The essays collected in this volume, written by well-known academics and policy analysts, discuss the impact of increased capital mobility on macroeconomic performance. The authors highlight the most adequate ways to manage the transition from a semi-closed economy to a semi-open one. Additionally, issues related to the measurement of openness, monetary control, optimal exchange rate regimes, sequencing of reforms, and real exchange rate dynamics under different degrees of capital mobility are carefully analyzed. The book is divided in four parts after the editor's introduction. The first part contains the general analytics of monetary policy in open economies. Parts two to four deal with diverse regional experiences, covering Europe, the Asian Pacific region, and Latin America. The papers on which the essays are based were originally presented at a conference on Monetary Policy in Semi-Open Economies, held at the Institute of Economic Research, Korea University in Seoul, Korea, in November 1992.
Capital controls, exchange rates, and monetary policy in the world economy
Capital controls, exchange rates, and monetary policy in the world economy Edited by SEBASTIAN EDWARDS University of California, Los Angeles The World Bank
CAMBRIDGE UNIVERSITY PRESS
PUBLISHED BY THE PRESS SYNDICATE OF THE UNIVERSITY OF CAMBRIDGE The Pitt Building, Trumpington Street, Cambridge CB2 1RP CAMBRIDGE UNIVERSITY PRESS The Edinburgh Building, Cambridge CB2 2RU, United Kingdom 40 West 20th Street, New York, NY 10011-4211, USA 10 Stamford Road, Oakleigh, Melbourne 3166, Australia © Cambridge University Press 1995 This book is in copyright. Subject to statutory exception and to the provisions of relevant collective licensing agreements, no reproduction of any part may take place without the written permission of Cambridge University Press. First published 1995 Reprinted 1997 First paperback edition 1997 Library of Congress Catalo gin g-in-Publi cation Data is available. A catalog record for this book is available from the British Library. ISBN 0-521-47228-8 hardback ISBN 0-521-59711-0 paperback Transferred to digital printing 2004
Contents
List of contributors
page vii
Introduction Sebastian Edwards Part I.
Monetary policy and stabilization in open economies 1. Stabilization and liberalization policies in semi-open economies Robert Mundell 2. Monetary regime choice for a semi-open country Jeffrey A. Frankel 3. Capital account liberalization: bringing policy in line with reality Manuel Guitidn
Part II. Capital mobility and macroeconomic policy in Europe 4. The lessons of European monetary and exchange rate experience Patrick Minford 5. Experience with controls on international capital movements in OECD countries: solution or problem for monetary policy? Jeffrey R. Shafer 6. Real exchange rates and capital flows: EMS experiences Alberto Giovannini 7. Monetary policy after German unification Wilhelm Nolling
1
19 35
71
93
119
157 181
vi
Contents
Part III. Capital controls and macroeconomic policy in the Asia-Pacific region 8. Capital movements, real asset speculation, and macroeconomic adjustment in Korea Yung Chul Park and Won-Am Park 9. The determinants of capital controls and their effects on trade balance during the period of capital market liberalization in Japan Shin-ichi Fukuda 10. Capital mobility and economic policy Michael Dooley 11. Monetary and exchange rate policies 1973-1991: the Australian and New Zealand experience Victor Argy Part IV. Capital mobility and exchange rates in Latin America 12. Exchange rates, inflation, and disinflation: Latin American experiences Sebastian Edwards 13. Capital inflows to Latin America with reference to the Asian experience Guillermo A. Calvo, Leonardo Leiderman, and Carmen M. Reinhart 14. Opening the capital account: costs, benefits, and sequencing James A. Hanson Index
199
229 247
265
301
339
383
431
Contributors
Victor Argyf School of Economics and Social Studies Macquarie University, Sydney Guillermo A. Calvo Department of Economics University of Maryland Michael Dooley Department of Economics University of California, Santa Cruz and National Bureau of Economic Research Sebastian Edwards Anderson Graduate School of Management University of California, Los Angeles The World Bank and National Bureau of Economic Research Jeffrey Frankel Department of Economics University of California, Berkeley Institute for International Economics and National Bureau of Economic Research
* Deceased.
vn
Shin-ichi Fukuda The Institute of Economic Research Hitotsubashi University Alberto Giovannini Graduate School of Business Columbia University and National Bureau of Economic Research Manuel Guitian Director, Monetary and Exchange Affairs Department International Monetary Fund James A. Hanson Resident Mission in IndonesiaJakarta The World Bank Leonardo Leiderman Department of Economics Tel Aviv University Patrick Minford Department of Economics and Accounting University of Liverpool and Cardiff Business School Robert Mundell Department of Economics Columbia University
viii
Contributors
Wilhelm Nolling East-West Consulting Agency, Hamburg
Carmen M. Reinhart Research Department International Monetary Fund
Won-Am Park Hongik University Yung Chul Park President, Korea Institute of Finance Korea University
Jeffrey Shafer Department of the Treasury Washington, D.C.
