting more labour free than is useful, and the growing sectors where there is a comparative advantage do not create enough additional jobs even though they grow faster than optimal. A growing level of unemployment results.
9780230_547018_06_cha04.indd 136
3/19/2009 3:48:40 PM
Outsourcing 137
In some European countries we see strong signs of such a deficiency of the adjustment process. Thus the EEAG advocates policies to make the labour market more flexible. The necessary measures include the measures it recommended in earlier years’ reports. They range from more limited job protection policies via opening clauses for collective wage agreements towards a policy of activating social aid that changes the role of the welfare state from a competitor to a partner of private enterprises. In addition, of course, increasing daily work times as recommended in Chapter 5, would be an easy way to alleviate the problems.
Appendix on Germany As argued above, for Germany the process of international outsourcing and offshoring seems to be particularly pronounced. Figure 4.6 illustrates that the value-added share of the manufacturing sector has decreased from more than 40 per cent in 1970, to less than 25 per cent in 2003. Comparably, the share of employment of the manufacturing sector in total employment (including government, without self-employed) has declined from 36 per cent in 1970 to only 20 per cent in 2002. Figure 4.7 illustrates the patterns of industrial changes in Germany in more detail. Particularly important are the developments of real output and real value added as reported in the input-output statistics of the German Federal Statistical Office (the nominal shares are also reported in Figure 4.6). Real output is the inflation-adjusted total value of industrial products in a given year. The value added is that part of the value that was generated by the manufacturing sector itself. Value added is equivalent to the primary income of the industrial production sector, plus taxes, which is in principle identical to the gross profits, interest payments, wage payments, gross salaries, including social insurance payments and indirect taxes. Not included are intermediate goods obtained elsewhere. Due to these intermediate goods, the real output is always larger than value added. This does not imply that the indexed real output figure need to lie above the index of value added. If both values were to grow at the same rate, the share of intermediate goods would remain constant over time. As Figure 4.7 illustrates, this is clearly not the case in Germany. Instead, one can see that the two lines drift apart in the course of the late 1990s. While real output has grown by 18.3 per cent, which is roughly equal to the EU average GDP growth, value added has only grown by 4 per cent. Apparently, a growing share of German manufacturing production is due to a process of outsourcing to other sectors and countries and to offshoring. In the light of these developments, it is not surprising that employment in the German manufacturing sector has been reduced by 8.3 per cent. The question is where the production of intermediate goods has moved to. In principle, both domestic and foreign sectors could be supplying the
9780230_547018_06_cha04.indd 137
3/19/2009 3:48:40 PM
138 Designing the European Model %
%
45
45
Employment 40
40 West Germany 35
35
30
30 Value added
Germany 25
25
20
20 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04
Figure 4.6 Share of manufacturing value added in total value added and manufacturing employment in total employment1 Notes: 1Total employment: Domestic employees in total economy, including public sector. Sources: Federal Statistical Office, National Accounts: Special series 18, Series S.21, Table 1.2.1 and 1.2.6.
intermediate products. Figure 4.7 shows that the latter of the two is more important. Imported intermediate products have grown by 45 per cent in the time period from 1995 to 2003, twice as fast as industrial production and about 10 times as fast as value added. Despite the cyclical downturn in 2002 and 2003, this shows that increasing fractions of industrial output have been moving abroad. The share that the German manufacturing industry had in its own increase in real output was merely 11 per cent; 38 per cent was due to intermediate products from other branches in the domestic economy, and a remarkable 51 per cent was due to intermediate products imported from other countries. This corresponds to the data provided by the Federal Statistical Office, as cited above in the main text, according to which 55 cents of each real additional euro of German export is directly used for the purchase of imported intermediate goods. The shaded parts of the pie-diagram in Figure 4.7 indicate which shares of manufacturing production have been crowded out by other sectors and other countries, respectively. Had all three components of real output (domestic production, foreign and domestic intermediate goods) changed proportionately, the share of the manufacturing sector’s own value added in the increase in production would have been 49 per cent. The fact that the share is actually only 11 per cent is explained by an 8 percentage point increase of domestic outsourcing, and a 30 percentage point increase in foreign outsourcing and offshoring. The reduction in the domestic depth of production
9780230_547018_06_cha04.indd 138
3/19/2009 3:48:40 PM
Outsourcing 139 160
Increase in real output 1995 – 2003
150 38%
140 130
Real purchases of intermediate products from the other domestic sectors 8%
Real imports of intermediate products
51%
11% Real value added
Real imports of intermediate products Real purchase of intermediate products from the other domestic sectors
120 30%
Real output
110 Real value added
100 90 80
Employees
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
Figure 4.7 Bazaar economy (Components of the change in German industrial prodution [manufacturing industry] 1995–2003, 1995 = 100) Note: Real output net of transactions within the manufacturing sector, as reported by the Federal Statistical Office. The shaded areas in the pie diagram show the extent to which the reduction in production depth can be traced to more than proportional increases in the intermediate products delivered from internal sectors and from other countries. 2003: Ifo estimate. Sources: Federal Statistical Office, Input–output tables 1995 and 2000: Special series 18, Series 2 and Sinn (2004a) on the basis of Ifo Institute calculations and estimates.
is therefore to an extent of four-fifths explained by shifting production to foreign countries and only to one-fifth by shifting production to other sectors in the domestic economy that are not part of the manufacturing sector.
Notes 1. Offshoring differs from foreign direct investment (FDI) in so far as FDI need not imply that the goods produced in an owned or partly owned company abroad are used in the domestic production process, although this could be the case as well. We follow the definition of Feenstra and Hansen. See also Feenstra (1996), who speak of outsourcing as international production sharing. 2. The term was first used in H.-W. Sinn, Deutsche Rede, Stiftung Neu Hardenberg, live radio transmission, Deutschland Radio, 15.11.2003, reprinted in Sinn (2004b). An extensive discussion of the critics’ arguments can be found in Sinn (2005). 3. For studies on service outsourcing, see Bhagwati et al. (2004) and Amiti and Wei (2004). The latter paper contains some data on international outsourcing of ser vices. The authors find that (i) service outsourcing is still low, though increasing, (ii) some rich countries, for example USA and UK, have had more “insourcing” of services than outsourcing, and (iii) in the UK, job growth at the sectoral level is not negatively related to service outsourcing. Jones and Kierzkowski (2003) emphasize that the IT revolution and reductions in regulations have allowed firms to exploit potential increasing returns in service links
9780230_547018_06_cha04.indd 139
3/19/2009 3:48:40 PM
140
4.
5. 6 7. 8.
9. 10.
11. 12.
13.
14.
Designing the European Model between different stages of production. A cautious conclusion on the significance of service outsourcing might be that while it is discussed very much in the media, its role in the aggregate is still relatively small. Naturally, the phenomenon is of far greater significance for specific firms, jobs and industries. See also the discussion of outsourcing services in Chapter 1 in 2005’s EEAG report. See e.g. articles “Banker zweifeln an Basar-Ökonomie”, Financial Times Deutschland 16.7.2004, p. 16, and “Der orientalische Basar und die deutsche Krise Ökonomen finden neue Erklärung für Wirtschaftsmisere”, Financial Times Deutschland, 25.6.2004, p. 18. Further indicators for Germany’s special role can be found in Sinn (2005). Statistisches Bundesamt (2004) and Sachverständigenrat zur Begutachtung der gesamtwirtschaftlichen Entwicklung (2004), pp. 361–4. See Flaig et al. (2004). Other factors contributing to the extraordinary pattern in Finland were the fall of exports to the Soviet Union that contributed to the large fall in 1989–91 and the recovery from the deep Finnish recession in the early 1990s, that contributed to the recovery in the share of manufacturing in GDP. See also the CESifo Forum (summer 2004) for a further discussion on the consequences of outsourcing. Wages for workers in the new industrial locations of Eastern Europe are higher than average wages as shown above. Outsourcing and offshoring from Western Europe imply substantial wage differences within the new accession countries. For a more detailed analysis for the economic performance of the accession countries, see our 2004 EEAG Report. In Chapter 5, we make suggestions on how to overcome the high labour cost problems of the Western European countries. Normally FDI statistics give little information of green-field investment. The lion’s share of measured investment is purchases of existing firms and retained earnings within such firms. In Eastern Europe, these items are unlikely to be large, however, due to the fact that communism did not leave many functioning existing firms. In the recent literature there is a consensus that domestic and foreign workers are substitutes. However previous studies, some came to different conclusions. Braconier and Ekholm (2000) found a substitutability for high income countries, but not for low income countries. Braconier and Ekholm (2001) found in the same data set, however, that outsourcing of Swedish firms to Eastern Europe has triggered both a relocation of jobs from Sweden to Eastern Europe and, to an even larger extent, a relocation of jobs of Swedish companies that where already outsourced to Southern Europe. Brainard and Riker (1997 and 2001) found for a set of US multinationals that parent employment and affiliate employment in low-wage countries are weak substitutes, while affiliate employment in different low-wage countries are strong substitutes. However, they also reported complementarity between affiliate employment in different types of countries. In particular in their 2001 paper, which uses an updated data set, they confirmed the evidence of substitutability. Konings and Murphy (2001) found evidence of substitutability, although only at the 10 percent significance level, for a set of EU firms. See Ekholm (2003) for an overview. Meshcheryakova (2004) attempts to estimate the income losses of low-skilled workers in a two-country version of a neoclassical growth model. In a calibration exercise with data from the United States and China she finds that outsourcing reduces the lifetime utility of low-skilled workers in the US.
9780230_547018_06_cha04.indd 140
3/19/2009 3:48:41 PM
Outsourcing 141 15. Katz and Autor (1999) give the estimates that between 1979 and 1995 the real wages of U.S. workers with more than 16 years of education rose by 3.4 percent, while those with less than 12 years of education fell by 20.2 percent. 16. Sinn (2005). 17. For a more extensive discussion, see Sinn (2005). 18. A current account surplus by definition is equal to an export of capital. Suppose we start with the identity Y-T=C + I + G + (X-M), where Y (income) – T (taxes) is the disposable income, C is consumption, I is investment, G is government expenditure and X (exports) – M (imports) is the current account surplus. A reformulation of this equation yields (Y-T-C) + (T-G) = I + (X-M), where private and public saving stand on the left hand side and investment and the current account surplus on the right hand side. National savings are obviously allocated in the open economy towards domestic investment and net exports. Therefore, by definition, an export surplus is a capital outflow. This identity of course says nothing about the causality. Part of the current European export surplus may result from the attempt to export capital, part may be due to low internal demand which implies low imports. However, the low internal demand may in turn result from the low investment volume which itself may reflect the availability of attracting outsourcing and offshoring alternatives. 19. The implications of labour market unionisation on the effects of outsourcing are considered by Skaksen (2004). He shows that potential outsourcing makes the members of trade unions worse off relative to capital, while the reverse is true after realised outsourcing. Zhao (2001) argues that unionisation in vertically related markets gives firms incentives to outsource. 20. See Chapter 1.
References Amiti, M. and S.-J. Wei (2004) “Fear of Service Outsourcing: Is it Justified?” IMF Working Paper WP704/186. Becker, S. O., K. Ekholm, R. Jäckle and M.-A. Muendler (2004) “Location Choice and Employment Decisions: A Comparison of German and Swedish Multinationals,” mimeo, Ifo Institute, November 2004. Bhagwati, J., A. Panagariya and T. N. Srinivasan (2004) “The Muddles over Outsourcing,” Journal of Economic Perspectives 18(4), 93–114. Braconier, H. and K. Ekholm (2000) “Swedish Multinationals and Competition from High- and Low-Wage Locations,” Review of International Economics 8, 448–61. Braconier, H. and K. Ekholm (2001) “Foreign Direct Investment in Central and Eastern Europe: Employment Effects in the EU,” Cepr Working Paper no. 3052. Brainard, S. L. and D. Riker (1997) “U.S. Multinationals and Competition from Low Wage Countries,” NBER Working Paper no. 5959. Brainard, S. L. and D. Riker (2001) “Are US Multinationals Exporting US Jobs?” in David Greenaway and Douglas R. Nelson, eds, Globalization and Labour Markets, vol. 2, E. Elgar, Cheltenham, UK and Northampton, 410–26. Brecher, R. A. (1974) “Minimum Wage Rates and the Pure Theory of International Trade,” The Quarterly Journal of Economics 88, 98–116. Casella, A. and J. E. Rauch (2002) “Anonymous Market and Group Ties in International Trade,” Journal of International Economics 58, 19–47. Dixit, A. and G. Grossman (2004) “Samuelson Says Nothing about Trade Policy,” mimeo, Princeton University.
9780230_547018_06_cha04.indd 141
3/19/2009 3:48:41 PM
142
Designing the European Model
Ekholm, K. (2003) “Industrial Structure and Industry Location in an Enlarged Europe,” SIEPS Report no. 19, November. Feenstra, R. C. (1996) “Integration of Trade and Disintegration of Production in the Global Economy,” Journal of Economic Perspectives 12, 31–50. Feenstra, R. C. and G. H. Hanson (2004) “Global Production Sharing and Rising Inequality: A Survey of Trade and Wages” in E. Kwan Choi and James Harrigan, eds, Handbook of International Trade 1, Wiley-Blackwell, Oxford. Feenstra, R. C., G. H. Hanson and S. Lin (2002) “The Value of Information in International Trade: Gains to Outsourcing through Hong Kong,” NBER Working Paper no. 9328. Flaig, G., W. Nierhaus, O.-E. Kuntze, A. Gebauer, S. Henzel, O. Hüllsewig, A. Kaltschütz, E. Langmantel, M. Ruschinski, H.-W. Sinn and T. Wollmershäuser (2004) “ifo Konjunkturprognose 2005: Abgehängt von der Weltkonjunktur”, ifo Schnelldienst 57(24), 15–56. Geishecker, I. and H. Görg (2004) “Winners and Losers: Fragmentation, Trade and Wages Revisited,” Discussion Paper 385, Deutsches Institut für Wirtschaftsforschung, Berlin. Grossman, G. M. and E. Helpman (2002) “Integration versus Outsourcing in Industry Equilibrium,” Quarterly Journal of Economics 117, 85–120. Grossman, G. M. and E. Helpman (2003) “Outsourcing versus FDI in Industry Equilibrium, Journal of the European Economic Association 1, 317–27. Hanson, G. H., R. J. Mataloni Jr. and M. J. Slaughter (2001) “Expansion Strategies of U.S. Multinational Firms,” NBER Working Paper no. 8433. Hild, R. (2004) “Produktion, Wertschöpfung und Beschäftigung im Verarbeitenden Gewerbe,” ifo Schnelldienst 57(7), 19–27. Jones, R. W. (2003) “Immigration vs. Outsourcing: Effects on Labour Markets”, mimeo, University of Rochester. Jones, R. W. and H. Kierzkowski (2003) “International Trade and Aggolomeration: An Alternative Framework”, mimeo, University of Rochester. Katz, L. F. and D. Autor (1999) “Changes in the Wage Structure and Earnings Inequality,” in O. Ashenfelter and D. Card, eds, Handbook of Labour Economics, vol.3A, Elsevier, Amsterdam, New York etc., 1463–555. Kohler, W. (2004) “International Outsourcing and Factor Prices with Multistage Production,” Economic Journal 114, 166–85. Konings, J. and A. Murphy (2001), “Do Multinational Enterprises Substitute Parent Jobs for Foreign Ones? Evidence from European Firm-level Panel Data,” CEPR Working Paper no. 2972, London. Marin, D. (2004) “A Nation of Poets and Thinkers? Less so with Eastern Enlargement, Austria and Germany,” Discussion Paper, Department of Economics, University of Munich, 2004–06. Marin, D., A. Lorentowicz and A. Raubold (2003) “Ownership, Capital and Outsourcing: What Drives German Investment to Eastern Europe?” Discussion Paper, Department of Economics, University of Munich, McLaren, J. (2000) “Globalization and Vertical Structure,” American Economic Review 90, 1239–54. Ng, F. and A. Yeats (2002) “Production Sharing in East Asia: Who Does What for Whom and Why?” in L. Cheng and H. Kierzkowski, eds, Globalization of Trade and Production in South-East Asia, Kluwer Academic Press, New York. Rauch, J. E. and A. Casella (2003) “Overcoming Informational Barriers to International Resource Allocation: Prices and Ties,” Economic Journal 113, 21–42.
9780230_547018_06_cha04.indd 142
3/19/2009 3:48:41 PM
Outsourcing 143 Sachverständigenrat zur Begutachtung der gesamtwirtschaftlichen Entwicklung (2004) Erfolge im Ausland – Herausforderung im Inland, Jahresgutachten 2004/05, Wiesbaden. Samuelson, P. (2004) “Where Ricardo and Mill Rebut and Confirm Arguments of Mainstream Economists Supporting Globalization,” Journal of Economic Perspectives 18, 135–46. Sinn, H.-W. (2004a) Ist Deutschland noch zu retten? 6th edition and higher, Econ, Berlin. Sinn, H.-W. (2004b) Mut zu Reformen. 50 Denkanstöße für die Wirtschaftspolitik, Beck (DTV), Munich. Sinn, H.-W. (2005) Basarökonomie Deutschland – Weltmeister oder Schlusslicht? in ifo Schnelldienst 58(6). Skaksen, J. R. (2004) “International Outsourcing when Labour Markets are Unionized,” Canadian Journal of Economics 37, 78–94. Statistisches Bundesamt (2004) Volkswirtschaftliche Gesamtrechnungen. Input-OutputRechnung. Importabhängigkeit der deutschen Exporte 1991, 1995, 2000 and 2002, Wiesbaden. Zhao, L. (2001) “Unionization, Vertical Markets and the Outsourcing of Multinationals,” Journal of International Economics 55, 187–202.
9780230_547018_06_cha04.indd 143
3/19/2009 3:48:41 PM
5 Longer Working Hours – The Beginning of a New Trend?*
A number of company-level agreements on longer working hours have recently been concluded in Germany. Such deals, for example, at Siemens and DaimlerChrysler have been widely publicised. The increases in working time have taken place without – or with only partial – pay compensation and thus represent cuts in hourly wages. They have in many cases occurred in response to employer threats of outsourcing production to lower-cost facilities abroad. The working-time agreements in Germany have provoked a lot of debate, not only in that country but also in France as well as in other Western European countries. A key question is whether the German agreements imply a reversal of the earlier trend towards shorter hours that could spread to other countries as well. Potentially, this could happen both via psychological demonstration effects and via direct competitive pressures as production sites in different countries compete for jobs. A controversial issue concerns the effects of longer working hours on jobs. It has been claimed both that a lengthening of working hours will raise employment and that it will reduce it. A related issue is whether or not an increase in working time will be an effective way of counteracting the future tendencies to falling labour supply in the ageing European societies (see Chapter 12). This chapter analyses both the causes and the consequences of the recent working-time agreements in Germany. The aim is to align the discussion better with available research than has so far been done.
5.1
Background
Recent research on income differences among countries has shown that the main factor behind the higher income per capita in the US than in Western Europe is higher labour input, not higher productivity (Gordon 2002; OECD 2003; Blanchard 2004). This is illustrated by Table 1. Whereas GDP per hour worked in the EU-15 was 90 per cent of the US level in 2002, the number of hours worked per capita in the EU-15 was only 79 per cent 144
9780230_547018_07_cha05.indd 144
3/19/2009 3:49:03 PM
Longer Working Hours – The Beginning of a New Trend?
145
Table 5.1 Real income, labour productivity, and labour input in EU-15 in percent of US levels 1970
2002
71 65 101
72 90 79
GDP per capita GDP per hour worked Hours worked per capita Source: Eurostat Structural Indicators (2004).
1200 1000 OECD 800 600 400 200
Lu
xe Ko m re bo a u S C w Ice rg ze it la z ch e nd r N Re lan ew p d Ze ubli al c a J nd U C ap ni a an te n d ad S a Po tate r s Au tug st al G ralia D ree e U nm ce ni te F ar d inl k Ki an ng d d Spom Ire ain H lan un d g Au ary Sw stri e a M den Sl ex ov ak P ico R ola ep nd ub lic N N o Italy et rw he a r y G lan er ds m Fr any a Be nc lg e i Tu um rk ey
0
Figure 5.1
Annual hours worked per capita, 2005
Note: 1) Data is calculated by average annual hours actually worked per worker multiplied with the ratio of total employment to total population. For the United States tbe total employment is substituted by the (1-unemployment rate as % of civilian labour force) multiplied with civilian Labour force. Source: OECD (2008).
of that in the US. The difference in income between the US and Western Europe was more or less unchanged between 1970 and 2002: GDP per capita in the EU-15 was about 70 per cent of the level in the US in both years. The unchanged income differential, however, masks two offsetting developments. Over the 1970–2002 period, labour productivity in Europe increased relative to that in the US, reflecting considerably faster productivity growth up until the mid-1990s and somewhat slower growth after that. In contrast, hours worked per capita fell continuously in Europe relative to the US. 5.1.1 Working time in different countries Figure 5.1 illustrates the differences in working time per capita among the OECD countries. The number of annual hours worked per capita is particularly
9780230_547018_07_cha05.indd 145
3/19/2009 3:49:03 PM
146
Designing the European Model
low in some of the continental Western European countries (France, Italy, Belgium, the Netherlands and Germany) as well as in Norway. Figure 5.2 decomposes the differences in hours worked per capita among OECD countries into: (1) differences in hours per employee; (2) differences in the employment rate (the ratio of employment to working-age population); and (3) differences in demography (the ratio of working-age population to total population). A low number of working hours per employee is the most important factor behind the low number of working hours per capita in Percentage difference with respect to OECD = Hours per worker effect + Employment rate effect3 + Demographic effect4 average hours per capita Korea Luxembourg Iceland Switzerland Czech Republic New Zealand Japan Canada United States Portugal Australia Greece Denmark Finland United Kingdom Spain Ireland Hungary Austria Sweden Mexico Poland Slovak Republic Italy Norway Netherlands Germany France Belgium Turkey –30 –20 –10 0 10 20 30 40 –30–20–10 0 10 20 30 40 –30–20–10 0 10 20 30 40 –30–20–10 0 10 20 30 40
Figure 5.2 Decomposition of differences in hours worked per capita (Percentage point differences in hours worked per capita with respect to the OECD average in 2005 1, 2 Notes: 1OECD averages are calculated as the population-weighted average for the countries shown for hours per capita and the demographic effects, employment average for hours per worker and working age populatian (15–64 years)-weighted average for the employment effect. 2Countries in descending order of the percentage difference from OECD average hours per capita. 3Based on the ratio of employment to worlking-age population (15–64 years). For the United States the total employment is substituted by the (1-unemployment rate as % of civilian labour force) multiplied with civilian Labour force. 4Based on ratio ofworking-age population (15–64 years) to total population. Source: OECD (2008).
9780230_547018_07_cha05.indd 146
3/19/2009 3:49:03 PM
Longer Working Hours – The Beginning of a New Trend?
147
most of the European countries at the bottom of the diagram (Belgium, France, Germany, the Netherlands and Norway). Figure 5.3 shows that hours per capita over the whole 1970–2005 period have declined in all OECD countries except the US, New Zealand, Canada, Iceland and Australia. Over the same period, hours per employee have fallen everywhere. The reductions have, however, been much larger in Western Europe than in the US (particularly so in Germany, the Netherlands, Ireland, Norway, France, Denmark, Spain, Belgium, Switzerland and Italy). Most of the reductions in both hours per capita and hours per employee took place in the 1970–1990 period. Working hours per capita actually rose in about half the OECD countries in the 1990–2005 period, mainly because of rising employment rates. Working hours per employee continued to fall in most European countries but at a slower pace than before. Actual hours worked per employee depend to a large extent on the incidence of part-time work, which we shall not discuss here.1 It is clear, however, that low standard working time for full-timers is an important explanation for the low working hours per employee in some Western European countries. This is illustrated in Table 5.2. Whereas annual standard working time for a full-time worker in 2003 was around 1900 hours in the US and around 1800 as an average in the new EU member states, the EU-15 average is only 1700 hours. The countries with the lowest annual standard working hours are France (1568), Denmark (1613), the Netherlands (1648), and (West) Germany (1648). As can be seen, the low annual standard working time in these countries reflects a short working week. In general, actual working time exceeds standard working time. One reason is the existence of both paid and unpaid overtime. Another reason, when working time is determined through collective agreements, is that not all firms are covered by such agreements and that working time tends to be longer in non-covered firms. The average actual as well as standard weekly hours for a number of Western European countries are shown in Table 5.3.2 The UK shows the largest discrepancy between the two measures with actual weekly working time exceeding standard working time by as much as 6.1 hours, but there are sizable differences in Germany (2.5 hours), France (2 hours), the Netherlands (1.9 hours) and Spain (1.9 hours) as well. Countries with shorter standard hours tend to have larger differences between actual and standard hours. A similar pattern appears to exist also within countries, with, for example, a larger discrepancy between actual and standard hours in Germany in the metal and engineering sector, which has the shortest standard working time, than in other sectors (Lehndorff 2004). Reductions in standard working time took place in a number of European countries in the 1980s and 1990s through either collective agreements or legislation. Table 5.4 gives an overview of cuts in the standard working week. The most far-reaching reductions took place in Germany and France. In Germany, the standard working week in the metal working and engineering
9780230_547018_07_cha05.indd 147
3/19/2009 3:49:04 PM
148 Designing the European Model Percent change
1970–2005
30 20 10 0 –10 –20
d
ly D en m ar k S U w ni ed te en d Ki ng do Sw m itz er la nd N or w ay Au st ra lia Ic el an U ni d te d St at es C an ad a
Ita
nl an
pa n
Fi
Ja
Fr
an
ce
–30
Percent change
1990–2005
40 30 20 10 0 –10
Ja p Fi an nl Po and rtu Sw ga ed l U e ni te Fr n d an Ki c ng e d Be om lg iu m Sw itz Ital er y la U N nd ni o te rw d a St y at e Ko s D re en a m A u ark st ra C lia an a G da N et ree he ce rla N Ic nds ew e l Ze and al a Lu Ire nd xe la m nd bo ur Sp g ai n
–20
Hours per capita Hours per worker
Employment rate
Working-age share of total population
Figure 5.3 Decomposition of the trend growth in labour utilisation 1970–2005 (The contribution of hours per worker; the employment rate and the age structure of the population)1 Note: 1Countries in ascending order by the trend growth in hours per capita duting the period indicated. Source: OECD (2008).
9780230_547018_07_cha05.indd 148
3/19/2009 3:49:04 PM
Longer Working Hours – The Beginning of a New Trend?
149
Table 5.2 Standard working time for full-time workers according to collective agreements and/or legislation, 2003 Per year, average for the whole economy
Per week, average for the whole economy
Per week, metal working
1904 1840 1840 1840 1840 1816 1803 1802 1801
40.0 40.0 40.0 40.0 40.0 40.0 39.2 39.0 39.6
– – 40.0 – – 40.0 – 39.0 –
1800 1776 1748 1748 1748 1730 1729 1728 1717 1710 1700 1693 1676 1673 1672 1648 1648 1613 1568
40.0 40.0 38.0 39.0 38.5 39.1 38.6 39.0 38.5 38.0 38.1 37.2 38.8 37.5 38.0 37.4 37.0 37.0 35.0
40.0 – 38.0 40.0 37.5 38.0 38.5 39.0 38.5 38.0 37.9 37.3 40.0 36.5 39.1 35.0 35.2 37.0 35.0
USa Estonia Hungary Latvia Poland Slovenia Japana Ireland EU-8 (new EU states) Greece Malta Belgium Portugal Slovakia Germany (east) Spain Luxembourg Austria Cyprus EU-15 UK Sweden Finland Italy Germany (west) Netherlands Denmark France Note: a The figure refers to 2002.
Source: All countries except Japan and the US: Working Time Developments (2003), EIROnline; Japan and the US: Deutschland in Zahlen (2004), Institut der Wirtschaft, Cologne.
industry was reduced from 40 to 35 hours in a series of collective agreements between 1984 and 1995. Working time reductions occurred also in most other sectors. In France, there were legislated cuts in the working week in 1982 (from 40 to 39 hours) and then again in 2000 and 2002 (from 39 to 35 hours). Cuts in the standard working week have also taken place in Austria, Belgium, Denmark, Greece, Hungary, the Netherlands, Norway and the UK.
9780230_547018_07_cha05.indd 149
3/19/2009 3:49:04 PM
150 Designing the European Model Table 5.3 Average actual and standard working time for full-time employees in EU-15, 2002 Actual working time
Standard working time
Difference between actual and standard working time
43.3 41.0 40.4 40.3 40.1 40.0 39.9 39.9 39.5 39.5 39.3 39.2 39.1 38.9 38.5 37.7
37.2 40.0 38.5 39.0 38.5 38.5 38.8 37.4 39.0 39.0 38.5 39.3 39.0 37.0 38.0 35.7
6.1 1.0 1.9 1.3 1.6 1.5 1.1 2.5 0.5 0.5 0.8 -0.1 0.1 1.9 0.5 2.0
UK Greece Spain Portugal Austria EU-15 Sweden Germany Ireland Luxembourg Belgium Finland Denmark Netherlands Italy France
Source: Actual working time: European Labour Force Survey (2002); Standard working time: Working-Time Developments (2003), EIROnline.
5.1.2 Theoretical considerations The differences in working time developments between Western Europe and the US over recent decades have initiated a lively research discussion. Basically, three types of theoretical explanations for these differences have been put forward. ●
●
Differences in preferences. For example, Blanchard (2004) argues that the most important cause of the differences in the development in working hours between Western Europe and the US are basic differences in the preferences regarding labour and leisure. According to this interpretation, Europeans have a stronger preference for shorter working hours than Americans. This view has received some support from survey studies of employee attitudes towards changes in working time (see Bell and Freeman 1994 and OECD 1998). Higher tax wedges in Europe. An alternative explanation focuses instead on the differences in tax wedges between Western Europe and the US. Labour taxes are higher and have risen by more in Europe than in the US over the last three decades. In a calibrated model of the labour-leisure choices of households, Prescott (2004) finds that the whole difference in the development of working hours per capita between a number of large European
9780230_547018_07_cha05.indd 150
3/19/2009 3:49:05 PM
Longer Working Hours – The Beginning of a New Trend?
151
countries and the US can be explained by the differential development of labour taxes. Olovsson (2004) obtains similar results for Sweden: focusing on households’ choices between market work and “home production,” he is able to explain all of the difference in working hours to the US with the higher Swedish tax wedges. These results are not undisputed however. Econometric estimates on panel data usually attribute a much smaller role to taxes in explaining the fall in hours per capita in Europe (Nickell 2003). Table 5.4 Major reductions in the standard work week in European economies, 1980–2004 Year
Change
Legislation
Austria
1990
40 → 38,5
Belgiumb
1999
40 → 39
x x
2003
39 → 38
Denmark
1987
39 → 37
France
1982
40 → 39
2000
39 → 35
x (large firms)
2002
39 → 35
x (all firms)
1984
40 → 38.5
1987
38.5 → 37.5
1989
37.5 → 37
1993
37 → 36
1995
36 → 35
1980
45 → 43
x
1981
43 → 42
x
1983
42 → 40
x
Hungary
2003
40 → 38
x
Ireland
1989–90
40 → 39
1982
40 → 38
1985
40 → 38
Germanya
Collective agreements x x (inter-industry agreement) x (inter-industry agreement) x (70% of employees)
x (metal working and engineering) x (metal working and engineering) x (metal working and engineering) x (metal working and engineering) x (metal working and engineering)
Greece
x (tripartite national framework agreement) x (Waasenaar agreement)
Nether-lands x (government civil servants)
Continued
9780230_547018_07_cha05.indd 151
3/19/2009 3:49:05 PM
152 Designing the European Model Table 5.4
Continued Year
Change
Legislation
Collective agreements
Norway
1987
40 → 37,5
x (blue-collar-workers in manufacturing)
UK
1979
40 → 39
1989–90
39 → 37
x (engineering) x (shipbuilding and engineering)
a
Notes: Working time reductions also occurred in other sectors than in the metal and engineering sector during the 1984–98 period, but are not shown in the table. b The entries in the table represent inter-industry agreements involving the government, which have been codified into law. The inter-industry agreements, have, however, only confirmed earlier concluded collective agreements at the sectoral level. For example, the reduction in the standard work week from 40 to 39 hours in such sectoral agreements took place mainly in 1980/81. Source: EIRO Online; Institut der deutschen Wirtschaft; http://www.eiro.eurofound.eu.int/2004/03/feature/tn0403108f.html http://www.reformmonitor.org/downloads/brochure/refmon_e.pdf http://www.reformmonitor.org/pdf-cache/doc_reports-cc-0-cm-3-cs-0.pdf http://www.reformmonitor.org/index.php3?mode=reform http://www.issa.int/pdf/jeru98/theme2/2–1b.pdf http://www.dgb.de/themen/Tarifpolitik/wo_arbeitszeit http://www.dgb.de/themen/Tarifpolitik/durchschnittliche_woaz
●
Higher coverage of collective agreements in Europe. A third possible explanation is the higher degree of unionisation and the larger role for collective agreements in Europe than in the US (see Chapter 3). Reductions in working time are a common objective of trade unions. Cuts in working hours can be viewed as a way of raising wages by restricting labour supply. They can also be seen as a method of distributing the unemployment following from wages above the market-clearing level more evenly among workers. Indeed, ideas of work sharing have been an important factor behind the working time reductions in continental European countries. There is also empirical research (from the US) showing that a higher coverage of collective agreements co-varies with shorter standard working hours (Earle and Pencavel 1990). This finding is consistent with a positive correlation among countries between, on the one hand, the working time reductions that took place in the 1980s and early 1990s and, on the other, union density and the coverage of collective bargaining (OECD 1998). Recent observations from Germany also show that weekly working time has increased in firms that have withdrawn from employers’ associations relative to firms that continue to be covered by a collective agreement (Kölling and Lehmann 2002).
5.2 The driving forces behind recent deals on longer working time In 2004, several widely publicised agreements on longer working hours were concluded in large German companies. These deals followed a debate
9780230_547018_07_cha05.indd 152
3/19/2009 3:49:05 PM
Longer Working Hours – The Beginning of a New Trend?
153
triggered by the Ifo Institute (Sinn 2003a, b).3 An agreement at Siemens (involving two mobile phone plants) raised standard weekly hours from 35 to 40 hours without pay compensation. At Daimler Chrysler there was an agreement (involving one car plant) to gradually raise the weekly working time from 35 to 39 hours without pay compensation for some workers and to increase the threshold at which overtime premia start to be paid. Another deal, encompassing all Volkswagen plants in Germany, included, beside a wage freeze, both a rise in the standard working time by 1.5 hours without pay compensation (but with deferred compensation in terms of enhanced possibilities to retire earlier) and an effective reduction of the threshold at which overtime premia are paid. Agreements on longer working hours have also been concluded at, for example, the truck manufacturer MAN, the Thomas Cook tourist group, Lufthansa, and many small and mediumsized firms.4 The state governments in Bavaria and Hessen have increased the weekly working time for civil servants and discussions on similar increases are also taking place in other states. The company-level agreements on longer working hours that have been concluded in the German metal working and engineering sectors were made possible by an opt-out clause in the latest industry collective agreement. The clause allows plant-level deals that deviate from the sectoral agreement on working hours in certain cases in order to safeguard jobs (Münchau 2004; EIROnline 2004). In line with this, several of the deals on longer working hours have also contained explicit employment guarantees on the part of firms and sometimes also commitments to invest in the existing production facilities. The recent agreements in Germany have intensified the debate on working time in that country as well as in other European countries. It appears, however, that the discrepancy between actual and collectively agreed working hours in Western Germany widened already in the second half of the 1990s. This is likely to have been the consequence of increases in working time both in firms leaving employers’ associations and in firms encompassed by sectoral collective agreements but entering into deals with their employees in violation of these agreements (Lehndorff 2004; Zimmermann 2004). The deals to lengthen working time are a response to pressures to reduce labour costs. High unemployment and international competitive pressures have, for a long time, exerted such pressures. The new feature is that employees are, to a larger extent than before, exposed to credible threats from employers that production sites will be closed down and jobs outsourced abroad, either to the new EU countries or elsewhere. This is a reflection of the ongoing “globalisation,” which has progressively lowered the obstacles to international capital mobility (see the discussion in Chapter 4). The enlargement of the EU may represent a “discontinuous jump” in this direction: by increasing access to EU-15 product markets as well as promoting a more stable institutional framework, the expected returns to investment in
9780230_547018_07_cha05.indd 153
3/19/2009 3:49:05 PM
154
Designing the European Model
the new EU states have increased and the risks associated with such investment decreased. Against this background, there are two useful – and complementary – ways of viewing the company-level agreements on longer working hours in Germany: (1) as a convenient way of reducing hourly wage costs; and (2) as an induced labour supply response to a reduction of the hourly wage. 5.2.1 Company deals on longer working hours as a way of cutting labour costs Both real and nominal wages tend to be rigid downwards. Most employees are likely to resist cuts in real wages that endanger the consumption standards to which they are accustomed. The resistance to real pay cuts through reductions in nominal pay may be even greater, because psychological selfesteem of employees often seems to be linked more to nominal rather than real pay (Bewley 1999). A lengthening of working time without pay compensation can obviously be a convenient way of reducing hourly wages, since it leaves the total pay per employee unchanged. No reduction in real consumption standards is then required. Nor has the nominal wage income per employee to be reduced. The option of reducing hourly wage costs through an increase in working time is likely to be seen as more favourable the shorter the working time is to start with, as marginal disutility of work is then lower. From this point of view it is logical that it is in Germany that working time increases have occurred. Increases in working time at unchanged pay can give rise to very substantial cuts in hourly wages. For example, an increase in weekly hours from 35 to 40 hours represents a reduction of hourly wages by as much as 12.5 per cent. An increase from 35 to 37 hours represents a reduction by 5.4 per cent. It is natural that the agreements to cut labour costs through longer working hours have come about at the company rather than the sectoral level. There are well-known obstacles to wage moderation at the latter level. Most union members in a sector will be employed “insiders,” whose jobs are not threatened. Hence, the majority of union members in a sector will be unwilling to concede across-the-board wage cuts in order to preserve threatened jobs in some firms or create new ones for the unemployed (which is a process that will take time and where the future winners cannot be identified ex ante).5 The incentives for wage moderation are, of course, much stronger at an individual production site that the employer threatens to close down: then all the employees can be identified as “winners” already in the short run. For this reason, concession bargaining resulting in wage cuts in order to preserve existing jobs typically take place at the company level. Chapter 3 pointed to the strong decentralisation forces set in motion by increased international competition. Employers in Western Europe used to
9780230_547018_07_cha05.indd 154
3/19/2009 3:49:06 PM
Longer Working Hours – The Beginning of a New Trend?
155
be favourable to industry-level collective bargaining because it provided them with a level playing field, eliminating domestic competition in terms of wages. But this attitude reflected a situation where the bulk of competition was domestic. When competition instead is mainly international, the uniformity of national sectoral collective agreements prevents adjustment of wage costs to the specific competitive situation of the individual firm. One way of viewing the company-level agreements on working time (as well as other cost-cutting measures) is therefore as a kind of “decentralisation revolt” from below. The agreements on working time at, for example, Siemens, DaimlerChrysler, and Volkswagen, have all included some form of employment guarantee from the employer. It is well-known from the theory of collective bargaining that bargaining between employers and unions over both wages and employment leads in general to more efficient outcomes (with higher employment) than bargaining only over wages (McDonald and Solow 1981; Layard et al. 1991). The explanation is that simultaneous bargaining about both wages and employment allows greater possibilities of trading off wage restraint against higher employment. Centralised bargaining at the sectoral (or multi-sectoral) level is not consistent with such efficient bargaining over both wages and employment, since it would require mechanisms, which do not exist, for allocating the aggregate employment agreed at the centralised level among firms. This is the reason why such wage-employment deals are observed only at the level of the firm. 5.2.2 Longer working hours as a labour supply response to a real wage reduction A complementary way of looking at the agreements on longer working hours is as a labour supply response to a reduction in hourly real wages necessary to preserve the competitiveness of domestic production facilities. Standard theory tells us that we should analyse the supply of working hours as the labour-leisure choice of individual employees (and/or as an issue of allocating time between market work and “home production”). According to standard analysis, a cut in the hourly wage has an income and a substitution effect. The income effect tends to increase the number of hours supplied, because the employee’s demand for leisure, as well as for all other goods, tends to fall when income is reduced.6 The substitution effect tends to reduce hours supplied, because (market) work becomes less attractive relative to leisure (and home production). Theoretically, one cannot tell which effect dominates, but empirical work suggests a small negative net effect of wage reductions on labour supply (for prime-aged males; the effect is larger for females).7 This is the same as saying that the labour supply curve is weakly positively sloped, as depicted in Figure 5.4a. According to this reasoning, one should thus expect a fall – and not a rise – in working hours as a response to a reduction in hourly wage rates.
9780230_547018_07_cha05.indd 155
3/19/2009 3:49:06 PM
156
Designing the European Model (a) On the part of an individual employee
Hourly real wage
Hourly real wage
Desired working time Figure 5.4
(b) On the part of a union
Desired working time
Desired working time
Source: EEAG Group.
However, the conventional labour supply framework is not the appropriate one for analysing recent increases in working time as these have resulted from collective, and not individual, bargaining between employers and employees. The appropriate framework is instead the theory of collective bargaining. It is standard to use this for analysing wage setting.8 According to this framework, unions try to trade off the benefits of real wage increases for the members who remain employed against the utility losses for the members that become unemployed if wages are raised (too much). The utility loss from unemployment for a worker equals the difference between utility when employed (which depends on both wage income and working time) and utility when unemployed (which depends on unemployment benefits). The theory of collective bargaining can also be used for analysing working-time decisions.9 Obviously, a union, which acts in the interest of its members, will care about the labour-leisure trade-offs of employees. But unlike individuals acting alone, it will also be concerned about the effect of working-time decisions on the number of jobs: in the jargon of economists, a union internalises the effects of a change in working time of one union member on other union members. To provide more jobs (or prevent job losses), the union (workers acting collectively) has an incentive to restrict working time as compared to what workers would do when acting individually (Calmfors 1985, 1987; Holmlund 1987): given the hourly wage, shorter working time forces employers to satisfy their total demand for working hours by employing more workers. The desire of unions to hold back working time is a way of reducing the negative employment effects of wages that are set above the market-clearing level. The union incentive to restrict hours
9780230_547018_07_cha05.indd 156
3/19/2009 3:49:06 PM
Longer Working Hours – The Beginning of a New Trend?
157
of work in order to raise employment is stronger, the more attractive it is to have a job (that is the higher the utility associated with a job).10 According to the described theory, unions will in general respond differently to changes in hourly real wages than individual employees acting on their own as regards desired working time. The ordinary income and substitution effects that arise for employees acting individually will also affect the desired working time on the part of a union. But there is also an additional effect. A cut in the hourly wage means that the attractiveness (utility) of holding a job decreases, since a job then tends to be associated with a lower income. This effect weakens the incentive of a union to restrict working hours in order to promote employment and thus tends to counteract the ordinary substitution effect of a wage change on working time. Therefore, there are stronger forces working in the direction of increasing working time when there is a cut in the hourly wage if working time is determined in collective bargaining between unions and firms than when it is determined in bargaining between individual employees and firms. Under some theoretical assumptions – most importantly that the total number of working hours demanded by employers depends only on the hourly wage rate but is independent of working time per employee11 – the change-in-the-attractiveness-of-a-having-a-job effect on the desired working time exactly offsets the ordinary substitution effect for a union (Calmfors 1985). Hence, in such an analysis only the income effect remains, so that a reduction in the hourly wage must lead to an increase in working time.12 This is equivalent to saying that the schedule showing how the desired working time of a union depends on the wage is negatively sloped, as shown in Figure 5.4b. Our theoretical reasoning provides a possible explanation of the recent agreements on longer working hours in Germany. These are then viewed as an endogenous labour supply response in collective agreements to a required reduction in the hourly real wage. Provided that the hourly wage must fall, longer working hours, which help maintain members’ incomes, are in the interest of unions.
5.3
The effects of working hours on employment and output
A key issue is how agreements on longer working hours of the type that have been concluded in Germany affect employment and output. The views in the public debate diverge fundamentally. Proponents of longer working time tend to argue that more jobs will be created, whereas opponents usually claim the opposite. To sort out the arguments, it is important to distinguish between short-run and long-run effects. In a short-run analysis, one can take the reductions in hourly wages implied by the agreements as given. In a long-run analysis, one must consider how wage-setting incentives will respond over time to longer working hours: in long-run equilibrium,
9780230_547018_07_cha05.indd 157
3/19/2009 3:49:06 PM
158 Designing the European Model
aggregate employment must be such that the parties to wage bargaining have an incentive to set wages that are consistent with the return to capital required by the international capital market. 5.3.1 Short-run employment and output effects of longer hours To illustrate the short-run effects of longer working hours, it is instructive to analyse a very stylised example. Think of a profit-maximising firm that produces an output the price of which the firm cannot affect. This would, for example, be the case if the firm is one of many producers of an identical good in the world market. To begin with, assume that the physical capital stock as well as the degree of capital utilisation in the firm are given. Assume also that hours and workers are perfect substitutes to the firm, so that it is indifferent to whether a given output is produced by more employees working fewer hours per employee or by fewer employees working more hours per employee. How would employment in such a firm respond to an increase in working time at unchanged pay per worker, where the pay per worker is the product of the hourly wage and the number of working hours? The firm employs workers up to the point at which the productivity of the marginal worker (the increase in output from an additional worker) equals the pay of a worker. The effect on employment of an increase in working time at unchanged pay depends on whether the productivity of a marginal worker rises or falls. There are two counteracting effects. ●
On the one hand, the productivity of a marginal worker tends to increase when he/she works more hours.
On the other hand, the productivity of a marginal worker tends to fall because longer working time for all workers implies a lower productivity of the marginal hour (the increase in output from an additional hour worked). The reason is that the productivity of the marginal hour depends positively on the ratio between capital and the total number of hours worked (the number of workers x working time) and that this ratio tends to fall when working time increases.13 The net effect on the productivity of a marginal worker of an increase in working time depends on which of the two effects dominates. If the productivity of a marginal hour falls only slowly when hours increase, the first effect dominates and the productivity of a marginal worker increases. This makes it profitable for the firm to increase employment. If instead the productivity of a marginal hour falls quickly as hours increase, the second effect dominates, so that the productivity of a marginal worker falls and the firm reduces employment. What is clear, however, is that, independently of how the number of workers is affected, output increases. This follows because a profit-maximising
9780230_547018_07_cha05.indd 158
3/19/2009 3:49:06 PM
Longer Working Hours – The Beginning of a New Trend?
159
firm chooses the total number of hours worked such that the productivity of a marginal hour worked equals the hourly wage. Obviously, when the hourly wage falls, it becomes profitable for the firm to increase the total number of hours. Another way of thinking about the employment effects of a lengthening of working time at unchanged pay per worker is in terms of the wage elasticity of labour demand (the percentage increase in the total number of working hours demanded by the firm when the hourly wage falls by one per cent). As discussed in Box 5.1, the condition for an increase in working time at unchanged pay per worker to raise employment in our stylised example is that the labour demand elasticity exceeds one (see also Sachverständigenrat 2003). When the capital stock is fixed, the labour demand elasticity equals the ratio between, on the one hand, the elasticity of substitution between capital and labour and, on the other, the profit share in output. Typical values for these parameters are 0.6 and 0.3, respectively, which would give a labour demand elasticity of around two. This presupposes, however, that the firm can restructure the given capital stock and adopt more labour-intensive production methods. Since this may be a time-consuming process, the elasticity is likely to be well below unity in the short run (say within a year). This could be taken to suggest that, in the short run, the output increase achieved through longer working hours must be bought at the cost of a reduction in employment. This conclusion does not follow, however. There are two main reasons for this. The first is that the degree of capital utilisation is likely to increase when working time increases. The second reason is that longer working time could affect the size of the capital stock. Working time and the utilisation of capital It is important to distinguish between the physical capital stock and the capital services produced by this physical capital stock – just as we distinguish between the “stock” of workers employed (employment) and the labour services produced (the total number of hours worked). The volume of capital services depends on the degree of utilisation of the capital stock, that is on the length of time during which the capital stock is operated (the operating time). It is reasonable to assume that longer working time for employees increases the operating time of capital. This can indeed be seen as one of the major advantages of longer working hours since it implies an output increase similar to that of an otherwise cumbersome accumulation of capital (Sinn 2004a, b). The increase in the volume of capital services produced by a given physical capital stock counteracts the tendency to a fall in the productivity of a marginal hour when the working time of all employees increases and thus makes a positive employment effect much more likely (see Box 5.1). A special case of interest is when the operating time of capital equals the working time of employees. This holds when there is no shiftwork (or, more
9780230_547018_07_cha05.indd 159
3/19/2009 3:49:07 PM
160
Designing the European Model
Box 5.1 Some formulas for the short-run employment effects of longer working hours at unchanged pay per worker If (1) a firm produces output with the help of (a fixed) capital (stock) and labour, (2) hours and workers are perfect substitutes in production within the ranges of variation in hours and employment that are relevant, (3) there are no fixed costs of employment (costs that are fixed per worker independently of the length of working time), (4) there is a given output price at which the firm can sell all that it produces, and (5) employment is determined by the demand for workers, the following formula for the employment effect of an increase in working time at unchanged (nominal) pay per worker applies: Percentage change in the number of workers = Percentage increase in working time x (Elasticity of labour demand with respect to the hourly real product wage – 1). The elasticity of labour demand measures the percentage increase in total working hours demanded when the hourly real product wage falls by one percent. The real product wage is the wage in units of the firm’s own output, that is the ratio between the nominal wage and the output price. With a given output price, the percentage change in the nominal wage equals the percentage change in the real product wage. The formula is easy to understand. If the hourly wage were held constant (which implies that the total number of hours worked = employment x working time is also constant), then a one percent increase in working time would reduce employment by one percent: a direct negative “work-sharing effect”. But when instead the lengthening of working time takes place at unchanged pay per employee, a one percent increase in working time implies a one percent decrease in the hourly wage. Hence, there is a percentage increase in the total number of hours worked which equals the elasticity of labour demand. The percentage change in employment is obtained as the difference between the percentage increase in total hours due to the fall in the hourly wage and the one percent reduction in employment that would occur at an unchanged hourly wage. When the operating time of capital depends on the working time of employees – so that the total amount of capital services produced by the fixed capital stock increases when working time increases – and the production function exhibits constant returns to scale, the change in employment following from an increase in working time at unchanged pay can be reformulated as: Percentage change in the number of workers = Percentage increase in working time x (Elasticity of labour demand with respect to the hourly real product wage + Elasticity of operating time of capital with respect to working time – 1). The difference to the first formula arises because, with a constant-returns-to-scale production function, a one percent increase in the volume of capital services increases the total number of working hours demanded also by one percent. The percentage increase in the operating time following from a one percent increase in working time is given by the elasticity of operating time with respect to working time. The larger is this elasticity, the greater is the probability of a positive employment effect. In the special case when the operating time of capital equals working time, the elasticity of operating time with respect to working time is unity and the formula becomes: Percentage change in the number of workers = Percentage increase in working time x Elasticity of labour demand with respect to the hourly real product wage.
9780230_547018_07_cha05.indd 160
3/19/2009 3:49:07 PM
Longer Working Hours – The Beginning of a New Trend?
161
In this case, an increase in working time at unchanged pay per worker always increases employment. The direct negative “work-sharing effect” on the demand for workers of longer working time is exactly offset by an increase in demand associated with the increase in the volume of capital services following from longer operating time. So, the only remaining effect on employment is the increase that follows from the lower hourly wage. The formulas above have rested on the assumption that the real pay of a worker in terms of the price of output produced is unchanged when working time increases. If the output price is given, constant nominal pay also holds real pay constant. Most firms cannot, however, sell an increase in output neither in domestic nor in world markets at an unchanged price. Instead, prices must be lowered relative to competitors in order to gain market shares. Such a relative price decrease is indeed the optimal response of a profit-maximising firm to a reduction in its relative wage cost vis-à-vis competitors. One has then to distinguish between changes in the nominal and in the real product wage, as a given percentage reduction in the nominal wage is associated with a smaller percentage reduction in the real product wage (the nominal wage deflated by the firm’s own output price) when the output price falls, too. An analysis of the employment effects of an increase in working time with constant nominal pay per worker in the case when output prices fall, thus requires a slight change in the formulas: the elasticity of labour demand with respect to the real product wage has to be multiplied by the elasticity of the real product wage with respect to the nominal wage (the percentage reduction in the real product wage when the nominal wage falls by one percentage point). The latter elasticity is a positive number below unity (but closer to unity, the closer substitutes in demand the outputs of different firms are). It follows that a higher elasticity of labour demand with respect to the real product wage is required for employment to rise when working time is increased at constant nominal pay per worker when the output price is flexible than when it is fixed. The conclusion that employment always increases when the operating time of capital equals the working time of employees still holds however.
generally, no overlapping of the working times of different employees in order to lengthen operating time relative to working time). In this case, the productivity of a marginal hour is independent of working time: a ceteris paribus change in working time changes the amount of capital services (operating time x the physical capital stock) and the total amount of labour (working time x the number of workers) proportionally by as much and therefore leaves the ratio between the amount of capital services and the total input of labour unchanged. The productivity of a marginal hour then depends only on the ratio between workers and physical capital. Under these conditions, it follows that an increase in working time at unchanged pay per worker must always raise employment: the productivity of a marginal worker, to be set against the constant pay, increases proportionally by as much as working time. This happens already in the short run without any need for restructuring of the capital stock and adoption of new production methods. Therefore, it pays unambiguously for the firm to hire more workers.
9780230_547018_07_cha05.indd 161
3/19/2009 3:49:07 PM
162
Designing the European Model
Working hours and the size of the capital stock A second reason why an increase in working time at unchanged pay could have a positive employment effect already in the short run is that the physical capital stock may not be fixed, even over this time horizon, but may respond to wages. Indeed, this is exactly the case when employees are (credibly) threatened by a closedown of the production site and an outsourcing of jobs abroad unless wage costs are lowered. Then the short-run wage elasticity of labour demand is infinite, that is all jobs will disappear unless hourly wages are cut. This is not, of course, the usual situation in most firms, but it will be the situation at some firms at some points of time: for example, when a car maker makes a “one-shot” decision on at which location to invest in the production of new car models that will replace older ones. Such oneshot investment decisions seem indeed to have been part of the picture in connection with some of the recent agreements on longer working hours in Germany. There is an additional advantage of an increase in working time over a longer time horizon. As the output increase that occurs with fixed pay per worker implies an increase in profits, firms may accumulate capital out of the retained earnings and expand their capital stock faster than would otherwise be the case. Other considerations A full analysis needs to consider a number of additional aspects. One is the use of overtime. Economic modelling explains the use of (paid) overtime with the existence of fixed costs of employment, that is costs per employee that are independent of the length of working time (these include the daily set-up and closing-down costs as well as in-work benefits and costs for training, human resource management, hiring, and firing etc.). Employers have an incentive to economise on these fixed costs through the use of overtime. More exactly, employers trade off the reduction in hourly wage costs that can be achieved by spreading the fixed employment costs over more hours against the rise in costs associated with overtime wage premia and the reduction in each worker’s productivity per hour that will ultimately set in when overtime increases. An increase in standard working time (above which overtime premia are paid) reduces the cost of a marginal standard hour (provided by a marginal worker) relative to the cost of a marginal overtime hour (provided by an intra-marginal worker), because fixed employment costs are spread over more standard hours. Employers then have an incentive to substitute workers for overtime hours,14 which will add to the positive employment effects arising from the incentives to expand output when hourly wages fall. It is difficult to evaluate how important these effects are. Several studies have indicated that actual working hours have fallen as much as standard hours when earlier reductions in working time took place.15 On the other hand,
9780230_547018_07_cha05.indd 162
3/19/2009 3:49:07 PM
Longer Working Hours – The Beginning of a New Trend?
163
it appears, as discussed in Section 1.1, that the difference between actual and standard working time is larger in countries and in sectors with shorter working time. Another complication is that the amount of shiftwork, or more generally the relationship between the operating time of capital and the working time of employees may be influenced by an increase in working time. According to Calmfors and Hoel (1989), an increase in working time makes it profitable for employers to substitute workers for shiftwork. Yet another aspect is that different types of labour may be complements in production. To the extent that this is the case, an increase in working time of, say, specialists who are in short supply may remove bottlenecks in production that increase the demand for other types of labour. A final consideration concerns the product demand side. A frequently asked question is how a firm can find the additional product demand to accommodate an output increase in response to longer working hours at unchanged pay per worker. In our stylised examples with perfectly competitive product markets, this additional demand would be automatically forthcoming as each firm is so small that it could sell any amount of output at the going market price. In a more realistic setting, the firm would have to reduce its output price relative to competitors in order to gain a larger share of the market. Indeed, this would be the profit-maximising response to the reduction in the marginal production cost that takes place when the hourly wage falls. However, the need for a firm to cut prices in response to an increase in working time is lower, the larger is the number of firms that simultaneously increase their output. More output with given pay per worker means higher profit incomes, and higher profit incomes imply higher expenditure of firms and their owners on other firms’ products. In a closed economy, prices on average might not have to fall at all in order to accommodate the output increase, even though relative prices would have to change so as to match the structure of additional demand with the structure of additional output. But in an open economy, which sells some of its products abroad and imports foreign products, a decline in the relative product price vis-á-vis other countries (a real exchange rate depreciation) would be an inevitable and optimal response to the increase in working time. Summarising our discussion of the short-run effects of longer working time at unchanged pay, it is clear that the total number of hours worked, and thus also output, increase. Whether or not the number of employed workers also increases depends on the situation of the individual firm. When the alternative to the agreements is a closing-down of production facilities and an outsourcing of jobs abroad already in the short run, the employment effects are by definition positive. They are also positive in firms where this is not the case, provided that an increase in working time leads to a large enough increase in the utilisation of the capital stock. But in other firms, the short-run employment effects are likely to be negative.
9780230_547018_07_cha05.indd 163
3/19/2009 3:49:07 PM
164
Designing the European Model
5.3.2 Long-run effects of working hours on employment and output What would be the consequences for output and employment in the long run of an economy-wide increase in working time? This question requires an analysis of how both wage-setting incentives and long-run capital accumulation are affected. The following analysis can serve as a benchmark for thinking about the issue. In the long run, the return to capital in the international capital market ties down the domestic return to capital in any open economy. With capital mobility, the capital stock in each country adjusts over time until it obtains the internationally determined rate of return (adjusted for differences in risk among countries). If we, to begin with, again assume a given degree of capital utilisation, this in turn ties down the domestic hourly real wage.16 The world market return to capital thus determines the hourly real wage that is feasible in the long run. A higher hourly wage than the feasible one would result in a progressive reduction in the capital stock, and a lower wage in a progressive increase. In long-run equilibrium, wage-setting behaviour has to be consistent with the rate-of-return requirements imposed by the international capital market. Aggregate employment must be such that it gives the parties in wage bargaining an incentive to choose the feasible hourly wage. An analysis of the longrun employment consequences of a change in working hours should therefore focus on how wage-setting behaviour is influenced (see, for example, Layard et al. 1991; Nickell and Layard 1999; or Calmfors and Holmlund 2000). If, for example, longer working time leads to less pressure to increase hourly wages, lower unemployment is needed in equilibrium to discourage wage setters from raising the hourly wage above the feasible level. A diagrammatical analysis The analytical framework is illustrated in Figure 5.5. The axes of the diagram show the hourly real wage and the number of employed workers, respectively. The horizontal line shows the long-run feasible hourly real wage, which is determined by the world market return to capital. The line can also be interpreted as a (completely elastic) long-run labour demand schedule. It should be distinguished from the short-run labour demand schedule, which is downward-sloping and the position of which depends both on the length of working time and the size of the capital stock (which for the economy as a whole is fixed in the short run). The upward-sloping curve is a wage-setting schedule: it shows how higher aggregate employment gives wage setters an incentive to set a higher hourly wage (because the bargaining position of employees in each firm/sector is strengthened when there are more alternative employment opportunities available). The intersection of the horizontal line, showing the feasible real wage, and the wage-setting schedule is the long-run equilibrium. In the long run, the capital stock adjusts, so that also the short-run labour demand schedule passes through the intersection of the other two schedules.
9780230_547018_07_cha05.indd 164
3/19/2009 3:49:07 PM
Longer Working Hours – The Beginning of a New Trend? Hourly real wage
165
Wage-setting schedule
A
Feasible real wage
B
Short-run labour demand schedule
Employment Figure 5.5
Real wages and employment in the long run
Source: EEAG Group.
One can think of “globalisation” and EU enlargement as an increase in the required long-run return to capital in Western Europe (and Germany), because real capital investment overseas, yielding higher returns, is now an option. This implies a reduction in the feasible hourly real wage, that is a downward shift in the horizontal line in Figure 5.5. At the initial equilibrium at A, a discrepancy between the domestic and the foreign returns to capital arises and capital is moved out. If the current account was initially in balance, the capital outflow is reflected in a current account surplus (as is now the case for Germany and EU-15). This shifts the short-run labour demand schedule gradually to the left. As a consequence, the hourly wage falls.17 This process continues until the capital stock has fallen so much that a new equilibrium is reached at B, where the domestic and foreign returns to capital are again equal. At B both employment and capital stock are lower than in the original equilibrium. Therefore output is also lower.18 Suppose now that working time increases in response to the fall in the feasible hourly wage. How would this affect the long-run equilibrium? In our example, the increase in working time has no repercussions on the feasible hourly wage, as this is uniquely determined by the international return to capital. Therefore, the long-run effect on equilibrium employment depends only on how the wage-setting schedule is affected. If it is left unchanged, equilibrium employment is unaffected. The effect will then simply be that the capital stock and output increase proportionally by as much as working
9780230_547018_07_cha05.indd 165
3/19/2009 3:49:07 PM
166
Designing the European Model
time (compared to what would otherwise be the case).19 If the wage-setting schedule shifts downwards, equilibrium employment increases relative to the situation in B. This implies an increase in output that is proportionally larger than the increase in working time. If the wage-setting schedule instead shifts upwards, equilibrium employment is reduced and output increases proportionally less than working time and may even fall. There will be a one-to-one correspondence between the effects of a change in working time on wage-setting incentives and on long-run equilibrium employment only if the degree of capital utilisation is not affected. The analysis becomes slightly more complicated if we, as in Section 3.1, take into account that longer working time may also increase the operating time of the capital stock. The reason is that such an increase would raise the feasible hourly wage. The intuition is straightforward. If there is an internationally required return to (physical) capital, an increase in the operating time of capital reduces the cost of capital services (the cost per operating hour of the capital stock). This makes it profitable for firms to use more capital services relative to labour. Such a more capital-intensive production raises the productivity of the marginal working hour and thus makes a higher hourly wage consistent with equilibrium in the international capital market. If this occurs – so that the horizontal line in Figure 5.5 shifts upwards again – an increase in working time would have a positive effect on long-run equilibrium employment also in the case of an unchanged wage-setting schedule. Theoretical analysis of working time and wage-setting incentives Our discussion above has shown that the long-run employment effect of longer working hours depends critically on how wage-setting incentives (in terms of the choice of hourly wages) are affected. What do we know about this? Theoretical analysis has pointed to a number of reasons why an increase in working time could give incentives for wage moderation.20 ●
●
●
With longer working hours, employees can attain a given total wage income with a lower hourly wage. Since longer hours at a constant hourly wage would be associated with higher wage income, they tend to make employment more favourable to the individual employee as compared to unemployment. This works in the direction of making unions more concerned about preserving jobs and therefore provides an incentive for them to accept lower hourly wages. An increase in working time increases the share of variable wage costs (wages paid in proportion to working time) relative to fixed costs per employee in total labour costs. This makes both employment and profits more sensitive to increases in the hourly wage. Hence, the costs of a wage rise in terms of lower employment (for unions) and lower profits (for employers) increase. As a consequence, there are incentives on both sides of the labour market to choose lower hourly wages.
9780230_547018_07_cha05.indd 166
3/19/2009 3:49:08 PM
Longer Working Hours – The Beginning of a New Trend? ●
167
Firms may be using high wages as a device to enhance the efficiency of the labour force, for example to discourage quits and in this way keep down hiring and training costs that reduce workers’ net productivity. By increasing output per worker, longer working hours may weaken the incentive of employers to set high hourly wages for such efficiencyenhancing reasons.
However, the literature has also identified effects through which longer working hours tend to strengthen the incentives for high hourly wages: ●
●
Longer hours mean more disutility from work, which works in the direction of making employment less favourable to the individual union member as compared to unemployment. This tends to make unions less concerned about jobs and thus provides an incentive for higher hourly wages on their part. Longer working time per employee also means that the reduction in the number of working hours demanded caused by a rise in the hourly wage is associated with a smaller fall in the number of jobs. This reduces the price of wage increases in terms of lost jobs and thus weakens union incentives for wage restraint.
As there are counteracting effects, theoretical analysis cannot give a clearcut answer to how changes in working time affect wage-setting incentives. The conclusions depend on the exact assumptions made, although there seems to be a bias in the theoretical research towards the conclusion that longer working hours lead to more wage restraint, in which case long-run equilibrium employment must increase. There is also a presumption that this outcome is more likely, the lower is working time initially. Empirical analysis of working time and wage-setting incentives There has also been a fair amount of empirical research on the effects of changes in working time on wage setting. Several studies have included working time as an explanatory variable in wage equations estimated on macroeconomic time series data. Such studies include among others Pencavel and Holmlund (1988) for Sweden, Nymoen (1989) for Norway, Calmfors and Nymoen (1990) for Denmark, Norway and Sweden, Lehment (1991) and Franz and Smolny (1994) for Germany, and Dur (1997) for the Netherlands. All of these studies, with the exception of Lehment, find that longer working hours co-vary with lower hourly wages.21 A problem with these studies, however, is that the strong trends in both wages (upwards) and working time (downwards) over the estimation periods can make the results unreliable. More emphasis should be placed on a few later studies that have used panel data, that is data with both time-series and cross-section variability.
9780230_547018_07_cha05.indd 167
3/19/2009 3:49:08 PM
168 Designing the European Model
Hunt (1999) estimated wage equations for individual employees in Germany, making use of the fact that both the timing and extent of working time reductions in the 1980s and early 1990s differed between sectors. Nordström Skans (2002) exploited instead the fact that working time was cut only for (some types of) shift workers in manufacturing and mining in Sweden in the 1980s, but not for other workers. Kapteyn et al. (2000) instead estimated wage equations for a panel of OECD countries. All three studies come up with very similar results, implying an elasticity of hourly wages with respect to working time of around – 1. This means that an increase in working time by one per cent tends to lower hourly wages also by one per cent. Put differently, the implication is that a working time increase – everything else equal – would leave total pay per worker more or less unchanged, as in the recent company-level agreements in Germany. The empirical research described thus supports the view that working time increases contribute to substantial reductions in wage pressure. This gives a strong presumption that longer working hours do indeed have longrun positive effects on employment. At the same time, a caveat is in place. The estimated wage equations are all in a sense partial-equilibrium analyses, which raises the – theoretical – possibility that a full general-equilibrium analysis could give other results.22 Some partial-equilibrium empirical results on the direct relationship between employment and working time also imply a presumption that longer working hours may help create more jobs when the effects on wagesetting behaviour are taken into account. When estimating reduced-form employment equations for a panel of industries in Germany, Hunt (1999) in many specifications found shorter working hours to reduce employment (for men) significantly. Kapteyn et al. (2000) found a negative, but insignificant, long-run effect on employment of working time reductions in their panel of OECD countries when wage effects were taken into account. According to Crépon and Kramarz (2002), using microeconomic data for individual employees in France, there was a higher unemployment incidence among workers who were affected by the 1982 reduction of the working week than among those who were not.
5.4
Conclusions
Both hours worked per capita and hours worked per employee are low in several continental European countries, such as Germany, France, Belgium and the Netherlands, as compared to the US. This accounts for a large part of the income difference between Western Europe and the US. The low working hours in Europe reflect to a large extent low standard working hours for full-time employees. Recent company-level deals in Germany on longer working time may represent a reversal of the earlier trend towards shorter working time that
9780230_547018_07_cha05.indd 168
3/19/2009 3:49:08 PM
Longer Working Hours – The Beginning of a New Trend?
169
could spread also to other Western European countries with low working hours. It seems that especially the French discussion has been very much affected by the working time developments in Germany, but also employer demands in Belgium and the Netherlands have been influenced. 23 The recent working-time agreements in Germany have implied longer working hours with no, or only partial, compensation and have thus reduced hourly wages. The deals can be seen as cost-cutting measures made necessary by both increased international competition in goods markets and credible employer threats to outsource jobs abroad associated with the on-going “globalisation” in general and EU enlargement in particular. One way of thinking about the deals on longer working time is as a convenient way of reducing hourly wages without reducing the pay per employee. Indeed, this was an important argument in the German discussion that anticipated these deals, However, one can also view them as a labour supply response in collective agreements to a required reduction in the hourly wage, brought about by the forces of globalisation. The desired working time on the part of trade unions is likely to respond differently to a wage change than the desired working time on the part of employees when acting on their own. When the hourly wage falls, the ordinary substitution effect – which tends to reduce desired working time – is counteracted by a weakening of union incentives to restrict working hours in order to promote employment. This weakening occurs because the value for a union member of having a job is smaller if the wage is lower. Under some theoretical assumptions, the response of desired working time on the part of a union is governed by an income effect only. If so, it is in the interest of unions to increase working time if hourly wages have to fall. When analysing the employment and output effects of longer working hours, it is essential to distinguish between the short run and the long run. Already in the short run, longer working hours at unchanged pay will by definition prevent job losses in firms where there is an acute risk of outsourcing production because costs are too high. Also in firms where this is not the case, such agreements will have positive employment effects, provided that longer working time of employees leads to a large enough increase in the utilisation of capital (because the capital stock can be operated for more hours). Indeed, such an increase in capital utilisation is one of the major advantages of longer working hours. However, in other firms where the operating time of capital cannot be increased and where it may take time to adopt new production methods, the employment effects are likely to be negative in the short run. Lower wages per hour will, however, always make it profitable for firms to expand output. In the long run, the hourly real wage level in an open economy must be such that capital earns the same rate of return as abroad. It follows that the longrun feasible hourly wage depends on the world market rate of return to capital. A critical factor for the long-run employment effects of an economy-wide
9780230_547018_07_cha05.indd 169
3/19/2009 3:49:08 PM
170 Designing the European Model
lengthening of working hours is therefore how wage-setting incentives are affected. If longer working time creates stronger incentives for wage moderation, lower unemployment is needed in equilibrium to discourage wages from rising above the feasible level. Although neither theoretical nor empirical research gives unambiguous conclusions, there is a presumption that longer working hours would contribute to wage moderation. If so, one should expect positive employment effects in the long run from longer working time. This would then add to the positive long-run output effects of an increase in working time that would arise already at an unchanged employment level.
Notes * We are grateful for comments on this chapter from Bertil Holmlund, Ann-Sofie Kolm, and Oskar Nordström Skans. 1. An increase in the incidence of part-time work can be associated with both rises and falls in working hours per capita. It will be associated with a fall if the employment rate is constant, but more employees choose – or are forced to choose – part-time instead of full-time work. It will be associated with a rise if it reflects an increase in labour market flexibility leading to a higher employment rate among the population. 2. The table refers to 2005, which is the latest year for which we have data on actual working time. 3. See also a sizeable number of Ifo newspaper interviews on this issue as recorded on www.ifo.de. 4. An agreement on an increase in working time without pay compensation was also concluded in 2004 at the German-owned Bosch factory in Venissieux in France. Currently, longer working hours are being discussed as a cost-cutting measure at the German Opel factory in Rüsselsheim, which is competing within the GM concern with the Swedish Saab factory in Trollhättan about future car production. 5. These obstacles to wage moderation at the sectoral level are similar to the political-economy obstacles to labour market reforms with the aim of promoting wage moderation at the national level. These obstacles were analysed in Chapter 2. 6. This assumes that leisure is a so-called “normal” good. 7. See, for example, Aronsson and Walker (1997) for a recent review of empirical work on labour supply. 8. See, for example, Oswald (1985), Layard and Nickell (1991), Nickell and Layard (1999), Calmfors and Holmlund (2000), or Naylor (2003) for reviews of this literature. 9. Contributions to this literature include among others Calmfors (1985, 1987), Holmlund (1987), Earle and Pencavel (1990), and Booth and Ravaillon (1993). 10. Under some assumptions, legislative regulation of working time produces similar outcomes as collective bargaining. The reason is that employees – making up the political majority – have similar incentives to restrict working hours when they act as voters in the political process as when they act as members of a trade union. This has been analysed by Marimon and Zilibotti (2000), who show that legislation restricting working time relative to the laissez-faire outcome from bargaining between firms and individual workers is in the interest of workers.
9780230_547018_07_cha05.indd 170
3/19/2009 3:49:08 PM
Longer Working Hours – The Beginning of a New Trend?
171
11. This is equivalent to assuming that working hours and workers are perfect substitutes (within a relevant range), so that employers are indifferent to whether a given output is produced by fewer employees working more hours or more employees working fewer hours. See Section 5.1 below. 12 The implicit assumption is then again that leisure is a “normal good,” that is a good for which demand increases when income increases. A similar theoretical result as above holds when working time is determined through legislation and employees make up the political majority. See also footnote 10. 13. Here, as well as below, we make the standard assumption that the production function exhibits constant returns to scale: a, say, ten percent increase in the inputs of both capital and labour (the total number of hours worked) increases output also by ten percent. Under this assumption, the marginal products of both labour (hours) and capital depend only on the capital-labour ratio. 14. See Calmfors and Hoel (1988). 15. This is the result in, for example, Hart and Sharot (1978), De Regt (1988), Hunt (1999), and Kalwij and Gregory (1999). An exception is Nordström Skans, who found that reductions in standard working time for shift workers in Sweden had only a partial effect on actual working time. 16. The assumption of constant returns to scale in production is crucial for this result. In long-run equilibrium, domestic firms use capital up to the point where the marginal product of capital is equal to the real return to capital. Hence, a given real return to capital in the world market determines the ratio between labour (the total number of hours worked = employment x working time) and capital (as the marginal product of capital depends on this ratio). Labour is used up to the point where the marginal product of labour (hours), which also depends on the capital-labour ratio, equals the hourly real wage. It follows that the internationally determined return to capital also determines the hourly real wage. The requirement that the domestic real return to capital (the nominal return deflated by the output price) must equal the world market real return under perfect capital mobility presupposes that domestic output prices follow foreign output prices, which is a reasonable benchmark in the long run. 17. A more elaborate anlaysis would distinguish between different categories of workers. It is then quite likely that only the wage of unskilled workers falls, whereas the wage of skilled workers could rise. See the discussion of outsourcing in Chapter 4. 18. If a fall in employment is to be avoided, the wage-setting schedule must shift downwards by a sufficient amount. A downward shift occurs, for example, if unemployment benefits are reduced, as this makes unemployment more unattractive and hence provides a stronger incentive for wage restraint. Such a reduction in the benefit level takes place automatically if benefits are indexed to wages. The output decrease associated with the movement from A to B in the diagram should be interpreted in a comparative-static sense only. In a growing economy, output at the time the economy has moved to B might be higher than at the time the economy was in A, but it would be lower than would be the case without a fall in the feasible real wage. 19. The hourly real wage given by the lower horizontal line in Figure 5 has to equal the marginal product of labour (hours), which depends on the ratio between capital and the total amount of labour (working time x employment). Longer working time at constant employment implies a larger total amount of labour. Hence, the capital stock has to increase by proportionally as much to keep the marginal product of labour equal to the given hourly wage. Thus, for example,
9780230_547018_07_cha05.indd 171
3/19/2009 3:49:08 PM
172
20.
21. 22.
23.
Designing the European Model ten percent longer working time implies ten percent higher total labour input and hence also ten percent higher capital stock in equilibrium than would otherwise be the case. Then output must also be ten percent higher. See, for example, Layard, Nickell and Jackman (1991) and Konjunkturinstitutet (2000). The theoretical literature includes contributions by, for example, Calmfors (1985, 1987), Hoel and Vale (1986), Holmlund (1987), Booth and Schiantarelli (1987), Booth and Ravaillon (1993), Houpis (1993), Marimon and Zilibotti (2000), and Nordström Skans (2002). Note, however, that Nymoen (1989) and Calmfors and Nymoen (1990) found only short-run but no long-run effects. For example, the empirical studies discussed do not take into account that the level of unemployment benefits is likely in the long run to adjust to the aggregate pay level of employed workers. This will happen, for example, if the replacement rate (the ratio of the unemployment benefit to the pay of an employed worker) is fixed and thus is not affected by a change in average working time. In this case, it is theoretically possible that an increase in working time in an individual firm only would reduce wages there, at the same time as an economy-wide increase in working time would reduce aggregate equilibrium employment. This can, for example, occur in the model of Nordström Skans (2002). Recently, for example, the French government has increased the ceiling for overtime from 180 to 220 hours per year. A new law also permits employers and employees to agree on overtime hours in excess of the legal ceiling.
References Aronsson, T. and J. R. Walker (1997) “The Effects of Sweden’s Welfare State on Labour Supply Incentives,” in R. B Freeman, R. Topel and B. Swedenborg, eds, The Welfare State in Transition. University of Chicago Press, Chicago. Bewley, T. F. (1999) Why Wages Don’t Fall during a Recession, Harvard University Press, Cambridge, MA. Blanchard, O. (2004) “The Economic Future of Europe,” Journal of Economic Perspectives 4, 3–26. Bell, L. and R. Freeman (1994) “Why Do Americans and Germans Work Different Hours?” National Bureau of Economic Research, Working Paper no. 4808. Booth, A. and F. Schiantarelli (1987) “The Employment Effects of a Shorter Working Week,” Economica 54. Calmfors, L. (1985) “Work Sharing, Employment and Wages,” European Economic Review 27. Calmfors, L. (1987) “Wages and Employment in Unionized Economies: Theory and Evidence,” Comment to Bertil Holmlund, in B. Holmlund, K.-G. Löfgren and L. Engström, eds, Trade Unions, Employment and Unemployment Duration, Clarendon Press, Oxford. Calmfors, L. and M. Hoel (1988) “Worksharing and Overtime,” Scandinavian Journal of Economics 90. Calmfors, L. and M. Hoel (1989) “Work Sharing, Employment and Shiftwork,” Oxford Economic Papers 41. Calmfors, L. and R. Nymoen (1990) “Real Wage Adjustment and Employment Policies in the Nordic Countries,” Economic Policy 5. Calmfors, L. and B. Holmlund (2000) “Unemployment and Economic Growth: A Partial Survey,” Swedish Economic Policy Review 7.
9780230_547018_07_cha05.indd 172
3/19/2009 3:49:08 PM
Longer Working Hours – The Beginning of a New Trend?
173
Crépon, B. and F. Kramarz (2002) “Employed 40 Hours or Not-Employed 39: Lessons from the 1982 Mandatory Reduction of the Workweek,” Journal of Political Economy 110. De Regt, E. (1988) “Labour Demand and Standard Working Time in Dutch Manufacturing, 1945–82,” in R. Hart, ed., Employment, Unemployment and Labour Utilization, Unwin Hyman, Boston. Dur, R. (1997) “Arbeidsduurverkorting en lonen in Nederland,” Economisch-Statistische Berichten 82. Earle, J. S. and J. Pencavel (1990) “Hours of Work and Trade Unionism,” Journal of Labor Economics 8. EIROnline (2004) Siemens Deal Launches Debate on Longer Working Hours, http:www.eiro.eurofound.eu.int/print/2004/07/feature/de0407106f.html. Franz, W. and W. Smolny (1994) “Sectoral Wages and Price Formation and Working Time in Germany: An Econometric Analysis,” Zeitschrift für Wirtschafts- und Sozialwissenschaften 114. Gordon, R. J. (2002) “Two Centuries of Economic Growth: Europe Chasing the American Frontier,” paper presented at the Economic History Workshop, Northwestern University. Hart, R. A. and T. Sharot (1978) “The Short-Run Demand for Workers and Hours: A Recursive Model,” Review of Economic Studies 45. Holmlund, B. (1987) “Wages and Employment in Unionized Economies: Theory and Evidence,” in B. Holmlund, K.-G. Löfgren and L. Engström, eds, Trade Unions, Employment and Unemployment Duration, Clarendon Press, Oxford. Hunt, J. (1999) “Has Work-Sharing Worked in Germany?” Quarterly Journal of Economics 114. Kalwij, A. S. and M. Gregory (1999) Overtime Hours in Great Britain over the Period 1975–1998: A Panel Data Analysis, Working Paper, University of Oxford. Kapteyn, A., A.S. Kalwij and A. Zaidi (2000) The Myth of Worksharing, IZA, Discussion Paper no 188. Konjunkturinstitutet (2000) Makroekonomiska effekter av en arbetstidsförkortning, Stockholm. Kölling, A. and K. Lehmann (2002) “Arbeitszeitregelungen und Tarifbindung,” BeitrAB 251. Layard, R., S. Nickell and R. Jackman (1991) Unemployment: Macroeconomic Performance, Oxford, University Press Oxford. Lehment, H. (1991) “Lohnzurückhaltung, Arbeitzeitverkürzung und Beschäftigung, Eine empirische Untersuchung für die Bundesrepublik Deutschland 1973–1990,” Die Weltwirtschaft, February. Lehndorff, S. (2004) “On the Rough Road: Collective Bargaining on Working Time in Germany,” paper presented at the WES 2004 Conference, Manchester. Marimon, R. and F. Zilibotti (2000) “Employment and Distributional Effects of Restricting Working Time,” European Economic Review 44. McDonald, I. M. and R. M. Solow (1981) “Wage Bargaining and Employment,” American Economic Review 71. Münchau, W. (2004) “Working Longer to Save Jobs: But Will It Help Europe Close the Productivity Gap?” Financial Times 23 July. Naylor, R. (2003) “Economic Models of Union Behaviour,” in J. T. Addison and C. Schnabel, eds, International Handbook of Trade Unions, Edward Elgar, Cheltenham. Nickell, S. (2003) “Employment and Taxes,” mimeo, Venice International Institute. Nickell, S. and R. Layard (1999) “Labor Market Institutions and Economic Performance,” in O. Ashenfelter and D. Card, eds, Handbook of Labor Economics vol. 3, North Holland, Amsterdam.
9780230_547018_07_cha05.indd 173
3/19/2009 3:49:09 PM
174
Designing the European Model
Nordström Skans, O. (2002) Labour Market Effects of Working Time Reductions and Demographic Changes, Dissertation Series 2002:2, IFAU, Uppsala. Nymoen, R. (1989) “Wages and the Length of the Working Day: An Empirical Test on Norwegian Quarterly Manufacturing Data,” Scandinavian Journal of Economics 91. OECD (1998) Employment Outlook, Paris. OECD (2003) The Sources of Economic Growth in OECD Countries, Paris. OECD (2004) Employment Outlook, Paris. Olovsson, C. (2004) Essays On Dynamic Macroeconomics, Doctoral Thesis, Institute for International Economic Studies, Stockholm University. Oswald, A. (1985) “The Economic Theory of Trade Unions: An Introductory Survey,” Scandinavian Journal of Economics 87. Pencavel, J. and B. Holmlund (1988) “The Determination of Wages, Employment and Work Hours in an Economy with Centralised Wage-Setting: Sweden 1950–1983,” Economic Journal 98. Prescott, E. C. (2004) “Why Do Americans Work So Much More than Europeans?” NBER Working Paper 10316, Cambridge, MA. Sinn, H.-W. (2003a) “Wieder 42 Stunden arbeiten,” Frankfurter Allgemeine Zeitung, 23 July 2003, also “Working Longer,” Ifo Viewpoint no. 47, 25 June 2003, www.cesifo.de. Sinn, H.-W. (2003b) Ist Deutschland noch zu retten? Econ, Munich, editions 1–3. Sinn, H.-W. (2004a) Ist Deutschland noch zu retten? Econ, Munich, editions 7 and 8. Sinn, H.-W. (2004b) “Warum wir länger arbeiten müssen,” Die Welt am Sonntag no. 46, 14 November 2004, also: “Why Extending Working Hours Will Create More Jobs,” Ifo Viewpoint no. 59, 18 November 2003, www.cesifo.de. Sachverständigenrat zur Begutachtung der gesamtwirtschaftlichen Entwicklung (2003) Jahresgutachten 2003/04, Wiesbaden. Saint-Paul, G. (1993) “On the Political Economy of Labour Market Flexibility,” NBER Macroeconomics Annual. Saint-Paul, G. (1998) “A Framework for Analyzing the Political Support for Active Labor Market Policy,” Journal of Public Economics 67. Zimmermann, K. F. (2004) “Does More Work Create More Work?” editorial, Economic Bulletin 8, DIW Berlin and Springer-Verlag.
9780230_547018_07_cha05.indd 174
3/19/2009 3:49:09 PM
Part II Macroeconomic Issues
9780230_547018_08_cha06.indd 175
3/19/2009 8:14:12 PM
This page intentionally left blank
6 Fiscal Policy and Macroeconomic Stabilisation in the Euro Area: Possible Reforms of the Stability and Growth Pact and National Decision-Making Processes
The recent economic-policy debate in the EU has largely focused on fiscal policy and the Stability and Growth Pact. The reason is the current budgetary problems of some member states. Portugal breached the three-per centof-GDP deficit ceiling in 2001 and 2002. Germany breached it in 2002, and may also do so in 2003. France and Italy have abandoned their commitments to earlier agreed budget objectives and there is a clear threat that they may violate the deficit ceiling, too. These events have contributed to a revival of the debate on the fiscal policy framework in the EU. The European Commission has recently proposed a number of changes in the Stability and Growth Pact (European Commission 2002b). There have also been calls for more fundamental revisions of the EU fiscal policy framework including proposals to scrap the Stability and Growth Pact altogether (see, for example, Financial Times 2002a, b, c; The Economist 2002; De Grauwe 2002; or Walton 2002). A key issue is the need to combine long-run sustainability of fiscal policy with short-run flexibility as a tool for macroeconomic stabilisation. The current EU fiscal rules mainly reflect a desire to enhance long-run fiscal discipline. This is explained by the earlier rapid accumulation of government debt, but also by the view that discretionary fiscal policy is unsuitable as an instrument of countercyclical stabilisation. At the same time, the risk of asymmetric cyclical developments in individual euro countries creates a potential need for using national fiscal policy as a stabilisation tool. This has led to a criticism that the EU fiscal rules are too rigid and hamper the use of fiscal policy for stabilisation purposes in an inappropriate way. The rules have also been criticised for being arbitrary. At the same time, the current 177
9780230_547018_08_cha06.indd 177
3/19/2009 8:14:12 PM
178 Designing the European Model
discussion illustrates very clearly that it may be very difficult to apply the rules in a sufficiently disciplining way when they are put to a real test. One aim of this chapter is to analyse what role fiscal policy should play as a stabilisation tool in the euro area. In line with other recent contributions, we argue that this role should be larger than according to the conventional wisdom that has prevailed in recent years. We also argue that this requires changes in the fiscal policy framework at the EU level. Recent proposals of the European Commission seek to make the EU fiscal rules more flexible through changes in the interpretation of the Stability and Growth Pact and a greater reliance on discretionary judgements. We argue that this approach is potentially harmful and that more fundamental reforms, implying changes in the Maastricht Treaty, are required. There is a case for letting the deficit ceiling in the Treaty depend on the debt level in a transparent way. Countries with low debt should be allowed to run larger budget deficits than three per cent of GDP. This would serve both to give low-debt countries greater scope for stabilisation policy and to enhance the incentives for fiscal discipline. Another recommendation is that the decisions on sanctions against EMU member states that violate the deficit ceiling should be moved from the political level of the Ecofin Council1 to the judicial level of the European Court of Justice. This would make the enforcement of the fiscal rules more credible. Changes in the EU fiscal rules involve difficult trade-offs. On the one hand, if the rules are perceived as too rigid and arbitrary by the public, their legitimacy will gradually evaporate and they will become unsustainable. On the other hand, the credibility of the fiscal policy framework must not be undermined by an impression that the rules are changed in a discretionary way as soon as they begin to bite, especially for the large EU countries. There is a continued need for fiscal rules at the EU level to enhance the incentives for fiscal discipline. The future burdens of ageing populations make large reductions of government debt highly desirable. Therefore, the medium-term budget objective of “close to balance or in surplus” should be maintained, although it should be made explicit that it is set in cyclically adjusted terms. Current events have, however, illustrated that there are limits to how much fiscal rules at the EU level can achieve on their own. Governments are likely always to shun away from political conflicts about each others’ fiscal policies. This is an argument for relying to a much larger extent than hitherto on national institutions that are conducive to both long-run fiscal discipline and effective short-run stabilisation policy. This can be seen as an application of the general principle of subsidiarity (see Chapter 10). One step in this direction would be to require the member states to adopt laws on fiscal policy that set well-defined long-run sustainability goals which are consistent with the common EU objectives, but also outline clear principles for the use of fiscal policy as a stabilisation tool.
9780230_547018_08_cha06.indd 178
3/19/2009 8:14:12 PM
Fiscal Policy and Macroeconomic Stabilisation 179
Although it is clearly not on the political agenda now, one should also explore the future possibilities of reforming the national decision processes for fiscal policy along the lines of monetary policy, as has recently been suggested in a number of contributions. Notwithstanding that the idea may seem unfamiliar to many, there is a case for trying to separate fiscal policy decisions with the aim of stabilising the economy from other types of fiscal policy decisions. One possibility might be delegation to an independent national fiscal policy authority. Such delegation would be in line with developments in other areas such as competition policy as well as financial regulation and supervision, where in many countries the operational conduct of policy has been delegated to various bodies, and politicians have focused more on setting the overall objectives. Delegation of national fiscal policy decisions could also be seen as an alternative to the larger role in fiscal policy surveillance desired by the European Commission (European Commission 2002b). The underlying motive for delegation at the national level is a desire to mitigate the problems hampering the use of fiscal policy as a stabilisation instrument: long decision lags, irreversibility of decisions, risks of contributing to a deficit bias, “confounding” of various objectives and so on. However, it remains an open question whether one could find forms of fiscal policy delegation that would be accepted by the general public. The most realistic possibility in the near future might be to require the government to consult with an independent fiscal policy committee before making its budget decisions and to base these on the committee’s estimates of the cyclical situation and its forecasts of government revenues and expenditures. The chapter is structured as follows. Section 1 reviews the role of fiscal policy as a tool of stabilisation policy both in general and in the specific EMU context. Section 2 analyses the case for modifications of the current fiscal policy framework in the EU. Section 3 discusses the possibilities to strengthen the fiscal policy framework at the national level and the case for letting an independent fiscal authority play a larger role.
6.1
Fiscal policy as a stabilisation tool
The perception of the role of fiscal policy has changed radically over recent decades. Discretionary fiscal policy to stabilise the economy has come to be regarded with great scepticism. Instead, the conventional wisdom today is that monetary policy should be the main stabilisation tool. One explanation of this development is, of course, the large accumulation of government debt in most OECD countries in the 1980s and early 1990s, which is unprecedented in peacetime (see Figure 6.1). As a consequence, fiscal sustainability has become the main fiscal policy issue, and major reforms of the fiscal policy framework have been undertaken in nearly all OECD countries. These reforms include both various
9780230_547018_08_cha06.indd 179
3/19/2009 8:14:12 PM
180
Designing the European Model in percent of GDP
80 75
European Union
70
OECD total
65
United States
60 55 50 45 40 35 30 70 Figure 6.1
72
74
76
78
80
82
84
86
88
90
92
94
96
98
00
02
Gross government debt/GDP, 1970–2003
Source: OECD, Economic Outlook 72 (December 2002).
fiscal policy rules, like the ones in the Maastricht Treaty and the Stabilisation and Growth Pact, and procedural changes in the national budget processes in various countries (see, for example, Kopits and Symansky 1998; and von Hagen et al. 2002). In terms of academic research, the problems with accumulating government debt have stimulated a large body of literature explaining why unconstrained fiscal policy may involve a permanent deficit bias (see Alesina and Perotti 1995, for a survey). Two major types of theoretical objections have been raised against using fiscal policy for stabilisation purposes. The first one questions the technical effectiveness of such policies. The second objection questions the ability of policy makers to use fiscal stabilisation policy in an effective way. The discussion of the technical effectiveness of fiscal policy takes its starting point in the notion of so-called Ricardian equivalence (see Elmendorf and Mankiw 1999, for a survey). The argument is that tax reductions or transfer increases that raise the disposable incomes of households will fail to increase private consumption if they involve larger budget deficits: households will realise that their life-cycle incomes have not increased, as they will have to pay for the deficits through higher taxes or lower transfers in the future. Another argument holds that tax reductions or government expenditure increases could even give rise to perverse negative demand effects if they are associated with credibility problems that lead to increased interest rates or to expectations of future “crisis adjustments” that will lower lifecycle incomes (see, for example, Giavazzi and Pagano 1996; and Giavazzi et al. 2000).
9780230_547018_08_cha06.indd 180
3/19/2009 8:14:12 PM
Fiscal Policy and Macroeconomic Stabilisation 181
There are a number of arguments why discretionary fiscal policy may be used in a less effective way as a stabilisation tool than monetary policy. ●
●
●
●
Decision lags are longer, as tax and expenditure changes have to go through a lengthy parliamentary decision-making process, which is usually annual in contrast to the almost continuous decision-making process for monetary policy. The political character of fiscal policy decisions makes it much harder to reverse decisions when circumstances change than is the case for monetary policy (Taylor 2000). Fiscal policy has other central goals than stabilisation, viz. income distribution and resource allocation. In addition, fiscal policy measures are often influenced by attempts of incumbent governments to enhance their reelection chances. Hence there is the serious risk that the stabilisation aspects will carry a low weight. The risk of an expansionary bias is much larger for fiscal policy than for monetary policy, as the former is run by policy-makers engaged in day-today politics, whereas the latter has been delegated to independent central banks, which can take a more long-run view.
The conclusion has been that fiscal stabilisation policy is likely to be badly timed and procyclical, especially in booms because it is politically much more difficult then to pursue restrictive policies than it is to pursue expansionary policies in recessions. 2 The current conventional wisdom is that fiscal policy should mainly be confined to letting the automatic stabilisers work, that is to let the automatic variations in the budget balance that follow from the cyclical variations in tax receipts and some government expenditures, such as unemployment benefits and costs for labour market programmes, dampen the business cycle (see, for example, Taylor 2000; Buti and Martinot 2000; Buti and Giudice 2002; or European Commission 2002d). Because of their “automatic” character, such policies do not raise the same problems of decision lags, deficit bias and problems of reversing policies in new situations as have traditionally been associated with discretionary fiscal policy. The size of the automatic stabilisers is positively related to the share of government expenditure in GDP, the degree of tax progressivity and the generosity of unemployment compensation. Recent empirical research does indeed confirm that large government sectors, and thus large automatic stabilisers, reduce output volatility to a significant degree (Galí 1994; Fatàs and Mihov 2001, 2002).3 6.1.1
The case for discretionary fiscal policy
Adopting a more long-term perspective, it has been noted that there have been large swings over time in the relative emphasis given to fiscal and monetary policy as stabilisation tools (Wyplosz 2002b). This raises the
9780230_547018_08_cha06.indd 181
3/19/2009 8:14:12 PM
182
Designing the European Model
possibility that the present downplaying of fiscal policy as a stabilisation instrument may have gone too far. Indeed, this has recently been suggested in a number of contributions, such as Ball (1997), Wren-Lewis (2000, 2002), von Hagen and Mundschenk (2001), Seidman (2001), Blanchard and Perotti (2002), Fatàs and Mihov (2002), and Wyplosz (2002b), who all have argued for a revival of the role of fiscal policy. There are several reasons why relying on discretionary monetary policy may not be enough in many situations. A well-known argument is that monetary policy may be impotent in a depression when it can be caught in a liquidity trap, because it is impossible to achieve negative real interest rates in situations with falling prices. Japan is an obvious case in point (see the Box on Japan in Chapter 1 of 2003 EEAG report). But also in more normal situations, there may be limitations to monetary policy because central banks are reluctant to change interest rates by much in the short run, as this would imply large variations in the prices of outstanding debt (interest rate smoothing). It has also been claimed that fiscal policy can more easily be targeted in a desirable way than monetary policy. Ball (1997) and Wren-Lewis (2000, 2002) argue that fiscal policy can be designed so as to have more even effects across the economy than monetary policy, which will have a greater impact on construction and investment goods sectors than on service sectors. These more even effects may be desirable in some situations. In other situations, one may want to target measures more specifically, for example to counteract a real property price boom, which may be easier through targeted fiscal policy (such as reductions of tax relief for mortgage interest rates) than through monetary policy. The strongest argument in favour of fiscal policy in the euro area is the risk of asymmetric cyclical developments in individual countries. In the event of such macroeconomic disturbances, national fiscal policy is the only remaining stabilisation policy tool. There is an obvious case for such stabilisation policy in recessions, as money wages, and thus also prices, tend to be rigid downwards (Calmfors 1998; Calmfors and Sonnegård 2001). The case for stabilisation policy in booms is somewhat more complex. A relative price increase vis-à-vis other countries (a real exchange rate appreciation) may be a proper adjustment to macroeconomic shocks that raise output growth. Whether or not this is the case depends, as was discussed in the 2002 EEAG report, on the character of the shocks. In the case of permanent structural changes, such as a permanent increase in the relative demand for a country’s output or a permanent increase in the relative productivity of a country, prices could just be left “to do their job.” But this is not an appropriate response in the case of a temporary asymmetric demand increase. The main reason is again downward money wage rigidity: inflation in a temporary boom tends to cause “permanent” wage increases that
9780230_547018_08_cha06.indd 182
3/19/2009 8:14:13 PM
Fiscal Policy and Macroeconomic Stabilisation 183
are hard to reverse and therefore “lock in” real exchange rate appreciations. This makes it more difficult to stabilise the economy in the next recession, as the real exchange rate appreciation requires a more expansionary fiscal policy with larger budget deficits than would otherwise be the case (Swedish Government Commission on Stabilisation Policy in the EMU 2002). The main rationale for fiscal stabilisation policy in a boom is thus intertemporal considerations relating to future stabilisation possibilities.4 A related intertemporal argument for fiscal stabilisation policy in booms in the EMU countries is the risk of strong asset price reversals, that is boombust cycles in asset prices where first large asset price rises reinforce the upswing and then large asset price falls exacerbate the downswing. Again, the most obvious recent example is Japan. The macroeconomic consequences of such excessive asset price volatility has recently been studied by Bordo and Jeanne (2002), who find boom-bust cycles to be much more common in real property prices than in stock prices and to be associated with large cyclical swings in output. This finding is highly relevant for stabilisation policy in the EMU countries, because cycles are much more likely to be country-specific in real property prices than in stock prices. Interestingly, Bordo and Jeanne (2002) also find boom-bust cycles in real property prices to be more common in small than in large countries, which they explain by the larger relative importance of local markets in small countries (the relative size of the Stockholm area in Sweden is larger than that of the Berlin area in Germany). Why are automatic stabilisers not likely to be a sufficient fiscal policy tool in the case of large cyclical asymmetries in the euro area? There are a number of reasons. ●
●
By their very nature automatic stabilisers can only cushion macroeconomic shocks, but not fully offset them. According to most estimates, automatic stabilisers reduce output fluctuations by around a third in the EU countries (see, for example, van den Noord 2000; and European Commission 2002d). As discussed above, the size of the automatic stabilisers is positively related to the share of government expenditure in GDP, the degree of tax progressivity and the generosity of unemployment compensation. But the decisions on such structural parameters have not been influenced much by stabilisation concerns: instead the size of automatic stabilisers is a by-product of other considerations, such as preferences over private versus public consumption or over income distribution versus allocative efficiency. There is no reason, therefore, to believe that the automatic stabilisers give an optimal degree of stabilisation. Nor should one expect that differences in the size of the automatic stabilisers among countries, for example due to differences in the share of government expenditures in
9780230_547018_08_cha06.indd 183
3/19/2009 8:14:13 PM
184
●
●
Designing the European Model
GDP, as shown in Table 6.1, reflect differences in the preferences for stabilisation. Countries with small automatic stabilisers may therefore want to use discretionary fiscal policy more than others. Structural reforms in the European economies with the aim of raising long-run employment and growth have weakened the automatic stabilisers. This is evident from Table 6.1, which shows the reduction in the size of government in recent years (from an unweighted average in the EU of 45.0 per cent of GDP in 1994 to 41.7 per cent in 2002). In addition, tax progressivity and the generosity of unemployment benefits (mainly in terms of coverage but also to some extent in terms of benefit levels and duration) have – for good reasons – been reduced (see EEAG 2002; as well as the Box on Germany in Chapter 1 of 2003 EEAG report). Finally, if there are permanent supply shocks, the automatic stabilisers tend to prolong the adjustment process and cause budget effects that must ultimately be eliminated through discretionary action.
Table 6.1 Government spending, excluding interest payments, as a percentage of GDP in the EU countries 1994
1998
2001
2002
Austria Belgium Germany Denmark Spain Finland France Greece Ireland Italy Luxembourg Netherland Portugal Sweden United Kingdom
49.2 41.5 43.1 54.7 na 56.4 48.5 32.1 36.6 41.7 43.7 43.2 36.6 62.9 40.0
47.0 40.7 42.7 51.5 35.2 46.4 46.7 34.9 29.9 39.8 40.9 39.2 36.7 52.7 34.6
47.7 40.3 42.7 48.9 34.6 43.6 45.9 36.6 29.9 40.5 39.5 39.3 38.9 50.1 36.3
48.2 40.4 43.0 49.5 34.8 44.2 46.6 37.2 31.4 40.8 43.3 40.3 38.4 50.9 37.0
Unweighted average Standard deviation Coefficient of variation
45.0 8.2 0.18
41.3 6.4 0.15
41.0 5.4 0.13
41.7 5.4 0.13
Note: The budget balance can be written B = tY – G, where B is the budget surpuls, t is the tax rate, Y is real GDP and G is real government expenditure. Dividing by Y, we have b = B/Y = t – G/Y, where b is the budget balance as a ratio of GDP. Differentiation with respect to b and Y gives db = (G/Y) dY/Y. So, G/Y is the semi-elasticity of the budget surplus as a percentage of GDP with respect to real output, that is it indicates by how many percentage points of GDP the fiscal balance improves when output rises by one percent. Source: OECD Economic Outlook 72 (December 2002).
9780230_547018_08_cha06.indd 184
3/19/2009 8:14:13 PM
Fiscal Policy and Macroeconomic Stabilisation 185
6.1.2 How effective is fiscal policy as a demand management tool? The discussion on Ricardian equivalence has contributed to the impression that fiscal policies are not very effective (see Elmendorf and Mankiw 1999). Much of this discussion is, however, rather superficial as it tends to lump together various fiscal policies that should theoretically be expected to have very different effects. The ineffectiveness postulate of Ricardian equivalence applies only to tax and transfer changes that affect the real disposable income of households. The postulate holds only under very restrictive assumptions: households must have a long enough time horizon for taking into account offsetting future tax and transfer changes and they must not be credit-constrained. One would expect the former assumption to be more valid in situations of pressing government debt problems when the public debate focuses on sustainability issues (as in the 1990s) than in more normal times (like now). One would always expect those tax and transfer changes to be effective that are targeted on low-income groups, which to a large extent are creditconstrained (Wren-Lewis 2000, 2002). Such targeting may also be motivated from a welfare point of view, as these groups are more exposed to cyclical income volatility than groups with higher incomes (Storesletten et al. 2001). Most empirical evidence seems to support substantial demand effects of tax changes. The evidence that automatic stabilisers, which work mainly on the tax side, reduce the volatility of output and consumption, is not consistent with Ricardian equivalence (Galí 1994; Fatàs and Mihov 2001, 2002). Blanchard and Perotti (2002) recently found a multiplier of close to one for discretionary tax changes in the U.S., whereas other studies have found somewhat lower multipliers (Wren-Lewis 2000, 2002; Wijkander and Roeger 2002; Swedish Government Commission on Stabilisation Policy in the EMU 2002; European Commission 2002d). Temporary changes in government consumption also have an effect on aggregate demand under Ricardian equivalence. This is obvious if an increase in current government consumption is financed through a reduction in future government consumption, as this involves no change in the taxes paid by households and hence no change in private consumption. But a similar conclusion holds also if a temporary increase in government consumption is financed through future taxes. The explanation is that the short-run direct demand effects are larger than the short-run effects on private consumption due to future tax changes: this is so because the changes in private consumption resulting from the changes in lifetime incomes will be spread over the whole future, as households want to smooth consumption over time, whereas the entire change in government consumption occurs in the short run. The positive output effects of increases in government consumption have been confirmed in a number of recent empirical studies (for example, Rotemberg and Woodford 1992; Ramey and
9780230_547018_08_cha06.indd 185
3/19/2009 8:14:13 PM
186
Designing the European Model
Shapiro 1997; Edelberg et al. 1998; Fatàs and Mihov 2002; and Blanchard and Perotti 2002). In most cases multipliers around one are found. Some of the studies find that increases in government consumption are associated with increases in private consumption – and not decreases as implied by Ricardian equivalence (Blanchard and Perotti 2002; Fatàs and Mihov 2002). Other work has found that the fiscal multipliers for government consumption are larger than for income taxes (Wren-Lewis 2000, 2002; Wijkander and Roeger 2001; European Commission 2002d). In the economic-policy discussion there is a tendency to associate fiscal policy mainly with measures that affect aggregate demand either via direct expenditure changes or via tax and transfer changes that have an impact on the disposable income of households. But fiscal policy can also work by changing relative prices. One such policy, which has been used at times in Sweden, is temporary changes in VAT (see also Blinder 2001; Wijkander and Roeger 2002; European Commission 2002d; Swedish Government Commission on Stabilisation Policy in the EMU 2002; and Wren-Lewis 2002). A temporary change in VAT affects private consumption in a similar way as a change in the real interest rate: by changing the relative price between consumption in different time periods, households are induced to reallocate consumption spending intertemporally. One could also conceive of a similar use of investment taxes or subsidies to affect the timing of private investment (Swedish Government Commission on Stabilisation Policy in the EMU 2002; WrenLewis 2002). There is some evidence that the fiscal multipliers are considerably larger for VAT than for income taxes (Wren-Lewis 2000, 2002; Wijkander and Roeger 2002; European Commission 2002d). Cross-border trade usually limits the possibilities to set VAT rates according to national priorities. But this does not apply in the same way to temporary VAT changes as a stabilisation tool in the case of asymmetric cyclical developments. On the contrary, if a temporary rise in national VAT in a boom shifts consumption purchases abroad, this, too, tends to reduce demand domestically. A potential risk of using temporary VAT increases to dampen an asymmetric boom in a euro country is, however, that they could trigger “permanent” wage increases, although the risk is much smaller than in the case of permanent VAT changes. Another possibility, which has also been overlooked in much of the international discussion, is to use temporary variations in the payroll taxes levied on employers as a discretionary stabilisation tool. By changing wage costs, such a policy directly affects the real labour cost and the real exchange rate. It thus represents a way of letting prices do their job in the case of asymmetric cyclical developments. However, since real exchange rate changes are known to affect trade volumes with substantial lags, such a policy would seem relevant mainly in the case of relatively drawn-out disturbances. It is not only temporary reductions in payroll taxes in downswings that may be of interest. In fact, temporary rises in employers’ payroll taxes may be a very
9780230_547018_08_cha06.indd 186
3/19/2009 8:14:13 PM
Fiscal Policy and Macroeconomic Stabilisation 187
appropriate policy if an individual euro country experiences an asymmetric boom. The reason is that higher payroll taxes for employers raise domestic wage costs and output prices, but not domestic wages. On the contrary, wages are likely to fall to the extent that the demand for domestic output falls and the tax is shifted backwards on to employees because the “room for wage increases” is reduced.5 The upshot is that a temporary increase in employers’ payroll taxes may be a desirable way of dampening a boom, because wage costs are temporarily raised at the same time as the risk is reduced that wages are bid up more permanently. The idea of using cyclical variations in employers’ payroll taxes in this way has large similarities with the system of so-called buffer funds that was set up in Finland in connection with the entry into the EMU. According to this system, funds have been built up through a temporary increase in various employer contributions to the social security system with the intention to use the proceeds of these funds to hold down contributions in downswings (Calmfors 1998; Holm et al. 1999; Swedish Government Commission on Stabilisation Policy in the EMU 2002).6 Our conclusion is that discretionary fiscal stabilisation policy is potentially effective and that occasions are likely to arise in the euro countries when its use would be desirable. We are not thinking about “fine tuning,” but about countercyclical stabilisation in the event of major macroeconomic disturbances. This does, however, require a fiscal policy framework that prevents misuse of stabilisation measures that cause excessive debt accumulation. It also requires addressing the problems of long decision lags and irreversibility of fiscal policy measures. The problem of decision lags is perhaps most obvious in the case of temporary VAT changes: it is a serious problem if, for example, a temporary increase in VAT in a boom can be decided only in a lengthy political process, as the anticipation of the measure will lead to effects that are the reverse of those desired in the period before the measure came into force. The problem of irreversibility is most clear-cut for increases in government consumption in a recession. As we have argued, this is likely to be a more effective stabilisation tool than general cuts in personal income taxes, but there is a serious risk of political “ratchet effects” making it impossible to reduce government consumption again in the next boom (Wijkander and Roeger 2002). We shall return to these issues in Section 3, after first having discussed the fiscal rules at the EU level and their impact on the possibilities to pursue stabilisation policies.
6.2
Possible reforms of EU fiscal rules
The raison d’être for the fiscal rules in the EU is the desire to ensure longrun sustainability of public finances, which came under threat in the 1980s and early 1990s because of the rapid build-up of government debt in most member countries. This is shown in Table 6.2.
9780230_547018_08_cha06.indd 187
3/19/2009 8:14:13 PM
188
Designing the European Model
Table 6.2 Gross government debt as a percentage of GDP in the EU countries, 1980–2003 1980
1990
1995
2000
2001
2002
2003
Belgium Denmark Germany Greece Spain France Ireland Italy Luxembourg Netherlands Austria Portugal Finland Sweden United Kingdom
78.6 36.5 31.7 25.0 16.8 19.8 75.2 58.2 9.3 46.0 36.2 32.3 11.5 40.3 53.2
129.2 57.8 43.5 79.6 43.6 35.1 101.5 97.2 4.4 77.0 57.2 58.3 14.3 42.3 34.0
133.9 69.3 57.0 108.7 63.9 54.6 82.6 123.2 5.6 77.2 69.2 64.3 57.2 76.2 51.8
109.2 46.8 60.2 106.2 60.5 57.3 39.1 110.6 5.6 55.8 63.6 53.3 44.0 55.3 42.1
107.6 44.7 59.5 107.0 57.1 57.3 36.4 109.9 5.6 52.8 63.2 55.5 43.4 56.6 39.1
105.6 44.0 60.9 105.8 55.0 58.6 35.3 110.3 4.6 51.0 63.2 57.4 42.4 53.8 38.5
101.7 42.4 61.8 102.0 53.2 59.3 35.0 108.0 3.9 50.1 63.0 58.1 41.9 51.7 38.1
Unweighted average GDP-weighted average Standard deviation Coefficient of variation
38.0 38.0 20.5 0.5
58.3 54.4 32.5 0.6
73.0 70.2 30.2 0.4
60.6 64.1 27.5 0.5
59.7 63.0 27.7 0.5
59.1 63.0 27.8 0.5
58.0 62.5 26.9 0.5
Note: The 1980 and 1990 figures for Germany refer to west Germany. Source: European Commission (2002c).
There are two main motives for fiscal rules at the EU level. The first – and in our view the most important – motive is to enhance fiscal discipline in general. The need for this has been emphasised by an extensive politicaleconomy literature, which has pointed out how a large number of factors may cause a deficit bias (see, for example, Alesina and Perotti 1995; or von Hagen et al. 2002). These factors include: (i) fiscal illusion on part of the general public; (ii) the fact that it is politically more popular to stimulate demand in recessions than to restrain it in booms; (iii) the use of debt by incumbent governments as a strategic variable to favour their own constituencies and constrain the policies of future governments in favour of other constituencies; (iv) distributional conflicts; (v) lobbying by local constituencies for targeted benefits, the costs of which are shared nationally; and (vi) problems of time inconsistency, according to which governments cannot resist ex post the temptation to abandon sound fiscal policy even if it is clear ex ante that this is inappropriate. The desire to strengthen budgetary discipline in general is not related to EMU per se. Rather, in a situation of general fiscal profligacy, monetary unification offered a unique opportunity to establish constraints on government budget deficits and debt accumulation at the EU level. As the creation of EMU required the set-up of new institutions anyway, it appeared much easier to establish such rules at the EU level than to initiate national reform
9780230_547018_08_cha06.indd 188
3/19/2009 8:14:13 PM
Fiscal Policy and Macroeconomic Stabilisation 189
processes, which could more easily be blocked by various vested interests. In this perspective, the role of an “external enforcer” of budgetary discipline that the EU has assumed can be seen as an outcome of very specific historic circumstances. The second motive for fiscal rules at the EU level is the moral-hazard problems that can arise in a monetary union because fiscal policies in one member state have spillover effects on the other states. A number of such spillover effects have been identified (see, for example, Buiter et al. 1993; or Beetsma 2001). There is a potential risk that other governments could in the end feel forced to bail out a bankrupt government of an individual member country. There is a risk for pressures on the ECB of both direct (buying up the debt of a highly indebted country in the secondary market) and indirect bail-outs (setting lower interest rates than are motivated by price stability considerations). Finally, the recent “fiscal theory of price determination” emphasises the risk that the ECB will be unable to control inflation if fiscal policies are not sustainable (Canzoneri and Diba 2001). The argument starts from the observation that solvency of the government requires the discounted value of future primary surpluses (including seigniorage revenue of the central bank) to match the outstanding real value of government debt. If fiscal policy violates this constraint, and monetary authorities do not relax their policy stance, solvency can only be maintained by an upward jump of the price level, which lowers the price of nominal government debt in terms of goods and thus reduces the real value of the debt. 6.2.1 The present framework The fiscal rules in the EU are determined mainly by the provisions in the Maastricht Treaty on the excessive deficit procedure (Article 104.3) and by the Stability and Growth Pact (SGP), which is embodied in two regulations of the Ecofin Council and two resolutions of the European Council (see, for example, Buti et al. 2002).7 The Treaty sets out the basic stipulations, whereas the SGP defines their operational content. The main rules are as follows: ●
The Treaty sets a deficit ceiling (a reference value) of three per cent of GDP for the actual government budget balance. Larger deficits are considered “excessive” unless “the excess over the reference value is only exceptional and temporary and the ratio remains close to the reference value.” The formal wording of this escape clause must be regarded as quite stringent, as all the conditions in it must in principle be fulfilled for it to apply (Balassone and Franco 2001; Buti and Giudice 2002). According to the SGP, the exceptionality condition can refer either to “an unusual event outside the control of the Member State in question which has a major impact on the financial position of the general government” or to an “abrupt cyclical downturn.” The formal decision on whether or not a
9780230_547018_08_cha06.indd 189
3/19/2009 8:14:14 PM
190
●
●
Designing the European Model
deficit should be considered “excessive” is taken by the Ecofin Council, acting on a recommendation from the Commission. An annual fall of real GDP of more than 2 per cent should automatically be considered as “abrupt” and a fall of between 0.75 and 2 per cent could be considered to be so after a discretionary judgement by the Council. The Council could also take into account the cumulative loss of output relative to past trends when deciding whether a member state has an “excessive deficit.” If a member state does not take corrective action to eliminate an “excessive deficit,” as recommended by the Council, it will be required to pay an annual interest-free deposit of 0.2–0.5 per cent of GDP. If the “excessive deficit” persists, this deposit will be converted into a fine, which is distributed among the other member states. The Treaty also stipulates that gross government debt should not exceed 60 per cent of GDP. If it does, the debt ratio should be decreasing “at a satisfactory pace.” The wording must be interpreted to mean that government debt is not allowed to increase when it is above the 60 per cent ceiling. Formally, no escape clause is associated with this stipulation, but there are no monetary sanctions in the case of violations. According to the SGP, countries should aim for a “medium-term” budgetary position of “close to balance or in surplus.” To ensure compatibility with this objective, there is a process of multilateral budgetary surveillance. EMU member states have to submit standardised stability programmes and non-EMU member states similar convergence programmes specifying budget targets. These programmes form the basis for the regular monitoring of the fiscal performance of individual countries by the Council, acting on recommendations of the Commission. In the case of a “significant divergence” of budgetary outcomes from targets, the Council can issue an early warning to a member state. According to the stated principles of the Commission, it takes both cyclical developments and the risk of breaching the three-per cent deficit ceiling into account when judging whether or not there is a “significant divergence” (European Commission 2002d). A downward deviation from the budget target due to the response of automatic stabilisers to unanticipated cyclical developments is not considered such a divergence for a country meeting the medium-term objective of “close to balance or in surplus,” whereas it is for a country that does not fulfil this criterion and approaches the deficit ceiling.
One ambiguity in the SGP has concerned the “close to balance or in surplus” budget objective. A common interpretation has been that this is a target for the cyclically adjusted budget balance (Balassone and Franco 2001; Buti and Giudice 2002; European Commission 2002d). On this interpretation, it has been argued that the medium-term budget target should be set such as to provide a safety margin for both cyclical developments and unanticipated budgetary risks. High-debt countries should, in addition,
9780230_547018_08_cha06.indd 190
3/19/2009 8:14:14 PM
Fiscal Policy and Macroeconomic Stabilisation 191
take into account the objective of reducing their government debt levels when setting their budget targets (European Commission 2002d). But, as noted in the 2002 EEAG report, the SGP does not state explicitly that the medium-term objective refers to the cyclically adjusted balance, and the budget targets in the stability and convergence programmes have been stated in actual rather than in cyclically adjusted terms. Recently, however, the Commission proposed that the medium-term budget objective should refer explicitly to the cyclically adjusted balance (European Commission 2002b). An important feature of the fiscal rules is that they attach greater importance to the current government budget balance (the net flow of receipts and expenditures) than to the stock of government debt. The monetary sanctions are related to violations of the deficit ceiling only but not of the debt criterion. The medium-term objective of “close to balance or in surplus” also refers to the current budget balance. However, as the longrun debt is determined by the cumulated sum of deficits over time, the budget balance objective implicitly defines a goal for long-run debt. If the “close to balance or in surplus” target is interpreted as a balanced cyclically adjusted budget, the implicit long-run target for government net debt is zero.8 In addition to this, recent Council Resolutions have suggested that the future strains on public finances from ageing populations may require even more ambitious targets (European Commission 2002d). Actual budget balance 8 6
CAB = 2
4
CAB = 0
2
CAB = –2
0 –8
–6
–4
–2
0
2
4
6
Output gap
–2 –4 –6 –8 Figure 6.2
The actual budget balance as a function of the output GAP
Note: CAB = cyclically adjusted budget balance in per cent of GDP. Source: European Economic Advisory Group (EEAG).
9780230_547018_08_cha06.indd 191
3/19/2009 8:14:14 PM
192
Designing the European Model
Although discretionary fiscal policy in the event of major cyclical disturbances are not explicitly ruled out by the Stability and Growth Pact, the consensus seems to be that the medium-term target (the target for the cyclically adjusted balance) should provide room primarily for the automatic stabilisers to work (Buti and Giudice 2002; European Commission 2002d). 6.2.2 The case for revisions of the EU fiscal policy framework The EU fiscal rules have been the subject of intensive discussion both in academic circles and in more popular contexts. A common criticism is the arbitrariness of the chosen deficit and debt ceilings as well as of the long-run budget balance and (implicit) net debt targets (see, for example, Buiter et al. 1993; or Wyplosz 2002b). Other types of criticism have pointed to the inappropriateness of focusing on gross rather than net government debt, which also takes into account government claims on the private sector or government net worth, which also includes government real assets (Buiter et al. 1993). A related argument suggests that it should be possible to finance government capital outlays through borrowing a so-called “golden rule” of the type presently implemented in the UK; see, for example, Blanchard and Giavazzi 2002). With respect to macroeconomic stabilisation, which is our main focus here, two main objections have been raised. The first objection is that the fiscal rules may hamper stabilisation efforts in downswings (see, for example, Calmfors et al. 1997; Eichengreen and Wyplosz 1998; Canzoneri and Diba 2001; Swedish Government Commission on Stabilisation Policy in the EMU 2002; or Wyplosz 2002b). This has been a common argument in the recent debate on the budget deficits in France, Germany, Italy and Portugal (De Grauwe 2002; The Economist 2002; Financial Times 2002a,b,c). A second objection is that the fiscal rules provide insufficient incentives for fiscal restraint in booms by not rewarding such policies enough (Bean 1998; Buti and Giudice 2002). As discussed in the 2002 EEAG report, the risk of fines if the deficit limit is violated in a recession may not influence government behaviour much in a boom, since the next recession may then appear very far-off and may even occur under another government, which the incumbent goverrnment may have no interest in helping.9 Such insufficient fiscal restraint in booms will increase output volatility (both directly and also indirectly because the scope for countercyclical fiscal policy in future recessions becomes smaller when the safety margin to the deficit ceiling is reduced10) and weaken the government budget balance over the cycle. Indeed, the current situation of Portugal, Germany, France and Italy provide good examples of how such insufficient fiscal restraint in the 1999–2000 upswings have created deficit problems in the subsequent recession (see Tables 6.3 and 6.4). A very different type of critique has focused on the difficulties of applying the fiscal rules in concrete situations. The argument is that it will be difficult in practice to fine a member state that exceeds the deficit ceiling (Calmfors et al. 1997; Uhlig 2002). Such actions are likely to arouse serious political
9780230_547018_08_cha06.indd 192
3/19/2009 8:14:14 PM
Fiscal Policy and Macroeconomic Stabilisation 193 Table 6.3 General government actual fiscal balance (net lending) as a percentage of GDP in the EU countries
Austria Belgium Germany Denmark Spain Finland France Greece Ireland Italy Luxembourg Netherlands Portugal Sweden United Kingdom GDP-weighted average Unweighted average
1998
1999
2000
2001
2002
2003
–2.4 –0.7 –2.2 1.1 –2.7 1.3 –2.7 –2.5 2.4 –2.8 3.1 –0.8 –2.6 1.9 0.2 –1.6 –0.6
–2.3 –0.5 –1.5 3.1 –1.1 1.9 –1.6 –1.9 2.2 –1.8 3.6 0.7 –2.4 1.5 1.1 –0.7 0.1
–1.5 0.1 1.1 2.5 –0.6 7.0 –1.3 –1.8 4.4 –0.5 5.6 2.2 –2.9 3.7 4.0 1.0 1.5
0.2 0.4 –2.8 3.1 –0.1 4.9 –1.4 –1.2 1.5 –2.2 6.1 0.1 –4.1 4.8 0.7 –0.8 0.7
–1.8 –0.1 –3.8 2.0 0.0 3.6 –2.7 –1.3 –1.0 –2.4 0.5 –0.8 –3.4 1.4 1.1 –1.9 –0.6
–1.6 0.0 –3.1 2.0 –0.3 3.1 –2.9 –1.1 –1.2 –2.2 –1.8 –1.2 –2.9 1.2 –1.3 –1.8 –0.9
Source: European Commission (2002c).
Table 6.4 General government cyclically adjusted fiscal balance as a percentage of GDP in the EU countries 1998
1999
Belgium Denmark Germany Greece Spain France Ireland Italy Netherlands Austria Portugal Finland Sweden United Kingdom
–0.6 0.5 –1.9 –1.9 –2.6 –2.6 1.9 –3.0 –1.9 –2.4 –3.0 –0.4 2.3 –0.3
–0.9 2.5 –1.4 –1.6 –1.5 –2.0 0.8 –1.9 –1.2 –2.5 –3.0 0.3 0.6 0.8
GDP-weighted average Unweighted average
–1.7 –1.1
–1.0 –0.8
2000
2001
2002
2003
–1.1 1.3 –1.9 –1.8 –1.4 –2.1 2.5 –2.1 –0.6 –2.5 –4.0 3.8 2.1 1.2
–0.3 2.6 –2.8 –2.1 –0.7 –2.0 0.2 –2.4 –1.2 0.0 –4.3 3.8 4.2 0.7
0.2 2.1 –3.3 –1.7 –0.1 –2.7 –1.4 –1.8 –0.6 –1.6 –3.0 3.7 1.3 –0.6
0.2 2.1 –2.4 –1.8 –0.2 –2.8 –0.8 –1.6 0.0 –1.4 –1.9 3.3 1.3 –0.9
–1.0 –0.5
–1.2 –0.3
–1.6 –0.7
–1.4 –0.5
Source: European Commission (2002c).
9780230_547018_08_cha06.indd 193
3/19/2009 8:14:14 PM
194
Designing the European Model
conflicts among the member states. It may also be difficult to explain to the general public that the proper way of handling a deficit situation in a country is to incur extra expenditures in the form of fines to the European neighbours. The decision of the Council in early 2002 to avoid giving Germany and Portugal early warnings for the deviations relative to the agreed budget targets, despite a recommendation from the Commission to do so, provides a clear illustration of the difficulties of applying the rules in practice. So does the present situation, in which the budgetary problems in some member states have led to demands that the SGP should be flouted. The objections we have summarised all raise highly relevant issues. But it is also clear that any fiscal rule has to reflect a difficult trade-off between what would be theoretically optimal and simplicity. A rule must be simple to facilitate monitoring and enforcement (Kopits and Symansky 1998). Simplicity is also required if the rule is to be understood by the general public. Otherwise the rule will not command the legitimacy necessary for it to be respected by policy makers and be sustainable in the long run. Simplicity of the rules is probably of extra importance in the context of the EU, which has often been accused of being too technocratic. In our view, the provisions in the Maastricht Treaty and the SGP do, on the whole, represent a reasonable trade-off between conflicting demands and have played a very useful role for strengthening fiscal discipline. The rules have become common knowledge and are a useful common benchmark for fiscal policy in the member countries. At the same time, it should be acknowledged that the fiscal rules were instituted in a specific historic situation. There was an urgent need for reversing the trend of rapidly accumulating government debt and to quickly establish credibility for the new currency in its initial phase. Once the monetary union has been shown to work, it might be possible to refine the fiscal rules more than was possible in the 1990s. One counter argument is that such modifications might undermine the credibility of any common EU fiscal rules by creating the impression that they can always be revised in response to the existing situation. These are crucial considerations, but it must also be recognised that if the rules are perceived as being too inflexible, they will lose their legitimacy. The likely consequence of this is either that the rules will be constantly bent or that they will at some point be abandoned altogether. A reasonable conclusion is that one should look for possible modifications of the fiscal rules that enhance their effectiveness and legitimacy without changing their main character. As argued in the 2002 EEAG report, there is an obvious case in favour of formulating the medium-term budget balance objective explicitly in cyclically adjusted terms in order to allow the automatic stabilisers to work. We therefore fully endorse the Commission’s recent proposals on this (European Commission 2002b). In addition, as discussed in Section 1, there are strong arguments for allowing discretionary
9780230_547018_08_cha06.indd 194
3/19/2009 8:14:15 PM
Fiscal Policy and Macroeconomic Stabilisation 195
Box 6.1 The cyclically adjusted budget balance Tax receipts in general and some government expenditures, such as unemployment benefits, vary automatically with the output and employment levels. To assess the underlying (structural) budgetary situation, one must therefore adjust the actual budget balance for the cyclical conditions. Computations of the cyclically adjusted budget balance require estimates of both the output gap, that is the extent to which actual output deviates from the equilibrium (potential) level, and of the sensitivity of the budget balance to such deviations. Technically, the cyclically adjusted budget balance as a ratio of GDP, bc, is calculated as: bc = b – a g, where b is the actual budget balance as a ratio of GDP, g is the deviation of actual from equilibrium GDP as a ratio of equilibrium GDP, and a is the effect on the actual budget balance of a one percentage point increase in the output gap. The lines in Figure 6.2 show how the actual budget balance depends on the output gap and the cyclically adjusted budget balance. The cyclically adjusted balances are given by the intersections of the lines with the vertical axis. In booms, when the output gap is positive, the actual budget balance is more positive (less negative) than the cyclically adjusted balance. In recessions, when the output gap is negative, the actual budget balance is less positive (more negative) than the cyclically adjusted one. Provided that cyclical deviations are symmetrically distributed around equilibrium output, a given annual cyclically adjusted budget balance implies the same given average annual actual budget balance. The largest problem in computing the cyclically adjusted budget balance is how to estimate the output gap. There exists no universally accepted way of doing this. Instead, different methods give different results and the estimates are often subject to large ex post revisions. One way of estimating equilibrium GDP for a country is to use purely statistical techniques to smooth the actual GDP series. This is typically done by applying a so-called Hodrick-Prescott filter. With this measure, one tries to strike a balance between obtaining a smooth time series for equilibrium (potential) GDP on the one hand and getting a reasonable fit to the actual data on the other hand. The main problem with applying purely statistical techniques for estimating the output gap is that the estimates of equilibrium (potential) GDP are also influenced by actual GDP developments when there are persistent deviations from equilibrium, for example because of prolonged demand disturbances. This is an argument for instead using a production function approach, according to which equilibrium output is estimated on the basis of assessments of trends in total factor productivity and of the equilibrium levels of inputs of capital and labour. The most critical factor with this approach is probably how to assess the equilibrium employment rate. This is typically done on the basis of some kind of Phillips curve approach, where one tries to estimate the equilibrium unemployment rate consistent with a constant rate of inflation (NAIRU) or a constant rate of wage increase (NAWRU) (see, for example, Calmfors and Uddén Sonnegård, 2001). Earlier, the Commission and the Ecofin Council based their calculations of cyclically adjusted budget balances mainly on Hodrick-Prescott estimations. But the new Code of Conduct on the content and presentation of stability and convergence programmes, which was adopted by the Ecofin Council in 2001, stipulates that
9780230_547018_08_cha06.indd 195
3/19/2009 8:14:15 PM
196
Designing the European Model
there should be a shift to a production function method and sets common standards for how the estimations should be made (European Commission, 2002a). The estimates of how the actual budget balance reacts to variations in the output gap are usually based on assessments of the response of various tax receipts and government expenditures (see, for example, van den Noord, 2000). As discussed in Section 1.1, these response parameters differ among countries, but an average value for α in the EU is around 0.5. This value is used in Figure 6.2. It must be acknowledged, however, that estimated budget response parameters reflect average cyclical variations, so that the actual response in a specific situation characterised by atypical shocks may differ substantially from the average pattern. This is another serious problem when estimating cyclically adjusted budget balances.
fiscal policy action, that is variations in the cyclically adjusted budget balance, in some situations. These could involve large asymmetries in cyclical developments among the euro countries or large common macroeconomic disturbances where monetary policy needs to be supported by fiscal policy. It would not be appropriate, however, to formulate the deficit ceiling in terms of the cyclically adjusted budget balance rather than in terms of the actual balance. The obvious reason is that there is no unique way of adjusting the actual budget balance for cyclical factors. Different methods of calculating the cyclically adjusted balance give different results, as discussed in Box 6.1. In addition, the calculations are frequently revised ex post. One cannot base sanctions on a measure that is so open to different interpretations. As argued in Box 6.2, there is a strong case in all EU member states for trying to attain average budget outcomes of “close to balance or in surplus,” or even more ambitious budget goals, over the coming decade(s), because of the future strains on government finances due to demographic developments. In this situation, it would seem very unwise to loosen the budget objectives, for example through the adoption of a so-called “golden rule,” which would allow borrowing for government investment or other selected government expenditures believed to promote growth. The arguments against a golden rule are developed further in Box 6.3. We view the existing EU fiscal rules as a valuable institutional framework that should be exploited, because other alternatives will involve new and high set-up costs. This is an argument of history dependence. Given that the present framework is there, there is a strong case for continuing to build on it. This requires that the credibility of the fiscal rules is maintained. In particular, any changes in the fiscal rules must not be perceived as giving in to claims from member states that have current difficulties. Recent proposals of the European Commission aim at making the fiscal rules more flexible through a reinterpretation of the Stability and Growth Pact, which would not require any Treaty changes (European Commission 2002b). The proposals focus mainly on the medium-term budget target, but not on the deficit ceiling. In our view, this is insufficient and potentially harmful. More fundamental changes, involving a revision of the Maastricht Treaty, are desirable.
9780230_547018_08_cha06.indd 196
3/19/2009 8:14:15 PM
Fiscal Policy and Macroeconomic Stabilisation 197
Box 6.2 Long-run government debt A common criticism of the Stability and Growth Pact is that the medium-term budget target of “close to balance or in surplus” is arbitrary. It is often claimed to be too ambitious as it implies that net government debt will over time converge to around zero (see, for example, de Grauwe, 2002; or Walton, 2002). It is true that theoretical analysis does not give much guidance on what is an optimal level of long-run government debt, although it points to various important aspects (Kell, 2001; Wyplosz, 2002): • From the point of view of minimising long-run tax distortions that reduce social efficiency, a low debt level (or a positive net financial position) for the government is desirable. • On the other hand, to the extent that households are credit-constrained, social welfare is increased if governments can borrow on their behalf. • Intergenerational equity is affected by the level of debt, since this influences how consumption possibilities are distributed across generations. None of these considerations have played a major role in the choice of budget targets (and thus implicitly also of debt targets) in the SGP. Instead, as discussed in Section 2, more pragmatic considerations relating to longrun fiscal sustainability have dominated. The aim has been to lower debt to prudent levels in order to reduce the risks of inflation and high interest rates. In such a perspective, the future strains on government budgets that can be expected from ageing populations (due both to higher pension payments and higher health-related government expenditures) become of paramount importance. Table 6.5 presents estimates by the Economic and Financial Committee in the EU of future expenditure increases due to demographic developments. As can be seen, such projections indicate an average increase in age-related government expenditures of 6.2 percent of GDP in the EU countries between 2000 and 2040. The estimated increases are largest in Greece, the Netherlands, Spain, and Finland (7–13 percent of GDP), but much smaller in the UK, Italy, and Sweden (0–4 percent of GDP). Needless to say, the calculations are based on a number of uncertain assumptions, for example regarding labour force participation and unchanged policies (implying, for example, substantial reductions of pension replacement rates in Italy and the UK). The projections do not take into account that there is some reduction in other age-related expenditures, like for childcare. The calculations are also before tax, so that they do not measure the net effects on the government budget balance. Still, the calculations illustrate clearly that ageing populations will result in large budgetary pressures. Further reductions of government debt, and thus also of interest payments, is one way of accommodating the tendencies to deteriorating primary budget balances (the balances excluding interest payments). Table 6.6 is an attempt to illustrate this in a very simplified manner. Column 1 shows the 2001 primary balances. Column 2 shows the primary surpluses necessary to service interest payments if the debt-to-GDP ratio were to stay at the 2001 level. The other columns show how much lower primary surpluses need to be from 2020 and onwards, as compared to this benchmark, under various assumptions on total budget balances (including interest payments) in the period 2001–20 if the debt-to-GDP
9780230_547018_08_cha06.indd 197
3/19/2009 8:14:15 PM
198
Designing the European Model
ratio is to be held stationary after this period. The assumptions that fiscal policies can be characterised by various assumptions on the total budget balance up to 2020 and that the debt levels after that are stabilised at the 2020 level are arbitrary, but serve as a crude illustration of the extent to which the future demands on primary budget surpluses are affected by present policies. Our calculations illustrate that the future reductions in required primary surpluses that follow from present policies (and also from more ambitious ones) are small relative to the budgetary strains imposed by ageing populations. The average reduction in required primary surpluses from the debt reduction associated with zero budget balances in 2001–20 as compared with present debt levels is only 0.7 percent of GDP (1.2–0.5). Budget deficits of 1.5 percent of GDP instead of zero balance would yield a reduction of the average required primary surplus by only 0.3 percent of GDP (1.2–0.9). A total budget surplus in 2001–20 of 3 percent of GDP implies a reduction of the average required primary balance by 1.5 percentage points, allowing a small average primary deficit after 2020. The effects are of different magnitude for different countries, the largest effects occurring for the countries with the largest initial debt levels (Italy, Belgium and Greece). In the perspective of Table 6.6, it is difficult to claim that the present “close to balance or in surplus” target is too ambitious. The projected increase in age-related expenditures is a strong argument against relaxing the budgetary objectives, for example by adopting a “golden rule”, according to which government investment can be financed through borrowing. Rather, if anything, there appears to be a strong case for sharpening the medium-term fiscal objectives in most countries as a complement to pension reform.
Table 6.5 Projected increases in age-related public expenditures in the EU countries in percent of GDP, 2000–2040 Pensions
Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden United Kingdom Unweighted average
Health care expenditures
Long-term care expenditures
Total
3.8 3.7 3.6 4.7 3.8 4.8 11.2 3.6 1.9 2.2 6.2 4.0 6.6 2.4 –0.5
1.6 1.3 0.7 1.2 1.2 1.4 1.5 1.9 1.4 na 1.0 0.6 1.5 0.9 0.8
0.7 0.7 1.8 1.6 0.4 na na 0.1 0.6 na 1.8 1.6 na 0.6 0.0
6.2 5.7 6.1 7.5 5.4 (6.2) (12.7) 5.6 3.6 (2.2) 8.9 6.3 (8.1) 3.9 0.3
4.1
(1.2)
(0.9)
(6.2)
Note: Figures are given in parenthesis when there are missing data. Source: Table II.7, European Commission (2002a).
9780230_547018_08_cha06.indd 198
3/19/2009 8:14:15 PM
Fiscal Policy and Macroeconomic Stabilisation 199
The most important stipulations on fiscal policy are those that refer to excessive deficits. The possibility of sanctions has much stronger incentive effects than other stipulations and forms the backbone of the fiscal rules. So, we believe that more of the discussion should focus on this aspect. We see two desirable changes in the excessive deficit procedure: ●
●
To condition the scope for stabilisation policy in downswings on the level of debt, so that low-debt countries are allowed to run larger deficits than high-debt countries. To depoliticise the decision-making process that establishes whether or not individual countries have violated the rules.
6.2.3
An enhanced role for the debt level as a fiscal policy criterion
As discussed in the previous section, the current EU fiscal rules assign more importance to the current budget balance than to the debt level. This can be criticised on the grounds that the amount of debt is a more relevant variable if one is concerned about long-run fiscal sustainability and price stability (Beetsma 2001; Canzoneri and Diba 2001). On the other hand, the current budget situation is likely to be a better predictor of future budget outcomes than the historic debt level (Perotti et al. 1998). From a more practical standpoint, the focus on the budget balance rather than on debt in the Maastricht Treaty was probably motivated to a large extent by the great dispersion in debt levels among the prospective entrants to EMU (see Table 6.2), which made it difficult to use debt as a convergence criterion if one wanted to achieve the joint objectives of giving everyone a reasonable chance of qualifying and disciplining fiscal behaviour. Another motivation was the larger ambiguities associated with measuring the debt level than the current budget balance. We find it a reasonable argument that the present fiscal rules do not allow countries with low government debt to reap the full benefits of this situation (Pisani-Ferry 2002). Indeed, a main benefit of low government debt should be to enhance the room for manoeuvre in stabilisation policy by allowing larger deficits in recessions than would otherwise be possible (Swedish Government Commission on Stabilisation Policy in the EMU 2002). This can be seen as a corollary to the common argument that a track record of low inflation for a central bank should enhance the scope for interest rate cuts in a downswing. According to the present fiscal rules, there is an association between the debt level and the scope for stabilisation policy, but it is implicit rather than explicit. One association is the stipulation that countries with a debt ratio higher than 60 per cent are not allowed to increase it, which may be a more binding constraint for these countries than the deficit ceiling, whereas there is no such stipulation for countries with lower debt ratios (Balassone and Monacelli 2000). Table 6.7 shows the maximum deficits at various debt
9780230_547018_08_cha06.indd 199
3/19/2009 8:14:15 PM
9780230_547018_08_cha06.indd 200
3/19/2009 8:14:16 PM
Finland France United Kingdom Greece Ireland
Austria Belgium Germany Denmark Spain
Country
6.3 1.5 3.0 5.5 1.4
–0.1 6.2 0.7 6.7 2.4
Current cyclically adjusted primary surplus (2001)
0.9 (43.6) 1.1 (57.2) 0.8 (39.0) 1.9 (99.7) 0.7 (36.6)
1.2 (61.7) 2.1(107.5) 1.2 (59.8) 0.9 (44.5) 1.1 (57.2)
Required primary surplus at current debt level (2001)
1.2 (60.1) 1.3 (65.0) 1.1 (55.8) 1.7 (87.3) 0.8 (38.4)
1.3(66.0) 1.7 (87.4) 1.2 (63.4) 1.2 (62.6) 1.2 (59.2)
–3
0.8 (40.0) 0.9 (45.1) 0.7 (36.2) 1.3 (66.9) 0.5 (23.2)
0.9(46.3) 1.3 (67.5) 0.9 (44.0) 0.8 (42.0) 0.8 (40.4)
–1,5
0.4 (19.8) 0.5 (25.2) 0.3 (16.6) 0.9 (46.6) 0.2 ( 8.0)
0.5 (26.6) 0.9 (47.6) 0.5 (24.7) 0.4 (21.4) 0.4 (21.7)
0
0.0 (–0.3) 0.1 (5.3) –0.1 (–3.0) 0.5 (26.2) –0.1 (–7.2)
0.1 (6.9) 0.5 (27.7) 0.1 (5.3) 0.0 (0.8) 0.1 (3.0)
1,5
–0.4 (–20.4) –0.3 (–14.5) –0.4 (–22.6) 0.1 (5.9) 0.4 (–22.5)
–0.2 (– 12.8) 0.2 (7.7) –0.3 (–14.0) –0.4 (–19.8) –0.3 (–15.7)
3
Required primary surplus at various annual total budget balances 2001–2020
Table 6.6 Required primary surpluses in the EU countries to stabilise the debt ratio after 2020 under various assumptions (debt ratios in parenthesis)
9780230_547018_08_cha06.indd 201
3/19/2009 8:14:16 PM
1.2 (59.1)
3.5
1.2 (63.7)
1.1 (56.9) 1.3 (68.0)
1.8 (93.0) 0.6 (32.5) 1.2 (60.0)
0.9 (44.5)
0.8 (38.5) 0.9 (47.4)
1.4 (72.5) 0.3 (16.9) 0.8 (40.8)
0.5 (25.3)
0.4 (20.1) 0.5 (26.9)
1.0 (52.0) 0.0 ( 1.3) 0.4 (21.6)
0.1 (6.2)
0.0 (1.7) 0.1 (6.3)
0.6 ( 31.5) –0.3 (–14.3) 0.0 (2.4)
–0.3 (–13.0)
–0.3 (–16.6) –0.3 (–14.3)
0.2 (11.0) –0.6 (–29.9) –0.3 (–16.9)
Source: Columns 1 and 2: European Commission. The other columns: computations by EEAG group and José Mauricio Prado.
Note: Column 1 gives the current (2001) primary budget balance (the budget balance excluding interest payments). Column 2 gives the primary balance necessary to stabilise government debt at its current (2001) value (with the debt level in parenthesis). The subsequent columns show the primary balances necessary to stabilise government debt from 2020 and onwards under various assumptions on the total annual budget balance (including interest payments) in 2001–2020. The stationary debt levels are given in parenthesis. Annual real GDP growth for each country in 2001–2020 is assumed to be the same as the average for 1985–2001. Annual inflation in 2001–2020 is assumed to be 2 percent. From 2020 and onwards, nominal GDP growth in all countries have been set equal to the average real GDP growth for all EU countries in 1985–2001 plus 2 percent. The nominal interest rate is assumed to be 2 percentage points higher than the nominal growth rate.
1.1 (55.6) 1.1 (56.0)
0.1 7.6
Portugal Sweden Unweighted average
2.1 (109.4) 0.1 (5.5) 1.0 (53.2)
4.9 3.8 3.1
Italy Luxembourg Netherlands
202
Designing the European Model
Box 6.3 The golden rule The “golden rule” in public finance is the notion that borrowing should be allowed for public investment. Such a golden rule for both the federal government and the states is formally enshrined in the German constitution. More recently, the UK has adopted such a rule, according to which deficit financing of government net investment is allowed, provided that the overall government debt is kept at prudent levels (at present defined as a ratio of net government debt to GDP below 40 percent) (see Buiter, 2001; or Kell, 2001). In the discussion of the Stability and Growth Pact, it has been argued that the present medium-term objective of “close to balance or in surplus” should be replaced by a golden rule, which would also require a redefinition of the deficit ceiling in the Treaty (see, for example, Blanchard and Giavazzi, 2002). The crucial issue when judging the future budgetary consequences of a public investment project is whether it generates a cash flow accruing to the general government, whose (appropriately measured) present value is at least equal to its financial costs for the government. If this is the case, deficit financing of public investment does not cause a deterioration of future budget balances and does not require any future rises in tax rates or cuts in spending. Otherwise, the structural deficit will be worsened to the extent that future additions to tax revenue fall short of interest costs of the additional debt. In principle, the statements above could be turned into a test to discriminate among different public investment projects. According to this test, projects meeting a minimum cash-flow requirement could be excluded from the deficit figure subject to the rules of the SGP. Government borrowing would be disallowed only for the costs of those projects failing the test. These considerations show that a “golden rule” should never be applied mechanically. Rather, it should always be made conditional on an assessment of the future financial flows from public investment. This is because nothing requires the public sector to undertake projects only if they satisfy the test specified above. The provision of public goods requiring public investment may well be motivated by their utility value, independent of whether or not it can generate a positive cash flow. Many projects may be highly desirable, yet require tax financing. But even such a stricter and sounder golden rule would run into a number of theoretical and practical objections that strongly discourage its application. First and foremost, it is very difficult to assess future revenues accruing to the general government. Such an assessment is necessarily based on arbitrary assumptions, and these assumptions are unavoidably open to manipulation. Typically, one should take both direct and indirect public revenues into account. An example of the latter is any increase in tax revenue due to incomes that would not be generated in the absence of public investment. However, proponents of a particular project, and groups benefiting directly from it, will have strong opportunistic motives to inflate the estimates of the indirect effects. Also, as amply documented, inadequate budgeting and implementation of public investment projects typically result in cost revisions, systematically reducing net cash flows well below the initial estimates. Moreover, the classification of public expenditures between “current expenditures” and “investment” is quite ambiguous. For example, should spending on public education be viewed as public investment in human capital? And why should tax cuts that stimulate private investment be treated differently from direct government investment? Adopting a golden rule will clearly create a
9780230_547018_08_cha06.indd 202
3/19/2009 8:14:16 PM
Fiscal Policy and Macroeconomic Stabilisation 203 strong incentive to reclassify many items in the budget, with no other purpose than to circumvent the Stability and Growth Pact. Second, suppose that the public sector can accurately predict cash flows, and is able to commit itself to extremely disciplined control procedures. Even so, it is quite difficult to identify the correct interest costs. For instance, a large programme of public investment could change the equilibrium interest rate in the economy: using interest rates prevailing before the implementation of the program would therefore not be appropriate. The recent revival of the debate on the golden rule may actually divert attention from a deeper issue. Fiscal retrenchment implemented by several countries in the euro area throughout the 1990s resulted in large cuts in public capital expenditure. As is well understood, the interest groups fighting cuts in public investment are not as strong and vocal as the interest groups opposing cuts in current transfers programs. Public capital investment tends to have diffuse effects and – more importantly – tends to benefit both current and future generations. Intergenerational redistribution via public capital is a theme that is often forgotten in the political debate. An inefficiently low level of public infrastructure, and a low quality, can harm future generations at least as much as higher future taxes financing present transfers. In many areas, a reduced presence of the public sector has crowded in private investment, substituting private for public capital. But especially for infrastructure, there is a widespread feeling that development and maintenance have been falling below efficient levels. Is this an argument to relax budget goals? What is called into question is not really budget discipline – but the political priorities in the national budget process as well as budget choices at the European level. If the investment in infrastructure is too low, which creates large welfare costs for current and future generations, governments can change spending plans or find proper ways to finance additional spending. In this respect, it should be kept in mind that deficit financing of investment generates future interest payments – payments that are avoided if investment is tax-financed. So, requiring tax financing does not impose any additional burden on taxpayers in the long run. It does, however, give rise to redistribution effects in favour of future generations in the short and medium run. In principle, the golden rule could be defended as a way to make future generations sustain the costs of infrastructure projects that will also benefit them. By the same token, it is well known that efficiency (tax-smoothing) arguments suggest the desirability of deficit financing at an early stage of development: countries starting with a low capital stock have a large need to build infrastructure, and should therefore be given the financial flexibility to do so. While all this is true, it should not be forgotten that what motivated the Stability and Growth Pact is exactly the argument according to which sound economic principles are seldom followed in the actual budget process. Recently, a common misinterpretation of public finance principles has been that there is a case for excluding military spending from the budget objectives according to the SGP. It is true that the tax-smoothing principle implies that any temporary upsurge in military spending should be financed by borrowing, and not by increasing taxes, because this avoids welfare-decreasing variations in private consumption. But in the case of Europe, those who believe in a larger military role for the EU advocate a permanent (rather than temporary) step-up of defence spending. While the choice of increasing military spending is a political one – and there is by no means an agreement on whether and how much the EU should change its course on this matter – there is no economic argument for deficit financing.
9780230_547018_08_cha06.indd 203
3/19/2009 8:14:16 PM
204
Designing the European Model
Table 6.7 Deficit levels consistent with a stable debt-to-GDP ratio Debt-to-GDP ratio Nominal growth rate 3% 4% 5%
60 1.7 2.3 2.9
70 2.0 2.7 3.3
80 2.3 3.1 3.8
90 2.6 3.5 4.3
100 2.9 3.8 4.8
Note: The change in the debt-to-GDP ratio, d, is given by dt bt (1 )dt 1 , where b is the budget balance in percent of GDP and is the rate of growth of nominal GDP. The deficits in the table are obtained by setting Δdt = 0 and solving for bt. Source: Calculations by EEAG group.
levels and growth rates consistent with the condition that the debt ratio must not increase. The debt change stipulation has the peculiar implication that it constrains the size of deficits more the closer the debt ratio is to the 60 per cent reference value. For example, assuming a 3 per cent nominal growth rate, the maximum deficit is 2.9 per cent of GDP at a 100 per cent debt ratio, whereas it is only 1.7 per cent at a 60 per cent debt ratio. This follows from the fact that nominal GDP growth automatically tends to reduce the debt-to-GDP ratio less the lower this ratio, requiring lower deficits if the ratio is not to increase. However, it is not clear how much emphasis is in practice likely to be put on the debt change stipulation. Against the letter of the Treaty, it was ignored as a convergence criterion for Germany, which violated it at the start of EMU. Nor has the debt change criterion received much attention in the recent discussion, although Italy is likely to have violated it in 2002 and Germany is likely to do so in 2003 (see Table 6.2). There is also an indirect association between the scope for stabilisation policy and the debt level, as low debt implies lower interest payments. For example, assuming a 4 per cent average interest rate on government debt, a reduction of the debt ratio from 50 per cent to 25 per cent of GDP would reduce interest payments as a ratio of GDP by 1 percentage point (from 2 to 1 per cent). Ceteris paribus this would improve the total budget balance. This assumes, however, that the lower interest payments have not been offset by a deterioration of the primary budget balance (the budget balance excluding interest payments) through tax cuts or expenditure increases. One could also argue that there should be a positive association between low debt and strong government budget positions because low debt can only have been achieved through small deficits or through surpluses in the past, and budget situations tend to exhibit a high degree of persistence (Perotti et al. 1998). According to this argument, a low debt level would be associated with a high probability of a strong current government budget position, which gives a large cyclical safety margin in a downswing. But on the other hand, we know from empirical studies of the determinants of the government budget
9780230_547018_08_cha06.indd 204
3/19/2009 8:14:16 PM
Fiscal Policy and Macroeconomic Stabilisation 205
balance that high debt is conducive to low deficits (high surpluses), as it creates pressure for adjustment (see, for example, von Hagen et al. 2002). Finally, the Council is likely in practice to take into account the debt position of a country when judging whether the escape clause allowing violations of the deficit ceiling in certain situations can be invoked, even if this is not formally stated (see Section 2.1). The recent Commission proposals on reinterpreting the SGP involve a greater emphasis on debt (European Commission 2002b). First, it is argued that the debt change criterion should be taken seriously and that breaches of it should trigger the excessive deficit procedure. Second, the Commission has proposed that member states with debt lower than 60 per cent of GDP should be given the possibility of small temporary deviations from the “close to balance or in surplus” target for the cyclically adjusted budget balance if these deviations derive from a “large structural reform” aiming at promoting growth. A third proposal is that “small deviations of a longerterm nature” from the “close to balance or in surplus” objective could also be envisaged for member states with debt ratios “well below the 60 per cent reference value.” We see two major problems with the Commission’s proposals. One is the increased complexity of the rules and the amount of discretionary judgements introduced. Another problem is that loosening the medium-term fiscal objective without changing the deficit ceiling reduces the safety margins and thus increases the risk that the ceiling is breached. In our view, a better plan for reforming the fiscal framework should instead focus on the deficit ceiling directly. There should be a clear and transparent rule. One possibility would be to condition the deficit ceiling explicitly on the debt level, allowing low-debt countries to run larger deficits in downswings than high-debt countries. More precisely, low-debt countries could be allowed to run larger budget deficits than three per cent of GDP. Such a Treaty revision would have several advantages. 1. The scope is widened for low-debt countries to pursue expansionary fiscal policy in a downswing. 2. The incentives for fiscal restraint in general are enhanced if the returns to such policies in the form of a greater scope for stabilisation policy in downswings become higher and more visible. 3. The stronger incentives for fiscal discipline imply smaller risks of procyclical policies in booms. 4. To the extent that the advantages of fiscal discipline become larger, the legitimacy of the fiscal rules and thus their credibility would be enhanced. Technically, a link between the deficit ceiling and the debt level could be established in several ways. One could simply stipulate different deficit
9780230_547018_08_cha06.indd 205
3/19/2009 8:14:16 PM
206
Designing the European Model
ceilings for different debt intervals. One proposal, which raises the deficit ceiling for low-debt countries but leaves it unchanged for high-debt countries, is given in the first column of Table 6.8 (see also Calmfors and Corsetti 2002, 2003). An alternative would be a scheme like the one in the second column, according to which the rises in the deficit ceiling for low-debt countries are matched by reductions for high-debt countries. The latter proposal may appear less politically realistic, but it could perhaps be made more attractive if it is linked to a formal abolition of the debt change criterion for the countries exceeding the 60 per cent debt-to-GDP reference value. A lower deficit ceiling for these countries would serve the same purpose as the debt change criterion, but do away with the anomaly that present rules formally require lower maximum deficits for high-debt countries the closer their debt ratio is to the 60 per cent value. A major advantage of such discontinuous “ladders” of deficit ceilings as shown in Table 6.8 is that they provide a strong incentive for fiscal discipline in normal times as well as in booms by allowing countries to move to categories with a higher “status.” Even if it is future governments that would get the advantage of a greater scope for stabilisation policy in recessions, it becomes much more visible to the general public that the incumbent government has made an investment that represents a future gain.11 An alternative set-up would be to retain the present three-per cent deficit ceiling, but allow countries to use extra-budgetary stabilisation funds in downswings that are not formally included under the deficit ceiling. Such so-called “rainy-day funds” exist in many US states and Canadian provinces as a cushion against unforeseen contingencies (Kopits and Symansky 1998; Knight and Levinson 1999; McGranahan 1999; Hemming and Kell 2000) and have been discussed in the European context by Buti and Giudice (2002) and Buti et al. (2002). A system with such funds could be constructed so as to mimic debt-deficit links of the type indicated by Rule 1 in Table 6.8. Countries with debts below given thresholds would be allowed to establish such funds and to draw maximum pre-specified amounts from them in recessions, in addition to running deficits in the normal budget up to three per cent of GDP. One possibility is to let countries with government debt ratios below certain thresholds establish the stabilisation funds immediately through borrowing, which would increase gross (but not net) government debt, or by transferring claims on the private sector to them. Another more demanding option is to let low-debt countries build up the funds over time by channeling government surpluses into them in good times. A system of extra-budgetary “rainy-day-funds” could formally maintain the three-per cent ceiling as the point of reference. It would “lock in” the assets put in the funds by earmarking them only for stabilisation of output and employment in recessions. Arguably, however, a system of extra-budgetary funds is less transparent than a system that explicitly conditions the deficit ceiling on the debt level.
9780230_547018_08_cha06.indd 206
3/19/2009 8:14:16 PM
Fiscal Policy and Macroeconomic Stabilisation 207 Table 6.8
Possible ways of conditioning the deficit ceiling on the debt ratio
Debt ratio (percent of GDP)
Deficit ceiling (percent of GDP) Rule 1
Rule 2
> 105 95–105 85–95 75–85 65–75 55–65
3.0 3.0 3.0 3.0 3.0 3.0
0.5 1.0 1.5 2.0 2.5 3.0
45–55
3.5
3.5
35–45
4.0
4.0
25–35 < 25
4.5 5.0
4.5 5.0
Countries in the debt range Italy Belgium, Greece
Portugal, France, Germany, Austria, Bulgaria Netherlands, Sweden, Spain, Hungary Ireland, UK, Finland, Denmark, Slovak Republic, Poland Czech Republic, Slovenia Luxembourg, Estonia, Latvia, Lithuania, Romania
Note: Accession countries in italics. These countries have been classified above according to their debt ratios in 2002. The incumbent EU member states have been classified according to predicted debt ratios in 2003. Source: See Tables 6.2 and 6.9.
How would the accession countries be affected by rules that explicitly condition the deficit ceiling on the debt level? The accession countries have on average much lower government debt than the present EU member states (see Table 6.9). Therefore, a rule that relates the maximum deficit to the debt level gives them more scope for stabilisation policy in downswings than the incumbent member states. This could be motivated to the extent that the accession countries are likely to be exposed to larger cyclical swings, as they are in a phase of transition to developed market economies. It is true that this could also involve risks that serious budgetary imbalances develop, as there are some tendencies to (see Table 6.9). But on the other hand the accession countries will also have a stronger tendency to reduce their debt ratios than the present EU member states because they will have higher nominal GDP growth. This is the consequence of both higher convergence-driven real GDP growth, as they catch up with Western Europe in terms of income per capita, and higher inflation due to the Balassa-Samuelson effect (according to which higher productivity growth in the tradables sector in catching-up countries than in already rich countries leads to higher wage growth and thus to higher price rises in the non-tradable sector, as discussed in EEAG, 2002).12 We believe there is a case for revision of the fiscal rules in the EU along the lines we have proposed. We have deliberately chosen the debt intervals in Table 6.8 such that our reform proposals would not accommodate the current budget problems of France, Germany, Portugal and Italy. The
9780230_547018_08_cha06.indd 207
3/19/2009 8:14:16 PM
208 Designing the European Model Table 6.9
The fiscal position of accession countries in 2002 Gross government debt as a percentageof GDP
General government actual fiscal balance in percent of GDP
Bulgaria Czech Republic Estonia Hungary Latvia Lithuania Poland Romania Slovak Republic Slovenia
58.1 25.6 4.4 52.9 16.8 23.6 43.3 24.6 39.3 27.9
–0.8 –6.4 –0.2 –5.7 –1.8 –1.9 –4.1 –2.7 –4.6 –1.8
Unweighted average
31.7
– 3.0
Source: Tables 9 and 10 in European Commission (2002d).
former three countries will all have debt-to-GDP ratios in 2003 of close to 60 per cent and Italy has a ratio of close to 110 per cent. It is true that the recent deterioration of the budget balances in these countries is associated with the workings of the automatic stabilisers, which dampen the present cyclical downswing. The root cause of the current budgetary problems is insufficient fiscal retrenchment in the preceding boom. Relaxing the rules such as to accommodate the current situation would, however, completely undermine the credibility of fiscal constraints at the EU level. It is not a good strategy to try to solve short-term problems by adjusting the long-term rules. The budget developments in some member states might imply that they are in the end fined if there is a drawn-out recession. This may not be all bad. Once a member state like Germany has been exposed to such fines, there would be little doubt that the sanction procedures are credible. The current German situation may seem particularly awkward, as there are indications of a much larger negative output gap than in other euro countries (see Chapter 1 of the 2003 EEAG report). One possibility that should be considered is whether demand could be boosted through a “tax shift,” that is through a reduction of employers’ payroll taxes that is financed by increases in taxes that fall on employees, such as employee contributions to the social security system, income taxes or VAT. Such a tax reform represents a fiscal policy to stimulate the economy without incurring a larger budget deficit. The policy is often labelled an “internal devaluation,” as it reduces the real labour cost and depreciates the real exchange rate in a similar way as a reduction in the external value of the currency (Calmfors et al. 1997; Calmfors 1998). Such internal devaluations were made in Denmark in the late 1980s and in Sweden in the early 1990s.
9780230_547018_08_cha06.indd 208
3/19/2009 8:14:16 PM
Fiscal Policy and Macroeconomic Stabilisation 209
The underlying assumption behind the proposal of a German internal devaluation is that appropriate reductions in the real labour cost and relative wage levels vis-à-vis other euro countries are at present prevented – or take a very long time to accomplish – because of downward money wage rigidity: it is very difficult to reduce money wage increases below those of the other euro countries at a low rate of inflation. The expenditure switch in favour of German products that a real exchange rate depreciation would achieve is motivated by a weaker demand situation in Germany than in the rest of the euro area and would most likely have occurred through an ordinary exchange rate depreciation in the absence of a common currency.
Box 6.4 Different measures of the government’s financial situation The gross government debt concept used in the Maastricht Treaty is only one of several possible measures of the government’s financial position. This box reviews various measures. • Gross government debt nets out all claims and liabilities within the government sector, but claims on the private sector are not included. • Another debt concept is net government debt, which deducts government claims on the private sector from the gross debt. • Conventional measures of government gross and net debt do not take account of pension obligations, but only refer to explicit debt. Pension obligations can be thought of as “implicit debt”. Yet another measure is thus total explicit and implicit debt. • If one adds in the real capital assets of the government, one obtains the net worth of the government. Theoretically, net worth is the most relevant measure of the government’s solvency (Buiter et al. 1993; Buiter 2001; Balassone and Franco 2000) Real capital assets must then be assessed according to market values and not according to historic costs, as it is the ability to generate future revenues that is of interest. However, in practice there are huge problems of evaluation. Theoretically, net debt is also a more relevant concept for government solvency than gross debt, as a government can in principle draw on claims on the private sector. But here, too, there may be problems of evaluation (although smaller than for real capital assets). For example, many government loans to the private sectors may be “soft ones” with a large ingredient of subsidisation (this is likely to be a severe problem in the transition economies in Eastern Europe) (Buiter et al. 1993; Föttinger 2001). It is not self-evident how implicit pension debt should be regarded, since pension obligations are less firm than ordinary debt obligations. On the one hand, pension obligations are a policy variable that is subject to possible change through reforms of pension systems. On the other hand, there is a political commitment to honour these obligations. Similarly, one could argue that there are very strong commitments also to some other government expenditures, such as health care
9780230_547018_08_cha06.indd 209
3/19/2009 8:14:16 PM
210
Designing the European Model
and long-term care expenditures. By also regarding the path of such expenditures as exogenous (for example, by assuming age-related spending increases, as discussed in Box 6.2), and by assuming unchanged policies with respect to taxes and other expenditure categories, one can forecast future budget balances. Computing the discounted present value of such tax and expenditure streams yields a “broader” measure of implicit debt, albeit one that can more easily be changed through policy action. Adding “implicit debt” calculated in this way to explicit debt is one way of assessing the need for fiscal adjustment in order to ensure fiscal sustainability. Alternatively, fiscal sustainability indicators can be expressed as the immediate and permanent change in the budget balance necessary to meet various definitions of long-run fiscal sustainability (Blanchard et al. 1990; Balassone and Franco 2000). It is a general rule that the more theoretically relevant the measure of the government’s financial situation is, the larger are the practical evaluation and measurement problems. So, there is a trade-off between theoretical relevance and practical applicability when choosing a measure as a basis for policy. Table 6.10 presents four different measures of the government financial situation in 2001 for the EU countries. The first column shows government gross debt according to the European Commission. The second column shows the same measure according to the OECD. Column 3 presents explicit government net debt according to the OECD. Column 4 shows “total” (explicit + implicit) net debt according to Frederiksen (2002), where the implicit debt has been calculated as the discounted value of future net expenditures “associated with current expenditure and tax policies”, thus taking into account inter alia pension obligations and expected increases in health-related expenditures. The numbers in parentheses rank the countries according to the various measures. Table 6.11, finally, shows the correlations between the different measures. Several observations may be made from the tables. The Commission and OECD measures of gross debt are highly correlated, but not identical. The average government net debt ratio is 15–20 percentage points lower than the gross ratios. There is a high correlation between the net and gross debt measures. The largest discrepancies refer to the Nordic countries (Denmark, Finland and Sweden), all of which have substantially lower net than gross debt. “Total net government debt”, including also implicit pension debt and the discounted present value of future age-related expenditure increases, is five to six times larger than explicit government net debt. There is a positive but weak correlation between explicit gross debt and “total net debt”, whereas there is a weak negative correlation between explicit net debt and “total net debt”. The weakness of the correlations is explained by a few countries, whose relative positions change dramatically when “implicit debt” is included. Italy, which has very low “implicit debt”, then moves from being the most indebted country to being a low-debt country, whereas Finland and Ireland move from relatively low to relatively high levels of government “debt”. As can be seen from Table 6.11, the correlations increase substantially if these countries are removed from the comparison.
9780230_547018_08_cha06.indd 210
3/19/2009 8:14:16 PM
Fiscal Policy and Macroeconomic Stabilisation 211 Table 6.10 Different measures of government debt as percentages of GDPfor the EU countries, 2001 Gross government debt EU Commission) Luxembourg Ireland United Kingdom Finland Denmark Netherlands Portugal Sweden France Spain Germany Austria Greece Belgium Italy GDP-weighted average Unweighted average Standard deviation Coefficient of variation
5.5 (1) 36.6 (2) 39.0(3) 43.6(4) 44.5(5) 53.2(6) 55.6(7) 56.0(8) 57.2(9) 57.2(9) 59.8(11) 61.7(12) 99.7(13) 107.5(14) 109.4(15) 63.0
Gross government debt (OECD)
5.5(1) 36.5(2) 52.5(5) 43.6(3) 46.4(4) 53.2(7) 55.6(8) 52.9(6) 64.8(11) 69.1(12) 60.3(9) 61.7(10) 99.7(13) 108.2(14) 108.7(15) 69.1
Explicit net Government debt
32.0(5) 30.9(4) – 47.9(1) 22.9(3) 42.1(7) 53.0(11) 1.0(2) 42.1(7) 39.8(6) 43.5(9) 47.0(10) 100.0(14) 98.9(13) 96.5(12) 48.8
“Total net government debt”
302(10) 102(1) 335(13) 165(2) 287(8) 222(4) 297(9) 280(7) 415(14) 222(4) 253(6) 329(12) 311(11) 174(3) 233
59.1
61.2
43.0
264
26.9
26.5
37.9
78
0.5
0.4
0.9
0.3
Note: Gross government debt is total government debt, where only claims and liabilities within the government sector have been netted out. Explicit net government debt deducts government claims on the private sector from gross liabilities. “Total net government debt” includes also “implicit” government debt associated with future net revenue consequences of current expenditure and tax policies (thus reflecting also inter alia future unfunded pension obligations and projected incresases in other age-related government expenditures). The numbers in parenthesis give country rankings. Source: Gross government debt: the EU Commission and the OECD. Explicit net government debt: the OECD. “Total net government debt”: Frederiksen (2002).
6.2.4 Depoliticising EU surveillance and excessive deficit procedures The decision of the European Council in early 2002 not to follow the Commission’s recommendation to give Portugal and Germany early warnings for their failure to meet their budget targets, after heavy lobbying on the part of these countries, have seriously undermined the credibility of the fiscal rules. Not least was the suspicion reinforced that it may be much more difficult to “shame” a large country than a small one. It goes without saying that the handling of the current deficits in Portugal, Germany, France and Italy will be an “acid test” of the credibility of the EU fiscal framework.
9780230_547018_08_cha06.indd 211
3/19/2009 8:14:16 PM
212
Designing the European Model
Table 6.11
Correlations between the government debt measures in Table 6.10
Gross government debt (EU Commission) Gross government debt (OECD) Explicit net government debt
Gross government debt (OECD)
Explicit net Government debt
“Total net government debt”
0.98
0.80
0.10 (0.52)
0.83
0.09 (0.52) – 0.09(0.32)
Note: The figures in parenthesis give the correlations when Finland, Ireland and Italy have been excluded. Source: Calculations by EEAG group and José Mauricio Prado.
A fundamental problem is the political character of EU decisions regarding the budget situation in individual countries. The finance ministers in the Ecofin Council have a strong incentive to act strategically as the budgetary surveillance process and the excessive deficit procedure have the character of a repeated game: by adopting a forgiving attitude towards colleagues with deficit problems, the risk of being branded oneself in similar situations in the future is reduced. If it is difficult to agree even on a relatively harmless early warning, it may be almost impossible to agree on harsher sanctions that involve both a financial cost and a loss of national prestige. Fines also risk triggering serious political conflicts among the member states, which goes directly against the aim that the EU should foster integration and common understanding (Uhlig 2002). The root of the problem is that EU monitoring of the fiscal situation in individual member states is in the end done by the politicians responsible for these very situations. This is an unsatisfactory state of affairs. Our political systems usually draw a sharp dividing line between making the laws (which is done by elected politicians) and applying them (which is done by an independent judiciary). The budgetary surveillance process and the excessive deficit procedure clearly violate this principle. The Commission has proposed that it alone should be given the power to issue early warnings in the budgetary surveillance process. The drawback of this proposal is that the Commission has weaker political legitimacy than the Council. Still, we endorse the proposal as a better alternative than the present arrangement. But it is less clear that a similar solution would be preferable for the excessive deficit procedure. Buti et al. (2002) have proposed that the Commission should be responsible for determining the existence of an excessive deficit and for giving a first warning to the member state in question. The next step would then be that the Council gives a second warning and requires corrective action to be taken. In the final step, the Council would take the decisions on sanctions, acting on a proposal from
9780230_547018_08_cha06.indd 212
3/19/2009 8:14:16 PM
Fiscal Policy and Macroeconomic Stabilisation 213
the Commission, which would have to be followed unless there is a unanimous decision not to do so. An alternative way of depoliticising the excessive deficit procedure would be to transfer the decisions on sanctions from the political level of the Council to the judicial level of the European Court of Justice. The natural procedure would then be for the Commission to take violations of the excessive deficit criterion to the Court, which would then make the ultimate decisions on deposits and fines, possibly after hearing a standing panel of independent economists. Specific procedures would then have to be followed to ensure a speedy process, which is necessary in order to create the proper incentives for avoiding excessive deficit situations. Our two proposals of making the deficit rules more flexible and revising the decision process on excessive deficits may appear unrelated. In fact, they are not. The more credibly the fiscal rules are enforced, the greater the scope for changes that introduce more flexibility.
6.3
Is there a case for delegation of national fiscal policy?
Section 2 focused on possible reforms of the EU fiscal rules. There is, however, a risk that too much of the fiscal policy discussion focuses on the EU. The foundations of good fiscal policy must be laid at the national level. If incentives for well-balanced policies are too weak there, political conflicts about the fiscal stance of individual member states will arise continuously at the EU level, which will undermine the legitimacy of the common rules. The risk is all the greater, as there is some evidence that peer pressure at the EU level exerts less influence on large than on small member states (von Hagen et al. 2002). The fiscal policy framework at the European level relies mainly on the common rules with numerical targets in the Maastricht Treaty and the SGP, whereas it has been left to the member states to decide on the national institutional frameworks to ensure compliance. Another strategy would have been to focus on common standards for the design of national fiscal institutions and decision procedures. The main reason why the latter method was not adopted is probably that it was considered to imply much greater interference with national sovereignty and to be associated with greater monitoring problems (Beetsma 2001; Buti and Giudice 2002). But the recent deficit experiences of some EU states have vividly illustrated the difficulties inherent in a system based mainly on the enforcement of common numerical targets. This raises the issue of whether one should not rely to a larger extent on common standards for national fiscal institutions. A parallel would be the common regulation of the legal status of the national central banks, which applies also to non-EMU members. The argument is that it might pay to take the one-off cost of reforming national institutions according to commonly agreed principles, because this would reduce the risks of
9780230_547018_08_cha06.indd 213
3/19/2009 8:14:17 PM
214
Designing the European Model
inappropriate fiscal policies in individual member countries and hence the risks of political conflicts at the EU level. Such common principles of national fiscal policy frameworks could take the form of each member state adopting a law on fiscal policy that must meet certain minimum standards (see also Swedish Government Commission on Stabilisation Policy in the EMU 2002; and Calmfors 2002). The law should specify long-run goals for the path of government debt and/or the medium-term (cyclically adjusted) fiscal balance. These goals must be consistent with (but could be more ambitious than) the requirements of the Maastricht Treaty and the SGP. Such a law should also define clearly the short-run stabilisation objectives of fiscal policy. It should be clarified to what extent one intends to rely on the automatic stabilisers and in what circumstances discretionary policy action will be undertaken. Similar to the Australian Charter of Budget Honesty, the government could be obliged to indicate which tax and expenditure changes are temporary (because they are undertaken for stabilisation purposes) and “the process for their reversal” (Business Council of Australia 1999). To shorten decision lags and reduce the risk that income distribution aspects dominate stabilisation considerations in concrete situations, a law on fiscal policy could also select in advance a small number of fiscal policy instruments to choose from if the need for discretionary measures were to arise (Swedish Government Commission on Stabilisation Policy in the EMU 2002; Calmfors 2002). In addition, a law on fiscal policy ought to regulate the budget policies of sub-national local governments such as to ensure compatibility with the overall national fiscal objectives. One should also try to ensure that the stabilisation programmes submitted to the Ecofin Council by the member states do not live a life of their own in the national decision process, but are approved by the parliament as part of the normal budget process, as has been suggested by von Hagen et al. (2002). Such an integration of the fiscal processes at the European and national levels would seem necessary to ensure consistency between national budget objectives and commitments at the European level. 6.3.1 Lessons from the decision-making process for monetary policy Section 1 discussed why monetary policy is today generally regarded as a much more effective stabilisation policy tool than fiscal policy. This is not because fiscal policy is intrinsically less effective in a technical sense, but because the decision-making process involves much greater risks of bad timing and expansionary bias. Whereas monetary policy has been delegated to independent central banks, which have been given clear stabilisation objectives, fiscal policy is decided in drawn-out parliamentary processes where stabilisation aspects become intertwined with income distribution and social efficiency aspects. This raises the question of whether one can learn some lessons for the fiscal policy decision-making process from the institutional changes that have been adopted in the field of monetary
9780230_547018_08_cha06.indd 214
3/19/2009 8:14:17 PM
Fiscal Policy and Macroeconomic Stabilisation 215
policy. Indeed, there exists recent literature which asks precisely this question. The contributions include von Hagen and Harden (1994), Eichengreen et al. (1995), Saint-Paul (1995), Calmfors (1995), Wren-Lewis (1996, 2000, 2002), Blinder (1997), Ball (1997), Business Council of Australia (1999), The Economist (1999), Eichengreen et al. (1999), Seidman (2001), Wyplosz (2002a), and the Swedish Government Commission on Stabilisation Policy in the EMU (2002). The main theme in this literature is whether one could improve both budget discipline and the effectiveness of fiscal policy as a stabilisation tool through delegation of some fiscal policy decisions to an independent agency that is assigned clear policy objectives by the political system, and whether this would be compatible with accepted principles of democratic governance. The fiscal agency would then be allowed to decide fiscal policy within predetermined limits without political interference in a way similar to how independent central banks pursue monetary policy. Different authors have used different terminologies for such a fiscal policy agency. We shall refer to it as a fiscal policy committee (FPC for short) to stress the parallel with monetary policy committees in central banks. The proposals vary as to the range of decision-making powers given to the FPC. The most far-reaching proposal is that of Blinder (1997) who proposes that an FPC should be given the power to decide the tax structure but not the level and composition of government expenditures, on the basis of general objectives for income distribution and social efficiency formulated by the legislature. Blinder supports the delegation with the argument that appropriate decisions on tax policy require a long-term perspective, technical expertise, and that undue influences from particularistic interest groups are avoided. Other far-reaching proposals are those of von Hagen and Harden (1994), Eichengreen et al. (1995), and Eichengreen et al. (1999). These contributions propose that a FPC should decide a “debt change limit,” which the government would be obliged to follow. Here, the motivation is to enhance fiscal discipline by taking the decisions on long-run debt accumulation out of the political sphere which is assumed to suffer from a deficit bias. However, most of the proposals focus strictly on the stabilisation aspect. This means that not only decisions on the size and structure of government expenditures, but also on the basic tax structure and long-run debt accumulation (the fiscal balance over the cycle or the cyclically adjusted annual fiscal balance) should remain in the political sphere.13 The FPC would only be delegated the power to decide how the budget balance should vary around the medium-term target depending on the cyclical situation or to vary certain tax rates around predetermined base levels. The underlying idea is that stabilisation of the business cycle is a commonly shared objective, which requires more technical expertise but fewer political trade-offs than other fiscal policy decisions. Another motive is to separate the stabilisation policy
9780230_547018_08_cha06.indd 215
3/19/2009 8:14:17 PM
216
Designing the European Model
aspects of fiscal policy from income distribution and resource allocation aspects such that stabilisation decisions are not “contaminated” by other considerations. The idea is also to reduce the risk that fiscal policy changes undertaken for stabilisation reasons involve a deficit bias. Other motives are to shorten decision lags and to make it easier to reverse fiscal policy decisions. 6.3.2 Two models of delegation The idea that part of fiscal policy could be delegated is no doubt unfamiliar and surprising to most people. To quote The Economist (1999), most people would probably regard “with horror” the idea that tax rates would be adjusted “by a band of unelected officials.” At the same time, there has been a general trend in many areas of economic policy making for politicians to focus more on setting the overall objectives and then delegate the operational decision-making to various bodies. The idea is to remove the actual implementation of policy from day-to-day politics. Areas where this has happened include, in addition to monetary policy, competition policy (Majone 1996) as well as regulation and supervision of financial markets (see Chapter 11). Even though delegation of national fiscal policy decisions to independent national bodies is not at present on the political agenda, there is a case for initiating a discussion of this possibility. It can be viewed as a contrast to the recent proposals that the European Commission should play a larger role in evaluating and approving national fiscal policies, which also represents a delegation to a non-elected body. If one takes the subsidiarity principle (see Chapter 10) seriously, one should explore solutions at the national level as an alternative to an enhanced role for the European Commission in the area of fiscal policy. In the literature on fiscal policy delegation, two basic models have been proposed. The first model implies that the FPC decides the annual budget balance, the second one that the FPC is allowed to vary specific tax rates or government expenditures. The FPC decides the annual budget balance According to this model, which has been proposed by Wyplosz (2002a, b), the parliament would decide a budget target over the cycle or a target for the cyclically adjusted budget balance. Given this constraint and some general guidelines on the roles of automatic stabilisers and discretionary policy (for example, specifying that discretionary fiscal policy measures should only be undertaken when output gaps are of a certain size, as proposed by the Swedish Government Commission on Stabilisation Policy in the EMU 2002), the FPC would be given the right to decide the annual budget target. The parliament would commit itself to follow the recommendations of the FPC. The FPC would also have to monitor both budget and cyclical
9780230_547018_08_cha06.indd 216
3/19/2009 8:14:17 PM
Fiscal Policy and Macroeconomic Stabilisation 217
developments over the fiscal year and have the power of requesting amendments to the budget. According to this model, the parliament retains the right to decide through which tax and expenditure changes the annual budget target should be met. This means that the parliament controls the income distribution and social efficiency aspects of fiscal policy. As the FPC would determine only the variations around the path of government debt over the cycle, but not the path itself, its decisions would have a negligible effect on intergenerational equity. The flip side of retained political decisions on all individual taxes and expenditures is that the FPC does not acquire full control over the stabilisation aspects of fiscal policy. To the extent that different taxes and expenditures have different multipliers, as discussed in Section 1.2, political decisions on tax and expenditure changes consistent with the budget target of the FPC can still affect aggregate demand to a significant degree. This may seriously complicate the task of the FPC, as it may not be possible to offset such effects without violating the long-run target for the fiscal balance. The FPC varies individual tax rates or government expenditures According to the second model (see for example Ball 1997; Business Council of Australia 1999; and Seidman 2001), the parliament would again take the decision on the medium-term target for the fiscal balance. But in contrast to the first model, the political sphere would only decide base rates for some taxes and base levels for some government expenditures (which would have to be consistent with the medium-term target for the fiscal balance). The FPC would then be granted the right to vary these tax rates and expenditure levels around their base values within prespecified limits in order to stabilise cyclical fluctuations. This could be done in two different ways. A first possibility is that the parliament determines in advance which fiscal instruments should be varied if the need arises. The simplest alternative is to give the FPC control over only one specific fiscal instrument. Alternatively, the parliament could prescribe ex ante that discretionary fiscal policy action should have a given composition (for example, 30 per cent of a fiscal stimulus should be a VAT decrease, 20 per cent a reduction in employers’ payroll taxes, 10 per cent a reduction in personal income taxes for lowincome earners, and 40 per cent an increase in government consumption), as proposed by Seidman (2001). This way the FPC decides only the overall size of fiscal stabilisation measures, but the political sphere retains control also over the short-run income distribution effects. A disadvantage of predetermining the composition of fiscal stabilisation measures is that different policy responses may be called for depending on the type of macroeconomic disturbance. For example, an increase in government employment may not be an appropriate response to a reduction in export demand. So, deciding on the composition of stabilisation packages once and for all may unduly constrain policy choices in a given situation.
9780230_547018_08_cha06.indd 217
3/19/2009 8:14:17 PM
218 Designing the European Model
Predetermining the composition of discretionary stabilisation measures would also require the parliament to form an informed view of which instruments are “on average” the best. This choice would be most important if one only delegates a single fiscal instrument to the FPC. Which one should then be chosen? There seems to be a presumption in the delegation literature in favour of taxes (Ball 1997; Blinder 1997; Business Council of Australia 1999). If so, one might argue that VAT changes could be a good candidate, as they affect private consumption in a similar way as interest rate changes, which are already subject to delegation (see Section 1.2). Delegation of the decision on VAT changes would shorten decision lags. This is likely to be particularly important for this instrument, because long decision lags could actually reverse the effect of policies: for example, a temporary VAT increase to cool off a boom will have an expansive demand effect in the period before it enters into force (see Section 1.2). However, variations in government consumption, for example through variations in general grants to regional and municipal authorities, might very well also be a suitable stabilisation policy parameter for an FPC. As discussed in Section 1, recent research suggests that expenditure multipliers may be larger than tax multipliers. Delegation to an FPC could mitigate the problem of irreversibility, which is usually regarded an important argument against increases in government expenditures in recessions (Wijkander and Roeger 2002; Swedish Government Commission on Stabilisation Policy in the EMU 2002).14 Another possibility would be that the parliament decides on a set of fiscal instruments that the FPC can vary within certain limits, but leaves the committee complete freedom to choose which of these instruments to be used in a specific situation. This would allow the FPC to use the instrument combination it finds most appropriate at each point of time. It would also allow the FPC to adjust the use of instruments to changes over time in the way economists judge their effectiveness.15 With such discretionary power over which instruments to use in specific situations, the FPC acquires a larger – but still limited – influence on income distribution (and social efficiency), as there is no longer a requirement that individual tax rates and government expenditures be changed symmetrically over the cycle. There is one important difference between letting the FPC vary the deficit target and letting it vary specific tax rates or expenditures. According to the former arrangement, the estimates of the FPC of the cyclical situation of the economy are automatically binding for the government. With the latter arrangement, one might fear that overoptimistic judgements of potential output on the part of the government causes it to systematically overestimate the cyclically adjusted balance, which might contribute to a deficit bias. A possible way of addressing this problem is to require that the estimates of the cyclically adjusted balance be based on the judgements of the FPC in the latter arrangement, too.
9780230_547018_08_cha06.indd 218
3/19/2009 8:14:17 PM
Fiscal Policy and Macroeconomic Stabilisation 219
6.3.3 Is there a democratic problem with delegation? To ensure that fiscal stabilisation policy decisions are taken at arm’s length from day-to-day politics, it has been suggested that the fiscal policy committee should have a similar degree of independence as a central bank. This would imply that the committee is not permitted to take instructions regarding individual decisions from the government or the parliament, and that the latter institutions are not permitted to give such instructions. Appointments should be long-term and non-renewable. Committee members should have professional competence: either earlier practical experience of economic policy making or analysis from ministries of finance, central banks, international organisations (like the IMF, the World Bank or the European Commission), private banks and so on, or academic competence in the field of stabilisation policy and macroeconomic analysis. The FPC should be granted a long-term budget, which could not be changed from year to year. The most common objection to delegation of fiscal stabilisation decisions is that it would interfere with conventional principles of democratic decision-making. How should one think about this? In any democratic society there exists the general problem of how to allocate decisions between the political and the technocratic spheres. This is done in different ways in different societies and the allocations also change over time. Different trade-offs are made in different areas of policy-making. There are no given standards, although there is a tendency to regard the current government institutions “as if they were the natural order of things” (Blinder 1997) and not subject to the possibility of reforms until the very moment when such reforms take place. The most important consideration for where to draw the line between political and technocratic decisions in a given area is the relative importance of value judgements and technical expertise (Majone 1996). The reason why most proposals on the delegation of fiscal policy exclude the size and composition of government expenditures, the tax structure and the size of long-run government debt is that value judgements on income distribution are crucial for these decisions. In contrast, macroeconomic stabilisation involves much less of value judgements and is more a question of technically finding the best ways of achieving commonly shared objectives (see Section 3.1). Indeed, it is difficult to find any fundamental difference in this respect between fiscal policy and monetary policy to stabilise the business cycle. As noted by Blinder (1997), most of the arguments against delegation of fiscal stabilisation policy decisions could also be used against the already existing delegation of monetary policy. To ensure the legitimacy of delegation of fiscal policy, an FPC would, of course, have to be subject to democratic oversight and accountability. Even if delegation is likely to improve fiscal policy “on average,” democratic control would be needed to reduce the risk that the FPC might at times pursue
9780230_547018_08_cha06.indd 219
3/19/2009 8:14:17 PM
220
Designing the European Model
idiosyncratic objectives or just make bad technical judgements. Some lessons could be learnt from monetary policy, but one could also go further in some respects. ●
●
●
●
●
Appointments of the members of the FPC would be made by the government and be subject to approval by the parliament. The candidates should be subject to questioning in parliament before they are confirmed, as is the case, for example, with members of the Federal Reserve Board in the United States (in the Senate). There is a similar procedure for the members of the Executive Board of the ECB in the European Parliament, although the parliament’s confirmation is not formally required. The objectives of the FPC should be determined by the legislature. A high degree of transparency should be required of the committee. It would have to explain all its decisions to the general public and to publish background fiscal reports (corresponding to the inflation reports of, for example, the Bank of England and Sveriges Riksbank) at regular intervals. The minutes of the FPC meetings and voting records should be published. The members of the FPC should regularly take part in public hearings in the parliament. The parliament should carry out ex post evaluations of the committee’s performance with the help of outside expertise. If the FPC fails over a period of years to achieve its objectives by a large margin – which needs to be given a clear operational definition ex ante – the parliament should have the possibility of dismissing the whole committee or individual members of it. Preferably, such dismissal should require a qualified majority to protect the FPC against misuse of this possibility. One could also conceive of an escape clause, which would enable the parliament to override an individual decision by the FPC. Again, this could require a qualified majority. The ultimate check on a system of delegation is, of course, the possibility to abolish the system altogether if it does not work in the desired way.
In any discussion of democratic control, it is important to distinguish between the formal aspects of accountability and how the process works in practice. It has been argued that in practice there may be more accountability with the delegation of a specific “technical” task to an independent committee, with clearly defined objectives against which to measure performance, than to have it executed as one of many simultaneous tasks by the government (Majone 1996). The argument is that the assignment of well-defined tasks to independent bodies makes it easier to “nail down” mistakes than if a government is at the same time to be held accountable for its performance in a large number of fields through the ordinary political process.
9780230_547018_08_cha06.indd 220
3/19/2009 8:14:17 PM
Fiscal Policy and Macroeconomic Stabilisation 221
6.3.4
Weaker forms of fiscal policy committees
It is an open question whether one could find forms for delegating the decisions on fiscal policy aiming at macroeconomic stabilisation that would be acceptable to the general public. However, there is good reason to initiate a discussion. Most people are likely to consider weaker forms of delegation politically more realistic, at least in a short-term perspective. One such possibility would be to give an FPC control only over a welldefined “rainy-day” stabilisation fund, but leave the political sphere in full command of the rest of fiscal policy (see also Section 2.3). Such a fund could be built up to a maximum level through specific tax receipts in booms and then run down through tax rebates in recessions. Many might regard this alternative as “less controversial” than the proposals in Section 3.2, because the powers of the FPC would be more clearly delineated and would not interfere with the normal budget process. The idea has some resemblance with the buffer funds in Finland that were described in Section 1.2. These funds differ, however, from the institutions discussed here, because they are of a corporatist nature: they are controlled by the central labour market organisations and not by an independent committee of experts.16 A general drawback of the stabilisation fund solution, as outlined here, is that it might introduce a “double command” to the extent that the government uses the fiscal parameters under its control to influence the cyclical situation. Another alternative has been suggested by Blinder (1997). According to this, the ultimate decision on a fiscal policy proposal of the FPC should be taken by the legislature, but be subject to a simple up-or-down vote. The proposal would thus have to be either accepted without any changes or rejected. In the latter case, one possibility could be to freeze tax rates and nominal government expenditures at last year’s level. Such an arrangement would give the political sphere more influence than with delegation of the actual decision-making, but the FPC would still have a strong hand, as automatic “fiscal drag” would tend to strengthen the budget balance in case the committee’s proposal is rejected by the parliament. Finally, one could give the FPC only advisory functions, but with some teeth as proposed by the Swedish Government Commission on Stabilisation Policy in the EMU (2002).17 This alternative is our preferred option for the near future. The FPC could then be assigned the task of independently estimating the cyclical situation of the economy (the output gap) and various tax and expenditure elasticities on which the government must base its budget calculations. The FPC could also be required to make public recommendations to the government on the annual budget targets and on specific tax and expenditure changes. The government could be more or less free to deviate from the recommendations of the FPC. This might be allowed only under exceptional
9780230_547018_08_cha06.indd 221
3/19/2009 8:14:17 PM
222
Designing the European Model
circumstances. Alternatively, the government might be free always to deviate from the recommendations, but would then be required to formally explain the reasons in a specific parliamentary session. Calmfors (2002) has proposed that the minister of finance then should be obliged to take part in a “reversed” public hearing, where he/she has to explain to the FPC why its recommendations are not being followed. The idea of these proposals is to increase the government’s reputational costs of deviating from the judgements of the FPC. The weakest form of an independent fiscal policy committee is just to let it take part as another voice in the public debate on the economy. This is more or less the way in which the Sachverständigenrat in Germany and the Economic Council in Denmark work. In that case the influence of the independent group of experts depends mainly on the reputation it can build up over time through its judgements and its ability to market itself. Its role can be enhanced by requiring the government to respond formally to the reports of the expert group (Swedish Government Commission on Stabilisation Policy in the EMU 2002). Some contributions have proposed that monetary policy committees (or executive boards) in existing central banks could function also as fiscal policy committees (Ball 1997; Seidman 2001; and Wren-Lewis 2002). One motive is that there would be small set-up costs if one uses an existing institution which already has an independent status and has acquired credibility for prudent stabilisation policies. Also, it might be regarded as less controversial to build on an already accepted institution rather than to establish a new one. Another argument has been that such an arrangement would facilitate co-ordination between fiscal and monetary policy. However, this argument does not apply to EMU, where monetary policy is centralised and fiscal policy decentralised. Here, it would rather be a question of finding a new role for the boards of the national central banks.18 National fiscal institutions that enhance the incentives for fiscal discipline and effective stabilisation policies reduce the risk of conflicts between national stabilisation objectives and the common EU fiscal rules. In a longer time perspective, the common fiscal framework in the EU might allow for and even encourage delegation of national fiscal stabilisation decisions along the lines we have suggested. One can conceive of several ways of doing this. For example, one could link our proposal of more flexible deficit ceilings in Section 2.2 to the existence of independent national fiscal authorities: low-debt countries could be allowed to use their higher deficit ceilings only if this is approved by the national FPC. Or if one adopts a system of extra-budgetary “rainy-day” funds, these might be used to avoid breaches of the three-per cent deficit limit only after a decision by an independent national fiscal authority. Recommendations and early warnings in the case of deviations from set budget targets could also take into account how these relate to the decisions of an independent fiscal authority.