Introduction Sebastian Edwards
The past few years have seen a remarkable expansion in international trade. Throughout the globe countries have liberalized their trade accounts, reducing import tariffs and eliminating quantitative restrictions. What is perhaps most remarkable about this process is that, after years of hesitation, many developing nations have joined the more advanced countries in implementing trade liberalization reforms. For example, during the late 1980s and early 1990s most of the Latin American nations, which since the 1930s had favored an import substitution development strategy, went through gigantic unilateral trade reforms. Similar processes are taking place in Asia, where countries that for decades had pursued highly protectionist policies - India, for instance - are implementing major trade liberalization efforts. There is little doubt that, as we enter the end of the century, the view that freer trade and liberalization is conducive to improved economic performance has become dominant in academic as well as in policy circles. The decades-old debate between outward orientation and inward orientation has been decisively won by the proponents of more open commercial policies. The approval of GATT's Uruguay Round in December of 1993 has added impetus to the global efforts to open international trade. Although impediments to international trade in goods and (to some extent) services have been greatly reduced, restrictions on capital movements continue to be widespread, especially among the newly industrialized and developing nations. In fact, during the past decade or so advanced and poorer countries have dealt with capital movements very differently. Restrictions to capital mobility have been greatly reduced in the most advanced nations, and especially in the European Community since 1992. It has been estimated that, largely as a result of the relaxation of capital restrictions in the industrial countries, the turnover in foreign exchange markets more than tripled between 1986 and 1993, with total (net) daily turnover exceeding U.S.$l,000 billion. As Svensson (1993) has
2
Sebastian Edwards
pointed out, this figure is remarkably high when compared with the total stock of reserves held by the industrial countries' central banks - in the spring of 1992 the G-10 held approximately U.S.S400 billion in official reserves. In his chapter in this collection Jeffrey Shafer provides an extensive and detailed analysis of the process of capital account liberalization in the OECD countries. He discusses the evolution of the policy and academic thinking on capital controls, assesses their early effectiveness (or ^effectiveness), and evaluates the most important consequences of their removal during the past few years. In stark contrast to the liberalization process in the advanced countries, capital controls and impediments continue to be pervasive among the developing nations. Manuel Guitian points out in his contribution to this volume that the multilateral institutions have traditionally considered restrictions on trade on goods to be significantly more harmful than impediments to capital mobility. For example, while the Articles of Agreement of the International Monetary Fund (IMF) condemns the use of trade restrictions, it condones the use of capital controls.1 Moreover, during the past few years most multilateral negotiations to reduce barriers on international exchange - including GATT's Uruguay Round - have concentrated on trade in goods and services, virtually ignoring capital controls. Recently the United States has urged a number of developing nations and in particular the East Asian countries - to relax capital controls and to open their capital accounts.2 However, this proposition has been met with skepticism and, in some cases, with resistance. The apprehension regarding the opening of the capital account - or at least its rapid opening goes beyond traditional nationalistic views, and is based on both macroand microeconomic arguments. It is often argued that under capital mobility the national authorities lose (some) control over monetary policy, and that the economy will become more vulnerable to external shocks. Also, policy makers have often expressed concerns over their (effective) freedom for selecting the exchange rate regime, if capital is highly mobile. Moreover, sometimes it has been argued that full capital mobility will result in "overborrowing" and, eventually, in a major debt crisis as in Latin America in 1982. Other concerns regarding the liberalization of capital movements relate to increased real exchange rate instability, and loss of international competitiveness. Still other analysts have pointed out that the premature opening of the capital account could lead to massive capital flight from the country in question. This type of discussion has led to a 1 2
Article 6, section 3. See, e.g., Ito (1993) for a discussion on the U.S. political pressure for economic liberalization in East Asia.