9780230_547018_08_cha06.indd 222
3/19/2009 8:14:17 PM
Fiscal Policy and Macroeconomic Stabilisation 223
Notes 1. The Council of Ministers of the European Union is termed the Ecofin Council when it is made up of the economics and finance ministers of the member states. 2. This seems indeed to be vindicated by past experiences according to, for example, Leibfritz et al. (1994), Buti et al. (1997), and European Commission (2001). But these empirical conclusions have to some extent been challenged by Mélitz (2002) and Wyplosz (2002). 3. According to Fatàs and Mihov (2002), an increase in the ratio of government expenditure from 40 to 50 percent of GDP reduces the standard deviation of real GDP growth among the OECD countries by around 0.5. 4. Some of the discussion on to what extent “prices should be left to do their job” has focused on whether demand shocks are internal or external. It has been argued that price adjustments are appropriate only in the latter case (Blanchard, 2001; European Commission, 2002d). In our view, these are not the relevant considerations, because a temporary increase in external demand is as problematic as a temporary increase in internal demand if it leads to a real exchange rate appreciation that is hard to reverse. 5. This latter tax-shifting effect has been shown to be empirically strong in the Nordic countries (see Calmfors and Nymoen, 1991; Rødseth and Nymoen, 1999; and Calmfors and Uddén Sonnegård, 2001). The effects mentioned in the text could be counteracted, because compensating wage claims are triggered by the CPI rises associated with higher output prices when payroll taxes are increased, but this effect is likely to be small compared to the other effects. 6. As unemployment benefits were earlier financed on a purely “pay-as-you-go” basis in Finland, variations in employer contribution rates had a procyclical impact on the economy with rates going up in recessions and down in booms. 7. The European Council consists of the heads of state or government of the EU countries. 8 See the note to Table 6.7. 9. One should note the similarity between this argument and the argument that incumbent governments may choose deficit policies to constrain the possibilities of future governments of other political colours to favour their constituencies. 10. See also the discussion in Section 1.1. 11. In addition, one can, of course, also institute a rule that a procyclical loosening of fiscal policy in a boom represents a violation of the budgetary requirements in the SGP, as proposed by the Commission (European Commission, 2002b). But one would expect our proposal to have more bite, as it does not require discretionary decision-making. 12. Due to the initially low debt levels and the tendency to large debt-to-GDP reductions following from high nominal GDP growth, our proposed link between debt and deficit ceiling is also likely to give the accession countries more scope for investment in public infrastructure than the incumbent EU member states, which would seem desirable. 13. Note that targets for the “fiscal balance over the cycle” and “the cyclically adjusted annual fiscal balance” are not identical requirements. A target for the fiscal balance over the cycle implies, for example, that past misjudgements of the cyclically adjusted fiscal balance should be compensated for in later phases of the cycle, whereas this is not the case with the second formulation.
9780230_547018_08_cha06.indd 223
3/19/2009 8:14:17 PM
224 Designing the European Model 14. Another proposal is that of Saint-Paul (1995), who proposed that the size and composition of active labour market programmes should be delegated to an independent labour market board. This idea has been analysed theoretically by Calmfors (1995) and Johansson (2002). 15. In view of the way “fashions” change over time among economists, this might not, however, be an unmixed blessing. 16. Formally, the use of the Finnish buffer funds also requires the formal consent of the government (the minister of social affairs). 17. See also Wren-Lewis (1996) and Wyplosz (2002a) for similar proposals. 18. However, a system where actual fiscal policy decisions are taken by an FPC in the form of a national central bank and where the decisions could be overridden by parliamentary decisions or the FPC be dismissed would not be desirable. It would not be consistent with the independence of the European Central Bank System, which now forms a corner stone of the monetary union.
References Alesina, A. and R. Perotti (1995) “Fiscal Expansions and Adjustments in OECD Countries,” Economic Policy 21, 205–48. Balassone, F. and D. Franco (2000) “Assessing Fiscal Sustainability: A Review of Methods with a View to EMU,” Fiscal Sustainability, Banca d’Italia, Rome, 21–60. Balassone, F. and D. Franco (2001) “EMU Fiscal Rules: A New Answer to an Old Question?” Fiscal Rules, Rome, Banca d’Italia, 33–58. Balassone, F. and D. Monacelli (2000) “EMU Fiscal Rules: Is There a Gap?” Working Paper no. 375, Banca d’Italia, Rome. Ball, L. (1997) “A Proposal for the Next Macroeconomic Reform,” Victoria Economic Commentaries March, 1–7. Bean, C. (1998) “Discussion,” Economic Policy 26, 104–7. Beetsma, R. (2001) “Does EMU Need a Stability Pact?” in A. Brunila, M. Buti and D. Franco, eds, The Stability and Growth Pact, Palgrave, Basingstoke, 23–52. Blanchard, O. (2001) “Country Adjustments within Euroland. Lessons after Two Years,” Working Paper, MIT. Blanchard, O. and F. Giavazzi (2002) “Reforms that Can Be Done: Improving the SGP through a Proper Accounting of Public Investment,” mimeo, MIT and Bocconi University. Blanchard, O. and R. Perotti (2002) “An Empirical Characterization of the Dynamic Effects of Changes in Government Spending and Taxes on Output,” The Quarterly Journal of Economics, MIT Press, 117(4), 1329–68. Blinder, A. S. (1997) “Is Government Too Political?” Foreign Affairs 76(6), 115–26. Blinder, A. S. (2001) “The Economic Stimulus We Need,” New York Times, 28 September. Bordo, M. D. and O. Jeanne (2002) “Boom-Busts in Asset Prices, Economic Instability and Monetary Policy,” CEPR Discussion Paper no. 3398. Buiter, W.H. (2001) “Notes on ‘A Code for Fiscal Stability’,” Oxford Economic Papers 53, 1–19. Buiter, W. H., G. Corsetti and N. Roubini (1993) “Excessive Deficits: Sense and Nonsense in the Treaty of Maastricht,” Economic Policy 8, 57–100. Business Council of Australia (1999) “Avoiding Boom/Bust: Macroeconomic Reform for a Globalised Economy,” Discussion Paper no. 2, Melbourne, Business Council of Australia.
9780230_547018_08_cha06.indd 224
3/19/2009 8:14:17 PM
Fiscal Policy and Macroeconomic Stabilisation 225 Buti, M. and B. Martinot (2000) “Open Issues on the Implementation of the Stability and Growth Pact,” National Institute Economic Review 174, 92–104. Buti, M., D. Franco and H. Ongena (1997) “Budgetary Policies during Recessions – Retrospective Application of the ‘Stability and Growth Pact’ to the Post-War Period,” Economic Papers no. 121, European Commission. Buti, M. and G. Giudice (2002) “Maastricht’s Fiscal Rules at Ten: An Assesment,” Working Paper, European Commission. Buti, M., S. Eijffinger and D. Franco (2002) “Revisiting the Stability and Growth Pact: Grand Design or Internal Adjustment?” mimeo, European Commission. Calmfors, L. (1995) “What Can We Expect from Active Labour Market Policy,” Beihefte der Konjunkturpolitik 43, 11–30. Calmfors, L. (1998) “Macroeconomic Policy, Wage Setting and Employment – What Difference Does the EMU Make?” Oxford Review of Economic Policy 14, 125–51. Calmfors, L. (2001) “Macroeconomic Policy Coordination in the EU: How Far Should It Go?” Swedish Economic Policy Review 8, 3–14. Calmfors, L. (2002) “Fiscal Policy as a Stabilisation Policy Tool in the EMU,” paper to the EPRU Network Conference, 23 May, Copenhagen University. Calmfors, L., A. Booth, M. Burda, D. Checchi, R. Naylor and J. Visser (2001) “The Future of Collective Bargaining in Europe,” in T. Boeri, A. Brugiavini and L. Calmfors, eds, The Role of Unions in the Twenty-First Century, Oxford University Press, Oxford, 1–155. Calmfors, L. and E. Uddén Sonnegård (2001) “Explaining Wage Developments,” Economic Review 4, 38–67. Calmfors, L. and G. Corsetti (2002) “A Better Plan to Loosen the Pact,” Financial Times, 26 November. Calmfors, L. and G. Corsetti (2003) “How to Reform Europe’s Fiscal Policy Framework,” World Economics Journal 4(1), 109–16. Calmfors, L., H. Flam, N. Gottfries, M. Jerneck, R. Lindahl, J. Haaland Matlary, C. Nordh Berntsson, E. Rabinowicz and A. Vredin (1997) EMU – A Swedish Perspective, Kluwer Academic Publishers, Dordrecht. Calmfors, L. and R. Nymoen (1991) “Real Wage Adjustment and Employment Policies in the Nordic Countries,” Economic Policy 5, 397–448. Canzoneri, M. B. and B. T. Diba (2001) “The SGP: Delicate Balance or Albatross?” in A. Brunila, M. Buti and D. Franco, eds, The Stability and Growth Pact, Palgrave, Basingstoke, 53–74. De Grauwe, P. (2002) “Europe’s Instability Pact,” Financial Times 22 July. The Economist (1999) “Economics Focus: Fiscal Flexibility: Could Finance Ministers Learn a Few Tricks from Central Bankers?” 27 November, 80. The Economist (2002) “The Euro-Zone’s Stability Pact: Rules Are Made to Be Bent, Aren’t They?” 27 July, 27–8. Edelberg, W., M. Eichenbaum and J. Fischer (1998) “Understanding the Effects of a Shock to Government Purchases,” mimeo, Northwestern University. EEAG (2002) Report on the European Economy, European Economic Advisory Group at CESifo, Munich. Eichengreen, B. and C. Wyplosz (1998) “The Stability Pact: More than a Minor Nuisance?” in D. K. H. Begg, J. von Hagen, C. Wyplosz and K. F. Zimmerman, eds, EMU: Prospects and Challenges for the Euro, Blackwell, Oxford, 65–114. Eichengreen, B., J. von Hagen and I. Harden (1995) “Hurdles too High: Improving Budget Procedures Is the Best Preparation for EMU,” Financial Times 28 November, 16.
9780230_547018_08_cha06.indd 225
3/19/2009 8:14:17 PM
226
Designing the European Model
Eichengreen, B., J. von Hagen and R. Hausman (1999) “Reforming Budgetary Institutions in Latin America: The Case for a National Fiscal Council,” Open Economies Review 10, 415–42. Elmendorf, D. W. and N. G. Mankiw (1999) “Government Debt,” in J. B. Taylor and M. Woodford, eds, Handbook of Macroeconomics vol. 1C, Elsevier, Amsterdam, 1615–69. European Commission (2001) “Public Finances in the EMU–2001,” European Economy 3. European Commission (2002a) “Autumn 2002 Economic Forecasts,” European Economy 5. European Commission (2002b) Communication from the Commission to the Council and the European Parliament: Strenghening the Co-ordination of Budgetary Policies, Brussels, 27 November. European Commission (2002d) “Public Finances in the EMU–2002,” European Economy 3. Fatás, A. and I. Mihov (2001) “Government Size and Automatic Stabilizers: International and Intranational Evidence,” Journal of International Economics 55, 3–28. Fatás, A. and I. Mihov (2002) “The Case for Restricting Fiscal Policy Discretion,” Working Paper, INSEAD. Financial Times (2002a) “Commission Chief Hints that Pact Is on Last Legs,” 18 October. Financial Times (2002b) “Fear of more EU Deficit Breaches,” 27–28 July. Financial Times (2002c) “Second Again,” 29 July. Frederiksen, N. K. (2002) “Fiscal Sustainability in the OECD, December 2001,” Working Paper, Ministry of Finance in Denmark. Föttinger, W. (2001) “Balanced Budget Versus Golden Rule: On the Remediability of Fiscal Restrictions,” in Fiscal Rules, Rome, Banca d’Italia, 331–66. Galí, J. (1994) “Government Size and Macroeconomic Stability,” European Economic Review 38, 748–56. Giavazzi, F. and M. Pagano (1996) “Non-Keynesian Effects of Fiscal Policy Changes: International Evidence and the Swedish Experience,” Swedish Economic Policy Review 3, 67–103. Giavazzi, F., T. Japelli and M. Pagano (2000) “Searching for Non-Linear Effects of Fiscal Policy: Evidence from Industrial and Developing Countries,” European Economic Review 44, 1259–89. von Hagen, J., A. Hughes and R. Strauch (2002) “Budgetary Institutions for Sustainable Public Finances,” in M. Buti, J. von Hagen and C. Martinez-Mongay, eds, The Behaviour of Fiscal Authorities, Basingstoke, Palgrave, 94–112. von Hagen, J. and I. J. Harden (1994) “National Budget Processes and Fiscal Performance,” European Economy Reports and Studies 3, 311–408. Hemming, R. and M. Kell (2000) “Promoting Fiscal Responsibility: Transparency, Rules and Independent Fiscal Authorities,” Fiscal Rules, Rome, Banca d’Italia, 433–59. Holm, P., J. Kiander and P. Tossavainen (1999) “Social Security Funds, Payroll Tax Adjustment and Real Exchange Rate: The Finnish Model,” VATT, Government Institute for Economic Research, Discussion Paper no. 198, Helsinki. Johansson, Å. (2002) “The Interaction between Labor Market Policy and Monetary Policy: An Analysis of Time Inconsistency Problems,” in Essays on Macroeconomic Fluctuations and Nominal Wage Rigidity, Stockholm, Institute for International Economic Studies, 15–55.
9780230_547018_08_cha06.indd 226
3/19/2009 8:14:17 PM
Fiscal Policy and Macroeconomic Stabilisation 227 Kell, M. (2001) “An Assessment of Fiscal Rules in the United Kingdom,” IMF Working Paper WP/01/91. Knight, B. and A. Levinson (1999) “Rainy Day Funds and State Government Savings,” National Tax Journal, LII 3, 459–72. Kopits, G. and S. Symansky (1998) “Fiscal Policy Rules,” IMF Occasional Paper no. 162. Leibfritz, W., D. Roseveare and P. van den Noord (1994) “Fiscal Policy, Government Debt and Economic Performance,” Working Paper no. 144, OECD Economics Department. Majone, G. (1996) Regulating Europe, London, Routledge. McGranahan, L. (1999) “State Budgets and the Business Cycle: Implications for the Federal Balanced Budget Amendment Debate,” mimeo, Federal Reserve Bank of Chicago. Melitz, J. (2000) “Some Cross-Country Evidence about Fiscal Policy Behaviour and Consequences for EMU,” European Economy 2, 3–21. van den Noord, P (2002) “Automatic Stabilizers in the 1990s and beyond,” in M. Buti, J. von Hagen and C. Martinez-Mongay, eds, The Behaviour of Fiscal Authorities, Palgrave, Basingstoke, 130–48. Nymoen, R. and A. Rødseth (1999) “Nordic Wage Formation and Unemployment Seven Years Later,” Memorandum no. 10/99, Department of Economics, Oslo University. Perotti, R. (1998) “The Political Economy of Fiscal Consolidations,” Scandinavian Journal of Economics, 100(1), 367–94. Pisani-Ferry, J. (2002) “Fiscal Discipline and Policy Coordination in the Eurozone: Assesment and Proposals,” Working Paper, Université Paris-Dauphine. Ramey, V. and M. Shapiro (1997) “Costly Capital Reallocation and the Effects of Government Spending,” Carnegie-Rochester Conference Series on Public Policy 48, 145–94. Rotemberg, J. and M. Woodford (1992) “Oligopolistic Pricing and the Effects of Aggregate Demand on Economic Activity,” Journal of Political Economy 110, 1153–207. Seidman, L. (2001) “Reviving Fiscal Policy,” Challenge 44(3), 17–42. Storesletten, K., C. I. Telmer and A. Yaron (2001) “The Welfare Cost of Business Cycles Revisited: Finite Lives and Cyclical Variation in Idiosyncratic Risk,” European Economic Review 45, 1311–39. Swedish Government Commission on Stabilisation Policy in the EMU (2002) Stabilisation Policy in the Monetary Union, SOU 2002:16, Stockholm, Fritzes. Taylor, J. B. (2000) “Reassessing Discretionary Fiscal Policy,” Journal of Economic Perspectives 14, 21–36. Uhlig, H. (2002) “One Money, but Many Fiscal Policies in Europe: What Are the Consequences?” CEPR Discussion Paper no. 3296. Walton, D. (2002) “Europe’s Stability Pact: In Need of New Clothes,” Global Economics Paper no. 81, Goldman Sachs. Wijkander, H. and W. Roeger (2002) “Fiscal Policy in EMU: The Stabilization Aspect,” in. M. Buti, J. von Hagen and C. Martinez-Mongay, eds, The Behaviour of Fiscal Authorities, Palgrave, Basingstoke, 149–66. Wijkander, H. and W. Roeger (2001) “Skill-Biased Technical Change, Sectoral Heterogeneity and Wage Setting: Unemployment or Wage Inequality?” in Sestito Buti and Wijkander, eds, Taxation, Welfare and the Crisis of Unemployment in Europe, Cheltenham Edward Elgar. Wren-Lewis, S. (1996) “Avoiding Fiscal Fudge,” New Economy 3, 128–32.
9780230_547018_08_cha06.indd 227
3/19/2009 8:14:17 PM
228 Designing the European Model Wren-Lewis, S. (2000) “The Limits to Discretionary Fiscal Stabilization Policy,” Oxford Review of Economic Policy 16(4), 92–105. Wren-Lewis, S. (2002) “Fiscal Policy, Inflation and Stabilisation in EMU,” Working Paper, University of Exeter. Wyplosz, C. (2002a) “A Better Way to Balance the Budget,” Financial Times, 4 December. Wyplosz, C. (2002b) “Fiscal Policy: Rules or Institutions?” Stabiliseringspolitik i valutaunionen. SOU 2002:16 Underlagsrapporter. Stockholm, Fritzes, 291–328.
9780230_547018_08_cha06.indd 228
3/19/2009 8:14:17 PM
7 Acceding Countries: The Road to the Euro
7.1
Introduction
The ten acceding countries are expected to join the euro area at some point in the next few years. Their participation in the monetary union is seen as the end-point of their process of integration in the EU. In the “pre-accession phase,” the acceding countries have fulfilled the acquis communautaire in the area of EMU: they have implemented reforms to make their institutions compatible with joining the European Union (for example, they have made their central banks independent) and liberalised capital flows. At the start of the “accession phase,” they are expected to pursue policies aimed at fulfilling the nominal convergence criteria established by the Maastricht Treaty and qualify for joining the euro area. The third and final phase will start with the formal replacement of domestic currencies with the European currency. The EU treaties do not contemplate the possibility that acceding countries opt out of the euro: joining the EU automatically requires these countries to take the necessary steps to enter the EMU. The Maastricht Treaty lays out the conditions and procedures to become a member of the euro area, which are the same for both old and new members of the EU. Technically, acceding countries will join the European Union with “a derogation.” Thus, the only policy decision by acceding countries is whether to try to join EMU at an early or a late stage after accession. In principle, a country could choose to delay participation in EMU indefinitely. In this respect, it is useful to recall that, contrary to the United Kingdom and Denmark, Sweden is staying outside EMU without having negotiated any “opt-out clause.” But the perspective of postponing EMU participation does not seem to be appealing to current governments in acceding countries: all of them have declared their willingness to adopt the euro as soon as possible. The traditional argument in favour of EMU participation is credibility of low inflation, which applies to the newly acceding countries as it did to the 229
9780230_547018_09_cha07.indd 229
3/19/2009 8:15:09 PM
230 Designing the European Model
Box 7.1
The nominal convergence criteria
• Deficit of the general government must be below three percent of GDP. Gross debt of the general government must be below 60 percent of GDP, or declining toward 60 percent of GDP at a satisfactory rate. • Inflation rate must not exceed the average rate in the three EU countries with the lowest inflation rate by more than 1.5 percentage points. • Long-term interest rates must not exceed the average rate in the three EU countries with the lowest inflation rate by more than two percentage points. • Two years of participation in the Exchange Rate Mechanism II (ERM II) without major tensions in the foreign exchange market. ERM II replaced the ERM of the European Monetary System created in 1979. ERM II was established in 1997 with a resolution of the European Council in order to link the currencies of EU member states outside the euro area and the euro. Like ERM I, ERM II is also a multilateral exchange rate arrangement with a fixed, but adjustable, central parity and a fluctuation band around it. Countries participating in ERM II peg their exchange rates to the euro, allowing for fluctuations within a symmetric band of 15 percent on each side of the central parity. Interventions at the margin are automatic, unless they conflict with the primary objective of price stability in the euro area. Decisions concerning central parity adjustment, or fluctuation within margins narrower than 15 percent, are taken by mutual agreement. Any member (including the ECB) can initiate a confidential procedure to reconsider central rates. As established by the Ecofin Council in 2000, currency boards are compatible with ERM II participation. The following regimes are incompatible with ERM II: floating, crawling peg, peg against a currency different from the euro, and unilateral euroisation (Ecofin Council 8/11/2000). Participation in ERM II is voluntary, but acceding countries are expected to join as a precondition to adopting the euro. The procedure to join ERM II can be initiated at any time by a EU member state. There are no formal criteria for joining ERM II. Whether a country satisfies the convergence criterion of exchange rate stability will be judged by looking at a range of indicators. As discussed below in the text, a country is expected to keep its exchange rate close to the central parity in ERM II or experience a currency appreciation. The assessment will also take into account short-term interest rate differentials and the frequency and intensity of foreign exchange market interventions.
Southern European countries in the 1990s. A common currency eliminates currency risk and reduces interest rate differentials. Such credibility gains are an advantage from the point of view of macroeconomic stabilisation that should be set against the loss of national monetary policy as a stabilisation instrument and of exchange rate flexibility as an adjustment mechanism.1 Recent literature has stressed a number of effects of joining EMU which are not analysed in the traditional debate. First, a common currency is
9780230_547018_09_cha07.indd 230
3/19/2009 8:15:10 PM
Acceding Countries: The Road to the Euro
231
likely to increase trade within the EU. In this respect, adopting the euro is equivalent to a drop in transaction costs in cross-border exchanges of goods and services within the EU economic area. Second, by reducing the stock of external debt denominated in a foreign currency, adopting the euro will substantially reduce vulnerability to currency and financial instability (although in principle EMU countries could still issue large stocks of dollardenominated debt). We will discuss this point at length below. Taking it as a political fact that the accession countries will ultimately join EMU, the question of the optimal timing of such a move becomes crucial. The timing directly impinges on the acceding countries’ ability to stabilise their economies in the next few years and build an economic environment that favours high rates of investment and growth, economic integration and financial stability. Fiscal and monetary authorities in acceding countries now operate in a regime of high capital mobility. This is the result of a relatively rapid process of liberalisation and deregulation implemented in the last few years. But their domestic institutions and markets have only recently started to operate in a fully liberalised and deregulated system. Whether or not the financial and legal systems of these countries can weather volatile capital movements is perhaps too early to say, but it would be naïve to hope for the better and envision years without large (global or region-specific) shocks. As discussed in Chapter 5 of 2004 EEAG report, structural imbalances in these economies may cause acute problems. Deteriorating fiscal conditions could constrain the use of budget policies for stabilisation purposes. Stabilisation is likely to fall disproportionately on monetary and fiscal authorities, both from a macro perspective and from a financial stability perspective. In such an environment, mandatory adoption of a regime of limited exchange rate flexibility (the ERM) for two years before entering EMU is quite controversial and has stirred a considerable debate in both policy and academic circles (see Buiter and Grafe 2002, Begg et al. 2003 and ECB 2003 among others). In this chapter, we will reconsider the main issues related to the choice among alternative paths of transition to the euro. Independently of the exchange rate regime, a high degree of capital mobility in the transition to the euro will entail high financial risk. The choice of inappropriate exchange rate regimes can, however, magnify this risk greatly. The experiences from recent crises and financial turmoil in emerging markets show that wrong decisions can be extremely painful – the punishment for apparently small mistakes can be enormous. Based on this experience, we will try to single out the policy that could reduce vulnerability to currency and financial instability. We will devote a large part of our discussion to the debate on the costs and benefits of participating in “ERM II,” which is the Exchange Rate Mechanism with large fluctuation bands. However, the exchange rate regime is only one
9780230_547018_09_cha07.indd 231
3/19/2009 8:15:10 PM
232 Designing the European Model
dimension of the policy framework. What is ultimately important for acceding countries is pursuing the right stabilisation policies, strengthening their fiscal, financial and monetary institutions, making their economies flexible and choosing a consistent level of social protection. For the EU, policy priority should be given to strengthening its financial architecture, along the lines extensively discussed in Chapter 11.
7.2 Exchange rate regimes with liberalised financial markets: the current policy framework of acceding countries In Chapter 5 of the 2004 EEAG report, we have seen that the acceding countries are rapidly integrating their markets for goods and services in the world and the EU economies. Acceding countries are highly open – the average degree of openness (exports plus imports as a share of GDP) is around 100 per cent of GDP, compared with 70 per cent of the present EU countries (Poland and Cyprus are significantly below average, however). Approximately 60 per cent of the acceding countries’ imports and exports is with EU countries. A significant share of EU trade with the acceding countries is intra-industry, clearly reflecting cumulated foreign direct investment (FDI) by EU countries. Openness is destined to increase after accession. Thus, these countries will be highly exposed to terms of trade shocks and international demand spillovers from both within and outside the EU area. Acceding countries have removed most restrictions on capital mobility, with the exception of restrictions on the purchase of land by foreigners. Slovenia also keeps some controls on short-term capital and direct investment. Financial integration with the EU is already quite high. Over the last few years, two-thirds of capital inflows to acceding countries consisted of FDI flows: approximately 80 per cent of these flows originated in the EU. Some indicators show that the financial systems of the acceding countries are moving towards the EU performance standard (see European Commission 2003). In Cyprus and Malta the size of the financial sector is close to the EU average. But compared to the rest of the EU, the domestic financial sector in the acceding countries in Eastern Europe is still underdeveloped. Banks dominate the financial sector. In the last few years, most banks were privatised– only in Poland and Slovenia do the governments still retain ownership of some large financial institutions. The privatisation process has coincided with a change in ownership from domestic to foreign. In 2000, domestic credit amounted to 60 per cent of GDP in the acceding countries, against an average of 140 per cent in the euro area. Gross debt of individuals was quite low: the average for the acceding countries was seven per cent of disposable income, against an average of 50 per cent for the euro area. Stock market capitalisation was also low. Several indicators show lack of funding for small and medium-sized firms, especially in the initial stages of their life (European Commission 2003).
9780230_547018_09_cha07.indd 232
3/19/2009 8:15:10 PM
Acceding Countries: The Road to the Euro
233
The room for financial deepening in the Eastern European acceding countries is large. In the next few years, credit to both households and corporations, stock market capitalisation as well as activities by financial intermediaries will probably experience very rapid growth. Financial deepening can be extremely beneficial: it can relax credit constraints, provide a sufficiently diversified supply of funds to finance projects with different risk profiles, and create opportunities for risk diversification and the reallocation of consumption over time. Yet, a high speed of expansion can also cause a deterioration of allocative efficiency if it leads to excessive risktaking, potentially undermining the contribution to welfare of financial market development. 7.2.1 Country size, capital mobility and exchange rate regimes Figure 7.1 plots the exchange rates of the acceding countries, while Table 7.1 reports the exchange rate regime adopted in 2003 as well as changes that have occurred since the beginning of the 1990s. Exchange rate regimes are classified according to the official IMF classification allowing for some suggestions by various researchers (see von Hagen and Zhou 2002 among others). There are two striking features in Table 7.1, already noted by many commentators. The first is a positive correlation between country size and the flexibility of the exchange rate regime. Looking at 2003, smaller countries like Estonia and Lithuania have currency boards (arrangements that constrain monetary authorities to expand domestic money supply in line with international reserves), Latvia pegs its currency to a basket of international currencies within very narrow bands. Malta pegs to a basket of currencies
125
125
120
120 Czech Republic
115 110
115 110
Cyprus Hungary Malta
105 100
105 100
Poland Slovak Republic
95
95
90
90
85
85 Euro area
80
80
75
75 1999
Figure 7.1
2000
2001
2002
2003
2004
Nominal effective exchange rates (trade weighted [Q1 1999 = 100])
Source: IMF, International Financial Statistics.
9780230_547018_09_cha07.indd 233
3/19/2009 8:15:10 PM
234 Designing the European Model Table 7.1
The evolution of exchange rate regimes in acceding countries 1990
1995
2000
2003
Peg
Managed float
Peg to the euro (±15% band, de facto ±1–2% band) Free float, inflation targeting
Currency board Peg
Currency board
Cyprus Czech Peg Republic Estonia Latvia Lithuania Peg Hungary
Peg
Currency board Crawling bands
Currency board
Crawling bands Peg Managed float
Crawling bands
Currency board with peg to the euro Peg to SDR (euro weight is 29%; ±1% band) Currency board with peg to the euro Peg to the euro (±15% band), inflation targeting Peg to currency basket (±0.25% band) Free float, inflation targeting
Managed float Managed float
Managed float Managed float
Peg
Crawling bands
Malta Poland
Peg
Slovakia Slovenia
Peg
Note: IT stands for inflation targeting. Source: Von Hagen and Zhou (2002); Begg et al. (2003).
with a 70 per cent euro share, also within very narrow bands. Cyprus pegs to the euro, officially adopting large bands of fluctuation, but de facto pegging within extremely narrow bands. Conversely, larger countries adopt more flexible regimes. The Czech Republic and Poland have opted for a free float – central banks in these countries have adopted inflation targeting as their monetary strategy. The Slovak Republic and Slovenia pursue some form of managed exchange rate float. Hungary has adopted a regime with somewhat limited exchange rate flexibility combined with inflation targeting. Fixed exchange rate regimes provide a nominal anchor to pin down the price level in the economy. For the peg to be viable, domestic prices cannot move too far from international prices, because real exchange rate changes would have destabilising effects on aggregate demand. The choice of a more flexible exchange rate regime, instead, raises the issue of choosing a nominal anchor to guide private sector expectations of inflation. The largest acceding countries have opted for some form of inflation targeting – requiring monetary authorities to set explicit, yet contingent, goals in terms of rates of increase of the consumer price index (CPI). The second feature of Table 7.1 is a tendency of governments to move away from intermediate regimes of limited exchange rate flexibility after
9780230_547018_09_cha07.indd 234
3/19/2009 8:15:10 PM
Acceding Countries: The Road to the Euro
235
capital market liberalisation. Indeed, most of the acceding countries had adopted some form of soft peg at the beginning of the 1990s. After liberalisation of capital flows during the 1990s, they have moved either to flexible exchange rate regimes, or to hard pegs/currency boards that in principle exclude any realignment. In 2003, five countries have a currency board or a hard peg, whereas five have a free float or some form of flexible regime. 7.2.2 Free capital mobility and fixed exchange rates: inconsistent policy regimes? According to a view that has gained more and more supporters after the global turmoil of the 1990s, free capital mobility is inconsistent with any form of limited exchange rate flexibility. The reason is as follows. Rates of return from short positions in currencies that fall by 10, 20 or 30 percentage points in a short time-span are extremely high. With high capital mobility, taking a speculative position against a currency has low costs: investors will miss no opportunity to test the resolve of governments in maintaining their exchange rate target. As long as market participants think that the government can “realign” the exchange rate (that is, devalue the currency), the targeted parity will not be completely credible. Investors know that, depending on the circumstances, policy makers will prefer to modify the exchange rate rather than to sustain the enormous costs of extreme defence. These costs possibly involve protracted periods of sky-high interest rates, with devastating effects on the health of the banking and financial system, and therefore on investment, growth and employment. Many factors determine the circumstances that may force governments to devalue. Some of them have a domestic origin, but very often the shocktriggering waves of speculative attacks originate in the global economy. Recurrent examples include monetary shocks in the OECD countries (higher US interest rates), as well as turmoil/crises in other countries or markets. What makes the problem worse is that speculative attacks may be driven by self-fulfilling beliefs. If investors consider an exchange rate policy credible, they will have no reason to speculate against the currency, and the current level of the exchange rate will be viable. Otherwise, they will launch speculative attacks that will undermine the stability of the peg. The ensuing currency and financial collapses validate ex post the initial expectations of turmoil. Different factors and events determine the ability of domestic policymakers to guarantee stability of their domestic markets and financial institutions, but the country’s vulnerability to them ultimately depends on its fundamentals. Sound domestic policies and strong institutions can eliminate the possibility of self-fulfilling crises. Low credibility of stabilisation policies and weak institutions cannot.
9780230_547018_09_cha07.indd 235
3/19/2009 8:15:10 PM
236
Designing the European Model
7.3
Vulnerability to crises: lessons from emerging markets
As extensively documented in Chapter 5 of the 2004 EEAG report, the acceding countries are emerging markets: their income levels are considerably lower than in the rest of Europe, and their economies are growing rapidly. What lessons can we learn from the recent experience with macroeconomic stabilisation and exchange rate regimes of other emerging markets? The 1990s are rich in cautionary tales. During this decade, several factors contributed to the rapid growth of markets for emerging market assets, including low interest rates in the industrialised countries as well as the growth of financial instruments and the diversification of financial institutions. But, as is well known, severe financial turmoil hit Mexico, Thailand, Indonesia, Korea, Russia, Brazil, Ecuador, Turkey, Argentina and Uruguay, to name the main ones. Before a crisis, these countries had all liberalised their capital accounts (although to different degrees) and adopted policies of low inflation and macroeconomic stabilisation. In addition, all of them had struggled to reduce their high country-risk premia charged in international capital markets. In all these episodes, a common pattern emerges. Capital mobility is associated with high volatility of capital flows. After capital account liberalisation, emerging markets pursuing macro stabilisation programs typically experience a very large upsurge of capital inflows. Such inflows are driven not only by high rates of return in economies with relative scarcity of capital, but also and to a large extent by the attractiveness of short-term profit opportunities from speculative positions. This is because, just after capital account liberalisation, country-risk and/or inflation tend to keep domestic interest rates high relative to international rates. Gains in the credibility of domestic policies drive the process of “convergence” of domestic to international interest rate levels over time, but usually at a slow pace. Large capital inflows into the country can just as easily and suddenly stop, opening up enormous financing gaps (see Calvo 2003). 7.3.1 Convergence can create vulnerability When a country commits to a peg, its exchange rate, riding the wave of the “convergence process” is very attractive from the point of view of international fund managers and investors. These invest in short-term debt issued by the country at high interest rates, expecting stable or appreciating exchange rates. When debt is denominated in foreign currency, the interest rate typically includes a substantial country-premium over international rates. Speculative positions are invariably short-term, since investors prefer to be able to withdraw quickly from the country if there is any sign of trouble (i.e. rumours about devaluation or default). What we have described above is labelled “convergence play” in the literature on currency and financial crises. It has been common in most episodes
9780230_547018_09_cha07.indd 236
3/19/2009 8:15:10 PM
Acceding Countries: The Road to the Euro
237
of exchange rate-based stabilisation in emerging markets as well as in advanced economies. The “convergence play” became notorious in Europe during the period preceding the currency crises of 1992–93, when fund managers massively bought high interest rate assets denominated in, say, Italian lira, sometimes pretending to cover their positions by selling short assets denominated in D-marks (see Buiter, Corsetti and Pesenti 1998). In emerging markets, large inflows driven by the “convergence play” are problematic in many dimensions. First, by feeding spending on both consumption and investment goods, they usually lead to an overheating of the economy. Second, by expanding the demand for short-term assets denominated in domestic currency at a rate several times higher than the growth rate of the economy, they provide a strong incentive for financial intermediaries to expand their activities without paying sufficient attention to prudential standards. To the extent that newly available funds drive up housing and land prices, the rising value of collateral assets further magnifies the incentive to create credit. Excessive credit creation exacerbates the fragility of the financial system (see Corsetti, Pesenti and Roubini 1999). In addition, with abundant capital inflows, policymakers perceive a softening of credit constraints. Not only does this create an incentive to borrow too much: abundant liquidity may also make governments more willing to extend public guarantees on private projects. Liberalisation and privatisation magnify the distortions due to public guarantees and “connected lending” (where credit is obtained through political links). This, too, contributes to excessive risk-taking. Finally, and most importantly, since most debt is short-term and denominated in a foreign currency, the country is highly illiquid and exposed to destructive debt runs. Are acceding countries likely to experience a “convergence play”? One may argue that these countries have already experienced large inflows of capital, most of which in the form of foreign direct investment (FDI). Moreover, interest rates in acceding countries are not far above the international rate. Why should the composition of capital flows change in the future? The problem is that participating in the EU will change the international assessment of risk in these countries. The EU “seal of approval” and the macroeconomic stabilisation programmes that accompany accession are likely to induce a new wave of capital inflows, this time with a stronger portfolio component. In other words, stabilisation and convergence policies are very likely to attract portfolio managers chasing short-term and medium-term profit opportunities. Even if the external debt of acceding countries now reflects to a significant extent the cumulated stock of FDI, the composition of external debt may change rapidly in the next few years.2 7.3.2 Is foreign direct investment a solution? Suppose, however, that capital will still flow into these countries mostly in the form of FDI. Would this mean that these countries are sheltered from
9780230_547018_09_cha07.indd 237
3/19/2009 8:15:11 PM
238 Designing the European Model
crises and/or their adverse consequences? Indeed, FDI flows have two major advantages over foreign debt as regards financial and currency stability. First, they are driven by real investment opportunities and therefore tend to be long-term. The data show that they are much more stable than portfolio flows. Second, the return on FDI depends on the profitability of real investment and is therefore pro-cyclical and ultimately contingent on the performance of the economy. In the presence of currency and financial turmoil, FDI investors typically suffer capital losses: the international value of the country’s liabilities drops in a crisis. Consider instead external debt denominated in foreign currency. Payments on debt are not contingent. In a crisis associated with a drop in output and devaluation, the burden of foreign debt increases: the larger the rate of devaluation, the sharper the revaluation of the country’s external liabilities. These two advantages of FDI over foreign debt clearly reduce the vulnerability to financial turmoil of countries with a large share of FDI in total capital inflows. However, the argument in favour of FDI is sometimes taken one step further. It is argued that, since FDI investors lose from a currency collapse, they will be unwilling to speculate against the country in which they have relatively illiquid assets (or perhaps that they will even be willing to take long positions in domestic currency in episodes of turmoil!). If this argument were true, large FDI investments could in principle shield a country almost completely against currency crises. Unfortunately, this argument is wrong. The point is that, once rumours of devaluation spread, investors who own domestic capital have the strongest incentive to hedge against capital losses due to the drop in the international value of their assets in the country. They will therefore take a short position against the currency and/ or the stock market index in the country. So, while FDI flows have indeed many properties that strengthen financial stability, FDI investors can also cause massive volatility in short-term capital flows as a result of hedging strategies by firms with capital in the country at the onset of a crisis! To put it simply, it is largely a myth that FDI can eliminate speculation in the currency and asset markets (see Guimaraes and Morris 2003). 3 7.3.3 Currency and maturity mismatch When most flows from abroad are short-term and are directed towards assets denominated in international currencies, the maturity structure and currency denomination of the country’s external debt create a strong imbalance for the financial system as a whole. Unless firms, households and banks hedge their positions (the evidence is that they hardly do so), fluctuations in exchange rates and asset prices have strong effects on the balance sheets of domestic agents and institutions. Currency crises may turn into widespread bankruptcy in both the banking and the real sector. These problems are quite compelling in acceding countries, since their financial systems are already operating to a large extent in a foreign currency.
9780230_547018_09_cha07.indd 238
3/19/2009 8:15:11 PM
Acceding Countries: The Road to the Euro
239
Thus, liabilities and financial assets are already to a large extent euroised, and the degree of euroisation may be expected to increase in the future. Table 7.2 and Figure 7.2 provide some evidence on the extent to which liabilities in selected acceding countries are denominated in a foreign currency. Could banks in acceding countries insure against these balance sheet effects by lending domestically in foreign currency? In this case there would be no currency mismatch in the balance sheet of the banking system, since both assets and liabilities would be denominated in foreign currency. The problem, however, is that many domestic firms borrowing in foreign currency obtain a large share of their revenues from sales in the domestic market. To the extent that domestic prices do not adjust one to one with the exchange rate, nominal devaluation would worsen these firms’ balance sheets: given the value of their cash flow in domestic currency, the localcurrency value of their debt would increase with a drop in the exchange rate. A high rate of bankruptcies of borrowing firms would directly and indirectly affect the health of financial intermediaries. In other words, euroisation of both assets and liabilities of the banking sector in the acceding countries does not solve the problem: exchange rate risk is simply translated into default risk for the banks. 7.3.4 Does foreign ownership of banks shield against crisis? Another often heard argument is that financial stability in acceding countries is not a concern since the financial sector in these countries is dominated by banks, and a large number of banks are owned by EU-based financial institutions, responsible to their home country’s supervisors and
Table 7.2
Outstanding euro-denominated bank-deposits As % of total deposits
Slovenia Estonia Turkey Bulgaria Latvia Hungary Romania Slovakia Czech Republic Malta Poland Cyprus Lithuania
37.9 19.2 15.6 15.3 11.6 11.2 9.3 8.5 7.6 6.6 5.3 5 1.9
Source: “Review of the International Role of the Euro”, ECB December 2002.
9780230_547018_09_cha07.indd 239
3/19/2009 8:15:11 PM
240
Designing the European Model %
0.6 0.5 0.4 0.3 0.2 0.1 0 Italy
Figure 7.2
Germany
France
Czech Republic
Poland
Hungary
Turkey
Share of foreign currency liabilities in the banking system in 1999
Source: OECD, Database on Bank profitability financial statements of banks, 1999.
regulators. Thus, EU standards in supervision and regulation apply to a large part of the financial sector in acceding countries. This point is well taken, but the argument misses an important element. Good supervision and regulation may constrain excessive risk-taking but cannot eliminate balance sheet problems and/or liquidity runs. Bank headquarters abroad cannot be expected to intervene and provide the necessary funds to their branches in the country should these experience difficulties or be hit by liquidity runs. The argument that the role of central banks as lenders of last resort is less crucial when financial intermediaries are owned by foreign institutions (weak on logical grounds) has been definitely proved wrong by the recent experience of Argentina. To put it simply, foreign bank owners do not guarantee any liquidity provision or capital injection. Most crises fall into the grey area between liquidity and solvency, so that it is difficult to assess whether additional funds will save a specific bank or simply be lost in a bankruptcy. But even when crises are close to those of illiquidity, systemic nation-wide contagion creates a co-ordination problem among liquidity providers. Either all foreign financial intermediaries provide liquidity and the crisis is prevented or no individual intermediary has an incentive by itself to bring more funds into the country, as its own contribution will be insufficient to avoid systemic financial collapse and will therefore be lost in the speculative run. The pattern of strong capital inflows, driven by “convergence plays,” is present in the experience of many countries, whether or not their governments
9780230_547018_09_cha07.indd 240
3/19/2009 8:15:11 PM
Acceding Countries: The Road to the Euro
241
commit to peg the exchange rate, and independent of the specific features of the exchange rate regime. However, commitment to a peg may exacerbate the intensity and the consequences of “convergence plays” and strengthen the expectations of public sector involvement in bailing out both financial and non-financial private enterprises. 7.3.5 Stabilising an economy on a fast convergence track: challenges to monetary policy In a boom with large capital inflows, monetary authorities have a hard time stabilising the economy. If they raise interest rates to reduce domestic overheating, they may exacerbate the “convergence play.” If they try to sterilise the capital inflow, reserve accumulation can become excessive by any standard. Nominal and real appreciation of the exchange rate, while reducing overheating, can harm exports and create the premise for currency and financial crises in the future. Regulatory institutions and supervisors have a hard time enforcing prudential standards, as price signals (for example, a high value of collateral) may create a misperception of risk. What is particularly disturbing for policymakers is that any policy aimed at slowing down an overheating economy may not be easily communicated to the public, since accession to the EU has created very optimistic expectations of fast convergence. How can one distinguish between an overheating economy and an economy that is simply on a fast convergence track? Efforts of policymakers to avoid overheating could be interpreted as misguided policy. In Chapter 5 of the 2004 EEAG report, we have shown that convergence of Eastern Europe will realistically take about 20 to 25 years in the best-case scenario for the most advanced countries. More likely it will take 50 to 100 years in most countries. What is not clear is whether cyclical stabilisation can be well defined and understood by the public in a country that is far away from its steady state growth rate.
7.4 Reducing vulnerability Recent experience with stabilisation and crises in emerging markets (including Eastern European countries) suggests two important considerations in designing stabilisation policies in acceding countries. 7.4.1 Financial frictions and the size of the business cycle First, business cycles in acceding countries are likely to be more pronounced than in the EU and perhaps comparable to boom-bust-cycle episodes recently experienced in emerging markets. An important reason for cyclical variability is the currency denomination of external debt. To the extent that debt is denominated in foreign currency, in periods of real exchange rate appreciation the value of liabilities in the banking system will fall, allowing banks
9780230_547018_09_cha07.indd 241
3/19/2009 8:15:11 PM
242 Designing the European Model
to lend more. In periods of sudden reversal and depreciation, banks will be experiencing stress and/or crisis. Both cyclical upswings and downturns are likely to be more pronounced than in the current EU member countries (see Tornell and Westermann 2002). Early adoption of the euro can in principle eliminate this magnification mechanism of the business cycle amplitude. Yet, even after euroisation, there could be other financial market imperfections that would still contribute to large fluctuations in economic activity. 7.4.2 Crises are not necessarily bad news for long-run growth Second, while convergence by the acceding countries is likely to be “bumpy” (that is, characterised by occasional episodes of financial turmoil and crisis), this is not necessarily bad for long-run growth. The main issue is whether, on average, a country could be better off by avoiding instability altogether, even if this happens at the cost of slowing down the process of deregulation/ liberalisation of goods and financial markets and full economic integration in the European and world markets. Theory and evidence in this respect are not conclusive. For instance, on the one hand, risky balance sheets may lead to fragility and crisis; on the other hand, if economic agents are credit constrained, risky balance sheets are an (admittedly imperfect) way to overcome credit constraints during the boom phase of the cycle. Looking at the experience of emerging economies over the last 20 years, Tornell, Westermann and Martinez (2003) point out that, despite episodes of major crisis, Chile, Thailand and Korea are among the fastest growing economies in the developing world. In these cases, financial fragility notwithstanding, financial market liberalisation appears to be good for growth in the long run. This is not to say that crises are necessarily good for growth and that countries should disregard financial fragility in pursuing their growth process. But the experience of the last decade also makes it clear that there is no easy solution to the problem of reducing a country’s vulnerability to financial shocks. The following lessons apply. 7.4.3
A sustainable fiscal stance
First, there is no monetary and exchange rate stability without a strong fiscal stance. One of the key factors underlying the crisis of Argentina’s currency board was the deterioration of fiscal conditions implied by a deep conflict between the central and the provincial governments. Such conflict undermined any expectation of fiscal discipline in the near and less near future. Even if current fiscal imbalances were still small at the time of the crisis, international investors and institutions came to realise that the path of Argentina’s public finances would not be consistent with exchange rate stability. In other words, current public deficits and the outstanding stock of public debt provide a partial picture of a country’s fiscal stance. Fiscal sustainability
9780230_547018_09_cha07.indd 242
3/19/2009 8:15:11 PM
Acceding Countries: The Road to the Euro
243
crucially depends on the market assessment of future and contingent liabilities. In many instances, this gives rise to the possibility of self-fulfilling crises. In anticipation of large deficits caused by a crisis, a speculative attack on a country’s assets can create macroeconomic imbalances that cause a crisis and thus validate ex post the initial forecasts of fiscal troubles. Implicit and contingent fiscal liabilities are a crucial determinant of currency and financial fragility. These are magnified by inconsistent fiscal policy that feeds expectations of public bailout of firms in trouble and creates incentives for the private sector to take excessive risk. 7.4.4 Well-functioning financial markets Second, there is no currency and financial stability without well-functioning financial markets. The main problem is that capital account liberalisation and deregulation magnifies the economic distortions associated with inconsistent financial policy. In practice, excessive risk-taking means that, when undertaking projects, firms, households and banks believe that under some circumstances they will be able to avoid the bill if things go wrong – a bill involving both monetary and non-monetary costs. Reducing vulnerability does require strong financial regulation and supervision, which in turn involves the development of strong and efficient institutions. It also requires a process of privatisation of the economy accompanied by a clear definition of rules, laws and policies concerning risk management. In principle, this can be achieved in different ways, not necessarily through strict external regulation and supervision of financial firms but also through self-regulating bodies set up at industry level. But simply asserting the unwillingness by EU governments to bail out private financial and non-financial firms would not be effective, as such announcements lack credibility: investors know that the political pressure to intervene in crisis situations is very strong. In Chapter 11 we analysed the incentives for EU governments to intervene excessively at the national level. These incentives will become even stronger once the acceding countries are in the EU and even more so once they are in EMU. The tension between many national regulators and one monetary policy will be aggravated by the entry of new members with weaker institutional structures. 7.4.5 An exit strategy from a fixed exchange rate regime Third, fixed exchange rate regimes suffer from an “escape clause” or “exit strategy” problem. Suppose a government adopts a hard peg as a means to reduce inflation. To be successful, the government needs to commit to keeping the current parity indefinitely. However, to the extent that the strategy is not immediately credible and/or there are contracts in the economy that predetermine inflation in the near future, the economy will suffer from increasing relative price imbalances during the first years of the strategy.
9780230_547018_09_cha07.indd 243
3/19/2009 8:15:11 PM
244 Designing the European Model
High rates of inflation relative to the international rate will translate into a deterioration of competitiveness. By the same token, to the extent that low credibility translates into high risk premia in the asset markets, the stock of public debt will rise rapidly, whereas the capital stock will not (as investment will be relatively costly). At a given exchange rate, the correction of these imbalances requires the government to reduce domestic inflation below the international rate for some time, adding to the cost of disinflation and therefore raising the attractiveness of a nominal realignment of the exchange rate. But this possibility undermines the credibility of the policy strategy in the first place. Expectations of devaluation raise the costs of exchange rate based stabilisation and can eventually become self-fulfilling. Either the exchange rate is kept fixed no matter what, or a fixed exchange rate policy becomes a recipe for crises. The longer the attempt lasts to resist devaluation, the greater is the magnitude of cumulated economic imbalances. As the recent experience of Argentina shows, the bill of delaying the exit from unsustainable policies can become so large as to cause a collapse of prices, incomes and production. 7.4.6 A high degree of wage and price flexibility Fourth, systems of inflexible exchange rates require some degree of price and wage flexibility. Those emerging market economies that weathered well the global shocks of the 1990s (such as Chile) experienced swings in the real exchange rate on the order of 20 to 30 per cent over a business cycle. In these cases, nominal exchange rate flexibility has arguably provided the country with an extra degree of freedom to adjust to shocks. Countries pegging their currency also experience large swings in the real exchange rate, but without this extra degree of freedom to adjust to shocks. As the Argentinean government pegged the peso to the dollar, Argentinean exporters were exposed to large destabilising fluctuations in the peso value of the euro, as well as to fluctuations in the peso value of other Latin American currencies (mainly the Brazilian currency). The same can be said for the Asian countries that were pegging to the dollar or to a basket of currencies, in which the dollar had a substantial weight, at the onset of the crisis of 1997–98. The appreciation of the US currency in that case translated into a loss of competitiveness vis-à-vis Japan and other countries in Asia and Europe. The current dollar depreciation vis-à-vis the euro is likely to raise similar problems for acceding countries by inducing sizeable changes in their terms of trade. 7.4.7 A consistent international financial architecture Fifth, vulnerability has a strong systemic dimension. The risk connected with maturity and currency mismatches in the external debt of emerging markets are understood by international investors, who are nonetheless willing to lend. Both domestic and international guarantees (in the form
9780230_547018_09_cha07.indd 244
3/19/2009 8:15:11 PM
Acceding Countries: The Road to the Euro
245
of international liquidity provision) induce creditors’ moral hazard: lenders have a weak incentive to differentiate among debtors of different quality if they believe that a combination of international bailouts and market timing (early withdrawal) can shield them against losses. The current debate on the reform of the international financial architecture has clarified the policy trade-offs in the bailout of countries. Large bailout packages can substantially reduce or eliminate altogether the large economic costs of capital outflows, costs falling disproportionately on workers and the weakest sector of the economy. A reduction of the costs associated with default, however, does induce moral hazard distortions. Debtors may be less ready to insure against crises: they may hold an insufficient level of international reserves; and most importantly, they may postpone important reforms or avoid the implementation of good policies when these entail political costs. This is “debtor moral hazard.” As mentioned above, international investors may lend while paying insufficient attention to countryrisk, as they expect bailout packages from international institutions to help rescue the country. This is “creditor moral hazard.” As a result of past FDI, EU-based corporations own a large number of financial intermediaries and firms in acceding countries. With financial deepening in these countries, households in other EU countries may be expected to include significant holdings of acceding countries’ liabilities in their portfolios. In the event of a default crisis, this implies that some nontrivial part of its cost will fall on households and firms in other EU countries. Political-economy considerations will strongly influence the balance
Box 7.2
Financial fragility and the sustainability of hard pegs
Gale and Vives (2002) provide a formal analysis of the costs and benefits of adopting hard currency boards (or unilateral euroisation) from a financial stability perspective. Three features characterise countries that stand to gain from a hard currency board and therefore from giving up their monetary policy altogether: these are countries that have (a) a weak institutional structure (including lacking truly independent central banks), (b) serious but not extreme moral hazard problems in the private sector, and (c) moderate liquidation costs of business projects. For these countries, the cost of operating without a lender of last resort (LOLR) is smaller than the benefits from the commitment to fight inflation. Note that low moral hazard implies that risk-taking is not that excessive (the number of firms that would default in the event of macro or sectoral shocks is therefore smaller than otherwise). In the case of default, economic and social costs are contained. A national monetary policy is better for countries with a stronger institutional structure. In their paper, Gale and Vives present an analysis of a sample of countries including Turkey. The authors conclude that Turkey is a candidate for euroisation despite important liquidation costs. Note that concerns about the stability of the banking system are a reason to diversify the choice of roads to EMU across acceding countries.
9780230_547018_09_cha07.indd 245
3/19/2009 8:15:11 PM
246
Designing the European Model
between “bailouts” and “bailins” (that is, how much international investors will lose) in the event of a crisis within the EU. In Chapter 11 we presented evidence of the fiscal costs of debt crises in OECD economies. Past experience shows that these costs can be substantial: the risks for financial stability and the presence of euro-area-wide spillovers are likely to play a key role in the decision process.
7.5 Are acceding countries different from other emerging markets? Recent studies on currency and financial crises pointed out a set of crucial institutional and economic features that characterise emerging markets. A consensus list (as in Mishkin 2003) includes 1. 2. 3. 4. 5.
weak fiscal institutions weak financial institutions low credibility of monetary institutions dollarisation/euroisation of liabilities overall, greater exposure to a sudden end of capital inflows and to liquidity crises.
To what extent are acceding countries different from typical emerging markets? There are at least two notable differences: the ultimate goal of full integration into the EU as well as into the euro area and participation in a multilateral exchange rate agreement/international monetary system. 7.5.1 The ultimate goal of full integration into the EU and into the euro area First, the stabilisation efforts of acceding countries are clearly driven and motivated by the ultimate goal of full integration into both the EU and the euro area. Not only does this end-point constraint create a strong incentive for governments to pursue reforms – in terms of expected benefits of closer ties to Europe. It also provides a clearly defined agenda of institutional and policy reforms, influencing private sector expectations. In the accession process, fiscal, monetary and financial matters are subject to multilateral surveillance. Integration into the European financial markets could, in principle, reduce the exposure of these countries to liquidity crises. A similarly clear end-point constraint cannot be found – at least in an equally strong form – in the experience of other emerging markets. 7.5.2 Participation in a multilateral exchange rate arrangement Second, governments of the acceding countries expect to play some role in the decision-making process of the ECB and other EU institutions. The extent to which this translates into an effective influence is uncertain. But
9780230_547018_09_cha07.indd 246
3/19/2009 8:15:11 PM
Acceding Countries: The Road to the Euro
247
no other emerging market in Latin America and Asia has any formal link with, say, the US Federal Reserve or the Bank of Japan. Specifically, participating in ERM II is different from unilaterally pegging a currency to the euro. Multilateral surveillance and integration into the European institutions strengthen the credibility of domestic monetary authorities. There are explicit mechanisms regulating liquidity provision in case of need. Yet, it is clear that liquidity support from the ECB will not be boundless but will be subordinated to maintaining price stability in the euro area. The two differences above may mitigate the credibility problem stemming from weak fiscal and financial institutions as well as the credibility problem of monetary authorities (points 1, 2 and 3 in the list above). 7.5.3 These differences however do not shield acceding countries from financial turmoil It is, however, unclear how institutions and markets in acceding countries will deal with financial stress if and when it comes. Possible shocks can take the form of higher interest rates with demand growth expansion in Europe, reversal of capital flows driven by domestic or foreign events and strong fluctuations in the terms of trade and commodity prices – think of the implications of strong dollar depreciation. Some imbalances and shocks are likely to originate domestically, during the process of convergence, as relative price and structural adjustments may produce changes in the production structure. As argued above, EU accession is not likely to mitigate the problems raised by currency mismatch in foreign liabilities. If anything, one may expect capital inflows to intensify after EU accession that provides these countries with a “seal of approval” Note that in the political debate, the issuance of euro-denominated liabilities could even be welcomed as a positive step towards full integration into the euro area. Stronger institutional ties with Europe could reinforce expectations of bailouts in the event of a crisis, leading to excessive risk-taking by both local and institutional investors. In Chapter 11, we stressed that the presently ill-defined procedures for dealing with financial crises within the EU create considerable uncertainty about policy responses to a crisis, well beyond what could be desirable in terms of “constructive ambiguity.” EU enlargement strengthens the case for a reform of the EU financial architecture. An often heard argument is that financial crises in any of the acceding countries are not a concern because each of them is economically very small relative to the EU. Thus, the argument goes, the risk of EU contagion is limited, and the EU can easily “afford” the costs of a regional crisis. This argument is not convincing. Even if the effects of a crisis in the richer regions of the EU can be contained, there could be “horizontal” contagion among new members. Acceding countries may be small in terms of GDP but not in terms
9780230_547018_09_cha07.indd 247
3/19/2009 8:15:12 PM
248
Designing the European Model
of population. Financial contagion can create widespread harm, generating political sentiments against European integration. Second, EU countries may have different and possibly conflicting views about the appropriate EU policy in the event of a crisis. This may make the EU response slow, uncoordinated and eventually ineffective (even if large), with adverse effects on the magnitude of the crisis. Eventually, the costs of a crisis for EU firms and institutions may be significant, and their distribution across EU member states is bound to create political conflicts.
7.6
Preparing for the euro
Acceding countries are expected to spend at least two years in ERM II prior to entering EMU. In light of the considerations above, two or more years in ERM II after accession could expose these countries to major currency and financial instabilities. 7.6.1 Acceding countries’ intended strategies In view of the risks associated with regimes of limited exchange rate flexibility, all acceding countries have stated their intention to participate in ERM II for as short a time as possible, that is, no more than the two-year requirement, before entering EMU. They see ERM II as a “waiting room,” with no recognisable merit or contribution to the convergence process. Strategies, however, differ across countries as regards the timing of EMU entry. One group of countries aims at joining ERM II as early as possible after accession. These are the countries that already have hard pegs or currency boards, including Cyprus, Estonia, Latvia and Lithuania. Currency boards and hard pegs to the euro as the reference currency have been declared compatible with participation in ERM II and therefore qualify a country for EMU membership (Governing Council of the ECB, April 13, 2000). Thus, these countries will not change their exchange rate regimes between EU accession and the adoption of the euro – some of them however have to revise the currency basket to which they peg, as to make the euro the only reference currency. For the group of countries that is instead pursuing some form of inflation targeting with flexible exchange rates, participation in ERM II is a clear change to an intermediate regime with only limited exchange rate flexibility before adopting the euro. They will therefore need to undertake a double regime switch, from the current regime to ERM II, and then from this to EMU. In some cases (the Czech Republic, the Slovak Republic and Poland) policymakers have expressed a preference for delaying ERM II participation for some time. This will give them time to achieve some progress in financial development, and, most importantly, to put their fiscal house in order. A relatively slow pace on the road to the euro could be appropriate for
9780230_547018_09_cha07.indd 248
3/19/2009 8:15:12 PM
Acceding Countries: The Road to the Euro
249
countries with mild fiscal problems – to the extent that mild fiscal problems are not priced in too harshly by international investors (i.e. to the extent that country and currency risk premia are not too high in international markets). However, countries with relatively bad fiscal fundamentals may also have a strong incentive to target EMU entry as early as possible, to achieve fast interest rate convergence with the euro area and in this way reduce the government’s interest bill. These countries are primarily worried about being exposed to fluctuations in risk and currency premia. This seems to be the reasoning underlying the strategy of Hungary, a country that is targeting early ERM participation despite apparent fiscal imbalances. 7.6.2 The ECB view In accordance with the spirit and the letter of the Treaty of Maastricht, the ECB sees ERM II as a means to achieve nominal convergence and macroeconomic stability, and ultimately to foster real convergence and growth. More precisely, the ECB sees ERM II as a catalyst, enhancing the discipline of stabilisation policies and domestic policy institutions (ECB 2003). Consistent with this view, there is no reason to limit participation in ERM II to two years only. The risks that we have discussed in this chapter are not inherent in the system: rather, they are inherent in premature participation in ERM II. The modalities to join should be decided on a case-by-case basis, looking at the progress of a country in implementing structural reforms, achieving policy credibility, and implementing stabilisation policies. Moreover, the ECB points out that real convergence may be associated with changes in the equilibrium real exchange rate, which are easier (less costly) to achieve via nominal realignment than domestic price adjustment. Once in ERM II or EMU, the adjustment will necessarily fall on prices. Joining ERM II and EMU at a later stage, after a country will have sufficiently advanced in the process of real convergence, may help reduce macroeconomic costs. The rationale of these different views of nominal convergence has been the focus of an intense debate. Specifically, there are strong concerns about three dimensions of nominal convergence: the ERM as a regime of intermediate exchange rate flexibility, the consistency of inflation and interest rate stability, and the extent of fiscal flexibility. 7.6.3
Exchange rate stability
As regards “exchange rate stability” as a criterion to qualify for EMU membership, an important issue is whether large exchange rate fluctuations within the official bilateral 15 per cent band around central parity would be considered an indicator of “tension” in the exchange market, disqualifying a country from EMU participation. Will the criterion be applied with reference to a much narrower band, say 2.25 per cent, the size of the band in the pre-1993 ERM? According to the ECB (2003), “the assessment of exchange
9780230_547018_09_cha07.indd 249
3/19/2009 8:15:12 PM
250 Designing the European Model
rate stability against the euro will focus on the exchange rate being close to the central rate.” This issue may be a minor concern for the set of countries adopting currency boards and hard pegs. It is, however, crucial for countries currently using some form of inflation targeting and therefore in need of undertaking a regime switch prior to entering the euro area. In light of the recent financial history summarised in this chapter, the ERM with narrow bands is the kind of intermediate fixed exchange rate regime that invites speculation and makes countries vulnerable to severe liquidity shocks. If the convergence criterion requires countries de facto adopting narrow bands, acceding countries will be forced to take unnecessary and useless risks. Several documents of EU institutions and the Eurosystem seem to define “exchange rate stability” as an asymmetric criterion, that is, compatible with appreciation but not with depreciation. But acceding countries are recommended to set their initial parity in ERM II according to their best guess of the currency’s fundamental value, based on a broad range of indicators, including market prices, rather than “playing games” with an eye on the final euro conversion rate. It is apparent that such a recommendation is not consistent with an asymmetric definition of exchange rate stability. Clearly, such a definition provides a strong incentive for acceding countries to choose a relatively weak central parity at the beginning of their participation in ERM II and let their currency appreciate over time. It is not surprising to see acceding countries’ governments strongly arguing that their currencies are overvalued (while perhaps intervening heavily to prevent appreciation). We should note here that uncertainty about the final euro conversion rate could actually damage the country as well as the stability of ERM II by creating a co-ordination problem among market participants: with which final conversion rate would markets co-ordinate their expectations?4 The benefits from a clear end-point exit from exchange rate based stabilisation would in part be eroded. 7.6.4 The (unfeasible) option of immediate euroisation Immediate euroisation, even in the form of unilateral adoption of the euro, would eliminate exchange rate risk and solve the problems raised by currency mismatches in the country balance sheets when external debt is denominated in a foreign currency. It would therefore close an important channel through which self-fulfilling prophecies in the exchange market and exchange rate crises can have devastating effects on the economy. Among the policy trade-offs of immediate euroisation, an important one concerns relative prices. An excessively appreciated (or depreciated) initial conversion rate between the domestic currency and the euro could create large and protracted real costs in terms of employment, investment and growth. But supporters of euroisation see the exchange rate as a potentially destabilising price (in the event of a crisis).
9780230_547018_09_cha07.indd 250
3/19/2009 8:15:12 PM
Acceding Countries: The Road to the Euro
251
A crucial dimension of this trade-off is the extent to which the elimination of exchange rate risk raises the default risk in the economy. With weak financial institutions and markets, immediate euroisation will possibly exacerbate moral hazard problems leading to excessive risk-taking. An ill-defined financial architecture for the euro area as a whole may magnify the distortion, offsetting the benefits of sheltering balance sheets from valuation shocks due to exchange rate movements. Moreover, in the political economy of EU accession, participating in EMU is a well-understood ultimate goal that can motivate reforms and good policy making in the third phase of the accession process. Immediate euroisation will substantially reduce the leverage of current EU members on acceding countries, as well as of domestic governments on domestic and international interest groups. As discussed in Chapter 1 of the 2004 EEAG report (Appendix 4 on the Past and Future of the Stability and Growth Pact), the goal of entering EMU can motivate large fiscal consolidation efforts, but common fiscal rules become much less binding once a country is in EMU. The European Council in Nice, however, excluded euroisation from the set of relevant policy options open to acceding countries on the ground that it would be inconsistent with the view underlying EMU as the end-point of a convergence process, adopted by the Treaty of Maastricht (Council of the European Union Press Release No. 13055/00; see also European Central Bank 2003). Euroisation is seen as a way to circumvent the convergence process. 7.6.5 Currency boards Would an early adoption of a currency board provide a good substitute for early euroisation to address the issue raised by the currency denomination of foreign debt and the escape clauses implicit in intermediate regimes of fixed exchange rates? There are strong reasons to be sceptical. First, markets may still attach some positive probability to devaluation. Second, (as in the case of unilateral euroisation) the ECB will not be required to act as de facto lender of last resort (although it may choose to do so if there is no danger for its price stability objective). Third, we have seen that many countries are suffering a deterioration of their fiscal state. Large deficits are not compatible with adopting a currency board. As experience shows, a currency board is not per se effective in forcing convergence of the country risk premium: interest rates may not fall at all if markets are not absolutely convinced about the sustainability of the fiscal and financial systems. These are among the reasons why a successful currency board requires a country to meet strict fiscal and macroeconomic conditions. If readiness for the euro is to be judged in terms of a country’s ability to sustain a currency board, the accession to the euro will be a longer and more risky process than most acceding countries would hope for.
9780230_547018_09_cha07.indd 251
3/19/2009 8:15:12 PM
252
Designing the European Model
7.6.6 Inflation versus exchange rate stability The second dimension of the convergence process under scrutiny concerns inflation. Given that acceding countries are growing at fast rates, and the price level tends to increase with income, is there a conflict between exchange rate stability and price stability? Many observers trace a possible conflict between exchange rate and price stability to the Balassa-Samuelson theory. The well known argument is as follows. Fixing the exchange rate pins down the price of tradables in domestic currency. As gains in productivity in the tradable sector cannot translate into lower prices, they translate into higher wages that, with sufficient labour mobility, will spread across industries in the non-tradable sector. But these sectors experience much lower productivity growth: higher wages can only be paid if the price of non-tradable output goes up. A high rate of price increase for nontradable goods (which is an equilibrium relative price adjustment) may raise observed CPI inflation above the convergence criterion. Suppose the Balassa-Samuelson effect was indeed the main determinant of inflation and real exchange rate appreciation in acceding countries. If monetary policy targets some low inflation rate, such policy would result in some moderate appreciation of the exchange rate. According to the BalassaSamuelson theory, an appreciating exchange rate would translate into a fall in the domestic price of tradables relative to non-tradables, a fall that is completely offset by gains in productivity. Thus, choosing an inflation target would not violate the exchange rate stability criterion (as the exchange rate would appreciate) and at the same time would produce a fall in the price of tradables in domestic currency (not to be confused with deflation). Conversely, if a country pegs the exchange rate, there would be some inflation differentials during the income convergence process. Available empirical studies produce a wide array of estimates of the size of inflation differentials attributable to Balassa-Samuelson effects. While most studies predict small differentials, unlikely to cause violation of the Maastricht inflation criterion, there are also much higher estimates. Critics of the convergence criteria point out that, in the presence of nominal rigidities in the economy, a binding inflation criterion would only produce unnecessary harm, as it would confuse equilibrium adjustment of relative prices with a general increase in the price level.5 Thus, according to these critics, the inflation criteria should be made more flexible for countries that choose hard pegs or currency boards. To sum up: during real convergence, sticking to a low inflation target would imply a moderate rate of exchange rate appreciation (which has been declared consistent with the exchange rate convergence criterion); adopting a hard peg would imply a rate of inflation which could violate the inflation convergence criterion. In principle, this criterion should be relaxed. Are there specific reasons to prefer one regime over the other?
9780230_547018_09_cha07.indd 252
3/19/2009 8:15:12 PM
Acceding Countries: The Road to the Euro
253
7.6.7 Inflation differentials depend on much more than BalassaSamuelson effects Even in the case that, at some stage of the convergence process, inflation differentials could be entirely attributed to desirable relative price adjustment, relaxing the inflation convergence criterion is risky. This is because high inflation rates could feed agents’ expectations of further price dynamics. The potential problems raised by such expectations include overvaluation of the real exchange rate and short-term excessive demand expansion. This argument stresses that in reality inflation and real appreciation during the convergence process reflect much more than the Balassa-Samuelson effect. There are also (a) changes in the terms of trade; (b) changes in the size of deviations from the law of one price for tradable goods; and (c) measurement errors that may be sizeable in acceding countries due to rapidly changing baskets of consumption goods and the structure of production. In addition there could be (d) inflationary effects of domestic overheating caused by capital inflows and private sector expectations. There is insufficient understanding of the relative importance of these elements in practice. Figure 7.3 plots the ratio between the producer price index (PPP) and the consumer price index (CPI) in the different accession countries. It is reasonable to expect that the PPI includes a larger share of tradable goods than the CPI, as services have a larger share in the latter. In that case, a strong Balassa-Samuelson effect would imply a fall in the PPP relative to the CPI, since a relative increase in the price of non-tradables would raise the CPI more than the PPI. The figure shows that a pattern consistent with the Balassa-Samuelson hypothesis can be detected for a few, but not all acceding countries. There are large differences in real exchange rate patterns. Overall, these considerations suggest caution in relaxing the inflation criterion as a precondition to participation in EMU. Overall, a regime of (some degree of) exchange rate flexibility coupled with low inflation dynamics seem to provide a safer path to joining EMU. 7.6.8 The need for fiscal discipline and the Stability and Growth Pact As regards the final concern with convergence criteria, some observers claim that the Stability and Growth Pact (SGP) is likely to become a straitjacket for fast-growing countries at low levels of income, in need of building their own infrastructure and implementing many reforms that could require temporary government spending. Given the political and institutional crisis concerning the SGP, critical views are likely to gain strength over time. It is quite obvious that fast-growing countries are in need of large public investment, and that there are reasons why it is neither optimal not equitable to finance infrastructure with current taxes. By the same token, institutional and structural reforms may create temporary spending or revenue
9780230_547018_09_cha07.indd 253
3/19/2009 8:15:12 PM
254 Designing the European Model
shortfalls, which would be optimally financed by borrowing. Yet one needs to be aware of the fact that local governments have a strong incentive to use fiscal flexibility for other purposes than the accommodation of a high rate of public investment or the facilitation of reforms (see the box on the SGP in Chapter 1 of the 2004 EEAG report). At times when governments are making an effort to consolidate their budgets and find viable fiscal paths, a relaxation of the fiscal rules may compromise the consolidation process, with negative effects on expectations affecting interest rates on the public debt. Our proposal for an improved Stability and Growth Pact (summarised in Chapter 1 of the 2004 EEAG report and extensively discussed in Chapter 2 of the 2003 EEAG report) can, however, suggest ways to grant some degree of flexibility in the design of fiscal policies. 7.6.9 The enlarged euro area Many of the problems discussed in this chapter will still persist at the end of the third phase of EU accession, when the euro area will consist of at least 22 countries. Enlargement obviously exacerbates issues regarding efficient monetary policy decision-making in the euro area: a very large decisionmaking body cannot work well. These issues will call for a much needed structural reform, which we do not see as a major issue, however. After all, monetary policy in the euro area is not a weighted average of national policies.
% 1.10 Cyprus 1.05 Euro area 1.00 Czech Republic Latvia Poland Estonia
0.95 Slovak Republic
Slovenia
0.90
Hungary
0.85 1999 Figure 7.3
2000
2001
2002
2003
2004
Ratio of tradables-to-non-tradables prices
Note: The figure shows the ratio between producer price index (PPP) and the consumer price index (CPI). Source: IMF, International Financial Statistics.
9780230_547018_09_cha07.indd 254
3/19/2009 8:15:12 PM
Acceding Countries: The Road to the Euro
255
Yet, enlargement raises new policy issues. In an enlarged euro area, the acceding countries will be small in terms of economic weight but not in terms of population. A small economic size means that acceding countries will not have much weight in the design of stabilisation policies. Even if inflation were tending to be higher in the new countries than in the current euro area, their influence on interest rate setting by the ECB would be negligible. First, their weight in the euro area Harmonized Index of Consumer Prices is very small. Second, a significant part of any inflation differentials is likely to be attributed to desirable relative price adjustments. A large population share in the EU, however, implies that the importance of acceding countries may become much larger in the event of financial turmoil and crises. Markets may expect political considerations to guide EU crisis management and resolution. Even if the ECB were to be able to pursue a consistent monetary policy through periods of turmoil, crises would be formidable challenges to national fiscal authorities and financial supervisors/regulators. As argued in chapter 4 of the 2003 EEAG report, defining procedures and intervention policies in the event of crisis at the euro area level may be extremely helpful in this respect.
7.7 Summary Overall, there is no single strategy that could be recommended to all acceding countries as regards macroeconomic stabilisation on the road to the euro. Arguments in favour of adopting the euro as early as possible include smaller financial risk due to the elimination of currency mismatch in the balance sheet of banks and firms (which implies the risk of a self-fulfilling run on the country debt); interest rate convergence (with the associated gains in terms of the interest bill for the government as well as investment financing by firms); and overall gains in monetary credibility. Arguments for a slower pace toward the euro stress the need to remove financial distortions creating moral hazard and therefore undermining the stability of the domestic financial sector and raising the country’s default risk; the advantage of relative price adjustments without the need of costly nominal wage and price adjustments; and the need to make fiscal and financial policy sustainable and compatible with a fixed exchange rate before participation in the EMU. At the end of 2003, some countries reiterated their willingness to enter ERM II on or shortly after accession: Hungary, Estonia, Cyprus, Lithuania in 2004, and Latvia and Slovenia in 2005. In this group of countries, Hungary has the most fragile fiscal situation. The other countries have chosen to delay their entry into ERM II, perhaps waiting to see whether the interpretation of the convergence criteria will be adapted in light of their arguments against the merits of ERM II.
9780230_547018_09_cha07.indd 255
3/19/2009 8:15:12 PM
256 Designing the European Model
Countries that are already able to sustain hard pegs should be helped to achieve a smooth and fast transition to the euro. In this set of countries, mainly small ones, priority should be given to institutional reforms and to building a policy framework consistent with participation in the euro area without suffering from major macroeconomic imbalance. Delaying participation in ERM II is a realistic option for countries that are currently unable to sustain hard pegs and have large domestic imbalances. The magnitude of domestic imbalances varies considerably across countries, so that ERM entry may be desirable at different times. Yet in all cases, the policy priority is achieving a sustainable fiscal situation and stabilising inflation at the correct relative prices, a task that requires both institutional and policy reforms. For both groups of countries, the convergence criteria in terms of inflation, interest rates, debt and deficit provide desirable targets to guide policy and should not be relaxed. Though they are not first-best targets, these convergence criteria should be judged relative to existing distortions that could derail the stabilisation efforts. The evidence reviewed in chapter 5 of the 2004 EEAG report shows that on average acceding countries are doing well as regards the two criteria of inflation and long-term interest rates. Once in the EU, it is even possible that the three countries with the lowest inflation rates will include acceding countries. The main issue is fiscal convergence (which of course may undermine the sustainability of the inflation and interest rate performance). As regards exchange rate stability, ERM II allows for large fluctuation bands around exchange rate parity. Once in ERM, a country should be able to use the exchange rate flexibility implied by such an arrangement, in the sense that exchange rate stability should not be mechanically assessed with reference to much narrower bands. Fluctuations in the exchange rate in response to domestic and foreign shocks are not necessarily indicators of tension in the exchange market but can be part of an efficient adjustment process. If the dollar continues to depreciate, it may be reasonable to expect exchange rate fluctuations within ERM II. Declaring that acceding countries will be accepted in the euro area only if they can peg to the euro within narrow bands may raise the possibility of speculative attacks driven by self-fulfilling prophecies. During the transition to the euro, strict domestic stabilisation with some exchange rate flexibility is better than exchangerate-based stabilisation with very limited flexibility. In practice, however, exchange rate flexibility will not be enough to shelter a country from financial turmoil. The risk of crisis is somewhat reduced, but not eliminated, by delaying participation in ERM II, or by making full use of the 15 per cent bilateral bands once a country is part of the ERM system. “Convergence play” and currency mismatches can still characterise the transition to the euro.
9780230_547018_09_cha07.indd 256
3/19/2009 8:15:12 PM
Acceding Countries: The Road to the Euro
257
A dangerous possibility during the transition is that markets do not learn to appreciate countries’ specific features and assess country risk based on domestic policy and real fundamentals. It would be extremely frustrating if the policy effort of one country were to be discounted in episodes of turmoil, whereas markets extend to all acceding countries the adverse assessment of a subset of them. The 1992–93 experience of the ERM as a multilateral system clearly shows that financial contagion is possible even in advanced countries (see Buiter, Corsetti and Pesenti 1998). The weakness of the old ERM is present, in magnified form, in ERM II. The experience with the ERM also points out that there is no stability without a consistent currency and financial policy framework for Europe as a whole.
Notes 1. The traditional theory of optimal currency areas (OCA) suggests a set of criteria to judge the costs of giving up exchange rate flexibility. These include the degree of price and wage flexibility, the extent of factor (labour) mobility, insurance via fiscal transfers or fiscal integration in general, and correlation of macroeconomic non-financial shocks. This last item can be attributed to several structural features of the macro economy: openness and economic size, degree of goods market integration, composition of production and trade specialization. Note that, per se, correlation of business cycles is not an OCA criterion, as a high correlation may not result from symmetric shocks but from symmetric policy responses to asymmetric shocks. Actually, national business cycles in Europe may well become more correlated after the creation of EMU, without implying that the cost of giving up exchange rate flexibility has fallen at all (see Corsetti and Pesenti 2002). The literature applying these criteria to judge whether acceding countries are ready for EMU is extremely vast, but its results are quite inconclusive (at least as inconclusive as the results of the literature on the same subject applied to EMU creation). The specific problem with the OCA approach applied to acceding countries is how to account for structural changes currently under way in the convergence process. We believe that these changes are exceedingly difficult to assess, and therefore provide shaky ground for empirical exercises. 2. The recent default crisis in Argentina affected investors’ attitude towards lending to sovereign states and private firms in emerging markets. To the extent that investors will be reluctant to engage in “convergence play”, the crisis in Argentina may turn out to have some beneficial implications for acceding countries. 3. Aside from the volatility aspect, there are also doubts that FDI is the most efficient way to channel capital to emerging economies, as with FDI the benefits from financial liberalisation accrue only to very few firms, mostly (but not exclusively) in the tradable sector. Conversely, bank flows – intermediated by the domestic banking system – are in practice the only source of external financing for firms in the non-tradable sector and/or small firms. Heavy reliance on stable FDI flows can easily lead to bottlenecks and strongly imbalanced growth. Moreover, in many cases FDI is mainly motivated by tax-saving schemes adopted by multinationals. 4. ERM realignments raise a number of well-known policy issues as regards their timing, size and guidance of market expectations (see Buiter, Corsetti and Pesenti 1998 for a discussion).
9780230_547018_09_cha07.indd 257
3/19/2009 8:15:12 PM
258
Designing the European Model
5. The problem could be mitigated by calculating the reference value for inflation using the euro area inflation rate, rather than an average for three countries with the lowest inflation in the EU.
References Begg, D., B. Eichengreen, L. Halpern, J. von Hagen and C. Wyplosz (2003) “Sustainable Regimes of Capital Movements in Accession Countries,” CEPR Policy Paper no. 10. Buiter, W., G. Corsetti. and P. Pesenti (1998) Financial Markets and European Monetary Cooperation, Cambridge University Press, Cambridge, UK. Buiter, W. and C. Grafe (2002) “Anchor, Float or Abandon Ship: Exchange Rate Regimes for the Accession Countries,” Banca Nazionale del Lavoro Quarterly Review, 221, June, 1–32. Calvo, G. (2003) “The Mirage of Exchange Rate Regimes for Emerging Market Countries,” NBER Working Paper no. 9808. Corsetti, G., P. Pesenti and N. Roubini (1999) “Paper Tigers: A Model of the Asian Crisis,” European Economic Review 43, 1211–36. Corsetti, G. and P. Pesenti (2002) “Endogenous Optimum Currency Area,” NBER Working Paper no. 8783. European Central Bank (2002) Review of the International Role of the Euro, December, Frankfurt. European Central Bank (2003) “Policy Position of the Governing Council of the European Central Bank on Exchange Rate Issues Relating to the Acceding Countries,” http://www.ecb.int (accessed on 18 December 2003). European Commission (2003) “Key Structural Challenges in the Acceding Countries: The Integration of the Acceding Countries into the Community’s Economic Policy Co-Ordination Process,” Occasional Paper, Directorate-General for Economic and Financial Affairs. European Economic Advisory Group at CESifo (2003) Report on the European Economy 2003, CESifo, Munich. Gale, D. and X. Vives (2002) “Dollarization, Bailouts, and the Stability of the Banking System,” Quarterly Journal of Economics 117, 467–502. Guimaraes, B. and S. Morris (2003) “Risk and Wealth in a Model of Self-fulfilling Currency Crises,” Cowles Foundation Discussion Paper no. 1433. IMF (2003) International Financial Statistics, Washington DC. Mishkin, F. (2003) Comments on Fraga, Goldfajn and Minella, “Inflation Targeting in Emerging Market Economy,” http://www.nber.org/books/macro18/mishkin814-03comment.pdf. OECD (1999) Bank Profitability: Financial Statements of Banks, database: http:// www.sourceoecd.org. Tornell, A. and F. Westermann (2002) “Boom-Bust Cycles: Facts and Explanation,” IMF Staff Papers 49 (special issue). Tornell, A., F. Westermann and L. Martinez (2003) “Liberalization, Growth and Financial Crisis: Lessons from Mexico and the Developing World,” Brookings Papers on Economic Activity, 34(2), 1–112. Von Hagen, J. and J. Zhou (2002) “De Facto and Official Exchange Rate Regimes in Transition Economies,” Center for European Integration Studies Working Paper B 13.
9780230_547018_09_cha07.indd 258
3/19/2009 8:15:12 PM
8 Economic Growth in the European Union
8.1
Introduction
Sluggish economic growth in many EU countries has been a major concern in Europe in the past ten to fifteen years. In the post-war period up to the 1990s European countries appeared to be catching up with the United States as the gap between GDP per capita in the US and West European countries gradually narrowed. This tendency was dramatically reversed in the 1990s. The catching-up process appears to have come to an end and several EU countries, in particular France, Germany and Italy, have started to fall further behind the US.1 The European growth problems have led to major political discussions within the EU and achievement of fast economic growth has become a key policy objective. A notable expression of the concern for growth was the March 2000 meeting of 15 EU leaders that was held in Lisbon. The agenda set in Lisbon is very clear in its emphasis on economic growth: by 2010 the EU should become “the most dynamic and competitive knowledge-based economy in the world capable of sustainable economic growth with more and better jobs and greater social cohesion, and respect for the environment.”2 The disappointing growth performance in several EU countries has, however, not been a universal phenomenon. Some EU countries – notably Ireland, Finland, Greece, UK, Spain and Sweden – have performed well in the last ten years. In addition, we are beginning to see “growth miracles” in several new EU member countries, though the short time-span since the start of the EU membership negotiations limits the possible conclusions about economic growth in the new member states. The striking differences in the growth experiences in the EU over the past decade are the motivation for focusing several chapters of the 2006 EEAG report on the topics that are important for growth and competitiveness of the EU. The purpose of this chapter is to set the stage for the subsequent chapters by providing an overview of economic growth that is at first hand descriptive but also provides insights into growth processes that are important for 259
9780230_547018_10_cha08.indd 259
3/19/2009 8:16:25 PM
260 Designing the European Model
policy relevant conclusions. More specifically, we will look at the following questions. 1. How fast is the current speed of convergence in per capita incomes in the EU, particularly in Eastern European countries? 2. How does Europe as a whole compare to the United States? 3. What are the main factors behind the different growth performances of the most successful and the most unsuccessful EU countries? As an answer to the first question, we show that convergence in per capita incomes in the EU is indeed taking place, although it is largely driven by the convergence between the old EU-15 and the new member countries. The annual convergence rate among the EU-25 is – depending on the measurement technique – between 1.7 and 1.9 per cent a year, while it is only 0.9 per cent among the old member countries. 3 All three figures fall somewhat behind the results typically reported for a larger set of OECD countries and over longer time periods. We then analyse the key determinants of economic growth by looking at growth accounting computations that for each country decompose GDP growth into the contributions of labour input, non-IT capital input, IT capital input and total factor productivity. To highlight the big differences between EU countries, we separately investigate the successful cases of Ireland, Greece, Finland, Spain, Sweden and the UK, and the countries that have since the 1990s experienced substantial growth problems, that is Germany, Italy and France. As a benchmark, we compare the EU with the US. While there exist, of course, many country-specific factors, one can argue that as a whole the unsuccessful cases have been growing mostly through traditional capital accumulation and somewhat through general technological progress (total factor productivity growth). Labour input, measured by total hours worked, plays a substantial negative role (except for Italy). Particularly in Germany, the decline in labour input has made a sizeable negative contribution to economic growth. There have been different roads to prosperity in the more successful countries. In Ireland, Finland, the UK and Sweden, there has been, since the mid 1990s, a large increase in the contribution by IT capital growth. However, it is remarkable that all production factors have made a positive contribution in these countries, including labour input for most episodes. In these countries, relatively rapid IT capital growth seems to have been coupled with relatively high total factor productivity (TFP) growth. On the other hand, two other success cases, Spain and Greece, have grown fast primarily due to more traditional factors, that is capital accumulation and labour input growth. Data limitations preclude a correspondingly detailed analysis of the new EU member countries and we consider their recent performance only in terms of crude indicators for sources of economic growth.
9780230_547018_10_cha08.indd 260
3/19/2009 8:16:26 PM
Economic Growth in the European Union 261
After the comparison between successful and laggard cases, we analyse further the performance of the EU countries in terms of the sources of economic growth. We emphasise a number of policy-relevant factors that can influence the rate of general technological progress. The concluding section draws together the different results and makes suggestions for re-orientation of the EU policies to improve economic growth.
8.2
Overview of growth in the EU
We start our analysis of economic growth in the EU-25 countries by looking at the paths of per capita income. Because comparisons of absolute levels of GDP per capita among countries are difficult due to differences in price levels, we normalise the value of per capita income in 1995 to an index number 100 and show how GDP per capita has evolved in each country.4 Moreover, the data are not adjusted for purchasing power parity (PPP), because PPP-adjusted data are available only with a lag. Figure 8.1a confirms that there are remarkable differences among the EU-15 countries in economic growth in the period 1995–2004. On the one hand, we have the Irish miracle, with Finland being a clear second. Economic growth in the UK, Greece, Spain and Sweden has also been fairly rapid. On the other hand, the large economies of Germany, France and Italy are the worst performers in terms of GDP growth among the EU-15 countries. However, country size per se cannot be an explanation for these 170 Ireland
160 150 140
United Kingdom Finland
Portugal Netherlands
130
Greece Spain Sweden Austria Belgium France Italy Germany
120 110 Denmark 100 1995 Figure 8.1a
1996
1997
1998
1999
2000
2001
2002
2003
2004
Index of real GDP per capita (EU15 countries 1995 = 100)
Source: Eurostat, calculations by the EEAG.
9780230_547018_10_cha08.indd 261
3/19/2009 8:16:26 PM
262
Designing the European Model
differences, as the UK and Spain were among the best performers and, among the smaller countries, GDP in Belgium and in Denmark grew rather slowly in the ten-year period starting from 1995. Denmark has been hailed as an example of a country that solved its unemployment problem, but it does not stand out as a good model for economic growth. There are also major differences in the growth performances of the new member countries, though on average these countries do grow faster than the old EU countries. Figure 8.1b shows that the Baltic countries, Estonia, Latvia and Lithuania, have had the best performance, while Malta and Cyprus have had the slowest growth. It should be noted that the starting levels of the latter have been higher. If we exclude Malta and Cyprus, then the Czech Republic has had the slowest growth among the new EU members. Figure 8.2, showing the cumulative growth in GDP per capita in 1995– 2004, puts all the EU-25 countries in a single diagram. The figure shows that the new EU members have mostly done well. They all start, of course, from low levels of living standards as compared to EU-15, a consideration that will be more closely investigated in the section on catching-up and convergence. To conclude this section, we compare the development of gross national versus gross domestic product. GNP is arguably a better measure of living standards of a country, as it takes into account incomes earned by factors of
190
Latvia Estonia
180
Lithuania
170 160
140
Poland Slovenia Hungary Slovakia
130
Czech Republic
120
Malta Cyprus
150
110 100 1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
Figure 8.1b Index of real GDP per capita (new EU countries 1995 = 100) Source: Eurostat, calculations by the EEAG Group.
9780230_547018_10_cha08.indd 262
3/19/2009 8:16:26 PM
Economic Growth in the European Union 263
production owned by the country. Figure 8.3 presents the annual average difference between growth in GDP and GNP for 1995–2004. This difference could be important especially for the countries that have invested abroad or whose residents work abroad. Figure 8.3 shows that GNP growth was indeed somewhat higher than GDP growth for the successful “high-tech” EU countries Finland, Sweden and the UK. In contrast, GDP and GNP growth was almost the same in France and Germany. Italy performed slightly better in terms of GNP than GDP growth. It is also seen that GDP growth for the Netherlands, Ireland, Spain and Belgium overstates the growth in living standards. Persistent movements in the terms of trade are another factor that can affect comparisons of different concepts of domestic product versus income. In the last fifteen years the prices of IT goods have fallen rapidly, which has reduced the benefits from fast productivity growth in the IT sector. Thus, real income growth in countries that have relied on IT sector exports has been slower than it would have been without adverse price developments. This effect can be non-trivial: for example, in 1998–2004 Swedish GDP at constant prices rose at an average annual rate of 2.8 per cent, whereas the rate of growth of its nominal GDP deflated by the price index for domestic absorption was only 2.2 per cent.5 in % Latvia Estonia Lithuania Ireland Poland Slovenia Hungary Slovakia Finland Greece Czech Republic Spain Sweden United Kingdom Malta Cyprus Austria Belgium Denmark Portugal Netherlands France Italy Germany 0
Figure 8.2
10
20
30
40
50
60
70
80
90
Cumulative real growth per capita, 1995–2004
Source: Eurostat, calculations by the EBAG Group.
9780230_547018_10_cha08.indd 263
3/19/2009 8:16:26 PM
264 Designing the European Model %
%
0.12
0.12
0.1
0.1
0.08
0.08
0.06
0.06
0.04
0.04
0.02
0.02
0
0
en Fi n ni la te nd d Ki ng do m
ly
U
Ita
Sw ed
k m
ar
an y
en D
m
ce
tri a
er G
Au s
la n
Fr an
Be
Ire
Sp
nd rla N
et
he
d
–0.06
lg iu m
–0.04
–0.06
ai n
–0.02
–0.04
s
–0.02
Figure 8.3 Difference between GDP and GNP growth from 1995 to 2004 (Average difference in aunualised growth in percentage) Note: For Ireland, where GNP data are only available from l997, the cumulative difference from 1997 to 2004 is shown. Source: Eurostat, calculations by the EEAG Group.
8.3 Convergence in economic growth A central question is whether economic growth in poorer countries is on average faster than in richer countries. If this is the case, it is said that there is convergence among countries (in levels of GDP per capita), which in turn is an indication that living standards tend to be equalised in the long run. The main reason for the convergence hypothesis is that the technologically most advanced countries are dependent on the development of new technologies, which is both a time- and resource-consuming activity, whereas technological followers can rely on imitation and technology diffusion to achieve technological progress with lower resource costs.6 In this section, we investigate the convergence properties of the EU-25 countries. The possibility of convergence can be examined by using different indicators. We look at some well-known concepts of convergence, such as the notions that countries with lower initial per capita incomes have higher growth rates on average (called absolute or beta convergence) and that there occurs a reduction in the dispersion of income levels across countries over time (called sigma convergence). Overall, we find that significant beta and sigma convergence is taking place in the EU, although the rate of convergence is somewhat slower than typically reported for the total OECD. As a first step, we perform a standard statistical regression analysis that tests for absolute (beta) convergence by regressing the growth rates over the
9780230_547018_10_cha08.indd 264
3/19/2009 8:16:27 PM
Economic Growth in the European Union 265 9.0
Latvia
8.0 7.0
Estonia Lithuania
Average growth
Ireland 6.0
Poland
5.0
Slovenia
Hungary
Slovakia
4.0 3.0
Czech Republic 2.0
Malta Portugal
1.0 0.0 7.0
7.5
8.0
United Kingdom
Greece
8.5
Spain
Finland Sweden Austria Denmark
Cyprus Belgium France Italy NL Germany
9.0
9.5
10.0
10.5
Log initial per capita income
Figure 8.4
Beta-convergence among EU25 countries
Note: The coefficience on the regression line it –1.82 (P-value: 0.00). The convergence rate that follows this coefficient, adjusted for sample size, is 1.93 per cent a year (Among the EUl5 countries only, the convergence rate is 0.86 per cent a year). For computations in finite samples, see Barro and Sala-i-Martin (2003), footnote 4, p. 422. Source: Eurostat, calculations by the EEAG Group.
last ten years on the initial (logarithmic) levels of per capita income in 1995 and an intercept. Figure 8.4 demonstrates that there indeed exists absolute (beta) convergence within EU-25. However, the speeds of convergence differ among the EU-25 and the old EU-15 countries. While in the EU-25, the countries are converging at an annual rate of 1.9 per cent a year, the convergence rate among the EU-15 countries is only 0.9 per cent a year. This suggests that convergence in the EU is mostly driven by the catching-up process of the Eastern European countries to the per capita income levels of Western Europe. An alternative, non-regression-based method to measure convergence is to examine the development of the dispersion of per capita incomes across countries over time (sigma convergence). This can be done by computing the standard deviation of the per-capita income distribution in the EU-25. Figure 8.5 shows the development of income dispersion over time and confirms the preceding evidence, as we see a clear tendency towards convergence within Europe. Dispersion has declined steadily, except between 1998 and 1999. The convergence rates that can be computed from this diagram are somewhat lower than the previous results from beta convergence. They indicate that the income dispersion is declining at a rate of 1.7 per cent a year.7
9780230_547018_10_cha08.indd 265
3/19/2009 8:16:27 PM
266
Designing the European Model
0.96
0.96
0.94
0.94
0.92
0.92
0.90
0.90
0.88
0.88
0.86
0.86
0.84
0.84
0.82
0.82 0.80
0.80 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 Figure 8.5
Income dispersion (sigma convergence)
Note: The figure shows the standard deviation of log per capita income in the EU25 countries in 1995–2004. Convergence rate 1.73 per cent (the dispersion of per capita income is reduced by 1.73 per cent a year). Source: EEAG Group, 2006.
Our findings suggest that convergence with the EU is taking place, though at somewhat lower rates than what has been found in other studies for a larger set of countries. Using both definitions of convergence, Barro and Sala-iMartin (2003) computed that the convergence rates are between 2 and 3 per cent among OECD countries as well as among US states. Figure 8.5, showing the declining dispersion of incomes per capita in the EU in terms of the standard deviation, does not take into account possible asymmetries. For distributions that are highly skewed, particularly flat or peaked, it is necessary to look at the higher moments, or better still at a histogram of the distribution of incomes per capita. The latter can be described by what is known as the kernel (smoothed density) of the distribution. Figure 8.6 shows the results of kernel density estimations for the distribution of per capita incomes in 1995 and 2004.8 From these graphs we can see that the distribution of income levels across EU countries has become less asymmetric over time. While in 1995, the income distribution was clearly skewed to the left and almost took the shape of a double-bell shaped function, its counterpart in 2004 indicates a more symmetric distribution around the mean. This again suggests that the convergence process is one of “East” converging to “West.” From this graph it is reasonable to expect that in the future there will emerge a clear single-peaked distribution of incomes per capita in the EU countries even before full convergence of the Eastern European countries has been achieved.
9780230_547018_10_cha08.indd 266
3/19/2009 8:16:27 PM
Economic Growth in the European Union 267 a) Log levels in 1995
b) Log levels in 2004 Kernel Density (Epanechnikov, h = 1.0000)
Kernel Density (Epanechnikov, h = 1.0000) 0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1 0.0
0.0 6
7
8
9
10
11
12
6
7
1995
Figure 8.6
8
9
10
11
12
2004
Comparison of income distribution across the EU25
Source: Eurostat, calculations by the EEAG Group.
Our findings can be summarised by noting that the growth process within the EU exhibits clear catching-up of the poorer countries, which are largely the new members from East Central Europe, towards the Western “old” EU countries. This convergence is very gradual and somewhat slower than what appears to hold for the OECD area as a whole. The results show that the major growth policy concerns in the EU should not be the differences between the “old” and “new” EU countries, but rather the sluggish performance of some key Western EU countries. Our descriptive analysis has revealed large differences between the EU-15 countries. It is instructive to evaluate the reasons behind these differences. We next examine growth in the old EU countries more closely.
8.4 Factors behind growth performance – a growth-accounting perspective In this section, we look more deeply into the sources of economic growth by performing a growth accounting analysis for selected countries. Our aim is to uncover the differences between the EU countries that have been, respectively, successful and unsuccessful in their growth performance. However, before examining the successful as well as the unsuccessful growth cases in the EU, we look at the US, which is the natural benchmark for growth comparisons. Quite appropriately, the US growth performance has been used as the reference point in policy discussions in Europe. In general, growth accounting tries to uncover the sources of economic growth by considering the production side of the economy, so that growth of aggregate output is decomposed into contributions from growth in factor
9780230_547018_10_cha08.indd 267
3/19/2009 8:16:28 PM
268
Designing the European Model
inputs (capital, labour and other factors) and from general technological change. This approach can be used in a flexible way depending on the availability of data on inputs of productive factors. It is not possible to measure technological change directly, so its effects are shown by the residual in the growth-accounting decomposition. Our analysis uses data provided by the Groningen Growth and Development Centre9 (see in particular Timmer, Ypma and van Ark 2003, who emphasise the role of information technology (IT) in economic growth). In our computation, overall GDP growth is decomposed into contributions from the growth of labour input, non-IT capital input, IT capital input and total factor productivity (TFP). TFP is a measure of general technological progress. The decompositions are based on the equation: ΔlnY = vLΔlnL + vKnΔlnKn + vKit ΔlnKit + ΔlnA, where the v’s denote the shares in total factor income, Y denotes GDP, L denotes labour input (measured as total hours worked), K denotes capital, the subscript it denotes the information technology sector, the subscript n denotes the non-IT sector and A denotes Hicks-neutral technological progress that augments the aggregate input.10 The results of this exercise can shed light on the question why the economies of some European countries – like Ireland, Greece, Finland, Spain, Sweden and the UK – grew very well, while others – in particular the large economies of Germany, Italy and France – experienced very sluggish growth over the last decade. We begin by establishing a benchmark for our analysis and first consider the US, which over the last ten years has grown remarkably well. 8.4.1 Benchmark: sources of growth in the US Applying the basic growth-accounting equation given above to US data, we obtain the results in Table 8.1. The results show that, for example, in the five-year period 1995–2000, GDP grew by an average annual rate of 4.2 per cent, of which growth of labour input contributed 1.3 per cent, growth of IT capital 0.9 per cent, growth of non-IT capital 0.6 per cent, while general technological progress (TFP growth) on average contributed 1.5 per cent per annum. In the latter period 2000–2004, growth was clearly lower than in the preceding five years and it came primarily from the growth of capital inputs (with IT capital again being somewhat more important than non-IT capital) and technological progress. The contribution of labour input was even slightly negative – probably as a result of the downturn in the US economy.11 The US benchmark yields several important results. First, the growth rate of the US economy has been quite high, which runs counter to usual notions of convergence. Second, the role of IT capital is quite strong as its growth
9780230_547018_10_cha08.indd 268
3/19/2009 8:16:28 PM
Economic Growth in the European Union 269 Table 8.1 Growth accounting for the US GDP growth
1995–2000 2000–2004
4.2 2.4
Contributions to GDP growth by IT capital growth
Non-IT capital growth
Labour growth
TFP growth
0.9 0.6
0.6 0.4
1.3 0.3
1.5 1.7
Note: The columns in the growth-accounting tables may not add because of rounding. Source: Groningen Growth and Development Center (GGDC), Total Economy Growth Accounting Database.
contribution has been higher than that of other capital. Such a key role of IT is a recent finding in studies of economic growth.12 A third striking feature of US growth is that growth in labour input played a strong positive role in the late 1990s even if it has had a minor negative contribution since 2000. 8.4.2 The laggard countries: Germany, France and Italy We next investigate economic growth in the large EU countries that have grown slowly over the period under investigation: Germany, Italy and France. The results of the growth-accounting exercises for these countries are reported in Tables 8.2a–c. The results suggest a number of important conclusions. The first striking observation is that labour growth contributed negatively in Germany throughout the ten-year period and in France after 2000. This negative contribution is likely to be due to increased unemployment as well as working time reductions. It will be examined further below. Second, conventional capital appears to have been more important than IT capital for growth in the laggard countries, with Germany being somewhat of an exception in 2000–2004. However, the contribution from growth in IT capital in Germany was very low anyhow. This is in marked contrast to the US, as shown in Table 8.1 above. Third, total factor productivity has not been a major source of growth in most cases, though France and Germany in 1995–2000 are exceptions. However, in the period 2000–2004, TFP growth was low in both France and Germany. We will discuss possible reasons for slow TFP growth below. The small role of technological progress is particularly marked for Italy, where non-IT capital growth has been the main source of growth, and TFP growth is even negative after 2000. The Italian experience (as well as that of other earlier high-interest-rate countries discussed below) may partly be explained by the introduction of the euro. The common capital market has induced capital flows from former low-interest-rate countries, like Germany, to former high-interest-rate countries, like Italy. With the introduction of the common currency, interest rates have been equalised across the eurozone.
9780230_547018_10_cha08.indd 269
3/19/2009 8:16:28 PM
270
Designing the European Model
Table 8.2a
Growth accounting for Germany GDP growth
1995–2000 2000–2004
1.7 0.5
Contributions to GDP growth by IT capital growth
Non-IT capital growth
Labour growth
TFP growth
0.3 0.2
0.3 0.2
0.4 0.5
1.5 0.6
Source: Groningen Growth and Development Center (GGDC), Total Economy Growth Accounting Database.
Table 8.2b
Growth accounting for France GDP growth
1995–2000 2000–2004
2.7 1.4
Contributions to GDP growth by IT capital growth
Non-IT capital growth
Labour growth
TFP growth
0.3 0.2
0.7 0.8
0.2 0.1
1.6 0.5
Source: Groningen Growth and Development Center (GGDC), Total Economy Growth Accounting Database.
Table 8.2c
Growth accounting for Italy GDP growth
1995–2000 2000–2004
2.0 0.9
Contributions to GDP growth by IT capital growth
Non-IT capital growth
Labour growth
0.4 0.4
0.7 0.8
0.5 0.8
TFP growth 0.4 1.1
Source: Groningen Growth and Development Center (GGDC), Total Economy Growth Accounting Database.
We observe today relatively low contributions of non-IT capital growth in the former low-interest-rate (capital abundant) countries and high contributions in the former high-interest-rate (capital scarce) ones. Fourth, the contribution of IT capital growth has been relatively small in the three laggard countries. Moreover, quite remarkably the share of TFP growth in total GDP growth has declined over time. This suggests that the laggard countries have not been successful in making use of the new opportunities provided by the IT revolution. 8.4.3 The successful cases: Ireland, Finland, Greece, Spain, Sweden and the UK Looking at the countries that are usually regarded as the European success stories, it is difficult to find clear patterns that are common to all of these
9780230_547018_10_cha08.indd 270
3/19/2009 8:16:28 PM
Economic Growth in the European Union 271
countries. It appears that we can identify two different groups of successful cases. The first group, consisting of Finland, Ireland, Sweden and the UK, has relied on the IT revolution as these countries have had relatively rapid growth of IT capital. Growth of IT capital has been more important than growth of conventional capital for Finland, Sweden and the UK. However, in Ireland non-IT capital has been relatively more important than IT capital, but the growth rate of IT capital has also been high. Tables 8.3a–d show the results of growth-accounting computations for Finland, Ireland, Sweden and the UK. This first group of countries is also characterised by the significant role of general technological progress, as indicated by the growth rates of TFP. We can also observe that the increase of labour input has been an important underlying source of growth in Ireland and in the UK. However, this observation does not hold for the Scandinavian countries, as in these countries growth in labour input was negative in 2000–2004. The positive contribution in the first period can be largely explained by increased labour utilisation when Finland and Sweden were emerging from the deep recessions in the first half of the 1990s.13 The second group of successful countries comprises Greece and Spain, which also grew clearly better than the average EU country. In these two countries, the sources of growth differ quite substantially from the growth patterns from the first group of countries discussed, where growth was largely driven by IT capital and TFP. Tables 8.4a–b give the results of the growth-accounting decompositions for Greece and Spain.
Table 8.3a
Growth accounting for Ireland GDP growth
1995–2000 2000–2004
9.7 5.0
Contributions to GDP growth by IT capital growth
Non-IT capital growth
Labour growth
TFP growth
0.6 0.4
2.3 2.3
2.1 0.5
4.7 1.9
Source: Groningen Growth and Development Center (GGDC), Total Economy Growth Accounting Database.
Table 8.3b
Growth accounting for Finland GDP growth
1995–2000 2000–2004
4.9 2.3
Contributions to GDP growth by IT capital growth
Non-IT capital growth
0.7 0.6
0.1 0.3
Labour growth 1.0 0.3
TFP growth 3.0 1.7
Source: Groningen Growth and Development Center (GGDC), Total Economy Growth Accounting Database.
9780230_547018_10_cha08.indd 271
3/19/2009 8:16:29 PM
272 Designing the European Model Table 8.3c
Growth accounting for Sweden GDP growth
1995–2000 2000–2004
3.5 2.1
Contributions to GDP growth by IT capital growth
Non-IT capital growth
0.8 0.4
0.4 0.2
Labour growth 0.7 0.4
TFP growth 1.7 1.9
Source: Groningen Growth and Development Center (GGDC), Total Economy Growth Accounting Database.
Table 8.3d
Growth accounting for the UK GDP growth
1995–2000 2000–2004
3.3 2.3
Contributions to GDP growth by IT capital growth
Non-IT capital growth
Labour growth
TFP growth
0.8 0.34
0.6 0.5
0.7 0.2
1.2 1.3
Source: Groningen Growth and Development Center (GGDC), Total Economy Growth Accounting Database.
Table 8.4a
Growth accounting for Spain GDP growth
1995–2000 2000–2004
4.0 2.5
Contributions to GDP growth by IT capital growth
Non-IT capital growth
Labour growth
0.3 0.3
1.1 1.2
2.8 1.6
TFP growth 0.3 0.6
Source: Groningen Growth and Development Center (GGDC), Total Economy Growth Accounting Database.
Table 8.4b
Growth accounting for Greece GDP growth
1995–2000 2000–2004
3.8 4.2
Contributions to GDP growth by IT capital growth
Non-IT capital growth
Labour growth
TFP growth
0.3 0.4
0.6 0.9
0.7 1.0
2.4 0.3
Source: Groningen Growth and Development Center (GGDC), Total Economy Growth Accounting Database.
9780230_547018_10_cha08.indd 272
3/19/2009 8:16:29 PM
Economic Growth in the European Union 273
In Spain, labour input has been by far the most important source of growth. This observation indicates that Spain was very successful in recent years in addressing its unemployment problem. Furthermore non-IT capital growth played a major role, which is partly due to a euro-driven single capital market effect, as previously discussed in the case of Italy. In both Greece and Spain, the perceived “country risk premium” in interest rates explained by exchange rate risk has disappeared after the introduction of the euro. The fall in interest rates has stimulated investment and explains the large contribution of non-IT capital growth. In Greece, the contribution from the individual factors of production has been of relatively similar magnitude, though in Greece the contribution of TFP growth was also quite large in the second half of the 1990s (and surprisingly small in the period 2000–2004).
8.5
Why are there differences in growth performance?
The preceding discussion has shown that, during the last ten years, there have been major differences in rates of economic growth and in the sources of growth between successful and unsuccessful EU countries. It is important to deepen our understanding of the possible reasons for these differences. Investigating them may also provide answers to the crucial question: to what extent can economic policy influence growth and what might be appropriate growth policies? We found in the previous section that the first group of the successful EU countries appeared to have strong growth in IT capital and strong overall technological progress (TFP growth), while the second group of successful countries relied more on traditional engines of growth: non-IT capital and labour. In this section we examine more closely the sources of economic growth for the Western EU countries. 8.5.1 Capital formation in the EU countries Table 8.5 shows the growth rates of IT and non-IT capital services for the EU-15 countries in the period 1995–2004.14 Capital services are measured using the methodology developed by Jorgenson and Griliches (1967), in which growth rates of different types of capital are weighted using average shares of each capital asset type in the value of the property compensation in terms of rental prices.15 Table 8.5 shows that most EU countries invested rather heavily in IT capital, which led to impressive growth rates in IT capital services in the boom period 1995–1999. In the period 2000–2004 IT capital growth slowed down. Interestingly, Ireland, Finland, Sweden and the UK do not stand out according to IT capital growth, even though they had significant contributions to growth from this source. The explanation, shown in Table 8.6, is instead that these countries already had a relatively high share of IT capital as their technologies had adapted to the relatively high use of IT capital before the mid-1990s. The same applies to the US. Differences in the shares of IT capital
9780230_547018_10_cha08.indd 273
3/19/2009 8:16:29 PM
274 Designing the European Model Table 8.5 Growth rates of IT and non-IT capital services in EU-15 countries, 1995–2004
Austria 95–00 00–04 Belgium 95–00 00–04 Denmark 95–00 00–04 Finland 95–00 00–04 France 95–00 00–04 Germany 95–00 00–04 Greece 95–00 00–04 Ireland 95–00 00–04
IT Capital
Non-IT Capital
14.5 10.9
2.0 2.1
20.3 11.0
0.7 0.4
10.9 17.6
1.4 2.9
13.8 10.8
0.2 0.2
16.9 8.6
2.2 2.4
13.4 7.8
1.0 0.3
18.9 15.1
3.8 5.3
34.6 13.5
5.9 5.0
IT Capital Italy 95–00 00–04 Luxembourg 95–00 00–04 Netherlands 95–00 00–04 Portugal 95–00 00–04 Spain 95–00 00–04 Sweden 95–00 00–04 UK 95–00 00–04 US 95–00 00–04
Non-IT Capital
13.6 10.0
2.4 2.5
17.0 12.9
5.9 4.3
21.1 9.0
1.8 1.1
21.6 10.2
4.4 2.3
15.8 9.4
3.8 3.8
19.1 6.8
1.7 0.7
20.1 8.3
2.1 1.1
17.9 9.1
2.8 1.8
Source: Groningen Growth and Development Center (GGDC), Total Economy Growth Accounting Database.
are an important explanation of why the recent growth contribution of IT capital has varied among countries. Early users of IT capital benefited from the high-tech boom of the 1990s. There is more variation among the EU countries in the growth rates for conventional capital. Ireland, Greece and Spain (as well as Luxembourg and Portugal) have had very high growth rates of non-IT capital. In contrast, the growth of conventional capital has been low for Germany and also Finland, though the latter country is among the success cases in the EU. Finland relied on increases in IT capital for its growth, but this was not the only reason for success. As can be seen from Table 8.3b, TFP growth was a major contributor to growth in Finland. This was in turn a result of structural
9780230_547018_10_cha08.indd 274
3/19/2009 8:16:29 PM
Economic Growth in the European Union 275 Table 8.6 Average share of GDP imputed to IT and non-IT capital, 1995–2004 Non-IT Capital
IT Capital
Austria
32.6
3.5
Belgium
25.1
4.4
Denmark
30.0
5.0
Finland
29.8
4.7
France
33.0
2.4
Germany
29.4
3.5
Greece
20.4
2.3
Ireland
41.4
2.3
Italy
34.9
3.7
Luxembourg
32.3
3.7
Netherlands
26.4
2.8
Portugal
26.2
3.0
Spain
28.9
2.7
Sweden
24.7
5.6
UK US
26.7 24.0
4.4 6.0
Source: Groningen Growth and Development Center (GGDC), Total Economy Growth Accounting Database.
changes in which low-productivity activities were replaced by new activities with higher productivity (see Honkapohja, Koskela, Leibfritz and Uusitalo 2005 for an analysis of the Finnish case). Diffusion of IT Another important aspect concerns how widespread the use of IT is. We compare EU countries, using two broad indicators of IT diffusion. Figure 8.7 describes the share of expenditure (as per cent of GDP) on IT in each country, while Figure 8.8 displays the number of personal computers per 1,000 persons. It is evident that Sweden is a clear leader in both these respects. The other high-tech based success cases, the UK, Finland and Ireland are not quite top performers according to these indicators, although on the whole they score well above the average. On the other hand, the other success cases, Greece and Spain, are below average in these indicators, which confirms the view that their success was not based on wide adoption of IT. Of the laggard countries, France and Germany score relatively high – somewhat above average – in terms of these indicators, while Italy is well below the average.
9780230_547018_10_cha08.indd 275
3/19/2009 8:16:30 PM
276 Designing the European Model 7.3 7.0 6.9 6.6 6.4 6.1 5.9 5.7 5.7 5.5 5.3 5.3
United Kingdom Sweden Finland Czech Republic Netherlands Hungary France Germany Denmark Belgium Slovak Republic Austria Poland Greece Portugal Italy Ireland Spain
4.5 4.3 4.2 4.1 3.9 3.8 0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
Figure 8.7 Information and communication technology expenditure, 2003 (as a percentage of GDP) Note: This figure gives total spending on IT and is different from investment into IT capital. Source: World Bank, World Development Indicators 2005.
8.5.2 Structure of labour input The previous section showed that there was no clear pattern among the various groups of countries of how labour growth has contributed to total growth in the economy. Among the successful cases, labour contributed positively to growth in Spain, Greece, Ireland and the UK over the whole 1995–2004 period, but had partly negative contributions in Finland and Sweden. In the group of lagging countries, the contribution was positive in Italy and negative in Germany and France. In this section, we decompose changes in the total hours worked into changes in annual hours per worker (working time) and changes in the total number of employed workers (employment). While a reduction in working time might be the outcome of negotiations between unions and employers, and therefore could sometimes even be interpreted positively as a welfare gain, comparable to a wage increase, a reduction in employment would clearly reflect problems in the labour market.16 Table 8.7 shows that total annual hours worked have decreased in almost all EU countries over the 1995–2004 period, Belgium, Denmark and Greece being the only exceptions with minor increases. Remarkably, total annual hours also decreased in the United States during the period 2000–2004. Employment (the total number of workers), on the other hand, has increased in almost all countries. The only instances where employment has
9780230_547018_10_cha08.indd 276
3/19/2009 8:16:30 PM
Economic Growth in the European Union 277 621.3 576.8
Sweden* Denmark* Germany Netherlands* Finland* Estonia Ireland* United Kingdom* Austria* France* Belgium Slovenia* Italy* Spain* Latvia Slovak Republic* Czech Republic* Poland Portugal Lithuania* Hungary* Greece*
484.7 466.6 441.7 440.4 420.8 405.7 369.3 347.1 318.1 300.6 230.7 196.0 188.0 180.4 177.4 142.0 134.4 109.7 108.4 81.7 0
Figure 8.8
100
200
300
400
500
600
700
Personal computers per 1000 people, 2003
Note: *2002. Source: World Bank, World Development Indicators 2005.
fallen are Denmark and Germany. This observation explains why Germany has experienced such a large negative contribution of labour growth to output growth as described in the earlier section. While Germany shared in the working time reductions that were common to most of the EU countries, it also experienced a reduction in the total number of employed workers, reflecting both increases in unemployment and reductions in labour force participation. The combination of a decline in working time and employment makes Germany unique in Europe. This development stands in sharp contrast to the US. As regards the successful countries, Table 8.7 is also interesting. Finland and Sweden, which also experienced negative contributions from total hours worked, both had substantial positive contributions from the total number of employed workers. In both countries, the negative contribution of labour to total growth after 2000 was in both countries due to substantial reductions in hours worked per employee. 8.5.3 Technological progress TFP growth can be thought of as a measure of general technological progress, which is not embodied in the explicit factors of production: labour and the various types of capital. The non-measurable factors in TFP include innovations and improvements in general knowledge and the organisation of production. However, since TFP growth is measured as a residual, it
9780230_547018_10_cha08.indd 277
3/19/2009 8:16:30 PM
278
Designing the European Model
Table 8.7 Changes in total annual hours worked, annual hours per employee and number of workers employed, 1995–2004
Austria 95–00 00–04 Belgium 95–00 00–04 Denmark 95–00 00–04 Finland 95–00 00–04 France 95–00 00–04 Germany 95–00 00–04 Greece 95–00 00–04 Ireland 95–00 00–04 Italy 95–00 00–04 Luxembourg 95–00 00–04 Netherlands 95–00 00–04 Portugal 95–00 00–04 Spain 95–00 00–04 Sweden 95–00 00–04 UK 95–00 00–04
Total hours
Annual hours per worker
Total number of workers
0.37 0.22
0.44 0.46
0.07 0.25
0.08 0.74
1.16 0.30
1.12 0.45
0.6 0.1
0.4 0.1
1.55 0.54
0.58 0.76
1.0 0.2 2.14 0.23
0.2
1.2
1.5
0.2
0.8
0.6
0.5 0.7
0.8 0.3
0.3 0.4
0.92 1.26
0.04 0.22
0.88 1.04
3.68 0.86
1.62 1.10
5.37 1.98
0.72 1.32
0.28 0.34
1.01 1.67
4.18 2.90
0.09 0.00
4.28 2.90
3.04 0.09
0.19 0.26
3.23 0.17
1.40 0.04
1.19 0.22
2.63 0.18
4.08 2.46
0.44 0.46
4.09 2.66
0.96 0.60
0.44 0.46
0.84 0.36
1.05 0.30
0.18 0.51
1.23 0.81 Continued
9780230_547018_10_cha08.indd 278
3/19/2009 8:16:30 PM
Economic Growth in the European Union 279
Table 8.7 Continued Total hours
US 95–00 00–04 EU-15 95–00 00–04
Annual hours per worker
Total number of workers
1.92 0.40
0.21 0.81
1.71 0.43
0.9 0.4
0.5 0.3
1.4 0.7
Source: Groningen Growth and Development Centre (GGDC), Total Economy Growth Accounting Database.
Table 8.8 TFP growth in EU-15 countries, 1995–2004
Austria Belgium Germany Denmark Spain Finland France Greece Ireland
1995–00
2000–04
1.7 1.7 1.3 1.4 0.3 3.3 1.4 1.9 4.4
0.2 0.3 0.6 0.3 0.5 2.0 0.5 1.8 2.0
Italy Luxemburg Netherlands Portugal Sweden UK EU-15 US
1995–00
2000–04
0.2 1.6 0.6 1.0 1.3 1.1 0.9 1.1
1.2 0.9 0.2 0.3 1.9 1.5 0.4 1.7
Source: Groningen Growth and Development Center (GGDC), Total Economy Growth Accounting Database.
also contains other effects such as cyclical ones, pure changes in efficiency, and measurement errors. It can also contain effects from improvements in labour and capital quality, since such quality improvements are difficult to quantify and may not be fully incorporated into factor shares and growth rates of the corresponding productive factors. Table 8.8 shows the TFP growth rates for the EU-15 countries.17 The high TFP growth in Finland, Ireland, Sweden and the UK is clearly visible. These countries do better than or about equally well as the US. Interestingly, Greece has also experienced high TFP growth, which has given an important contribution to its high growth. As discussed above, TFP growth in principle measures general technological progress, including structural change. However, innovations and improvements in the general knowledge and organisation of economic
9780230_547018_10_cha08.indd 279
3/19/2009 8:16:31 PM
280 Designing the European Model
activities are not directly measurable. This means that one has to restrict the analysis to only indirect indicators of determinants of TFP. It is usually thought that high quality of the education system, strong competition and deregulation, and innovation and entrepreneurship can enhance TFP growth. Therefore, we next look at indicators of these factors. Education Education is often considered a key determinant of economic growth. It is regarded as one of the most important potential policy instruments for raising both TFP growth and economic growth in general.18 Education has also been subject to intensive policy discussion in the EU, as evidenced by, for example, the emphasis on education and the information society in the Kok report (EU 2004). A traditional way of studying the role of education in economic growth is to allow for human capital as an explicit determinant of economic growth (which we did not do above). Human capital is then usually measured as the average number of years in schooling. With this measure, education has been found to have a clear positive effect on growth.19 Figure 8.9 provides basic data on the educational expenditures in EU-25 countries. We see that some countries, particularly Sweden and Finland, which were found to have large contributions of IT capital and TFP growth, also have large shares of expenditure on education in GDP. However, the
Denmark Sweden Cyprus Finland Belgium Slovenia Lithuania Portugal Latvia France Estonia Austria Poland Hungary United Kingdom Netherlands Germany Italy Malta Spain Czech Republic Slovakia Ireland Luxembourg Greece 0.0
8.5 7.7 6.8 6.4 6.3 6.0 5.9 5.8 5.8 5.8 5.7 5.7 5.6 5.5 5.3 5.1 4.8 4.8 4.5 4.4 4.4 4.4 4.3 4.0 4.0 1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
Figure 8.9 Expenditure on education in EU countries, 2002 (as a percentage of GDP) Source: Eurostat, population and social conditions. Education data.
9780230_547018_10_cha08.indd 280
3/19/2009 8:16:31 PM
Economic Growth in the European Union 281
correlation between TFP growth and education spending is not that strong. For example, Ireland and the UK, which also ranked among the highest in terms of the contribution of IT capital and TFP to overall growth, are among the countries with the lowest expenditure on education. The possible links between education spending and growth may be indirect and work through other variables. An analysis of education systems is made in Chapter 13. Competition and Regulation As technological change is to a significant degree associated with the emergence of new and more productive firms, the degree of competition is potentially an important element behind TFP growth.20 One way for public policy to influence competition is through regulatory policies – a less regulated economy makes it easier to establish new firms and thereby enhance competition in the economy. The findings of Alesina et al. (2005) suggest that regulatory reform leads to increased investment of firms, so that effects of competition can work through increased capital accumulation and not only through TFP growth. To examine the strictness of regulation in the EU countries, we use the structural indicators on product market regulation and employment protection constructed by the OECD. Indicators of product market regulation and employment protection are reported in Tables 8.9a–b, respectively. As a general tendency, the EU countries have been moving towards less regulation in product markets, but this has been happening in varying degrees. The successful high-tech EU-15 countries, Finland, Ireland, Sweden and the UK, appear to be among the countries with lowest degrees of product market regulation. Their levels of regulation are nowadays close to that of the US. In contrast, France and Italy, and also Germany (although the difference here is smaller), have a higher degree of product market regulation. The same seems to be true for the smaller EU countries that have not done so well in terms of economic growth. We also note that, of the EU success cases, Greece and Spain also have levels of regulation that are comparable to those of the laggard countries. The overall picture is thus not clear-cut. It appears that the connection between competition and growth can depend on the nature of the growth process. Conventional sources of growth might be less sensitive to regulatory intervention than growth relying on high-tech and new products, where competition should be particularly encouraged.21 In Greece for instance, a high degree of state control accounted for the high number in 1998 (see Conway et al. 2005). On average, the progress in lowering barriers to competition is due less to increased entrepreneurship and more to reducing state control (like price control, command-and-control measures or direct control of business enterprises) and barriers to trade and investment (like declining average most-favoured-nation tariff rates or diminishing restrictions on FDI). According to Conway et al. (2005), the
9780230_547018_10_cha08.indd 281
3/19/2009 8:16:31 PM
282
Designing the European Model
Table 8.9a Product market regulation in EU-15 countries and the US Product market regulation
Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden UK US
1998
2003
1.8 2.1 1.5 2.1 2.5 1.9 2.8 1.5 2.8
1.4 1.4 1.1 1.3 1.7 1.4 1.8 1.1 1.9 1.3 1.4 1.6 1.6 1.2 0.9 1.0
1.8 2.1 2.3 1.8 1.1 1.3
Note: A higher number reflects stronger regulation. The indicator measures the degree to which policies promote or inhibit competition and summarise a large set of regulations and formal rules. The data for the indicators derive from answers to questionnaires sent to OECD member governments. The questionnaire contained questions spanning from general and sectoral regu latory policies (firm ownership, state control, market access, entry requirements, regulation in transport industries etc.) to industry structure. “YES/NO” answers are coded by assigning a numerical value to each possible response to a particular question. Quantitative information is subdivided into classes using a system of thresholds. The coded information is then normalised over a scale of zero to six reflecting increasing restrictiveness of regulatory provisions for competition. Source: Conway, Janod and Nicoletti (2005).
progress in France and Spain was especially driven by a reduced administrative burden for firm start-up, whereas Italy removed legal barriers for entry to some sectors, and Finland, Greece and Sweden improved the system of licence and permits, thus reducing barriers to entrepreneurship between 1998 and 2003. Table 8.9b looks at regulation in the labour markets using indicators of employment protection for both regular and temporary employment. The tendency towards less regulation is visible, but it is arguably much weaker than for product market regulation. In a number of cases, regulation has remained unchanged or has even tightened somewhat. There is also quite a lot of variation even among the successful high-tech countries Finland,
9780230_547018_10_cha08.indd 282
3/19/2009 8:16:31 PM
Economic Growth in the European Union 283 Table 8.9b Strictness of employment protection legislation in EU-15 countries and the US Regular employment
Austria Belgium Denmark Finland France Germany Greece Ireland Italy Netherlands Portugal Spain Sweden UK US
Temporary employment
Late 1980s
Late 1990s
2003
Late 1980s
Late 1990s
2003
2.9 1.7 1.5 2.8 2.3 2.6 2.5 1.6 1.8 3.1 4.8 3.9 2.9 0.9 0.2
2.9 1.7 1.5 2.3 2.3 2.7 2.3 1.6 1.8 3.1 4.3 2.6 2.9 0.9 0.2
2.4 1.7 1.5 2.2 2.5 2.7 2.4 1.6 1.8 3.1 4.3 2.6 2.9 1.1 0.2
1.5 4.6 3.1 1.9 3.1 3.8 4.8 0.3 5.4 2.4 3.4 3.8 4.1 0.3 0.3
1.5 2.6 1.4 1.9 3.6 2.3 4.8 0.3 3.6 1.2 3.0 3.3 1.6 0.3 0.3
1.5 2.6 1.4 1.9 3.6 1.8 3.3 0.6 2.1 1.2 2.8 3.5 1.6 0.4 0.3
Note: The overall summary measure of EPL strictness is based on three components related to specific requirements for collective dismissals, protection of regular workers against (individual) dismissal (which constitutes the core component of the overall summery index) and regulation of temporary forms of employment. There are 18 items that describe these areas. These are expressed either in units of time (for example months of notice and severance pay), as a number or as a score on an ordinal scale specific to each item. All these measures were converted into cardinal scores that were normalised to range from 0 to 6. A higher number reflects stronger regulation. The weighted average was constructed for the average indicator where the measure for collective dismissals was attributed 40 percent of the weight assigned to regular and temporary contracts. Source: OECD Employment Outlook 2004.
Ireland, Sweden and the UK. The levels of employment protection are rather low in Ireland and the UK, though they are clearly higher than in the US. Finland and Sweden have higher levels of employment regulation than Ireland and the UK. However, Finland and Sweden have taken some steps towards lowering employment protection: Finland with respect to regular employment and Sweden with respect to temporary employment. The other EU success cases of Greece and Spain, as well as the laggard cases of France, Germany and Italy, have on the whole clearly higher employment protection regulation than high-tech successful countries. Some theoretical studies suggest a negative relationship between employment protection and growth (see, e.g., Bertola 1994 and Boone 2000), but empirical evidence has not been studied.22
9780230_547018_10_cha08.indd 283
3/19/2009 8:16:32 PM
284
Designing the European Model
Innovation Promotion of innovations and facilitating start-ups of new production activities are another possible policy tool for improving TFP growth. It is, however, difficult to find good measures of innovative activities and start-ups of new production that are relevant for long-term growth. We consider two indicators, venture capital financing and investment in R&D. 23 As to venture capital financing, the statistical data distinguish between “early-stage” and “expansion and replacement” (or late-stage) venture finance. Table 8.10 shows that the US is a clear leader according to both indicators. Finland, Sweden and the UK also do well in terms of both indicators. The performance of Ireland is close to the EU average: it is above average in early-stage and below the average in late-stage venture financing. The picture for the rest of the EU-15 countries is far more varied. Spain is doing fairly well in terms of the late-stage venture finance indicator, but poorly for early-stage finance. Greece is relatively low on both indicator counts. Of the laggard countries, Germany is close to or above average in terms of early-stage venture financing, but it does not do so well in
Table 8.10 Early-stage versus (expansion and replacement) venture capital investment in EU-15 countries and the US (percentage of GDP) Early-stage venture capital investment
Expansion and replacement venture capital investment
Average 1995–2000
Average 1995–2000
Average 2000–2004
Average 2000–2004
0.0025 0.0313 0.0060 0.0237 0.0128 0.0168 0.0060 0.0143 0.0080 0.0432 0.0080 0.0062 0.0203 0.0093 0.0132 0.0400
0.0025 0.0313 0.0060 0.0237 0.0128 0.0168 0.0060 0.0143 0.0080 0.0432 0.0080 0.0062 0.0203 0.0093 0.0132 0.0400
0.0164 0.0584 0.0498 0.0790 0.0428 0.0458 0.0128 0.0466 0.0186 0.0576 0.0198 0.0178 0.0896 0.0510 0.0416 0.1384
0.0164 0.0584 0.0498 0.0790 0.0428 0.0458 0.0128 0.0466 0.0186 0.0576 0.0198 0.0178 0.0896 0.0510 0.0416 0.1384
Austria Belgium Denmark Finland France Germany Greece Ireland Italy Netherlands Portugal Spain Sweden UK EU-15 US Source: Eurostat.
9780230_547018_10_cha08.indd 284
3/19/2009 8:16:32 PM
Economic Growth in the European Union 285
terms of late-stage venture finance. France is above the EU-15 average for late-stage venture finance, but does fairly poorly in early-stage venture financing. Italy does poorly on both counts. Overall, EU-15 countries are well behind the US in venture financing. In the EU, it appears that the importance of venture capital financing correlates fairly strongly with the relative importance of high-tech industries, but otherwise the picture is not so clear-cut. Table 8.11 shows R&D spending as a fraction of GDP as another indicator of innovative activity that contributes to general technological progress. On this count, two EU success cases, Finland and Sweden, do particularly well. Especially in the period 2000–2004, R&D spending as a share of GDP in these countries was even higher than in the US. The other high-tech EU economies, Ireland and the UK, are not very big spenders on R&D and are below the EU-15 average. Interestingly, of the laggard countries, France and Germany are ahead of Ireland and the UK according to this indicator. On the other hand, Italy does poorly: its share of R&D in GDP is quite low and is not much higher than the figures for Greece and Spain, which are relying on traditional sources of growth and not on high-tech. Overall, R&D spending seems to have some relationship to fast growth, but the relationship is not very strong.24
Table 8.11 Expenditure on R&D in EU-15 countries and the US (percentage of GDP)
Austria Belgium Denmark Finland France Germany Greece Ireland Italy Netherlands Portugal Spain Sweden UK EU-15 US
Average 1995–2000
Average 2000–2004
1.73 1.88 2.01 2.82 2.20 2.29 0.28 1.25 1.04 1.99 0.31 0.86 2.36 1.86 1.89 2.58
2.10 2.01 2.49 3.44 2.18 2.48 0.37 1.14 0.90 1.79 0.49 0.77 2.40 1.50 1.96 2.13
Source: Eurostat.
9780230_547018_10_cha08.indd 285
3/19/2009 8:16:32 PM
286
Designing the European Model
8.6 Eastern Europe The data for the new EU member countries are less complete than for the EU-15 countries. However, an analysis that is comparable to Section 4 has been conducted by van Ark and Piatkowski (2004). They look at growth in labour productivity in the Central East European (CEE) countries in the period 1995–2001. In this paper, the authors show that, similar to Ireland and Finland (which were the EU-15 countries with the highest TFP growth) during this time period, the Eastern European countries experienced large increases in total factor productivity, which has been the largest contributor to overall growth in GDP per capita. The Czech Republic is an exception to this pattern. It is also shown in their study that IT capital has played a smaller role in the CEE countries than in the EU-15 countries. Not surprisingly, economic growth in these countries is relying on traditional means of growth in conventional capital, labour and total factor productivity. It is also instructive to look at the development of total hours worked, which are shown in Table 8.12 below. The countries for which the more recent data are available are the Czech Republic, Hungary, Poland and the Slovak Republic. Table 8.12 shows that total hours worked have fluctuated substantially over the last years. In Hungary, total hours worked have contributed positively to GDP growth in all years except 2000–2001. In the other countries, the contribution of labour to growth has been quite sizable and mostly negative. Looking at the determinants of growth for the new member countries, we must note that there are major gaps in the data. Data on diffusion of IT are available for several new member countries and were included in Figures 8.7 and 8.8. The results vary from country to country. The data on education expenditures is relatively good for the new members and they are included in Figure 8.9 above. It is seen that most of these countries spend above or close to the EU-25 average share on education, with the Czech and Slovak Republics having lower spending on education. Data on other factors that could influence TFP growth – shown in the Appendix – are incomplete.
Table 8.12
Growth rates of total hours worked in Eastern Europe
Czech Republic Hungary Poland Slovak Republic
98–99
99–00
00–01
01–02
02–03
03–04
3.5 4.0
0.3 0.7
2.1
1.7
4.1 1.8 2.9 0.1
0.6 1.1 2.8 3.4
0.6 1.9 1.0 1.1
3.4 1.1 1.3 1.9
Source: OECD, Productivity data base, July 2005, and IMF, International Financial Statistics.
9780230_547018_10_cha08.indd 286
3/19/2009 8:16:32 PM
Economic Growth in the European Union 287
Tables A8.1a–b show that there tends to be more of employment protection and product market regulations than in the EU-15 countries, though the protection indicator for temporary employment tends to have lower values than for EU-15 countries. The new member countries receive low scores on venture capital indicators (see Tables A8.2a–b). For R&D expenditure shown in Table A8.3, the new member countries have somewhat lower spending relative to GDP than the EU-15 countries, though the Czech Republic and Slovenia are exceptions in this respect. Overall, for the new member countries it is difficult to draw strong conclusions with respect to these indicators. Not surprisingly, these countries tend to score lower than the old EU members, though according to some indicators, such as education, their performance is good. These countries are likely to continue to grow through traditional means – capital investment and TFP growth associated with structural change.
8.7
Policy challenges for the EU
Our analysis of economic growth in the EU has yielded many results that bear on current policy discussions in Europe. The most striking conclusion is that the Lisbon strategy should be modified. The Lisbon strategy argues for the creation of a uniform model of a high-tech information society for the EU. The problem with this line of thinking is the restrictive focus on a single model; the model is designed to imitate the success of the US economy in creating and making use of the IT revolution. The European experience in the last ten years suggests that this is not the right approach. There are different routes to success, as is witnessed by the experience of the successful EU countries. Some of the countries – Finland, Sweden and the UK – have focused on technological transformation and structural change involving increased use of advanced technologies, in particular IT. The successes of Finland, Sweden and the UK have indeed many similarities with the US model. But other successful countries have had a different strategy for growth. Ireland has had great success on many fronts and not only in the development and use of IT. Spain and Greece have relied on traditional sources of growth, capital accumulation and increasing labour input, and not on high technology. The different routes to success show that a growth strategy for the EU countries should not be based on a uniform model. Some of the countries are on the frontier of creation and adoption of new technologies. It is natural for these countries to continue with this strategy for growth. However, it must be recognised that the high-tech strategy involves major risks and it is unlikely to be successful for all EU states. Major failures may result if EU-driven technology policy is made the main part of the path forward. The public sector bureaucrats and politicians are probably not the right
9780230_547018_10_cha08.indd 287
3/19/2009 8:16:32 PM
288
Designing the European Model
people for picking future winners in the high-technology businesses. It is better to rely on private profit motives and finance for the promotion of high-technology industries. Instead the EU should allow for a flexible strategy for growth, in which there is scope for both high-tech-driven growth as well as growth based on more traditional means of capital accumulation, increased labour input and the imitative adoption of new technologies from the leaders. The examples of Spain and Greece demonstrate that the latter approach can also lead to success. Moreover, this is a natural strategy for the new EU member countries to follow, as they are currently well behind the high-technology frontier. Reaching the frontier is a gradual process, which will take many years. The key elements of growth policy lie elsewhere. First, policies should focus on improving the education systems, and this should be done at both the national and EU levels. The traditional studies of human capital and economic growth show the significance of education and, in addition, there are important complementarities between education systems and the ease of adopting innovations and new technologies.25 Diffusion of new technologies such as IT involve learning costs that decrease over time with the increasing number of users, and this process is facilitated by a well-qualified labour force. The evidence of high educational attainment in Finland and Sweden supports this conclusion. Both countries are on the IT technology frontier. Clearly, EU countries should direct major efforts to improving their education systems. There are important differences in the performance of the secondary education systems among EU countries as is discussed in Chapter 13. An important question in education policies concerns the level of education at which improvements should be focused. The answer appears to depend on whether the country is close or far from the technology frontier.26 Countries that are close to the frontier should specifically focus on improving the tertiary education systems, as high-technology innovations appear to require more advanced skills than lower-level innovations. The latter are often process improvements and rely on imitative adoption of known technologies. While the US does not stand out in the quality of secondary education, it is obviously well ahead of EU countries in universities, the part of the education system that matters the most for economic growth of the advanced countries. A larger proportion of an age class goes to higher education in the US than in the EU. In 2002 in the US, the percentage of age classes attending tertiary education was 38 per cent, while it was 33 per cent in Finland and Sweden, 28 per cent in the UK and 23 per cent in France and Germany. The university system in the US is quite varied, but the best universities compete strongly with each other for the best graduate students and researchers. In European countries, the university system does not generally work this way,
9780230_547018_10_cha08.indd 288
3/19/2009 8:16:32 PM
Economic Growth in the European Union 289
as there is no intensive competition for the best researchers and students. The UK is partly an exception to other European countries, as research and teaching quality audits there have increased competition. Nevertheless, even the best UK universities find it difficult to compete globally with the US universities. A second policy conclusion concerns the potential to increase labour input to enhance economic growth. The growth accounting in Section 4 of this chapter showed that labour input has not grown much and in some countries labour input growth was even negative for some periods. Labour market reforms are an appropriate means of raising labour input. Such reforms should include lowering unemployment benefits, introducing employment tax credits and reducing marginal tax rates on labour. Decentralised collective agreements lengthening working hours in firms exposed to heavy international competition (as in Germany) and reversals of earlier legislated working time reductions (e.g., in France) are other appropriate measures. So are reforms making pension systems more actuarial and increases in the retirement age. We have recommended these kinds of measures in part 1, chapters 2 and 5, and in part 3, chapter 12. The aging of population in EU countries makes these proposals particularly pertinent. The third policy conclusion concerns the easing of regulatory policies in the EU. Europe has relatively high levels of regulation that limit competition in various markets. The regulations concern limitations on entrepreneurial activities, entry restrictions and restrictions on labour market adaptability in hiring and firing, which tend to suppress innovation and technological advancements. As noted above, in terms of OECD structural indicators on product market regulation and employment protection regulation, the euro area scores much worse than the US or the UK. Interestingly, Finland and Sweden do well with respect to product market flexibility, but not so well with respect to labour market regulation.27 The results in the literature are somewhat tentative, but suggest that regulatory reforms tend to increase TFP growth and investment, which in turn should promote faster economic growth.28 The effects of deregulation work naturally via intensified competition, which in turn leads to increased entry and exit. Recent evidence suggests that the growth effects of entry and exit depend on the industry, more precisely on the distance of the industry from the technology frontier. 29 Industries that are close to the frontier can compete in the environment of intensified competition, whereas industries or sectors that are far behind the frontier face difficulties in such an environment. The exit of less efficient firms and their replacement by more efficient ones tends to have positive effects on economic growth. Technology policy should thus focus on the provision of opportunities for creation of new firms and industries and less on glorifying national
9780230_547018_10_cha08.indd 289
3/19/2009 8:16:32 PM
290 Designing the European Model
champions. Improvements of venture capital financing and R&D continue to be important policy areas for the EU countries. There are big variations in the amount of venture capital investments in the EU, and Europe is lagging behind the US in this respect. 30 Correspondingly, competition policies should focus attention on facilitating the entry of new firms. There are particular problems concerning competition and entry in the service sector. It is well established that very significant barriers to trade in services still exist in the EU. 31 These barriers derive from the fact that the cross-border provision of services requires the presence of service providers in the importing country. As a consequence, exporters of services tend to be subjected to national regulations in both the country of origin and in the host country. In view of the great importance of the services sector – making up around 70 per cent of both GDP and employment in the EU-15 – a lowering of trade barriers for services would potentially have large growth effects. It is therefore very important that the new EU Services Directive being discussed is not watered down, but is instead designed to open up the market for services for cross-border competition. The most significant barrier to intra-EU trade in services is that host countries can impose national pay conditions on posted workers from other EU member states (a right given by the so-called Posted-Workers Directive), as this prevents effective cross-border price competition. In the presence of such pay regulations, the gains from trade in services will be limited to those that can be derived from economies of scale, more effective organisation and greater product diversity. But one will not obtain the bulk of potential gains unless EU-15 states allow service providers from the new EU member states to compete effectively by compensating for lower productivity through lower wages. This is not “unfair wage dumping,” but a necessary precondition for the exploitation of different comparative advantages in old and new EU member states. There are no strong reasons why one should not allow wage competition among countries in trade with services in the same way as one does in trade with goods. The preceding conclusions on education, regulation and competition policies are in particular directed at improving the current growth performance of the old EU member countries. However, they also apply, to some extent, to the new member countries as well, though the policy recipes vary somewhat. The main concern of the new EU members is how to catch up best with the Western EU countries. The growth-enhancing policies for catching up include, in particular, facilitating technology transfer and improvement of productivity in industries that are mostly below the high technology frontier. Education policy and financing of new firms and innovations continue to be major items on the policy agenda for the new EU members.
9780230_547018_10_cha08.indd 290
3/19/2009 8:16:32 PM
Economic Growth in the European Union 291
Appendix: Selected indicators on new EU members Table A8.1a Strictness of employment protection legislation, selected new EU members Regular employment
Temporary employment
Late 1990s
2003
Late 1990s
2003
3.3 1.9 2.2 3.6
3.3 1.9 2.2 3.5
0.5 0.6 0.8 1.1
0.5 1.1 1.3 0.4
Czech Republic Hungary Poland Slovak Republic
Note: A higher score reflects stronger regulation. Source: OECD Employment Outlook 2004.
Table A8.1b
Product market regulation, selected new EU members Product market regulation
Czech Republic Hungary Poland Slovak Republic
1998
2003
3.0 2.5 3.9
1.7 2.0 2.8 1.4
Note: A higher score reflects stronger regulation. See also Table 8.9a. Source: Conway, Janod, and Nicoletti (2005).
Table A8.2a Early-stage venture capital investment in selected new EU members (percentage of GDP) Average 1999–2004 Czech Republic Hungary Poland Slovakia
0.008 0.006 0.015 0.003
Source: Eurostat.
Table A8.2b Expansion and replacement venture capital investment in selected new EU members (per mill of GDP) Average 1999–2004 Czech Republic Latvia Hungary Poland Slovakia
0.49 0.42 0.38 0.63 0.11
Source: Eurostat.
9780230_547018_10_cha08.indd 291
3/19/2009 8:16:32 PM
292
Designing the European Model Table A8.3 of GDP)
Expenditure on R&D in selected new EU members (percentage
Czech Republic Estonia Cyprus Latvia Lithuania Hungary Malta Poland Slovenia Slovakia
Average 1995–1999
Average 2000–2004
1.10 0.32 0.12 0.42 0.53 0.71 0.00 0.67 1.42 0.84
1.24 0.77 0.31 0.42 0.68 0.92 0.17 0.60 1.54 0.60
Source: Eurostat.
Notes 1. For example, see EEAG (2002), Chapter 5 for the data documenting convergence before the 1990s and divergence since the 1990s. 2. See p. 8 of EU (2004). 3. Various concepts of convergence will be discussed in more detail in the following sections. 1.7 refers to the annual reduction in the variance of the growth rates across countries, while 1.9 is the rate at which the initially lagging countries are closing the gap to the leading countries in terms of per capita incomes per year. 4. Thus, the time series of different countries cannot be used for comparing living standards. 5. Measuring this effect requires an appropriate price deflator for domestic absorption, which is not always available. 6. Neoclassical growth models also predict, ceteris paribus, convergence in levels of GDP. See, for example, Barro and Sala-i-Martin (2003) and Jones (2002) for a further discussion of these concepts. 7. The standard deviation of log per capita income across countries declines by 1.7 percent a year. Note that this is not the same as the convergence rate reported above. In the case of beta convergence, the convergence rate measures the rate at which the lagging countries are expected to catch up with higher-income countries. Only under specific assumptions about the distribution of growth rates and income levels would the two convergence rates be identical. 8 The kernel densities are the smoothed versions of a histogram of real per capita incomes. 9. Their data is publicly available at their web-page at: http://www.ggdc.net/dseries/ growth-accounting.shtml. 10. This is equation (6) of Timmer et al. (2003). 11. Although the total number of hours fell, as indicated in the table, the total number of employed workers increased slightly also in the period 2000–2004.
9780230_547018_10_cha08.indd 292
3/19/2009 8:16:32 PM
Economic Growth in the European Union 293 12. For a long time, growth accounting studies had difficulties in showing the importance of IT. 13. See Honkapohja, Koskela, Leibfritz and Uusitalo (2005) for a discussion of the 1990s crises in Finland and Sweden. 14. The data here, too, are from the Groningen Growth and Development Centre. See footnote 9. 15 See Timmer et al. (2003) for further details on the method of construction of the data. 16. Note, however, that in the case of France for instance the reduction in working hours was legislated, rather than the outcome of negotiations. Furthermore, the reduction in working hours is often the response to high unemployment, as work-sharing may be one way to alleviate the consequences. Chapter 5 discusses these issues in detail. 17. Due to some differences in the methods of computations, these numbers do not exactly match those in earlier tables in Section 4. 18. Griffith et al. (2004) provide recent evidence of the importance of education for innovation and absorptive capacity. 19. See, for example., Mankiw et al. (1992), Barro (1997) and Hall and Jones (1999). 20. Using British industry data, Nickell (1996) provides empirical evidence that higher competition is associated with higher rates of TFP growth. Nicoletti and Scarpetta (2003) also provide strong support for this relationship in OECD countries. 21. Nicoletti and Scarpetta (2003) provide empirical evidence for this argument. 22. Empirical studies (see, e.g., Nickell 2003 and chapter 2 for overviews) usually focus on other labour market institutions and their effects on (un)employment rather than growth. 23. High intensity of entrepreneurial activity has also been considered a possible determinant of TFP growth; see Achs et al. (2005) and Audretsch and Keilbach (2004) for empirical results on the connection between entrepreneurship and growth. The concept of entrepreneurship is even more subject to the caveat about difficulties of quantification and measurement than the measures discussed in the text. 24. Griffith et al. (2004) provide evidence on the importance of R&D for technological catch-up and innovation. 25. For empirical evidence on complementarities between IT expenditure and spending on IT personnel, see Kaiser (2003). 26. See Vandenbusscheet al. (2004) and Aghion et al. (2005) for detailed results and Aghion and Howitt (2005) for a summary. 27. See Annenkov and Madaschi (2005), Table 6 for detailed results. 28. See, for example, Alesina et al. (2005) and Nickell (1996). 29. See Aghion and Howitt (2005) for a summary of the recent evidence. 30. See Figure 6 in Annenkov and Madaschi (2005) for details. 31. See, for example, European Commission (2002) or OECD (2005).
References Achs, Z., D. Audretsch, P. Braunerhjelm and B. Carlsson (2005) “Growth and Entrepreneurship: An Empirical Assessment,” Discussion Paper no. 3205 on Entrepreneurship, Growth and Public Policy, Max Planck Institute of Economics. Aghion, P., L. Boustan, C. Hoxby and J. Vandenbussche (2005) “Exploiting States’ Mistakes to Evaluate the Impact of higher Education on Growth,” mimeo, Harvard University.
9780230_547018_10_cha08.indd 293
3/19/2009 8:16:32 PM
294
Designing the European Model
Aghion, P. and P. Howitt (2005) “Appropriate Growth Policy: A Unifying Framework,” mimeo, Harvard University. Alesina, A., S. Ardagna, G. Nicoletti and F. Schiantarelli (2005) “Regulation and Investment,” Journal of European Economic Association 3, 791–825. Annenkov, A. and C. Madaschi (2005) “Labour Productivity in the Nordic EU Countries. A Comparative Overview and Explanatory Factors 1980–2004,” Occasional Paper no. 39, October, European Central Bank. Ark, B. van and M. Piatkowski (2004) “Productivity, Innovation and ICT in Old and New Europe,” GGDC Research Memorandum GD-69, March. Audretsch, D. P. and M Keilbach (2004) “Entrepreneurship Capital and Economic Performance,” Discussion Paper no. 0104 on Entrepreneurship, Growth and Public Policy, Max Planck Institute of Economics. Barro, R. J. (1997) Determinants of Economic Growth: A Cross-Country Empirical Study, MIT Press, Cambridge, MA. Barro, R. J. and X. Sala-i-Martin (2003) Economic Growth, 2nd edition, MIT Press, Cambridge, MA. Bertola, G. (1994) “Flexibility, Investment and Growth,” Journal of Monetary Economic 34, 215–38. Boone, J. (2000) “Technological Progress, Downsizing and Unemployment,” Economic Journal 110, 581–90. Conway, P., V. Janod and G. Nicoletti (2005) “Product Market Regulation in OECD Countries, 1998 to 2003,” OECD Economics Department Working Paper no. 419. http://www.olis.oecd.org/olis/2005doc.nsf/linkto/ECO-WKP(2005)6. EEAG (2001) Report on the European Economy 2001, CESifo, Munich, Germany. EEAG (2002) Report on the European Economy 2002, CESifo Munich, Germany. EEAG (2004) Report on the European Economy 2004, CESifo, Munich, Germany. EEAG (2005) Report on the European Economy 2005, CESifo, Munich, Germany. EU (2004) Facing the Challenge, the Lisbon Strategy for Growth and Employment, Report from the high Level Group Chaired by Wim Kok, http:// www.europa.eu.int/lisbon_ strategy. European Commission (2002) Report on the state of the Internal Market for Services, IP/02/1180, Brussels. Griffith, R., S. Redding and J. van Reenen (2004) “Mapping the Two Faces of R&D: Productivity Growth in a Panel of OECD Countries,” Review of Economics and Statistics 86, 883–95. Hall, R. and C. I. Jones (1999) “Why Do Some Countries Produce So Much More Output per Worker Than Others?” Quarterly Journal of Economics 114, 83–116. Honkapohja, S., E. Koskela, W. Leibfritz and R. Uusitalo (2009) Economic Prosperity Recaptured: Finland’s Path from Crisis to High Growth, MIT Press, Cambridge. Jones, C. I. (2002) Economic Growth 2nd edition, Norton Company, New York City. Jorgenson, D. and Z. Griliches (1967) “The Explanation of Productivity Change,” Review of Economic Studies 34, 249–83. Kaiser, U. (2003) “Strategic Complementarities between Different Types of ICTExpenditures,” ZEW Discussion Paper no. 03–46. Mankiw, N. G., D. Romer and D. N. Weil (1992) “A Contribution to the Empirics of Growth,” Quarterly Journal of Economics 107, 407–37. Nickell, S. (1996) “Competition and Corporate Performance,” Journal of Political Economy 104, 724–46. Nickell, S. (2003) “Labour market institutions and unemployment in OECD countries,” CESIfo DICE Report 2/2003.
9780230_547018_10_cha08.indd 294
3/19/2009 8:16:32 PM
Economic Growth in the European Union 295 Nicoletti, G. and S. Scarpetta (2003) “Regulation, Productivity and Growth: OECD Evidence,” Economic Policy 36, 11–72. OECD (2005) Enhancing the Performance of the Services Sector, OECD, Paris. Timmer, M., G. Ypma and B. van Ark (2003) “IT in the European Union: Driving Productivity Divergence?” GGDC Research Memorandum GD-67, October. Vandenbussche, J., P. Aghion and C. Meghir (2004) “Growth, Distance to Frontier and the Composition of Human Capital,” mimeo, Harvard University and University College London.
9780230_547018_10_cha08.indd 295
3/19/2009 8:16:32 PM
9 Global Imbalances
9.1
Introduction
The large and persistent current account deficits run by the United States from the second half of the 1990s have generated widespread concerns about the sustainability of current macroeconomic imbalances at the global level. To what extent is the US trade deficit sustainable? If not, what will global adjustment require? In particular, to what extent will the dollar depreciate? Will adjustment lead to global recession? What are the appropriate fiscal, monetary and financial policies to minimise the risks of disruption? Many observers (for example, Roubini and Setser 2004a, b, 2005a, b) fear that the correction of global imbalances could lead to a period of disorderly adjustment, characterised by turmoil in currency and asset markets, a slowdown in economic activity, and ultimately large welfare costs for the world economy as a whole. Currently, large external deficits in the US are matched by large surpluses in Japan, Asian emerging markets, oil-producing countries and a few European countries. However, the euro area as a whole is close to external balance. In light of this, the question is whether adjustment of global imbalances will affect Europe only marginally, since the heart of it will consist in rebalancing the position of the US vis-à-vis the surplus regions, especially Asia. The goal of this chapter is to review the current debate on the causes and nature of global imbalances, assess policy options currently on the table, and more specifically discuss the implications of global adjustment for the European economy and European policy-making.1 We argue below that despite the close-to-external-balance position of Europe a rebalancing of the US deficits will create major policy challenges to European policy-makers.
9.2
Basic facts about global imbalances
We begin by considering three facts defining the nature of current global imbalances: the size and persistence of the US current account deficit; 296
9780230_547018_11_cha09.indd 296
3/19/2009 8:19:24 PM
Global Imbalances 297
Box 9.1 External imbalances, the current account balance and the balance of payments To understand global imbalances, it is useful to keep in mind that there are three ways of looking at the current account balance. First, the current account balance of a country is the sum of the trade balance (exports minus imports of goods and services), income from foreign assets held by residents (net of interest paid on the country’s foreign liabilities), and net labour income from residents working abroad: Current account balance = Trade balance (goods and services) + Income from net foreign assets + Net labour income This accounting relation makes it clear that the accumulation of foreign wealth by a country is related to its capacity to generate positive net exports of goods and services, income from capital lent abroad and labour services supplied by domestic residents employed in a foreign country. Movements of the current account are usually dominated by the trade balance component, but not to the extent suggested by national accounts. The reason is that official statistics include income payments from net foreign assets, but do not account for capital gains and losses on such assets. At times, these may be substantial (throughout this chapter we use the dataset constructed by Philip Lane and Gian Maria Milesi-Ferretti, who reconsider current account balance and portfolio positions accounting for capital gains and losses on foreign assets and liabilities). Second, the current account is also equal to the difference between national saving and investment: Current account balance = Saving – Investment = Private saving + Public saving – Investment This accounting relation makes it clear that external imbalances result from intertemporal choices by firms and households regarding how much to consume and invest in the current year as opposed to future years, as well as by government decisions about the size of the budget deficit (that is, the time profile of taxes and expenditure). Clearly, for the world economy as a whole, saving must equal investment, although the sum of current account balances rarely adds up to zero because of statistical errors. Third, the balance-of-payments identity equates surpluses in the current account balance to the accumulation of net foreign assets, recorded in the capital account: Current account balance = Increase in private net foreign assets + Increase in official net reserve holdings (Official reserve settlement balance) This identity makes it clear that current account deficits must be financed by capital inflows and/or changes in the stock of reserves held by monetary authorities. So, a current account deficit by one country generates a demand for foreign
9780230_547018_11_cha09.indd 297
3/19/2009 8:19:24 PM
298
Designing the European Model
capital, which must be matched by the portfolio decisions of foreign private agents and public institutions to acquire assets issued by that country. The above are not three alternative views of the current account balance; they are three identities. A country is solvent when, at the market interest rate, the present discounted value of future surpluses of the balance of trade in goods and services and net income from labour supplied abroad, is not smaller than the current net value of liabilities: Value of net foreign liabilities ≤ Present discounted value of surpluses in the current account excluding income from net foreign assets So, if a country is currently running current account deficits and is a net debtor, foreign lenders expect that country to generate positive net exports (in trade and services) and net labour income in the future, corresponding to a positive difference between net saving and investment. If this were not the case, some creditor country must be willing to finance the entire interest bill by the debtor country in the indefinite future. The debtor country would be playing a so-called Ponzi game: it would try to finance the interest bill on an ever-increasing stock of debt by further borrowing. In the international financial markets, the supply of assets by a borrower must be matched by the world demand for them. Sustainability of a country’s debt cannot be defined independently of the equilibrium structure of international portfolios. In particular, given demand and supply conditions, the equilibrium price at which a country’s debt is traded determines the risk premium that a country must pay on its external liabilities. The main challenge in understanding sustainability of external debt thus consists in understanding the factors underlying the desired portfolio composition by international investors. Recent episodes of financial and currency crises have arguably pointed to the possibility that frictions in financial markets may cause sudden changes in asset demands and the emergence of binding constraints limiting the extent to which a country can borrow. This could be the case, for instance, if co-ordination problems among international investors long in short term debt issued by a country lead to liquidity runs similar to bank runs. When markets co-ordinate on an equilibrium characterised by a run, the debtor country is forced to come up quickly with external resources to close any “financing gap” which may result. International runs (panics) can easily have potentially high costs in terms of consumption, output and overall welfare for international debtors.
the rising share of official capital flows from emerging markets and Japan, and the increasing importance of “valuation effects” of exchange rate and asset prices movements in determining the real burden of a country’s external debt. Throughout this chapter we will complement data from traditional sources (such as OECD and IMF) with the dataset on the “Wealth of Nations” computed by Lane and Milesi-Ferretti (2006). The distinctive feature of this data set is that (estimated) capital gains and losses on the
9780230_547018_11_cha09.indd 298
3/19/2009 8:19:24 PM
Global Imbalances 299
external portfolios of financial assets and liabilities have been taken into account when calculating net foreign asset positions. 9.2.1 The size and persistence of the US current account deficit The single most quoted fact characterising global imbalances is the size of the US current account deficits (relevant definitions are presented in Box 9.1). The US current account deficit grew from 1.6 per cent of US GDP in 1997, to 4.2 per cent in the year 2000. It kept increasing afterwards: at the time of writing, the 2005 deficit is estimated at above six per cent of US output, around 800 billion dollars. To get a sense of how large the US current account deficit is, consider that the US has about 110 million households: thus, an 800 billion dollar deficit means that, during 2005, the net external debt per household increased by 7,200 dollars! As a result of current account deficits, and changes in the values of US assets and liabilities, the estimated value of US net debt at the beginning of 2004 was around 23 per cent of US GDP. If the US keeps borrowing at the current rate, the external net debt of the US could approach 100 per cent of GDP in about a decade. Figure 9.1a shows the evolution of the current account balance in percentage of world GDP between 1996 and 2004, for the US, the euro area, Japan, Switzerland and the Nordic countries, Asian emerging markets, and oil-producing countries. In terms of world GDP, the US external deficit grew from less than half a per cent before 1997 to 1.6 per cent in 2004. Figure 9.1a emphasises that all other regions shown have recently been in surplus. In 2004, the US deficit was as high as 668 billion dollars. The combined surplus of Asian emerging markets and oil-producing countries (358 billion dollars) accounts for more than 50 per cent of it. Japan’s surplus (172 billion dollars) accounts for about 25 per cent; the surplus by Norway, Sweden and Switzerland (141 billions) accounts for another 20 per cent. The small positive current account surplus for the euro area accounts for the remaining five per cent of US deficit. From a global perspective, the US is borrowing resources from all the other five regions in the diagram, including regions at a relatively early stage of industrialisation. However, the US is not the only industrial country to run a current account deficit. Persistent imbalances are also run by Australia and New Zealand (together they borrowed 45 billion dollars in 2004), and by a few countries in the euro area. The deficits by Greece, Portugal, Spain and the United Kingdom altogether totalled 124 billion dollars in 2004. Italy’s deficit has also been increasing steadily, reaching 15 billion dollars in 2004, and showing no sign of reduction in 2005. While deficits by some of these countries are not too distant from the US if, for instance, net borrowing is
9780230_547018_11_cha09.indd 299
3/19/2009 8:19:24 PM
300 Designing the European Model
1.0
%
% Emerging Asia
Japan
0.5
Swi + Nordics
1.0 0.5 0.0
0.0 Euro area
Oil exporters –0.5
–0.5
–1.0
–1.0 –1.5
–1.5 United States
–2.0
–2.0 1996
Figure 9.1a
1997
1998
1999
2000
2001
2002
2003
2004
Current account balances (as a percentage of world GDP)
Note: The emerging Asia group includes China, Hong Kong, SAR, Taiwan, Korea, Malaysia, Singapore and Thailand. The Swi + Nordics group includes Norway, Sweden and Switzerland. Oil exporters includes Algeria, Bahrain, Egypt, Iran, Jordan, Kuwait, Lebanon, Libya, Saudi Arabia, Syria, United Arab Emirates and Yemen. Source: Lane and Milesi Ferretti (2006).
6.0
%
% 6.0 Japan
4.0
4.0
Emerging Asia Swi + Nordics
2.0
2.0 Oil exporters
0.0
0.0 –2.0
–2.0 Euro area
–4.0
–4.0
–6.0
–6.0 United States
–8.0
–8.0
–10.0
–10.0 1996
Figure 9.1b
1997
1998
1999
2000
2001
2002
2003
2004
Net foreign assets (as a percentage of world GDP)
Note: The emerging Asia group includes China, Hong Kong, SAR, Taiwan, Korea, Malaysia, Singapore and Thailand. The Swi + Nordics group includes Norway, Sweden and Switzerland. Oil exporters includes Algeria, Bahrain,Egypt, Iran, Jordan, Kuwait, Lebanon, Libya, Saudi Arabia, Syria, United Arab Emirate and Yemen. Source: Lane and Milesi Feretti (2006).
9780230_547018_11_cha09.indd 300
3/19/2009 8:19:24 PM
Global Imbalances 301
scaled by the size of the population of GDP, their sizes are typically small in absolute terms. 9.2.2 International reserves accumulation and the rising share of official financing after 2000 An important change in the composition of external financing of the US deficit occurred around 2000. From 1997 to 2001, that is, between the Asian crisis and the end of the period of asset market exuberance, private investors mostly financed the US deficits as they systematically rebalanced their portfolios in favour of US dollar-denominated assets, especially equities. The importance of private inflows has diminished substantially since 2001. As shown in Table 9.1, in the period 1997–1999 average net capital inflows into the US were as high as 176 billion dollars per year: private inflows accounted for 90 per cent of the total. In 2003 and 2004, average net capital inflows into the US were as high as 573 billion dollars: the share of private capital dropped to 40 per cent. In terms of net flows, the US external imbalance is now mainly financed through foreign official lending, in large part corresponding to the build-up of official reserves by five Asian economies: Japan, China, Hong Kong, Taiwan and South Korea. The above observation stresses a second striking dimension of current global imbalances, that is, the high level of international reserves in the form of dollar assets. Official reserves held by the five Asian countries mentioned above (consisting to a large extent of dollar-denominated assets) grew from 1.16 trillion dollars in 2000 to 2.66 trillion dollars in 2004. China is reported to own reserves up to 800 billion dollars in the last months of 2005 (see Genberg et al. 2005 for a detailed analysis of reserves policies).
Table 9.1 Private and official financing of the US current account deficit (in billions of dollars)
Current account deficit Total net capital inflows Net private inflows Of which: Direct investment Portfolio investment Net official inflows
1997–1999
2000–2002
2003–2004
218 176 160
427 462 404
594 573 233
34 126 15
38 366 59
–109 343 340
Source: Own calculations based on Congressional Budget Office (2005).
9780230_547018_11_cha09.indd 301
3/19/2009 8:19:24 PM
302
Designing the European Model Table 9.2 Official reserves in percent of total foreign liabilities for developing countries 1980–84 1985–89 1990–94 1995–99 2000–04
14.7 13.5 15.6 18.6 26.4
Source: Own calculations based on Congressional Budget Office (2005).
%
%
35
35
30
30
25
25
20
20
15
15
10
10 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004
Figure 9.2 countries
Official reserves as a percentage of total foreign liabilities, developing
Note: Ratio of official reserves in per cent of total foreign liabilities for the group of non-industrial countries. Source: Lane and Milesi Ferretti (2006).
From a more general perspective, Table 9.2 and Figure 9.2 show the growth of total official reserves by developing countries in per cent of their liabilities: official reserves have grown from 15 per cent in the 1980s, to an average of 26 per cent after the year 2000, up to 32 per cent in 2004. Now, developing countries pay a high risk premium on their liabilities but earn a low interest rate on their official reserves. Since in 2004 foreign official reserves accounted for about one-third of developing countries’ foreign liabilities, one out of three dollars that relatively poor countries borrowed from rich countries at high interest rates was thus lent back to rich countries at relatively low interest rates. Rodrik (2006) estimates that the financial
9780230_547018_11_cha09.indd 302
3/19/2009 8:19:25 PM
Global Imbalances 303 %
%
120
120
100
100 Gross external liabilities
80
80
60
60
40
40
Gross external assets
20
20 Current account balance
0 –20
0 –20 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98 00 02 04
Figure 9.3 US current account balance versus US stock of foreign assets and liabilities (as a percentage of US GDP) Source: Lane and Milesi Ferretti (2006).
cost of holding reserves is now currently close to one per cent of developing countries’ GDP. 9.2.3 Financial globalisation and the increasing importance of capital gains and losses due to exchange rate movements The emergence of external imbalances at the end of the last decade occurred in the context of a strong expansion of cross-border holdings of financial instruments. Indeed, in terms of world GDP, the total stock of foreign assets (= liabilities) in the world is currently above 120 per cent, twice as much as at the beginning of the 1990s. So, while the US current account deficit is large in terms of the US GDP, it is small relative to the stock of US foreign assets. This point is clearly shown by Figure 9.3, plotting the US current account between 1970 and 2004 together with the stock of US gross external assets and liabilities. In 2004, the US “owed” more than 100 per cent of its GDP to foreigners, but also owned claims to foreign output equivalent to about 80 per cent of its own output. The difference is the US net debt. The large expansion of gross portfolio holdings is responsible for a third, important dimension of current global imbalances. The change in net external debt between two dates depends not only on the flow of net exports during the period but also on changes in the value of the country’s foreign assets and liabilities, reflecting both asset price and exchange rates movements: the larger the stock of foreign assets and liabilities, the larger are these “valuation effects.” Valuation effects in the asset markets were clearly
9780230_547018_11_cha09.indd 303
3/19/2009 8:19:25 PM
304 Designing the European Model
much smaller in previous episodes of current account adjustment, before capital liberalisation had led to large cross-border holdings of financial assets. Paradoxically, in a financially globalised economy, a country with a balanced current account may be subject to large swings in its net external position, because of price fluctuations in financial and currency markets. Figure 9.1b shows the evolution of net foreign assets for the same regions as in Figure 9.1a. As mentioned above, the figures underlying this graph are from the dataset on the “Wealth of Nations” by Lane and Milesi-Ferretti (2006), specifically built to account for capital gains and losses on the external portfolios. Figure 9.1b shows that, in the second half of the sample period, both the euro area and the US had negative net asset positions (they are net debtors), whereas the other macro regions had positive net asset positions (they are net creditors in the world economy). Comparing Figures 9.1a and 9.1b, one can appreciate an apparent anomaly in the evolution of the net foreign asset position of the US and the euro area after the year 2002. After 2002, the current account of the US is negative and large, yet its foreign asset position was stable, even improving. The euro area has been running a surplus, yet its foreign asset position has worsened. This observation provides a striking illustration of the effects of capital gains and losses from exchange rate movements on external imbalances. The US typically borrows from international markets by issuing dollardenominated assets but lends abroad mostly by acquiring equities and foreign-currency denominated bonds. About two-thirds of US foreign assets are denominated in foreign currency. Because of this particular currency composition of the US external portfolio, a falling dollar has an automatic stabilising effect on the real value of the US net foreign debt. Dollar depreciation leaves the dollar value of US foreign liabilities unaffected but raises the dollar value of US foreign assets: overall, the US owes less to the rest of the world in terms of real resources (see the discussion in Tille 2003). To get a sense of the potential magnitude of these effects, recall that, at the end of 2004, US gross foreign assets were about 80 per cent of US GDP. As two-thirds of these were denominated in foreign assets, a 10 per cent depreciation of the dollar would reduce US net debt by 0.1 * 0.8 * 2/3, equivalent to more than five per cent of US GDP: approximately the size of the US external deficit!2 Note that these gains are increasing in the size of US gross assets, independently of US gross and net debt. The above net gains are however calculated ex post, that is, for given stocks of assets and liabilities. The gains are clearly smaller if currency depreciation is anticipated by financial markets. If this is the case, ex ante US interest rates would rise relative to foreign ones. This would raise the growth rate of liabilities in terms of GDP, so that depreciation-related capital gains on US assets would be at least in part compensated by a higher stock of US gross debt. To complete our back-of-the-envelope calculations, suppose that, at
9780230_547018_11_cha09.indd 304
3/19/2009 8:19:25 PM
Global Imbalances 305
Box 9.2 Real return differentials in favour of the US The US enjoys an important advantage in international capital markets. Historically, the rates of return earned by the US on its external assets are above the rates of return the US pays on its liabilities. For instance, taking five-year averages, the return differentials in favour of the US between 1995 and 2004 vary between one and approximately eight percentage points. What explains these positive return differentials? Potential explanations include the following. First, a large share of US foreign assets consists of equities, while US liabilities consist mostly of debt instrument with a large short-term component: the US benefits from the fact that equities earn a premium over bonds. Second, since dollar-denominated bonds are traded in deep liquid markets, they earn a liquidity premium: the US can borrow at particularly low interest rates. Third, dollar monetary assets are an important component of international liquidity, providing the US with seigniorage revenue. Because of positive rate of return differentials, total income from net foreign assets earned by the US was still positive in 2005, even though the country is a large net debtor in the world economy. This may soon change as the stock of US liabilities keeps increasing. The fact that the US has long earned a positive income from its net foreign assets has recently been used to suggest the following provocative thesis: because an international debtor should pay an interest income to its creditor; the fact that the US is actually receiving income from abroad means that it cannot be a net debtor. By way of example: in 2004, the US earned 300 billion dollars. Capitalised at 5 percent, this means that the US should actually have positive net foreign wealth of 600 billions (see Cline 2005 and Hausmann and Sturzenegger 2006). But how can the US be a net foreign creditor after running large current account deficits for so many years? Hausmann and Sturzenegger call the difference between recorded US net debt and their estimate of positive US net wealth “dark matter”. Dark matter is a colourful label pointing to under-reporting of US exports of knowledge (via foreign direct investment), liquidity (the US issues widely traded assets) and insurance (dollar assets are sought after as official reserves). The existence of “dark matter” implies that the value of US foreign assets and liabilities must be well above their market value (estimated by Lane and Milesi-Ferretti 2006). Does “dark matter” exist? If it does, there would be no need for global adjustments of the type discussed in the text. Unfortunately, the calculations underlying estimates of “dark matter” completely ignore the basic fact that different assets can pay different rates of return, depending on their risk and maturity. Why should one apply the same rate of capitalisation (5 percent) to all the assets and liabilities in the US external portfolio? Moreover, the debt figures presented in this chapter are all estimated taking into account market valuation in stocks and bonds markets, as well as exchange rates in the currency markets. Why should one distrust market prices completely, and put one’s faith on a simple capitalisation exercise at an arbitrary rate of return?
9780230_547018_11_cha09.indd 305
3/19/2009 8:19:25 PM
306
Designing the European Model
the beginning of 2004, markets attached a 25 per cent probability to a ten per cent fall in the value of the dollar by the end of the year. Abstracting from any risk premium, one-year interest rates on US liabilities would have increased by 2.5 percentage points. Now, at the beginning of 2004 the stock of US gross liabilities was close to 95 per cent of GDP (clearly higher than the stock of assets). Assuming for simplicity that all US liabilities had one-year maturity, higher interest rates would have raised the stock of US gross debt by an extra 2.4 percentage points, halving the net ex-post gains from dollar depreciation.3 Moreover, it is well known that US foreign assets have a large equity component. This component exposes the US to market risks, due to sizeable changes in asset prices that may accompany rebalancing. In fairly extreme scenarios of the adjustments (for example, a worldwide recession), it is likely that there would be significant declines in equity values. Then, for a given exchange rate, US assets abroad would fall in value relatively more than foreign investors’ holding of US assets. This would seem to offset some of the advantages that the US has from exchange rate depreciation.
9.3
What has caused the current imbalances?
There are a number of views on the causes of current imbalances, with quite different implications for the need for corrective policy measures and different predictions about the costs of adjustment. In this section we briefly discuss a representative set of these theories, grouping them under four main headings depending on their focus: insufficient US savings, productivity growth differentials, excessive savings outside the US, and exchange rate policies pursued by Asian countries. 9.3.1 Low US saving A widespread view attributes the persistent US current account imbalances to structural factors and policies lowering US national savings. As is well known, private savings in the US have been trending downward for quite some time. Possible factors likely to have influenced this trend include population ageing; structural changes in financial markets, boosting the use of credit cards and consumer credit; and, in the framework of deregulated credit markets, the recent strong dynamics of housing prices. Last, but not least, private spending in the US has long been sustained by easy monetary policy. While the US monetary stance has been progressively tightened in 2005 (see Section 1.2 and Figure 1.9 in Chapter 1 of the EEAG report 2006), long-term interest rates have remained low. For this reason, and because of tax advantages benefiting mortgages, the US housing market remained overall strong through 2005, helping to offset the demand effect of the monetary contraction.4
9780230_547018_11_cha09.indd 306
3/19/2009 8:19:25 PM
Global Imbalances 307
US private savings have, however, followed a rather stable pattern, compared to US public savings, which have deteriorated markedly since 2000 (see Chapter 1 of the 2006 EEAG report). The relaxed fiscal policy adopted by the Bush administration has been blamed for worsening the external position of the country, when other factors (essentially, exuberant expectations in the asset market) were no longer influencing domestic spending and international investment. Recent imbalances would thus validate the “twin deficits” hypothesis, that is, the idea that fiscal shocks raising the budget deficit also raise the current account deficit. An important question is whether the recent US budget and current account deficits are efficient ways to finance the costs of the wars in Afghanistan and Iraq, as well as the costs of dealing with terrorist threats and unexpected events such as the Katrina hurricane in 2005. Through domestic and foreign borrowing, US residents can in fact smooth their consumption and investment in the case of government spending hikes, avoiding highly distortionary peaks in tax rates. The argument of tax and consumption smoothing has strong theoretical foundations. Yet the implied benign view of the US external imbalance is not warranted. In particular, the argument disregards the basic fact that most of the US budget deficits result from tax cuts which the Bush administration has been struggling to make permanent. While current tax cuts mainly benefit current generations, future generations will have to service the interest bill on the higher domestic and external debt. The argument of efficient tax smoothing in the face of temporary spending hikes does not apply. Rather, what is at stake is re-distribution across income classes and across generations in a direction that amplifies long-run fiscal and macroeconomic concerns about population ageing (see Chapter 12). The view attributing external imbalances to low public savings in the US has been recently challenged by some researchers, who point out that fiscal shocks and autonomous changes in spending appear to have only limited quantitative effects on the current account. According to that argument, the impact of fiscal shocks on US investment and saving is so high that at the margin only 20 cents out of each deficit-financed dollar translates into excess demand for foreign imports (see Bussière et al. 2005; Chinn and Ito 2005; Erceg et al. 2004 and Ferguson 2005 among others). Some authors go as far as to question the validity of the “twin-deficit hypothesis” altogether (see Kim and Roubini 2003). However, even if current fiscal changes that increase budget deficits had no sizeable contemporaneous effects on the current account, budget deficits would still have important effects on the sustainability of the US current account. As argued by Corsetti and Mueller (2005), the return to capital in an open economy generally responds to fluctuations in the real exchange rate: fiscal shocks leading to real appreciation lower the return to current investment and hence cause crowding-out effects. To the extent that fiscal
9780230_547018_11_cha09.indd 307
3/19/2009 8:19:25 PM
308 Designing the European Model
deficits crowd out private investment, a lower stock of capital in the future would reduce the ability of the US economy to meet its external interest bill without reducing domestic consumption. In other words, consumption of goods or leisure would have to be cut in the future to service foreign debt. Thus, whether or not a policy correction to the US fiscal stance has an immediate impact on the US external trade, greater fiscal discipline would surely help reduce imbalances in an intertemporal perspective.5 9.3.2 Expectations of sustained US productivity growth A different argument emphasises expectations of sustained productivity growth differentials in favour of the US. Expectations of high productivity growth have arguably played an important role in generating strong US domestic demand in the second half of the 1990s, while making investment in the US relatively attractive to foreigners. The question is whether and to what extent this factor is still important.
Box 9.3 Useful exchange rate and international relative price definitions The nominal exchange rate is the price of one currency (the dollar) in terms of another (the euro). In January 2006, one euro is worth approximately 1.20 dollars. An increase in this figure would correspond to a nominal appreciation of the euro (one euro buys more dollars), that is, a nominal depreciation of the dollar. The external value of a currency can be calculated with reference to many currencies. In this case one talks of a multilateral (as opposed to a bilateral) exchange rate. Multilateral effective exchange rates are calculated as weighted averages of bilateral exchange rates (the euro against the dollar, the yen, the sterling pound etc.), weighted by importance of foreign trade with different trade partners. Alternative weighting schemes can be based on GDP or financial portfolios. The real exchange rate is the price of consumption in one country relative to the price of consumption in another country. A real depreciation (or a depreciation in real terms) indicates that the consumption basket in one country become less expensive relative to that in another country (or group of countries). According to its definition, the real exchange rate is calculated using consumer prices. As an indicator of competitiveness, it is sometimes calculated using producer prices or labour costs (usually per unit of product). The consumer price index includes the price of both goods that are traded internationally and goods that are not traded internationally (commonly refereed to as non-tradables or non-traded goods). A good is not traded internationally when, given technology and relative prices, its value is small relative to transportation and trade costs, so that its shipment abroad is not economically viable. Given the price of tradable goods, a fall in the price of non-tradables in a country (which lowers the domestic consumer price index) implies a real depreciation (that is a depreciation of the real exchange rate). The terms of trade are the price of exports relative to the price of imports. The terms of trade worsen, or deteriorate, when the price of imports rises, or the price of exports falls.
9780230_547018_11_cha09.indd 308
3/19/2009 8:19:25 PM
Global Imbalances 309
We have seen in the previous section that private capital inflows into the US have fallen after 2000. As shown by Table 9.1, the average net inflow of foreign direct investment actually turned negative after 2003, while foreign demand for US equities levelled off, mostly because of a shift in the demand by Europeans (Congressional Budget Office 2005). Figure 9.4 shows that, on balance, the stock of US net equity and cumulated FDI positions decreased rapidly in the 1990s and became negative after 2000; since then, however, it has been increasing again. This evidence is inconsistent with the hypothesis that international investors are currently “chasing investment opportunities” in the US, motivated by superior productivity performance. Nonetheless, expectations of high productivity and income gains in the future may be a factor underlying the low saving rates (discussed above), a point stressed by Ferguson (2005). In the same vein, Engel and Rogers (2005) focus on the US share in output produced by advanced countries (including the G7 plus Switzerland, Sweden and Norway).6 This share fluctuates between 38 and 40 per cent before 1990; it then trends upward, reaching 44 per cent in 2004. Most importantly, current forecasts imply a further increase along the same trend. According to the calculations by Engel and Rogers, the US is expected to account for 49 per cent of developed countries output by 2017. In light of such forecasts, current account deficits up to five to seven per cent of US GDP may be rationalised in terms of efficient consumption smoothing: US households are simply taking advantage of borrowing opportunities to consume part of their anticipated future income gains now. The view that US external deficits are essentially driven by expectations of high future growth in income has two important policy implications. First, it is not appropriate to talk about “imbalances,” as trade flows are in fact balanced in an intertemporal perspective. Running a deficit today, US residents are increasing current expenditure by borrowing resources from foreign residents, in exchange for future resources capitalised at the market interest rate. Second, little or no dollar depreciation in real terms may be required for some time. In the Engel and Rogers’ version of conventional open macro analysis, the dollar is strong (in real terms) during the phase of high external deficits, and will weaken once the US share in the advanced countries’ GDP stabilises. As stressed by these authors, such stabilisation will happen quite a few years from now. Moreover, when it comes, there would be nothing dramatic in the real dollar depreciation that will accompany the US current account reversal.7 But, as mentioned above, it is unclear why the optimistic forecasts of future growth driving US consumption would not also cause foreign direct investment into the US and/or acquisition of US equities by foreigners. Most importantly, current expectations about US differential growth may be too optimistic (after all, expectations systematically underestimated prospective US growth rates in the early 1990s). If and when expectations
9780230_547018_11_cha09.indd 309
3/19/2009 8:19:25 PM
310
Designing the European Model %
%
10 5
10 Net portfolio equity + FDI
5
0
0
–5
Net other claims
–5
–10
–10
–15
–15
–20
–20 Net portfolio debt –25
–25
–30
–30 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004
Figure 9.4 United States: Net portfolio debt and equity cum FDI positions (as a percentage of US GDP) Note: Net portfolio equity + FDI equals the difference between the sum of FDI and portfolio equity assests and the the sum of FDI and portfolio liabilities. Net other indicates the difference between the stock of other assests and other liabilities, and net portfolio debt assets and reserves and the stock of portfolio debt liabilities. Source: Lane and Milesi Ferretti (2006).
are revised downwards, restoring US external balance (in an intertemporal perspective) would require a sharp correction of spending plans, possibly implying large movements in exchange rates and relative prices (as discussed later on in this chapter). Note that the above view completely downplays the role of US government budget deficits in generating the current account deficits. 9.3.3 Excess saving outside the US: the “saving glut” or “investment drought” view Another view, which also downplays the idea of insufficient US savings, interprets the US current account imbalance as the mirror image of excess supply of saving in the rest of the world: according to Bernanke (2005), the US current account deficit is the counterpart of a global saving glut. This corresponds to an increase in saving in excess of investment in emerging markets after the series of currency and financial crises throughout the 1990s. The glut is magnified by rising surpluses in oil-producing countries that benefit from high oil prices. The saving glut is essentially caused by “self-insurance” policies pursued by many emerging-market economies to minimise the risks of future crises and liquidity runs. In practice, many countries have pursued macroeconomic policies that turned external deficits into surpluses, while building extremely large stocks of international reserves.
9780230_547018_11_cha09.indd 310
3/19/2009 8:19:25 PM
Global Imbalances 311
One may object that excess saving in emerging-market economies could be matched by relatively small deficits in all industrial countries, rather than by a large deficit in one country only. To address this objection, the saving glut view points to differences in the macroeconomic, legal and institutional environment in which national financial markets operate: because of these differences, international investors perceive US assets to have higher “quality” than the assets of other countries. For instance, asset quality depends on the extent to which investors’ rights are protected: since the US offers a high level of such protection, US assets are preferred by world savers looking for opportunities of portfolio investment and diversification. According to this view, the US deficits mainly depend on the fact that excess world savings are channelled preferentially to the US. An obvious problem with this view is that while it can rationalise the increasing role of monetary institutions in providing financing to the US as a consequence of self-insurance policies pursued by many governments in emerging markets, it cannot explain the increasing disaffection of private investors towards US equities (see Table 9.1 and Figure 9.4). While the saving glut idea may have had some merit before 2000, it needs to be refined to fit recent global portfolio patterns.8 Moreover, some observers (notably Roubini and Setser 2004a, b and 2005a, b) emphasise that the saving-investment imbalance outside the US is to a large extent due to abnormally low investment rates: thus, it should be labelled an “investment drought” rather than a “saving glut” (see also the evidence in Chinn and Ito 2005). With the exception of China and a few other countries, investment rates have fallen markedly across emerging markets.9 In South-East Asian economies, the drops have been as high as 10 percentage points of GDP from the peak in the first half of the 1990s. One may argue, however, that the rate of investment corresponding to those peaks was not sustainable (see Corsetti et al. 1999). Investment rates have also fallen in Japan and most noticeably in the euro area.10 Low capital accumulation may be due to the need by many corporations to clean up their balance sheets after the financial turmoil around 2000 (International Monetary Fund 2005) or simply to “animal spirits” of entrepreneurs. In either case, the “investment drought” view offers a potential explanation of the observed low levels of real interest rates. The standard textbook model predicts that an exogenous drop in investment demand indeed reduces the equilibrium rate of interest (by how much depends on the interest elasticity of savings). As an implication, one may expect interest rates to rise as soon as investment picks up again. 9.3.4 Chinese economic policy and Asian currency pegs The view that Asian emerging markets have substantially contributed to generating current global imbalances emphasises both a trade channel (related to the size of the Asian external surplus) and a financial channel
9780230_547018_11_cha09.indd 311
3/19/2009 8:19:26 PM
312
Designing the European Model
(related to the increasing weight of official lending by Asian countries in international net capital flows). The main focus is, however, on China’s exchange rate policy and its strong influence on the policies pursued by the other emerging markets in Asia. Formally, China abandoned its inflexible peg against the US dollar in July 2005, when it switched to a managed float, allowing the renminbi to fluctuate inside a small band around the dollar parity (see Box 1.2 in Chapter 1 of the EEAG report 2006). Despite such reform, the Chinese renminbi hardly appreciated through the second half of 2005. Reserve accumulation has kept outpacing the trade and FDI surpluses by a large amount.11 The dollar peg regime is an important element in China’s strategy to achieve rapid industrialisation, which also includes strict capital controls de-linking the domestic financial and banking sector from the rest of the world, thus allowing Chinese authorities to pursue country-specific credit policies and retain some control over domestic monetary policy.12 Standard growth models predict that a financially closed economy (such as China), converging to the higher income level of industrial countries should generate high investment and saving rates (see, for example, Cuñat and Maffezzoli 2004). Indeed, Chinese investment and saving rates are high by international standards: official sources reported gross investment to be 43 per cent of GDP in 2003 (recent GDP revisions, however, may lower this percentage significantly). Obviously, to generate external surpluses, China has engineered even higher saving rates. As the Chinese economy has been growing at a sustained rate of around 9 per cent per year, households can reasonably be expected to have a strong incentive to borrow against higher future income. How can extraordinarily high growth rates be reconciled with low consumption and excess saving? Traditional explanations point to credit policies pursued by Chinese banks, firmly directed towards the growth objectives of the government in terms of industrialisation and export. Recent views also stress that with the recent transformation of the business sector, Chinese employees face the need to finance their retirement, the education of their children, and health services, as state-owned companies no longer provide support in these areas (Blanchard and Giavazzi 2005; Chamon and Prasad 2006). With an increase in lifetime income uncertainty, high savings may correspond to an inefficiently high level of self-insurance. Finally, by worsening the country’s terms of trade, undervaluation of the exchange rate reduces the purchasing power, and therefore the wealth of domestic households. Overall, a strategy of export-led rapid industrialisation appears to be accompanied by policies discouraging domestic (consumption) demand. These considerations help address a rather puzzling feature of the Chinese dollar peg, that is, the extent to which Chinese authorities have managed to avoid overheating and relative price correction for so many years. Despite the high GDP growth rates, there has so far been little evidence of inflationary
9780230_547018_11_cha09.indd 312
3/19/2009 8:19:26 PM
Global Imbalances 313
pressure and overheating leading to revaluation in real terms: in 2004 overall CPI inflation rose significantly (reaching a peak as high as 5 per cent in the third quarter of the year), but it subsequently fell below 2 per cent in 2005. According to available statistics, wages and non-traded goods prices do not show appreciable changes. As often argued, an important reason has been an extremely elastic supply of labour (see, for example, Dooley et al. 2005). But in light of the arguments above, structural factors and policies containing domestic demand have also played a role. Overheating and inflation risk, however, is only one possible undesired effect of the Chinese exchange rate and export promotion policies. Price competitiveness as well as tax and credit incentives for exporting firms have arguably distorted the allocation of capital and employment. In this respect, some studies report that Chinese total factor productivity has fallen between the first and the second half of the 1990s. Blanchard and Giavazzi (2005) attribute such a fall mostly to misallocation, that is excessive investment in the export sector. The financial side of real distortions is excessive exposure of Chinese banks and financial institutions to low-return firms, whose profitability would be completely compromised were the exchange rate to appreciate. Many observers argue that the stock of non-performing loans by the state-owned Chinese banking system is already large: the persistence of distorted relative prices may bring it to quite dangerous levels. Rising financial risks imply a rising fiscal risk for the Chinese government. The policy pursued by China and other Asian countries has global implications for world demand and international prices. High rates of Chinese growth have raised world demand for some capital goods as well as for commodities, especially energy, which are necessary to sustain the expansion of infrastructure and productive capacity. On the other hand, excessive saving (relative to investment) has limited the Chinese contribution to the world demand for consumption goods at large, possibly reducing the relative price of consumption in terms of investment goods. This has had a selective impact on the profitability of investment in industrial countries (countries specialised in light manufacturing and consumer goods have obviously suffered the most). It is important to distinguish between long-run effects of the ongoing integration of China and other emerging markets into the world economy and the short- and medium-run effects of the exchange rate and macroeconomic policies described above. As regards the long run, classical trade theory offers precise predictions about the economic repercussions of integrating large regions with abundant labour and a small capital stock: as the global supply of labour (and especially of low-skilled labour) rises faster than global capital, the world economy will experience a fall in the relative price of (unskilled) labour relative to capital and a fall in the relative price of labour-intensive goods. The integration process could make unskilled workers in industrial countries worse off while making capital owners better off
9780230_547018_11_cha09.indd 313
3/19/2009 8:19:26 PM
314
Designing the European Model
than would otherwise be the case (the policy issues implied by these movements were addressed in Chapters 4 and 5). In the short and medium run, most estimates of the equilibrium exchange rates between China and the rest of the world point to undervaluation of the renminbi in real terms: estimates vary between 20 and 40 per cent. Undervaluation creates a cost advantage to Chinese exporters on top and above what is implied by their comparative advantages but also keeps Chinese terms of trade abnormally low and distorts the internal relative price between traded and non-traded goods. A 20 to 40 per cent real exchange rate appreciation would not wipe out export growth of China nor eliminate the need for adjustment in the production structure of industrial countries. Such a correction would nonetheless rebalance the Chinese macroeconomy in a decisive way. There are several reasons to expect revaluation of the Chinese currency in the coming months. First, after a period of extraordinary expansion of China’s manufacturing base, the benefit of further expansion may be low relative to the costs of allocation distortions (including environmental costs). Second, revaluation could prevent the build-up of protectionist pressures in the US, sheltering the Chinese government from charges of currency manipulation. While improving the US trade deficit, however, a renminbi revaluation may also be associated with a substantial slowdown in the rate of reserve accumulation, that is, it may reduce the contribution of official inflows to finance current US imbalances. It is unclear at what speed, if any, China will undertake some steps towards liberalisation of capital flows. Concerns about the health of its financial institutions may induce caution in exposing Chinese financial markets to the risks of volatile capital flows.13 At this stage, an asymmetric relaxation of controls on capital inflows is a relatively low-risk option because of widespread revaluation expectations. In general, capital controls are notoriously difficult to implement: they may become less and less stringent over time, or even force Chinese authorities to accelerate the pace of liberalisation of capital movements.