Introduction 3 growing literature on the most adequate sequencing and speed of liberalization and stabilization reforms.3 The essays collected in this volume were presented at a conference on Monetary Policy in Semi-Open Economies, in Seoul, Korea, during November 6 and 7, 1992. The purpose of the conference - jointly organized by the Institute of Economic Research, Korea University (Yung Chul Park, Director), and the Ministry of Finance of the Korean government was to discuss candidly the consequences of increased capital mobility on macroeconomic performance, and the most adequate way to manage the transition from a "semi-closed" economy into a "semi-open" one. Issues related to the measurement of openness, monetary policy and control, optimal exchange rate regimes under capital mobility, sequencing of reform, and real exchange rate behavior under alternative degrees of capital mobility, among others, were discussed by a group of academics and policy analysts from virtually every part of the world. As the reader will find out, in spite of some differences of opinion, the authors found significant common ground during the discussion. In preparing this collection I have divided the volume into four parts. The first one, with essays by Robert Mundell, Jeffrey Frankel, and Manuel Guitian, deals with the general analytics of monetary policy in open and semi-open economies. The other three parts concentrate on case studies for distinct geographical areas: Part II, which includes chapters by Patrick Minford, Jeffrey Shafer, Alberto Giovannini, and Wilhelm Nolling, deals with Europe. Part III concentrates on the Asia-Pacific region and includes essays by Yung Chul Park and Won-Am Park, Shin-ichi Fukuda, Michael Dooley, and Victor Argy. Finally, Part IV contains chapters by myself; Guillermo Calvo, Leonardo Leiderman, and Carmen Reinhart; and James Hanson deals with the case of Latin America. In the rest of this introduction I discuss some of the most important analytical and policy issues related to the relaxation of capital controls in the world economy. Measuring the degree of capital mobility The fundamental purpose of capital controls is to interfere with the free international exchange in financial assets (Einzig 1934). A key issue, however, is how effective capital controls are in practice. Ample historical evidence suggests that there have been significant discrepancies between the legal and the actual degree of controls. In countries with severe impediments to capital mobility - including countries that have banned capital movement - the private sector has traditionally resorted to the overinvoicing of imports and underinvoicing of exports to sidestep legal 3
See, e.g., McKinnon (1991).
4
Sebastian Edwards
controls on capital flows. The massive volumes of capital flight that took place in Latin America in the wake of the debt crisis clearly showed that, when faced with the "appropriate" incentives, the public can be extremely creative in finding ways to move capital internationally. A number of authors have resorted to the term "semi-open" economy to describe a situation where the existence of taxes, licenses, or prior deposits restricts the effective freedom of capital movement - see, for example, Robert Mundell's contribution to this volume. Harberger (1978, 1980) has argued that the effective degree of integration of capital markets should be measured by the convergence of private rates of return to capital across countries. In his empirical analysis Harberger used national accounts data for a number of countries - including eighteen LDCs - to estimate rates of return to private capital, and found out that these were significantly similar. More important, he found that these private rates of return were independent of national capital-labor ratios. He interpreted these findings as supporting the view that capital markets are significantly more integrated than what a simple analysis of legal restrictions would suggest. Harberger (1980) also argued that remaining (and rather small) divergences in national rates of return to private capital are mostly the consequence of country risk premia imposed by the international financial community on particular countries. These premia, in turn, are determined by the perceived probability of debt default or rescheduling, and depend on a small number of "fundamentals," including the debt/GDP ratio and the international reserves position of the country in question. In trying to measure the effective degree of capital mobility, Feldstein and Horioka (1980) analyzed the behavior of savings and investments in a number of countries. They argue that if there is perfect mobility of capital, changes in savings and investments will be uncorrelated in any particular country. That is, in a world without capital restrictions an increase in domestic savings will tend to "leave the home country," moving to the rest of the world. Likewise, if international capital markets are fully integrated, increases in domestic investment will tend to be funded by the world at large, and not necessarily by domestic savings. Using a data set for sixteen OECD countries, Feldstein and Horioka found that savings and investment ratios were highly positively correlated, and concluded that these results strongly supported the presumption that long-term capital was subject to significant impediments. Frankel (1989) applied the FeldsteinHorioka test to a large number of countries, including LDCs, and largely corroborated the results obtained by the original study, indicating that savings and investment have been significantly positively correlated in most countries.