9.4 What does global adjustment require? Concerns about US current account imbalances are often played down by stressing that, thanks to financial globalisation, markets can finance increasingly large imbalances, and let “adjustment” proceed smoothly and gradually (an influential view voiced by Greenspan 2004). According to this argument, relative to the pre-globalisation world, market depth and efficiency reduce the need of domestic governments to implement policy corrections. As discussed above, financial globalisation has led to unprecedented cross-border holdings of foreign assets and liabilities, whose value fluctuates with the exchange rate. The importance of capital gains and losses in
9780230_547018_11_cha09.indd 314
3/19/2009 8:19:26 PM
Global Imbalances 315
determining the real burden of a country’s debt has led many observers to talk about a new “valuation channel” in the adjustment process, which complements adjustment via net exports. In the case of the US, the valuation channel functions as a shock absorber, providing an additional reason to expect a smooth resolution to the problem of stabilising the US external debt, as was discussed in Section 2.3. However, financial globalisation raises the possibility of large reversals in capital flows, because international investors (perhaps led by hedge funds and other large active players in international markets) may all at the same time attempt to shift out of US short- and medium-term bonds. In the presence of sudden capital flow reversals, current account and trade adjustment become central. If the US needs to increase its net exports over a short period of time, reducing imports and boosting exports will require a downturn in economic activity and/or a sharp downward movement of the US dollar. In this section we reconsider this debate and its implications for the external value of the dollar in detail.14 9.4.1 Real dollar depreciation: the goods market and the domestic demand channel What is the size of dollar real depreciation consistent with correcting US imbalances? In a series of papers, Obstfeld and Rogoff (2001, 2004, 2005) have addressed this question by focusing on the equilibrium relative price adjustment required to eliminate the US current account deficit (say, because a sudden reversal of capital flows prevents the US from rolling over its debt). The relative prices of interest include the terms of trade, that is the price of exports in terms of the price of imports, and the price of non-traded goods in terms of the overall CPI, or in terms of the price of internationally traded goods entering the US CPI. Relevant definitions of these prices are provided in Box 9.3 of this chapter. To focus sharply on movements in these relative prices, Obstfeld and Rogoff propose a stylised model calibrated to the US economy in which employment and capital in the traded goods and in the non-traded goods sectors, and therefore also the outputs produced in the two sectors, are kept constant. The authors study the changes in the consumption level, consumption composition (between the two types of goods) and relative income (the value of US output relative to the rest of the world) necessary to eliminate the current account deficit. The adjustment mechanism is as follows. To fill its external financing gap, the US needs to raise its net exports, that is, export more of the tradables produced and import less foreign tradables. This means that the demand for tradables by US firms and households must fall and that the demand for US tradables by the rest of the world must increase. Selling more US output abroad requires a drop in the relative price of US tradable goods in the world market. By definition this is a deterioration of
9780230_547018_11_cha09.indd 315
3/19/2009 8:19:26 PM
316
Designing the European Model
the US terms of trade. As traded output in the US and abroad is held constant in the calculation, the size of price adjustment will depend on the price elasticity of the world demand for US tradables. However, note that a fall in the price of US tradables per se would raise, instead of reduce, the demand for them by US firms and households. This is the reason why adjustment also requires an even larger fall in the price of US non-tradables, redirecting US demand towards these goods. As a result, real depreciation “switches” US consumption demand away from both US and foreign tradables, in favour of US non-tradables. This consumption “expenditure-switching” effect corresponds to a change in the composition of consumption. Moreover (and this is perhaps the most important point), once the dollar has fallen in real terms US households are poorer: the value of US nontradable output falls in terms of foreign goods, as does the value of US tradables (the value of the latter falls with the deterioration of the terms of trade). As US income falls relative to the rest of the world, US consumption also falls. In this model, a real depreciation thus causes a US income and demand slowdown. According to Obstfeld and Rogoff, most of the required adjustment in the US real exchange rate is attributed to the need for a fall in the relative price of US non-tradables. In extensive quantitative experimentation, these two authors calculate the overall depreciation of the dollar in real tradeweighted terms required to improve the US trade balance by about five percentage points of GDP. The required real rate of depreciation ranges between 15 and 34 per cent, depending on the elasticity of substitution between tradables and non-tradables as well as between domestic and foreign tradables. Conversely, adjustment in the terms of trade is quite contained, ranging between four and seven per cent. In other words, at most one-third of the adjustment can be attributed to adjustment in the international prices of US tradable goods. The size of adjustment estimated by Obstfeld and Rogoff is quite large, but not unusual as compared to the swings that major currencies have experienced over the last decades. Between its peak in 2002 and the end of 2005, the dollar depreciated in real effective terms by 24 per cent (based on IMF data). The corresponding nominal depreciation was as high as 31 per cent. A revaluation by China and other Asian countries will also contribute significantly to correcting the external value of the dollar in real effective terms.15 The time horizon for the correction makes a difference. In the above model, import demand from the US falls with a large real depreciation of the dollar, because a real depreciation (at constant output and employment) implies a contraction in US income relative to the rest of the world. But in the short run, adjustment in US external demand may well be driven by a slowdown in output and employment (see Edwards 2005). A contraction in
9780230_547018_11_cha09.indd 316
3/19/2009 8:19:26 PM
Global Imbalances 317
the production of non-traded goods would reduce, at the margin, the pressure on the exchange rate: with less non-tradables produced, their relative price will have to fall by less to match the increased demand by US households. However, if the slowdown spills over to the traded good sector (despite the favourable relative price movements), this will add to depreciation pressures. This is because with less tradables to exports, US imports must fall by more, creating the need for sharper depreciation in equilibrium. In the medium run, employment and capital allocation are bound to change (with consumption). First, part of the adjustment may take the form of an increase in hours worked and labour participation, instead of a drop in consumption (after all, a deterioration in the US terms of trade means that households are poorer relative to other countries: labour supply may increase in response to this negative income shock). Second, production will be re-allocated across sectors, in response to the increase in the relative price of tradables, raising the overall supply of US exports. As regards the dynamics of net exports and terms of trade, an important issue is the extent to which external adjustment will occur via an increase in the quantity of goods already exported, as opposed to an increase in the range of exports (that is, the extent to which adjustment will be at the “intensive” as opposed to the “extensive” margin). The main point here is that any exogenous shock to the ability of the US to borrow ultimately leads to a larger external demand for US goods. This can stimulate the supply of US goods and goods varieties that were not previously exported. In equilibrium, the terms of trade are likely to move less when new goods are exported than when adjustment only takes place at the intensive margin (exporting more of a given set of goods). In other words, adjustment at the extensive margins can further reduce or prevent altogether a fall in the terms of trade, reducing the required equilibrium real depreciation (Corsetti et al. 2005, 2006). Unfortunately, empirical studies on the US point to worrisome regularities regarding the response of US imports to exchange rate and income movements. Recent studies confirm the asymmetry between the US and other countries first noted by Houthakker and Magee (1969): the income elasticity of US imports well exceeds the income elasticity of other advanced countries’ imports from the US. Thus, a given fall in the imports-to-GDP ratio requires a much stronger income slowdown in the US than abroad. By the same token, US import price elasticities are quite low.16 Chinn (2005) finds that one sixth of US imports are apparently insensitive to exchange rate variations (although another finding is that non-oil, non-computer imports are much more sensitive to exchange rate changes than aggregate imports). An important lesson from these considerations is that adjustment will require a protracted period of real dollar weakness, but the magnitude of further dollar depreciation (that is whether and by how much the dollar should further depreciate) is quite uncertain.
9780230_547018_11_cha09.indd 317
3/19/2009 8:19:26 PM
318 Designing the European Model
The analysis by Obstfeld and Rogoff points to the need for further real depreciation, but their approach focuses on price elasticities between traded and non-traded goods, whose estimates in the literature vary markedly. Other complementary studies look into other adjustment margins: consumption, the level of employment, sectoral allocation of production, as well as the composition of exports (intensive versus extensive margins). The implications for equilibrium movements in the dollar real exchange rate are quite disparate. Second, the main reason for expecting sharp dollar depreciation, laid out in detail above, is that a sharp dollar fall is required to drive down the price of US non-tradables. So, while adjustment may well require large depreciations of the dollar in real trade-weighted terms, movements in the relative price of US exports may actually remain quite small. What needs to be sizeable is the correction in US net external demand, but not necessarily the correction in US export prices. 9.4.2 Dollar depreciation: portfolio valuation effects We have already observed that, because of the particular currency composition of US foreign asset and liabilities, dollar depreciation automatically reduces the real net-debt burden of the US.17 The larger the valuation effects from depreciation, the larger the fall in the real value of US net liabilities. Note that a fall in the dollar helps the US external position through two channels. The first is the traditional channel, through which dollar depreciation encourages US net exports, improving the competitiveness of US exporters, while discouraging US imports. The second channel consists of valuation effects, which raises the dollar value of US foreign assets, improving the net external position of this country. For this reason, valuation effects tend to reduce the magnitude of dollar depreciation required to achieve external adjustment, relative to the case in which only the first channel is active. For instance, in the model by Obstfeld and Rogoff described above, valuation effects from dollar depreciation could reduce the required rate of real effective dollar depreciation by about five percentage points. These effects may help address current imbalances but clearly are no substitute for net export correction (see Obstfeld 2004). Interestingly, as argued by Cavallo and Tille (2005), valuation effects may do more than reducing the overall magnitude of adjustment in trade: they can actually play a substantial role in smoothing the dollar decline along the path of adjustment. In the experiments by these two authors, the US is assumed to stabilise its stock of net external debt relative to GDP. In the long run, the required rate of real dollar depreciation is 27 per cent (in their baseline estimate). In the short run, sizeable capital gains on gross external liabilities allow US households and firms to sustain current imports, reducing pressures on the exchange rates. Hence dollar depreciation is below 10
9780230_547018_11_cha09.indd 318
3/19/2009 8:19:26 PM
Global Imbalances 319
per cent, and around 15 per cent in the first two years of external adjustment. An important difference between short- and long-run effects is that, over time, depreciation expectations driving interest rates raise the cost of debt, and therefore tend to reduce the overall magnitude of valuation effects (see Section 2.3 of this chapter). Nonetheless, some authors claim that advantages of valuation effects are a stable long-run feature of US borrowing, allowing the US to pay effectively negative returns on its net liabilities, a point forcefully stressed by Gourinchas and Rey (2005, 2007). In some respects, this position is consistent with the evidence on rate of return differentials in favour of the US, which imply that this country can borrow on better terms (discussed in Box 9.2 of this chapter). But the fundamental issue is the extent to which a debtor can count on valuation effects to depreciate systematically the real value of its liabilities. While asset pricing may have played a large role in adjustment historically, it is unlikely that a permanent net-debt devaluation strategy is sustainable. Sooner or later, international investors will price the risk of valuation effects, leading to higher interest rates on the US debt. So far, however, US long-term interest rates have not moved significantly away from comparable euro area rates. In the same spirit as a previous exercise by Krugman in the 1980s, Obstfeld (2005) looks at the return on inflation-indexed bonds issued in dollars and in euros: in 2005 the return differentials between the US and French 30-year debt instruments are just a few basis points. In other words, markets do not seem to attach any significant probability to the event of a sizeable real depreciation of the dollar. The dollar even appreciated during 2005 and earlier forecasts of sharp redirections of portfolio flows away from the US have so far been proven wrong.18 This is clearly a puzzle for those observers who believe that further dollar depreciation and international portfolio rebalancing is needed. If, as many believe, markets are indeed mispricing adjustment risks (it would not be the first time), the macroeconomic consequences of market ‘inattention’ can be quite far-reaching. To the extent that asset prices suggest to policy-makers that no correction is urgent or necessary, imbalances may keep growing, making adjustment much harsher and deeper in the future. 9.4.3 Demand policies (fiscal correction in the US) Even if revaluation of Asian currencies realign international relative prices in a way that is consistent with a reduction in the current global imbalances, it is doubtful that substantial correction will take place without appropriate demand policies. Conventional wisdom suggests that “expenditure-switching policies” (essentially, exchange rate revaluation by Asian countries and benign neglect by the US vis-à-vis a weak dollar) should be complemented by “expenditure changing policies.” In China, for instance, a renminbi revaluation should be accompanied by measures to sustain domestic demand (or at least
9780230_547018_11_cha09.indd 319
3/19/2009 8:19:26 PM
320 Designing the European Model
to remove current distortions that generate very high savings). A reduction of Chinese national saving would contribute to world demand, lowering this country’s surplus more than implied by the loss of “competitiveness” due to revaluation. It could also help contain the strain on the Chinese economy due to relative price changes. The most important contribution to adjustment should, however, come from a reduction in the US fiscal deficits, which requires a revision of tax policy. Without any fiscal rebalancing in the US, a reduction in Asian saving, possibly associated with a slowdown or reversal in reserve accumulation, increases the risks of financial strain in the global currency and asset markets, due to disorderly adjustment characterised by a loss of confidence in the dollar and financial turmoil. In principle, a pick-up in European demand could also provide a valuable contribution to global rebalancing. There is, however, considerable scepticism about such a possibility. In many European countries, private consumption growth has been persistently low (see Chapter 1 of the 2006 EEAG report), for reasons that are not entirely understood. As argued in Chapter 1of the 2006 EEAG report, high debt levels in the euro area in combination with future demographic strains imply that there is little or no room for fiscal stimulus in Europe.
9.5
A European perspective
This chapter has analysed various adjustment scenarios and policy options for reducing the macroeconomic risks associated with increasing global imbalances. Reducing the size of the US current account deficit does require an increase in US saving (both public and private) relative to the rest of the world, and/or an increase in spending in the rest of the world relative to the US. Depending on its intensity, adjustment may produce a slowdown in US growth and/or a prolonged period of dollar weakness in real terms. Both factors will contribute to a drop in US net imports. As is well known, the response of the trade balance to real exchange rate changes usually takes time: in the case of the US the lag is traditionally quite long (see Krugman 1991) for an analysis of the so-called J-curve effect in the US). While there is considerable uncertainty as to the timing and intensity of adjustment, most of the scenarios reviewed above have potentially negative consequences for the European macroeconomy. Europe is likely to face a further weakening of external demand for its products, as well as increasing competitive pressure from the US (although the overall consequences of adjustment for Europe will also depend on the intensity of policy correction in Asian countries: higher demand in these countries would obviously have some positive effect on European net exports). In addition, portfolio rebalancing, redirection of capital flows and the associated swings in asset prices
9780230_547018_11_cha09.indd 320
3/19/2009 8:19:26 PM
Global Imbalances 321
and exchange rates may have important effects on the value of European assets and liabilities, raising the likelihood of financial turmoil involving European firms and banks. We consider these arguments in detail below. To begin with, even though the overall European current account imbalance with the US is small relative to other macro regions, the external performance of individual European countries is quite diverse. The new EU members (as expected) run external deficits.19 Some old EU members (notably Germany) enjoy a strong export performance, while some others (notably Italy) have experienced a deterioration of their competitiveness. In this context, the real and financial dimensions of global adjustment are likely to have asymmetric effects on the European economy. Adjustment of global imbalances may or may not require further dollar depreciation vis-à-vis the euro. However, even if adjustment takes place with no further fall in the dollar and/or with limited movements in international prices, correcting the US current account deficit does require an improvement in US net exports. It follows that Europe is likely to experience a drop in external demand even if the associated exchange rate movements are not as large as many fear. In addition, an important question is whether the US will decide to reverse its attitude towards free trade, generating a new wave of protectionism. If this is the case, we may witness some reduction in trade among macro regions, with uncertain effects on the cohesion of the European economy. Through valuation effects, further dollar depreciation will make Europe poorer relative to the US. It is true that Europe has a small net foreign asset position, but the magnitude of valuation effects depends on the size of the stocks of gross assets and liabilities. With financial globalisation, these gross stocks are several times larger than the net asset position. These effects may also create asymmetric effects across Europe, depending on the size of a country’s total foreign assets and liabilities, as well as on the currency and maturity structure of these portfolios. The last few years have been characterised by very low long-term interest rates in real terms, lower than predicted by standard economic models (and conventional wisdom). This may change with the start of an adjustment process leading to a reduction of the US current account. The reversal of capital flows and portfolio allocation may lead to the emergence of larger risk premia (also within the euro area) and upward pressure on interest rates. The risks associated with high prices in the housing markets of many countries have been frequently discussed (see Chapter 5 of the 2005 EEAG Report). Increases in long-term interest rates could clearly cause substantial falls in housing prices. Further depreciation of the dollar in real effective terms, associated with rising interest rates and the emergence of interest rate differentials also among European countries, could clearly exacerbate business cycle and inflation differentials in Europe. Past experience and common sense suggest that
9780230_547018_11_cha09.indd 321
3/19/2009 8:19:26 PM
322 Designing the European Model
consumer prices and growth may respond more intensely to euro exchange rate movements in smaller and more open European economies than in the large economies. The resolution of current imbalances may well proceed rather smoothly. But it is also possible that the current build-up of imbalances will lead to ‘hard landing’ scenarios. What risks do European policy-makers face? Consider first the possibility of a disorderly adjustment, if and when international financial markets become unwilling to roll over their credit to the US. This means a US current account reversal associated with strong relative price and exchange rate movements, creating financial turmoil across markets: risk premia will rise markedly, housing markets may collapse, US demand could falter, and the dollar may fall dramatically. In this scenario, it is highly plausible that European financial and nonfinancial firms would suffer from strong deterioration of their balance sheet and liquidity shortages. This scenario would call for European monetary and supervisory authorities to stress-test their institutional framework. Chapter 11 analysed the regulatory and supervisory framework for European financial markets, assessing its effectiveness in intervening in defence of the European payment and financial system, and in reducing the liquidity costs for firms of financial turmoil. The report pointed out concerns related to the decentralised structure and complexity of the framework. Technically, interventions providing emergency liquidity to firms and financial markets do not need to compromise the ability of the ECB to retain control over aggregate liquidity in the euro area. Injections in one region could be compensated with opposite interventions somewhere else. On the other hand, if the magnitude of financial crisis is large enough to generate substantial uncertainty about default rates by firms and banks, monetary authorities may face difficult trade-offs between financial stability and price stability, as monetary interventions may not be effective in preventing widespread default. Governments may then have to shoulder large fiscal costs to avoid a chain of destabilising bankruptcies. This raises important issues about the distribution of possible fiscal costs across countries. The deteriorating public finances in many European countries, which were discussed at length in Section 3.3 in Chapter 1 of the 2006 EEAG report, are an aggravating factor in this context. Weak public finances may create undue constraints on emergency financing in the case of a crisis associated with a “hard” unwinding of global imbalances. This provides yet another argument for fiscal discipline now as a precaution against future financial crisis. In the event of a sharp correction of the dollar and a deep US recession, monetary authorities in Europe (the ECB and the national central banks in the countries outside the eurozone) must react to deflationary pressure (coming from likely falls in export prices and export volumes) by loosening the monetary stance. The timing of intervention will be an issue, as proactive pre-emptive interest rate cuts may be warranted in such a situation.
9780230_547018_11_cha09.indd 322
3/19/2009 8:19:26 PM
Global Imbalances 323
Overall, however, even if European monetary authorities are successful in fighting financial contagion and other undesired effects of liquidity shortages due to large price swings in asset markets, Europe would still face a severe aggregate demand problem well beyond the reach of monetary policy and, as argued in our previous reports, also of fiscal policy. Perhaps the most important risk for Europe associated with global imbalances is that of facing a severe crisis without effective policy instruments to stabilise the European economy.
Notes 1. Recent theoretical, empirical and policy-related contributions to the debate on global imbalances can be found on the website ‘Current Account Sustainability of Major Industrialized Countries’ at the University of Wisconsin, Madison, http:// currentaccount.lafollette.wisc.edu/. 2. The mechanism benefiting the US is the same (but with an opposite sign) as the mechanism raising financial and macroeconomic risks in emerging markets: as these borrow by issuing debt denominated in foreign currency, domestic devaluation in response to negative shocks magnifies macroeconomic adjustment problems by raising the real burden of external debt (see Chapter 7). 3. When market expectations anticipate depreciation, the size of net gains also depends on the maturity structure of debt. 4. Some sign of house price stabilisation has been detected throughout the year (see the Economist 2005). House prices have been high and rising through the last few years also in some of the other industrial countries that, like the US, have a persistent current account deficit. This applies to Australia, New Zealand, Spain and the UK. See also Chapter 5 of the 2005 EEAG Report. 5. An interesting analysis discussing US fiscal policy in relation to the possible asset market bubble in the 1990s is Kraay and Ventura (2007). 6. Engel and Rogers (2005) carry out their analysis in terms of net GDP. This is derived by subtracting investment and government final consumption of goods and services from output. In intertemporal models of the current account, net GDP measures the flow of resources that households can devote in each period either to current consumption or to the acquisition of foreign assets to finance future consumption. 7. Such a portfolio perspective on dollar adjustment dynamics is discussed by Blanchard et al. (2005). 8. Caballero et al. (2006) reconsider this issue in a model with three regions: a fastgrowing US-type region, a slow-growing Europe-type region, and an industrialising region with exceptional growth opportunities, modelled to reflect Asian emerging markets. Notably, in the latter region, financial markets cannot supply quality assets because there are frictions generating constraints on asset supply. For this reason, in this region investment is mainly financed by firms’ managers/ owners, without the participation of savers. The model can do reasonably well in accounting for recent patterns of global financial flows. According to this analysis, both a slowdown in the Europe-type region and a crash in the asset market in the emerging-market region can cause an external deficit in the US-like region. Either shock produces a prolonged period of low real interest rates. 9. See Chapter 2 of International Monetary Fund (2005) for a detailed analysis.
9780230_547018_11_cha09.indd 323
3/19/2009 8:19:26 PM
324
Designing the European Model
10. Some of the investment fall can, however, be attributed to a secular decline in the relative price of investment goods. 11. Dooley et al. (2003 and 2004) interpret the current international monetary and exchange rate regime with dollar pegs and large dollar reserve accumulation as a revised Bretton Woods regime. See Roubini and Setser (2005b) for a critical view. 12. A weak currency feeds a strong external demand for Chinese products, encouraging investment, but it also raises prices of imported intermediate and capital goods. However, the bulk of infrastructure building is based on local and nontraded goods, and FDI flows have remained substantial. 13. By reducing profitability of firms now exporting thanks to subsidies and a low exchange rate, a revaluation may generate bankruptcies and costs for financial institutions. To the extent that the banking system is public, these costs will deteriorate the fiscal balance of China. In addition, a revaluation will create capital losses on the vast reserve holdings by the Chinese central bank (now around 40 percent of GDP). 14. Recent contributions discussing alternative scenarios of adjustment include Adalet and Eichengreen (2007), Clarida et al. (2007), Croke et al. (2005), Edwards (2005), Faruqee et al. (2007), Freund and Warnock (2007), Hunt and Rebucci (2003), Mann (1992) and Mann and Plueck (2007). 15. The share of Asia in US imports was in fact as high as around 28 percent in 2004, up from 25 percent in 2000 (the share of US exports to Asia is much lower and quite variable). 16. See Hooper et al. 1998 and the quantitative analysis by Corsetti et al. 2004. 17. See Section 2.3 above. 18. In his blog (http://www.rgemonitor.com/blog), Roubini lists a series of contingent factors which may explain the strengthening of the dollar in 2005. The list includes: widening of short-term interest rate differentials between, on the one hand, the US and, on the other, the euro area and Japan; growth differentials in favour of the US; the effect of temporary measures, such as the Homeland Investment Act, providing a tax incentive for profit repatriation; political factors related to the constitutional referendum failures in Europe; and increasing reservations about the pace and depth of the European integration process. 19. See Chapter 5 of the 2004 EEAG report.
References Adalet, M. and B. Eichengreen (2007) “Current Account Reversals: Always A Problem?” in R. Clarida, ed., G3 Current Account Imbalances, Chicago, University of Chicago Press, pp. 205–46. Bernanke, B. (2005) “The Global Saving Glut and the U.S. Current Account Deficit,” Sandridge Lecture, Virginia Association of Economics, Richmond, Virginia. Blanchard, O. and F. Giavazzi (2005) “Rebalancing Growth in China: A Three-Handed Approach,” CEPR Discussion Paper Series no. 5403. Blanchard, O., F. Giavazzi and F. Sa (2005) “The U.S. Current Account and the Dollar,” NBER Working Paper no. 11137. Bussière, M., M. Fratzscher and G. J. Müller (2005) “Productivity Shocks, Budget Deficits and the Current Account,” ECB Working Paper no. 509. Caballero, R., E. Farhi and P. O. Gourinchas (2006) “An Equilibrium Model of ‘Global Imbalances’ and Low Interest Rates,” mimeo, MIT.
9780230_547018_11_cha09.indd 324
3/19/2009 8:19:26 PM
Global Imbalances 325 Cavallo, M. and C. Tille (2005) “Current Account Adjustment with High Financial Integration: A Scenario Analysis,” mimeo, Federal Reserve Bank of New York. Chamon, M. and E. Prasad (2006) “Determinants of Household Saving in China,” IMF Working Paper (Washington: International Monetary Fund, forthcoming). Chinn, M. (2005) “Still Doomed to Deficits: An Update on US Trade Elasticities,” mimeo, University of Madison, Wisconsin. Chinn, M. and H. Ito (2005) “Current Account Balances, Financial Development and Institutions: Assessing the World Saving Glut,” mimeo, University of Wisconsin, Madison. Clarida, R., M. Goretti and M. Taylor (2007) “Are There Thresholds of Current Account Adjustment in the G7?” R. Clarida, ed., G3 Current Account Imbalances, Chicago, University of Chicago Press, pp. 169–204. Cline, W. (2005) The United States as a Debtor Nation, Institute for International Economics. Congressional Budget Office (2005) “Recent Shifts in Financing the U.S. CurrentAccount Deficit,” Economic and Budget Issue Brief, 12 July 2005. Corsetti G. and G. Mueller (2005) “Twin deficits: squaring theory evidence and common sense,” paper prepared for the Economic Policy Panel in April 2006. EUI Economics Working Paper 2005–22. Corsetti G., L. Dedola and S. Leduc (2004) “International Risk Sharing and the Transmission of Productivity Shocks,” European Central Bank Working Paper no. 308. Corsetti G., P. Martin and P. Pesenti (2005) “Productivity Spillovers, Terms of trade and the Home Market Effect,” CEPR Discussion Paper 4964. Corsetti G., P. Pesenti and N. Roubini (1999) “What Caused the Asian Currency and Financial Crisis?” Japan and the World Economy 3, 305–73. Croke, H., S. B. Kamin and S. Leduc (2005) “Financial Market Developments and Economic Activity during Current Account Adjustments in Industrial Economy,” International Finance Discussion Paper no. 82.7. Cuñat A. and M. Maffezzoli (2004) “Neoclassical Growth and Commodity Trade,” Review of Economic Dynamics 7(3), 707–36. Dooley, M., D. Folkerts-Landau and P. Garber (2005) “Direct Investment, Rising Real Wages and the Absorption of Excess Labor in the Periphery,” in R. Clarida, ed., G7 Current Account Imbalances: Sustainability and Adjustment, NBER. Dooley, M., D. Folkerts-Landau and P. Garber (2004) “The U.S. Current Account Deficit and Economic Development: Collateral for a Total Return Swap,” NBER Working Paper no. 10727. Dooley, M., D. Folkerts-Landau and P. Garber (2003) “An Essay on the Revised Bretton Woods System,” NBER Working Paper no. 9971. Edwards, S. (2005) “Is the U.S. Current Account Deficit Sustainable? And If Not, How Costly Is Adjustment Likely to Be?” NBER Working Paper no. 11541. Engel, C. and J. H. Rogers (2005) “The U.S. Current Account Deficit and the Expected Share of World Output,” Paper presented at the Carnegie-Rochester Conference Series on Public Policy, November 18–19. Erceg, C. J., L. Guerrieri and C. Gust (2004) “Expansionary Fiscal Shocks and the Current Account,” mimeo, Board of Governors of the Federal Reserve System. European Economic Advisory Group at CESifo (2003) Report on the European Economy 2003, CESifo, Munich. European Economic Advisory Group at CESifo (2004) Report on the European Economy 2004, CESifo, Munich.
9780230_547018_11_cha09.indd 325
3/19/2009 8:19:26 PM
326 Designing the European Model European Economic Advisory Group at CESifo (2005) Report on the European Economy 2005, CESifo, Munich. European Economic Advisory Group at CESifo (2006) Report on the European Economy 2006, CESifo, Munich. Faruqee H., D. Laxton, D. Muir and P. Pesenti (2007) “Current Accounts and Global Rebalancing in a Multi-Country Simulation Model,” in R. Clarida, ed., G3 Current Account Imbalances, Chicago, University of Chicago Press, pp. 377–456. Ferguson, R. W. (2005) “U.S. Current Account Deficit: Causes and Consequences,” Remarks to the Economics Club of the University of North Carolina. Freund C. and F. Warnock (2007) “Current Account Deficits in Industrial Countries: The Bigger They Are, The Harder They Fall?” in R. Clarida, ed., G3 Current Account Imbalances, Chicago, University of Chicago Press, pp. 133–168. Genberg, H., R. McCauley, Y. Chul Park and A. Persaud (2005) Official Reserves and Currency Management in Asia: Myth, Reality and the Future, Geneva Reports on the World Economy 7, CEPR. Gourinchas, P.-O. and H. Rey (2005a) “International Financial Adjustment,” International Finance no. 0505004 and NBER Working Paper no. 11155. Gourinchas, P. O. and H. Rey (2007) “From World Banker to World Venture Capitalist: The US External Adjustment and the Exorbitant Privilege,” in R. Clarida, ed., G3 Current Account Imbalances, Chicago, University of Chicago Press, pp. 11–66. Greenspan, A. (2004) “Evolving U.S. Payments Imbalances and Its Impact on Europe and the Rest of the World,” Cato Journal 24, 1–11. Hausmann R. and F. Sturzenegger (2006) “Global Imbalances or Bad Accounting? The Missing Dark Matter in the Wealth of Nations,” Center for International Development at Harvard University, Working paper no. 124, January. Hooper, P., K. Johnson and J. Marquez, (1998) “Trade Elasticities for G-7 Countries,” Federal Reserve Board of Governors, International Finance Discussion Paper no. 609. Houthakker, H. and S. Magee, (1969) “Income and Price Elasticities in World Trade,” Review of Economics and Statistics 51, 111–25. Hunt, B. and A. Rebucci (2003) “The U.S. Dollar and the Trade Deficit: What Account for the 1990s?” IMF Working Paper no. 03/104. International Monetary Fund (2005) World Economic Outlook, September. Kim, S. and N. Roubini (2003) “Twin Deficits or Twin Divergence? Fiscal Policy, Current Account, and Real Exchange Rate in the US,” mimeo, New York University. Kraay, A. and J. Ventura (2007) “The Dot-Com Bubble, the Bush Deficits, and the U.S. Current Account,” in R. Clarida, ed., G3 Current Account Imbalances, Chicago, University of Chicago Press, pp. 457–96. Krugman, P. (1991) Has the Adjustment Process Worked? Institute for International Economics, Policy analyses in international economics, 34. Lane, P. and G. M. Milesi-Ferretti (2005) “A Global Perspective on External Positions,” IIIS Discussion Paper no. 79 and NBER Working Papers no. 11589. Lane, P. and G. M. Milesi-Ferretti (2006) “The External Wealth of Nations Mark II: Revised and Extended Estimates of External Assets and Liabilities,” Unpublished: Washington, International Monetary Fund. Mann, C. L. (1992) “Perspectives on the US Current Account Deficit and Sustainability,” Journal of Economics Perspectives 16, 131–52. Mann, C. and K. Plueck (2007) “The US Trade Deficit: A Disaggregated Perspective,” in R. Clarida, ed., G3 Current Account Imbalances, Chicago, University of Chicago Press, pp. 247–82. Obstfeld, M. (2004) “External Adjustment,” Review of World Economics no. 4.
9780230_547018_11_cha09.indd 326
3/19/2009 8:19:26 PM
Global Imbalances 327 Obstfeld, M. (2005) “America’s Deficit, the World’s Problem,” Keynote speech, Monetary and Economic Studies, Special Edition, October. Obstfeld, M. and K. Rogoff (2001) “Perspectives on OECD Economic Integration: Implications for US Current Account Adjustment,” Proceedings, Federal Reserve Bank of Kansas City, pp. 247–82. Obstfeld, M. and K. Rogoff (2004) “The Unsustainable US Current Account Position Revisited,” NBER Working Paper no. 10869. Obstfeld, M. and K. Rogoff (2005) “Global Current Account Imbalances and Exchange Rate Adjustments,” Brookings Papers on Economics 1, 67–123. Rodrik, D. (2006) “The social cost of Foreign Exchange Reserves,” mimeo, Harvard University. Roubini, N. and B. Setser (2004a) “The Sustainability of the US External Imbalances,” CESifo Forum, 2005, 1, 9–15. Roubini, N. and B. Setser (2004b) “The US as a net debtor: The Sustainability of the US External Imbalances,” mimeo. Roubini, N. and B. Setser (2005a) “How scary is the deficit?” Foreign Affairs, 84(4), 194–98. Roubini, N. and B. Setser (2005b) “Will the Bretton Woods 2 Regime Unravel Soon? The Risk of a Hard Landing in 2005–2006,” Proceedings, Federal Reserve Bank of San Francisco, Feb. issue. Tille, C. (2003) “The Impact of Exchange Rate Movements on U.S. Foreign Debt,” Current Issues in Economics and Finance, Federal Reserve Bank of New York, 9(1), January 2003.
9780230_547018_11_cha09.indd 327
3/19/2009 8:19:27 PM
This page intentionally left blank
Part III Institutions
9780230_547018_12_cha10.indd 329
3/19/2009 8:22:26 PM
This page intentionally left blank
10 Rethinking Subsidiarity in the EU: Economic Principles
10.1
Introduction
The European Union (EU) is currently facing major challenges, since the enlargement to East Central European and some other countries necessitates considerable changes in the structure of decision-making and the operation of the union. The establishment of the Economic and Monetary Union (EMU) and the introduction of the euro as the common currency in EMU countries testify to a stride towards deeper political integration. These developments suggest that a reconsideration of the principles of public decision-making is timely, and they have led to the establishment of the European Convention, which will submit its proposals to the European Council sometime in 2003. The tasks of the Convention concern the division of competence between member states and the Union, better definitions of tasks for EU institutions, coherence and efficiency of EU external action and democratic legitimacy of the Union.1 The current times provide a good opportunity to reconsider the public governance and organisation of economic activities within the EU. The key question is how decision-making for public sector activities should be distributed among different levels of government. The EU is only one layer of government and there are other levels of decision-making for public sector activities, including the national and local governments in all EU member countries. Some EU countries have an explicit federalist structure, so that their governments have local, state and national levels of public decisionmaking. The non-federalist countries, for example Finland and Sweden, have also some intermediate levels of public administration, but these bodies do not have any legislative power. The principle of subsidiarity is one of the key concepts in political decision-making in the EU. This principle stipulates that public sector decisions and problem solutions should be kept at the lowest layer of government that is appropriate for that specific task. However, the assessment of what is appropriate must be specified by reference to further principles developed 331
9780230_547018_12_cha10.indd 331
3/19/2009 8:22:27 PM
332
Designing the European Model
in the theory of fiscal federalism, such as international policy spillovers and international public goods (see Oates 1972 and Shah 1994 for a general discussion of this theme). Subsidiarity is a widely debated and even disputed concept (see, for example, Bermann 1994; Begg et al. 1993 and Alesina et al. 2002 and the references therein).2 Subsidiarity raises several questions. First, is it possible to provide more concrete and operational content or guidelines to implementation of subsidiarity for specific public sector activities? Second, if this more precise content of subsidiarity can be found, how well is the EU currently matching these guidelines? Third, what might be the directions for the EU to take in order to improve its operation? This chapter considers these questions from the perspective provided by economic analysis. Since the EU is to a large extent an economic union, an analysis of the concept of subsidiarity from the economic viewpoint can provide useful input to the current discussion about the Constitution of the EU and the future course of the development of the EU. We will argue that here are useful economic principles for assessing the application of subsidiarity in specific governmental activities.3
10.2
Economic principles for subsidiarity
Consider a group of countries which have formed an economic and political union in which there is a union level of government besides the national (and possibly lower level) governments in each country. The formation of an economic union implies the existence of important economic activities and policies that provide important benefits to the members of the union. The single market is probably the most fundamental economic factor in an economic union (see, for example, Weingast 1994; Tabellini 2002). Economic theory suggests that significant benefits can accrue from free trade in goods and from free movement of economic resources among countries. An economic union yields free trade and mobility only within its jurisdiction and not necessarily globally, but regional free trade areas are usually viewed as beneficial.4 Since global free trade and mobility are extremely hard to come by through negotiations, regional integration may provide a way for reaping some of the benefits and it may also lead to further global integration in the future. A strong regional union can also provide impetus to the process toward global free trade, though this need not always be the case. Regional unions are only second-best solutions, since they can lead to trade diversion rather than the creation of new trade. Trade diversion is evident in agriculture as a result of the EU Common Agricultural Policy, as witnessed for example by the very high food prices in the EU (Corsetti et al. 2002, Table 7.1). These considerations suggest that protection of the single market with internal free trade and factor mobility and the enhancement of global free trade constitute the single most important economic function of an economic union. By itself, the single market and common trade policy provide a reason for the centralisation of specific economic interventions. The
9780230_547018_12_cha10.indd 332
3/19/2009 8:22:27 PM
Rethinking Subsidiarity in the EU
333
protection of the single market can be in conflict with the notion of subsidiarity that, in contrast, suggests the greatest possible decentralisation of public sector economic activities. Are there economic principles that can provide justifications for the decentralisation of public sector activities? From an economic point of view, the notion of subsidiarity is largely the question of allocating public sector tasks among the different levels of government. Centralisation versus decentralisation of public sector tasks raises many different issues and the choice is not a routine matter. Centralisation of decision-making can, in principle, provide savings of administrative costs, and problems of co-ordination of decisions and activities among the different levels of government can also be minimised. However, centralisation can also lead to rigid governance that, in turn, implies inefficiency when there is significant heterogeneity between countries in terms of the preferences of the citizens and also in terms of the costs and productivity of public activities. If centralised public sector decisions are uniform across the different member countries, as they tend to be in practice, they may bring about considerable losses of efficiency because they artificially enforce homogeneity on heterogeneous countries. Box 10.1 illustrates potential inefficiencies arising from rigid centralised decisions.5 It may also be true that bureaucracy increases in line with more centralised public sector governance, leading, for example, to excessive delays in decision-making. The current EU is often criticised in this respect. Centralisation of public sector decisions can also lead to problems of accountability of political decision-makers. A centralised government structure means that the citizens are further apart from the decision-makers than if public decisions are carried out in a more decentralised structure. Accountability would seem to favour decentralisation of public decision-making. The preceding considerations confirm that subsidiarity, that is decentralisation of public sector decision-making, should be favoured as a basic principle. However, there are economic activities in which centralised governance can bring forth benefits in terms of improved efficiency and equity among citizens. Tax systems are a case in point. Decentralised taxation in a federation of countries can lead to serious biases that, in turn, result in inefficiencies and inequities. In general, economic analysis suggests a number of reasons for centralising the governance of some public activities and decentralising others. There is no simple or uniform answer to the question of centralisation versus decentralisation of public sector decisions. The answer depends on the nature of the economic activity under scrutiny. This chapter examines a number of economic activities more closely and discusses criteria for deciding when to apply the principle of subsidiarity. 10.2.1
Public goods, externalities and spillovers
Public goods are commodities or activities in which the benefits accrue jointly to the inhabitants of an area or jurisdiction. Consumption or use of the good by one economic agent (consumer or firm) in the jurisdiction does
9780230_547018_12_cha10.indd 333
3/19/2009 8:22:27 PM
334
Designing the European Model
Box 10.1
Heterogeneity and inefficiency of uniform decisions
To illustrate the significance of heterogeneity consider a public good, say a cultural activity, which has the characteristic that citizens of a country or region consume it jointly. In this case we can speak of total or marginal value of the public activity. Consider the following example: Suppose that for the citizens of country A any amount X A of a public activity has a marginal value 6-X A to them and that the marginal cost of production is 2. The efficient level of provision XA to citizens of country A is X A = 4. Suppose also that, if the citizens of country B choose any amount XB of the same activity that has a marginal value, say (1/2) (6-XB) to them. Note that, at each level of supply, the citizens of country B value the public activity less than those of country A. The marginal cost of production is assumed to be the same as in country A. For country B the efficient amount of supply is XB = 2, that is half of the efficient amount for country A. Any uniform supply of the public good decided by a union level government, for example the mid-point X = 3, is inefficient. This level is too low for the citizens of country A and too high for the citizens of country B. In other words, a uniform centralised decision about the level of public activity does not respect the preferences of the citizens in the two member countries. It would be better to leave the decision about this activity to each country, as the two countries would presumably choose the efficient levels 4 and 2 of output.
not reduce the amount available to other economic agents in the area or jurisdiction. Public goods are goods that are jointly consumed, and the consumption of the goods by individuals is non-rival.6 The costs of provision of such goods are then to be shared among the inhabitants of the area or jurisdiction. National defence, works of art in public places and the preservation of the cultural heritage are standard examples of such public goods. The benefits from a public good are geographically limited in most cases. However, if the public good character of an activity extends beyond the nation state, then it is efficient to make decisions about its provision, including the sharing of costs, at a higher level of government. The appropriate level can even be the supra-national European (or even global) level. If the beneficiaries of the public good make the decisions about the level of provision and payment of the production costs of the good, then efficiency can be achieved. Suppose that the decisions about the provision of the public good are made in a smaller area and thus by a smaller group than the geographical distribution of benefits would dictate. If this is the case, then some benefits from the good and some potential contributors to its costs are not taken into consideration and the outcome is usually inefficiently low provision. Generally, public goods are provided most efficiently by a jurisdiction that has control over the minimum geographic area that would internalise the benefits and costs of its provision, as suggested by Oates (1972, p. 55). There are probably not that many goods or activities for which the public good character or non-excludability of that activity extends well beyond
9780230_547018_12_cha10.indd 334
3/19/2009 8:22:27 PM
Rethinking Subsidiarity in the EU
335
national jurisdictions.7 Countries and political unions of countries are different as the former is likely to be the natural jurisdiction for more public goods than the latter. We will come back to this case and the examples later when we discuss the activities of the EU. In addition to union level public goods, there are other forms of economic activity where public involvement in appropriate ways can be justified. Generally speaking, this occurs when economic activity by some economic agent generates significant externalities or spillovers on other agents in the sense that the latter get unpaid benefits or incur unaccounted costs from this activity. If these externalities or spillovers extend well beyond the national jurisdictional borders, then involvement of union level public administration can be justified from the efficiency viewpoint. It is easy to think of examples of externalities and spillovers. An important case is communication networks. The benefits from membership in a network to any single agent typically depend on the size of the network. Any particular method of communication is not useful to somebody if he cannot contact many other relevant people through it. Establishment of a common standard is often important for networks. Competition between, say different communication technologies, can be wasteful since the establishment of a network often involves significant investments. Competing standards can lead to wasteful investments; one of the duplicate investments in related technologies is often redundant. Moreover, economies of scale in production can possibly be achieved through the adoption of a common technology or standard. Public decision-making can play an important role in the setting of a common standard and, for cases of networks or standards that naturally extend beyond national boundaries, the appropriate level of governmental decisions will be the union level. Economies of scale and benefits from common standards may also exist in public sector activities. If these are obtainable by joint decisions beyond national jurisdictions, then union level decisionmaking is potentially justifiable on the basis of efficiency considerations. The general lesson from the preceding discussion is that the appropriate level of government intervention – global, union, national, regional or local – depends on the nature of the economic activity. This must be assessed case by case, and there are few general rules about the most appropriate level of governmental action. Moreover, in practice compromises must be made since the geographical dispersion of benefits for public goods and spillovers need not follow the geographical boundaries of the EU. It might also be impractical to have too many levels that are geared just to decisions about specific public goods or spillovers. Administrative efficiency must also be taken into consideration. It is important to point out that, even if a particular good is a public good, its provision need not always be handled by a government agency. The forms of government intervention, be it at union, national or local level,
9780230_547018_12_cha10.indd 335
3/19/2009 8:22:27 PM
336 Designing the European Model
can vary a great deal from case to case. The forms of intervention can range from common regulation to the provision of the good or activity through public funds and agencies. It must also be emphasised that governmental decision-making covers much more than the provision of physical goods and services. For example, ensuring a stable financial system for the operation of a single market is an activity in which externalities are not very tangible but nevertheless important. As discussed in Chapter 11, financial regulation is basically a union level activity for the EU. Another example that respects subsidiarity is the establishment of national fiscal councils which is suggested in Chapter 9. The notion of a single market for the EU has many dimensions and, as a common goal, implies different types of decisions and activities. The question of subsidiarity can be raised with respect to many different types of public activity, including taxes and subsidies, redistribution and the welfare state, labour market standards, regulatory standards for industry and competition, environmental standards and so on. We will analyse subsidiarity in relation to different types of activity further in Section 4. 10.2.2 Politico-economic aspects The above discussion about the allocation of tasks to different levels of government has been conducted from the basic viewpoint of economic efficiency. Subsidiarity can, however, also be considered from a different economic viewpoint that combines the functioning of the political and economic systems. This is the so-called political economy approach. Competition between jurisdictions One strand in the political economy approach emphasises competition between national jurisdictions. It is often argued that intergovernmental or “systems” competition can involve a “race to the bottom” where, in terms of economic efficiency, too low levels of public goods or activities are provided. A well-known example is tax competition. It is suggested that, in order to attract more capital and investment, countries compete by lowering their taxes on capital to zero, since capital is internationally mobile and would move to the country with lowest taxation. Labour is also potentially mobile and can possibly move in search of the highest net social benefits.8 A counterpart to tax competition is subsidy competition for particular industries. Shipbuilding provides a current example. Various governments in the world subsidise shipyards, and even within the EU it has been very difficult to abolish these subsidies. This seems in part to be caused by subsidisation of the activity in other countries outside the EU. These subsidies run seriously counter to the key principle for the EU, the protection of the single market in which the competitive conditions for different firms should be uniform.
9780230_547018_12_cha10.indd 336
3/19/2009 8:22:27 PM
Rethinking Subsidiarity in the EU
337
The creation of unequal terms for competition within the single market caused by public subsidies by national governments is a risk that is potentially present in various activities. One case is the possible reintroduction of the national subsidisation of agriculture that has been suggested in current discussions of agricultural reform in the EU. Natural conditions for agriculture differ among countries, but this is not a cause for concern. On the contrary, this is an instance of comparative advantage, the existence of which is the key basis for benefits from free international trade. More generally, national regulatory systems can lead to “systems competition,” in which different regulatory standards can result in important externalities between countries and/or industries. However, in some cases systems competition can lead to more efficient government decisions. Solutions to these kinds of systems competition need to be worked out case by case, as discussed by Sinn (2003). It should also be noted that the arguments about intergovernmental competition and the “race to the bottom” or “race to the top” are not easy to assess empirically. There is clear evidence that, for example, corporate taxes have fallen and labour taxes have risen in many countries during the 1990s after liberalisation of capital movements (see Sorensen 2000; Wildasin 2000; Sinn 2003 and Devereux et al. 2002). Nevertheless, the evidence on the mobility of productive resources across countries and tax competition is far from systematic, and it is not currently clear what the final verdict will be, as has been recently argued for example by Oates (1999) and Bhagwati (2002). It is probable that even within the EU this mobility is less than perfect, and thus the forces driving intergovernmental competition may not always be very strong. The goals of government A different view on intergovernmental competition arises from a fundamental tension in government objectives (see Weingast 1995). On the one hand, a sufficiently strong government is needed to enforce private contracts and to protect the economic rights of private economic agents. These government activities are important for a proper functioning of markets. On the other hand, a government may not be benevolent but instead “Leviathan,” aiming to confiscate the wealth of its citizens. Strong governments may be successful in this. Intergovernmental competition for mobile economic resources can provide limits to Leviathan governmental behaviour, though the desirability of the outcome depends on the circumstances, as has been pointed out by Edwards and Keen (1996). Improved accountability in a more decentralised government structure can also provide limits to Leviathan behaviour. A union of countries or a federalist structure with decentralised governmental decision-making can be a vehicle for an efficient system of markets. Three dimensions are pertinent here, according to Weingast (1995) and others. First, decentralised governments should have primary responsibility
9780230_547018_12_cha10.indd 337
3/19/2009 8:22:27 PM
338
Designing the European Model
to regulate the economy unless specific circumstances call for more centralised decision-making. Second, the union of countries has the task of running a common market with no barriers to trade. The good functioning of the single market is the key governmental function at the union level. Third, the low levels of government should face “hard budget constraints” whereby they cannot have access to unlimited credit or the printing of money.9 These aspects suggest an important role for subsidiarity as a general guideline, from which departures should be considered only for specific activities. Different ways of facilitating the single market are then the main reason for exceptions to the principle of subsidiarity. For macroeconomic management these three criteria are relatively well in line with the current division of responsibility between the EU and national levels of decision-making. The monetary union with centralised decisionmaking for monetary policy can be seen as a method both for promoting the common market in the EU and for preventing soft budget constraints for EMU member countries. The fiscal provisions in both the Maastricht Treaty and the Stability and Growth Pact are a further means for preventing unlimited credit by EMU member countries, even if concerns are occasionally expressed about the restrictive nature of the Pact from the point of view of macroeconomic stabilisation policy (see Chapter 9). The actual operation of the EU can vitiate the economic principles of subsidiarity versus centralisation. Political bargaining and lobbying can lead to inefficiencies in public decision-making. Decision-making at the EU level is not free from these phenomena. On the other hand, subsidiarity may be exploited by national lobbies, and moving certain decisions to the Union level can provide benefits to member states. EMU and common monetary policy illustrates this phenomenon; it enhanced the credibility of monetary policy and has ensured low interest rates for some countries (see, for example, Corsetti et al. 2002, Figure 4.1). Overall, the political economy aspects of governmental decisions in a union of countries suggest that the implications of intergovernmental or systems competition must also be assessed for specific functions and activities of the government. These include tax competition, competition in regulatory systems and issues concerning migration and labour. A further controversial subject of subsidiarity in decision-making concerns redistribution. We will take up these topics below.
10.3
What does the European Union currently do?
The previous section considered the basic economic principles of subsidiarity. It was suggested that the important criterion is the existence of public goods or externality effects with strong geographical dispersion of benefits or costs across a number of EU countries. While we still need to consider further the specific activities in the light of the general principles, it is also
9780230_547018_12_cha10.indd 338
3/19/2009 8:22:27 PM
Rethinking Subsidiarity in the EU
339
important to take a look at what the EU currently does. This will provide a basis for a further discussion of how different types of activities might be placed in the different levels of government. Here we consider empirically the current EU activities and policies. We look at them from two different angles. First, we take up the traditional approach and consider the structure of EU level public spending. We compare the structure to that of existing federalist countries, such as the US, Canada and Germany. This comparison is useful even if the EU is not and is not meant to be a federation, but rather a union of independent countries. This is so because an efficient distribution of tasks among different levels of government does not require the formation of a federation from a legal point of view. Second, we consider the activities and policies of the EU from the point of view of regulation of the economy that does not often lead to significant public spending. This is important since the budget of the EU is far more limited than that of individual federalist or unitary government countries. Focusing only on public spending and revenues would give a misleading picture of EU activities and policies. 10.3.1 Public spending The EU budget is shown in Table 10.1. It is evident that EU budgetary spending focuses on very specific tasks. Moreover, these tasks are quite surprising when they are compared to what economic principles would suggest. Even today, nearly half of the EU budget is devoted to agricultural subsidies and guarantees, even though the share has decreased during the 1990s. For 2002, the share is 46.5 per cent, down from 58.5 per cent in 1991. Structural funds and operations are the second largest item in the EU budget, with its share at approximately 34 per cent in 2002. These two main items are largely redistributive in nature and, as discussed below, it is far from obvious that they should be the responsibility of the EU level of government. The remaining important items are external action, that is policies towards non-EU countries (for example development aid and pre-accession strategy) and international operations with a budget share of 7.7 per cent in 2002, research and technological development with a share of 3.9 per cent and administration with a budget share of 5.4 per cent in 2002. EU spending on research and technological development can to an extent be viewed as a natural EU level spending item. However, this is so only as long as the expenditure is an attempt to correct for externalities in research, as will be discussed below. The EU budget is quite small in comparison with the government budgets of the member countries as illustrated by the following comparisons. In the period 1996–2000 the EU budget was only 2.4 per cent of the total of government budgets of the EU member countries. In 2000, the EU budget was only 1.1 per cent of joint EU gross domestic product whereas the share of national budgets of EU member countries in GDP was 46.8 per cent.
9780230_547018_12_cha10.indd 339
3/19/2009 8:22:28 PM
9780230_547018_12_cha10.indd 340
3/19/2009 8:22:28 PM
EU expenditures by function
100
3.2 4.0 0.00 2.4 3.0 1.7
0.4 0.5
0.7
55.4 58.7 25.8
1991
100
0.7 3.6 4.8 0.00 0.5 3.5 1.7
100
0.9 3.8 5.8 0.04 0.5 3.4 1.8
0.9 0.2
32.3
30.8 0.8 0.2
82.4 50.1
1997
70.0 53.5
1994
100
1.1 4.0 6.1 0.03 0.2 3.4 1.8
0.8 0.2
35.8
89.4 46.4
2000
100
1.2 3.9 6.7 0.04 0.2 3.5 1.8
0.8 0.2
34.3
92.6 47.3
2001
100
1.2 3.9 7.7 0.04 0.4 3.5 1.9
0.9 0.2
33.8
95.7 46.5
2002
Source: Eurostat, Statistisches Jahrbuch für das Ausland, Statistisches Bundesamt 2001, Amtsblatt der Europäischen Gemeinschaften; Rechtsvorschriften, Haushaltspläne, Amt für Amtliche Veröffentlichungen der Europäischen Gemeinschaften 1991–L30; 1994–L34, 1996–L22; 1997–L44; 1998–L44; 1999– L39; 2000–L40; 2001–L56; 2002–L2.
Total
Total expenditure in million EUR EAGGF-Guarantee Section Structural operations, structural and cohesion funds: financial mechanism, other agricultural and regional operations, transport and fisheries Training, youth, culture, audio-visual media, information, social dimension and employment Energy, Euratom nuclear safeguards and environment Consumer protection, internal market, industry and transEuropean networks Research and technological development External action Common foreign and security policy Guarantees and reserves Administrative expenditure of the Commission Administrative expenditures (of all other institutions = Parliament, Council, Court of Justice, Court of Auditors, Economic and Social Committee, Committe of the Regions, European Ombudsman, European Data Protection Supervisor)
Function
Table 10.1
Rethinking Subsidiarity in the EU
80
341
%
70 60 50 40 30 20 10 0 central state local Germany
central state local Spain
central state local Australia
central state local USA
Figure 10.1 Public expenditure of country, state and local governments in selected federalist nations in 1996 Source: IMF, Government Financial Statistics Yearbook 2002, own calculations.
The structure of the EU budget can be contrasted with the division of public expenditure between national, state and local government budgets. Figure 10.1 shows this division for some federal states. Public expenditures are spread widely across different activities ranging from key public services, education and health, social security and welfare to a number of other sectoral activities. There are notable differences in public expenditures among the different levels of government in the federalist countries, as shown in Table 10.2. For example, public safety is a federal matter in Spain whereas it is more the responsibility of the state or local government in the other countries shown. Another example is education, which is a state responsibility in Germany and a shared task in other countries. In general, for each function, the division of responsibilities between the three levels of government shows significant variability, such that federal countries do not present a unified picture in this respect. A very important observation is that the federal government accounts for a relatively high share of total public spending. This share ranges from about 52 to 69 per cent in the countries shown in Figure 10.1. This is in marked contrast to the EU budget. 10.3.2
Regulatory activities and policies
The regulatory instruments of EU policies can be divided into Treaties, Secondary Legislation and Other Acts. Treaties are negotiated at Intergovernmental Conferences to be ratified by all member countries. Treaties are the ultimate source of mandate and legitimacy for all EU institutions and their legislative and judicial authority.
9780230_547018_12_cha10.indd 341
3/19/2009 8:22:28 PM
9780230_547018_12_cha10.indd 342
3/19/2009 8:22:28 PM
24.1 33.5 0.0 71.7 68.4 12.5 10.9 26.4 32.9 35.5 52.6 75.1 6.1 33.3 30.8 31.1 0.0
59.6 38.1 100.0 7.6 4.2 72.4 78.6 7.4 4.4 46.1
36.1 15.1 93.9 43.4 50.0 63.1 100.0
state
11.3 9.8 0.0 23.3 19.2 5.8 0.0
16.2 28.4 0.0 20.7 27.4 15.1 10.5 66.2 62.7 18.4 91.6 31.8 47.4 49.6 60.5 84.5 100.0
69.0 46.8 100.0 60.6 30.8 36.7 94.3 7.6 21.4 49.8 7.2 60.2 46.1 19.5 24.4 6.9 0.0
19.3 14.8 0.0 12.8 63.0 60.4 3.7 18.3 26.9 29.6
local central state
Spain
Australia
1.2 8.0 6.6 31.0 15.2 8.6 0.0
11.7 38.5 0.0 26.5 6.2 2.9 2.0 74.1 51.7 20.6 66.1 49.9 43.5 17.1 66.4 77.9 100.0
58.8 48.7 100.0 15.1 29.4 53.3 91.2 23.9 26.4 38.3 33.9 49.5 40.4 58.6 27.0 19.7 0.0
36.9 42.3 0.0 81.6 70.5 46.0 7.4 44.2 37.1 47.1
local central state
Source: Government Finance Statistics Yearbook 2002, IMF; calculations by the Ifo Institute.
Note: *From 1996 onward all government data are compiled on a year ending June 30.
1. Total Expenditure 2. General Public Services 3. Defense 4. Public Order & Safety 5. Education 6. Health 7. Social Security & Welfare 8. Housing & Commun. Amen. 9. Recr., Cultr., Relig. Affrs. 10. Econ. Affairs & Services (11–15) 11. Fuel & Energy 12. Agric., Forestry, Fishing, Hunt 13. Mining, Manufac. & Construc. 14. Transportation & Comm. 15. Oth. Econ. Affairs & Serv. 16. Other Expenditures of which Interest Payments
central
Germany
0.0 0.6* 16.0* 24.4* 6.6* 2.3 0.0
5.1* 9.1* 0.0 3.4* 0.1* 0.7* 1.4* 31.9* 36.5* 14.6*
local
Table 10.2 Central, state and local government expenditures by function in % of total expenditure – 1996
97.9 48.3 100.0 27.8 66.4 62.9 100.0
52.6 68.2 100.0 16.6 5.2 56.7 70.5 70.3 17.3 42.9
2.1 40.8 0.0 44.7 25.1 18.5 0.0
25.9 12.4 0.0 28.1 43.3 33.0 21.2 9.4 11.4 36.2
central state
USA
0.0 11.0 0.0 27.5 8.6 18.6 0.0
21.5 19.4 0.0 55.3 51.5 10.3 8.3 20.3 71.3 20.9
local
Rethinking Subsidiarity in the EU
343
Treaties are sometimes very general and at other times specific, so that much of the governance by the EU is based on secondary legislation and other acts. Binding legal acts are divided into: (i) Regulations that are directly applicable without national implementation, (ii) Directives that are binding but require national implementation and (iii) Decisions that bind all the parties concerned. In addition, the EU Commission issues a number of softer non-binding documents. White Papers, signalling legislative strategies, are an example of such documents. Besides the activity that is legislative in nature, the European Court of Justice has a key role as it both interprets EU Law and seeks its application and enforcement. Finally, international agreements negotiated by the EU are a further form of EU activity that affects its member states.10 The legislative activity of the EU has grown significantly over the years, as is indicated by recent studies by Alesina et al. 2002 and Pollack, 2000. In their paper the authors show the number of EU legal acts (Directives, Regulations and Decisions), EU Court Decisions, international agreements and non-binding documents for five year periods. There has been a very rapid growth in the regulatory activities of the EU, while the expenditure share has risen only very slightly. This indicates that a very significant part of EU intervention has the form of regulation rather than direct spending. Alesina et al. (2002) provide information only on the total number of activities and policies of the EU and it is, of course, limited in three important respects. First, pure numbers do not indicate the importance of specific regulatory acts or public expenditures. Second, the legal acts reported are not good indicators of the acts that were actually implemented. Many directives have not been implemented in national laws and, what is worse, some countries have an extremely poor record in terms of enforcing the laws and directives. Third, the data have not been classified according to the functions that might or might not properly belong to the domain of the EU as discussed above. This limitation is difficult to overcome, but fortunately the authors have considered the activities and policies of the EU from the viewpoint of economic functions. We will next describe the regulatory activities of the EU using this study. Alesina et al. (2002) divide the functions of the EU into policy domains as follows: 1. International trade: the creation of the common market and external trade policy; 2. Common Market: free mobility of goods, services, capital and people; 3. Money and finance; 4. Education, research and culture; 5. Environment; 6. Sectoral business relationships: agriculture (and fishing), industry (including energy) and transport;
9780230_547018_12_cha10.indd 343
3/19/2009 8:22:28 PM
344
Designing the European Model
7. Non-sectoral business relations: laws, market competition and state subsidies; 8. International relations: defence and foreign policy; 9. Citizen and social protection: home affairs, justice, consumer protection, civil rights, health, labour relations and so on.11 This categorisation reflects only partially the basic duties of the EU as outlined in Section 2 and below in Sections 4–7. Categories (1) and (2) include primary functions of the EU and for most of them the allocation of the tasks to the EU level of governance is largely evident.12 In category (3) monetary policy has been delegated to the European Central Bank, a EU level institution, but the “lender of last resort” functions have been left to national authorities. In contrast, the appropriate level of public decision-making on fiscal policy is far from obvious and is currently a hotly debated issue, as discussed in Chapter 9. In categories (4) and (5) there are clearly variations, as some particular issues can be naturally conceived to be governed at the EU level, while others are mostly national or even local. In category (6) agriculture is the oldest, most active but very controversial area of EU policy. By the criteria in this chapter it is not a clear EU level task. In contrast, some other sector policies in (6) may provide public goods or spillovers. Category (7) contains both natural EU level activities (for example competition policy), but state aids and subsidies can also be counterproductive, possibly undermining the single market programme. Finally, category (9) contains both policy domains with clear supra-national aspects (such as migration and justice) and others with far less clear EU level governance requirements (such as social protection and structural and regional funds). Alesina et al. (2002) show in their paper that category (6), Sectoral business relations, constitutes the most active area of EU activities and policies. Moreover, category 6a (agriculture and fishery) is by far the largest subgroup within it. These activities are not obvious areas of EU intervention and we will consider them further below. In the other domains of EU policy, categories (1) and (2) International trade and the Single Market are sizeable and quite naturally so since they concern the key EU function, the creation and protection of the single market. There is significant growth in the share of some other categories, especially in (9) Citizen and social protection and (7) Non-sectoral business relations. Taken together, the total number of EU legislative and non-binding acts has shown significant growth over the years. 10.3.3
An assessment
EU activities and policies are regularly assessed by the citizens of its member countries in the opinion survey conducted annually by the Eurobarometer. Table 10.3 reports the average results for the EU as a whole from the latest survey.13 Table 10.4 reports the country-specific results as deviations from the corresponding EU average in Table 10.3. The two columns for the EU in both
9780230_547018_12_cha10.indd 344
3/19/2009 8:22:28 PM
Rethinking Subsidiarity in the EU
345
tables describe, respectively, the percentages of people favouring either decision-making at national government level, marked as ‘Nat’, or decision-making as a shared responsibility of the EU and national government, marked as ‘EU’. Table 10.3 shows that citizens of the EU member countries would clearly like to give a shared role to the EU with national governments in monetary ● ● ●
and fiscal matters, environmental issues, international relations (including humanitarian aid, poverty and exploitation of human beings), Table 10.3 Eurobarometer-Policies: National or EU level decision-making – Results in % of EU averagea EU 15
Defence Environment Currency Humanitarian aid Health and social welfare Media Poverty/social exclusion Unemployment Agriculture & Fishing Regional aid Education Research Information EU b Foreign policy Cultural policy Immigration Political asylum Organised crime Police Justice Accepting refugees Juvenile crime Urban crime Drugs Exploit. Hum. Beings Terrorism
NAT
EU
45 33 31 24 59 56 30 44 40 32 61 27 20 22 49 48 45 25 63 58 43 51 56 26 16 12
51 64 65 72 37 38 67 53 54 63 36 68 74 71 44 49 51 72 34 38 53 45 40 71 80 85
Note: aDifferences between “NAT” and “EU” and 100 is the percentage of don’t know. bInformation about the EU, its policies, institutions and bodies. Source: Eurobarometer Spring 2002, p. B 43, 44, http://www.gesis. org/en/data_service/eurobarometer/ index.htm.
9780230_547018_12_cha10.indd 345
3/19/2009 8:22:28 PM
346
Designing the European Model
Table 10.4
Eurobarometer-Policies: National or EU level decision-making – Results B
DK
D
GR
E
F
IRL
Nat EU Nat EU Nat EU Nat EU Nat EU Nat EU Nat Defence –10 Environment 2 Currency –9 Humanitarian aid –4 Health and social 3 welfare Media –3 Poverty/social 3 exclusion Unemployment 0 Agric. & Fishing –10 Regional aid –7 Education 8 Research –6 –4 Information EU b Foreign Policy –8 Cultural Policy 3 Immigration –10 Political asylum –9 Organised crime –7 Police 6 Justice 2 Accepting refugees –9 Juvenile crime 5 Urban crime 1 Drugs –1 Exploit. Hum. Beings –4 Terrorism –4 Note:
a
EU
10 –1 11 5 –2
12 11 14 13 23
–11 –9 –13 –11 –21
–7 –5 –5 1 0
6 6 5 –4 5 –8 –2 –2 0 –23
–4 6 10 4 25
–7 –7 –8 –6 –7
6 6 7 6 8
3 –1 –4 –4 10
–3 2 6 5 –9
18 10 –1 –3 –1
–21 –13 –2 –1 –1
4 –3
18 –15 8 –7
2 –3
–3 –12 1 –5
15 6
–4 –7
3 6
–1 3
2 –2
–7 1
3 –4
0 12 9 –8 7 5 9 –1 10 10 8 –6 0 9 –5 1 2 5 5
18 4 –3 11 5 15 16 31 17 18 3 8 25 25 22 19 1 1 1
–2 –7 –5 –1 6 –3 –3 2 8 9 –6 –2 3 6 –3 –2 –4 –1 –1
1 6 4 –2 –7 2 3 –2 –9 –8 6 0 –3 –5 2 2 3 0 0
16 3 9 18 10 6 2 8 14 14 6 13 10 9 25 26 6 0 0
–4 –7 –7 –7 –7 –4 –7 –12 –17 –15 –7 –11 –11 –11 –17 –13 –2 0 0
3 7 7 6 6 3 6 12 14 13 6 9 10 10 16 12 1 –1 –1
1 –1 0 3 –1 4 11 –10 –10 2 –1 1 –3 5 –8 1 0 –1 0 1 1 2 2 10 –4 4 10 –4 4 9 –3 3 10 –1 1 0 1 0 0 –7 8 17 7 –6 5 4 –3 –1 –4 4 2 –1 1 0 –1 1 0
–3 –6 7 –4 3 –1 –4 –11 –15 –13 –12 –3 –2 –20 –7 –1 –4 –3 –3
–17 0 5 –10 –3 –13 –13 –28 –17 –16 –1 –7 –25 –24 –20 –16 1 1 1
–14 1 –6 –17 –8 –2 2 –4 –15 –14 –4 –12 –8 –8 –22 –24 –4 1 1
Difference between ‘Nat’ and ‘EU’ and 100 is the percentage of don`t know.
Source: Eurobarometer Spring 2002, p. B 43, 44;
9780230_547018_12_cha10.indd 346
3/19/2009 8:22:28 PM
Rethinking Subsidiarity in the EU
347
in % by countrya (deviation from EU average) I
b
L
NL
A
P
FIN
S
UK
Nat
EU
Nat
EU
Nat
EU
Nat
EU
Nat
EU
Nat
EU
Nat
EU
Nat
EU
–14 –6 –20 –10 –14
15 6 21 10 13
–8 0 –15 5 8
7 –1 16 –5 –8
–11 –3 –11 1 5
9 3 11 –2 –4
12 14 0 17 9
–12 –14 1 –17 –8
16 14 10 9 4
–19 –17 –12 –11 –6
46 25 15 24 29
–44 –23 –13 –23 –26
33 14 14 20 30
–31 –12 –14 –18 –27
13 12 35 6 4
–13 –11 –37 –8 –4
–9 –9
8 9
–8 4
9 –4
4 12
–5 –13
10 4
–9 –6
10 8
–11 –11
–10 17
11 –17
19 7
–15 –5
4 6
–5 –7
–17 –5 –1 –16 –12 –6 –10 –13 –19 –20 –3 –10 –12 –18 –14 –6 –8 –5 –5
17 3 2 15 13 7 10 13 18 18 3 9 12 17 14 6 8 5 5
10 2 –6 9 –14 –3 –6 5 4 –1 –3 2 4 12 1 –3 12 –4 –4
–11 –1 7 –10 15 3 7 –3 –6 0 2 –3 –4 –12 –1 3 –12 3 3
12 0 –1 12 –1 3 –1 10 –10 –7 –7 2 –8 –5 8 8 8 0 0
–12 1 0 –11 –1 –2 2 –10 9 8 7 –2 8 5 –6 –7 –8 0 0
–1 16 1 5 4 5 6 14 20 21 2 10 14 22 –5 –13 2 3 3
0 –17 –2 –5 –3 –5 –5 –12 –20 –20 –2 –9 –13 –21 5 14 –2 –2 –2
6 11 3 8 9 21 9 13 9 8 6 8 10 7 2 4 11 9 9
–8 –11 –5 –10 –11 –22 –10 –13 –13 –11 –8 –10 –11 –10 –5 –7 –13 –11 –11
21 33 7 16 14 8 10 13 35 24 2 15 16 39 13 –1 6 3 3
–21 –29 –5 –16 –14 –7 –10 –12 –34 –23 –2 –14 –15 –37 –12 1 –7 –2 –2
8 16 4 7 11 11 13 24 18 14 –3 17 17 25 20 14 10 –2 –2
–6 –12 –1 –6 –9 –8 –11 –21 –17 –12 5 –15 –16 –23 –17 –11 –8 3 3
14 12 1 12 8 4 12 3 17 16 20 13 9 16 18 15 11 6 6
–15 –13 –4 –13 –8 –7 –14 –5 –17 –16 –20 –12 –8 –16 –18 –14 –11 –7 –7
Information about the EU, its policies, institutions and bodies
http://www.gesis.org/en/data_service/eur obarometer/index.htm
9780230_547018_12_cha10.indd 347
3/19/2009 8:22:29 PM
348 ● ● ●
Designing the European Model
research, global crime protection and regional aid.
According to the poll, EU activities should not be focused very much on ● ● ● ●
education, agriculture and fishery, health and social welfare and unemployment.
The survey results in Table 10.4 for individual countries indicate some interesting country-specific opinions. First, citizens’ opinions in some countries follow the EU average rather closely. Taking (plus or minus) 20 point deviation from the EU as a criterion, it can be seen that opinions in Belgium, Germany, Spain, France, Luxembourg and the Netherlands do not differ radically from the EU average opinion. Second, citizens in the Scandinavian countries show a strong preference for respecting subsidiarity, and this is so especially for foreign policy, humanitarian and social issues, refugee policy, justice and crime. Third, there are some areas for which country opinions do not differ radically from the EU average. These areas are media, poverty, regional aid, education, research, foreign policy, police, drugs, exploitation and terrorism. The empirical data yield a rather clear total picture. First, in terms of public spending, the EU is quite far from the expenditure patterns of federalist nations. The EU budget share in GDP is quite small – indeed it is much smaller than the corresponding share in federalist countries. Thus the balance of government tasks in the EU is very much geared towards national governments, which accords well with the principle of subsidiarity. The EU is not close to being a federation, even though the non-budgetary interventions of the EU run to some extent counter to this conclusion. Second, the tasks that the EU level public administration is currently undertaking do not accord that well with the opinions of its citizens. Third, the current tasks of the EU do not match very closely the economic principles for decentralisation of public sector responsibilities, which are considered further in Sections 4 to 6 below.
10.4
Activities for an economic union: examples and cases
We have seen that, for a number of activities, EU level public intervention can be justifiable for efficiency reasons. We will now discuss key activities for which union level government intervention is clear-cut in our opinion. However, we will also consider some activities that are more controversial in this respect.
9780230_547018_12_cha10.indd 348
3/19/2009 8:22:29 PM
Rethinking Subsidiarity in the EU
349
10.4.1 Public goods Section 2 argued that, for efficiency reasons, there should be union level decision-making for public goods for which the geographical distribution of benefits extends widely across different nations within the union. The clearest example of public goods with such benefits is probably national defence and foreign policy associated with external security. Strong defence by a country can bring major benefits to friendly neighbouring countries as it enhances their security, for example, against threats against the territory of the former. Significant cost savings exist moreover in compatible weapons systems and common weapons development, exchange of military and security information and a common EU foreign policy diplomatic corps for external security. The existence of significant public goods aspects in national defence and external security policy are evidently areas where decision-making at the EU level is justifiable from an economic viewpoint. Efficient provision of these services can nevertheless take a variety of different forms and it need not necessarily be fully centralised in an EU agency even if overall responsibility should rest with EU administrative structures. At the moment the role of the EU is quite limited in national defence and security, since NATO has in effect had this role already for many years. Some EU countries are, however, not members of NATO, and this situation makes the role of the EU fairly complex. In the Eurobarometer opinion survey, defence was not seen as a top priority for EU level action. Evidently, there is major heterogeneity in preferences across citizens of different member countries and the conclusion is less clear-cut than economic considerations suggest (see, however, Persson et al. 1998). Other related goods and services, of which public goods aspects can be geographically widely dispersed within the EU, are internal security and border control, as discussed, for example, by Tabellini (2002). With the Schengen agreement in place, the benefits from joint decision-making about the forms of border control are apparent. Correspondingly, there are potentially large benefits from creating joint mechanisms in internal security such as Europol and close co-operation between national police in the member countries. The importance of joint action in internal security has become apparent after recent terrorist attacks. Apart from national defence and security policy, it appears that there are relatively few public goods for which the benefits accrue to the EU members jointly. Other public goods, such as those associated with culture, are often national or even more local in character.14 Provision of cultural services raises difficult conceptual issues. Some forms of culture can be provided through the market system, while other forms seem to require some form of public intervention. It appears that there is little need to have union level decision-making or intervention for local or national public goods and services.
9780230_547018_12_cha10.indd 349
3/19/2009 8:22:29 PM
350
Designing the European Model
10.4.2
Externalities and spillovers
Cross-border externalities and spillovers are the other main category for possible governmental decision-making and intervention at the EU level. This is the case when benefits and/or costs from externalities and spillovers are widely dispersed across the member states, so that a national jurisdiction (or contracting between a limited number of countries) may not be able to internalise these benefits and costs. Financial markets with interbank cross-border exposures provide an example of an externality which is potentially EU wide. This case will be discussed in Chapter 13. Here we take up other economic activities in which geographically widely dispersed spillovers or externalities can also arise. Networks Trans-European networks, for example in transport and in research, are good examples of operational structures that try to reap benefits from spillovers arising as an integral part of the formation of a network. Compatibility of different national parts of a network and free or low-cost use of such networks leads to reductions in operating costs and can bring economic benefits from the existence of a larger number of nodes in the network. Free access to national networks of for example roads and railways for users from different countries is very important for reaping the full potential benefits from these networks. If investment into and provision of a national network is limited to the country in which the network lies, then risks of under-investment are a real concern if important benefits from the network accrue to users from other countries. Freeways and rail transport network with intensive traffic across national borders by compatible equipment (for example trains) may need to be supported by intergovernmental intervention, possibly at the EU level.15 It is, however, important to distinguish clearly between network activities that benefit all EU countries and others that affect only two or a few countries. Investments in telecommunications, for example, belong to the former category, since each country will contact every other country. Road and railway networks, on the other hand, belong to the second category. The border crossing connections will typically help two countries only, and sometimes a third or a fourth country is affected, but rarely if ever do they involve a higher number of countries. Thus the EU should not subsidise such connections. The EU could, however, have a useful role in terms of coordinating and facilitating the negotiations between the countries involved and to design common standards for inter-community networks. Research is an activity that can involve significant externalities. For example, a specific scientific discovery can provide input to additional discoveries. The additional discoveries may well take place in different countries, which indicates that the externalities from research satisfy the requirement of geographically wide dispersion of benefits. In any case, the
9780230_547018_12_cha10.indd 350
3/19/2009 8:22:29 PM
Rethinking Subsidiarity in the EU
351
initial producer usually incurs all of the costs of the discovery but may not be the recipient of the benefits from the additional discoveries. Moreover, co-operation of different researchers can lead to increases in productivity, which suggests that networking by researchers can facilitate research. In these situations the research system may not operate at an efficient level with too little investment in research. On the other hand, research often contains an element of a race to be the first to find a result which in itself implies overinvestment and unnecessary parallel activities. Moreover, very successful research is carried out by private universities in the United States, which is obviously a decentralised way of organising research. These elements diminish the rationale for EU support and centralised action through grants as a natural vehicle for internalising the benefits and costs of research activity and for achieving productive efficiency in research. This impression is strengthened if account is taken of the fact that, at the moment, the administration of these tasks in the EU is very bureaucratic, involving very long administrative lags and insufficient quality control, which can nullify the benefits from the intervention. Moreover, it seems that current EU rules for research grants imply a bias in the distribution of the available funds in favour of smaller countries, which for natural reasons have more cross-border connections than larger ones. This bias leads to a systematic redistribution at the expense of large countries. We conclude that trans-border networks and public goods with geographically dispersed beneficiaries are important parts of EU level public infrastructure whose governance can be facilitated if the EU helps the countries involved to co-ordinate their decisions. However, normally no financial implications should be involved. In some other activities, the spillover or public good aspects can be much less dispersed. If so, centralising the public intervention to the top level of government is not the right answer from the viewpoint of economic efficiency. We next take up a few other examples of spillovers or externalities, in which the EU-dimension in the governance of these activities is even less clear cut. Environment Environmental concerns have an international dimension in a number of economic activities. Trans-border pollution provides examples of negative externalities that extend beyond national jurisdictions. When such externalities exist, there is a case for negotiation between the countries concerned. It appears that relatively few environmental externalities specifically have an EU wide geographical dimension. In many cases the externality is much more local and in some cases, such as global warming and climate control, the externality is instead global. The principle of subsidiarity suggests decentralisation of the regulation of environmental externalities that are local, national or a concern to only a
9780230_547018_12_cha10.indd 351
3/19/2009 8:22:29 PM
352 Designing the European Model
few neighbouring countries, though there can be benefits from having the EU set up a common framework for such situations. The regulation appropriate for global warming and any other global externality is an EU level duty. The EU alone is not the sufficiently high level of public decision-making, but it is the natural party in international negotiations to control a global environmental externality. Ecological dumping as a result of too lax regulation has been raised as a possible concern when independent countries set their own environmental standards. Opinions differ between academic researchers about the consequences of regulatory competition in environmental standards. These differences are in part due to different circumstances about possible environmental damage. If the latter occurs only within the borders of countries and if the profit from being able to freely use the environment accrues to domestic residents only, then national regulation will not lead to ecological dumping (see Long and Siebert 1991 and Oates and Schwab 1988). This reinforces the previous observation that geographically limited pollution problems do not require the centralisation of public regulation. In contrast, the outcome is not clear-cut when there are international spillovers of environmental damage and if foreigners are, to an important degree, owners of polluting firms as is shown in Sinn (2003). The outcome will then very much depend on the type of regulations that are used by the different countries. If international co-operation is required, then a suitably designed system of tradable pollution permits, which last only a limited time and are regularly sold by the governments, can in principle lead to an outcome that internalises the spillovers and externalities. The administration of such a system requires international co-operation, and the principle of subsidiarity does not imply a decentralisation of this activity. Natural resources Management of commercial fishing, where EU interventions have been attempted, provides a complicated example, in which the fishing activity by a single fishing unit can exert a negative externality on other fishing units. This is a result of missing property rights in the fish population itself. The latter are an example of a common property resource since in the water the fish migrate across national borders. Such resources can be subject to over-extraction.16 Any single fishing unit does not take fully into account the effect of its fishing on the fish population since the fishing unit does not own the resource. Over-fishing results when all fishing units behave in this way. This is a clear case in which centralised action is necessary. Fishing may be contrasted with agriculture, which is a very different type of activity exploiting a natural resource, land. In contrast to fishing, it is easy to define property rights to land and these rights were indeed defined in Europe a very long time ago. The justification for public regulation of fishing does not apply to agriculture, which is a standard form of
9780230_547018_12_cha10.indd 352
3/19/2009 8:22:29 PM
Rethinking Subsidiarity in the EU
353
production and business activity. Thus the common agricultural policy of the EU (CAP) cannot be defended on the basis of a spillover or externality. Traditional landscape and other similar concerns can be seen as a public good, but if so, certainly as one whose benefits accrue primarily to local or national citizens. Subsidies leading to increased production are not a proper way to take account of such an “aesthetic” public good produced at the national level. On the contrary, such subsidies have led to an overly extensive form of agriculture, which has created substantial environmental damage (for example pollution of ground water with nitrates) and has often contributed to a destruction rather than preservation of the landscape. Support should be geared to the preservation of the landscape itself or environmental improvement and not to agricultural production. These arguments should be kept in mind if the discussed reform of CAP is geared to the introduction of national support to agriculture. National subsidies should not be tied to production or exports but rather to the preservation of the rural way of life if a member country finds national support to be in its interest. Production or export subsidies would also work against a unified single market in which such subsidies would frustrate the forces of comparative advantage and destroy the “level playing field” of the single market. The 2002 Report of the European Economic Advisory Group discusses further a blueprint for reform of the Common Agricultural Policy. Quotas and markets for licences for the extraction of a common property resource can provide an efficient solution to the problem of over-extraction of fish populations. The management of such schemes requires public sector intervention, and the market for fishing rights should be administered by some public agency. The level of government for the management of fishing rights seems to depend on the geographical area where the fish population resides and to which fishing fleets from different countries have access. It is not obvious that the EU administration is always the appropriate level of public intervention in all cases. Some fish populations are relatively local, though free access of other EU member countries can make management of a local fish population an EU level concern. For some other fish populations a global level would, in principle, be the right level of public intervention. In the latter, the EU can be the right type of public body for international negotiations and contracts for management of common property resources. To sum up, while there are environmental reasons for common EU policies to regulate fishing, there is no economic reason whatsoever for agricultural policy to be carried out by the EU. The subsidiarity principle dictates that agricultural policy be shifted to national authorities as quickly as possible. Standardisation and product quality Provision of a common standard for economic activities can be viewed as a particular type of networking. There are clear benefits from well-chosen
9780230_547018_12_cha10.indd 353
3/19/2009 8:22:29 PM
354
Designing the European Model
standards, since wide usage in different countries can provide benefits to individual users or lead to lower costs of production and information to users. In many activities there is a clear international dimension to standardisation, so that its governance at the EU level can be warranted. Harmonisation of product quality standards across the member countries can be efficient, since it lowers the cost of gathering and understanding of product information to consumers. Moreover, competition in product quality regulation can lead to too lax product standards, since individual countries can try to minimise production costs for domestic firms. When all countries engage in this race, a non-optimal outcome can result. Thus for many goods and services, common standards and the harmonisation of product quality information across the EU appears to be an efficient level of regulation. For some commodities, in which for example the use of hazardous substances is possible, product information and safety is a major concern, while for other goods this issue is of little importance. In the former case, it can be necessary to establish Europe-wide standards in which case an EU level supervisory administrator is required. Food and drugs are a case in point.17 In other cases, private provision of the evaluation activity for standardisation of products and assessment of product quality may be preferable to public provision. Michelin guides for restaurants provide a simple example of private supply of information on product quality for those who prefer the French cuisine. The wide range of cases in product standards indicates that there is no single model of regulation or level of regulatory administration that fits all the different possibilities.
10.5 Tax and infrastructure competition The international mobility of some productive factors, especially capital, has widely raised concerns about tax competition and about under-provision of public goods. It is argued that independent countries have strong incentives to reduce tax rates for mobile factors and that this would lead to too low levels of some forms of taxation and too high levels for others.18 An important related argument is that, in order to attract businesses, independent countries also compete for mobile factors of production by providing excessive public infrastructure for businesses. These two arguments, low tax rates for mobile capital and excessive public infrastructure to attract businesses, raise a number of different issues and require a closer analysis (see Sinn 2002, chapter 2 for a clear exposition). A key starting point is the nature of public goods that the public sector needs to finance. One possibility is that the public goods solely benefit wage earners or consumers. In this case tax competition prevails: a tax on the mobile factor is entirely shifted to the immobile factors, the income of the immobile factor is reduced, and the tax revenue from such a tax is insufficient to compensate for the losses of the immobile factor.
9780230_547018_12_cha10.indd 354
3/19/2009 8:22:29 PM
Rethinking Subsidiarity in the EU
355
In contrast, the outcome is different when the public good is an infrastructure good that benefits the firms: tax rates will not be driven to zero and competition in taxation and infrastructure can yield productive efficiency in both the international allocation of capital and the provision of public infrastructure. In this last case, concerns may still remain about the financing of the public sector budget, as the revenues from taxing the mobile factor need not be sufficient. A proper resolution to the financing problem is to use a co-ordinated requirement of full self-finance of infrastructure goods from taxes on the mobile factor by all countries. The EU could, for example, extend its subsidy ban to the case of implicit subsidisation through under-priced infrastructure provision. This would resolve the issue of financing difficulties without necessitating a co-ordination of infrastructure. A difficulty with this solution is that it needs even more extensive monitoring than the present interpretation of the subsidy prohibition. It should be stressed that tax harmonisation is not a proper remedy for the financing problem, since it intensifies jurisdictional competition by providing strong incentives for excessive investment in public infrastructure. When the tax rates on mobile capital are fixed at the EU level, each single country has an incentive to compete for mobile capital by providing the needed infrastructure, but the amount of infrastructure investment resulting from this competition will be larger than what is efficient from an international perspective. The harmonisation of capital income taxes would require infrastructure harmonisation to prevent this overinvestment effect, but in view of the prevailing economic and geographical divergences among the EU countries, the latter cannot meaningfully be achieved. Alternatively, all taxes would have to be harmonised such that there are no sources of funds for excessive infrastructure investment. All this is hard to imagine for the time being. Are there any good methods for resolving the financing problem with little international co-ordination? One possible remedy is the residence principle of taxation whereby taxes are levied at the recipient of the income rather than at the source of income. With resident taxation it is not possible to evade taxes by simply transferring the capital to another location. It may, however, be difficult to administer residence taxation, since it requires reporting of income earned abroad. Moreover, tax competition can also ensue through investors shifting residence rather than the location of mobile capital. Another possibility is to design the corporate tax system with zero marginal taxes on new investment but with positive average taxation. Such a system attempts to tax only dividends but not retained earnings, so that investment financed at the margin from retained earnings are tax-free. However, such tax forms entail the difficulty that new investment need not always be tax-free. While retained earnings provide marginal finance for new investment for mature firms, start-ups usually require new equity and dividend taxes raise their cost of capital.
9780230_547018_12_cha10.indd 355
3/19/2009 8:22:30 PM
356
Designing the European Model
In general, the principle of subsidiarity in the design of the tax system is facing serious difficulties due to the increased mobility of productive factors. Different degrees of mobility lead to both low taxation and less redistribution among productive factors. With a deepening of integration it is possible that, in addition to capital, labour will also become much more mobile in the long run. If this happens, decentralised taxation by member countries will face increasing difficulties from tax competition and widespread erosion of tax bases. A well-functioning central government is an answer to ruinous competition when all key factors of production are fairly mobile.
10.6 Labour markets, social standards and the welfare state With the enlargement of the EU, questions about social standards in employment and labour markets and the future of the welfare state, that is redistributive capability of the state have become subject to active discussion. The coming enlargement is different from the earlier ones in a number of respects. First, migration to the existing EU countries has been severely constrained, first by the Iron Curtain and then by Western immigration laws, such that the migration pressure will not have been relieved at the time of enlargement. Second, accession countries have, with some exceptions, quite low GDP per capita and low wages relative to the EU average. Currently, monetary wage costs per hour are one fifth to one tenth of those in the richer EU countries. Moreover, social standards in employment and redistribution and welfare benefits offered by these countries are much lower than in the current EU member countries. These differences imply that there are tensions concerning the welfare state, social standards and the migration of productive factors. The differences in wages and social standards can be the source of potentially large migration flows of productive factors. It is likely that some capital will migrate from the current EU member countries to the accession countries, while the reverse is true for labour. It is difficult to forecast the magnitude of possible migration flows.19 However, it is important to understand the nature of the tensions and discuss the appropriate policy responses. 10.6.1 The welfare state From a personal ex ante viewpoint, income redistribution and the welfare state are to be seen as insurance systems. They protect those citizens who will experience unfavourable personal circumstances, such as long-term illness or unemployment. The welfare state is a response to these insurance needs,20 and if there is no factor mobility we can think of citizens as contributors to the insurance system and beneficiaries when they encounter hardships. This kind of closed system no longer works well when there is factor mobility. Mobility can provide a way to enjoy the benefits of the welfare
9780230_547018_12_cha10.indd 356
3/19/2009 8:22:30 PM
Rethinking Subsidiarity in the EU
357
state, while at least partly avoiding the insurance payments. When borders are open and people are mobile, differences in welfare systems can induce migration. People in need of benefits move to the country with relatively high benefit levels, while “healthy and lucky” people (who are net payers to the system) move out of the high benefit and high cost country. With such movements, the funding of an advanced welfare state runs into trouble and there are pressures to reduce both taxes and benefits of the system. The magnitude of these pressures depends on the degree of factor mobility and is difficult to forecast. There is probably little mobility among the current EU citizens. However, the differential mobility to various Western EU countries of people deciding to emigrate from Eastern Europe because of the huge current income differentials could be extremely high, implying a high sensitivity of migration to even small differences in living standards among the target countries. One policy response to the pressures resulting from the mobility of people would be to change the current principles of eligibility for welfare benefits. Currently, EU welfare states practice a dichotomous approach to migrants within the EU. People who come to live in another country for reasons other than working there, are usually excluded from welfare benefits of any kind and have to rely on the benefits received from the home country. On the other hand, people who come with the intention to work are fully included.21 Thus, a home-country principle applies to non-working migrants and a residence principle applies to working migrants under current EU law. If the migration-induced pressures on the welfare state stem exclusively from the application of the residence principle, a potential solution could be a strengthening of the home-country principle by applying this principle to at least some of the benefits received by working migrants. Such a reform would reduce artificial migration incentives due to differences in welfare systems, though there would still be migration for normal economic reasons. Of course, as with all measures discussed in this chapter, such a move would require a substantial reform of the existing EU laws, which might meet with resistance on equity grounds. However, it seems to us that the alternative, which currently is being envisaged by the EU Commission, that immigration will be held in check by quotas and other quantity constraints for many years to come, would be a much more severe infringement on the right of free migration granted in the Treaty of Rome. The Treaty of Rome is not, however, in conflict with a solution which permits every EU citizen who wants to migrate to migrate, but without receiving any gifts. In practice, the home-country principle is difficult to be fully implemented. The provision of free public goods with benefits geographically accruing to only that country, is naturally funded by taxes from factors actively working in that country. However, it might be possible to try to apply the homecountry principle in a limited way, whereby migrants would be immediately
9780230_547018_12_cha10.indd 357
3/19/2009 8:22:30 PM
358 Designing the European Model
entitled to contribution-financed benefits, but only gradually entitled to social benefits that are funded from general tax revenues.22 The idea of gradual access to social benefits is designed to minimise the fiscal implications that arise from free mobility and differences in the welfare systems between countries. These implications are difficult to assess, as illustrated, for example, by the case of differences in the skill levels of immigrants. Highly skilled workers who have obtained their education in their home country can be a net benefit to the country of immigration. Correspondingly, they represent a cost to their home country. Such a brain drain is sometimes seen as an important policy concern (see Chapter 14 for a discussion of the question of a brain drain from Europe to the US). 10.6.2
Social standards
A related concern about the coming EU enlargement involves social standards of work, including fringe benefits, work safety, pension schemes and welfare benefits. It is often argued that poorer countries practice “social dumping” by having lower social standards and thereby gain an unfair competitive advantage by avoiding the costs of higher standards. Harmonisation of these standards is proposed in order to eliminate the seemingly unfair cost advantages of low standards. Such arguments are problematic as they originate from a static view of the world. They ignore the fact that different countries are in very different stages of economic development and that wages are also very different. It is well known that non-wage benefits and work standards are positively related to wages. Thus it seems likely that as poorer countries become more advanced and catch up with richer countries, internal competitive and social pressures in these countries will lead to both higher wages and higher non-wage labour costs. It is natural to ask whether harmonisation of labour standards would facilitate the process of economic development and catching-up or whether it might even have harmful effects on the development process. An answer to this question requires an analysis of the dynamic forces of development that occur when a relatively poor country joins an area of well-developed countries.23 A typical model of the development process predicts that in the short run there will be migration of part of the work force of the poor country to the richer area. This will reduce labour supply, raise wages, destroy less productive jobs and induce the (supposedly benevolent) government of the poor country to raise work standards in line with wages. The poor country will also enjoy a gradual inflow of capital from the rich area. This will increase demand for labour leading to a further increase in wages and social standards. In due course, some of the guest workers from the poor country will gradually return to their home country. Eventually, the poor country will catch up with the rich area. When this happens there will be factor price equalisation with wages and social standards reaching the level of the rich area.
9780230_547018_12_cha10.indd 358
3/19/2009 8:22:30 PM
Rethinking Subsidiarity in the EU
359
Box 10.2 German unification and social dumping From an economic perspective, German unification has failed. There was an initial catching-up until 1996, which was largely induced by massive tax incentives and investment subsidies that made the cost of capital negative for most investment projects and induced excessively capital intensive investment. However, the East German economy has stagnated and the gap between East and West Germany has widened. GDP per person of working age in the East has been falling relative to the corresponding value in the West for the last six years, and currently stands at 58 percent. However, due to massive public transfers, nominal incomes per capita are at a level of more than 80 percent of that of the West and pensions are at 110 percent of those in the West. In real terms, about 10 percentage points can be added to these numbers since the prices of non-traded goods, in particular the government-controlled rents, are substantially lower than in the West. The East German economy absorbs much more resources than it produces, the current account deficit being about 45 percent of GDP. This is huge. Even the current account deficit of the Italian Mezzogiorno is only 13 percent of GDP. Two-thirds of the East German current account deficit – which amounts to nearly 5 percent of West Germany’s GDP – is financed by public transfers via the federal budget, via revenue-sharing agreements among the Laender (Finanzausgleich) and, primarily, via unemployment and pension benefits. One-third is financed by private capital flows, of which, however, a substantial fraction feeds an increasing stock of East German public debt which has reached a higher per capita value than that in the West. Unemployment in East Germany currently hovers around 17 percent, and there are regions where it is way above 20 percent, even though many unemployed have been hidden in early retirement schemes and training programs. Regular employment has been shrinking at an annual rate of nearly 2 percent since the middle of the 1990s, long after three-quarters of East German industry had closed down. While the industrial sector is growing at a solid rate, albeit from a very low base, the overall prospects for the East German economy are far from satisfactory. The aggregate net public resource transfer from West to East has been about € 800 billion, which is ten times more than the amount even the most pessimistic politicians had dared to forecast at the time of unification. One reason for the economic disaster in East Germany is an excessive fear of social dumping which led to rapid wage convergence and an immediate jump in social standards. The cost of labour increased much faster than aggregate productivity, resulting in mass unemployment. High wages combined with tax incentives induce overly capital intensive investment, and they also made East Germany an unattractive location for international investment. Before unification, Eastern wage costs stood at about 7 percent of those in the West at the then prevailing exchange rate (4.3:1). With the 1:1 currency conversion in the summer of 1990, wage costs jumped to about 30 percent. Wage negotiations that followed in late 1990 and early 1991 specified a wage adjustment path reaching the West German standard wage level in only five years. Actual wages initially followed this path, but eventually increased more slowly, since the privatised and newly founded firms decided to leave the employers’ associations or not to participate in these associations in the first place. Currently, about
9780230_547018_12_cha10.indd 359
3/19/2009 8:22:30 PM
360 Designing the European Model 85 percent of East German firms with a majority of all employees are not covered by union contracts. The wage negotiations that followed unification had been proxy negotiations. They were carried out by the newly founded East German trade unions, which were completely under Western control, and the employers’ associations which had come from the West. Prior to privatisation there had been no private employers in the East who could have participated. Both bargaining parties agreed that rapid wage adjustment was needed to safeguard West German jobs and prevent foreign competitors from buying up East German firms and thus entering the German market. In doing so, they were supported by a firework of superficial arguments provided by politicians who forecast “flourishing landscapes” in “three to five years” and warned of massive migration flows which otherwise would have had to be expected. The results of the proxy wage negotiations were flanked by the social union which implied high replacement incomes and forced market wages upward. The social union was introduced in the summer of 1990 in addition to the monetary union. East Germans were included in the Western pension system, received Western type unemployment benefits and were entitled to nearly the Western level of social aid. Social aid was initially higher than Eastern wages and has remained high. A family of four is entitled to social aid and housing grants amounting to 75 percent of the average East German wage. Social aid and similar benefits implied excessively high wage costs and a replacement income with which the market economy was unable to compete.
The preceding reasoning suggests that it would be a mistake to impose the social standards of the rich area on the poor accession country. The forces of development should be left to run their course. A policy of an early and quick equalisation of wages and social standards between rich and poor areas would have disastrous consequences. A large fraction of the jobs in the poor country would become unprofitable following the imposition of high wages and social standards, leading to mass unemployment and emigration out of the poor area. It is also likely that political pressures would emerge for massive transfers from the rich area to the poor country. Most likely, a policy of harmonisation would greatly slow down or even prevent the development process whereby the poor country gradually reaches higher standards of living. These events are well illustrated by the experiences of German unification, discussed in Box 10.2. Differences in preferences of social standards and wages among different countries do not change the preceding conclusions. If one country prefers higher social standards, then wages should be correspondingly lower, other things being equal. With heterogeneity in preferences, the principle of subsidiarity should be applied, that is each country should be allowed to choose its preferred combination of wages and social standards (see Jackman 2001). These considerations suggest that harmonisation of social standards is an incorrect policy in EU enlargement. Instead, it is important to focus on ways to facilitate the process of development that enables the poor countries to catch up with the rich area. Free trade and the provision of new markets to
9780230_547018_12_cha10.indd 360
3/19/2009 8:22:30 PM
Rethinking Subsidiarity in the EU
361
the poor countries is the most important policy in support of this process. With new markets, the poor accession countries will have increased economic opportunities that will enable them to speed up economic growth and gradually raise the living standards of their citizens. The process of economic growth is necessarily slow by its very nature, but centralised interventions in the form of prematurely harmonised social standards, say by the current EU, can only risk a slowing down of this process. There is no miracle cure for an instantaneous closing of differences in productivity, wages and living standards that have built up in the past over many years. 10.6.3
Redistribution between people and between states
The governance of interpersonal income redistribution in the enlarged EU is a delicate matter. The conventional public finance view is that redistribution should primarily be the concern of the highest level of government.24 However, such a view implicitly assumes a certain degree of regional homogeneity, as otherwise the interpersonal redistribution becomes an interregional redistribution or one between countries. This is certainly a matter of concern for the EU, where the income disparities between rich and poor regions will differ by a factor of four or more after enlargement. Any universal redistribution scheme, say one that is based on a common progressive income tax schedule, would systematically distribute income from the richer to the poorer countries. For example, even an uneducated Swedish worker is rich relative to a Spanish government clerk. There would be redistribution of funds from Sweden to Spain. Such redistribution might also be legitimated from a deeper understanding of European solidarity, but it clearly goes beyond the insurance motive that can explain and justify interpersonal redistribution within a country. Above, we warned against ignorance of the differences in the stages of economic development and against premature harmonisation of social standards that can bring serious harm to the process of development and catching-up of the poorer countries. These considerations also suggest that income redistribution through the tax and transfer systems should mostly be left to the national decision-making of EU member countries. Intergovernmental competition by well-meaning governments taking into consideration the productivity differences is probably the best way to achieve systems of redistribution that are in line with differences in labour productivity and also to achieve fast convergence in the growth processes of the different countries. This is particularly true if adverse migration effects are minimised by changing the welfare state as was suggested above. If there is any role for EU level public decision-making, it is in providing opportunities and incentives for fast growth in the poorer countries. Nevertheless, there is the difficult question of whether the EU should purposely redistribute resources from the richer to the poorer member countries. Much of the preceding analysis has viewed the EU primarily as an
9780230_547018_12_cha10.indd 361
3/19/2009 8:22:30 PM
362 Designing the European Model
economic union with the single market as the main objective. However, it is sometimes suggested that the EU is more than a pure economic union even if it is not a federation, namely one that effectively shifts resources so as to even out any pre-existing income differences. Of course, opinions also differ on this question. In practice, the EU carries out large redistributive schemes, which suggests that the EU is something more than a common market. Thus it is possible to think of giving at least part of the redistributive task to the EU level. Looking at current practice, much of the common agricultural policy appears to be motivated by distributional concerns, although such redistribution can hardly be defended since rich countries like France are among the biggest beneficiaries. The various structural and cohesion funds that are the second largest item in the EU budget can more easily be justified by redistributive goals. Their purpose is to make richer countries pay for the funds flowing to the poorer countries, and it seems that this purpose has been achieved. However, the funds should be more than just income transfers in that they help the less developed countries to improve their infrastructure and to support these countries’ own forces of economic growth. The evidence on whether they achieved this goal is mixed. Some schemes appear to have reduced regional disparities (see, for example, De la Fuente and Vives 1995 for a study of Spain). Yet, on balance, it seems that the current EU regional policies have not contributed positively to the catching-up processes of economic growth for the poorer regions of the EU, as discussed in Boldrin and Canova (2001). These results suggest that a reconsideration of the current EU redistributive role would be worth while.
Notes 1. See http://european-convention.eu.int for further information. 2. Devolution of the tasks of the federal government is in some respects an analogous debate in the United States (see, for example, Inman and Rubinfeld, 1997 and the discussion by Musgrave, 1997; Donahue, 1997 and Qian and Weingast, 1997). 3. The chapter focuses on different government tasks. We will not discuss the economic analysis of the formal structure of political. See, for example, Baldwin et al. (2000, 2001) and Leech (2002) for the latter. 4. Baldwin and Venables (1995) review the empirical evidence on the benefits from regional integration agreements. 5. With heterogeneity, efficiency requires differentiation of benefits and costs between citizens or regions and this is often difficult to achieve in practice. 6. For private goods the consumed or used amount of the good by any economic agent is not available to other agents. 7. For many public goods the minimum geographic area of benefits and costs is smaller than a nation or state. In these cases, lower levels of jurisdictions should provide the goods.
9780230_547018_12_cha10.indd 362
3/19/2009 8:22:30 PM
Rethinking Subsidiarity in the EU
363
8. A recent OECD study (1998) discusses tax competition and ways for managing its possible harms. 9. State and local governments in the United States frequently borrow for longlasting capital projects but operate in private credit markets in the funding of this debt (McKinnon, 1997). 10. Another important EU activity nowadays is the European Central Bank, which is responsible for the EU activities in the sphere of monetary policy. We will not consider the ECB activities. 11. The Social Cohesion chapter and attached social, structural and regional funds are included here. 12. It should be noted though that (2) includes some regulatory issues, whose conclusion is not obvious. 13. See Alesina, Angeloni and Schuknecht (2002) for further discussion using earlier data. 14. However, a few cultural activities, such as preservation of key monuments of human cultural heritage can be seen as even global public goods. 15. See, for example, Shah (1994) for a discussion of the principles for the design of such grants. 16. See for example Dasgupta (1982), chapter 2, for a good introductory discussion of the “problem of the commons”. 17. These considerations imply that the EU needs to be concerned with agricultural production, but the EU agricultural responsibilities would then be focused on food safety (see the 2002 EEAG Report for further discussion). 18. As noted above, there is empirical evidence about tax competition even if the evidence is limited and may not support extreme “race to the bottom” arguments. 19. According to an extensive empirical study of the Ifo Institute, 4 percent of the population of the accession countries will migrate to Germany over a period of 15 years if wages converge at an annual rate of 2 percent (Barro-Sala-i-Martin rule). This will imply an emigration of 6 percent to the total of the old EU countries if the current proportions of migration from the eastern countries to Germany and the rest of the EU (two-thirds Germany, one-third rest) remain stable. See Sinn et al. (2001). 20. In practice also further re-distributive goals as a result of particular ethical viewpoints are stressed in political debates. 21. Sometimes the criterion for inclusion is actual work rather than intention to work. There may also be special provisions for refugees. 22. Such a system has recently been recommended by the Scientific Council of the German Ministry of Finance (2001). 23. See chapter 4 of Sinn (2003) for a detailed analysis of this process. 24. See, for example, Musgrave (1959, 1997) for the conventional view.
References Alesina, A., I. Angeloni and L. Schuknecht (2002) What Does the European Union Do? mimeo, ECB. Baldwin, R. E. and A. J. Venables (1995) “Regional Economic Integration,” in G. Grossman and K. Rogoff, eds, Handbook of International Economics, Volume III, Elsevier Science, Amsterdam, 1598–644. Baldwin, R. E., E. Berglöf, F. Giavazzi and M. Widgren (2000) “EU Reforms for Tomorrow’s Europe,” CEPR Discussion Paper no. 2623.
9780230_547018_12_cha10.indd 363
3/19/2009 8:22:30 PM
364 Designing the European Model Baldwin, R. E., E. Berglöf, F. Giavazzi and M. Widgren (2001) Nice Try: Should the Treaty of Nice Be Ratified, Monitoring European Integration 10, London, CEPR. Begg, D., J. Crémer, J.-P. Danthine, J. Edwards, V. Grilli, D. J. Neven, P. Seabright, H.-W. Sinn, A. J. Venables and C. Wyplosz (1993) Making Sense of Subsidiarity: How Much Centralization for Europe? Monitoring European Integration 4, London, CEPR. Bermann, G. (1994) “Taking Subsidiarity Seriously: Federalism in the European Community and the United States,” Columbia Law Review 94, 331–456. Bhagwati, Y. (2002) Taking Free Trade Seriously, Princeton University Press, Princeton NJ. Boldrin, M. and F. Canova (2001) “Inequality and Convergence in Europe’s Regions: Reconsidering European Regional Policies,” Economic Policy 32, 207–53. Corsetti, G., J. Flemming, S. Honkapohja, W. Leibfritz, G. Saint-Paul, H.-W. Sinn and X. Vives (2002) Report on the European Economy 2002, European Economic Advisory Group at CESifo, Munich. Dasgupta, P. (1982) The Control of Resources, Basil Blackwell, Oxford. De la Fuente, A. and X. Vives (1995) “Infrastructure and Education as Instruments of Regional Policy: Evidence from Spain,” Economic Policy 20, 13–51. Devereux, M., B. Lockwood and M. Redoano (2002) “Do Countries Compete Over Corporate Taxes?” CEPR Working Paper no. 3400. Donahue, J. D. (1997) “Tiebout? Or Not Tiebout? The Market Metaphor and America’s Devolution Debate,” Journal of Economic Perspectives 11, 73–82. Edwards, S. and M. Keen (1996) “Tax Competition and Leviathan,” European Economic Review 40, 113–34. Eurobarometer, http://www.gesis.org/en/data_service/eurobarometer/index.htm(04.12.02). European Commission, http://european-convention.eu.int (04.12.02). European Economic Advisory Group at CESifo (2002) Report on the European Economy 2002, CESifo, Munich. IMF (2002) Government Finance Statistics Yearbook 2002. Inman, R. P. and D. L. Rubinfeld (1997) “Rethinking Federalism,” Journal of Economic Perspectives 11, 43–64. Jackman, R. (2001) “The Pros and Cons of a Common European Employment Policy,” Swedish Economic Policy Review 8, 203–34. Leech, D. (2002) “Designing the Voting System for the Council of the European Union,” Public Choice 113(3–4), December, 437–64(28). Long, N. V. and H. Siebert (1991) “Institutional Competition Versus ex-ante Harmonisation: The Case of Environmental Policy,” Journal of Institutional and Theoretical Economics 147, 296–311. McKinnon, R. (1997) “Market-Preserving Fiscal Federalism in the Am erican Monetary Union,” in M. Blejer and T. Ter-Minassian, eds, Macroeconomic Dimensions of Public Finance: Essays in Honour of Vito Tanzi, Routledge, London, 73–93. Musgrave, R. A. (1959) The Theory of Public Finance, McGraw-Hill, New York. Musgrave, R. A. (1997) “Devolution, Grants and Fiscal Competition,” Journal of Economic Perspectives 11, 65–72. Oates, W. E. (1972) Fiscal Federalism, Harcourt Brace & Jovanovich, New York. Oates, W. E. (1999) “An Essay on Fiscal Federalism,” Journal of Economic Literature 37, 1120–49. Oates, W. E. and R. M. Schwab (1988) “Economic Competition among Jurisdictions: Efficiency Enhancing or Distortion Inducing?” Journal of Public Economics 35, 333–54. OECD (1998) Harmful Tax Competition, An Emerging Global Issue, Paris, OECD.
9780230_547018_12_cha10.indd 364
3/19/2009 8:22:30 PM
Rethinking Subsidiarity in the EU
365
Persson, T., G. Roland and G. Tabellini (1998) “The Theory of Fiscal Federalism: What Does it Mean for Europe?” in H. Siebert, ed., Quo Vadis Europe? Tübingen, J.C.B. Mohr. Pollack, M. A. (2000) “The End of Creeping Competence? EU Policy-Making Since Maastricht,” Journal of Common Market Studies 38, 519–38. Qian, Y. and B. R. Weingast (1997) “Federalism as a Commitment to Preserving Market Incentives,” Journal of Economic Perspectives 11, 83–92. Scientific Council of the German Ministry of Finance (2001) Freizügigkeit und Soziale Sicherung in Europa, Wissenschaftlicher Beirat beim BMFi, Berlin. Shah, A. (1994) “The Reform of Inter Governmental Fiscal Relations in Developing and Emerging Market Economies,” World Bank Policy and Research Series no. 23. Sinn, H.-W. (2003) The New Systems Competition, Basil Blackwell, Oxford. Sinn, H.-W., G. Flaig, M. Werding, S. Munz, N. Düll and H. Hofmann (2001) EU-Erweiterung und Arbeitskräftemigration, Ifo Institute, Munich. Sorensen, P. (2000) “The Case for International Coordination Reconsidered,” Economic Policy 31, 429–72. Tabellini, G. (2002) “Principles of Policy Making in the European Union: An Economic Perspective,” CESifo Forum 3(2), 16–22. Weingast, B. R. (1995) “The Economic Role of Political Institutions: Market-Preserving Federalism and Economic Development,” Journal of Law, Economics and Organization 11, 1–31. Wildasin, D. (2000) “Factor Mobility and Fiscal Policy in the EU: Policy Issues and Analytical Approaches,” Economic Policy 31, 339–78.
9780230_547018_12_cha10.indd 365
3/19/2009 8:22:30 PM
11 Financial Architecture
11.1 Introduction There has been much progress in the integration of European financial markets since the lifting of capital controls during the 1980s, the First and Second Banking Directives (1977 and 1989), the Single Market Programme and monetary union. However, the European banking market, especially in its retail sector, continues to be segmented, and many obstacles remain on the way. For example, most of the mergers and acquisitions (M&A) activity is domestic. Furthermore, the emergence of pan-European stock exchanges has been delayed for a variety of reasons, but especially because of regulatory barriers. Indeed, there seems to be growing disenchantment because of the resiliency of the obstacles to the integration of financial markets. Arguably, regulatory fragmentation constitutes the major obstacle to the integration of financial and banking markets in Europe. Questions have also been raised about the adequacy of present arrangements to preserve stability in the euro area. According to the “official” view, the present decentralised supervisory arrangements of the banking and financial markets in Europe are adequate because of the existing segmentation of business by country. Despite this, several committees and groups (Brouwer, Lamfalussy, Giovannini) have been set up to study the obstacles to financial market integration and propose solutions to improve the regulatory and supervisory frameworks. By and large, those committees have recommended more co-operation among supervisors (Brouwer in particular) as well as the adoption of a “comitology” procedure (delegation of powers to define rules to a committee) to speed up the implementation of financial legislation (Lamfalussy in particular). Many political economy issues are at the heart of the regulatory fragmentation problem, in particular the tension between economic integration and the lack of willingness to relinquish national political control. But while these political economy issues slow down the pace of regulatory and institutional innovations, there are important sources of systemic risk to which 366
9780230_547018_13_cha11.indd 366
3/19/2009 8:24:23 PM
Financial Architecture
367
the European markets are exposed. Recent events have stressed the threat of terrorist action, and possible financial weakness associated with the current slow-down. Some European banks are heavily exposed to emerging markets and to particular sectors, such as telecoms, which have recently experienced deep crises. The process of consolidation within countries has led to the creation of many “national champions” that may create incentives for national authorities to provide excessive guarantees. At the same time, the expansion of cross-border activities may increase potential spillovers and externalities across countries, while creating incentives for the underprovision of supervision and liquidity support by national authorities. In this chapter we review the financial architecture of the euro area, take stock of some of the proposals for reform, and suggest ways to progress. We argue that there are at least three open problems with the present financial architecture arrangements in the euro area. First, these arrangements may not be adequate for financial stability. For instance, in the event of a crisis, there is no clear chain of command among the institutions potentially involved in the intervention. How would the Eurosystem react to the threat of a major disruption like LTCM? Who in Europe would have the responsibility to organise a rescue like that of the president of the New York Fed in the United States? A response based on improvised co-operation may not be enough and may come too late. Moreover, there could be misaligned incentives for national supervisors dealing with transnational firms, as they do not internalise cross-border spillovers from the crisis of such firms. Conversely, national authorities may have strong incentives to provide excessive help to national champions. Second, to a large extent, the present arrangements hinder European financial market integration. Legislation is slow, rigid, and lags behind market developments. Regulatory fragmentation prevents the emergence of deep, liquid European markets (see, for instance, the failure of iX). Protection of national champions and regulatory barriers prevent the emergence of panEuropean banks. Finally, the present arrangements hinder the competitiveness of EU financial markets and institutions. The present gradualist approach may yield more costs than benefits in the longterm and may end up proving ineffective. It would be better not to wait for a major crisis to strike in order to put the house in order. To have a passive policy regarding the lack of financial integration of certain segments of the financial market may backfire, as it fails to remove obstacles to integration and endangers stability. While endorsing in general the well-intentioned recommendations of the committees and groups seeking to remove the obstacles to European financial integration, we feel that a more ambitious approach is needed. This is so because alternative models for reforming the financial architecture in Europe will have profound implications for the degree of financial market
9780230_547018_13_cha11.indd 367
3/19/2009 8:24:23 PM
368 Designing the European Model
integration, competitiveness in the financial industry, and financial and monetary stability. Reform proposals should be assessed in terms of their contributions to the welfare of European citizens, including the price they will pay for financial and payment services, the range of opportunity for insurance and portfolio diversification, the reliability and trustworthiness of the financial institutions in the area. Those criteria lead us to propose some reforms in the European financial architecture, distinguishing short-run measures and calling for a debate on the basic framework with a long-run view. In the short run, clear procedures should be established for crisis lending and crisis management with the European Central Bank at the centre. The crisis framework should be put in place now and the fiscal issue of how to provide help to a transnational institution confronted. A debate should be opened with a view towards evaluating the benefits of more centralised supervisory arrangements in banking, insurance and securities in the medium and long run. In particular, in addition to the current decentralised regulatory competition frame, two basic long run models should be discussed: ●
●
In the first model, the ECB, in the context of the ESCB, would be given a larger role in the supervision of banking, with the contemporaneous creation of separate specialised European-wide supervisors in securities and insurance. In the second model, an integrated supervisor would be constituted, a European Financial Services Authority, and the ECB would have access to supervisory information in order to maintain systemic stability.
In either of the two models supervision need not be completely centralised at the European level. First, because national supervisors will be involved in the day to day supervisory operations. Second, because European level agencies could leave entities trading mostly within one national jurisdiction to be supervised by the appropriate national regulator (under the homecountry principle). An implication of our vision is that the door should be left open in the Convention on the Future of Europe for the necessary institutional changes to implement more centralised regulation, perhaps along the lines of one of the models above. Last but not least, the EU wide competition policy in the banking sector, which limits help to national champions that are “too big to fail,” and removes obstacles to cross-border mergers, should continue. At the same time, domestic competition policy should be reinforced to keep in check local market power in national markets. Reforms of the financial architecture are admittedly complex, as technical aspects are strictly interwoven with legal and institutional aspects.
9780230_547018_13_cha11.indd 368
3/19/2009 8:24:23 PM
Financial Architecture
369
Given the large interests at stake, the process of reform is the target of particularly strong lobbies, both private and public. It would be a great cost for society if the need to reconcile conflicting special interests resulted in lower protection of European citizens against the many risks that an inefficient and vulnerable financial system entails. The plan of the chapter is as follows. Section 2 will look at the state of European financial integration and the impact of the euro while Section 3 will describe present arrangements in terms of financial stability, regulation and supervision, and competition policy. Section 4 will deal with problems of the present regulatory framework and Section 5 will look ahead, rethinking European financial architecture.1
11.2 European financial integration and the impact of the euro The process of European financial integration coincides with the general trend in the financial services industry towards liberalisation, regulatory reform and globalisation (encompassing advances in information technology and communications). These changes have increased competition, as well as the weight of markets in relation to financial intermediaries (“disintermediation”), although banking is not receding in real terms. For the banking sector, the result is a move from the traditional business of taking deposits and granting loans (earning money on the financial margin) to the provision of services to investors and firms (earning money by charging fees and commissions). There is a move from investment in branches to investment in communication networks, information technology and specialised human capital. In general, the transformation of the banking sector (in terms of development of mutual and pension funds, insurance, a corporate debt market and venture capital) is less advanced in Europe than in the United States. The measures undertaken to date to foster the integration of financial markets in Europe (from the Banking Directives and the Single Market Programme to the introduction of the euro) have produced mixed results. The euro-area money market has become substantially integrated, although the degree of integration varies in the different segments. For example, integration is complete for unsecured interbank deposits, as well as for euro-derivatives. Integration is less pronounced for secured money market segments, where liquidity is exchanged for collateral (like commercial paper, CDs, Treasury bills and private repurchase agreements). National disparities in cross-border clearing and settlement are an obstacle to integration. The euro-denominated bond market has become much more homogeneous since the introduction of the euro and has increased in depth and liquidity. Sovereign issuance remains a dominant source of supply and until
9780230_547018_13_cha11.indd 369
3/19/2009 8:24:24 PM
370 Designing the European Model
1998 was associated with the convergence of yields. However, yield convergence has virtually stopped since 1998, as differentials have been reduced to those related to the size of individual issues.2 Smaller member states are not able to provide enough issuance volume in all maturities to reap the full benefits of the unified yield curve. Co-ordinated issuance could alleviate this problem. The euro seems to have stimulated cross-border equity investment and the consolidation of stock exchanges (a successful example is the merger of the Amsterdam, Brussels and Paris exchanges to form Euronext in 2000). Despite the trend towards integration in capital markets, the European banking retail market continues to be segmented, and the degree of crossborder penetration is small.3 In addition to regulatory barriers, existing branch networks and relationships with clients are important obstacles to entry in the retail sector, and there are significant switching costs for customers.4 The lack of integration is most apparent in the markets for consumer credit and mortgages. Regulatory restrictions governing the composition of the portfolio of institutional investors, such as pension funds and insurance companies, are a source of market segmentation in asset management. The lack of integration of the retail market also characterises electronic banking, which remains very limited (with some exceptions in the Nordic countries and the United Kingdom). As the European Commission has stated, crossborder retail fees are high and have maintained a high degree of dispersion in the last decade. In 2001, the European Commission introduced a regulation of cross-border payments in euros because of the little progress observed in reducing price differentials. At the same time, cross-border securities trade is much more expensive than trade within the national boundaries (see Economic Financial and Committee, 2002.) In general, important differences among countries remain in terms of the degree of competition, amount of rationalisation of the banking sector, financial strength of banks, and progress in the transformation towards a services industry.5 Consolidation among banks is taking place in Europe mostly through domestic mergers (see Table 11.1). In contrast to the United States, obstacles to cross-border mergers in Europe consist of restrictions on labour mobility, differences in corporate culture, and political interference (for example, promotion of national champions).6 Nevertheless, international deals predominated among insurance companies in the period 1985–1997. Furthermore, in the same period, mergers and acquisitions (M&A) deals tend to be concentrated among the same type of institutions, rather than being targeted to create conglomerates (Berger et al. 2000). An interesting feature of cross-border banking in Europe is that it often takes the form of subsidiary instead of branch. This is the case, for example, of the crossborder mergers and acquisitions involved in the formation of Nordea AB, ING Group and HypoVereinsbank (see Dermine, 2002) as holding companies with subsidiaries.
9780230_547018_13_cha11.indd 370
3/19/2009 8:24:24 PM
9780230_547018_13_cha11.indd 371
3/19/2009 8:24:24 PM
7,974 378 9
Non-bank financial – non-bank financial 1998 6 7,299 2 1999 11 15,508 4 2000 e 4 5,071 1 7 19 5
3 12 4
12 15 5
Nb.
1,201 21,888 545
897 4,130 653
13,787 7,495 11,654
Valued
Other non-euro country
15 34 10
8 16 13
20 28 8
Nb.
Valued
16,474 37,774 5,534
31,147 25,746 7,052
22,379 58,202 16,182
Total
11.8 15.7 23.3
24.2 20.8 48.1
12.7 15.9 26.7
Nb.
13.8 40.7 18.8
37.9 56.4 39.1
13.0 34.2 62.0
Valued
As a percentageb
Source: BIS (2000), p.134. See also www.bis.org.
Note: a Either acquirer or target company is resident in the euro-area. Only completed or pending deals, announcement date volumes. bOf mergers and acquisitions in all countries. cNb. = Number. dIn millions of US dollars. e1 January to 10 April.
646 800 1,631
1 1 1
147 9,465 0
Valued
Banks-non-banks financial 1998 7 28,604 1999 3 20,816 2000 e 8 4,768
Nb.
1 4 0
7 9 3
Valued
Other euro country
8,445 41,242 4,528
Banks-banks 1998 1999 2000 e
Nb.c
Same country
Table 11.1 Merger and acquisition activity in the euro-area financial industrya
372 Designing the European Model
One issue with domestic mergers is that they tend to increase local concentration, which is what matters for the exercise of monopoly power in retail banking. In 1997, the C5 deposit ratio (the share of deposits of the five largest institutions) had a value which was similar for the EU and the United States (around 12 percent).7 Yet, because of the weight of interstate mergers, the current consolidation process in the United States has not generated a clear trend towards local concentration (Berger et al. 1999). With very limited cross-country mergers, the situation in Europe is more worrisome. 11.2.1 Has the level of risk increased? Consistent with the international evidence (Demirgüç-Kunt and Detragiache 1998), liberalisation in Europe has also been associated with bank failures. Table 11.2 shows the fiscal costs of selected banking crises in Europe, compared with Japan and the United States. The table shows that the costs of European banking crisis have been comparable to the experience of other countries.8 Because of pressure on margins due to disintermediation and the general increase in competition in local markets, European banks have increasingly looked for markets with larger margins abroad. As a result, in the second half of the 1990s, the exposure of European banks to emerging markets was several times larger than that of US banks. If we break down the income of large European banks by geographical origin (including offbalance sheet activities), we see that a substantial part is earned abroad (about 33 percent in 1998, more than half of which is earned outside the EU). The largest Spanish banks, for instance, have very high exposure in Latin America (SCH in Brazil and BBVA in Mexico). By the same token, international interbank claims of EU banks have grown substantially. In 1998 international claims by banks located in the EU on banks located
Table 11.2 Fiscal costs of select banking crisis Period
Spain France Finland Sweden USA Japan
1977–85 1994–95 1991–94 1991–94 1981–91 1992 present
Fiscal cost (% of GDP)
5.6 0.7 11.0 4.0 3.2 20.0
Blanket Extensive liquidity guarantee for support to depositors and financial creditors intermediaries No No Yes Yes No Yes
Yes No Yes No No Yes
Source: The EU Economy 2001 Review (2001) and Honohan and Klingebiel (2001).
9780230_547018_13_cha11.indd 372
3/19/2009 8:24:24 PM
Financial Architecture
373
outside (in) the EU represented 7 percent (12 percent) of the total balance sheet of the EU banking system (see Economic and Financial Committee, 2000.) These data point to high risk-taking, especially by large banks, leading to non-negligible systemic risk. At the same time the wave of domestic consolidation has created banks that are large in relation to some national economies, particularly in small economies like Switzerland and the Netherlands, but also in larger ones like Spain. This means that trouble in some of these “national champions,” with its possible systemic consequences, may come at a high cost. For example, the book value of equity to national GDP ratio (2000) for UBS and Crédit Suisse in Switzerland is 12.4 percent and 10.5 percent, respectively; for ING Group in the Netherlands it is 6.6 percent and for SCH in Spain it is 4.3 percent. In other words, while financial market integration provides opportunities for better diversification (for example with cross-border M&As),9 it also provides incentives for higher risk-taking, increasing the level of systemic risk and vulnerability to contagion.
11.3
The present arrangements
In addition to the development of national legislation, financial regulatory institutions in the euro area derive from the Treaty of the European Union and European Commission Directives. Competition policy also goes back to the Treaty of Rome. This section will illustrate the present regulatory situation in terms of crisis lending and management, regulation and supervision, and competition policy. We leave to Boxes 11.1 to 11.3 a brief discussion of the theoretical arguments for financial regulation, the regulatory facilities in place to guarantee financial stability, and the moral hazard problem motivating the need for prudential supervision. 11.3.1
Crisis lending and management
The monetary authority under the Treaty of the EU is the European System of Central Banks (ESCB or Eurosystem), made up of the European Central Bank (ECB) and the national central banks (NCBs). The ESCB has the narrow mandate to maintain price stability, and without prejudice to this objective, it should support the general economic policies of the EU (Article 105(1) of the Treaty). The ESCB is subordinate to the national governments and to other European institutions in the area of financial supervision and the stability of the European financial and banking systems: “The ESCB shall contribute to the smooth conduct of policies pursued by the competent authorities relating to the prudential supervision of credit institutions and the stability of the financial system” (Article 105(5)). In summary, the Treaty does not put the ESCB explicitly in charge of the stability of the financial system, although there is recognition of the
9780230_547018_13_cha11.indd 373
3/19/2009 8:24:24 PM
374 Designing the European Model
Box 11.1
The rationale for financial regulation
• Fragility and its consequences. Because of currency and maturity mismatch between assets and liabilities, the banking and financial system is vulnerable to sudden losses of funds resulting in the failure of fundamentally solvent intermediaries. Experience shows that panics and systemic crises compromising the banking and financial system may have a major impact on the real sector of the economy (as suggested by the examples of the Great Depression of the 1930s, the 1998 international financial crisis, or the on-going crisis in Japan). • Co-ordination failure of investors and runs. In the case of a purely speculative panic, depositors withdraw their funds and force the bank to early and costly liquidation of assets. A panic can be generated by news regarding bank solvency problems. In this case, the possibility of depositors’ runs may have a disciplinary effect on risk taking by financial intermediaries (see Diamond and Dybvig, 1983; Jacklin and Bhattacharya, 1988; and Postlewaite and Vives, 1987.) • Contagion and systemic risk. The bankruptcy of one financial intermediary can have systemic consequences, owing to contagion effects which may give rise to strong negative externalities for both the financial sector and the real sector of the economy. The failure of one institution may jeopardise the solvency of other institutions via default on commitments assumed in the interbank market. Large variations in the price of assets such as an abrupt fall of stock prices or the failure of a major intermediary, may generate a domino effect and systemic crisis affecting the payment system. • Why regulation? Regulation aims at providing the banking and financial systems with stability to elude the negative effects associated with failing institutions and systemic crises. A second aim is to protect small investors and customers of firms providing financial services.
ESCB’s task of promoting the harmonised operation of the payments system (Article 105(2)). However, the ECB has a consulting role in legislation regarding financial institutions in so far as they may affect stability (Article 105(4) and EU Council Decision 98/415/EC), and its role with respect to questions of supervision can be larger: “The Council may, acting unanimously on a proposal from the Commission and after consulting the ECB and after receiving the assent of the European Parliament, confer upon the ECB specific tasks concerning policies relating to the prudential supervision of credit institutions and other financial institutions with the exception of insurance undertakings” (Article 105(6)). This means that the ECB could be assigned supervisory powers, with the exception of insurance, without the need to reform the Treaty of the EU. Only relatively recently has the ECB raised its profile in crisis management. The first official statement we are aware of is Duisenberg’s October 1999 declaration in the European Parliament:10 The main guiding principle within the Eurosystem with reference to the provision of emergency liquidity to individual financial institutions is
9780230_547018_13_cha11.indd 374
3/19/2009 8:24:24 PM
Financial Architecture
Box 11.2
375
Financial stability facilities
Crisis lending and the central bank: An important discretionary activity of the central bank consists in helping banks experiencing temporary liquidity problems via the discount window or open-market operations. The central bank can create liquidity as needed, and can credibly commit to unlimited lending and fast reaction because of its control of high-powered money. Alternative arrangements to provide liquidity involving private money (lifeboats, liquidity consortia) or funds raised with taxes (via deposit insurance funds, building “war chests”, or direct recourse to the finance ministry) are costly and in general can be at best part of a solution in which the central bank is also involved. Crisis management: A crisis manager helps to solve the co-ordination problem among creditors that a crisis entails. In many instances the lender of last resort (LLR) manages the crisis but does not put up its own funds, which may be private money (as in the rescue of Long-Term Capital Management (LTCM) co-ordinated by the Federal Reserve Bank of New York, see next paragraph), or money from the deposit insurance fund or the taxpayer (Goodhart and Shoenmaker, 1995). Examples of crisis management: The stock-market crisis of 1987 provoked problems in the clearing systems of the derivative markets and was overcome thanks to an injection of liquidity by the Federal Reserve. Financial intermediaries required additional funds to meet the needs of their clients with margin calls. Indeed, intermediaries in the capital and money markets were assisted by bank credit lines in providing liquidity. In the crisis of the hedge fund LTCM, after Russia’s default in 1998, LTCM had to be recapitalised in order to meet the margin requirements in derivatives when the market spreads moved adversely to the position of the fund. The Federal Reserve Bank of New York organised a rescue operation with investment banks that were investors in the fund. According to the Fed, the hasty liquidation of the (large) fund positions could have caused a major disruption in world financial markets. The classic Lender of Last Resort (LLR): The classic prescription for the LLR (associated with Bagehot, see Meltzer, 1986) is that funds should be provided only to solvent banks with liquidity problems. These banks are to be helped with loans at a penalty rate and against good collateral, evaluated in “normal” conditions. The solvency and collateral terms under which help will be given must be clearly stated and the LLR must announce its readiness to lend without limit. Goodfriend and King (1988) have disputed this “banking policy” view arguing that in developed financial systems a solvent bank cannot be illiquid and therefore only open-market operations are needed. Rochet and Vives (2002) provide a modern justification of Bagehot’s view. Deposit insurance: Deposit insurance is a non-discretionary activity by means of which deposits are protected up to predetermined limits. If the limits are not very high, it meets the aim of protecting the small investor. Too-Big-To-Fail (TBTF) policy: Often banks and depositors are protected above the levels required by the deposit insurance scheme. Under the TBTF policy a large insolvent bank will be rescued (and its uninsured depositors will be protected) whenever its failure is likely to affect other banks, via the interbank market, and the real economy.
9780230_547018_13_cha11.indd 375
3/19/2009 8:24:24 PM
376 Designing the European Model
Box 11.3
Regulation, prudential supervision and moral hazard
• Moral hazard. Both the Too-Big-To-Fail policy and the deposit insurance system reduce the incentive of depositors and investors to monitor bank performance. Excessive risk taking may then derive from the bank’s limited liability charter and the non-observability of the risk level in the bank portfolio. Moral hazard also arises because the effort of bank managers in monitoring projects is not observable. • Time-inconsistency. A well-intentioned LLR will find it optimal ex post to help a bank whenever this salvages the value of projects that the bank is financing. Indeed, ex ante the central bank may want to commit to close the bank if the returns are very low (signalling a solvency problem) while helping the bank if the returns are only moderately low (signalling a liquidity problem). Such a commitment provides incentives for bank managers to monitor the projects they finance. However, ex post, costly liquidation of the projects will not be optimal, so the central bank may hesitate to carry out its threat. The commitment problem is compounded by the interest of the bank management in the continuation of the bank. Bankers, anticipating the help, will tend to exert suboptimal effort in monitoring projects and take excessive risk. • Alleviating the excessive bailout problem. • A central bank with a “tough” reputation can alleviate the time-inconsistency problem. Credible central banks typically adopt a “constructive ambiguity” policy, not making explicit the criteria under which entities with problems will have access to help. • Alternatively, external discipline can be imported into a small open economy by adopting another (stable) currency (“dollarising”), entering into a credible monetary union (like EMU), and/or acquiring foreign short-term debt. In all those instances a firm commitment is acquired (with dollarisation because recourse to the LLR is drastically limited, with a monetary union because of the credibility of the central bank, and with foreign short-term debt because it cannot be inflated away, see Gale and Vives (2002) and Vives (2002)). • Another way to import discipline for countries which have difficulties building a reputation for the central bank, is by transferring political sovereignty, forming a monetary union and establishing an independent and credible central bank. For this to succeed, some of the participants’ central banks must have already established a credible reputation. • Prudential supervision. The general trend in prudential supervision is to check risk-taking with capital requirements and appropriate supervisory controls. Both risk-based deposit insurance and disclosure improvements have been proposed to limit risk-taking behaviour. However, while it is feasible to introduce disclosure requirements of the banks’ market positions, it is more difficult to assess the risk level of the illiquid loan portfolio of a bank. (See Matutes and Vives, 2000, and Cordella and Yeyati, 2002.) Furthermore, more disclosure may in fact induce information-based runs of depositors generating instability.
9780230_547018_13_cha11.indd 376
3/19/2009 8:24:24 PM
Financial Architecture
377
that the competent national central bank would be responsible for providing such assistance to those institutions operating within its jurisdiction. The ECB does, however, have to be informed of this in a timely manner. In addition, in operations of relevance to the single monetary policy, the decision-making bodies of the Euro-system will be involved in assessing the compatibility of the envisaged operations with the pursuit of monetary stability. In the case of a general liquidity crisis resulting from a gridlock in the payment system, for instance, the direct involvement of the Eurosystem could be expected. The central bank is the natural candidate for the lender of last resort function (LLR) in a financial system (see Box 11.2). The Federal Reserve and the Bank of England are explicitly in charge of the stability of the financial system (but the Bundesbank was not). For example, the Federal Reserve Board (FRB) determines the policy regarding supervision and last-resort lending on the part of the banks of the Federal Reserve System. The FRB determines the conditions under which discount-window loans will be granted by the Federal Reserve banks and, in practice, the FRB is consulted regarding any major loan. Most likely, the reason behind the lack of formal responsibility of the ECB on stability matters is that there is no central European fiscal authority. Typically, a central bank turns to the finance ministry or specialised agencies, like a deposit insurance fund, when an assisted bank turns out to be insolvent. 11.3.2 Regulations and Supervision The home country control principle and regulatory competition The Single Market in financial services builds on the single banking licence, together with the principles of home country control and of mutual recognition (Second Banking Directive, effective since 1993).11 If a financial institution is authorised to operate in one European country, it may offer or establish financial services anywhere else. That is, the financial institution can branch from one member country to any other member country.12 The Second Banking Directive establishes the control of the home country (that is the member state in which the financial institution has been authorised) for the prudential supervision of solvency and of major risks, and a minimum harmonisation between countries in other areas, such as minimum capital requirements, concentration of risks, and protection of investors.13 The Directive regarding deposit insurance proposes a minimum coverage (up to €20,000), which tends to reflect an interest more in protecting the small investor than in protecting the stability of the banking system. Deposit insurance is organised according to the home country principle: a bank granted a licence in an EU country is insured by the deposit insurance system of the home country when it operates in another EU country. However, a foreign branch may join a more favourable host country scheme.
9780230_547018_13_cha11.indd 377
3/19/2009 8:24:24 PM
378 Designing the European Model
The principles of home country control and mutual recognition lay out a regulatory competition framework. This framework may encourage information production and limit the potential opportunism of the national regulators. Country discretion ranges from legal differences in financial contracts, the organisation and conduct of banking supervision, the structure of deposit insurance schemes, and the institutions and procedures to restructure banks. For example, the administration of deposit insurance may be in the hands of either the government or the banking sector, or both. In general, deposit insurance premia are a flat percentage of deposits but some consideration to risk is given in Italy, Portugal and Sweden. Funding is secured in some countries with ex ante contributions and in some others with ex post levies. Diversity of regulatory institutions A recent development is the establishment of universal regulators for banking, insurance and financial markets. This is the approach taken in the United Kingdom (in 1997), the Scandinavian countries (Norway in 1986, Denmark in 1998 and Sweden in 1992), and Japan. Let us describe the UK approach. The Bank of England Act (1997) sets up the Financial Services Authority (FSA) that integrates responsibility for the supervision of markets (securities), financial intermediaries and insurance. The FSA undertakes the authorisation and prudential supervision of all financial entities, the supervision of financial markets, regulatory policy, and the response to problems in institutions and markets that do not enter into conflict with the competence of the Bank of England on the stability of the financial system and systemic risk. The Bank of England and the FSA must work jointly, but each institution has a leadership role in its field of responsibility. The Bank of England, the FSA and the UK Treasury have signed a Memorandum of Understanding (MOU) that delineates their respective responsibilities. In particular, when dealing with an emergency situation: “The Bank and the FSA would need to work very closely and they would immediately inform the Treasury, in order to give the Chancellor of the Exchequer the option of refusing support action” (MOU, par. 13).14 It is specified also that the Bank and the FSA must share information and work jointly to avoid duplication in the gathering of information. The Bank of England has free and open access to supervisory records (MOU, par. 21). In the EU, there are six countries in which the central bank is the main supervisory authority: Greece, Ireland, Italy, the Netherlands, Portugal and Spain. These are the only countries in the EU-15 that maintain different supervisors for banking, insurance and securities markets. Germany15 (2002), Austria (2002), Denmark, Sweden and the United Kingdom have embraced the FSA model. Belgium, Finland and Luxembourg have an integrated banking and securities supervisor. However, in Belgium, as well as Ireland, the central bank will take over and integrate financial supervision.
9780230_547018_13_cha11.indd 378
3/19/2009 8:24:24 PM
Financial Architecture
379
France is the only country with a specialised banking supervisor, who shares responsibility with the central bank (see a recent paper by Lannoo (2002) for further details). Overall, many central banks have moved away from banking supervision. However, as shown in Lannoo (2002), disparity still exists, and there are other contending models. For example, Australia has three supervisors with horizontally assigned tasks: systemic stability for the central bank, prudential supervision for a specialised agency, and conduct-of-business rules (disclosure, level playing field, transparency, market integrity) for another agency. In any case, it is to be noted that bank supervisors will focus on prudential supervision (control of credit and market risk), securities supervisors on investor protection and market integrity, while insurance supervisors will worry about the long-term sustainability of the insurers (and hence monitoring asset-liability management). Box 11.4 provides an analysis of the arguments in favour or against separation of regulatory institutions. The European regulatory and supervisory maze Supervision remains decentralised at the national level and national supervisors operate mostly within borders. The main institutional channel of the ECB for obtaining information regarding the banking and financial system is the Banking Supervision Committee of the ECB. The BSC also serves as an advisory body to the ECB when the latter forms opinions on EU and national legislation. It is in the BSC, where the national supervisors of EU countries (the central banks and other agencies) are represented, that the supervision of euro (and EU) countries must be co-ordinated via the exchange of information and co-operation of supervisors. It is worth pointing out that EU directives do not impose information sharing obligations on national supervisors in times of crisis. However, the so-called BCCI Directive of the EU has removed obstacles to the exchange of confidential information from supervisors to central banks. Moreover, there is bilateral co-operation between supervisors, who negotiate information exchange and supervisory procedures about cross-border activities in a Memorandum of Understanding (MOU). A more informal (and lower level) multilateral arrangement is the Groupe de Contact, a group of EU banking supervisors from the EEA (European Economic Area), which deals with individual bank problems. Parent to the Groupe de Contact is the EU Banking Advisory Committee (BAC) that has mainly a legislative role in advising the European Commission. Parallel groups in insurance are the Conference of Insurance Supervisors and the Insurance Committee. In February 1999, a multilateral MOU among European security supervisors representing members of FESCO (Forum of European Securities Commissions) was signed. The Lamfalussy Committee of Wise Men (see Section 4.3) proposed the creation of the Committee of European Securities Regulators (CESR) to replace FESCO and to strengthen
9780230_547018_13_cha11.indd 379
3/19/2009 8:24:24 PM
380
Designing the European Model
Box 11.4
Optimal regulatory design
The case for central bank supervision Arguments for the central bank to have supervisory capacity. A central bank is best placed: • to distinguish between problems of liquidity and of solvency in order to minimise the losses associated with loans granted and making possible a role as crisis manager; • to determine the best kind of intervention (open-market or discount operations); • to profit from economies of scope in the acquisition of information between the function of providing liquidity and that of supervising (for example, the first of these functions requires a detailed familiarity with the banks’ liquidity requirements); • to exploit synergies between the conduct of monetary policy and information collected with supervisory purposes. Indeed, banking supervisory information (early warning of problems with non-performing loans or changes in the lending pattern of banks) may improve the accuracy of macroeconomic forecasts. Arguments against the central bank having supervisory capacity: • The combination of control of monetary policy and the role of LLR at the central bank raises an inflationary concern. However, a central bank committed to price stability will sterilise the injections of liquidity necessary for the stability of the system in the event of crisis (as the Federal Reserve did in 1987) so that there is no undesired increase in the money supply. In practice matters may not be so simple and intervention as LLR may give rise to confusion in the expectations of the private sector regarding the central bank’s monetary policy stance. • There may be a conflict of interest between the reputation of the central bank as guarantor of currency and financial stability. For example, concern for the reputation of the central bank as supervisor may encourage an excessive use of the LLR facility so that bank crises will not put its supervisory capacity in question. Underlying the conflict-of-interest concern there are incentive problems among regulators related to their career concerns, accountability and monitoring of their multiple tasks, allocation of control, incentives to produce information and potential capture (see Vives, 2000). • Some preliminary evidence indicates that central bank involvement in supervision may increase inflation (see Bini Smaghi, 2000, and Di Noia and Di Giorgio, 1999). The case for an independent FSA Arguments for the separation of supervision from the central bank: • Separation facilitates the optimal provision of incentives to self-interested bureaucrats so as to minimise conflicts of interest. • The convergence between the activities of financial institutions and markets points to the need for the combined regulation of banking, insurance and securities. It is becoming increasingly difficult to separate market-derived risk
9780230_547018_13_cha11.indd 380
3/19/2009 8:24:24 PM
Financial Architecture
381
from credit risk. Banking crises that involve operations with financial derivatives (such as Barings or LTCM) seem to require specialised knowledge of market regulators. At the same time banking and insurance tend to converge. • There are also EU-related political economy considerations. In a system in which the ECB is perceived as having already too much power and faces accountability questions, the creation of an independent regulatory agency may help lessen both concerns. It is easier to hold accountable an agency with a well-defined mission.
co-operation among national regulators. The CESR was established in 2001. The Lamfalussy Committee also recommended the establishment of an EU Securities Committee with implementing powers to interpret and adapt legislation. Additionally there are more committees in the EU. There is the Financial Services Policy Group (FSPG) to set strategic lines for financial services regulation, and the Economic and Financial Committee (EFC), which discusses financial stability and other issues in ad-hoc committees. There are also some cross-sectoral committees: a Mixed Technical Group of Financial Conglomerates and a Cross-Sectoral Round Table of Regulators. The latter was set up to foster information exchange among supervisors following the recommendation of the Brouwer Report on Financial Stability by the Economic and Financial Committee. The maze of committees is summarised in the paper of Lannoo (2002). 11.3.3 Competition policy European competition policy also plays an important role in shaping the European financial architecture. Two important instances are bank rescues and state aids, and cross-border mergers. First, the European competition policy Commissioner can intervene to examine whether a bank rescue with public money is compatible with competition policy towards state aids. Assistance to the French national champion Crédit Lyonnais was challenged exactly on this basis. Public rescue of Banesto (Spain) and Crédit Lyonnais provide additional examples of the Too-Big-To-Fail policy in Europe. European competition policy over state aids (complementing the EU Directive on reorganisation and winding-up of credit institutions 2001) allows prompt corrective actions. The intervention of the European competition policy authority may be desirable even if there are no negative cross-border externalities from the state aid. The reason is that the European competition policy authority may strengthen domestic policy makers’ commitment to screen state aids according to market failure principles, away from local lobbying pressures. Secondly, the European competition policy authority can play an important role in facilitating cross-border mergers and acquisitions by removing obstacles established by national authorities. Indeed, political obstacles to cross-border mergers have been pervasive – as suggested by the BBVA’s failed
9780230_547018_13_cha11.indd 381
3/19/2009 8:24:24 PM
382
Designing the European Model
attempt to take over Unicredito in Italy, or the problems of former BSCH (now SCH) in Portugal while attempting, and finally succeeding, in acquiring the Champalimaud group. In the latter case, the European Commission challenged the Portuguese regulator, who stated its opposition to the takeover because of “stability concerns.” Another example is provided by the attitude of the French authorities, looking for a “French” solution in the triangular battle of BNP-SG-Paribas that ended with the merger of BNP and Paribas. Responsibility for the control of domestic mergers, which are so far predominant in Europe, varies from country to country. In many countries, responsibility lies with the competition authority, sometimes shared with the regulator (United Kingdom, Switzerland, Scandinavia, France, Greece), but in practice the central bank/supervisor carries a lot of weight. In Italy, the central bank approves bank mergers and the competition authority has only a consulting role. European practice contrasts with that in the United States, where banking mergers must receive approval of the regulator (be it the Federal Reserve, the FDIC or the OCC) but the Department of Justice (DOJ) can (and does) challenge mergers that threaten to reduce competition substantially. Typically, the DOJ uses more stringent criteria.
11.4.1
Plans, reports and problems
Several reports and studies on financial market integration in Europe, by the European Commission or by committees and groups specifically formed to address this issue, have pointed at several pending problems and have produced recommendations. In this section, we take stock of these concerns and add a few more. 11.4.1
Integration of financial markets: regulatory barriers
As we have seen in Section 2, the integration of financial retail markets is far from complete. In addition to natural barriers (like language, culture, information), there are regulatory barriers. An important one is that the legislation on consumer protection is in the hands of the host country. Financial entities still have to design 15 different products for 15 different markets (member states). This extends to e-banking. While the e-commerce Directive calls for the supply of services based on the rules of origin, in the draft of the Directive on distance selling of financial services things are much more complicated. Differential tax treatments are another obstacle to integration (as regards, for example, pension funds and life insurance). Regulatory barriers are still in place as pan-European institutions are confronted with multiple rules and reporting requirements. For example, a typical large financial institution has to report to more than 20 supervisors in the EU (out of the 39 existing). To this we should add the political obstacles to cross-border mergers.
9780230_547018_13_cha11.indd 382
3/19/2009 8:24:25 PM
Financial Architecture
383
In 1998, the European Council adopted the Financial Services Action Plan for 1999 to 2005, comprising 41 separate measures (EU Directives and Commission Communications) with the aim of completing the legislative framework for market integration in financial services. Three main objectives are ● ● ●
a single EU wholesale market, open and secure retail markets, state-of-the-art prudential rules and supervision.
There has been progress in the implementation of the 41 measures, but not without important setbacks. Examples of setbacks are the failure of the Takeover Bids Directive, the standstill on pension funds, and tensions between the European Commission, the Council and the Parliament in implementing the recommendations of the Lamfalussy Report (further discussed below in Section 4.3). 11.4.2 Crisis management and cross-border risk: What framework? Under present arrangements, it is up to national central banks (NCB) to undertake the LLR function and provide emergency liquidity assistance to financial institutions. They are responsible for decision-making in crisis situations, and they have to bear the eventual cost of the intervention. So, if a bank develops solvency problems and ends up being rescued, the cost is paid either by the national deposit insurance fund or the national budget, or both. The responsibility for intervening falls on the “host” country central bank when a crisis hits a subsidiary and will be likely to be shared between home and host country central banks when it affects a foreign branch. If liquidity assistance has monetary consequences for monetary policy, then the ECB and the Eurosystem will be involved. Clearly, the involvement of the Eurosystem is to be expected in the presence of a general liquidity crisis, such as a gridlock of the payment system. This policy is consistent with the principle of home country control for supervision and deposit insurance. In response to criticisms that the present arrangements were not adequate to guarantee stability in the euro area, the Economic and Financial Committee of the EU was asked in 1999 to check “whether the existing regulatory and supervisory structures in the EU can safeguard financial stability.” An ad-hoc working group chaired by Henk Brouwer was formed. In its Report on Financial Stability (Economic and Financial Committee 2000) this group concludes that the existing institutional arrangements provide a coherent and flexible basis for safeguarding financial stability in Europe, and make some recommendations to enhance their smooth functioning. A second report of the EFC (Economic and Financial Committee 2001) assesses whether the current arrangements for crisis management are appropriate, and whether any progress has been made on the recommendations
9780230_547018_13_cha11.indd 383
3/19/2009 8:24:25 PM
384
Designing the European Model
of the first report. The report concludes: “Substantial progress is being made by the various supervisory committees and the national authorities in the EU in implementing the recommendations of the first report on financial stability.” The main recommendations of these reports are to enhance co-operation among different authorities (supervisors, central banks, and finance ministries), and to foster convergence of supervisory practices. Supposedly, these recommendations have been advanced with the help of a plethora of committees (see Section 3). To deal with major financial institutions (including conglomerates) domiciled in the EU, it was recommended to reach an agreement on one co-ordinating supervisor with well-defined responsibilities. Accordingly, the draft directive on financial conglomerates (April 2001) prescribes the mandatory appointment of one (or more) supervisory coordinator(s) of qualifying conglomerates as well as his (their) tasks.16 The key question is whether the existing co-operative framework of crisis management is up to the task of preserving financial stability in the euro area. To a large extent, the reason why present arrangements were considered adequate is that financial consolidation has so far taken place mostly at the domestic level. However, there are a number of open questions and issues regarding intervention policies in the event of the failure of some large financial entity, possibly causing systemic problems with cross-border spillovers in the euro area. A number of these questions are listed below: ●
●
The chain of command and crisis procedures. Who is in charge of monitoring cross-border crises? Who will take the lead in a crisis with systemic consequences? What are the crisis management procedures at the ESCB? We have seen above that supervision is in the hands of NCBs. The response to a crisis is one of “improvised co-operation” anchored at the BSC of the ECB, where information should be exchanged and decisions taken. In response to concerns about the effectiveness of this arrangement, the BSC has developed a set of prudential indicators trying to capture aggregate risk exposures of EU banks. The goal of these indicators is that – should problems arise in, say, a large group, potentially leading to contagion effects in many EU countries – these problems are reported to the authorities in all the countries concerned (Economic and Financial Committee 2001, p. 7). Similarly, the Brouwer Report II has requested national supervisors to add crisis management procedures to their bilateral MOU (as well as to remove the remaining legal obstacles to information sharing among supervisors). Conflict between national supervisors in a transnational crisis. Central banks and/or national supervisors may pay too little attention to problems of foreign clients of domestic banks, or to systemic international effects of a domestic crisis. For instance, in principle they could focus exclusively on the consequences of financial failure for the national economy, ignoring
9780230_547018_13_cha11.indd 384
3/19/2009 8:24:25 PM
Financial Architecture
●
●
385
spillovers to other countries.17 In addition, there is the possibility that the failure of a foreign bank will have systemic consequences in the host country.18 Those conflicts of interest will likely impair information exchange among national supervisors.19 Excessive help at the country level and insufficient help at the European level? National regulators may be pressured to help ailing national champions, while they may be less willing to help transnational banks (intervening too little and perhaps also too late). The issue is who will internalise the cost of failure of a pan-European bank given that any single country would not reap the full benefits of a bailout. Under the present rules it is not clear who would pay for a failed insolvent transnational institution that has gone bankrupt after being helped; that is, how would the losses be eventually shared among the fiscal authorities? Excessive help to national banks can be partially controlled by European competition policy (state aids). A low level of help to transnational banks has two sides: on the positive side it helps to keep moral hazard problems in check (see Box 11.3); on the negative side it may dangerously increase systemic risk. 20 EU versus euro area. Some thought should be given to the co-ordination issues between the euro area and the whole EU. This will be particularly important after enlargement.
11.4.3 Integration of securities markets: Regulatory fragmentation Despite many obstacles, some segments of the securities markets are integrating quickly in Europe. The formation of Euronext (the joint venture of the Paris Bourse and the Amsterdam and Brussels stock exchanges) is a positive example. The failure of the merger of the London Stock Exchange and the Deutsche Börse into iX (international exchanges) is a negative one. National regulation and lack of harmonisation in settlement systems, disclosure and supervision are obstacles to the integration of stock markets. The iX project illustrates the complexity of the arrangements devised to circumvent regulatory hurdles in different countries. According to the original plan, secondary-market trading on the pan-European blue-chip market was to be regulated by the UK FSA, while trading on the pan-European highgrowth market was to be regulated by the German federal equities regulator (at the time BAWE). Furthermore, existing companies could keep their home-country listing, while newly admitted companies would have been listed through the competent authority of their choosing. How to achieve settlement consolidation and a central counterparty system was, however, left vague in the merger proposal. As mentioned before, the Lamfalussy Committee of Wise Men was set up to diagnose the regulatory mechanisms in the EU, and to propose measures to speed up the development of European financial markets. The report (February 2001) identifies a set of obstacles to integration of securities
9780230_547018_13_cha11.indd 385
3/19/2009 8:24:25 PM
386
Designing the European Model
markets in Europe related to inadequate regulation (in addition to differences in legal systems, taxation, as well as political, trade, and cultural barriers). It also lists the most important gaps in European regulation and priorities for reform (see Box 11.5). The major problem according to the Lamfalussy Report lies in the regulatory process because: ●
● ● ●
it is too slow (the Takeover Directive has been discussed for more than a dozen years and not yet been adopted, Basel I took 4 years, how long will it take for Basel II?); it is too rigid; it is too ambiguous; it fails to distinguish between core principles and implementation rules.
The Lamfalussy Report proposes a four-level approach to securities market legislation: ●
●
●
●
definition of a framework legislation (for example directives setting the general principles); establishment of an EU Securities Committee with implementing powers to interpret and adapt legislation; creation of the Committee of European Securities Regulators (CESR, done in 2001), replacing FESCO, to strengthen co-operation among national regulators; and adoption of stricter enforcement procedures.
The creation of the EU Securities Committee was contested by the European Parliament (EP) on the matter of the degree of implementation powers of such a committee (“institutional balance” in European parlance). But even independently of political and institutional issues, the approach envisioned by the report is not easy to apply. This is because it is not obvious how principles should be distinguished from implementation rules. For example, the draft prospectus21 and market abuse22 directives have gone into quite a bit of detail. The EP has proposed more than 100 amendments to the market abuse draft proposal (because the EP would not be able to amend the “implementation” decisions of the Securities Committee at a later stage). It is to be expected also that enlargement of the EU will compound this kind of problem. The Lamfalussy Report has contributed to what is called (in European parlance) the “comitology” procedure. According to this procedure, while directives establish general principles, adopting the implementation and adaptation procedures of the general principles is left to a committee with broad interpretative powers (the “comitology powers”) – an example being the proposed EU Securities Committee. This has clearly clashed with the
9780230_547018_13_cha11.indd 386
3/19/2009 8:24:25 PM
Financial Architecture
Box 11.5
387
Lamfalussy diagnosis and priorities
Obstacles to development of European securities markets (p. 10): • The absence of clear Europe-wide regulation on a large number of issues (for example prospectuses, cross-border collateral, market abuse, investment service provision) which prevents the implementation of the mutual recognition system. • An inefficient regulatory system. • Inconsistent implementation, in part due to lack of an agreed interpretation of the rules that do exist. • A large number of transaction and clearing and settlement systems that fragment liquidity and increase costs, especially for cross-border clearing settlement. • The inadequate development of funded pension schemes in most Member States. The most important gaps in European regulation (p. 12): • Lack of commonly agreed guiding principles covering all financial services legislation. • Failure to make the mutual recognition principle work for the wholesale market business in the context of the Investment Services Directive (ISD); for regulated markets themselves; for the retail sector; or for a single passport prospectus working for cross-border capital raising. • Outdated rules on listing requirements, no distinction between admission to listing and to trading, and lack of a definition of a public offer. • Ambiguity over the scope and application of conduct of business rules (Article 11 of the ISD) as well as on the definition of who is a professional investor. • No appropriate rules to deal with alternative trading systems. • Potential inconsistencies between the E-commerce Directive and financial services directives. • No comprehensive market abuse regime. • No cross-border collateral arrangements. • No set of common European-wide accepted international accounting standards. • Outdated investment rules for UCITS and pension funds. • Unresolved public policy issues for clearing and settlement activities. • No agreed takeover rules. • No high and equivalent levels of consumer protection and no efficient methods for resolving cross border consumer disputes. The main priorities for reform (p. 13): • A single prospectus for issuers, with a mandatory self-registration system. • Modernisation of admission to listing requirements and introduction of a clear distinction between admission to listing and trading. • Generalisation of the home country principle (mutual recognition) for wholesale markets, including a clear definition of the professional investor. • Modernisation and expansion of investment rules for investment funds and pension funds. • Adoption of International Accounting Standards. • A single passport for recognised stock markets (on the basis of the home country control principle).
9780230_547018_13_cha11.indd 387
3/19/2009 8:24:25 PM
388
Designing the European Model
Box 11.6
Basel II
Basel II will reform the 1988 Basel Accord on Capital Requirements to adjust them better for risk. Capital requirements, supervision/intervention and market discipline/disclosure are the three pillars of regulatory reform. As regards capital requirements, banks have two options. The first (standard approach) consists of relying on credit rating by external agencies to set the risk weights for different types of loans (say corporate, banks and sovereign claims). The second consists of relying on internal rating: banks themselves estimate probabilities of default, and assess the loss given default in an advanced version of the method. The idea is to calibrate the capital requirement so that it covers the Value at Risk from the loan (expected and unexpected losses from the loan) under some appropriate set of assumptions. The implementation of Basel II will require a complex and technical directive (given that a lot is at stake for financial institutions).
desire of the EP to retain control over the process, but eventually the EP settled on the proposal, in exchange for the promise that the Commission would take “utmost account” of its views. We note here that the Commission had already promised not to go against the predominant views of the Council as regards key implementation issues. The EU Securities Committee is chaired by the European Commissioner in charge, and takes decisions by qualified majority voting. Its decisions are prepared by the independent Committee of European Securities Regulators (CESR, see Lannoo (2002)). But as regards the implementation of the revisions of the 1988 Basel Capital Accord (see Box 11.6), an empowered Banking Advisory Committee (BAC) could end up having a similar role as the EU Securities Committee. Both will be in charge of interpreting and adapting the EU directive that establishes the framework for the application of Basel II. The Giovannini Group was formed in 1996 to advise the Commission on financial integration. The latest report of the group concentrates on existing problems in cross-border clearing and settlement in the markets for fixedincome securities, equities and exchange-trades derivatives. The group has proposed that systems should be judged against criteria of cost efficiency, accessibility, and safety and soundness. Two recent initiatives of the European Commission deserve to be mentioned. The first is a proposal concerning regulation of listed companies – the so-called single European prospectus. The second proposal is about allowing investors to by-pass stock exchanges. The proposal for a single European prospectus, agreed in November 2002 by EU finance ministers, but still to be approved by the European Parliament, allows securities (equity and bonds) to be issued with a single prospectus approved by the home regulator. A listed company, for example, would be regulated by the authorities of the country where the stock exchange is situated. The same applies to bonds under a value threshold. This is consistent
9780230_547018_13_cha11.indd 388
3/19/2009 8:24:25 PM
Financial Architecture
389
with the “home country principle” in financial supervision and is designed to help firms raise capital with a single document. Once approved by domestic regulators, the “prospectus” (or main document for listing) has to be accepted by all EU exchanges. However, Germany and the United Kingdom would prefer to let companies choose the listing authority by which to be regulated. Freedom of choice corresponds to current practice. Very recently, in its proposal for an Investment Services Directive, the European Commission has envisaged scrapping the rules forcing investors to trade only via stock exchanges. Investment banks would be allowed to compete with stock exchanges by trading shares for their clients inhouse, disclosing prices before the market. 23 Investment banks would also be able to operate across the EU when authorised in their home country. The idea is to allow investors to by-pass stock exchanges and trade directly via investment banks so that regulatory fragmentation of stock exchanges would not prevent cross-border trade. At present, banks are already allowed to trade in-house in some countries such as the United Kingdom and Germany, whereas in other countries such as France, Italy and Spain all major trades have to go through the organised stock market. Large investment banks and stock exchanges (like London and Frankfurt) are to benefit from these measures while smaller banks and national exchanges may suffer. The evaluation of such a proposal comes down to gauging the trade-off between more competition, enhanced by the rivalry between investment banks and exchanges, and the decrease in liquidity in stock exchanges as well as the lack of transparency of in-house trading. At first blush, the proposed measure seems to accept defeat conceding that the emergence of integrated European stock markets is difficult despite the fact that markets are superior precisely in price discovery and facilitating transactions.
11.5 European financial architecture: diagnosis and proposals for reform 11.5.1
Diagnosis
Potential increase in risk The introduction of the euro implies the consolidation of deep and liquid financial markets in the euro area as well as in the EU. As financial integration advances, it is likely that the relative weights of financial intermediaries and markets in continental Europe will shift towards the latter. Deeper and more integrated markets increase diversification possibilities, but at the same time raise potential problems of contagion and liquidity crises. Indeed, as European financial markets become integrated, cross-border externalities increase: the failure of an institution in one country may have effects on the financial system of other European countries. This may happen either
9780230_547018_13_cha11.indd 389
3/19/2009 8:24:25 PM
390
Designing the European Model
because of default in interbank commitments, or via problems in the payments mechanism. Furthermore, credit risk may increase in the national economies because the exchange rate and monetary policy buffers are no longer in place (although diversification possibilities may increase and exchange rate risk eliminated in the euro area). At the same time, stronger competition will impinge upon the restructuring of the banking sector creating difficulties for weak institutions and/or enhancing the incentives for banks to take more risk. The exposure of European banks to emerging markets may be an example of the latter. It is even questionable whether the recent wave of domestic mergers adds to stability. This is so because enhanced diversification possibilities (which are relevant given the diversity of regions inside European nations24) through consolidation might be more than compensated by the perverse incentives induced by the TBTF policy applied to national champions. Note that, to the extent that regulatory and political obstacles hinder cross-border consolidation, they end up exacerbating this problem. Overall, these considerations lead us to conclude that the fragility of the banking system may well increase in the short term. The consequences of regulatory fragmentation Regulatory fragmentation in Europe is a major obstacle to financial integration. It reduces the international competitiveness of European markets and institutions, and poses a threat to the stability of the financial sector. There is a wide consensus on the first and second issue (as clearly shown by the Lamfalussy and Giovannini reports). It could be argued that Europeans should not be too concerned with the stability of the financial sector, precisely because European financial markets remain segmented. Indeed, one may interpret the statements of the European Commission and the ESCB along this line. For example, the Brouwer Report (2000) on financial stability concluded: “The existing institutional arrangements provide a coherent and flexible basis for safeguarding financial stability in Europe. No institutional changes are deemed necessary.” However, capital markets are integrating steadily. Although it is true that the retail business remains segmented, changes may happen relatively fast (with the expansion of electronic banking, for example). The role of disclosure and market discipline It has been argued that disclosure requirements and market discipline are a substitute for financial architecture design. 25 For example, in the present decentralised supervisory framework, an increase in disclosure by financial intermediaries would contribute to increase market discipline and reduce information asymmetries among European supervisors.26 The LTCM crisis provides a paradigmatic example: If the banks that had lent to LTCM had declared their positions, then supervisors and market agents could have
9780230_547018_13_cha11.indd 390
3/19/2009 8:24:25 PM
Financial Architecture
391
acted upon it. However, relying on transparency and market discipline alone is not without problems. First, more transparency may increase, rather than decrease instability.27 Second, a problem of relying on market discipline is that agents, small investors in particular, have an incentive to free-ride on the information generated by others on financial institutions.28 11.5.2
Thinking ahead
Crisis lending and crisis management in the euro area The present system of “improvised co-operation” in a crisis situation may not be adequate and put the stability of the system in danger. The value of centralised authority with appropriate information is enhanced in crisis situations. This responsibility for stability can only be assumed by the ESCB and the ECB in particular. The ESCB should explicitly assume the function of guarantor of the system. 29 This would probably only require a broad interpretation of the Treaty (Article 105(2 and 5)) on the contribution of the ESCB to the smooth operation of the payment system and the stability of the financial system. At the same time the ESCB should establish and make public a formal framework of crisis resolution. The chain of command in a crisis situation should be clearly identified. Duisenberg’s declaration of October 1999 in the European Parliament, on the division of responsibilities between national central banks and the ECB, is a step in this direction but what is to be done with transnational institutions should be clarified. By leaving open the resolution of the many problems raised by the presence of transnational financial institutions, the present system imposes discipline (controlling moral hazard) at too high a cost in terms of systemic stability. An explicit recognition of the role of the ECB could instead enhance the response to systemic financial stability concerns, counting on the ECB’s reputation not to create moral hazard problems (due to expectations of excessive help). The ECB should be able to develop such a reputation given its strong credentials. The formal recognition of the role of the ECB as a lender of last resort is not in contradiction with maintaining a degree of “constructive ambiguity” about the circumstances of intervention. Indeed, transparency in the procedures to follow in crisis situations provides a reference point for the markets, and minimises costly bargaining ex post among authorities. It also provides a decision-making framework that should guarantee fast responses, with clearly defined responsibilities for the different institutions involved. Crisis lending cannot be separated from fiscal issues when liquidity problems end up in insolvency. When this happens to a transnational financial institution, a procedure must be devised to share the fiscal costs of the intervention. A formal mechanism of co-operation should be established between the ECB, the NCBs and/or national supervisors, and the national treasuries to clarify responsibilities, establish information sharing protocols, and elucidate who would pay for failed (insolvent) institutions that have
9780230_547018_13_cha11.indd 391
3/19/2009 8:24:25 PM
392 Designing the European Model
been helped. The European Union Council of Finance Ministers (Ecofin) could have a consultative role when the ECB initiates interventions that may end up in losses to be paid with tax money. This proposal is in line with the idea launched in April 2002 by Mr. Eichel of Germany and Mr. Brown of the United Kingdom to establish a “European stability forum.” Supervision and political economy The review of supervisory arrangements in Sections 3 and 4 points to the need for more centralised supervision mechanisms in order to internalise cross-border effects and foster financial integration. In an integrated market the mere co-ordination of financial supervision may prove insufficient. The question then is how to devise a supervisory system for the euro area (as well as the EU at large) that promotes financial integration and the competitiveness of European institutions and markets, while at the same time guaranteeing financial stability in a long-run perspective. There are at least two alternative models, apart from the current decentralised arrangement. In the first model prudential supervision of banks is in the hands of the ESCB with the ECB having a central role while Europeanwide specialised regulators in insurance and securities are constituted. In the second model, an integrated regulator of banking, insurance and markets – a European Financial Services Authority (EFSA) – is formed, while the ECB (in the ESCB) is responsible for systemic problems. In either model it must be noted that the lender of last resort function would require the ECB to have some monitoring powers. This concerns in particular the power to access supervisory records and gather information. This seems possible without amending the Treaty of the EU. A central bank in charge of systemic stability needs access to supervisory information. For instance, suppose that facing a major threat to financial stability and lacking supervisory capacity, the ECB will have to base its actions on information provided by national authorities. Not only might national authorities be tempted to under-report problems; greater access to information for the ECB would save costs in communication and negotiation, as well as facilitating the exchange of information. The first solution centralises supervision of banking in the ECB, but maintains the implementation in the decentralised structure of the ESCB. This solution would probably be favoured by the ECB, but disliked by the NCBs and national governments. The attempt at the Nice EU summit to enlarge the supervisory responsibilities of the ECB failed because of pressure from NCBs. (It was proposed but not accepted to extend the majority voting decision procedure to the article in the Treaty of the EU that envisages a larger role of the ECB in banking supervision.) As regards the establishment of a European Securities and Exchange Commission as a supervisory body for European financial markets, the EU Securities Committee proposals in the Lamfalussy Report could be seen
9780230_547018_13_cha11.indd 392
3/19/2009 8:24:25 PM
Financial Architecture
393
as a first significant step in this direction. But the main message of the Lamfalussy Committee is that a lot of preliminary harmonisation work among the different national authorities remains to be done in such disparate areas as legal frameworks, surveillance of settlements systems, disclosure and enforcement. The challenge is to develop a common framework that allows different market institutions and trading systems to compete. The case for a European Financial Services Authority (EFSA), with authority over banking, insurance and securities, is based on the trend toward integration of intermediaries and market operations, which makes it increasingly difficult to separate credit and market risk. Such an independent agency would bring relief also to the potential conflict between monetary policy and supervision of the financial system. The EFSA could have a horizontal structure with one division in charge of prudential supervision (monitoring credit and market risk), and another in charge of investor protection and conduct-of-business rules. An alternative model could have three divisions for banks, insurance companies and markets, but then the synergies of working with well-defined objectives might be lost. Political-economy considerations indicate that an independent EFSA, along with the ECB itself, might better resist local pressure to assist particular institutions. In principle, an EFSA would facilitate accountability, as both the ECB and the EFSA would then have well-defined missions, and would not increase the power of the ECB, which is already perceived as very powerful. However, note that such an agency would face the same accountability problem as the ECB, namely the lack of a well-defined European political principal. In either of the two models, supervision need not be completely centralised at the European level. First, because national supervisors will need to be involved in the day-to-day supervisory operations. Second, because a twotier system with some scope for regulatory competition can be envisioned because European level agencies could leave entities trading mostly within one national jurisdiction to be supervised by the appropriate national regulator (under the home-country principle). Neither an EFSA nor centralisation of supervision at the ECB level are proposals for the immediate future. The first would require a Council decision, the second a change in the Treaty of the EU. However, an open debate about this long-term aspect of European financial architecture is needed as well as leaving the door open in the Convention on the Future of Europe for the necessary institutional changes to implement more centralised regulation.
Notes 1. The reader is referred to Vives (2000, 2001) for a development of some of the arguments presented in this chapter. 2. See Chapter 4 in European Commission (2001).
9780230_547018_13_cha11.indd 393
3/19/2009 8:24:25 PM
394 Designing the European Model 3. For example, the market share of subsidiaries and branches of foreign credit institutions as a percentage of the total assets of domestic credit was less than 13 percent for the euro-area average as at the end of 1997 (see ECB, 1999). The exceptions are Belgium, Ireland and Luxembourg. 4. In spite of this, French and German banks have foreign assets in branches and subsidiaries amounting to about a third of domestic assets (see ECB, 1999). 5. For example, some countries lag behind in the move towards services, like France, Italy and Spain, as compared to Germany or the United Kingdom. 6. In the United States, recent studies indicate that there are sizeable economies of diversification in macroeconomic risk that can be exploited by means of mergers of entities in different states (Hughes et al. 1996, 1998). In Europe these economies of international diversification are partly limited by the increasing correlation in the business cycles of different countries (and the reduction in correlation between regions belonging to the same country). However, Berger, DeYoung, and Udell (2000) report that correlations of bank earnings across European nations are low, or even negative, relative to those across states in the United States. 7. National concentration levels in banking are much higher in European countries than in the United States at large, and they have tended to increase, particularly for smaller countries. For example, the concentration ratio C5 for deposits ranges from 30 to 80 percent in EU countries, with the exception of Germany which is less concentrated. 8. In the crises in Spain and in Scandinavia, also factors other than financial liberalisation were involved, that is the economic recession in Spain and, in Scandinavia, errors in fiscal and monetary policies which helped to inflate the speculative bubble. In all cases there was poor management, along with deficiencies in banking supervision. 9. There is an argument pointing at a stronger need for diversification of credit risk in a single currency area. As a single monetary policy responds to an average of shocks hitting the different regions of the euro area, it becomes less effective (relative to national monetary policies) in stabilising local demand conditions. Hence, after the introduction of the euro, the possibility of asymmetric business cycle developments increases the credit risk in any specific region of the Union. Obviously, this effect has to be set against the smaller exchange rate risk between euro countries. 10. See also Padoa-Schioppa, 1999, member of the Executive Board of the ECB in charge of prudential supervision. The quotation in the text is from the introductory statement delivered on the occasion of the Presentation of the ECB’s Annual Report 1998 to the European Parliament in Strasbourg, 26 October 1999. 11. Other relevant Directives are that of investment services, (implemented in 1995), and those on own funds, solvency ratios and large exposures. The Directive on the Winding-Up of Credit Institutions was finally approved in 2001. It states that when a bank with branches in other member states goes bankrupt, the winding-up process will be governed by the bankruptcy proceedings of the home country. 12. Furthermore, the legal obstacles to the setting up of subsidiaries have practically disappeared, although there are still restraints on the takeover of domestic institutions by foreign banks (need for approval by the supervisory authority and other restrictions in some countries). 13. The harmonisation of minimum capital requirements may be needed to avoid the distortions induced by regulatory competition among national authorities.
9780230_547018_13_cha11.indd 394
3/19/2009 8:24:26 PM
Financial Architecture
14.
15. 16. 17.
18. 19. 20.
21. 22. 23. 24. 25. 26.
27. 28. 29.
395
For example, undersupply of capital regulation may follow from the fact that national solvency regulations create a positive international policy externality on foreign lenders of domestic banks (see Sinn, 2003). At the same time some ambiguity about the character of the intervention is maintained: “The form of the response would depend on the nature of the event and would be determined at the time”. (MOU, par. 12). However, the supervision of securities markets is in the hands of the Länder in Germany. Proposal for a directive on the supplementary supervision of credit institutions, insurance undertakings and investment firms in a financial conglomerate. For example, in the case of the failure of the Bank of Credit and Commerce International (BCCI), many of the clients were not residents of the country in which the bank was authorised to operate, that is Luxembourg, and hence the costs of failure were borne to a considerable degree by foreign clients or their insurers. Thus, for example, some local authorities in the United Kingdom withdrew their funds from small banks that had contracted risks with the BCCI. See Holthausen and Ronde (2001). The fact that national authorities cannot discriminate against foreign creditors in a winding-up process of a bankrupt bank (according to the Directive on the Winding-Up of Credit Institutions) may contribute to the undersupply of help to an international bank by the home country. The draft prospectus directive requires that member states de-couple listing from trading and to have the listing authority as an independent supervisory agency. The market abuse draft proposal requires that member states appoint one independent authority to deal with insider trading. In a first proposal, the European Commission allowed investment banks to disclose prices after the trade was conducted. See Danthine et al. (1999). See Favero et al. (2000). This is inspired by the New Zealand experiment where quarterly disclosure of relevant bank information is mandatory and there is no deposit insurance. A system of penalties, including the possibility of unlimited civil liability of banks’ directors for losses caused to creditors, enforces the disclosure requirements. See Mayes (1997) and Mayes and Vesala (1998). See Rochet and Vives (2002). New Zealand’s reliance on market discipline to control risk has the particularity that most banks are foreign and therefore supervised abroad. This has been argued by Chiappori et al. (1991), Vives (1992), Folkerts-Landau and Garber (1994), and more recently by Pratti and Schinasi (1999).
List of abbreviations BAC BAWE BBVA BCCI BNP SG
Banking Advisory Committee Bundesaufsichtsamt für Wertpapierhandel Banco Bilbao Vizcaya Argentaria Bank of Credit and Commerce International Banque Nationale de Paris Société Générale
9780230_547018_13_cha11.indd 395
3/19/2009 8:24:26 PM
396
Designing the European Model
BSC BSCH CB CD CESR DOJ ECB EEA EFC EFSA EP ESCB EU FDIC FESCO FRB FSA FSPG IC ISD LCTM LLR M&A MOU NCB OCC SCH TBTF UCITS
Banking Supervision Committee Banco Santander Central Hispano Central Bank Certificate of Deposit Committee of European Securities Regulators Department of Justice (US) European Central Bank European Economic Area Economic and Financial Committee European Financial Services Authority European Parliament European System of Central Banks European Union Federal Deposit Insurance Corporation Forum of European Securities Commissions Federal Reserve Board Financial Services Authority Financial Services Policy Group Insurance Committee Investment Services Directive Long-Term Capital Management Lender of Last Resort Mergers and Acquisitions Memorandum of Understanding National Central Bank Office of the Comptroller of the Currency Santander Central Hispano Too Big to Fail Undertakings for Collective Investments in Transferable Securities
References Bank of International Settlements (2000) Annual Report 2000. Berger, A., R. Demsetz and P.E. Strahan (1999) “The Consolidation of the Financial Services Industry: Causes, Consequences, and Implications for the Future,” Journal of Banking and Finance 23, 135–94. Berger, A., R. DeYoung and G. Udell (2000) “Efficiency Barriers to the Consolidation of the European Services Industry,” European Financial Management, 7(1), 117–30. Bini Smaghi, L. (2000) “Who Takes Care of Financial Stability in Europe?” in Chapter 2 of Open issues in European Central Banking, Palgrave Macmillan. Cordella, T. and E. Levy Yeyati (2002) “Financial Opening, Deposit Insurance, and Risk in a Model of Banking Competition,” European Economic Review 46(3), 471–85. Chiappori, P., C. Mayer, D. Neven and X. Vives (1991) “The Microeconomics of Monetary Union,” in: Monitoring European Integration: The Making of Monetary Union, CEPR Annual Report.
9780230_547018_13_cha11.indd 396
3/19/2009 8:24:26 PM
Financial Architecture
397
Danthine, J-P, F. Giavazzi, E-L von Thadden and X. Vives (1999) Monitoring European Integration: The Future of European Banking, CEPR, London. Demirgüç-Kunt, A. and E. Detragiache (1998) Financial Liberalization and Financial Fragility, mimeo, The World Bank. Dermine, J. (2002) “European Banking: Past, Present and Future,” paper prepared for the Central Banking Conference, ECB, Frankfurt am Main, October. Di Noia C. and G. Di Giorgio (1999) “Should Banking Supervision and Monetary Policy Tasks be Given to Different Agencies?” International Finance 2, 1–18. Diamond, D. and P. Dybvig (1983) “Bank Runs, Deposit Insurance and Liquidity,” Journal of Political Economy 91, 401–19. Economic and Financial Committee (2000) “Report on Financial Stability,” (Brouwer Report), Economic Papers no. 143. Economic and Financial Committee (2001) “Report on Crisis Management,” Economic Papers no. 156. Economic and Financial Committee (2002) “Report on EU Financial Integration,” Economic Papers no. 171. European Commission (2001) “The EU Economy 2001 Review European Economy,” Economic Papers no. 73. European Central Bank (1999) “Possible Effects of EMU on the EU Banking System in the Medium to Long Term,” February. Favero, C., X. Freixas, T. Persson and Ch. Wyplosz (2000) “One Money, Many Countries,” CEPR Monitoring the European Central Bank 2, January, 74. Folkerts-Landau, D. and P. Garber (1994) “The ECB: A Bank or a Monetary Policy Rule,” in M. Canzoneri, V. Grilli and P. Masson, eds, Establishing a Central Bank: Issues in Europe and Lessons from the US, Cambridge, Cambridge University Press. Gale, D. and X. Vives (2002) “Dollarization, Bail Outs, and the Stability of the Banking System,” Quarterly Journal of Economics 117(2), 467–502. Giovannini Group (2002) “Cross-border Clearing and Settlement Arrangements in the European Union,” Economic Papers no. 163, February. Goodfriend, M. and R. King (1988) “Financial Deregulation Monetary Policy and Central Banking,” in W. Haraf and R. M. Kushmeider, AEI Studies, no 481, Lanham, MD, UPA. Goodhart, C. and D. Schoenmaker(1995) “Should the Functions of Monetary Policy and Banking Supervision be Separated?” Oxford Economic Papers 46, 539–60. Holthausen, C. and T. Ronde (2001) “Cooperation in International Banking Supervision: A Political Economy Approach,” mimeo, ECB, 1–43. Honohan, P. and D. Klingebiel (2001) “Controlling the Fiscal Costs of Banking Crises,” World Bank Working Paper 29. Hughes, J., W. Lang, L. Mester and C.-G. Moon (1996) “Efficient Banking under Intestate Branching,” Journal of Money, Credit, and Banking, 28, 1045–71. Hughes, J., W. Lang, L. Mester and C.-G. Moon (1998) “The Dollar and Sense of Bank Consolidation,” Federal Reserve Bank of Philadelphia W.P. no. 98/10. Jacklin, C. and S. Bhattacharya (1988) “Distinguishing Panics and InformationBased Bank Runs: Welfare and Policy Implications,” Journal of Political Economy 96, 568–92. Lamfalussy Report (2001) Final report of the committee of wise men on the regulation of European securities markets, February 15. Lannoo (2002) “Supervising the European Financial System,” CEPS Policy Brief no. 21. Matutes, C. and X. Vives (2000) “Imperfect Competition, Risk Taking and Regulation in Banking,” European Economic Review 44, 1–34.
9780230_547018_13_cha11.indd 397
3/19/2009 8:24:26 PM
398 Designing the European Model Mayes, D. (1997) “A Market Based Approach to Maintaining Systemic Stability: Experiences from New Zealand,” Bank of Finland Discussion Paper 18/97. Mayes, D. and J. Vesala (1998) “On the Problems of Home Country Control,” Bank of Finland, Discussion Paper no. 20/98. Meltzer, A. (1986) “Financial Failures and Financial Policies,” in G. G. Kaufman and R. C. Kormendi, eds, Deregulating Financial Services: Public Policy in Flux, Cambridge, MA, Ballinger. Memorandum of Understanding (MOU) between HM Treasury, the Bank of England and the Financial Serv. Authority, http://www.bankofengland.co.uk/financialstability/mou.htm. Padoa-Schioppa, T. (1999) “EMU and Banking Supervision,” lecture at the London School of Economics, 24 February. Postlewaite, A. and X. Vives (1987) “Bank Runs as a Equilibrium Phenomenon,” Journal of Political Economy 95, 485–91. Pratti, A. and G. Schinasi (1999) “Financial Stability in European Economic and Monetary Union,” IFS, no. 86, Department of Economics, Princeton University. Rochet, J-C. and X. Vives (2002) “Coordination Failures and the Lender of Last Resort: Was Bagehot Right After All?” CEPR DP 3233. Sinn, H. W. (2002) “Risk Taking, Limited Liability and the Competition of Bank Regulators,” FinanzArchiv, 59(3), 305–29. Vives, X. (1992) “The Supervisory Function of the European System of Central Banks,” Giornale degli Economisti e Annali di Economia 51, 523–32. Vives, X. (2000) “Central Banks and Supervision,” in Santomero, Viotti and Vredin, eds, Challenges for Modern Central Banking, Boston, Kluwer International. Vives, X. (2001) “Restructuring Financial Regulation in the European Monetary Union,” Journal of Financial Services Research 19(1), 57–82. Vives, X. (2002) “External Discipline and Financial Stability,” European Economic Review 46(4–5), 821–28.
9780230_547018_13_cha11.indd 398
3/19/2009 8:24:26 PM
12 Pensions and Children
12.1
Introduction
The crisis of the pensions system in Europe is primarily a demographic crisis, resulting from falling fertility and increasing longevity. Fertility rates fell from 2.1 children per female in the 1970s to 1.7 children per female in the 1990s, and they are not expected to pick up again in the future. The problem is compounded by increasing longevity. The ratio of people above 65 to people between 15 and 65 (one way to measure the so-called dependency ratio) in most European countries is expected to more than double, on average, in the EU countries between now and 2050. In percentage terms, such a ratio is expected to rise by at least 60 per cent in the case of the UK, up to 125 per cent in the case of Italy. In 2050 the dependency ratio is projected to be around 70 per cent in Italy and Spain; between 45 and 50 per cent in Germany and France; and just below 40 per cent in the UK. Changes in the dependency ratio are expected to be even more dramatic for the newly acceded countries – currently featuring relatively young populations, but expected to follow the same demographic pattern as the rest of Europe at a very fast pace. The magnitude of the demographic crisis is strikingly similar across countries. But its impact on pension systems differs depending on the structure and generosity of these systems. In some countries, the issue is amplified by other sources of macroeconomic fragility, namely, fiscal problems. A proper assessment of public pension liabilities needs to account for contingent implicit debt that derives from implicit guarantees of the welfare of pensioners (minimum pension, bailouts of bankrupt private funds and so forth). From a fiscal perspective, addressing the pension crisis requires reforms of the public social security systems in light of demographic dynamics that were not foreseen at the time when the systems were introduced.1 In many countries, demographic developments have been irresponsibly ignored for at least three decades in most countries. The extent of required reforms varies 399
9780230_547018_14_cha12.indd 399
3/19/2009 8:34:59 PM
400
Designing the European Model
Box 12.1
Why a social security system?
Why not leave retirement saving decisions to individuals rather than organising mandatory public systems or private systems strictly regulated by law? There are four main arguments. Historically, the argument was to avoid undesirable behaviour within a family. Either rotten children refuse to take care of their parents – an argument invoked by Bismarck – or parents may free ride on the altruism of their children, and consume too much, without taking any measures to sustain their lives as retired people. Free riding on the altruism of others may also have a collective dimension if people expect help from the community in case of need in their old age. This is nothing but a well-understood moral hazard issue in setting up the welfare state. A different and most powerful argument stresses myopia in savings behaviour. A wealth of empirical evidence suggests that people start to save for their retirement at a quite advanced age – raising an issue as of whether this would lead to irrational under-saving. A mandatory pension system would therefore be a paternalistic measure to force myopic workers to protect themselves, and save enough for their own good. A recent view stresses that people do not make life-time plans or revise their plans at different stages of their life in an inconsistent way. In particular, people seem to use a higher discount rate for the near future than for the far future. As a consequence, although they care about their retirement, they end up postponing the beginning of a serious savings plan. It is interesting to note that, while this view is shared by many economists and policy analysts, projection models set up to calculate welfare improvements from pension reform almost never account for individuals with such characteristics; see Angeletos et al. (2001) and Choi et al. (2002). However, one could note that what this argument really calls for – instead of a public pension system – is the existence of illiquid assets which would allow people to commit not to dissave. A third set of arguments refers to inefficiency and limitations of financial markets. An efficient market for annuities – where people can convert their pension funds accumulated throughout their working life into a stream of monthly payments up to their death – is an essential element of saving for retirement. It is well known that such markets may not work well because of adverse selection problems and because the size of the financial intermediary that would provide annuities would be too large, thus granting monopoly power. However, the adverse selection argument may be overstated, as characteristics that affect life expectancy are well observable. Moreover, the performance of a private market for annuities should be assessed against the yield of public pension systems. A fourth argument is that a pay-as-you-go pension system insures people against not having children. In principle people could be sustained in their old age by their children. However, some people cannot have children either for biological reasons or because they do not find an appropriate partner. The public pension system can be seen as a device to protect these people against low income in old age. This argument is particularly strong when capital markets are inadequately developed so that only children are a realistic means to ensure
9780230_547018_14_cha12.indd 400
3/19/2009 8:35:00 PM
Pensions and Children
401
sufficient resources in old age. However, the argument would hold through even with perfect capital markets insofar as education offers infra-marginal returns above the market rate of interest – implying families with children are better off. Note that this view presupposes a commitment by children to sustain their parents, to which they may not agree. With a pay-as-you-go pension system, a similar duty is extended to the society as a whole. In any case, it is necessary for a society to produce enough members of future generations so as to sustain the needs of its members.
across countries, but in most cases radical steps are needed in reconsidering the coverage of the public systems, the degree of redistribution and fairness within the system, and the average level of benefits. It is also necessary to reduce or prevent altogether the abuse of social security as an improper instrument of industrial or redistributive policy. It is important to note that balancing the public pension system from an accounting perspective does not automatically imply that the system is optimal, nor moving towards optimality. Therefore, it is important to focus the debate on what the desirable characteristics of a pension system should be. From a macroeconomic perspective, fewer workers relative to non-active people means that, for a given productivity, domestic output is on average lower than it would be if there was a younger population. Even if the government takes the necessary steps to ensure fiscal sustainability, it would still be true that citizens will face the challenge to maintain and raise their average standard of living. There is a trade-off between consumption and leisure: either people should work longer and/or save more to accumulate more human and physical capital, or living standards should fall relative to the economy’s productivity trend. In light of these considerations, addressing the pension crisis entails much more than quantitative fiscal policy: it also requires the design of policies that create incentives (or remove disincentives) to innovate and finance innovations, to participate in the labour force, etc. In designing fiscal measures, microeconomic and incentive-related issues play a dominant role. When addressing social security issues, distribution and efficiency are strictly interconnected: any proposal of reforms will involve some tradeoffs between different objectives, shifting the incidence and magnitude of distortions in different areas of the economy, including the labour and financial markets, but also fertility choice. An analysis of the principles that should guide reforms needs to focus on the relevant trade-offs, rather than addressing each distortion on its own. In this chapter, we reconsider the root of the current crisis in the social security systems in Europe, stressing its macroeconomic dimensions. We then consider different proposals to reform the system.2
9780230_547018_14_cha12.indd 401
3/19/2009 8:35:00 PM
402 Designing the European Model
12.2 The demographic challenge to pension systems Table 12.1 and the Appendix convey the striking dimension of the demographic challenge to social security in an immediate way. In 2050, the lowest old-age dependency ratio (ratio of people above 65 to people between 15 and 65) is projected to climb just above 30 per cent for the US and the UK. Currently, the highest dependency ratio – recorded for Italy – is below that figure.3 For the EU-15, the average dependency ratio is currently 21 per cent. It will more than double, to about 50 per cent in 2050. Strikingly, these averages are essentially identical for the newly acceded countries as a group. Even within Europe, individual countries differ in the extent of the demographic change. Due to very low fertility rates, the situation is extremely alarming for Italy and Spain, which are expected to achieve the highest dependency ratios – close to Japan. On the other side of the spectrum, there are Denmark, Ireland, the Netherlands and the UK, which are quite far from those levels. Yet with dependency ratios around 40 per cent, the situation is critical enough. It should be noted that a somewhat different picture evolves when one looks at completed fertility rates, calculated using the number of children during a woman’s reproductive lifetime. Completed fertility rate is sometimes around or above two. This means that part of the observed low fertility rates in many European countries is a transitory phenomenon, due to the fact that women currently in their reproductive lifetime tend to have children at a later stage relative to previous generations. So the low fertility rates in the data reflect the coexistence of older cohorts who have already completed their fertility, with younger cohorts who have not yet reached their peak rates, which is expected to occur at a higher age than for older cohorts. This point is well illustrated by means of simulations for a “typical” European country. As a benchmark, we consider the case in which the fertility rate is 2.1 with a peak in fertility for women between 20 and 30 years old: in our simulations for this case, dependency ratios in the long run would equal 33.9 per cent, i.e. they would be barely affected. They would be somewhat higher than current dependency ratios only because of lower mortality rates among pensioners. Against the above scenario, suppose that the fertility rate were to fall to 1.3, leaving the peak of fertility for women between 20 and 30 years old: dependency ratios would then go up to 54 per cent in the long run – and population would shrink by 1.3 per cent a year. Finally, suppose the fertility rate were to remain at 2.1 but women experienced a peak in fertility between 28 and 38: in this case the long-run dependency ratio would be 34.0 per cent, remaining virtually unchanged relative to the benchmark above.4 Yet the current demographic indicators would record a temporary deterioration. However, one should not count too much on a recovery of fertility. While it is true that in some countries completed fertility is around 2 (the UK,
9780230_547018_14_cha12.indd 402
3/19/2009 8:35:00 PM
9780230_547018_14_cha12.indd 403
3/19/2009 8:35:00 PM
0.50 0.47 0.68 0.48 0.55 0.57 0.58 0.52 0.49 0.55 0.48 0.65 0.53 0.57 0.57 0.42
1950
((0–15)+ (65–100))/ (15–64)
0.16 0.16 0.10 0.12 0.14 0.17 0.11 0.17 0.14 0.11 0.11 0.18 0.13 0.18 0.15 0.14
(65–100)/ (15–64)
Age groups
0.47 0.53 0.50 0.41 0.51 0.46 0.50 0.53 0.49 0.50 0.44 0.47 0.50 0.46 0.49 0.49
2004
0.24 0.27 0.18 0.20 0.23 0.24 0.23 0.25 0.27 0.28 0.22 0.17 0.29 0.24 0.22 0.21
(65–100)/ (15–64)
Age groups ((0–15)+ (65–100))/ (15–64)
Dependency ratios in selected countries
Austria Belgium Cyprus Czech Republic Denmark Estonia Finland France Germany Greece Hungary Ireland Italy Latvia Lithuania Luxembourg
Table 12.1
0.56 0.62 0.58 0.55 0.60 0.56 0.67 0.63 0.60 0.57 0.53 0.54 0.59 0.54 0.55 0.50
2025
((0–15)+ (65–100))/ (15–64)
0.36 0.37 0.29 0.35 0.35 0.33 0.41 0.36 0.38 0.38 0.32 0.25 0.40 0.33 0.30 0.25
(65–100)/ (15–64)
Age groups
0.79 0.74 0.66 0.84 0.68 0.84 0.72 0.73 0.76 0.88 0.75 0.69 0.90 0.82 0.74 0.62
Continued
0.55 0.47 0.39 0.59 0.42 0.57 0.46 0.46 0.49 0.62 0.50 0.41 0.65 0.56 0.43 0.35
(65–100)/ (15–64)
2050
((0–15)+ (65–100))/ (15–64)
Age groups
9780230_547018_14_cha12.indd 404
3/19/2009 8:35:00 PM
0.53 0.57 0.55
0.53 0.52 0.51
0.49 0.68
0.54
Poland Portugal Slovakia
Slovenia Spain Sweden
UK Japan
US
0.13
0.16 0.08
0.11 0.11 0.15
0.08 0.11 0.10
0.10 0.12
0.51
0.52 0.50
0.42 0.45 0.54
0.42 0.49 0.41
0.47 0.48
2004
0.18
0.24 0.29
0.21 0.25 0.27
0.18 0.24 0.16
0.19 0.21
(65–100)/ (15–64)
Age groups ((0–15)+ (65–100))/ (15–64)
Source: United Nations; Population Division, 2004; EEAG.
0.69 0.59
1950
(65–100)/ (15–64)
Age groups
((0–15)+ (65–100))/ (15–64)
Continued
Malta Netherlands
Table 12.1
0.60
0.56 0.70
0.56 0.53 0.67
0.53 0.53 0.50
0.63 0.58
0.28
0.31 0.50
0.38 0.35 0.40
0.31 0.32 0.27
0.36 0.33
(65–100)/ (15–64)
2025
((0–15)+ (65–100))/ (15–64)
Age groups
0.31
0.65 0.98
0.89 0.93 0.73
0.75 0.79 0.73
0.73 0.68
0.32
0.38 0.72
0.64 0.68 0.47
0.50 0.53 0.47
0.46 0.42
(65–100)/ (15–64)
2050
((0–15)+ (65–100))/ (15–64)
Age groups
Pensions and Children
405
Norway and France) for the latest available cohorts, that is, those born in 1960, it is only equal to 1.6 in Germany, 1.7 in Italy and 1.5 in Greece.5 Furthermore, it is likely that cohorts born after 1960 will have even fewer children. Otherwise, it would be very difficult to explain the observed age pyramids, with cohort size steadily going down with the date of birth. Because the most numerous cohorts are those between 30 and 40 years of age, if these cohorts had their maximum fertility in these ages, we should currently observe much higher birth rates. In principle, one could explain the currently low birth rate if these cohorts had their fertility peaks below 30 years of age, while younger cohorts of women ones have later peaks. But if this were the case, birth rates should have been much higher than observed ten years ago.
12.3 Some unpleasant growth arithmetic To convey the implications of adverse demographic developments in an immediate way, it is useful to build a simple numerical example showing the implications of ageing for a country’s productive capacity. In particular, assume that workers’ productivity growth is two per cent per year in EU-15 as well as in Japan and the US, while it is 2.5 per cent in the new EU states. Assume also that the participation rate in the labour force is either constant at the current rate, or increases from the current rate up to 80 per cent in 2050: for European countries this implies eight percentage points average rise in participation. Based on this assumption, we calculate the change in per capita output for the population above the age of 15. This measure of per capita output is a rough measure of output available for workers and retired people. Needless to say, the results are very sensitive to the parameters assumed in the exercise. Results are shown in Table 12.2. As our benchmark, consider a hypothetical situation with no ageing. For the EU-15, other things equal, a two per cent productivity growth per year would raise average output per adult by a factor of 2.44 between 2004 and 2005. Accounting for population ageing, however, the same average output only grows by a factor of 1.63. This figure is raised to 1.82 if substantial gains in the participation rate are achieved (that is, more people of working age actually do work). Note that, relative to the scenario with no adverse demographic development, these average figures are equivalent to a productivity slowdown (for a given population structure) from 2 to 1.1 or 1.3 per cent per year. The outlook is particularly grim in Italy, Spain and Greece. In the case of Spain, for instance, a two per cent productivity growth barely compensates for the projected population ageing, implying a quasi-stagnation of output per capita. Even when extreme gains in participation rates are assumed, income gains are only of the order of 50 per cent over the 45-year period.
9780230_547018_14_cha12.indd 405
3/19/2009 8:35:00 PM
406 Designing the European Model Table 12.2
The effects of ageing on per capita output Dependency Participation ratio rate
Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden UK EU 15-Average Cyprus Czech Republic Estonia Hungary Latvia Lithuania Malta Poland Slovakia Slovenia Average Japan US
Gains in per capita output
2004
2050
Constant Increasing If no participation participation ageing
0.24 0.27 0.23 0.23 0.25 0.27 0.28 0.17 0.29 0.21 0.21 0.24 0.25 0.27 0.24
0.55 0.47 0.42 0.46 0.46 0.49 0.62 0.41 0.65 0.35 0.42 0.53 0.68 0.47 0.38
80.00 67.00 81.00 74.00 70.00 76.00 66.00 70.00 63.00 67.00 70.00 76.00 71.00 76.00 76.00
1.44 1.77 1.84 1.71 1.76 1.70 1.29 1.73 1.20 2.01 1.79 1.51 1.04 1.77 1.99
1.44 2.11 1.81 1.85 2.01 1.79 1.56 1.98 1.53 2.40 2.05 1.59 1.17 1.86 2.09
2.44 2.44 2.44 2.44 2.44 2.44 2.44 2.44 2.44 2.44 2.44 2.44 2.44 2.44 2.44
0.24 0.18 0.20
0.49 0.39 0.59
72.20 NA NA
1.64 2.26 1.56
1.82 – –
2.44 3.04 3.04
0.24 0.22 0.24 0.22 0.19 0.18 0.16 0.21 0.21 0.29 0.18
0.57 0.50 0.56 0.43 0.46 0.50 0.47 0.64 0.51 0.72 0.32
NA NA NA NA NA NA NA NA NA 78.00 85.00
1.72 1.95 1.76 2.22 2.03 1.85 1.92 1.38 1.86 0.96 2.02
– – – – – – – – – 0.99 2.16
3.04 3.04 3.04 3.04 3.04 3.04 3.04 3.04 3.04 2.44 2.44
Note: NA: not available. Source: EEAG.
Table 12.2 shows that raising participation rates does make a difference. If the gain in participation rate is achieved via a delay in retirement age, this result is just a way to reiterate that the impact of longevity on the pension system can be mitigated by adjusting the retirement age. Figure 12.1, taken from
9780230_547018_14_cha12.indd 406
3/19/2009 8:35:00 PM
Pensions and Children
407
70 68 66 64 62 60 58 56
Sl
H un ov ak T gar R ur k y ep e u y Po blic Lu Be lan xe lg d m ium b Fi our nl g F an Au ran d s ce N A tral et u ia he st rla ria nd C ze It s U c ni h K aly te R or d ep ea Ki u n bl G gdo ic er m m an D Spa y en i N Cm n ew a ar Ze nadk U S ala a ni w nd te e d de St n Sw G ate itz ree s er ce N lan or d Ire way Po la rtu nd Ja gal Ic pan e M lan ex d ic o
54
Figure 12.1
Estimated effective retirement age of older male workers in 2000
Source: OECD.
the OECD, shows a wide dispersion in the average effective retirement age for male workers across countries in 2000, which varies between 56 and 62.6 There are two main lessons from our numerical example. First, in the next few decades, when the impact of demographic changes on pension systems will be at its peak, raising retirees’ living standard in line with productivity will only be possible if the economy as a whole consumes its wealth, i.e. consumes its stock of domestic and foreign capital. The numerical example clearly shows the large extent to which the decline in working-age population offsets output gains from productivity growth. The transition to a new stable population is likely to coincide with large decumulation of aggregate wealth stocks, especially starting in the 2030s, when the baby boomers born in the mid-1960s will receive their pensions.7 Making sure that our economies arrive at that time with a sufficiently large endowment of domestic and foreign capital is a necessary condition to prevent a crisis and manage economic and social issues that could possibly arise in the process Second, while running down the stock of national wealth will help smooth consumption during the transition to a new stable population structure, consuming capital stock cannot, however, be a permanent solution to the pension crisis. A higher stock of capital (human and physical) per worker is instead required in the long run when the population structure stabilises, reflecting lower fertility and increasing longevity. A higher capital per worker raises productivity levels and sustains living standards.
9780230_547018_14_cha12.indd 407
3/19/2009 8:35:01 PM
408 Designing the European Model
12.4 Issues facing the design of a pension system We now discuss major issues that, in our view, should be policy priorities in dealing with pension reform. These issues are often overlooked or misunderstood, and they inform the current debate on social security systems virtually in all European countries, despite the differences across national systems. In Europe, there is a wide variety of pension institutions, ranging from purely redistributive pay-as-you-go systems, to systems with an important funded component. Many countries are undertaking pension reforms that develop social securities in different directions, not only because there are differences in existing pension institutions and macroeconomic conditions but also because reforms may be inspired by alternative principles. In what follows, we will focus on a few general dimensions of pension reforms and then analyse possible solutions to the issues raised.8 12.4.1 Fairness “Equity” is a central concern of redistributive policies. Yet in many dimensions, equity concerns are ignored in designing these policies and reforming them. As far as pensions are concerned, there are inequities between individuals of the same generation as well as inequities between generations. Intragenerational inequities come from not taking individual preferences and characteristics into account. Those who systematically have a lower life expectancy, for example, because of their occupation, gender, or other observable characteristics, do not earn a higher pension per period.9 If their pensions were managed by a private insurance company, it would offer them a higher return on their pension wealth than to other agents, because it is known that this return would have to be paid only for a shorter period of time. As people who die earlier are “worse-off” than people who die later, market discrimination, in this case, tends to make outcomes fairer, while non-discrimination by the state makes outcomes less fair.10 A related issue in pension reform that is recently subject to widespread debate is that, in many social security systems, people who want to work longer typically lose the extra pension wealth they accumulate, while people who want to retire earlier often lose a greater share of their pension wealth than the one corresponding to their foregone contributions. The issue is how to make sure that the system can accommodate individual preferences about retirement age, without either penalising or favouring those who want to deviate from the legal retirement age. A more general problem is due to intergenerational inequities coming from differences in cohort size. Generations who work and retire in periods when the old-age dependency ratio is low are typically better off than generations who work and retire in periods with a high dependency ratio. Clearly, a low old-age dependency ratio reduces the workers’ contributions that are needed to sustain a given pension level; alternatively, it increases the pension level
9780230_547018_14_cha12.indd 408
3/19/2009 8:35:01 PM
Pensions and Children
409
that can be sustained at a given contribution rate. To our knowledge, none of the existing pay-as-you-go systems embodies an explicit mechanism to correct for these inequalities – requiring substantial smoothing of consumption across generations. This could be achieved if “luckier” generations accumulate buffer stocks of pension wealth to be eaten up by less lucky generations, according to a principle of intergenerational insurance (see, for instance, Allen and Gale (2000) Chapter 6). Another aspect of intergenerational equity is related to economic growth, which tends to make younger (and future) generations better off relative to older (and current) generations. We think these considerations should be taken into account when implementing the needed reform of the system, as financial considerations alone may lead to an unfair distribution of the burden of adjustment between generations. 12.4.2
Funded vs. pay-as-you-go systems
An important element of the debate concerns the choice between funded and unfunded pension systems. In an unfunded system, the contributions by the young are directly paid to the old. In a funded system, the contributions are invested in assets and the principal and return to these assets are used to pay for future pensions. Thus, under a funded system, the young are paying “for themselves” whereas under an unfunded system, they are paying “for the old.” As shown below, while an unfunded system is an intergenerational redistribution device that forces all generations to make gifts to previous generations, a funded system involves no such redistribution. The “returns” in an unfunded system are determined by the ratio between active contributors and retired beneficiaries. The higher that ratio, the greater the amount that can be paid per unit of contribution, and the greater the financial returns from one euro “invested” in the social security system. The returns in a funded system, by contrast, are simply determined by the rate of returns on financial assets. Does that mean that if the economy had been under a funded system from the start, one would not have to worry about an increased old-age dependency ratio? Not quite! While under an unfunded system, ageing of the society reduces the returns on contributions directly through the ratio between contributors and recipients, under a funded system it reduces the rate of return on capital, as the size of savings in search of productive investment opportunities goes up relative to the workforce. In fact, one can show that the dynamic response of the economy to an increase in the oldage dependency ratio is qualitatively similar under a funded system and an unfunded system. Thus, ageing of society would also be a matter of worry for retirees if pensions had been funded from the start. There are reasons to believe, however, that the implied reduction in pensions would be much lower under funding: for instance, a fraction of pension wealth can be consumed every year, in addition to the return on pension assets.11
9780230_547018_14_cha12.indd 409
3/19/2009 8:35:01 PM
410 Designing the European Model
Moreover, a fully-funded pension system can be insulated by the effects of domestic population changes on the returns on capital to the extent that capital markets are integrated, and pensions contributions are invested in foreign assets – provided the rest of the world does not have an ageing problem. With integrated capital markets, arbitrage links the domestic rate of returns to the international rate (note that this is true even when the two rates are not equalised). International diversification reduces the exposure of pension fund holders to domestic risk – including demographic risk. A funded system also yields a larger level of wealth and capital than an unfunded one. That is because social security contributions raise saving and therefore net investment in productive assets. Therefore, the economy is “richer” in the long run than under an unfunded system, meaning that it
Box 12.2 Shifting to a fully funded system cannot improve everybody’s welfare: an example It is sometimes heard that the transition to a fully funded system would pay for itself, because the contributions enjoy a greater rate of return than under an unfunded system. As discussed in the text, this argument is incorrect, as one generation must necessarily lose. A simple example will make this very clear, following Sinn (2000). In a pay-as-you-go system, a worker’s contributions to the system earn an ‘internal’ rate of return that can be easily calculated as the rate that equates the value of contributions C to the future pension benefits P C . (1i ) P where i denotes the “internal rate of return” of the pay-as-you-go system, referred to the entire lifespan of an individual, rather than to a single year. Since future pensions are paid with the contributions of future workers, the internal rate of return is just the rate of growth of total contributions from one generation of workers to another – proportional to the rate of growth of the economy. In the last decades, the yearly average of this rate has been of the order of two per cent, with some fluctuations depending on the growth rate of the economy (and also some disparity across individuals in the same generation). Now, if the worker could invest his/her retirement savings S in the financial markets, these would gain the market rate r. Thus, we can calculate the amount of saving invested in the market that would yield the same pension wealth P as follows: S
P 1 r
An estimate of the average yearly market rate is of the order of four per cent, twice as much as the internal rate of return of social security. Obviously, with these rates it will take much less resources to guarantee a pension P with the fully
9780230_547018_14_cha12.indd 410
3/19/2009 8:35:01 PM
Pensions and Children
411
funded system. Roughly: for a working life of 35–40 years, if the yearly return is four per cent for r and two per cent for i, then S is about one half of C. The alleged efficiency of the fully funded system stems from this simple consideration. But let us think of a fundamental difference between the two systems. In a fully-funded system, every generation accumulates its pension assets. In a payas-you-go system, the first generation of old people that receive their pension financed by active workers receive a “gift.” Thus, the contribution C paid by a worker to the social security system is the sum of an implicit tax T used to finance the initial “gift” when the system was set up, plus savings for retirement, S. Such a tax component is just the difference between the contribution to the pay-asyou-go system and the amount of saving S which would yield a pension as high as P at market rates T=C–S whose size is directly related to the difference between i and r. It is not difficult to see that the sum of the tax payments T over all generations contributing to the system is equal, in present discounted value, to the initial gift to the first generation of pensioners in the system. Denote by P0 the pension received by the first generation of retired people when the pay-as-you-go system is introduced, financed by the contribution Cl by the first generation of workers in the system P0 = C1 Assume that, after that, the pension P paid to each generation of retired people is also equal to the contribution paid by the current young people, although pensions are no longer a gift: each worker contributes into the system expecting to be rewarded in the future. So the first generation of workers who pays social security contributions equal to Cl expects pensions as high as P l = C2, where the subscript 2 indicates the second generation of workers in the system. Hence we can write C1 T1
P 1 r
C2
T1
1 r
Since this is true also for the second generation of workers and so on we can write C1 T1
T2 1 r
T3 2
(1 r )
...
Tn (1 r )
n
...
present discount value of Taxes
Cn1 (1 r )
n
Cn1 (1 r )
n
Because the growth rate of contributions (= pensions) in a pay-as-you-go system is linked to the economy’s growth rate, and is therefore lower than the market rate of return, the last term vanishes when we consider a long horizon n. Hence the initial net transfer when the system was created is exactly equal, in present value terms, to the cash flow that can be attributed to the tax component of the social security contribution; that is, T = S (r–i).
9780230_547018_14_cha12.indd 411
3/19/2009 8:35:02 PM
412 Designing the European Model Obviously, a pay-as-you-go system is an intergenerational redistribution device with a strict present value equivalence between all gains and all losses. This is bad news for those who believe in a low cost and easy transition from a pay-as-you-go to a funded system on the basis of superior efficiency of the latter. It can be shown that, at each point in time, the implicit pension debt in terms of the then-existing pension claims, is equal to the present value of all future implicit taxes. Suppose that a country decides to move from a pay-as-you-go to a fully funded system, financing the current pension payments by issuing public debt. This would make the implicit debt an explicit one. Government solvency requires explicit tax revenue to increase against the explicit public debt. The present value of the explicit taxes necessary to service the explicit debt is exactly equal in present value terms to the implicit taxes in an ongoing pay-as-you-go pension system. Thus nothing can be gained in present value terms by moving from one system to the other. It would even be possible with an appropriate borrowing strategy to choose a time path of the explicit tax fully in line with the implicit pay-as-you-go taxes. This is just another aspect of the equivalence between pay-as-you-go pensions and debt discussed in the text.
has accumulated more productive assets, implying higher wages and higher living standards. Furthermore, without a social security system, a rational worker would save for his retirement and invest in a well-diversified portfolio of financial and real assets earning the market rate. In an actuarial pay-as-you-go system, legal arrangements force him/her to save into an “asset” yielding a lower rate of return. This implies a constraint on portfolio investment, such that a component of the worker savings has to be put into what financial researchers call a “dominated asset,” that is, an asset whose returns are systematically below other assets available in the market. Do the considerations above imply that funded systems are uniformly superior to pay-as-you-go ones? The answer is, surprisingly, no. While all current and future generations are better off if the economy accumulates more productive assets (because they get higher wages during their work life), asset accumulation must be paid for by foregone consumption of previous generations. To understand this, assume the economy has had no pension system until now. If an unfunded system is introduced, it is possible to pay pensions to the existing old immediately. On the other hand, if the system is funded, one will have to wait until the proceeds of the investments are realised to start paying pensions, and one is unable to pay pensions to the existing generation of old. In other words, when one starts a pension system, initial tax proceeds are consumed by the old under an unfunded system, while they are invested under a funded system. Thus, the initial generation of old prefers the unfunded system, while all subsequent generations prefer the funded system. The argument also applies when the system is initially unfunded and one contemplates moving to a funded one: some generations have to invest instead of consuming, and it is impossible to make everybody better off. If for example one stops paying pensions and
9780230_547018_14_cha12.indd 412
3/19/2009 8:35:02 PM
Pensions and Children
413
invests contributions in productive assets instead, retirees obviously lose. If on the other hand, one increases contributions to finance existing pension claims, while at the same investing in a funded system to pay for future pensions, current generations of workers lose, relative to the status quo, as they have to pay twice: once for the old, once for themselves. In fact, the pay-as-you-go pension system can be interpreted as an intergenerational redistribution device: the gifts to the first generation of retirees (receiving a pension without having contributed to the system) are fully matched by losses of later generations – who receive a rate of return below the market rate of interest.12 The difference between the market rate of return, and the lower return granted by the social security system can be interpreted as an implicit tax that all future generations have to pay. The present value of this implicit tax is equal to the gains of the first generation (see Box 12.2).13 In general, at each point in time the implicit pension debt in terms of the then-existing pension claims is equal to the present value of implicit taxes yet to be paid by future generations of participants. This has important fiscal implications when an economy chooses to move from an unfunded to a funded system without negating the existing pension claims: such a reform requires an immediate redemption of the implicit debt by levying an explicit tax, and/or issuing explicit debt. If the choice is to finance the transition with a tax, the revenue from such a tax is equal, in present value terms, to the current and future revenues from the tax previously implicit in the payas-you security system. Then, the move just concentrates the time path of implicit taxes on the transition generation without changing the size of the tax burden in present value terms.14 So, while in the long run the funded system yields more wealth and capital accumulation, transition from one system to the other involves trade-offs: moving from one system to another implicitly, and inevitably, redistributes welfare among generations. 12.4.3 Fiscal budget process and pension reforms An important caveat in assessing the difference between pension systems is that one has a truly funded system only if contributions ultimately finance productive assets. One can show that an unfunded system is equivalent to public debt. That is not surprising: public debt is held by current generations, who buy it (the equivalent of their contributions), and it is paid back, with interest, by future generations who pay taxes (thus these taxes are equivalent to future generations’ contributions, while repayments are equivalent to pensions). Thus there is no difference between issuing public debt to give money to the existing retirees, and setting up a pay-as-you go system. Consequently, if one has a funded system where contributions are invested in newly issued public debt (i.e., correspond to a fiscal deficit), it is in fact not different from an unfunded system. The counterpart of
9780230_547018_14_cha12.indd 413
3/19/2009 8:35:02 PM
414 Designing the European Model
contributions is in the form of government expenditures rather than productive capital, and it is the taxes paid for by future generations, rather than the return to the investment, that provide the basis of future pensions. The problem would be slightly less severe if pension contributions matched new public capital (rather than public consumption), but this is unlikely to be a plausible scenario, given the limited role played by public expenditures on public capital in a typical fiscal budget. To sum up: a funded system whose counterpart is public debt is indeed not funded at all, in the sense that the consumption possibilities of future generations will be no different from what an equivalent pay-as-you-go system would have yielded. It follows that the goal of promoting capital accumulation could be pursued, instead of moving to a funded system, by reducing public debt. A fiscal consolidation would have the same effect on capital. In this sense, the trade-off between pay-as-you go and funded systems is not particularly important from a macroeconomic perspective, as there are alternative fiscal strategies to pursue similar macroeconomic goals. However, the two systems have quite different properties from a political and distributional viewpoint. The pay-as-you-go system typically leaves less room for individual choice (if workers were allowed to decide to contribute less, the system could not meet the claims of the old), and is therefore more politicised. It automatically encounters financing problems when the population becomes older and there is a political conflict between generations over how to fix the pension system’s budget. Each generation wants taxes to be increased immediately after it has retired. Under a funded system, that conflict can be solved by market forces alone, as equilibrium between supply and demand determines the change in pension levels – that is rates of return adjust downwards for investment opportunities to absorb the increase in savings brought about by ageing. If people can freely determine how much they contribute and when they retire, these variables will adjust as well; for example, if rates of return fall, people will decide to work a little bit longer to offset the adverse effects on their pensions. A funded system does away with collective decision-making and relies on individual decisions and market mechanisms to absorb demographic shocks. While the political conflict is avoided, it is not clear that the resulting allocation of the burden between generations is “fair.” Small cohorts, for example, will benefit from both higher wages (because they will work with the capital accumulated by previous, more numerous generations) and a higher return on capital (because their savings will be invested in productive assets that will be operated by the more numerous, subsequent generations), so they will be disproportionately better off relative to other generations. Under a pay-as-you-go system, they will command less political influence, and have to pay the pensions of the larger previous generation. Therefore, they will have to contribute more when active (for both reasons) and perhaps even get lower pensions when old (because of reduced political
9780230_547018_14_cha12.indd 414
3/19/2009 8:35:02 PM
Pensions and Children
415
Table 12.3 The relative performance of a pension fund invested in stocks, relative to a pension fund invested in bonds Year of retirement
Ratio, fund A/fund B
1985 1989 1994 1999 2004
0.68 0.97 1.02 2.43 1.83
Source: EEAG.
influence). These effects run against the market effect of higher wages and higher return to capital, and may contribute to even out the distribution of the gains and losses from demographic fluctuations across generations. 12.4.4 Where to invest contributions to pension funds Another important issue, when one considers moving to a funded system, is the assets in which the contributions should be invested. What should the composition of the pension funds portfolios be? Given that people contribute during their working time, the returns to that portfolio should be negatively correlated with the labour market risk of contributors. So it is not a very good idea to have an employer-based pension scheme where a large fraction of the wealth is invested in the firm’s own shares. Such employee stock ownership may perhaps provide good work incentives – as a good performance of the firms would raise its workers pension wealth (although there are free-riding issues to take into account). But it has very poor insurance properties, since it exposes workers to the risk of experiencing both job loss and a capital loss on their pension wealth should their employer encounter trouble. Ideally, pension wealth should be invested in assets whose return goes up when business conditions deteriorate in the contributor’s industry or occupation. Unfortunately, as business conditions tend to move together in all sectors, it is not easy to find such assets. In principle, one should also relate the optimal portfolio composition to a worker’s age, skills, occupation and industry, using standard finance tools. At a minimum, a well-diversified portfolio with little or no assets in the industry where the person is working is advisable; the portfolio should be readjusted when the person’s job or labour market status changes. An important but tricky question is whether wealth should be invested in equities or bonds. It is often argued that investment in equities yields a much higher return than bonds. However, they are much more volatile, and can yield a lower return for long periods if there is a persistent bear market. The following table compares the performance of two pension funds: one (fund A) is fully invested in a portfolio indexed on the Dow-Jones Industrial
9780230_547018_14_cha12.indd 415
3/19/2009 8:35:02 PM
416 Designing the European Model
Box 12.3 Pensions reforms as a way to circumvent inefficiencies coming from taxes and regulations By adopting a fully funded system, an economy can raise its savings rate, therefore building a higher capital stock. One may argue that, in a closed economy, a higher capital stock is a potential source of net gains for the economy as a whole because the marginal return on (real) capital is typically higher than the market rate of return, earned on individual savings. Thus, when one uses the latter to calculate the present discounted value of current and future output flows, adopting a fully funded system clearly raises domestic wealth that is it drives up the current value of current and future output. The problem with this argument is that it ignores the reason for the wedge between the marginal return on capital and the market rate. If this is due to different tax rates on individuals and corporations, there is a simpler and more direct way to achieve an equivalent increase in domestic welfare consisting of a tax reform that eliminates the source of inefficiency. It should be noted here that the difference in rates of return in the above argument should not be confused with compensation differences in risk properties of different assets. Clearly, there is no room for welfare improvement following from these price differences. Appropriately adjusting the discount factor to calculate the present value of output makes clear that no gain in domestic wealth can be reaped by exploiting them. A different efficiency-related argument stresses that the social security contribution in a pay-as-you-go system introduces a wedge between net wages and labour costs, thus creating distortions in the labour market. The above argument would of course apply in full force if social security contributions were totally de-linked from benefits. If pensions were universally granted to old people, independently of past contributions, any payment into the social security system would indeed be a tax distorting the labour/leisure choice. Most importantly, the incentive to evade would be very strong: systems with these features are a powerful reason for a thriving informal sector in the economy. It should be noted that a fully funded system may not be immune to the problem if it is run on a collective basis with redistribution goals that weaken the link between retirement saving and pension payments at the individual level. In practice, however, most systems link between benefits and contributions. Clearly, the stricter the link is, the lower the distortionary effects of social security payments are. This is because a strict link would induce workers and employers to correctly consider social security as part of the compensation of labour – although deferred in time. Some tax distortions are inherent in social security systems: first, people are forced to save rather than spend their income as they wish, which reduces the value of working; second, the rate of return on social security contributions in pay-as-you-go systems is below the market rate. Yet the magnitude of such distortions should not be exaggerated, as if pensions were not linked to contributions at all. We return on this issue in Box 12.4 below.
9780230_547018_14_cha12.indd 416
3/19/2009 8:35:02 PM
Pensions and Children
417
Average; the other (fund B) is invested in a safe asset yielding an annual real return of two per cent. Table 12.3 reports the ratio of real total pension wealth between fund A and fund B, at the age of retirement, for an individual who invests 20 per cent of his yearly income during 40 years.15 While fund A would have left cohorts retiring in 2004 and 1999 much richer than fund B, it barely makes a difference for those retiring in 1989 (who lose a bit relative to fund B) and 1994 (who gain a bit), while those retiring in 1985 are much worse off than if they had invested in bonds. A key reason is that they suffered a lot from the bear market of 1971–73, during which the Dow-Jones lost almost half its value. Another lesson from the exercise is that fund A generates huge inequalities among cohorts: the pension wealth of workers retiring in 1999 is more than twice the pension wealth of workers retiring in 1994. From this exercise we draw three conclusions. First, it is unwise to use realised stock returns over, say, the last ten years, to evaluate the performance of a funded system in the future. This argument is reinforced if one further notes that funded systems are likely to generate a massive supply of savings to the market, and thus depress rates of return. Second, one should therefore make sure that a significant fraction of pension portfolios is invested in safe assets; otherwise, there is the risk of massive social unrest should the market turn “bearish” for a number of years. Third, one should consider stimulating the development of a market for corporate bonds to absorb the demand for safe, productive assets that would result from the rise of private, funded pension schemes.16
12.5
Possible solutions
We now discuss a number of solutions that are typically proposed to fix the pension problem. A pension reform should not only make the system viable in the long run by correcting financial imbalances but also aim at designing it in the most efficient way. That is, a system may be inefficient even if it is not in financial trouble, and the financial crisis is an opportunity to deal with such inefficiencies. 12.5.1
Raising contributions
The most straightforward solution, from a pure financial perspective, is an increase in the level of contributions, computed to maintain the pensioners’ living standards unchanged, relative to GDP. This possibility raises a number of questions. First, is it fair? Raising contributions puts the burden of adjustment on some generations but not on others: current generations of retirees will not participate at all in the adjustment effort. The generations of workers who will have to provide for the baby boomers’ pensions will suffer the larger loss. Subsequent generations who will provide for cohorts of retirees less
9780230_547018_14_cha12.indd 417
3/19/2009 8:35:02 PM
418
Designing the European Model
numerous than the baby boomers will suffer a smaller loss, but still contribute more than current retirees. In short, this option may imply an arbitrary distribution of the adjustment burden across generations. Second, an increase in contributions would take place in the context of already high tax rates that discourage employment, investment and innovation. Increasing payroll taxes further may have severe distortionary effects on the economy. In fact, some studies (Laroque 2004) find that for some categories of workers, one is close to the top of the “Laffer Curve,” meaning that taxing these workers further would actually reduce tax receipts. But, as discussed in the box, the severity of this problem crucially depends on the design of actuarial pay-as-you-go-systems. Distortions can be reduced by designing systems in which contributions are effectively deferred wage payments, accumulated at an internal rate of return that is not too far from the market rate of return. A possible argument in favour of raising contributions is that the cost for workers associated with an increasing old-age dependency ratio may be partly offset by the fact that fewer children may mean a lower burden related to caring and education, i.e. by a lower youth dependency ratio. However, the extent to which the cost of raising children falls with the lower number of children is unclear, because of raising schooling levels and education costs. Moreover, the above ignores differences between those who have children and those who do not, an issue we will discuss extensively below. To summarise: we believe that there is some room for increasing contributions. The adverse distortionary effects of higher contribution rates in the labour and financial markets can be contained by making social security systems more actuarially fair. But in light of our previous consideration about the fiscal dimension of pension reform, governments should be prevented from undoing the beneficial effects of such a policy by raising fiscal deficits: increases in contributions should be matched by equivalent savings in public budgets. 12.5.2
Lowering pension benefits
An alternative route is to lower pension benefits, which has the merit of not increasing fiscal distortions. Again, if done blindly, holding contribution rates constant, cutting pension benefits of present and future retirees has controversial distributional consequences: it is fairer than increasing contributions, as the burden is spread among all current and future generations; however, it also imposes an excess burden on abnormally large cohorts – as opposed to increasing contributions that impose an excess burden on abnormally small cohorts. Also, in order to gauge the desirability of such an option, one must look at the living standards of retirees. In the EU as a whole, the retirees’ median income is about 83 per cent of the median income of people aged less than 65. One may consider that as a rather high number: while retirees spend more on health care, they typically do not
9780230_547018_14_cha12.indd 418
3/19/2009 8:35:02 PM
Pensions and Children
419
bear all the costs of a family (childrearing, housing, etc.). Furthermore, from a “fairness” point of view, retirees do not bear the disutility of work. One argument in favour of reducing pension benefits is that declining fertility rates were not appropriately accounted for in the original design of the systems. To the extent that a lower fertility rate can be interpreted as insufficient investment in human capital, current generations who have not borne the costs of raising children, on average, should be entitled to a lower return on their social security saving. But by the same token there is no reason to penalise individuals in these generations who do have enough children – an argument that underlies the proposal to differentiate pensions by the number of children discussed below. Overall, we think that there is some room for manoeuvre in reducing pension benefits, especially in light of the fact that the required reduction in the purchasing power of pensions is only relative: their absolute purchasing power can still grow. 12.5.3
Raising the retirement age
Another natural option is to raise the retirement age. This option has been largely neglected until the most recent reforms. Average retirement age has actually been falling rather than rising. Yet, from an arithmetical point of view, such a reform makes a lot of sense. Average life expectancy at 65 was about 12 years in 1960 (for men) and should be about 20 years in 2040. Thus if the average length of time spent in retirement were held constant, the retirement age should increase to 73 in 2040. If one targets the share of a lifetime (from age 20) spent in retirement rather than its absolute length, one still reaches the conclusion that the age of retirement should increase to 71½. Therefore, if one had indexed the retirement age on life expectancy one way or another, the “pension problem” would simply be non-existent. A recent British report calculates that from a balanced budget perspective (using again men as a benchmark), one should increase the age of retirement to 69 by 2050.17 Instead of increasing, the actual retirement age has fallen below 60 in many countries, because of generous pre-retirement policies. For example, the employment rate of the 55–64 age group does not exceed 50 per cent in most European countries, being as low as 25–28 per cent in Austria and Belgium and 34–38 per cent in France and Germany. Therefore, as far as the arithmetic is concerned, increasing longevity is not necessarily a problem for pension systems to the extent that the retirement age correspondingly rises. But this does not imply that a substantial rise in retirement age is necessarily desirable, or that it is the only policy that should be pursued. The extent to which this margin should be used clearly depends on whether people are willing and/or able to work the required extra years. It may be that, faced with a choice, people would prefer lower consumption when retired and/or during their work life rather than working longer.
9780230_547018_14_cha12.indd 419
3/19/2009 8:35:03 PM
420
Designing the European Model
After all, the secular trend of reduced working time in Europe – engineered by public policies such as pre-retirement and the reduced workweek – must have to do with the preferences of the political majority. It remains to be seen, though, whether these preferences reflect a genuine taste for more leisure rather than misperception of its costs. Moreover, incentive for shorter working time may also be explained by high tax wedges, rather than preferences for leisure, as discussed in Chapter 5. As far as feasibility is concerned, one may ask whether people are in good enough condition, say between 60 and 70, to perform jobs. Here the answer is probably yes. First, most jobs are less physically demanding than 50 years ago; second, the increase in life expectancy has been associated with a reduction in morbidity rates: one is in better health at 68 now than 50 years ago. Finally, since societies have chosen to adapt the workplace to make it more adequate for the disabled, there is no reason why similar steps – likely to be less costly than for the disabled – could not be taken for older workers as well. That being said, by how much the retirement age should increase is less clear, as it depends on preferences. In particular, people are impatient and prefer to enjoy leisure earlier rather than later. This puts limits on the desirability of increasing the retirement age. As preferences and individual situations in that respect are heterogeneous, we advocate pension systems where people can trade later retirement ages for higher pensions at an actuarially fair rate, making people free to choose their retirement age according to their individual preferences. Nonetheless, given that the pension finance problem comes from a longer lifetime, increasing the age of retirement should naturally be part of any solution. More generally, a higher rate of participation in the workforce would clearly help. As is well known, labour force participation is lower in Europe than in other advanced regions.18 Reducing distortions in labour markets, which may include making pension rights actuarially linked to contributions as we recommend in this report, would therefore be an important step in tackling the pension finance problem. 12.5.4 Growth The crisis of pension systems due to a rising old-age dependency ratio is particularly apparent if one wants pensions to grow in line with other incomes. This is apparent when pension benefits are indexed to wages: improvements in the quality of the workforce that raise wages also raise pension claims. This is why reducing or eliminating wage indexation of pensions may ease transitional problems: to the extent that pensions of a given generation of retirees are indexed to their own past wages, raising the human capital of the young will raise the tax base but not the claims of these retirees (only the future claims of these young workers increase). While welfare across cohorts of retirees will rise with the standard of living and therefore the
9780230_547018_14_cha12.indd 420
3/19/2009 8:35:03 PM
Pensions and Children
421
productivity of individuals during their own working lifetime, increasing growth will tend to reduce the pension burden relative to GDP. Observe that in principle a pension system could be designed to guarantee only a constant pension in real terms – perhaps according to an “assistance” philosophy. In this case, productivity growth will induce a downward trend in the value of pensions relative to GDP, making it quite easy to cope with the increasing number of retirees. However, such an option has not been seriously debated, mainly because it would entail a widening inequality between pensioners and workers, with adverse social and political consequences despite the fact that the pensioner’s living standards would not fall over time. To the extent that policies that promote growth increase the amount of resources available for redistribution, they will also help alleviate the pension problem. Thus one may consider investing in a better educated, high-skilled workforce – endowed with a higher level and better quality of physical capital – so that higher productivity would compensate for the loss in the size of the workforce. However, these policies are no free lunch and involve a reduction in consumption for current generations, just like any other increase in savings to provide for future retirement. 12.5.5 Immigration A policy that is often advocated in the context of debates about pension crisis is to open up immigration, to make up for demographic and financial deficits with new workers contributing to the system. We have a number of reservations about this solution. First, it is strange to increase migration for the sole purpose of financing pensions: one could in principle pursue the same goal by enrolling foreign workers into national pay-as-you go schemes without them actually living in the country. In fact, as argued above, this would not be different from selling the national debt to foreigners. Of course, one may want to have a younger population and/or a stronger production basis for other reasons. But as far as financing pensions is concerned, the geographical origin of contributions is irrelevant. Given the social strains created by excess immigration and the controversies it generates, the pension finance argument does not seem very appealing. Second, immigration is likely to have only transitory benefits: as immigrants age and adjust their fertility behaviour to that of natives, a pension finance problem pops up again. Only if immigrants systematically die earlier or make more children than natives would immigration permanently solve the pension problem, and each of these hypotheses would mean that they fail to integrate into society, which creates problems of its own. Finally, whether the immigrants’ contribution to public finances is positive overall is unclear and depends on their skill level. While the inflow of young legal immigrants raises overall contributions to the pensions system, to the extent that they are unskilled they may also be net recipients from
9780230_547018_14_cha12.indd 421
3/19/2009 8:35:03 PM
422
Designing the European Model
other welfare programmes such as unemployment or child benefit. The net effect depends on the skill composition of immigration and on the structure of the welfare state. To be sure, taking in skilled immigrants would clearly be a positive contribution to public finance. However, one does not select the skill level of migrants by decree. It depends on the incentives that they face when choosing a destination, incentives that reflect the levels of taxation and social protection, as well as immigration policy. Countries with a high level of redistribution in favour of the unskilled are more likely to attract unskilled immigrants. To conclude, immigration is an important dimension of policy which should be discussed on its own merits, rather than being advocated as a mere fix for the pension system’s financial problems. 12.5.6 Increasing the margin of individual choice There are some good reasons to have a state-sponsored pension system, most prominently the fact that a fraction of the population does not save adequately. However, this argument does not justify uniform pension
Box 12.4
An actuarial system reduces the tax wedge
To what extent are social security contributions as distortionary as a tax on labour income? To provide an intuition, consider the extreme example of a hypothetical reform from a system with no link between contributions and pension, to a system of individual accounts such as the German system set up by Adenauer in 1957 or the new Swedish one. Keeping the average contribution rate fixed at 20 percent of the wage, we can easily calculate the difference between tax wedges in the two systems. Consumption (including bequests) during retirement will be equal to the pension benefits plus any private saving accumulated over the lifespan, and capitalised at the market rate r. In an extreme non-actuarial system, pension payments are independent of contributions. Hence pension contributions drive a tax wedge distorting the labour/ leisure choice as high as 20 percent. In an actuarial system, pension benefits are linked to contributions via the internal rate. The contribution is perceived as deferred wage income, capitalized at a rate of return that is, however, lower than the market interest rate. Hence the tax wedge in pension contributions is less than 20 percent: it is proportional to the difference between the market rate of return and the internal rate of return of the social security system. For instance, if the market rate is four percent, and the internal rate of return is two, the tax wedge would only be two percent of the contribution rate. Clearly pension reforms strengthening the link between benefits and contributions in a credible way can lead to a marked reduction in labour market distortions. This could result in higher working hours and a higher participation rate (depending on the strength of the substitution effect from higher wages). As a note of caution, we should point out however that the overall effect of a reform should be assessed by considering the entire structure of taxation.
9780230_547018_14_cha12.indd 422
3/19/2009 8:35:03 PM
Pensions and Children
423
systems. Therefore, one of our key recommendations is to increase the margins of individual choices. Pensions should be linked to contributions in an actuarially fair way. This recommendation has a number of implications: People should be allowed to retire at the age they wish, provided they get correspondingly higher or lower pensions, at an actuarially fair rate. Similarly, they may choose to contribute more or less, provided they contribute more than a pre-defined floor. Making pensions more actuarial also reduces the distortions induced by the pension system on the labour supply, as illustrated in Box 12.4 The example of Sweden and Italy, moving towards notionally defined contributions, can be interpreted as a step in this direction: any payment to the social security system over a worker’s life would be contributing to determine his/her pension, reducing the incentive to evade. An actuarial system would of course limit the role of social security in redistributing resources within the same generation. But this is not necessarily bad news, in light of the evidence on current systems. In systems that are not actuarially fair it is often the case that the within-cohort redistribution is highly regressive. For instance, when pension benefits are calculated based only on the wage income in the last few years of employment (or based on the best wage over a few years), social security systems tend to favour high-wage workers who usually have steeper age profiles of wage income than low-wage workers. Allowing for this margin of manoeuvre may in principle generate problems for the pay-as-you-go system if individual decisions do not square with the system’s overall commitments. To analyse these problems, we need to take into account the fact that one euro invested in the pay-as-you-go system does not yield the same return as one euro invested in capital markets. Suppose that an individual decides to retire early, hence contributing less to the pay-as-you-go system. If his/her pension benefits are reduced calculating the current market value of his/her missing contributions, there is no negative financial implication on social security. In fact, the government can make up for the shortfalls in revenues by borrowing from financial markets a sum that, in present value terms, is identical to the reduction in individual pension benefits. In other words, the value of incremental future interest payments on the additional debt is exactly equal to the reduction in the individual’s pension claims. By the same token, the financial balance of a well-designed pension system would not be affected if individuals who choose to postpone their retirement beyond the statutory age would be entitled to receive pension benefits while still working (say, above 65) without paying further contributions to the system. Therefore, as long as changes in benefits are calculated at the market rates, any shortfall in contributions (if any) that raise the stock of public debt correspond to an equivalent decrease in future pension benefits: introducing that margin of manoeuvre does not jeopardise the pay-as-you-go system. Note that allowing for such a margin is equivalent to adding a small implicit
9780230_547018_14_cha12.indd 423
3/19/2009 8:35:03 PM
424
Designing the European Model
fully-funded component to the pay-as-you-go system that can be activated based on individual preferences around the retirement age. Nonetheless, while we advocate the possibility for individuals to contribute less, and earn a correspondingly lower pension, this should only be allowed for those whose pension benefits are high enough relative to the social assistance level. Simple rules should deter individuals from opting out of the system and then claim assistance benefits because of insufficient retirement savings. 12.5.7 Reforming the labour market and avoiding pre-retirement The pension crisis in Europe is aggravated by ill-functioning labour markets. On the one hand, labour market rigidities lower employment rates, which reduces the tax base for contributions. On the other hand, rigidities increase the number of claimants for various welfare programmes. Especially absurd, in this light, is the practice of pre-retirement, which artificially depresses retirement age, thus contributing to the pension crisis. While dismissals of prime-aged workers are extremely costly in many European countries, they are almost subsidised for workers near retirement. It is not clear what the justification of such policies is (unemployability or “making room for the young”?). In practice, dismissing older workers with a generous, publicly financed pre-retirement package is a way for firms to restore some flexibility in managing their workforce, in the face of stringent employment protection legislation. Clearly, a comprehensive labour market reform would help in many ways. First, the burden of flexibility would be more evenly distributed, and firms would be less inclined to use pre-retirement. Second, employment would go up, and so would the tax base. But, even if such reform does not take place, it would be very useful to eliminate the pre-retirement trap. For example, as has been proposed by Sinn (2003), one could make early retirement schemes less attractive by reducing pension benefits before retirement age in an actuarially fair way (in line of our proposal), while allowing retirees to cumulate their pension with a secondary job. 12.5.8 Introduce private pension funds As discussed at length above, the current pension crisis is rooted in demographic problems. The extent of these problems will be magnified in the next decades, when the baby boomers will reach retirement age. This cohort of people is large but has much fewer children than the previous generation. The pay-as-you-go pension system is based on human capital investment: if too little human capital has been formed, it is unable to provide enough pensions. Clearly, pensions come from human capital and/or real capital. A possible reaction to the lack of human capital is therefore the formation of more real capital. To the extent that human capital is missing, more real capital is
9780230_547018_14_cha12.indd 424
3/19/2009 8:35:03 PM
Pensions and Children
425
needed to fill the gap. This is the rationale for adding a funded pillar to the pay-as-you go pension system. The argument is not that the funded system is inherently more efficient because it offers a higher rate of return. We explained the fallacy of that argument in Box 12.2. The rationale is crisis management by increasing the real capital endowment of the society. Specifically a fully-funded component of social security is needed to induce those baby boomers who are now around 40 and will not (or choose not to) have children to substantially increase their savings for the two decades that remain until retirement. There still is enough time to save. Private pension funds are an effective way of decentralising the additional savings efforts required. The state should encourage private-sector pension schemes by certifying that these are sound, as well as advise people to invest in assets with an appropriate risk structure. This may involve a deterrent from employer-based pension schemes and a stimulus to pension schemes with a broad, diversified portfolio. We have also seen that despite the higher average return on stocks, pension funds fully invested in them would have a return that fluctuates widely across cohorts. Therefore, there should be a minimum fraction in private pension schemes that have to be invested in safe assets. Simple legal rules should also be designed to supervise and regulate fund management so as to minimise budget risks and social costs associated with financial instability and moral hazard in financial markets. These rules should also provide strong incentives to contain the managing costs of pension funds. Especially in the initial phase of a reform, these costs may levitate in a privatised system because of aggressive advertising by an excessive number of providers. Contributions to the fund should be mandatory in order to prevent free riding on the generosity of society. As long as there are other safeguards against poverty in old age, such as a general system of social aid, private incentives to save may not be sufficient: when deciding how much to save, low-income individuals will know that social aid will be lower, the larger the funds accumulated. If households are guaranteed a minimum level of income via public transfers, participation in funded pension plans should be mandatory, to prevent possibly large distortions on savings behaviour. 12.5.9 Differentiating pensions by the number of children Above we have advocated the introduction of mandatory pension funds in addition to the current pay-as-you-go system to increase savings by the baby boomers. Forcing people to save more may be seen as a burden: many may consider such a burden unfair, arguing that they already paid for their old-age pension by contributing to the pension system. However, this argument is quite weak, once it is taken into account that, on average, such a
9780230_547018_14_cha12.indd 425
3/19/2009 8:35:03 PM
426 Designing the European Model
generation has substantially reduced, or avoided altogether, the burden of raising children. Since many individuals have chosen not to have children, i.e. not to invest in human capital, it is fair to require them to sustain a compensatory burden in terms of additional savings towards investment in real capital. While the rationale of the above argument applies for the generation of baby boomers as a whole, a well-designed policy reform inspired by it must take into account differences within such a generation. After all there are still families with children. These then run the risk of bearing three different burdens. They nourish the generation of their parents with their payas-you-go contributions; they have borne the cost of raising their children, thus safeguarding future pay-as-you-go systems; and they may face the burden of additional indifferentiated mandatory saving plans, as pension systems are reformed along the lines discussed in our chapter. To avoid a triple burden for workers with children, pension reforms could differentiate according to the number of children. A simple way to implement this is to make savings plans mandatory only for childless workers. To be concrete: the contribution rates in the current European pension systems could be frozen despite the adverse demographic development. Other things equal, then, constant contribution rates, will substantially reduce replacement rates: as the old-age dependency ratio doubles in thirty years’ time, the replacement rates will be cut in half unless other measures to alleviate the problem are taken. To compensate for the decline in replacement rates two new pillars of the pension system could be introduced. One pillar is the mandatory savings plan as described above. Another is a supplementary pay-as-you-go pension for parents financed with a general income tax – a “child pension.”19 According to this proposal, every person entering the labour force participates in the mandatory savings plan. As soon as a child is born, the savings obligation is reduced by some fraction, while the same fraction of accumulated savings is paid out to the individual. The same happens when a second child is born, and so on, until some target number of children is reached. Such a “child pension” compensates for the missing participation of parents in the funded pension scheme. The size of the child pension should be designed such that it ensures today’s replacement rate when all pension elements are taken together: the pension from the existing pay-as-you-go system, the funded pension and the child pension. Basically this plan means differentiating pay-as-you-go pensions by the number of children and compensating the pension gap for the childless with mandatory private savings. The plan is fair because it reduces the extent to which the fruits of human capital investment are socialised by the
9780230_547018_14_cha12.indd 426
3/19/2009 8:35:03 PM
Pensions and Children
427
public pension system. The plan may even contribute to revitalise the desire to have children. Nonetheless, it should be made clear that increasing the fertility rates will not really help solve the imminent pension crisis in the 2030s. For this the policy would come too late. However, in the longer run, the pension system and the European society as such will only be able to function better if the population size can be stabilised. Our proposal may help achieve that goal. Other measures that may also help achieve the same goal are special child benefits in the tax system or government-financed child-care facilities. These measures implicitly follow the logic of double interventions: given that the fruits of human capital investment are socialised with the pension system, the investment outlay is socialised too.
12.6 Conclusions In this chapter, we have assessed the demographic challenge to the European pension systems that in most countries are based on the pay-as-you-go principle. Under current conditions, most pay-as-you-go pension systems in Europe are not sustainable: the old-age dependency ratios are forecasted to grow from the current 0.2–0.3 range to as high as 0.4–0.68 pensioners per worker in 2050, which would eventually require a very large increase in tax rates, and/or a reduction in pensions. Reform is required and it should aim not only at fixing the budget problem but also at designing a more efficient pension system. A general slowdown in the growth of living standards associated with ageing is inevitable. Pension reform that entails a move to a partially funded system (in which workers make savings in personal accounts toward their future pensions) will not avoid the slowdown and cannot benefit all generations. However, such a move may help stimulate national savings and smooth the pension burden across generations. The pension crisis results from a lack of human capital. Partial funding means filling the human capital gap with real capital. It thus helps mitigate the provision crisis to be expected when the baby boomers receive their pensions. Funded pension components may also increase the room for individual flexibility by allowing people to choose their pension level and retirement age at an actuarially fair rate, and thus alleviate political conflicts associated with ageing. One should ensure that private pension funds have an appropriate risk structure. That includes limiting exposure to stock market fluctuations and minimising the correlation between the financial risk of pension wealth and labour market risk. Thus portfolios of pension funds should
9780230_547018_14_cha12.indd 427
3/19/2009 8:35:03 PM
428
Designing the European Model
be adequately diversified, with a critical mass of risk-free assets and a very limited exposure to assets in the firm and sector in which the worker is employed. Simple legal rules should be designed to supervise and regulate fund management so as to minimise budget risks and social costs associated with financial instability and moral hazard in financial markets. These rules should also provide strong incentives to contain the managing costs of pension funds. Especially in the initial phase of a reform, these costs may rise in a privatised system because of aggressive advertising by an excessive number of providers. The introduction of an individually based, privately managed, funded pillar of the pension system would allow for a great deal of individual flexibility, provided it satisfies these requirements. It would be a good idea for those European countries that have not already done so to complement the existing pay-as-you-go system with such a pillar. A number of other margins of manoeuvre also exist that would contribute to fixing the problem of sustainability of the pension system. To the extent that part of ageing is due to an increase in life expectancy and that people are healthier, it is perfectly natural to raise the retirement age, which has trended downwards for many years in most countries. Pre-retirement schemes that are meant to artificially reduce unemployment statistics, while increasing the burden on pensions, should be avoided altogether. Structural reform in the labour market, although desirable in its own right, will also have a positive effect on pension finance by increasing employment, thus increasing the tax base for contributions. The fiscal system could be amended so as to reduce its distortionary impact on people’s decisions to have children. Specifically, when deciding on the number of children, people ignore the fiscal benefits brought by children to society in the form of contributions to pensions and may therefore have fewer children than is socially desirable. One could envisage reforms to address this issue. A partial indexation of pay-as-you-go pension claims on the number of children is one possibility. Additional selffinanced mandatory funded pensions for those who have no or only few children could then supplement the pay-as-you-go pension for those with no or only few children. People who do not raise children have, on average, more funds to save for their old-age pension. Alternatively, personal income taxation can be differentiated according to the number of children and systems of child allowance be used to provide stronger incentives towards having children. However, there are reasons to believe these alternative measures to be less effective in addressing the distortions that undermine the viability of social security system, contributing to the substantial drop in fertility experienced in our countries. Much of our future welfare, and the welfare of our children and offspring, is at stake with the current pension reforms.
9780230_547018_14_cha12.indd 428
3/19/2009 8:35:03 PM
Pensions and Children
429
Appendix: Age pyramids in selected countries AUSTRIA
BELGIUM
2050
2050
100 Male
Female
80
60+ 0–59
60
Age
Age
80
100
40
20
20
5
0 Percentage
5
0 10
10
60+ 0–59
5
2025
80
Female 60+ 0–59
Age
Age
Male
40
20
20 5
0 Percentage 2004
5
0 10
10
100 Female
80
60+ 0–59
Age
Age
Male
40
20
20 5
0 Percentage
5
0 10
10
60+ 0–59
5
5
10
Female 60+ 0–59
5
0 Percentage
5
10
1950 100
Male
Female
80
60+ 0–59
60
Age
Age
0 Percentage 2004
Male
1950 100
40
20
20 5
0 Percentage
5
10
Male
Female 60+ 0–59
60
40
0 10
Female
60
40
80
10
100
60
0 10
Male
60
40
80
5
100
60
0 10
0 Percentage 2025
100 80
Female
60
40
0 10
Male
0 10
5
0 Percentage
5
10
Figure 12.A1 EU15 Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 429
3/19/2009 8:35:04 PM
430 Designing the European Model DENMARK
FINLAND
2050
2050
100 Male
Female
80
60+ 0–59
60
Age
Age
80
100
40
20
20
5
0 Percentage
5
0 10
10
60+ 0–59
5
2025
80
Female 60+ 0–59
Age
Age
Male
40
20
20 5
0 Percentage 2004
5
0 10
10
100
80
60+ 0–59
Age
Age
Female
40
20
20 5
0 Percentage
5
0 10
10
60+ 0–59
5
5
10
Female 60+ 0–59
5
0 Percentage
5
10
1950 100
Male
Female
80
60+ 0–59
60
Age
Age
0 Percentage 2004
Male
1950 100
40
20
20 5
0 Percentage
5
10
Male
Female 60+ 0–59
60
40
0 10
Female
60
40
80
10
100 Male
60
0 10
Male
60
40
80
5
100
60
0 10
0 Percentage 2025
100 80
Female
60
40
0 10
Male
0 10
5
0 Percentage
5
10
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 430
3/19/2009 8:35:04 PM
Pensions and Children FRANCE
GERMANY
2050
2050
100
100 Male
Female
80
60+ 0–59
60
Age
Age
80
40
20
20
5
0 Percentage
5
0 10
10
Male
60+ 0–59
5
2025
80
Female 60+ 0–59
Age
Age
Male
40
20
20 5
0 Percentage 2004
5
0 10
10
100
80
60+ 0–59
Age
Age
Female
40
20
20 5
0 Percentage
5
0 10
10
60+ 0–59
5
5
10
Female 60+ 0–59
5
0 Percentage
5
10
1950 100
Male
Female
80
60+ 0–59
60
Age
Age
0 Percentage 2004
Male
1950 100
40
20
20 5
0 Percentage
5
10
Male
Female 60+ 0–59
60
40
0 10
Female
60
40
80
10
100 Male
60
0 10
Male
60
40
80
5
100
60
0 10
0 Percentage 2025
100 80
Female
60
40
0 10
431
0 10
5
0 Percentage
5
10
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 431
3/19/2009 8:35:04 PM
432 Designing the European Model
GREECE
IRELAND
2050
2050
100 Male
Female
80
60+ 0–59
60
Age
Age
80
100
40
20
20
5
0 Percentage
5
0 10
10
60+ 0–59
5
2025
80
Female 60+ 0–59
Age
Age
Male
40
20
20 5
0 Percentage 2004
5
0 10
10
100
80
60+ 0–59
Age
Age
Female
40
20
20 5
0 Percentage
5
0 10
10
60+ 0–59
5
5
10
Female 60+ 0–59
5
0 Percentage
5
10
1950 100
Male
Female
80
60+ 0–59
60
Age
Age
0 Percentage 2004
Male
1950 100
40
20
20 5
0 Percentage
5
10
Male
Female 60+ 0–59
60
40
0 10
Female
60
40
80
10
100 Male
60
0 10
Male
60
40
80
5
100
60
0 10
0 Percentage 2025
100 80
Female
60
40
0 10
Male
0 10
5
0 Percentage
5
10
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 432
3/19/2009 8:35:04 PM
Pensions and Children ITALY
LUXEMBOURG 2050
2050 100
100 Male
60
Female
80
60+ 0–59
60
Age
Age
80
40
40
20
20
0 10
5
0 Percentage
5
0 10
10
Male
5
80
Female 60+ 0–59
Age
Age
Male
20
20 5
0 Percentage 2004
5
0 10
10
100
80
60+ 0–59
Age
Age
Female
40
20
20 5
0 Percentage 1950
5
0 10
10
100
60+ 0–59
5
0 Percentage 2004
5
10
Male
Female 60+ 0–59
5
0 Percentage 1950
5
10
100 Male
Female
80
60+ 0–59
60
Age
Age
Female
60
40
40
20
20 5
0 Percentage
5
10
Male
Female 60+ 0–59
60
40
0 10
Male
100 Male
60
80
10
60 40
0 10
5
2025
40
80
0 Percentage
100
60
0 10
Female 60+ 0–59
2025 100 80
433
0 10
5
0 Percentage
5
10
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 433
3/19/2009 8:35:05 PM
434 Designing the European Model NETHERLANDS
PORTUGAL
2050
2050
100 Male
Female
80
60+ 0–59
60
Age
Age
80
100
40
20
20
5
0 Percentage
5
0 10
10
60+ 0–59
5
2025
80
Female 60+ 0–59
Age
Age
Male
40
20
20 5
0 Percentage 2004
5
0 10
10
100
80
60+ 0–59
Age
Age
Female
40
20
20 5
0 Percentage
5
0 10
10
60+ 0–59
5
5
10
Female 60+ 0–59
5
0 Percentage
5
10
1950 100
Male
Female
80
60+ 0–59
60
Age
Age
0 Percentage 2004
Male
1950 100
40
20
20 5
0 Percentage
5
10
Male
Female 60+ 0–59
60
40
0 10
Female
60
40
80
10
100 Male
60
0 10
Male
60
40
80
5
100
60
0 10
0 Percentage 2025
100 80
Female
60
40
0 10
Male
0 10
5
0 Percentage
5
10
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 434
3/19/2009 8:35:05 PM
Pensions and Children SPAIN
SWEDEN
2050
2050
100
100 Male
Female
80
60+ 0–59
60
Age
Age
80
40
20
20
5
0 Percentage
5
Male
60+ 0–59
0 10
10
5
2025
80
Female 60+ 0–59
Age
Age
Male
40
20
20 5
0 Percentage 2004
5
80
60+ 0–59
Age
Age
Female
5
40
20
20 5
0 Percentage
5
5
Male
10
Female 60+ 0–59
0 10
10
5
1950
0 Percentage
5
10
1950
100
100 Male
Female
80
60+ 0–59
60
Age
Age
0 Percentage 2004
60
40
40
20
20 5
0 Percentage
5
10
Male
Female 60+ 0–59
60
40
0 10
Female
100 Male
60
80
10
60+ 0–59
0 10
10
100
0 10
Male
60
40
80
5
100
60
0 10
0 Percentage 2025
100 80
Female
60
40
0 10
435
0 10
5
0 Percentage
5
10
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 435
3/19/2009 8:35:05 PM
436 Designing the European Model UNITED KINGDOM 2050 100
Age
80
Male
Female 60+ 0–59
60 40 20 0 10
5
0 Percentage
5
10
2025 100
Age
80
Male
Female 60+ 0–59
60 40 20 0 10
5
0 Percentage 2004
5
10
100
Age
80
Male
Female 60+ 0–59
60 40 20 0 10
5
0 Percentage
5
10
1950 100
Age
80
Male
Female 60+ 0–59
60 40 20 0 10
5
0 Percentage
5
10
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 436
3/19/2009 8:35:05 PM
Pensions and Children CYPRUS
CZECH REPUBLIC
2050
2050
100
100 Male
Female
80
60+ 0–59
60
Age
Age
80
40
20
20
5
0 Percentage
5
Male
60+ 0–59
0 10
10
5
2025
80
Female 60+ 0–59
Age
Age
Male
40
20
20 5
0 Percentage 2004
5
80
60+ 0–59
Age
Age
Female
5
40
20
20 5
0 Percentage
5
10
Female 60+ 0–59
5
1950
0 Percentage
5
10
1950 100
Male
Female
80
60+ 0–59
60
Age
Age
5
Male
0 10
10
100
40
20
20 5
0 Percentage
5
10
Male
Female 60+ 0–59
60
40
Figure A2
0 Percentage 2004
60
40
0 10
Female
100 Male
60
80
10
60+ 0–59
0 10
10
100
0 10
Male
60
40
80
5
100
60
0 10
0 Percentage 2025
100 80
Female
60
40
0 10
437
0 10
5
0 Percentage
5
10
New EU Member States.
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 437
3/19/2009 8:35:06 PM
438 Designing the European Model ESTONIA
HUNGARY
2050
2050
100 Male
Female
80
60+ 0–59
60
Age
Age
80
100
40
20
20
5
0 Percentage
5
60+ 0–59
0 10
10
5
2025
80
Female 60+ 0–59
Age
Age
Male
40
20
20 5
0 Percentage 2004
5
80
60+ 0–59
Age
Age
Female
5
40
20
20 5
0 Percentage
5
5
Male
10
Female 60+ 0–59
0 10
10
5
1950
0 Percentage
5
10
1950
100
100 Male
Female
80
60+ 0–59
60
Age
Age
0 Percentage 2004
60
40
40
20
20 5
0 Percentage
5
10
Male
Female 60+ 0–59
60
40
0 10
Female
100 Male
60
80
10
60+ 0–59
0 10
10
100
0 10
Male
60
40
80
5
100
60
0 10
0 Percentage 2025
100 80
Female
60
40
0 10
Male
0 10
5
0 Percentage
5
10
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 438
3/19/2009 8:35:06 PM
Pensions and Children LATVIA
LITHUANIA
2050
2050 100
100 Male
Female
80
60+ 0–59
60
Age
Age
80
40
20
20
5
0 Percentage
5
0 10
10
Male
60+ 0–59
5
2025
80
Female 60+ 0–59
Age
Age
Male
40
20
20 5
0 Percentage 2004
5
0 10
10
100
80
60+ 0–59
Age
Age
Female
40
20
20 5
0 Percentage
5
0 10
10
60+ 0–59
5
5
10
Female 60+ 0–59
5
0 Percentage
5
10
1950 100
Male
Female
80
60+ 0–59
60
Age
Age
0 Percentage 2004
Male
1950 100
40
20
20 5
0 Percentage
5
10
Male
Female 60+ 0–59
60
40
0 10
Female
60
40
80
10
100 Male
60
0 10
Male
60
40
80
5
2025
60
0 10
0 Percentage
100
100 80
Female
60
40
0 10
439
0 10
5
0 Percentage
5
10
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 439
3/19/2009 8:35:06 PM
440 Designing the European Model MALTA
POLAND 2050
2050 100
100 Male
Female
80
60+ 0–59
60
Age
Age
80
40
20
20
5
0 Percentage
5
0 10
10
60+ 0–59
5
2025
80
Female 60+ 0–59
Age
Age
Male
40
20
20 5
0 Percentage 2004
5
0 10
10
100
80
60+ 0–59
Age
Age
Female
40
20
20 5
0 Percentage
5
0 10
10
60+ 0–59
5
5
10
Female 60+ 0–59
5
0 Percentage
5
10
1950 100
Male
Female
80
60+ 0–59
60
Age
Age
0 Percentage 2004
Male
1950 100
40
20
20 5
0 Percentage
5
10
Male
Female 60+ 0–59
60
40
0 10
Female
60
40
80
10
100 Male
60
0 10
Male
60
40
80
5
100
60
0 10
0 Percentage 2025
100 80
Female
60
40
0 10
Male
0 10
5
0 Percentage
5
10
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 440
3/19/2009 8:35:06 PM
Pensions and Children SLOVAKIA
SLOVENIA 2050
2050 100
100 Male
Female
80
60+ 0–59
60
Age
Age
80
40
20
20
5
0 Percentage
5
0 10
10
Male
60+ 0–59
5
2025
80
Female 60+ 0–59
Age
Age
Male
40
20
20
5
0 Percentage 2004
5
0 10
10
100
80
60+ 0–59
Age
Age
Female
40
20
20 5
0 Percentage
5
0 10
10
60+ 0–59
5
5
10
Female 60+ 0–59
5
0 Percentage
5
10
1950 100
Male
Female
80
60+ 0–59
60
Age
Age
0 Percentage 2004
Male
1950 100
40
20
20 5
0 Percentage
5
10
Male
Female 60+ 0–59
60
40
0 10
Female
60
40
80
10
100 Male
60
0 10
Male
60
40
80
5
100
60
0 10
0 Percentage 2025
100 80
Female
60
40
0 10
441
0 10
5
0 Percentage
5
10
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 441
3/19/2009 8:35:07 PM
442 Designing the European Model UNITED STATES
JAPAN
2050
2050
100 Male
Female
80
60+ 0–59
60
Age
Age
80
100
40
20
20
5
0 Percentage
5
0 10
10
60+ 0–59
5
2025
80
Female 60+ 0–59
Age
Age
Male
40
20
20 5
0 Percentage 2004
5
0 10
10
100
80
60+ 0–59
Age
Age
Female
40
20
20
60+ 0–59
5
5
0 Percentage
5
10
10
10
Female 60+ 0–59
5
0 Percentage 1950
5
10
100 Male
80
Female 60+ 0–59
60
Age
Age
5
Male
1950 100
40
20
20
0
Male
Female 60+ 0–59
60
40
Figure A3
0 Percentage 2004
0
0
10
Female
60
40
80
10
100 Male
60
10
Male
60
40
80
5
100
60
0 10
0 Percentage 2025
100 80
Female
60
40
0 10
Male
0 5
0 Percentage
5
10
10
5
0 Percentage
5
10
Others.
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 442
3/19/2009 8:35:07 PM
Pensions and Children RUSSIA
TURKEY
2050
2050
100
100 Male
Female
80
60+ 0–59
60
Age
Age
80
40
20
20
5
0 Percentage
5
0 10
10
Male
60+ 0–59
5
2025
80
Female 60+ 0–59
Age
Age
Male
40
20
20 5
0 Percentage
5
0 10
10
Male
60+ 0–59
5
Female
80
60+ 0–59
Age
Age
Male
Male
40
20
20
60+ 0–59
0
0 5
0 Percentage
5
10
10
5
1950 100
0 Percentage 1950
5
10
100 Male
80
Female 60+ 0–59
60
Age
Age
10
Female
60
40
40
20
20
0
Male
Female 60+ 0–59
60
40
10
5
100
60
80
0 Percentage 2004
2004 100
10
10
Female
60
40
80
5
100
60
0 10
0 Percentage 2025
100 80
Female
60
40
0 10
443
0 5
0 Percentage
5
10
10
5
0 Percentage
5
10
Source: United Nations Population Division; CESifo calculations.
9780230_547018_14_cha12.indd 443
3/19/2009 8:35:07 PM
444 Designing the European Model
Notes 1. See for instance Oksanen (2003, 2004) among others. 2. A useful glossary of pension terms and a taxonomy can be found at http:// www.oecd.org/dataoecd/5/4/2496718.pdf and http://www.oecd.org/dataoecd/ 34/23/2488707.pdf, respectively. 3. The same picture emerges if one looks at the dependency ratio including young people of a non-working age (0–15), although the deterioration of this indicator appears less dramatic. 4. If the completed fertility rate is, say, 1.7 instead of 2.1, then postponement of child-bearing actually slightly reduces dependency ratios in the long run, from 0.435 to 0.42. 5. Source: Eurostat, http://www.un.org/esa/population/pubsarchive/gubhaju/ table7.htm 6. The new EU member states suffer markedly from the rapid deterioration of the demographic outlook. In some of our examples, average income per adult is 50 per cent of what it could be for a stable population structure. For Japan, our exercise suggests quasi stagnation: annual productivity growth of two per cent is not enough to compensate for population ageing in terms of average resources, although the level of income in this country is high in absolute terms to start with. A relatively better demography explains the relatively good outlook for the US. 7. A country can turn the stock of national wealth into consumption (consume its stock of capital) by selling domestic and foreign assets to non-residents (including the stock of housing) as well as by letting capital depreciate in line with a falling population. 8. Diamond (2004) and Diamond and Orzag (2004) propose a related exercise, with a somewhat different emphasis relative to our text. See also Casey et al. (2003) among the publications from “ageing society” at the OECD (www.oecd.org). 9. There are some exceptions in the form of specific retirement ages for some occupations; yet these exceptions often reflect the recipient’s ability to bargain collectively rather than anything else, and are seldom adjusted for evolutions in working conditions. 10. An important issue that receives little attention is the treatment of surviving (non-working) spouses. There are vast differences across systems and within systems (as in many cases somebody who becomes a widow when his/her partner is one day away from retirement may receive a much better treatment than somebody who becomes a widow one day after his/her partner has retired). Reformers may be tempted to “save money” to the system by unduly reducing the pension income accruing to surviving spouses – just because these are not a politically well-defined and vocal group in the national political arena. 11. To assess the qualitative effect of variations in the population size on the rate of return to capital, consider the textbook example of a closed economy where the real rate of returns on capital (and financial assets) is four per cent. Assume returns to scale in production are constant, with capital and labour as production factors; the capital share in GDP is 30 percent and the elasticity of substitution between the two factors is 0.7. Under the assumption of a constant contribution rate and retirement age, a 10 per cent unexpected decline in the size of the working population will cut the pay-as-you-go pension by 10 per cent, but it will cut the return to capital only by a small fraction (10 per cent) of four per cent. While
9780230_547018_14_cha12.indd 444
3/19/2009 8:35:07 PM
Pensions and Children
12.
13.
14. 15. 16.
17. 18. 19.
445
both systems are affected by the population shock, the funded system is obviously much less vulnerable to adverse demographic developments. As is well known, intergenerational redistribution is an advantage if the economy is “dynamically inefficient”, i.e. it has over-accumulated capital. In this case, it has been shown that the pay-as-you-go system can increase consumption of all generations, by making the currently old generation “eat” the capital stock (the same could be achieved by granting transfers to the current generation, financed with the accumulation of public debt). Note that the implicit (or explicit) accumulation of public debt would not correspond to an increase in the debt to GDP ratio, as long as the interest rate is lower than the growth rate: in such a situation the rate of debt accumulation is slower than growth. A situation of dynamic inefficiency is more likely, the greater the population growth rate and the lower its productivity. While the pay-as-you-go system obviously is not inherently inferior to a funded system when it offers a rate of return below the market rate of interest, there may be periods in which the growth rate is above the market rate of interest. The 1960s may have been such a period. However, economists tend to be quite sceptical about the idea that a negative difference between the rate of interest and the rate of growth could last forever. If this were the case, one may argue that the price of assets whose returns tend to match aggregate growth rate (for instance, land) would be infinite (or indeterminate). People would feel so rich that no one would think it necessary to save: thus the over-accumulation of capital would disappear, driving the return to capital above the rate of growth. See Sinn (2000) for the general proofs. See also Fenge and Werding (2003a, b) for an empirical assessment. We have assumed that this income grows at a real rate of two per cent per year. We have argued above that investing pension contributions in newly issued non-productive assets backed by future domestic tax liabilities makes a funded system equivalent to a pay-as-you-go system. We should stress here that this is not an argument against investment in government bonds by pension funds. Provided that the government does not issue new debt (i.e. does not run a budget deficit), a pension fund that invests part of new pension contributions in government bonds “frees” private financial resources that can be redirected towards the accumulation of domestic capital and foreign wealth. See Pensions Commission (2004). Nevertheless, the corresponding contraction of non-market production in the household sector should not be disregarded. See Sinn (2004).
References Allen, F. and D. Gale (2000) Comparing Financial Systems, MIT Press, Cambridge, MA. Angeletos, G. M., A. Repetto, J. Tobacman and S. Weinberg (2001) “The Hyperbolic Buffer Stock Model: Calibration, Simulation, and Empirical Evaluation,” Journal of Economic Perspectives 15(3), 47–68. Casey, B., H. Oxley, E. Whitehouse, P. Antolin, R. Duval and W. Leibfritz (2003) Policies for an Ageing Society: Recent Measures and Areas for Further Reform, Working Paper 369, OECD Economics Department.
9780230_547018_14_cha12.indd 445
3/19/2009 8:35:07 PM
446 Designing the European Model Choi, J. J., D. Laibson, B. C. Madrian and A. Metrick (2002), “Defined Contribution Pensions: Plan Rules, Participant Decisions, and the Path of Least Resistance,” Tax Policy and the Economy 16, 67–114. Diamond, P. A. and P. R. Orszag (2004) Saving Social Security: A Balanced Approach, Brookings Institution Press, Washington, DC. Diamond, P. A. (2004) “Social Security,” American Economic Review 94(1), 1–24. Fenge, R. and M. Werding (2003a) “Ageing and the Tax Implied in Public Pension Schemes: Simulations for Selected OECD Counties,” CESifo Working Paper no. 841. Fenge, R., and Werding, M. (2003b) “Ageing and Fiscal Balances Across Generations: Concepts of Measurement,” CESifo Working Paper no. 842. Laroque, G. (2004) “Income Maintenance and Labor Force Participation,” mimeo, 2004 CREST-INSEE, Paris. OECD (2004) “The labour force participation of older workers,” Economics Department, May issue. Oksanen, H. (2003) “Population Ageing and Public Finance Targets,” European Commission, Directorate-General for Economic and Financial Affairs, Economic Papers no 196, Brussels. Oksanen, H. (2004) “Pension Reforms: An Illustrated Basic Analysis,” European Commission, Directorate-General for Economic and Financial Affairs, Economic Papers no 201, Brussels. Also published in CESifo Economic Studies, 3/2004. Pensions Commission (2004) Pensions: Challenges and Choices, the first report of the Pensions Commission, www.pensionscommission.org.uk. Sinn, H.-W. (2000) “Pension Reform and Demographic Crisis. Why a Funded System is Useful and Why it is Not Useful,” International Tax and Public Finance 7, 389–410. Sinn, H.-W. (2003) “Ist Deutschland noch zu retten?” Econ, Munich. Sinn, H.-W. (2004) “Europe’s Demographic Crisis. A Plea for Child Pension System,” Tinbergen Lecture 2004, forthcoming in De Economist.
9780230_547018_14_cha12.indd 446
3/19/2009 8:35:07 PM
13 Prospects for Education Policy in Europe
13.1
Introduction
Does education matter or is it just a consumption good like other goods? There is a large consensus among economists that education is an important productive input into the wealth of a nation. For years, people have consistently shown that education enhances individual productivity, which shows up in higher wages by 5 to 10 per cent per extra year of education.1 And, when one tries to explain why some countries are richer than others, the rate of secondary enrolment comes out as one of the most robust determinants.2 Education is an investment: by increasing the total labour input that individuals supply to the market, it boosts GDP per capita and living standards. Furthermore, it is also widely considered that an educated workforce is a valuable asset at times of rapid technological change, because educated workers are better at adopting new technologies.3 Another aspect of education is that it affects the distribution of income. When the supply of educated workers goes up, to restore equilibrium in the labour market, their wage must fall relative to uneducated workers. As the latter tend to earn less, that reduces wage inequality. Lower inequality may in turn be valuable because it reduces social conflict, that is, voters have lower incentives to support costly redistributive policies,4 and individuals have lower incentives to engage in crime and social unrest. In most European countries, the public sector holds a quasi-monopoly on the provision of education. In most countries, the share of private expenditures in total educational spending does not exceed 10 per cent. It is not totally clear why this should be so, but at least there are good arguments that justify public intervention in the educational sector. First, parents may not make the right educational decisions for their children, especially if they come from disadvantaged backgrounds that are poorly informed about the costs and benefits of alternative options. Second, education is more valuable for an individual if he or she interacts with individuals who are themselves educated. For example, an elaborate vocabulary is only useful if one talks to 447
9780230_547018_15_cha13.indd 447
3/19/2009 3:55:41 PM
448 Designing the European Model
people who can understand that vocabulary. By acquiring education, individuals increase the value of education for others, but they are not remunerated for that positive effect (hence called an externality by economists). Thus, individuals may spontaneously acquire too little education. One key question is whether direct provision is an adequate form of government intervention, or whether one could promote a more decentralised approach that would contain costs and allow for greater diversity of individual choices. Historically, public education has not been put in place to tackle the market failures just discussed, but rather to attain political goals. For example, the public education system in France was established in the context of a struggle of influence between the Church and the State. And educational curricula played an important role in promoting French national identity and the French language, for example preparing for “revenge” for the loss of Alsace and Lorraine in 1870. If government involvement in education is now viewed as an economic intervention rather than shaping the citizenship’s beliefs for political reasons, one may well reconsider the working of the public education system. Direct provision of public education transforms decisions into collective decisions that, arguably, might better be left at the individual level, such as: what should children learn, how fast, where, and with which techniques? Furthermore, such provision often eliminates useful competitive mechanisms that help contain costs and correct errors. For example, in a public education system, it is difficult for voters and tax-payers to evaluate the system’s efficiency. A deterioration of outcomes can always be blamed on resources being insufficient rather than misallocated. This chapter discusses these issues. It starts by providing a panorama of Europe’s educational achievements in light of the recent OECD PISA study. The chapter then discusses the cost of education and casts doubts on the usefulness of often-advocated costly policies, such as reductions in class size. The final part reviews evidence suggesting that competition between schools and parental choice bring economic benefits, especially if institutions such as central exams allow for transparent choice.
13.2
How do European countries perform?
Before starting the discussion, it is interesting to point out that there now exist measures of student achievement that are comparable across countries, in particular the OECD-sponsored PISA (Programme for International Student Assessment) study. This is an internationally standardised assessment that has been jointly developed by participating countries and administered to 15-year-olds in schools. The survey is repeated every three years. It was implemented in the first assessment in 2000 and in the second assessment in 2003. Tests are typically administered to between 4,500 and 10,000 students in each country.
9780230_547018_15_cha13.indd 448
3/19/2009 3:55:41 PM
Prospects for Education Policy in Europe 449
How do European countries fare in that comparison? Table 13.1 summarises the average math proficiency of European students, comparing them with the other countries participating in that study. It is not obvious how to interpret these scores. To make them more transparent, the OECD has made six groupings, from level 1 to 6. The median level, level 3, ranges from 483 to 544 points: because of averaging, most countries fall into that interval. The corresponding competences are described as follows: At level 3, students can execute clearly described procedures, including those that require sequential decisions. They can select and apply simple problemsolving strategies. Students at this level can interpret and use representations based on different information sources and reason directly from them. 5 In some countries, the average student is near the top of that level or even at level 4, which is defined as follows: At level 4, students can work effectively with explicit models for complex concrete situations that may involve constraints or call for making assumptions. They can select and integrate different representations, including symbolic ones, Table 13.1
PISA ranking in average math proficiency
Hong Kong Finland Korea Netherlands Liechtenstein Japan Canada Belgium Switzerland Macao Australia New Zealand Czech Republic Iceland Denmark France Sweden Austria Germany Ireland
Math score
Level
550 544 542 538 536 534 532 529 527 527 524 523 516 515 514 511 509 506 503 503
4 3/4 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3
Slovak Republic Norway Luxembourg Hungary Poland Spain United States Latvia Russia Italy Portugal Greece Serbia Turkey Uruguay Thailand Mexico Indonesia Tunisia Brazil
Math score
Level
498 495 493 490 490 485 483 483 468 466 466 445 437 423 422 417 385 360 359 356
3 3 3 3 3 3 2/3 2/3 2 2 2 2 2 2 2 1 1 1 1 1
Source: OECD (2004a), Table 2.5c, p. 356.
9780230_547018_15_cha13.indd 449
3/19/2009 3:55:41 PM
450 Designing the European Model
linking them directly to aspects of real-world situations. (...) They can construct and communicate explanations and arguments based on their interpretations, arguments, and actions. Conversely, in 11 countries the average student falls short of level 3, and in a number of them he virtually is on the border between level 3 and level 2, which involves the following skills: At level 2, students can interpret and recognise situations in contexts that require no more than direct inference. They can extract relevant information from a single source and make use of a single representational mode. Students at this level can employ basic algorithms, formulae, procedures or conventions. They are capable of direct reasoning and making literal interpretations of the results. Table 13.1 suggests a large amount of heterogeneity in performance. Despite the averaging, four European countries fall short of level 3: Italy, Portugal, Greece and Serbia. Latvia and Spain are almost at the frontier between levels 2 and 3. At the other end of the spectrum, Finland, the Netherlands and Liechtenstein come close to reaching level 4. If we now look at reading proficiency, as reported in Table 13.2, we see that the ranking and the degree of heterogeneity between countries is quite similar to math. Country average performance ranges from level 1 (337–407 points) to level 3 (481–552), with level 2 between 408 and 480 score points. Level 3, the most frequent one, is defined as follows: Students proficient at level 3 on the reading literacy scale are capable of reading tasks of moderate complexity, such as locating multiple pieces of information, making links between different parts of a text and relating it to familiar everyday knowledge. 6 On the other hand: Students proficient at level 2 are capable of basic reading tasks, such as locating straightforward information, making low-level inferences of various types, working out what a well-defined part of a text means and using some outside knowledge to understand it. Finally, level 1 is the lowest understanding level; literacy problems start when one is below that level: Students proficient at this level are capable of completing only the simplest reading tasks developed for PISA, such as locating a single piece of information, identifying the main theme of a text or making a simple connection with everyday knowledge.
9780230_547018_15_cha13.indd 450
3/19/2009 3:55:42 PM
Prospects for Education Policy in Europe 451 Table 13.2
PISA ranking in average reading proficiency Reading score
Level
Reading score
Level
Finland
543
3
Austria
491
3
Korea
534
3
Germany
491
3
Canada
528
3
Latvia
491
3
Australia
525
3
Czech Republic
489
3
Liechtenstein
525
3
Hungary
482
3
New Zealand
522
3
Spain
481
3
Ireland
515
3
Luxembourg
479
2
Sweden
514
3
Portugal
478
2
Netherlands
513
3
Italy
476
2
Hong Kong
510
3
Greece
472
2
Belgium
507
3
Slovak Republic
469
2
Norway
500
3
Russia
442
2
Switzerland
499
3
Turkey
441
2
Japan
498
3
Uruguay
434
2
Macao
498
3
Thailand
420
2
Poland
497
3
Serbia
412
2
France
496
3
Brazil
403
1
United States
495
3
Mexico
400
1
Denmark
492
3
Indonesia
382
1
Iceland
492
3
Tunisia
375
1
Source: OECD (2004a), Table 6.2, p. 444.
The average student is at least at level 2 in all European countries. Nevertheless, the breadth of score levels across countries, in light of the above definition, can be considered as large as for math. Another striking fact is that countries which do well in math also do well in reading, and the laggards are the same in both fields. The correlation coefficient, across countries, of the math and the reading scores is as high as 0.94. A natural question is: how can we explain such important differences among European countries in proficiency levels? That heterogeneity does not seem to come from genetic differences, nor cultural differences between people: countries that are quite similar genetically and/or culturally seem to experience wide differences in achievement levels. Finland does much better than neighbouring Sweden, while Norway does poorly. Furthermore, it is unlikely that countries that are, for some reason, very “talented” in math could also be very “talented” in reading. These are two rather different
9780230_547018_15_cha13.indd 451
3/19/2009 3:55:42 PM
452 Designing the European Model
kinds of skills – as suggested by the fact that girls are better than boys in reading, but boys perform better in math. Thus it is unlikely that “talent,” wherever it comes from, explains these differences. It does not seem that a relationship between GDP per capita and achievement can explain the difference between European countries, although it may help explain why the worst performers – Thailand, Uruguay, Turkey, Mexico, Indonesia, Tunisia and Brazil – are all emerging economies. Among the richest European countries, Liechtenstein ranks 5th in mathematics, Switzerland 9th, Luxembourg 23rd, and Norway 22nd. Finally, as we argue below, there is only a moderate influence of spending per pupil on these achievement measures. Thus, the most likely explanation is that differences in achievements across countries are due to differences in the way their school system is managed. That important conclusion suggests that substantial improvements in schooling achievements can be obtained by learning from the experience of other countries.
13.3 The cost of education Education accounts for a large share of GDP: in 2001, from some 4 per cent in Greece to as much as 7 per cent in Denmark. That is not specific to Europe: in the United States, the fraction is as high as 7.3 per cent. Thus, from a cross-country comparison perspective, the cost of education does not seem abnormally high in Europe. However, in the United States there is a growing sense of an “education crisis,” based on the observation that costs are growing, with little impact on achievement levels that remain mediocre and seem to deteriorate if one uses standardised tests. We want to know whether these problems may harm Europe, too. This section discusses the basic economics of the cost of education, the next one asks whether increased spending is an efficient investment. From an economic viewpoint, the key property of the educational market is that it is affected by the so-called cost disease. The technology of education has been virtually unchanged in the last thousand years: teachers lecturing in front of an audience. That makes education similar to a performing art like theatre or opera and stands in contrast to the production of industrial goods, where a single worker, thanks to improvements in technology and machinery, can produce a far larger quantity of goods than in the past. It implies that the unit cost of education should roughly grow like wages, which themselves roughly grow as GDP per capita. In contrast, for industrial goods, prices fall relative to wages because of productivity improvements. Thus education becomes progressively more expensive relative to industrial goods. That is more or less borne out by the data: Americans spent 30 per cent of GDP per capita on each student in 2001, and that figure is exactly the same as in 1991. Thus, parents have to give up a greater amount of physical
9780230_547018_15_cha13.indd 452
3/19/2009 3:55:42 PM
Prospects for Education Policy in Europe 453
consumption to get the same education level (in years) for their children. As for the European Union, that figure is 25 per cent, down two points from 27 per cent in 1991.7 In the performing arts, the cost disease naturally drives consumers away; they move to substitutes that are not to the same extent affected by the cost disease, like movies, TV, and DVDs. In the realm of education, two things may prevent that from happening. First, such substitutes may not be available, a point to which we return below. Second, educational expenditures are socialised, so that parents do not see the cost of their children’s education. As a consequence of the cost disease, we expect the share of GDP devoted to education to remain constant as long as the school population remains constant, and it should go up as long as one increases the size of that population. According to this logic, pledges to bring more people to upper education, in order to catch up with more advanced economies like the US, can only be met at a substantial cost for the taxpayer. One may well accept that as a fact and conclude that it is desirable for countries to spend a growing share of their income on education. But, alternatively, one may challenge the “cost disease” view and explore the extent to which one can increase productivity in the educational sector. That brings us into a realm of controversy: While the rise of education suggests that costs could now be cut drastically, many people still recommend policies which precisely go in the other direction, such as costly reductions in classroom size. We discuss these policies in the next section. Box 13.1
The cost disease in economics
The cost disease has been analysed by the American economist William Baumol from Columbia University. It affects goods like the performing arts where there is no – or only small – room for productivity improvements. As other sectors of the economy see their productivity rising, one hour of labour can produce more goods, so that the relative price of the performing arts goes up. A simple example is as follows: assume that one hour of work produces 2 shirts, and that 1000 people can see a play by 10 actors which lasts 2 hours. These 10 actors could alternatively produce 10 2 2 40 shirts, so that each spectator must pay the equivalent of 40/1000 0.04 shirts to see the play. Suppose that a new invention raises productivity tenfold in the textile industry. The actors could now produce 400 shirts. The textile industry is willing to pay its workers 10 times more than before; to match it, theatres must increase their price to the equivalent of 400/1000 0.4 shirts. The price of theatre has risen tenfold relative to the price of shirts. As the economy grows, the performing arts, whose productivity cannot keep up with other sectors, become ever more expensive in relative terms. If people can find substitutes for the performing arts that are not affected by the cost disease, they will gradually shift to these alternatives as the economy grows, and the performing arts will slowly disappear, in the sense that they will employ fewer and fewer workers. That is what seems indeed to be happening for theatre and
9780230_547018_15_cha13.indd 453
3/19/2009 3:55:42 PM
454
Designing the European Model
opera. But, if that is not the case, people may actually consume more of them, despite the fact that they cost more. This is because less can be spent on the goods that become cheaper, so that more money can be devoted to the goods affected by the cost disease (In technical terms, it is said that the income effect then dominates the substitution effect). That seems to hold for goods like health and education, whose expenditure shares go up as the economy grows. In the case of education, people have been studying longer; given that the cost per year per pupil has grown in line with GDP, the share of educational expenditures in GDP has trended upwards.
13.4 Spending – what does it buy us? The evidence is clear that an extra year of education is a productive investment, both at the individual and economy-wide level. However, many advocate policies that increase spending given the number of students and the number of years they spend at school. That raises the following questions: What is the effect of increased spending per student? How should a given amount of spending be allocated between say, classroom size, teachers’ skills, books, computers, etc.? To answer that question, we may again use the PISA study, which reports correlations between, on the one hand, achievement measures and, on the other hand, school resources and organisation. The lesson from this study is that spending indeed seems to affect performance positively (see OECD 2004a, figure 2.20, p. 102), but the effect is not very strong and it only accounts for a small proportion of the crosscountry variation in performance. Furthermore, the effect is probably overstated, because richer countries spend more, and that simple correlation may also capture other effects of living standards on performance (through nutrition, social norms, and so on). A lot of variation in achievement levels is not due to spending: the Slovak Republic spends one fourth of the United States’ expenditure on each student and yet fares better in mathematics. Figure 13.1 plots cumulated spending on education per student between 6 and 15 years, adjusted for PPP,8 against the PISA math score.9 There is no tight connection between spending and performance. We can observe a “top league” of efficient providers, such that no other country performs substantially better on both proficiency and spending: these countries are the Slovak Republic, the Czech Republic, Korea, Finland and the Netherlands. Similarly, we can spot clear underperformers: those countries that seem most remote from the efficiency frontier, getting the lowest value per dollar spent: Mexico, Portugal, Spain, Greece, Italy and the United States(!). Thus the effect of spending on performance is not very strong. Can we make more detailed statements? For example, do specific types of spending, like reducing classroom size, have a more significant effect of their own? That is not what the data seem to indicate. In fact, the raw cross-country correlation between students per teacher and reading proficiency is positive: countries
9780230_547018_15_cha13.indd 454
3/19/2009 3:55:42 PM
Prospects for Education Policy in Europe 455 600
Finland
PISA mathematics proficiency
550
Japan Belgium Netherlands Switzerland Denmark Sweden Germany Ireland France Austria Norway Hungary Spain United States Korea
Czech Republic 500 Slovakia
Poland
Italy
Portugal 450 Greece Spain 400 Mexico 350 0.000
20.000
40.000
60.000
80.000
100.000
Spending (US dollars converted using PPPs)
Figure 13.1
Cumulative spending per student vs. mathematic proficiency
Source: OECD, PISA 2003, Table 2.6, p. 358.
with more students per teacher perform better. The same is true for mathematics (see Figures 13.2 and 13.3). While such raw correlations are replete with biases and should be taken with caution, one may think of a number of explanations: low classroom size may come at the expense of other kinds of investment, or there may be a trade-off between quality and quantity of teachers. The economics literature has obviously gone beyond these simple correlations: an abundant literature, based on US data, has analysed the costs and benefits of alternative educational policies. Hanushek (2002, chapter 30) has produced an extensive meta-analysis of that literature, using 376 different studies. He finds that only a small number of spending items have a significant effect on student performance. The pupil-teacher ratio, in particular, has an effect that is not statistically different from zero in 72 per cent of the studies, while the effect is statistically positive in 14 per cent of them and statistically negative in the remaining 14 per cent. Yet costly reductions in that ratio are constantly advocated as the path to success. Table 13.3 documents its downward trend and the associated increase in the cost of education in the United States. The Hanushek meta-study also suggests that facilities do not have any noticeable impact on educational achievement. Teacher salary and expenditure per pupil seem to matter a little bit more, but only marginally: in both cases, some 75–80 per cent of the studies find negative or insignificant effects. What seems to matter most for outcomes is the teacher’s ranking on an IQ test! If teacher quality matters, then it is not surprising that reducing classroom size is counterproductive: reducing classroom size means appointing more teachers, which is likely to imply a reduction in the average quality of
9780230_547018_15_cha13.indd 455
3/19/2009 3:55:43 PM
456
Designing the European Model 550 Finland Netherlands United Kingdom
PISA mathematics proficiency
540 530
Belgium
520
Denmark
510 Austria
Czech Republic
Sweden
France
Germany
Ireland
500 Norway 490
Poland
Hungary
United States
Spain
480 470 Portugal
Italy
460 450 0
Figure 13.2
5
10 15 Students per teacher
20
25
Students per teacher vs. mathematics proficiency
Source: OECD, PISA 2003, Table 2.5c, p. 356; OECD, Education at a Glance 2004, Table D2.2, p.377.
550 Finland
540
PISA reading proficiency
530 United Kingdom
520 Sweden 510
Austria
490
Hungary 480
Portugal Greece
470
Ireland
Belgium
Norway 500
Netherlands
United States Poland Germany Denmark Spain Czech Republic
France
Italy
460 450 0
Figure 13.3
5
10 15 Students per teacher
20
25
Students per teacher vs. reading proficiency
Source: OECD, PISA 2003, Table 6.2, p. 444; OECD, Education at a Glance 2004, Table D2.2, p.377.
teachers. That suggests that one should actually pursue the opposite policy: select fewer, but higher quality teachers, reward them accordingly, and increase classroom size to make sure that all pupils can access the high quality teachers.
9780230_547018_15_cha13.indd 456
3/19/2009 3:55:43 PM
Prospects for Education Policy in Europe 457
Table 13.3
The soaring cost of education in the United States Pupil-teacher ratio
Expenditure per pupil (constant 2001 $)
1960
25.8
2,275
1970
22.3
3,849
1980
18.7
5,146
1990
17.2
6,996
1995
17.3
7,090
2000
15.8
8,044
Source: Hanushek (2002), National Center for Education Statistics.
Box 13.2 The French ZEP experience In 1982, the French ministry of education put in place a programme of increased support for schools in disadvantaged neighbourhoods, called ZEP (zones d’éducation prioritaire). The measure was meant to be temporary, but, as is often the case with costly administrative programmes, is still in force today. Over the years, different waves of ZEPisation have taken place, so that the fraction of ZEP schools has increased. Schools in a ZEP zone have greater financial means than schools outside a ZEP zone. The bulk of this money is used to reduce classroom size, from a national average of 25 to 23, which increases teaching costs per pupil by 8 percent. Furthermore, teachers who teach in those schools have a higher salary (which boosts costs by another 4–5 percent) and are promoted more quickly. While most studies typically have failed to find a significant effect of ZEPs, they often fail to control for the fact that students in ZEPs are from more disadvantaged backgrounds. A recent study by Bénabou, Kramarz and Prost (2004) is, however, immune to that criticism because, using the 1989-90 ZEPisation wave, it looks at the difference in performance between students in a ZEP school and students in that same school before if joined a ZEP, and compares it to schools whose status was unchanged during that wave. They use results at the national Baccalauréat degree as a measure of performance. Their key finding is that being in a ZEP has essentially a zero effect on performance. While that confirms the results of Hanushek’s meta-study, there is some room to believe that they may be due to the negative stigma associated with ZEP status, with the most motivated parents taking their kids to private school when their district’s school becomes classified as ZEP. In any case the presumption holds that the policy has failed, especially given its high cost.
The finding that indiscriminate spending and class size reduction are inefficient is confirmed by a recent event study by Bénabou et al. (2005) on the French experience of ZEP (Zones d’Education Prioritaire) (see Box 13.2) and by several recent studies such as, for example, Woessman and West (2005).
9780230_547018_15_cha13.indd 457
3/19/2009 3:55:43 PM
458 Designing the European Model
We conclude from this section that there is little empirical support for the popular view that increased spending per student, especially in the form of reduced classroom size, has had a significant effect on educational quality. That does not imply that it could not. To quote Hanushek’s words: The evidence does not say that money and resources never matter. Nor does it say that money and resources could not matter. It simply describes the central patterns of results given the current organisation and incentives in schools. Indeed, a plausible interpretation of the evidence is that some schools in fact use resources effectively, but that these schools are counterbalanced by others that do not. At the same time, the expansion of resource usage unaccompanied by performance gains implies a high level of inefficiency in the current operations of schools.
13.5
Meritocracy and selectivity
An issue in defining an educational system is: how demanding should it be on pupils? How much should one emphasise the achievements of some pupils relative to others? How large should differences be between “good schools” and “bad schools” and how selective should the “good schools” be? That is, how meritocratic should the system be? From an economic perspective, a number of issues are involved. At a first level of analysis, one may just view education as an input in the production of human capital. If people were just buying it on the market, no selection would be required. Indeed, that is what is happening in private, continuing education classes when people take classes in foreign languages, driving, computers or math. Nobody is being turned down, entry tests are used to determine which level is appropriate, and exit tests give the customer a signal that he may use on the job market. This suggests that selectivity is the by-product of the artificial scarcity created by the fact that education is free. However, full-curriculum private schools are often selective, despite often charging high tuition. That is because education is not only an input in human capital, but also a signal about one’s intrinsic productivity.10 An institution that becomes less selective allows more people to get education, which raises their productivity, but at the same time it performs a poorer job at signalling the intrinsic productivity of its graduates. For that reason it may not accept applicants even though they might want to pay the full cost of education. Another aspect of meritocracy is that it provides incentives to work hard in order to gain entry into the good schools. However, note that too much selectivity may be counter-productive in that respect, as one does not want to invest too much into winning a contest if the probability of winning it is too low. Also, critics argue that meritocracy is inegalitarian because
9780230_547018_15_cha13.indd 458
3/19/2009 3:55:44 PM
Prospects for Education Policy in Europe 459
it favours students from privileged backgrounds. That argument may hold only if one believes that the alternative would be to put all pupils at the same level by some coercive means. However, the likely alternative is that if one reduces academic meritocracy as an engine of social promotion, it will be replaced by money and social networks and social mobility will be even lower.
Box 13.3
Main features of the Finnish education system
Figure 4.4 gives an overall picture of the Finnish education system. Municipalities are responsible for providing and running the basic education and upper secondary schools. Basic education is based on comprehensive schools and on the principle of equality. Pupils usually attend the local school in the neighbourhood where their family lives, though there is some movement from neighbourhood schools to other schools especially between the 6th and 7th grades when students move from primary to lower secondary school. The basic education is mandatory, the curriculum is largely the same for everybody, and the schools are publicly funded and run. There are a very small number of private schools, mainly a few foreign schools in Helsinki. As regards educational achievement in basic education, it has been found that there are some differences in learning results between schools, between boys and girls and between different regions of the country. These differences are, however, relatively minor. Moreover, it has been found that in Finland the influence of students’ socio-economic background on learning performance is among the lowest among countries in the PISA system.a About 55 percent of the students finishing basic education entered the upper secondary school system and 35 percent the vocational school system (in 1999). Three percent of students continued an additional, voluntary 10th grade of basic education, while seven percent did not continue in the education system in the year after they finished basic education. Both upper secondary and vocational schools provide qualification for continuing into tertiary education, which splits into university and polytechnic education. Entry to the upper secondary schools is based on the final grades that students achieve in the basic education system. This has led to significant competition among students for places in the best upper secondary schools. The competition is very visible in the bigger cities like Helsinki. The grades required for achieving entry to one of the best high schools in Helsinki are quite high and the entry thresholds make news every year. Naturally, the competitive elements have led to significant differences in the education results among schools, though even at this stage students’ socio-demographic background is the most important explanatory factor behind differences in educational attainment in upper secondary schools. About 65 percent of an age cohort enters the tertiary system and about 43 percent of the entrants to tertiary education go to the polytechnics. The emphasis on competition between students continues at the level of entry to the university system. The universities exercise a system of entry exams and entry can be very difficult, depending on the subject that a student finishing upper second schooling wants to study. There is significantly less competition among students for places in the poly technics. a
See Asplund and Leijola (2005) for a summary of these results and references.
9780230_547018_15_cha13.indd 459
3/19/2009 3:55:44 PM
460
Designing the European Model
UNIVERSITIES
POLYTECHNICS
UPPER SECONDARY SCHOOL
VOCATIONAL SCHOOL
Matriculation examination
Professional qualification
BASIC EDUCATION Lower secondary school (3 years) Primary school (6 years)
Figure 13.4
The Finnish education system
The Finnish education system, discussed in Box 13.3, involves stiff competition from pupils to get into the best high schools, while remaining quite egalitarian in terms of the resources given to each student. This may be an important explanation of why Finland fares extremely well in the PISA study in terms of overall achievements (see Tables 13.1 and 13.2). Moreover, outcomes are not more unequal than in other countries, and socio-economic background does not matter more than in other countries. That experience suggests that meritocracy is a powerful tool for boosting educational achievements, while not generating more inequality than the feasible alternatives.
13.6 Fighting the cost disease: school competition and parental choice In light of the rising costs of education, which is a particularly salient phenomenon there, the US has in recent years seen a heated debate on the problem of school quality. This has triggered a wealth of empirical studies as well as interesting proposals regarding, for example, school vouchers, greater parental choice between schools, greater accountability of public schools, incentive schemes based on academic achievements, and the like. Even though education in Europe is not as costly as in the US, we believe it is time to ask the same question here, rather than wait for costs to become unbearable. The idea behind parental choice is simple: parents would be free to choose the school which gives the best results given their children’s needs. Schools that are successful in attracting pupils would be allowed to grow
9780230_547018_15_cha13.indd 460
3/19/2009 3:55:44 PM
Prospects for Education Policy in Europe 461
accordingly, as their resources would be adjusted upwards. Schools that lose children would experience a proportional reduction in resources and can eventually be forced to close. For this system to work, sufficient autonomy must be granted to schools in their organisational and pedagogical choices. The outcome would somewhat mimic a market outcome, but one need not resort to full-scale privatisation to achieve it, and one can avoid the adverse consequences for equal opportunity and social mobility of a pure private system. In most current public systems, pupils are allocated to a school depending on their residence, and schools are run in a centralised way. On paper, this is supposed to guarantee maximum equality. In practice, there is no incentive for either cost reductions or quality improvements. In addition, disadvantaged families suffer most from inefficient public schools, as opting out into a private school is too costly for them. Thus they might be the individuals who gain most from increased school competition. Against these arguments stand the two traditional ones that (i) parental choice need not be the best solution from the child’s perspective (some parents, for example, may, place excessive weight on religious instead of academic content), and (ii) pupils may be sorted into schools in a segregated and/or inegalitarian way. The available amount of evidence on the role of parental choice and competition is not huge, because of the scarcity of real world voucher programmes and similar school competition policies. Furthermore, that evidence is not always easy to interpret. For example, a prominent study by Rouse (1998) finds that a Milwaukee voucher program that allowed a small number of poor children to attend private schools was successful in that participants fared better than a “control group” of non-participants. That tells us that private schools are “better” than public schools, which is not very surprising since a “worse” private school could not survive competition from free public schools. But it tells us little about what would happen if the educational system was entirely redesigned to be based on vouchers. Angrist et al. (2002) studied a truly random experiment in Colombia, the so-called Programa de Ampliacion de Cobertura de la Educacion Secundaria, which gave vouchers covering 50 per cent of the cost of private schools to 125,000 students selected by a lottery mechanism. They found that lottery winners had better achievements, confirming the view that the private schools were better. Again it is hardly surprising that those being allowed to choose between a private and a public school achieve better results than if they are confined to a public school. We expect that to always be the case unless people grossly misperceive school performance.11 These studies cannot answer the key question, which is: does parental choice create competitive pressure that leads to increased efficiency or does it only lead to sorting of pupils by skills, parental background, religious beliefs, income and other characteristics, with potential adverse effects on social cohesion and also
9780230_547018_15_cha13.indd 461
3/19/2009 3:55:44 PM
462
Designing the European Model
(to the extent that there are peer effects) on the overall performance of the school system? Fortunately, there are studies – a number of them by Hoxby (2004)12 – that directly document such competitive pressure by looking at the effect of parental choice programmes on achievement in public schools – that is in the schools that the choice students are allowed to leave. For example, the Milwaukee voucher programme (see Box 13.4) was extended in 1998, covering now 15 per cent of school enrolment. That means that public schools can lose a substantial fraction of their students. Furthermore, the programme is designed so that these schools also lose resources – thus public schools have a genuine incentive to improve to retain students. Box 13.4
The Milwaukee Voucher Programme
“Vouchers for poor students in Milwaukee were enacted in 1990 and were first used in the 1990–91 school year. Currently, a family is eligible for a voucher if its income is at or below 175 percent of the federal poverty level (at or below 17,463 dollars for a family of four). For the 1999–00 school year, the voucher amounted to 5,106 dol lars per student or the private school’s cost per student, whichever was less. For every student who leaves the Milwaukee public schools with a voucher, the Milwaukee public schools lose state aid equal to half the voucher amount (up to 2,553 dol lars per voucher student in 1999–00). Milwaukee’s per pupil spending in 1999–00 was 8,752 dollars per pupil, so the district was losing 29 percent of the per pupil revenue associated with a voucher student. Currently, the vouchers may be used at secular and non-secular private schools” (Hoxby 2004, p. 24). Using differences between eligibility levels to identify the degree of competition faced by schools, Hoxby finds that achievement (as measured by test scores) improved in the schools that faced more competition from private schools relative to those that faced less competition. Evidence from parental choice programmes in Michigan and Arizona (Charter schools) points to similar results. These are important findings suggesting that introducing competition between schools would improve the efficiency of the educational system. What about potential adverse effects? Can we fear, for example, that vouchers would lead to increased segregation, and that the most disadvantaged would suffer because they would stay in degraded public schools? We do not have a clear answer on that. Epple and Romano (1998) analyse the impact of vouchers in a theoretical model. They predict that introducing vouchers will mostly benefit bright students from poor families, who would be able to move to private schools that will not charge them a high price (thus remunerating them for the positive peer effects they create), while low-ability low-income students would lose, but not by much. However, their analysis ignores any positive effect of school competition on the efficiency of public school. If these effects are strong enough, voucher systems could in fact benefit everybody. Enhancing parental choice and school competition is not absent from the European debate either, although many countries maintain an egalitarian, rigid
9780230_547018_15_cha13.indd 462
3/19/2009 3:55:44 PM
Prospects for Education Policy in Europe 463 approach to schooling. In particular, in the UK, the scope for competition and parental choice was greatly enhanced by the 1988 Education Act. A recent study suggests that this had positive effects, in that achievement improved substantially in the schools that were granted more autonomy. However, there is no evidence of competition improving performance in schools that remained in the traditional, non-autonomous system.
Box 13.5
Education reform in the UK
In 1988 the Thatcher government enacted an Education Act, which substantially increased the scope for parental choice. The key ingredient of the reform was to force schools to accept pupils indiscriminately up to some limit, and to tie a school’s financial resources to its number of pupils. Thus, it was felt, better schools will attract more pupils, which would trigger an increase in their resources, allowing them to grow at the expense of less efficient schools. At the same time, school autonomy was enhanced by transferring decision making from the district (Local Education Authorities, LEAs) to the school level. Finally, new schools (so-called “Grant Maintained” GM) were created. They enjoyed even greater autonomy and were totally independent of the LEAs. Schools under the supervision of LEAs could opt out of that system and become grant-maintained. In particular, that procedure involved parental vote. Clark (2005) identifies the effects of the reform on educational achievements by looking at differences in outcomes among schools that decided to opt out of the LEA system by a narrow margin, and schools that decided to stay in by a narrow margin (thus mimicking a random experiment). He finds large positive effects of the GM status on graduation rates, and shows that these effects are not due to student selection. On the other hand, contrary to some findings by Hoxby, he does not find large spill-over effects on schools that remained in the LEA system – they do not seem to have been ‘disciplined’ by competition from GM schools.
Box 13.5 gives basic information and results about the UK education reform under the Thatcher government, which gave more scope for parental choice and more independence to schools with respect to local education authorities.
13.7 Putting the evidence together: designing an efficient schooling system The preceding discussion suggests that school autonomy associated with parental choice has positive effects on the performance of the educational system, while centralised meritocratic systems also work well. More systematic cross-country studies based on PISA-type data can correlate achievement measures with data on how the system is managed in each country. These data are summarised in the appendix to this chapter in Table A13.1, and capture characteristics such as the degree of school autonomy, the prevalence of central exams, the importance of private vs. public financing and
9780230_547018_15_cha13.indd 463
3/19/2009 3:55:45 PM
464
Designing the European Model
management, and so on. They suggest that the combination of autonomy and meritocracy works best. Indeed, these studies show that school autonomy typically works best in countries where a central nation-wide exam takes place at the end of high school, while it has insignificant effects in the absence of a central exam.13 While this type of evidence is not as convincing as natural experiments, since it ignores the reasons why a given country has chosen a given system, it makes a lot of sense. Results at national exams provide a common metric by which parents can evaluate the performance of alternative schools. Furthermore, by designing national exams, governments can set standards for what schools are supposed to achieve, and (provided the central exams are properly designed) reduce parental incentives to put their children into schools with useless or biased curricula. In principle, however, one could also envisage a market-based system for evaluating schools, based on subsequent labour market outcomes, as is the case for higher professional education. As for the financing of education, things are more complex. At one extreme, pure public financing favours equal opportunity, since students can in principle access the same educational resources regardless of their family background. On the other hand, it gives little incentives to cut costs. At the other extreme, pure private financing is inegalitarian: there is indeed evidence that achievements are more likely to depend on family background, the lower the share of public financing.14 But it leaves more room for cost-cutting and price competition. However, price competition does not seem to have managed to bring down the costs of private schools, in part because in most countries the poor, who are the ones who should care about costs, cannot afford private schools and send their children to public schools instead. This suggests that a generous voucher system, in which a large fraction of costs is covered by vouchers, might be a good mechanism to reconcile price competition with equal opportunity. The PISA study suggests that the organisation of public schools has a large impact on achievements. Furthermore, naïve increases in spending, in particular in the form of smaller classes, seem to be an inefficient way of raising achievements. On the other hand, we have argued that substantial improvements can be obtained if one fosters competition, both among students to get into the good schools and among schools to attract the good students. The available evidence suggests that while raising performance, such policies would not be particularly “unfair” or “inegalitarian” relative to current practices. Equal opportunity can be preserved if the financing of education remains public or if private financing is sustained by vouchers for a large enough amount. A national central exam will allow society to make most out of school competition, by giving parents a clear, uncontroversial way of measuring school quality.
9780230_547018_15_cha13.indd 464
3/19/2009 3:55:45 PM
9780230_547018_15_cha13.indd 465
3/19/2009 3:55:45 PM
0 0 100 100
100
0
35
0 100 100
100 100
100
Finland
France
Germany
Greece Hungary Ireland
Italy Japan
Netherlands
90.3
92.6 74.5
95.4 89.8 93.4
83.3
92.1
97.8
93.3 94.2 94.5 96.1
Share of Share of students public covered expendiby ture central (2002) exams (Science & Math)
31.3
93.2 99.1
92.4 94.7 99.1
97.3
85.4
98.8
95.6 45.4 98.9 88.0
Public institutions
Distribution of students in lower secondary education Distribution of students in upper secondary education
68.7
0.0 0.0
0.0 5.3 0.0
2.7
14.3
1.2
4.4 54.6 1.1 12.0
23.8
96.6 94.0
94.5 93.7 100.0
92.9
78.8
95.8
92.3 43.2 98.2 76.9
76.2
0.0 0.0
0.0 6.3 0.0
7.1
21.0
4.2
7.7 56.8 1.8 23.1
7.8
93.7 69.8
93.7 85.9 98.5
92.5
69.5
89.6
90.0 42.1 87.4 97.5
92.2
0.7 0.0
0.0 14.1 0.0
7.5
29.7
10.4
10.0 57.9 12.6 2.5
GovernmentPublic GovernmentPublic Governmentdependent institutions dependent institutions dependent private private private institutions institutions institutions
Distribution of students in primary education
The institutional characteristics of educational systems
Austria Belgium Czech Rep. Denmark
Table A13.1
Appendix
12
14 15
15 11 15
10
15
16
10 12 11 16
First age of selection
Continued
100
48 23
13 68 n.a.
32
31
27
29 43 60 44
Percentage of decisions at school level, lower secondary education (2003)
9780230_547018_15_cha13.indd 466
3/19/2009 3:55:45 PM
Continued
73.8
84.4
n.a.
96.7
88.4
95.3
98.4
96.2
89.2
95.1
96.3
94.9
66.6
95.8
89.5
98.2
Public institutions
Distribution of students in lower secondary education Distribution of students in upper secondary education
0.0
0.0
1.3
5.1
30.1
4.2
0.0
1.8
90.8
93.2
93.0
94.6
67.2
94.9
88.7
97.8
0.0
0.4
2.5
5.4
29.7
5.1
0.0
2.2
90.9
26.9
93.1
96.6
76.9
92.0
81.8
90.1
0.0
70.4
3.2
3.4
12.0
8.0
0.0
9.9
GovernmentPublic GovernmentPublic Governmentdependent institutions dependent institutions dependent private private private institutions institutions institutions
Distribution of students in primary education
16
n.a.
15
16
16
11
15
16
First age of selection
n.a.
85 (England)
n.a.
47
28
50
41
37
Percentage of decisions at school level, lower secondary education (2003)
Sources: Central Exams: Woessmann (2003), (*) only math. Share of public expenditure for all levels of education, OECD (2005), table B3.1; Distribution of students by type of institution, OECD (2005), table D5.1 (remainder to 100% are independent institutions); First age of selection, OECD (2005), table D6.1; Decision level, OECD (2004), Table D.6.6.W
7
US
0
Switzerland
100
50
Sweden
UK
0
Spain
100
0
Portugal
Slovak Rep.
100(*)
Norway
Share of Share of students public covered by expendicentral ture exams (2002) (Science & Math)
Table A13.1
Prospects for Education Policy in Europe 467
Notes 1. 2. 3. 4. 5. 6. 7. 8.
9. 10. 11.
12.
13. 14.
See, for example, Psacharopoulos (2003). See Barro and Lee (1994). See Benhabib and Spiegel (2002). See Saint-Paul and Verdier (1993). Source: OECD (2004a), Figure 2.2, p. 47. PISA (2003), p. 278. Wasmer et al. (2005). Not adjusting for purchasing power parity is likely to be misleading. For example, consider a country where wages are lower than elsewhere, because the cost of a standard basket of goods is lower. This country would be classified as spending less per pupil than elsewhere even though it would hire the same number of teachers per pupil and grant the same living standards to the teachers. Adjusting for PPP allows to compute a measure of the true inputs into education, rather than just their dollar value. Based on OECD (2004a), Table 2.6, p. 358. See Akerlof (1970) and Spence (1973). However, the authors also argue that a cost-benefit analysis can be performed on the basis of their results and that this suggests that shifting from public provision to vouchers has a positive net social value. In an earlier paper, Hoxby (2000) uses instrumental variable techniques rather than natural experiments, reaching similar conclusions. While there is controversy about the robustness of these results (Rothstein 2005), which the authors of the present report are not able to evaluate, the 2004 paper we refer to uses a number of different empirical studies to validate the claim that school competition has a positive impact on productivity, including the Milwaukee voucher programme mentioned in the text. See Woessmann (2005), Bishop (1997), and Juerges et al. (2005). See Schuetz et al. (2005).
References Akerlof, G. A. (1970) “The Market for Lemons: Quality, Uncertainty and the Market Mechanism,” The Quarterly Journal of Economics 84(3), 488–500. Angrist, J., E. Bettinger, E. Bloom, E. King and M. Kremer (2002) “Vouchers for Private Schooling in Colombia: Evidence from a Randomized Natural Experiment,” American Economic Review 92(5), 1535–58. Asplund, R. and L Leijola (2005) “Education and Wage Inequality in Finland: a Review of the Empirical Evidence,” in Asplund, R. and E. Barth, eds, Education and Wage Inequality in Europe, Report B 212 of the Research Institute of the Finnish Economy, Helsinki, Finland, 57–122. Barro, R. J. and J.-W. Lee (1994) “Losers and Winners in Economic Growth,” World Bank Economic Review, Washington, DC., World Bank, 267–97. Benhabib, J. and M. M. Spiegel (2002) “Human capital and technology diffusion,” Federal Reserve Bank of San Francisco, Working Papers in Applied Economic Theory, 2003–02. Bénabou, R., F. Kramar and C. Prost (2004) “Zones d’éducation prioritaire: quels moyens pour quel résultats?” Economie et Statistique 380, 3–30.
9780230_547018_15_cha13.indd 467
3/19/2009 3:55:46 PM
468
Designing the European Model
Bishop, J. H. (1997) “The Effect of National Standards and Curriculumbased Exams on Achievement,” American Economic Review 87(2), 260–64. Clark, D. (2005) “Politics, markets and schools: Quasi-experimental evidence on the impact of autonomy and competition from a truly revolutionary UK reform,” Working Paper, Nuffield College, Oxford. Epple, D. and R. Romano (1998) “Competition between Public and Private Schools, Vouchers and Peer Effects,” American Economic Review 88(1), 33–62. Finnish Ministry of Education (2004) Universities 2004, Annual Report, Helsinki. Hanushek, E. (2002) “Publicly provided education,” in Alan J. Auerbach, ed., Handbook of Public Economics, vol. 4, chapter 30, Elsevier, Amsterdam, North Holland. Hoxby, C. M. (2000) “Does Competition among Public Schools Benefit Students and Taxpayers?” American Economic Review 90(5), 1209–38. Hoxby, C. M. (2004) “School Choice and School Competition: Evidence from the United States,” Swedish Economic Policy Review 10(2), 11–67. Juerges, H., K. Schneider and F. Buechel (2005) “The effect of central exit examinations on student achievement: quasi-experimental evidence from TIMSS Germany,” Journal of the European Economic Association 3(5), 1134–55. OECD (2004a) “Learning for Tomorrow’s World. First Results from PISA 2003,” Paris. OECD (2005) “Education at a glance 2005,” Paris. Psacharopoulos, G. (2003) “The economics of schooling and school quality,” in Labor markets, distribution and growth 1, Elgar Reference Collection, 54–72. Rothstein, J. (2005) “Does competition among public schools benefit students and taxpayers? A comment on Hoxby (2000),” NBER Working Paper no. 11215. Rouse, C. (1998) “Private school vouchers and student achievement: an evaluation of the Milwaukee parental choice program,” Quarterly Journal of Economics, May, 553–602. Saint-Paul, G. and T. Verdier (1993) “Education, democracy, and growth,” Journal of Development Economics 42(2), 399–407. Schuetz, G., H. Ursprung and L. Woessman (2005) “Education policy and equality of opportunity,” CESifo Working Paper 1518, Munich, CESifo. Spence A. M. (1973) “Job Market Signalling,” Quarterly Journal of Economics 87, 355–74. Wasmer, E., P. Fredriksson, A. Lamo, J. Messina and G. Peri (2005) The Macroeconomics of Education, report presented at Fondazione Rodolfo De Benedetti, Venice, June 2005. Woessmann, L. (2003) “Central Exit Exams and Student Achievement: International Evidence.” In: P.E. Peterson, M.R. West, eds, No Child Left Behind? The Politics and Practice of School Accountability, Washington, DC, Brookings Institution Press, 292–323. Woessman, L. (2005) “The Effect Heterogeneity of Central Exams: Evidence from TIMSS, TIMSS-Repeat and PISA,” Education Economics 13(2), 143–69. Woessman, L. and M. West (2005) “Class-size effects in school systems around the world: Evidence from between-grade variation in TIMSS,” European Economic Review 50(3), 695–736.
9780230_547018_15_cha13.indd 468
3/19/2009 3:55:46 PM
14 Mergers and Competition Policy in Europe
14.1
Introduction
Merger activity is gathering pace in Europe. 2005 saw large value mergers or acquisitions such as Italy’s Unicredito of Germany’s HVB in the banking industry and France’s Pernod Ricard of the UK’s Allied Domecq in the food and drink sector. The pace of activity in utilities has been especially hectic: France’s Suez acquired Belgium’s Electrabel, France Telecom bought Spain’s Amena and Telefónica (Spain) has launched a bid for O2 (UK). Within Spain, Gas Natural has also announced its intention to take over Endesa. Private equity firms (mostly British and American) have been active, especially in the profitable restructuring of conglomerates. Not so long ago mergers were basically an Anglo-Saxon phenomenon, but now they are a European phenomenon. Cross-border mergers are an increasing proportion of the total, and activity within the EU-15 is now the most important component of this trend. This reflects the long-term effects of market integration in Europe. But broader trends in the world economy are also important – the revolution in information technology, the widening of markets through globalisation, the strength of corporate profits and the availability of cheap credit. Globalisation, especially in the form of competition from emerging economies, like China and India, has induced restructuring and redeployment to increase productivity, and mergers are an integral part of such a process. Mergers raise many public policy issues. It is not clear that mergers always create value for either shareholders or customers. Consolidation poses a threat to competition, the main driver of efficiency and productivity growth, and domestic competition is the best school for international competitiveness. Domestic mergers are generally more threatening to competition than cross-border ones, and it may be agreed that globalisation lessens the need for merger control. But it is important to establish that European merger control is up to the task of ensuring that the merger wave is beneficial to consumers as well as to investment bankers. 469
9780230_547018_16_cha14.indd 469
3/19/2009 3:56:04 PM
470
Designing the European Model
Many European governments have a protectionist instinct and view with suspicion the foreign takeover of their national champions. They also tend to give special attention to what are considered “strategic” sectors, such as banking and utilities. These pressures are particularly strong in France and Italy, as the discussion over whether French Danone could be taken over by PepsiCo, and the obstacles put by the (former) governor of the Bank of Italy to the foreign takeover of Antonveneta and BNL show. France has issued a list of strategic sectors where national interests have to be protected (although it seems that yoghurt finally has not been included in the list).1 These actions have not, however, halted the steady rise in cross-border mergers; and, somewhat paradoxically, acquisitions by French public or semi-public companies (for example, by France Telecom and EDF) have been particularly extensive. There are two closely related public policy questions in this context. Does ownership matter? Does Europe need either national or European champions? In this chapter we will look first at some of the recent evidence on merger activity, focusing on the rationale of mergers and evaluation of their effects. We go on to survey the evolution of competition policy towards mergers in Europe and the new regulation that the EU has put in place. We will also explore the tension between industrial policy and competition policy.
14.2 Mergers and acquisitions in figures The world has never before seen a boom of mergers and acquisitions (M&A) on the scale of the late 1990s. According to the Thomson Financial Securities Database, 1999 marked the peak of the recent mergers and acquisitions boom with an aggregate value of nearly 4,000 billion dollars. For comparison, in 1989 – the peak of the preceding boom in terms of merger value – the figure reached 747 billion dollars. The peak in the number of deals was reached in 2000 when more than 37,000 were reported (see Figures 14.1 and 14.2). The magnitude of the recent boom means that the rising trend evident in Figure 14.1 disguises the tendency for mergers to occur in waves: the scale of activity at the end of the 1990s dwarfs the peak of ten years before, itself a substantial increase on previous levels. However, the last century has seen merger waves – one at the beginning of the century and then at the end of the 1920s, 1960s, 1980s and 1990s. Until the 1980s, most M&A activity was undertaken by American and British firms. But during the latest wave, mergers and acquisitions also played an increasing role in other industrialised countries, especially in continental Europe. This can (in part) be attributed to the introduction of the Single Market in 1993, but also to the fall of the Iron Curtain and the intensified competition from low-wage countries. The decline in value since 2000 is more accentuated than the decline in the number of deals makes clear. This indicates the extent to which the
9780230_547018_16_cha14.indd 470
3/19/2009 3:56:04 PM
Mergers and Competition Policy in Europe
471
Billion US$ 4000 Global United States EU15 Japan
3500 3000 2500 2000 1500 1000 500 0
81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04
Figure 14.1
Value of M&A deals by region of the sellers, 1981–2004
Source: Thomson Financial Securities Database.
Billion US$
in thousand
40 Global United States EU15 Japan
35 30 25
40 35 30 25
20
20
15
15
10
10
5
5 0
0 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04
Figure 14.2
Number of M&A deals by region of the sellers 1981–2004
Source: Thomson Financial Securities Database.
recent boom featured so-called mega-deals, defined as transactions with a value of more than one billion dollars. According to the World Investment Report 2004 these transactions made up 40 per cent of the entire crossborder M&A value in 1987, whereas its share rose to 75 per cent in 2000. Prominent examples of such mega-deals have been mentioned in the introduction. Other examples include the purchase of Italy’s Banca Antonveneta by the Dutch ABN Amro and the German deal between Viterra and Deutsche Annington Immobilien (IBO).
9780230_547018_16_cha14.indd 471
3/19/2009 3:56:04 PM
472
Designing the European Model
As the number of international transactions demonstrates, cross-border M&A has gained importance and contributed more than proportionately to the increase in overall merger value. In 1986, 17.6 per cent of all M&A value was incorporated in cross-border transactions, the share rose to 39.9 per cent in 1990 and to 37.2 per cent in 2001.2 Figure 14.3 shows that cross-border activity was especially vigorous in EU-15 countries. The exceptional size of the 1999–2000 boom is almost entirely accounted for by this explosion in activity within the EU. US companies were also major players in mergers and acquisitions, but the cyclical upswing there was much less pronounced, as was also true in Japan (where the scale of M&A activity is much smaller). Within Europe, the United Kingdom has always been the most active purchaser of foreign firms, as shown in the lower part of Figure 14.3. France was next, with purchases rising from a value of 21.8 billion dollars in 1990
by region of purchaser
1200
Billion US$ World EU15 United States Japan
1000 800 600 400 200 0
400
87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04
Billion US$
by country of purchaser
United Kingdom France
300
Germany Netherland
200
Spain
100
0 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 Figure 14.3
Value of cross-border M&A purchases
Source: UNCTAD (2005).
9780230_547018_16_cha14.indd 472
3/19/2009 3:56:05 PM
Mergers and Competition Policy in Europe
473
to 168.7 billion dollars in 2000. In 2005, France even became the leading player with an aggregate deal value of 59.5 billion dollars. German companies only started to purchase foreign firms in the recent merger wave: 1999 was the peak year so far with an aggregated deal value of 85.5 billion dollars. The Netherlands and Spain are the next largest acquirers. The growth of European merger activity means that the US is now a net seller of firms and Europe a net buyer: in 2000, European firms were sold, to the value of $587 billion, but purchases by European firms totalled $802 billion. The equivalent figures for the United States are $324 billion and $159 billion respectively.3 China basically only plays a role as a target region of M&A. In 2001, M&A sales reached a local peak of 10.6 billion dollars. The purchases, however, peaked in 2001 with only 1.6 billion dollars. Surprisingly for its size, Japan literally does not participate in international cross-border merger activity. Even in the boom year 2000 Japanese companies undertook M&A deals of “only” 20.8 billion dollars. For comparison, the US reached a value of 159 billion and the UK one of 382 billion dollars.4 The new EU member states can be neglected as purchasers, but they are important as a target region. In these countries merger activity increased from literally zero in 1989 to a total value of about 18 billion dollars in 2001.5 Hungary, Poland and the Czech Republic are the major target countries, with a combined deal value of almost 13 billion dollars in 2001, while the entire group of new members reached a level of 14.7 billion dollars. This makes up 88 per cent of all cross-border M&A sales in these three countries. The development was almost entirely driven by foreign investors, especially in the eight Eastern European states which became recipients of flows of foreign capital after the Iron Curtain was lifted. In 1990, 100 per cent of all merger value there was classified as cross-border activity. Thereafter, this share decreased to 64 per cent in 2003. However, the trend pointed upwards again in 2004. The share of cross-border M&A is still at a very high level compared to more developed economies. Hostile activity has declined since the 1980s. In 1985, about 30 per cent of the value (although a much smaller proportion of the number) of acquisitions were hostile,6 and this figure was even close to 50 per cent within the EU-15; the share of contested bids had fallen to 8 per cent in 2004 (there is a spike in 1999 due to the takeover of Mannesmann by Vodafone for $202.8 billion, a deal so large that it significantly distorts all figures for that year7 (see Figure 14.4). Among cross-border transactions, Evenett (2003) points out that the service sector has played a more prominent role in M&A activity in recent years. While in 1990 about 62 per cent of the value of mergers was in the manufacturing sector (with only 35 per cent in the tertiary sector), this ratio was reversed ten years later. In the recent wave, the proportion of horizontal mergers has increased: Andrade et al. (2001) report that nearly 50 per cent
9780230_547018_16_cha14.indd 473
3/19/2009 3:56:05 PM
474
Designing the European Model %
%
50
50
40
40 United States
30
30 EU15
20
20 Global
10
10
0
0 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04
Figure 14.4 Note:
1
Value share of hostile M&A sales by region of the seller
Cross-border and non-cross border.
Source: Thomson Financial Securities Database.
of all mergers in the US take place between firms from the same industry. But the industries within which such activity has been concentrated have changed substantially over time (see Andrade et al. (2001)).8 Although the recent trend of M&A underlines the increased global perspective of firms by both a higher value of mergers and a wider participation of countries, M&A activity is still concentrated in a few regions. The US and the UK are still major players if one looks at total value of M&A deals. However, continental European countries have caught up in recent years and even outperformed the US and the UK in cross-border M&A recently.
14.3 The rationale for mergers Globalisation is associated with technological change (particularly in information technology), with decreases in trade and transport costs (in goods, capital, people and information), and with liberalisation and market integration, which simultaneously enlarge the market and increase competitive pressure. Productive systems require revision to accommodate these changes. In many sectors the number of firms will have to be reduced in an integrated or enlarged market in order to reap economies of scale. In many industries, in particular in those subject to network externalities and learning curves, activities that demand or give advantage to scale, such as investment in R&D, innovation and securing a large customer base, become central to competitive strategy. Mergers are a prime instrument of industrial restructuring.
9780230_547018_16_cha14.indd 474
3/19/2009 3:56:05 PM
Mergers and Competition Policy in Europe
475
For example, in banking there is a move from the traditional business of taking deposits and granting loans to the provision of services to investors (investment funds, advice and insurance) and to firms (consulting, insurance, advice and preparation of M&A, underwriting of equity and debt issues, and risk management). As a source of revenue, the margin – that is the difference between borrowing and lending rates – makes way for fees and commissions, and investment in “bricks and mortar” (the branches) to investment in communication networks, information technology, and highly specialised human capital.9 This change may mean that size is important, especially in wholesale and investment banking. The outcome of this process is an overhaul of the banking sector that, in general, is more advanced in the US than in Europe. Merger and acquisition activity may enhance profits by improving the efficiency of firms or increasing their market power. Firms may merge to obtain synergies, exploit scale economies and rationalise production, or to reduce managerial inefficiency by taking over a poorly run firm. If capital markets are imperfect, it may pay also to merge to gain financial muscle and to diversify to undertake major and/or risky investments. This may apply in particular to important R&D projects. But a horizontal merger may also increase the market power of the merging firms, enabling them to raise prices. A vertical merger may have efficiency benefits. It may eliminate the “double margin” which arises when two successive firms in the chain of production each have market power: it may also enable complementary assets to be brought under single control. Vertical mergers may, however, also raise rivals’ costs or foreclose a competitor. This may be the case, for example, if the merged entity controls some essential input for downstream suppliers. There may also be other motives for merger that need not increase profits or enhance shareholder value. These result from the ambitions of managers and inefficiencies in the relationship between them and the owners of the firm. Managers may seek higher status by promoting size rather than profitability, may protect their private benefits of control with entrenchment strategies, and may engage in empire building. Size may lead to inefficiency as management becomes bureaucratic and the loss of control by owners leads to agency problems. A typical defensive strategy to keep a firm independent is to engage in unprofitable mergers to increase firm size and make a takeover by another company less likely. Mergers may be the consequence of hubris: managers are overconfident and assess the potential value of a merger incorrectly, thus undertaking unprofitable mergers (Roll 1986).10 Regulation, and the relationship between economics and political structures, may also provide a motive for merger. A large firm may be “too big to fail”; this is typical of the banking business but applies to other industries as well. It may have a larger capacity to influence regulation or more capacity to obtain subsidies through lobbying and political connections. For international firms, size may be important to obtain government protection of
9780230_547018_16_cha14.indd 475
3/19/2009 3:56:05 PM
476
Designing the European Model
their interests abroad. These arguments will tend to have more weight for domestic mergers than for cross-border mergers. Almost all these reasons for mergers – good and bad from the perspective of shareholders, desirable and undesirable from the perspective of public policy – can be seen in the banking industry. Size offers the possibility of exploiting scale economies in administrative and back-office operations, information technology, and in investment banking type operations (related to information gathering and fund management). Size may help in realising scope economies (through combining different product lines because, for instance, it increases the value of customer relationships and decreases average marketing costs). Consolidation may deliver these advantages, eliminating excess capacity in the branch network when the networks of the merging banks overlap, and improving diversification, particularly if the banks operate in regions with non-synchronised cycles. Furthermore, consolidation may provide a way to cut excess labour and to access the mass retail market in a foreign country. As we have seen, mergers tend to happen in waves. Only recently have theoretical models attempted to account for this phenomenon. The explanations involve the exploitation of market inefficiencies (Shleifer and Vishny 2003; Rhodes-Kropf and Viswanathan 2004), the desire of managers to preserve their independence (Gorton et al. 2005), cyclical phenomena (Lambrecht 2004), strategic considerations (Faulí-Oller 2000 and Toxvaerd 2004) or capital reallocation due to technological shocks (Jovanovic and Rousseau 2002, 2003). Although all these factors may help explain merger waves, the argument based on technological shocks seems the most compelling. The different theoretical models are described in Box 14.1. What are the consequences for concentration? Globalisation and market integration imply effective enlargement of the market, which has consequences for the equilibrium level of concentration in a market. In an industry characterised by the presence of a fixed and sunk cost of entry, concentration decreases as the ratio of the market size to the sunk cost of entry increases. For example, when markets with the same number of firms become integrated, the total number of firms in the free entry equilibrium falls, but to a number which is larger than the initial number of firms in either of the original markets. Concentration in the integrated market is thus lower than in any of the original markets. However, in industries in which the sunk cost is endogenous – that is controlled by the firm, as in investment in R&D or in advertising or other expenditures not related to output but designed to reduce costs, boost demand, or to improve the quality of the services offered – an increase in market size need not lower concentration (Sutton 1991). For example, in banking the cost of establishing a branch network is a fixed and sunk cost, and the transformation of banking towards a service industry may increase the sunk cost required from the bank through
9780230_547018_16_cha14.indd 476
3/19/2009 3:56:05 PM
Mergers and Competition Policy in Europe
Box 14.1
477
Why mergers happen in waves
Exploitation of market inefficiencies. In Shleifer and Vishny (2003), stock-financed mergers are driven by relative market valuations of merging firms. Stock market inefficiencies lead to misvaluation. Rational managers exploit these inefficiencies, in part through mergers, and takeovers are more likely when misvaluations are high. In a similar vein Rhodes-Kropf and Viswanathan (2004) construct a rational model of stock mergers with false market valuations. While managers of the bidder have private information about both the stand-alone value of their own firm and the potential value of the merged firms, target firms’ managers are only aware of the stand-alone value of their firm. To assess the synergies, they have to rely on market valuation through the bids they receive. This is the crucial mechanism of the model that generates merger waves. Since there is a market-wide and a firm-specific effect of the misvaluation, the target tries to filter out the marketwide effect. When the market-wide overvaluation is high, the estimation error of the synergy is also high. Hence, bids appear much more attractive in overvalued than undervalued markets. Empirical evidence for the impact of stock market valuation on M&A activity can be found in Ang and Cheng (2003), Dong et al. (2004) and Rhodes-Kropf et al. (2005). Manager independence. Gorton et al. (2005) explain merger waves by the managerial interest in keeping the firm independent. Since larger size decreases the probability of acquisition by another firm, protection against mergers may be pursued at the expense of profitability, and a potentially efficient merger within one industry may result in a merger wave with unprofitable mergers. The first merger sets off a chain reaction; as the authors put it, the defensive actions of all companies result in a “race for firm size”. Cyclical phenomena. Lambrecht (2004) provides a model to explain why mergers occur more frequently in times of economic expansion than recession. Takeovers promise the exploitation of scale economies, which are correlated with the market demand for the firm’s product and hence its market price. Given that mergers always involve costs, transactions become more likely if product prices, and hence the benefits of the merger, rise. Consequently, product markets that show cyclical behaviour are more likely to be subject to merger waves. Strategic considerations. Toxvaerd (2004) considers a finite number of acquiring firms competing for a scarce number of target firms. Each acquiring firm can either undertake the transaction immediately or wait for better market conditions. This, however, involves the risk of being pre-empted by competitors. The author shows that, in equilibrium, acquirers undertake their transactions simultaneously, thus generating merger waves. The intuition for this result is that while waiting to merge is only optimal when other firms wait as well, there is a risk of not finding a target firm to acquire when other companies have already acted. Thus, a merger wave is anticipated and all firms buy their predetermined target at the same time. Non-strategic and strategic considerations. Faulí-Oller (2000), in a Cournot model with cost asymmetries, combines both non-strategic and strategic explanations for merger waves. The former comprise exogenous determinants that make takeovers profitable. The latter include interrelations among firms making mergers
9780230_547018_16_cha14.indd 477
3/19/2009 3:56:05 PM
478
Designing the European Model
profitable only if other firms merge as well. The author shows that an exogenous reduction in demand stimulates merger activity as cost asymmetries are thereby accentuated and mergers become more profitable. The strategic effect is that non-participating firms create a negative effect on the merger since they respond to it by increasing their output. The fewer competitors there are, the less intense is this effect. Under certain circumstances, an initial merger can thereby make further mergers more attractive.a Capital reallocation due to technological shocks. According to the theory of Jovanovic and Rousseau (2004), merger waves coincide with epochs of technological change, for example the spread of electricity in the period 1890–1930 or of information technology between 1970 and 2002. Technological progress forces firms to restructure and make their production process more efficient. As some firms do better than others, technological change naturally generates a higher dispersion between the efficiency of different firms. The analysis relies on Tobin’s q as a measure of the market’s evaluation of the relative efficiency of a firm. Following this reasoning, a technological shock increases the q for successful firms and lowers it for less successful ones. Hence, high-q firms will purchase low-q firms and thus ensure a better allocation of capital. Using US stock market data, this straightforward theory provides a reasonable fit for four out of five merger waves that occurred in the twentieth century.b Mergers and acquisitions have become a much more important device for capital allocation relative to ordinary entry and exit of firms in the market. a Faulí-Oller explains merger activity by a reduction in demand (in declining industries and refers to Dutz 1989 for examples), while Lambrecht (2004) points out that merger activity is positively related with booms (like the last wave which peaked in 2000). There need to be no contradiction since an overall merger wave can rather be explained by Lambrecht’s view, whereas it may be true that there is some need for restructuring in declining branches (which would probably be too small to create a merger wave). b In an earlier study Jovanovic and Rousseau (2002) demonstrate that there is a clear empirical relationship between the reallocation of capital via M&A and the dispersion of q among companies. It is shown that the response of investment through M&A to a change in the dispersion of q is 2.6 times higher than the response of regular investment.
its investment in communication networks/information technology or specialised human capital. We might envision competition proceeding in three stages: (1) entry decisions which demand certain expenditure to be present in a market at all; (2) investment decisions which imply fixed costs, as in R&D, advertising, and information technology; and (3) following these preparatory expenditures, competition in the marketplace. Under these conditions increasing the size of the market does not generate more entry in equilibrium, in fact it may generate exit, because competition at the investment stage is very fierce. The required circumstances are that the fixed expenditure in (2) must loom large in relation to the variable one at the production and market stage and that the market share must be sufficiently sensitive to the investment effort. A larger market leads only to increased expenditures by a few firms and there is typically an upper bound to the number of active
9780230_547018_16_cha14.indd 478
3/19/2009 3:56:06 PM
Mergers and Competition Policy in Europe
479
firms in the market (no matter how large that market) and we may speak of a natural oligopoly (see Sutton 1991; Schmalensee 1992). It is an empirical question to what degree sunk costs in a particular industry are “endogenous” in the sense described above. It is likely, however, that the increased importance of investment in information technology, in information acquisition, in building a customer base in markets with network externalities and/or learning curves has increased the importance of endogenous sunk costs. This means that fixed expenditures are now larger relative to variable ones, and that the market share has increased its sensitivity to investment in fixed costs. If this is so, then it may be true that in some industries in the global market place there is only room for a few players. In financial services this might apply to wholesale and investment banking (providing services – underwriting, trading, brokerage, rating, and advice and preparation of M&A – to the top tier of multinational corporations and medium-sized firms with international operations).
14.4
Merger performance
14.4.1 Competitive impact A merger will affect the profits of the merged entity (the insiders), its rivals (outsiders) and will influence the prices, quality and variety of products available to consumers. Welfare assessment of a merger is thus complex. When mergers increase the profits of the merging firms, there are generally potentially two opposing effects: welfare losses from a reduction in effective competition and welfare gains from scale economies and unit cost reductions. Increased concentration will tend to enhance welfare overall only if the merged firm gains market share through lower prices. In the absence of cost reduction effects, a merger that is profitable for insiders tends to be profitable for outsiders also because it tends to raise price and restrict output. Consumers are typically hurt. If firms are symmetric – that is of similar sizes – then the effect is a reduction in total welfare (the sum of producer and consumer surplus). However, if firms are asymmetric, then a merger of small firms may improve productive efficiency (by making the capacities of the firms in the market more equal and shifting production from high-cost to low-cost firms) and profits so much that total welfare may rise even though prices and concentration increase.11 In a study of mergers around the world based on the global Thomson Financial Securities Database for the period 1981 to 1998 (which covers all transactions of at least 1 million dollars in size), Gugler et al. (2003) found that on average mergers do increase profits but reduce sales of the merging firms. The authors did not find large differences between geographic areas, manufacturing or services, or between domestic and cross-border operations. Conglomerate mergers decrease sales more than horizontal mergers.
9780230_547018_16_cha14.indd 479
3/19/2009 3:56:06 PM
480 Designing the European Model
The authors differentiate four categories of merger: Market power effects (27.6 per cent) is when sales are reduced and profits increase. Efficiency improvement (29.0 per cent) occurs when both sales and profits rise. Efficiency decline (28.2 per cent) is where both sales and profits fall, while a fourth category involving higher sales but lower profits (15.1 per cent) may be explained by the pursuit of size at the expense of profits. The study reveals a rich variety of outcomes consistent with a similar variety of motives for merger. Before concluding that a majority of mergers in the 15-year period examined were welfare-reducing, as the authors do, a full welfare analysis should consider effects on rivals and consumers. Pesendorfer (2003) does this for horizontal mergers in the US paper industry in the mid-1980s. He found that both efficiency and welfare increased after this industry-specific M&A wave. Merged firms reduced capacity and generally lost market shares with an overall positive welfare effect as a result since the mergers increased producer surplus without affecting consumer surplus. Maksimovic and Phillips (2001) provide empirical evidence that the majority of firm asset transactions generate productivity gains and better allocative efficiency in a sample of around ten thousand transactions. Using data from the Longitudinal Research Database (LRD) of the US Bureau of the Census for the years 1974 to 1992, the authors calculate total factor productivity and compare that measure at the firm level one year before a transaction with two years after. The mean industry-adjusted change in productivity is significantly positive with a 2 per cent increase. During the 1990s, many national and cross-border mergers took place in the banking sector, complemented by strategic alliances and joint ventures that led to a consolidation of the industry. Evenett (2003) seeks to distinguish efficiency and market power effects by analysis of the interest rate spread (the difference between the average interest rate paid by borrowers and the average interest rate the bank pays to depositors). An increase in the interest rate spread indicates that banks gained more market power and did not pass lower costs due to efficiency gains onto customers. Using data from the World Bank and the Bank for International Settlements, Evenett finds mixed results. Cross-border strategic alliances within the EU seemed to be associated with an increase in the interest rate spread, whereas intra-EU cross-border mergers reduced this spread. Outside the EU, both strategic alliances and M&A have decreased interest rate spreads, pointing to positive welfare effects. Another interesting stylised fact is that mergers between asymmetric firms (for example, a large and a small firm) tend to fare better than those among more symmetric firms (see Capron 1999 and Conn et al. 2003). 14.4.2
Winners and losers
Andrade et al. (2001) seek to identify winners and losers of M&A between publicly traded US firms in the period 1973–1998.12 The impact of mergers
9780230_547018_16_cha14.indd 480
3/19/2009 3:56:06 PM
Mergers and Competition Policy in Europe
481
on stock prices is analysed by employing short-window event studies that monitor stock market reactions one day before the announcement of the merger to one day after. In a second step, this window is also extended to 20 days prior to the announcement to the completion of the merger. The basic result is that owners of target firms’ stock clearly benefit from the merger. The value creation amounts to 16 per cent, on average. On the other hand, the impact of the merger on stock owners of the acquiring company is not clear cut (on average, stock value is reduced by 0.7 per cent; however, this result is insignificant). Combining both the target and acquiring firm effect, stockholders on average gained 2.6 per cent in the 1980s and 1.4 per cent in the 1990s. If markets are sufficiently well informed about the effects of the merger, these increases in value would reflect the anticipated returns. However, the high degree of uncertainty about estimates of expected returns incorporated in stock prices must be translated into a similarly high degree of uncertainty in the estimation of abnormal returns. The empirical evidence for an important fraction of mergers seems to be inconsistent. For those mergers, event studies find that the stock market value of the merged firms increases, while profits decline. This leaves three puzzles: (1) Why do unprofitable mergers occur? (2) How can the value of the merged firms increase while profits decrease? and (3) Why do firms acquire other firms if the gain is almost wholly derived by stockholders of the target company? Fridolfsson and Stennek (2001) try to resolve the puzzles by proposing a single explanation. In a coalition bargaining model, the authors show that pre-emptive mergers are rational if it is disadvantageous to become an outsider, that is to be excluded from the merger wave. The negative externality imposed on non-participating firms creates an incentive to engage in pre-emptive mergers. Although the merger results in lower profits, the value might increase since the pre-merger stock value takes into account the risk of becoming an outsider, an even worse result for profits, which is thereby averted. 14.4.3 Mergers and R&D The dynamic consequences of mergers on innovation are among the most important effects. Innovation is the engine of growth and it may well be that a (negative) static welfare effect is overturned by a (positive) dynamic merger-induced efficiency effect or vice versa. Mergers might have scope and scale effects in R&D, generating sharper incentives to engage in that activity (for example, a firm producing a larger output will have more incentive to invest in cost reduction, as in Vives 2004). Also, positive spillover effects visà-vis related firms can be internalised, bringing social and private returns more closely in line. Mergers may also provide a firm with more financial resources and this may allow it to undertake riskier and more ambitious R&D projects in the presence of capital market imperfections. At the same time,
9780230_547018_16_cha14.indd 481
3/19/2009 3:56:06 PM
482
Designing the European Model
mergers may avoid the duplication of effort and, in fact, reduce R&D effort while maintaining R&D output. On the other hand, mergers may soften competition, and the empirical literature as well as some theoretical models have concluded that a degree of competition is needed for innovation to occur and that competitive pressure tends to foster innovation (see Baily and Gersbach 1995; Nickell 1996; Aghion et al. 2005 and Vives 2004). The strength and net impact of these effects is an unresolved empirical question. The evidence available on the R&D performance of mergers is fragmentary and mixed. Note, however, that a result that indicates that R&D effort is reduced by mergers could be compatible with welfare-enhancing operations, since the elimination of duplication efforts may be good. Several studies focus on the effect of mergers on R&D investment. In a study for the biotech and pharmaceutical industries, Danzon et al. (2004) show that large firms in this sector often merge to eliminate overcapacity, which may emerge on the expiry of patents. Controlling for the propensity to merge, the authors find that there is no difference between R&D expenses of merged firms compared to similar firms in the same industry that did not merge. Smaller firms in this industry often see mergers as an exit strategy when under financial pressure. Cassiman et al. (2004) argue that technological and market relatedness should determine the impact of a merger on the level of R&D. Firms that use complementary technologies should increase their joint level of R&D after merger, whereas firms with substitutive technologies might be expected to reduce it. Aggregation of product market activities can yield economies of scale and scope with indirect consequences for R&D expenditures. In a small sample of 31 EU mergers, Cassiman et al. (2004) find that R&D levels rise where technologies are complementary and fall when they are substitutive. Where technologies are substitutive, R&D reductions are larger where firms are product market rivals. Overall the authors find empirical evidence for the scope effect, whereas the scale effect does not appear in the data. The semiconductor industry ranks high in R&D spending (amounting to 13 per cent of sales). Gugler and Siebert (2004) looked at efficiency versus market power effects in this industry and compared mergers with research joint ventures (RJV). RJVs should also be capable of internalising positive spillover effects yielding a higher R&D level. Indeed, Gugler and Siebert (2004) find that RJVs generate higher welfare gains since the market power effect, which could potentially decrease the efficiency gain effect, is lessened. Hence, from a public policy perspective, such RJVs offer the advantage of not reducing competition on the product market. 14.4.4 Summary The assessment of the performance of mergers is complex and there is evidence consistent with different explanations of their origin. On balance, however, the most plausible explanation is that mergers respond
9780230_547018_16_cha14.indd 482
3/19/2009 3:56:06 PM
Mergers and Competition Policy in Europe
483
to technological shocks and are an instrument of restructuring, and this explains merger waves. Other motives are superimposed on this central influence of technology: to gain market power, to benefit managers or managerial overconfidence. These latter factors explain why merger waves may overshoot and some mergers may destroy value.
14.5 Competition policy and merger control Competition policy has as its main goal the protection of social welfare by maintaining a healthy competitive process. There is a debate over whether competition policy should be directed only to the consumer interest (that is consumer surplus) or should encompass also profit or producer surplus (often described as a total surplus standard). In any case, competition policy is today directed towards economic efficiency. In the US, it took some time to get to this point, overcoming populist attitudes according to which mere size was an offence. Now only market power is. In Europe, the efficiency objective has been intertwined with other goals like the promotion of small and medium-sized firms, innovation and the external competitiveness of European firms. An added objective has been to promote European economic integration. Unique among competition policy authorities, the European Commission has the duty of monitoring industrial rationalisation programmes and state aid to industry. All in all, however, the efficiency objective has gained weight. The foundation for competition policy is that competitive pressure is the guiding force towards economic efficiency. There are general arguments in favour of competition that, in principle, apply to any industry. Indeed, the benefits of competition for allocative efficiency are well established since Adam Smith. It must be noted, however, that competition is in general imperfect because of entry barriers, switching costs, product differentiation and asymmetric information. The result is that there is room for firms to exercise market power. A consequence is that the welfare theorems associated with perfect competition are not directly applicable to any real industry. All in all, however, competition is perceived to be good for both allocative and productive efficiency. The pressure of a competitive market provides incentives to managers to perform and information to design appropriate incentive schemes.13 Monopoly power induces inefficiency and waste, and a healthy degree of rivalry is necessary to keep a vigorous pace of innovation in an industry, that is, for dynamic efficiency. Competition policy tries to prevent adverse consequences of market power by controlling ex post restrictive practices (like price-fixing or market-sharing arrangements in cartels) and abusive practices (like attempts to monopolise or exclude rivals from the market). It also seeks to control merger activity ex ante: preserving market structures conducive to competition or preventing market structures that preclude effective competition.
9780230_547018_16_cha14.indd 483
3/19/2009 3:56:06 PM
484 Designing the European Model
Merger control tries therefore to anticipate the consequences for competition of the restructuring induced by a merger. Merger analysis distinguishes between the “unilateral effects” and the “co-ordinated effects” of a merger. The unilateral effects reflect the consequences of raising prices above costs and are measured with the standard oligopoly-pricing static models of competition among the few. Co-ordinated effects arise when firms agree explicitly, or implicitly via reward and punishment strategies, to keep prices high. Co-ordinated effects reflect actual or tacit collusion. An analysis of entry conditions, potential efficiency gains, and dynamic effects (on investment and innovation for example) is also relevant. The procedure in the US and the EU is similar. It starts by defining the relevant (geographic and product) market and proceeds to compute market share and concentration indicators (like the Herfindahl index).14 Those are used to define safe harbours, basically stating that mergers with small combined market shares should be allowed to proceed, in particular in unconcentrated industries. There is a major difference between the EU and the US. In the EU, the jurisdiction that should deal with the merger must be established: the European Commission or a national authority.15 The allocation of jurisdiction may be contentious as the recent Spanish Gas Natural–Endesa case shows, with the Commission taking more than two months to decide. The definition of the relevant market, in product or geographic space, is crucial and typically contentious. For example, in the blocked proposed merger of the two Swedish truck manufacturers, Volvo and Scania, the European Commission concluded that each of the individual countries (such as Sweden, Denmark, Norway, Finland, and Ireland) was a market within which the merged company could exercise excessive market power. However, the merging parties argued in favour of the European market being the relevant one (the European Economic Area). If the relevant market had a European rather than a national dimension, the decision would have been different. This case makes clear that companies that seek to gain size to compete internationally may be prevented from merging if this is considered to raise concentration too much in a national market. A rise of concentration due to the merger must then be checked against a specific analysis of unilateral and co-ordinated effects and entry conditions. To check for unilateral effects, quantitative and simulation techniques, based on oligopoly models, are increasingly used. The analysis of co-ordinated effects is more qualitative, relying on market structure conditions (like market transparency, asset distribution in the industry in terms of capacities of production and product portfolios of the firms, concentration and number of firms, multi-market contact, asymmetries in cost and demand, entry conditions and buyer power) and facilitating practices (like a history of cooperation in the industry, communication of plans, exchange of information on prices and quantities, and pricing policies) that may impinge on the capacity of firms to collude in sustaining prices above a competitive level.
9780230_547018_16_cha14.indd 484
3/19/2009 3:56:07 PM
Mergers and Competition Policy in Europe
485
In the US and the UK, a merger would be challenged if it “substantially lessens competition.” The substantive test in the EU under the old merger regulation (from 1989) was that: “A concentration which does not create or strengthen a dominant position as a result of which competition would be significantly impeded in the common market ... shall be declared compatible with the common market.” This is a “twoprong” test referring to the “creation or strengthening of a dominant position” and checking whether “competition is significantly impeded.” The problem was that this test is not well adapted to deal with unilateral effects because a merger may raise prices and diminish welfare even though no dominant position for a single firm is created. The Commission then tried to block mergers because they would create a collective dominant position (that is foster collusion) when in fact what potentially was at stake were unilateral effects (that is non-collusive exercise of market power), which could not be challenged using the old regulation. An example is provided by the blocking in 1999 of the merger of Airtours and First Choice, because it would have created a collective, dominant position in UK short-haul foreign package holidays. However, in 2002 the Court of First Instance (CFI) concluded that the Commission had made errors of assessment and had not proved to the requisite legal standard that the merger would give rise to collective dominance (that is the collusive potential according to the co-ordinated effects analysis). Under the old merger regulation, there was a period of vigorous intervention at the end of the 1990s. Prohibited or abandoned transactions tripled from an average of two per year in 1990–1998 to an average of six per year in 1999–2000. From 1998 onwards there was also an increase in transactions subject to structural remedies. However, about half of the prohibition decisions of the Commission have been challenged by notifying parties, including, for example, Gencor/Lonrho, Airtours/First Choice, Worldcom/MCI Sprint and GE/Honeywell, and in 2002 the CFI overturned the Commission’s decision in Airtours/First Choice (as discussed above), Schneider-Legrand and Tetra Laval Sidel (see Table 14.1). Duso et al. (2003) have claimed that a stock-market event study can in principle detect the welfare impact of mergers. The authors reached their conclusions based on the stated insight that in some mainstream oligopoly models with substitute products,16 consumer surplus increases if the profits of outsider firms, that is firms that are not involved in the horizontal merger, decline. A merger is pro-competitive if and only if it decreases the value of rivals. It must be noted, however, that this only applies to horizontal mergers, and several of the mergers considered have either vertical or complementary market components. Furthermore, a decline in the stock market valuation of rivals as an outcome of the merger announcement may also come about if the market anticipates predation and exclusionary strategies against outsiders.
9780230_547018_16_cha14.indd 485
3/19/2009 3:56:07 PM
486
Designing the European Model
In 2004 the Commission introduced a new merger regulation with a reformed substantive test, a strengthening of the parties’ procedural rights, and internal controls and investigative powers of the Commission. The procedure in the EU differs from that in the US, which is of an adversarial nature: antitrust agencies must challenge a merger in court. In the EU the procedure is administrative: a merger of a European dimension has to be notified to the Directorate General of Competition, which examines it and makes a recommendation after which the Council of Commissioners takes the final decision on approval. The new procedure in the EU adds check and balances and establishes the appointment of a Chief Competition Economist to enhance economic analysis and the creation of a Scrutiny Panel, which will review cases that go over a second phase of investigation and report to the Director General.17 The new merger regulation of 2004 introduces a new substantive test to assess the anticompetitive impact of concentration: “A concentration which would not significantly impede effective competition on the common market ... shall be declared compatible with the common market.” In this Table 14.1
M&As formally blocked by EU
Year
Deal
1991 1994 1995
Aerospatiale/Alenia bid for de Havilland (Canada) Bertelsmann, Kirch, Deutsche Telecom MSG deal (digital pay TV) Dutch Holland Media Group venture between RTL4, Veronica and Endemol Nordic satellite distribution joint venture between Norsk Telecom, TeleDanmark and Kinnevik Saint-Gobain and Wacker-Chemie silicon carbide joint venture Finnish retail deal between Kesko and Tuko Blokker’s acquisition of Dutch operations of Toys’R’Us Proposed digital TV alliance of CLT-UFA and Kirch Acquisition by Deutsche Telecom and CLT-UFA of stake in Kirch’s BetaResearch (decoders for pay TV) Airtours’ bid for tour operator First Choice (travel agencies), (turned down by CFI, 2002) MCI Worldcom & Sprint/USA (internet access) Volvo and Scania (cars and trucks) GE – Honeywell (appeal pending to CFI) Scheider – Legrand in electrical equipment (turned down by CFI, 2002) SCA-Mölnlycke & Metsä Tissue (paper) CVC and Lenzing (synthetic fiber) Tetra Laval and Sidel (drink packaging), (turned down by CFI, 2002) ENI –EDP –GDP (energy)
1995 1996 1996 1997 1998 1998 1999 2000 2000 2001 2001 2001 2001 2001 2004
Source: European Commission.
9780230_547018_16_cha14.indd 486
3/19/2009 3:56:07 PM
Mergers and Competition Policy in Europe
487
Significant Impediment of Effective Competition (SIEC) test, the first prong of the old regulation test, “... which does not create or strengthen a dominant position ...”, is disposed of and only the part “effective competition is significantly impeded” is kept. This makes the test closer to the US and UK practice and allows the Commission to deal with market power issues in non-collusive oligopolies. At the same time the Commission has issued horizontal merger guidelines in the style of the US guidelines. In those guidelines, also potential efficiencies of a merger are considered as long as they are beneficial to consumers, merger-specific, and verifiable. When the antitrust agency in the US – be it the Department of Justice or the Federal Trade Commission – decides to challenge a merger, the parties may decide not to pursue the transaction, given that the judicial process may drag on for quite some time. The new substantive test in the EU is closer to economic analysis and less rigid. It allows the elimination of distortions in the use of the concept of collective dominance, which was creating uncertainty in the procedure. The merger guidelines should also help reduce uncertainty for parties contemplating a merger. The changes introduce more checks and balances and provide an enhanced role for economic analysis. However, the Commission staff dedicated to economic analysis is still quite limited in comparison to the US antitrust agencies, and the imbalance of resources relative to the private sector is marked. Still, an open question is whether the new checks and balances will be enough to avoid the cases being overturned by the CFI, which seems to have tightened the standards of proof required from the Commission. Further rejections may imply that the system effectively changes to one of judicial review more similar to the US. In an inquisitorial procedure, the prosecutor/judge may not look for all sides of the argument and seek only reinforcing information. In the CFI decisions on Airtours/First Choice, Tetra Laval/Sidel and Schneider/Legrand, the Court criticised the Commission for being one-sided and suppressing conflicting evidence. When the prosecuting and judging functions are separated, the parties will generate information on all sides of the argument. There is thus a case for a more explicit adversarial procedure. The introduction of a scrutiny panel has gone some way, but it would make sense to go further and consider an internal team that makes the procompetitive case for the merger.18 A further step would be to consider the establishment of an administrative tribunal, independent of the investigators, but still within the Commission, that makes a public recommendation on the merger to the College of Commissioners. The Commissioners might still disagree with the panel’s recommendation but would have to explain why. A further step would be for the administrative tribunal to take the final decision. An even bolder step would be to set up an independent European Competition Agency similar to the US Federal Trade Commission (FTC). At the FTC, the challenge to a merger is brought to the decision of
9780230_547018_16_cha14.indd 487
3/19/2009 3:56:07 PM
488 Designing the European Model
an administrative law judge and the decision can be appealed to the full Commission; if the merging parties are not happy, they can then go to a circuit court of appeals. In the European case the parties could appeal the decision of the European Competition Agency to the CFI and the ECJ (the European Court of Justice).
14.6 Industrial policy and competition policy Industrial policy often conflicts with competition policy. National governments may want to help declining industries (like textile, coal, shipbuilding etc.) or national champions such as Crédit Lyonnais, Bull or MG Rover. Examples abound with declining industries and national champions. This may reflect lobbying efforts of local constituencies or a belief that some sectors (like banking or energy) are strategic and need well-entrenched domestic firms to defend the national interest. This conflicts with competition policy, which limits state aids, and may slow the integration of European markets. 14.6.1
Why ownership matters
Does the national ownership of firms matter? It seems hard to argue that Belgium is worse off because it has no national car producers. However, is the same true of other industries like banking? Is a country worse off if all its banks are foreign-owned? This is close to happening in some developed countries like New Zealand as well as some emerging markets (including new entrants to the EU such as Estonia). The banking example In the financing of domestic economic activity, and in particular in relationship banking like lending to small and medium-sized firms, proximity matters for long-term commitments. Foreign ownership may reduce the commitment of domestic banks to domestic borrowers because distant headquarters may use hard information and rigid protocols instead of soft information and may have less tendency to internalise the welfare of local stakeholders.19 In the US, this has been a concern when large out-of-state banks took over local institutions. However, as the Japanese experience has painfully pointed out, close relationships may result in high costs of finance (because banks cannot exit from transactions), delay the closure of non-viable firms, lead to collusive arrangements that prevent entry, and eliminate healthy competition and innovation from foreign institutions. All in all, in developed economies with a well-diversified range of institutions, national protectionism does not seem warranted even in banking. In emerging economies – as in Eastern Europe where financial integration has been achieved mostly through ownership by West European banks – the alternative to foreign bank ownership
9780230_547018_16_cha14.indd 488
3/19/2009 3:56:07 PM
Mergers and Competition Policy in Europe
489
may be semipublic banks, vulnerable to political influence and soft budget constraints. Most mergers in the banking sector have been domestic. There are obstacles to cross-border mergers in Europe that do not affect cross-state mergers in the US: more limited economies of international diversification, labour market rigidities as well as differences in language, regulation, and corporate culture. Political interference and the fostering of national champions are also prevalent. Recent examples of this include BBVA with first Unicredito and very recently with BNL in Italy, and ABN Amro’s aquisition of Antonveneta. The French authorities adopted a protectionist attitude in the triangular battle BNP-SG-Paribas that ended up in the merger BNPParibas by insisting on a “French” solution to the case. Domestic mergers cut costs by reducing overlap in branches (and overcoming labour market rigidities) and many reap benefits from a financial conglomerate. But such mergers may also increase or maintain market power and prevent hostile takeovers. Cross-border mergers may help to acquire local expertise, access high-margin deposits or diversify, while size is gained to compete in global markets. In fact, cross-border regional mergers, where cultural and legal differences are smaller, took place in late 1990s in the Scandinavian and the Benelux countries. More recently, Banco Santander (Spain) was able to acquire Abbey because the UK does not have a protectionist attitude: the antitrust authority did block the takeover of Abbey by Lloyds TSB in 2001 but not the takeover by Banco Santander. Here a vigorous national competition policy promoted a cross-border merger. Santander comes from an increasingly competitive domestic market in Spain that has induced efficiency gains and allowed international expansion (mostly in Latin America). We have described how in the banking sector the location of headquarters of banks may matter because proximity is important for long-term credit relationships. More generally, the location of headquarters of a firm matters because headquarters create agglomeration effects for both other headquarters and business services. One example of such positive external effects is the density of the market for highly qualified labour. Most regions are therefore prepared to subsidise the location of new headquarters (see Strauss-Kahn and Vives 2005 for evidence on location factors and external effects in the US.) Other advanced activities like R&D also tend to be located close to headquarters. Ownership also matters because proximity is relevant for the protection of the interests of the different stakeholders (such as workers, suppliers, small shareholders and communities) in a firm. In bad times the firm may tend to minimise staff cuts in its country of origin. Large international firms typically develop a corporate culture with a national base and do need the support of a government, in terms of influence activities and protection of property rights, to compete truly internationally
9780230_547018_16_cha14.indd 489
3/19/2009 3:56:08 PM
490
Designing the European Model
(for example, US “multinationals” typically call on the State Department when in trouble). This implies that companies remain “national” for good reasons that may still be important in a globalised world. In summary, ownership matters in so far as it influences the location of corporate control centres and the associated externalities. Local and regional authorities have incentives to retain and attract corporate headquarters. There is a lot of path dependence, that is, history matters, and there may be potential multiple equilibria. This gives room for policy intervention. However, if all regions give subsidies or protect their firms, they may neutralise each other and imply both large budgetary costs and welfare losses. Such protectionist attitudes may be self-defeating. 14.6.2
From national to European champions?
Cross-border mergers may lead to the formation of European champions. Competition concerns should be less in such cases, because of the size of the European market. But a trade-off may exist: extracting rents from abroad against exploiting European consumers. However, those champions may be helped unduly to start with or, perhaps worse, are too big to fail and are still provided with subsidies when they should be closed down. In some sectors where the learning curve is very steep like aerospace, the commitment power that comes with public help may prove crucial in international competition. In this case, according to Neven and Seabright (1995), help to Airbus basically entailed a transfer of rents from US producers Boeing and McDonnell-Douglas to Airbus, leaving the consumers with small gains. Cross-border mergers require the combination of very different corporate cultures: to be successful in the end, one culture has to predominate over the other. Asymmetric mergers or absorption seem to work better than mergers among equals. European competition policy regarding state aid may be effective in checking support to national champions (as with Crédit Lyonnais) and serve as an external commitment to not keep inefficient institutions in business (see Besley and Seabright 1999). However, it is not so clear that it can prevent the support of pan-European champions: the Commission cannot easily resist the simultaneous pressure of France and Germany, as shown by the dilution of the Stability Pact (see Box 1.4 in Chapter 1 of the EEAG report 2006). Can the independence of competition policy be maintained given the politics of the Commission since states can lobby Commissioners and other Directorates (like Industry or Energy) to further national policies? This tension between competition and industrial policy has often surfaced, as in proposals to create a Super Commissioner or Vice Presidency that would oversee both industrial and competition policy, with the obvious objective to keep competition policy in check. Independent institutional bodies like an administrative panel within the Commission or even an own European
9780230_547018_16_cha14.indd 490
3/19/2009 3:56:08 PM
Mergers and Competition Policy in Europe
491
Competition Agency might be ways of protecting competition policy from these industrial policy pressures. 14.6.3
Privatisation and regulated sectors
Network industries such as electricity, gas and telecoms have been liberalised and the incumbent monopolies privatised in most countries (with the conspicuous exception of France). In those industries, regulation is maintained, even after liberalisation, because they have some segments that are a natural monopoly (for example, transmission and distribution of electricity). The way the privatisation process has been accomplished has very important consequences for the level of competition and performance in the industry. To privatise a monopoly is one thing: to introduce a degree of competition and then privatise is another. The evidence suggests that what really matters for performance is the level of competition in the industry (see, for example, Armstrong et al. 1994) rather than the structure of ownership. However, in some instances the desire of governments to obtain cash from privatising a monopoly has been a more important motive than to ensure competition. Some public companies have even been allowed to merge before privatisation (this was the case of the formation of the Spanish Endesa, for example). In those network industries, it is important that a potentially competitive structure is created before privatisation. Once firms with monopoly power are privatised it is very difficult to change the market structure with divestitures to enhance competition. Merger proposals may offer an opportunity to rearrange assets in a pro-competitive way. Network industries, such as those in the energy sector, are typically considered strategic by many countries, which resist privatisation or hold on to “golden shares” even after privatisation. France has resisted domestic liberalisation at the same time as publicly controlled French firms have gone shopping in other European markets.
14.7
Conclusions
Globalisation, accompanied by the information technology revolution and consequential lowering of trade costs and market expansion, imposes restructuring in many sectors and mergers are a prime instrument. Size is necessary to compete globally in many segments of industry and services, but consolidation may pose a threat to competition. Competition is a necessary prerequisite for economic efficiency: sufficient competition is needed for innovation and the timely termination of bad projects drives productivity growth. Domestic competition is key to international success and competitiveness: fostering national champions defeats this objective. The policy challenge is to allow the needed restructuring and potential increase in firm size in some sectors while at the same time protecting competition.
9780230_547018_16_cha14.indd 491
3/19/2009 3:56:08 PM
492 Designing the European Model
Our first conclusion is that a vigorous competition policy is needed, but care must be taken not to try to enforce low concentration in natural oligopoly industries, where the dynamics of investment is such that only a limited number of players can survive. Furthermore, merger control should take into account the need for larger firm size in several industries and the potential dynamic efficiencies, benefiting innovation, generated by merger proposals. A second conclusion is that artificial obstacles to hostile and cross-border mergers should be removed in Europe. Hostile takeovers are a sign of health of the market for corporate control. Cross-border mergers should proceed without regulatory obstacles, as they may keep in check the increase in domestic concentration. We acknowledge that ownership is not neutral, in particular in some industries like banking where relationships are important, but on balance this is insufficient justification for protectionism. European as well as national competition policy must play a major role in keeping markets open. A third conclusion is that care must be taken in not promoting European champions that end up effectively protected from closure. The political economy of European champions may imply that the powers of European competition policy, with the present institutional structure, are very limited to deal with those cases. Indeed, this is one instance where global coordination of competition policies may help. Fourth, the 2004 reform of the merger control procedure in the EU went a step in the right direction, increasing checks and balances for merging parties and the role of economic analysis. However, the guarantees for the parties, the quality of analysis and decision-making, and the protection against the lobbying pressures of national governments and firms could still be improved. Merger decision proposals should be taken by an administrative panel, independent of prosecutors and investigators. Failing this, a debate should be opened about the need of an independent European Competition Agency similar to the US Federal Trade Commission.
Notes 1. Mr Riboud, chairman and CEO of Danone, stated: “It is the duty of governments and political representatives in all parts of the world to do what they can to see that the decision-making centres of large businesses stay in their home countries. There is nothing shocking about that. I don’t think there was anything out of place in the way the French government and politicians showed concern over the possibility of a hostile bid.” 2. According to the Thomson Financial Securities Database. 3. According to UNCTAD (2005). 4. UNCTAD Cross-Border M&A Database (2005). The figure for the UK is exceptionally high for the year 2000 due to the takeover of Mannesmann by Vodafone. The deal itself comprised 202.8 billion dollars. 5. These figures are from the Thomson Financial Securities Database.
9780230_547018_16_cha14.indd 492
3/19/2009 3:56:08 PM
Mergers and Competition Policy in Europe
493
6. A merger is classified as hostile if there is resistance to the takeover on the part of the target company’s board of directors and management. 7. UNCTAD and Thomson Financial count this merger in different years. 8. Further evidence is provided by Mitchell and Mulherin (1996) and Andrade and Stafford (1999). 9. However, we should expect that banks will maintain a liquidity insurance provision role. 10. See also the survey article on mergers by Mueller (2003). 11. In fact, under some technical conditions on the profit functions under quantity (Cournot) competition, a merger with a small initial joint market share of the insiders, which is profitable and increases price, also raises the total surplus (Farrell and Shapiro 1990). With product differentiation, the average shareweighted price may fall post merger if there is a significant shift in output towards non-merging lower-cost firms. If a merger generates synergies, then (under some technical conditions on the profit functions) under quantity (Cournot) competition with homogenous product or price (Bertrand) competition with differentiated products, a merger will benefit consumers if and only if it hurts competitors (see Farrell and Shapiro 1990 and Vives 1999). 12. They employ the stock database from the Center for Research in Security Prices (CRSP) at the University of Chicago, including pricing information for all firms listed in the New York Stock Exchange, the American Stock Exchange and the Nasdaq. 13. See, for example, Hart (1983), Schaferstein (1988), Hermalin (1990), Schmidt (1997) and Vives (2000). 14. The Herfindahl index is defined as the sum of the squares of the market shares of firms in a particular market (see Chapter 4 in Vives 1999). 15. The EC will have authority over concentrations having a “Community dimension”, that is of operations with combined annual turnover larger than 5 billion euros or EU-wide turnover of each of at least two of the firms larger than 250 million euros. If each of the firms involved has more than 2/3 of its EU-wide sales in one member state, then this country has jurisdiction. 16. Namely, Cournot markets with homogeneous products or Bertrand markets with differentiated products and under some technical conditions on payoffs. 17. In fact, the group in charge of merger analysis – the Merger Task Force – was dissolved as a separate unit and integrated into the already existing industryspecific units of DG Competition. 18. See Baker (2005). 19. If a domestic firm borrows from a foreign bank, the lending bank’s headquarters will be presumably located in another country. The lending officers responsible for the loan will respond, directly or indirectly, to headquarters located far away geographically and organisationally. The foreign lending bank will presumably be a more complex organization, which may be less able to offer the same services and respond with the same flexibility as a local domestic bank. This implies that large multinational banks, to ensure effective internal controls, may be obliged to operate with internal procedures that are quite standard across countries. Thus, large multinational banks may not be flexible enough to adapt themselves to the specific needs of local borrowers. Local banks, instead, are more able to respond to the specific needs of local firms, and may be better partners in situations where relationship banking is important. An open issue is why multinational banks do not develop internal organization structures to cope with this problem. See Berglöf et al. (2005).
9780230_547018_16_cha14.indd 493
3/19/2009 3:56:08 PM
494
Designing the European Model
References Aghion, P., N. Bloom, R. Blundell., R. Griffith and P. Howitt (2005) “Competition and Innovation: An Inverted-U Relationship,” Quarterly Journal of Economics 120(2), 701–28. Andrade, G and E. Staffor (1999) “Investigating The Economic Role of Mergers,” Working Paper, Harvard Business School. Andrade, G., M. Mitchell and E. Stafford (2001) “New Evidence and Perspectives on Mergers,” Journal of Economic Perspectives, 15(2), 103–20. Ang, J. S. and Y. Cheng (2003) “Direct Evidence on the Market Driven Acquisition Theory,” unpublished working paper, Florida State University, Tallahassee, Florida. Armstrong, M., S. Cowan and J. Vickers (1994) Regulatory Reform: Economic Analysis and British Experience, MIT Press, Cambridge, MA. Baily, M. and H. Gersbach (1995) “Efficiency in Manufacturing and the Need for Global Competition,” Brooking Papers on Economic Activity: Microeconomics, 307–47. Baker, J. (2005) “My Summer Vacation at the European Commission,” The Antitrust Source (www.antitrustsource.com), September. Berglöf, E., P. Fulghieri, J. Gual, C. Mayer, P. Pita Barros and X. Vives (2005) “Integration of European Banking – The Way Forward,” Monitoring European Deregulation, Banking, CEPR. Besley, T. and P. Seabright (1999) “The Effects and Policy Implications of State-Aids to Industry,” Economic Policy, 14(28), 15–53. Capron, L. (1999) “The Long-term Performance of Horizontal Acquisitions,” Strategic Management Journal, 20, 987–1018. Cassiman, B., M. G. Colombo, P. Garrone and R. Veugelers (2004) “The Impact of M&A on the R&D Process. An Empirical Analysis of the Role of Technological and Market Relatedness,” mimeo, IESE Business School and Catholic University of Leuven. Conn, R., A. Cosh, P. Guest and A. Hughes (2003) “The Impact on U.K. Acquirers of Domestic, Cross-Border, Public and Private Acquisitions,” Working Paper no. 276, ESRC Centre for Business Research, University of Cambridge. Danzon, P. M., A. Epstein and S. Nicholson (2004) “Mergers and Acquisitions in the Pharmaceutical and Biotech Industries,” NBER Working Paper no. 10536. Dong, M., D. Hirshleifer, S. Richardson and S. H. Teoh (2004) “Does Investor Misvaluation Drive the Takeover Market?” Working Paper, Finance 0412002, Econ WPA. Duso, T., D. J. Neven and L.-H. Röller (2003) “The Political Economy of European Merger Control: Evidence using Stock Market Data,” Discussion Paper FS IV 02-34, Wissenschaftszentrum Berlin. Dutz, M. A. (1989) “Horizontal Mergers in Declining Industries,” International Journal of Industrial Organisation, 7, 11–33. Evenett, S. J. (2003) “The Cross-Border Mergers and Acquisitions Wave of the Late 1990s,” NBER Working Paper no. 9655. Farrell, J. and C. Shapiro (1990) “Horizontal Mergers: An Equilibrium Analysis,” American Economic Review 80, 107–26. Faulí-Oller, R. (2000) “Takeover Waves,” Journal of Economics and Management Strategy 9(2), 189–210. Fridolfsson, S.-O. and J. Stennek (2001) “Why Mergers Reduce Profits and Raise Share Prices – A Theory of Pre-emptive Mergers,” JEEA.
9780230_547018_16_cha14.indd 494
3/19/2009 3:56:08 PM
Mergers and Competition Policy in Europe
495
Gorton, G., M. Kahl and R. Rosen (2005) “Eat or Be Eaten: A Theory of Mergers and Merger Waves,” NBER Working Paper no. 11364. Gugler, K., D. C. Mueller, B. B. Yurtoglu and C. Zulehner (2003) “The Effects of Mergers: An International Comparison,” International Journal of Industrial Organization 21, 625–53. Gugler, K. and R. Siebert (2004) “Market Power versus Efficiency Effects of Mergers and Research Joint Ventures: Evidence from the Semiconductor Industry,” NBER Working Paper no. 10323. Hart, O. (1983) “The Market Mechanism as an Incentive Scheme,” Bell Journal of Economics 14, 367–82. Hermalin, B. (1990) “The Effects of Competitive Pressures on Executive Behavior,” Working Paper no. 148, U. C. Berkeley. Jovanovic, B. and P. L. Rousseau (2002) “The Q-Theory of Mergers,” American Economic Review 92(2), 198–204. Jovanovic, B., and P. L. Rousseau (2003) “Two Technological Revolutions,” Journal of the European Economic Association 1, (2–3) (April–May), 419–28. Jovanovic, B. and P. L. Rousseau (2004) “Mergers as Reallocation,” mimeo, New York University. An earlier version appeared 2002 as NBER Working Paper no. 9279. Lambrecht, B. M. (2004) “The Timing and Terms of Mergers Motivated by Economies of Scale,” Journal of Financial Economics 72, 41–62. Maksimovic, V. and G. Phillips (2001) “The Market for Corporate Assets: Who Engages in Mergers and Asset Sales and Are There Efficiency Gains?” Journal of Finance LVI(6), 2019–65. Mitchell, M. L. and J. H. Mulherin (1996) “The Impact of Industry Shocks on Takeover and Restructuring Activity”, Journal of Financial Economic 41, 193–229. Mueller, D. C. (2003) “The Financial Literature on Mergers: A Critical Survey” in Waterson, M., ed., Competition, Monopoly and Corporate Governance, Elgar, Cheltenham, 161–205. Neven, D. and P. Seabright (1995) “European Industrial Policy: The Airbus Case,” Economic Policy, 10(2), 315–58. Nickell, S. J. (1996) “Competition and Corporate Performance,” Journal of Political Economy 104, 724–46. Pesendorfer, M. (2003) “Horizontal Mergers in the Paper Industry,” RAND Journal of Economics 34(3), 495–515. Rhodes-Kropf, M. and S. Viswanathan (2004) “Market Valuation and Merger Waves,” Journal of Finance 59(6), 2685–718. Rhodes-Kropf, M., D. T. Robinson and S. Viswanathan (2005) “Valuation Waves and Merger Activity: The Empirical Evidence,” Journal of Financial Economics, 77(3), pp. 561–603. Roll, R. (1986) “The Hubris Hypothesis of Corporate Takeovers,” Journal of Business 59(2), 197–216. Schaferstein, D. (1988) “Product-Market Competition and Managerial Slack,” Rand Journal of Economics, 19, 147–55. Schmalensee, R. (1992) “Sunk Costs and Market Structure: A Review Article,” Journal of Industrial Economics, XL(2), 125–34. Schmidt, K. (1997) “Managerial Incentives and Product Market Competition,” Review of Economic Studies 64(2), 191–213. Shleifer, A. and R. W. Vishny (2003) “Stock Market Driven Acquisitions,” Journal of Financial Economics 70, 295–311. Strauss-Kahn, V. and X. Vives (2005) “Why do Headquarters Move?” CEPR Working Paper.
9780230_547018_16_cha14.indd 495
3/19/2009 3:56:08 PM
496 Designing the European Model Sutton, J. (1991) Sunk Costs and Market Structure: Price Competition, Advertising, and the Evolution on Concentration, The MIT Press, Cambridge, MA. Toxvaerd, F. (2004) “Strategic Merger Waves: A Theory of Musical Chairs,” Working Paper, Hebrew University of Jerusalem. UNCTAD (2005) Cross-Border M&A Database, www.unctad.org. Vives, X. (1999) Oligopoly Pricing, MIT Press, Cambridge MA. Vives, X. (2000) “Corporate Governance: Does It Matter?” in Xavier Vives, ed., Corporate Governance. Theoretical and Empirical Perspectives, Cambridge University Press, 1–21. Vives, X. (2004) “Innovation and Competitive Pressure,” CEPR Working Paper no. 4369f.
9780230_547018_16_cha14.indd 496
3/19/2009 3:56:08 PM
Index activity guarantee, 53 Agenda 2010, 4, 77 antitrust agencies, 486–7 Asian Crisis, 301 assistance to needy families, temporary, 48, 54–5 automatic stabilisers, 181, 183–5, 192, 208, 214–16 balance sheet effect, 239 Balassa-Samuelson-Effect, 207, 252–3 Banking Advisory Committee, 379, 388 Banking Directive, 366, 369, 377 Banking Supervision Committee, 379 Bank of International Settlements, 480 bargaining collective, 5–9, 75, 83–90, 94–6, 104, 111–12, 152–5 enterprise-level, 5–6, 93–4, 115 multi-employer, 5, 86, 93, 99, 107 pay, 5, 77, 83, 103, 111 sector-level, 6, 86, 93–6, 112 single-employer, 5, 86, 104–7, 112 bazaar activity, 123, 132, 136 bazaar economy effect, 7, 123, 136 boom-bust cycle, 183, 241 brain drain, 358 Brouwer Report on Financial Stability, 25, 381–4, 390 capital market imperfections, see financial market imperfections capital mobility (international), 103, 153, 233 class size reduction, 29, 448, 455–7 collective agreement sectoral, 5, 84, 106, 153 Committee of European Securities Regulators, 379, 386–8 Committee of Wise Men, 379, 385 company-level agreement, 144, 153–5, 168 competition policy, 12, 21, 27, 30, 216, 301, 373, 469
compulsory activation, 52 convergence beta, 264–5 criteria, 13, 199, 229, 250–3, 255, 256 play, 236–7, 240–1, 256 sigma, 264–5 corporate tax system, 355 Council of Economic Advisors, 124 credit constraints, 233, 237, 242 cross-border externalities, 350, 381, 389 currency board, 245, 248, 230, 250, 251 currency mismatch, 239, 256 current account balance, 297, 299 deficit, 18–19, 296, 299, 307, 310, 315, 320, 359 cyclically adjusted balance, 191, 195, 218 debt, implicit, 209–10, 399, 412–13 debt-to-GDP ratio, 197, 204 deficit twin, 307 demographic change, 402, 407 demographic crisis, 399 Directorate General of Competition, 486 earned income tax credit, 2, 48–51, 54–8, 75 modified, 57 ECB, 26, 96, 189, 220, 230–1, 246–51, 255, 322, 344, 368, 373–84, 391–3 Ecofin Council, 178, 189–90, 195, 212–14, 230 emploi jeunes programme, 62, 73 employee stock ownership, 101–2, 110–13, 415 employer-based pension scheme, 415, 425 employers’ associations, 5, 62, 88, 106–8, 152–3, 359–60 employment initiative contract, 53 measures, 89 part-time, 103, 147 497
9780230_547018_17_ind.indd 497
3/19/2009 8:30:03 PM
498 Index employment – continued protection, 3, 61–70, 74–83, 281–3, 287–91, 424 protection legislation, 74–5, 424 protection provision, 3, 61, 70, 76–9 temporary, 3, 103, 282–3, 287 employment-based tax credit, 53 employment-conditional benefits, 48 EU common agricultural policy, 332 EU enlargement, 9, 165, 169, 247, 358–60 European Central Bank, see ECB European Commission, 11–12, 177–9, 196, 210, 216, 370–3, 379–83, 388–90, 483–4 European Competition Agency, 32, 487–92 European Competition Policy, 381, 385, 490–2 European Council, 189, 211, 230, 251, 331, 383 European Court of Justice, 11, 178, 213, 343, 488 European merger control, 31, 469 European Parliament, 220, 388, 374, 386, 391 European Stock Exchange, 366 European System of Central Banks (ESCB), 26, 368, 373–4, 384, 390–2 excess wage pressure, 75 exchange rate depreciation, 89, 163, 209, 306 policy, 19, 235, 312, 344 regime, 231–6, 241–3, 248–50 factor price equalisation, 8, 131, 134–6, 358 factor productivity, 16, 195, 260, 268–9, 286, 313, 480 Family Credit, 51 see also Working Families’ Tax Credit favourability principle, 111 Federal Collective Agreement Act, 111 financial market imperfections, 242, 481 integration, 24–5, 366–7, 373, 382 Financial Services Action Plan, 383 Authority, 368, 378, 392–3 Policy Group, 381 firing costs, 61, 63
9780230_547018_17_ind.indd 498
Fiscal Policy Committee (FPC), 12, 179, 215–23 fiscal policy delegation, 179, 216 fiscal sustainability, 179, 197–9, 210, 242, 401 foreign assets, 297–8, 303–6, 314, 318, 321, 410 German Council of Economic Experts, 79, 99, 111 German Federal Statistical Office, 124–5, 137 Giovannini Group, 388 growth accounting, 16, 260, 267–79, 289 long-run, 82, 242 Heckscher-Ohlin trade model, 133 home-country principle, 24, 27, 357, 393 human capital accumulation, 82 human resource management, 101, 162 Ifo Institute, 2, 10, 21, 128, 153 income redistribution, 356, 361 inflation targeting, 96, 234, 248, 250 interest elasticity of savings, 311 equilibrium rate of, 311 intergenerational insurance, 409 intermediate production chain, 122 internal devaluation, 208–9 international reserves, 18, 233, 245, 301, 310 accumulation, 301, 312 interregional redistribution, 24, 361 intra-industry trade, 122 Investment Services Directive, 387–9 in-work benefits, 2, 51–2, 56, 75, 162 J-curve effect, 320 joint bargaining association, 107 labour cost, 3, 8, 54, 63, 69, 77, 128, 153–4, 166, 186, 208, 308, 358, 416 formal, 42–3 informal, 42 labour demand effective elasticity of, 47, 96, 162 elasticity of, 159
3/19/2009 8:30:03 PM
Index 499 labour demand – continued long-run, 164 short-run, 164–5 wage elasticity of, 159 labour-leisure trade-off, 156 labour market deregulation, 69–70 inflexible, 134 policy, 48, 74, 78, 83 reform, 2–3, 6–7, 17, 64, 69–70, 75–8, 105, 111–14, 289, 424 rigidities, 3, 61–4, 77, 424, 489 lender of last resort, 245, 251, 344, 375–7, 391 liquidity trap, 182 Lisbon strategy, 15–16, 287 low-wage sector, 47, 78 Maastricht Treaty, 11–13, 178–80, 189, 194, 199, 209, 213–14, 229, 338 merger cross-border, 27, 31, 368, 370, 381–2, 469–70, 476, 480, 489–90 horizontal, 473–5, 479–80, 485–7 vertical, 475 minimum income, 36, 51, 57–8 minimum wage, 1, 39, 45, 53–4, 58, 61–9, 72, 89, 133 constraint, 133 monetary policy autonomy, 96 committee, 215, 222 monetary unification, 96, 188 moral hazard creditor, 245 network externalities, 474, 479 new deal programme, 48 non-performing loans, 313, 380 offshoring activities, 122–3, 126–9, 132–6 old-age dependency ratio, 27, 402, 408–9, 418–20, 426–7 opt-out clause, 153, 229 pan-European banks, 25, 367 parental choice programme, 462 pay incentive pay mechanism, 110 incentive pay systems, 102
9780230_547018_17_ind.indd 499
performance-related, 97, 100–2, 110–13 profit-related, 101, 111–12 pay-as-you-go system, 28, 408–18, 423–8 pay compression general, 98 union policies of, 100 payroll-tax rebate, 48 pension debt, implicit, 209–10, 412–13 PISA study, 448, 454, 460–4 Plan d’Aide au Retour à l’Emploi, 74 portfolio rebalancing, 319–20 valuation effect, 318 Posted-Workers Directive, 290 pre-retirement policy, 28, 419–24 principle of subsidiarity, 12, 21, 178, 331–3, 338, 348, 351–6, 360 private equity firms, 30, 469 product market deregulation, 2, 70–1, 102 profit-sharing arrangements, 97 Programme for International Student Assessment, see PISA study prudential supervision, 373–9, 392–3 public education system, 448 public transfer, 35, 359, 425 pupil-teacher ratio, 455–7 rainy-day funds, 206, 222 regulatory fragmentation, 25, 366–7, 385, 389–90 replacement income, 24, 134–5, 360 rate, 38–9, 43, 54–6, 197, 426 residence principle, 23–4, 355–7 retirement age average effective, 407 legal, 408 Ricardian equivalence, 180, 185–6 saving-investment imbalance, 311 Schengen agreement, 349 school competition policies, 461 self-employment, 72, 79 self-fulfilling crises, 235, 243 service-based economy, 132 Single Market Programme, 344, 366, 369 skill-biased technical change, 131
3/19/2009 8:30:04 PM
500 Index social aid, 2, 8, 37–41, 45, 56–8, 134–7, 360, 425 social benefit, 23–4, 35–7, 48–9, 53, 56, 336, 358 social insurance contribution, 51, 89 social pact, 4, 88–9, 112–14 Bündnis für Arbeit, 89 social protection, 37, 232, 344, 422 social standard, 24, 36, 356–60 Stability and Growth Pact, 10–11, 177, 189, 196–7, 253–4, 338 structural imbalances, 14, 231 systemic risk, 26, 366, 373–4, 385 tax credit, 1, 47–53, 289 harmonization, 23, 355 too-big-to-fail policy, 375–6, 381 transaction cost, 13, 127, 231 trade gains from, 7–8, 18, 122, 131–6, 290 Treaty of Rome, 357, 373 underground economy, 42 unemployment assistance, 37–9, 56, 78 benefit programme, 62 benefits, 4, 17, 37–9, 52–6, 76–9, 90, 134, 181–4, 289 compensation, 36–7, 62, 98, 181–3 insurance, 37–9, 43–5, 76, 83 insurance benefit, 78 long-term, 38, 45–8, 52–4, 73, 77–8 low-skilled, 40 union density, 84–6, 95 unionisation rate, 83, 85–6, 103 unit cost of education, 452 US Bureau of the Census, 480
9780230_547018_17_ind.indd 500
vertical disintegration, 127 vertical integration, 127 voucher programme, 461–2 wage contract, 97, 106 dispersion, 90, 98–102, 109–11 flexibility, 4–8, 82, 89, 97–8, 108–10, 135–6, 244 indexation of pensions, 420 moderation, 3–9, 63, 71, 76, 82, 91, 96–7, 104, 112–14, 154, 166, 170 policy, 99 replacement policy, 36–7, 45–7, 56 subsidies, 2, 48, 53 supplementation policy, 45 wage-setting system, 105 welfare benefit, 1, 36, 51, 356–8 payments, 2, 40, 57, 75 programme, 35–6, 45, 51, 422–4 spending, 63 state, 1–4, 8, 23–4, 35–7, 56–8, 77, 131–7, 336, 356–7, 361, 400, 422 system, 2, 4, 23–4, 37–8, 58, 69, 74, 357–8 welfare-to-work policies, 45–8, 52–3 programme, 37, 48–9, 54 workfare system, 48 Working Families’ Tax Credit, 48–52 working time reduction, 17, 65–8, 86, 147–52, 168, 269, 277, 289 World Bank, 219, 480 youth dependency ratio, 418
3/19/2009 8:30:04 PM