Introduction
5
Although Harberger (1980) and Feldstein and Horioka (1980) used different methodologies - the former looking at prices and the latter at quantities - and reached opposite conclusions regarding the degree of integration of world capital markets, they agreed on the need to go beyond legal restrictions in assessing the extent of capital mobility. In a series of studies, Edwards (1985, 1988) and Edwards and Khan (1985) argued that time series on domestic and international interest rates could be used to assess the degree of openness of the capital account. Using a general model that yields the closed and open economies cases as corner solutions, they estimated the economic degree of capital integration. They argued that capital restrictions play two roles: First, they introduce divergences to interest rate parity conditions and, second, they tend to slow down the process of interest rate convergence. Results obtained from the application of this model to the cases of a number of countries (Colombia, Singapore, Chile, Korea) support the idea that, in general, the actual degree of capital mobility is greater than what the legal restrictions approach suggests. Haque and Montiel (1990) and Reisen and Yeches (1991) have provided expansions of this model that allow for the estimation of the degree of capital mobility even in cases when there are not enough data on domestic interest rates, and that considered the possibility of a changing degree of capital mobility through time. Their analyses also suggest that in these developing countries the degree of capital mobility has been less than full. In his contribution to this volume Michael Dooley further expands this approach in an effort to measure the degree of effective capital mobility in Korea. Dooley compares the evolution of capital controls in Germany during the early 1970s and in Korea in the early 1990s. Dooley shows that a strict application of the semi-open economy model suggests that Korea has had an implausibly high degree of capital mobility. He argues that the estimated degree of capital mobility is highly sensitive to which series on domestic interest rates is used to depict "the" domestic interest rate. He argues that when the "curb" rate prevailing in the unregulated segment of capital markets is used, the effective degree of capital mobility in Korea appears to have been limited between 1970 and 1990. In his chapter in this collection Fukuda argues that deviations from interest rate parity provide an effective way of measuring the degree of capital mobility. If the capital market is fully liberalized from an economic perspective, covered interest parity must always hold. He argues that after December 1980 deviations from interest parity became negligible in Japan. Fukuda goes on to argue that in the case of Japan the extent of effective capital controls was endogenously determined by the behavior of the current account. Moreover, he points out that - in spite of the existence of generalized capital controls - until 1980 Japanese multinational trading
6
Sebastian Edwards
companies were given de facto preferential treatment. This allowed those firms to circumvent many of the restrictions to capital mobility. Nominal exchange rate regimes, monetary policy, and capital controls The international monetary system forged in Bretton Woods experienced a final collapse in 1973, when the industrial nations decided to adopt freely floating exchange rates.4 In spite of this significant change in the international financial system, throughout the 1970s most of the developing countries continued to rely heavily on fixed exchange rates. For example, by December 1979 85 percent of the developing countries had some variant of fixed exchange rates. During the 1980s and early 1990s, however, a large number of advanced nations moved slowly toward reducing the degree of exchange rate flexibility. The exchange rate mechanism (ERM) of the European Monetary System, with its narrow + / - 2.25 percent bands, represented the institutionalization of a system of limited flexibility. It was thought that by reducing the extent to which nominal exchange rates could fluctuate it was possible to combine the best features of purely floating and purely fixed exchange rate regimes (Giavazzi and Giovannini 1989). The crisis of the ERM in 1993 has introduced, however, very serious doubts on the desirability of fixed exchange rates in a world with a very high degree of capital mobility. Interestingly enough, as the industrial countries were shying away from exchange rate flexibility, more and more developing countries abandoned fixed exchange rates and adopted more flexible regimes. For example, according to the December 1990 issue of International Financial Statistics (IFS), the proportion of LDC members of the IMF under some type of fixed exchange rate regime had declined to 67 percent.5 This movement on behalf of the LDCs toward greater exchange rate flexibility was, to a considerable extent, associated with the debt crisis unleashed in 1982. Those countries that had to cope with sudden cuts in external financing had very limited policy options. In an effort to engineer gigantic resource transfers to their creditors, most of these countries adopted adjustment 4 5
Parts of this section draw on Edwards (1993). The IFS distinguishes several categories of fixed exchange rate countries, including those pegged to the US. dollar, those pegged to the French franc and those pegged to a "composite of currencies." It is unclear, however, to what extent the countries in this latter group have indeed followed a policy of pegging their currency to a basket. For all practical purposes, if a country alters continuously the composition of the basket, the resulting policy will not be one of a pegged exchange rate, but rather a form of exchange rate management.
Introduction 7 packages that included, as an important component, large nominal devaluations. It is in this context that in the mid-1980s we saw the end of long experiences with fixed exchange rates in countries such as Venezuela, Paraguay, and Guatemala. In the late 1980s and early 1990s, a number of observers and experts including prominent members of the IMF executive board - began to argue that the enthusiasm for an active and flexible exchange rate policy in the developing countries had gone too far. It was pointed out that by relying too heavily on exchange rate adjustments, and by allowing countries to adopt administered systems characterized by frequent small devaluations, many adjustment programs had become excessively inflationary. According to this view exchange rate policy in the developing countries should move toward greater rigidity - and even complete fixity - as a way to introduce financial discipline and provide a nominal anchor. This position was largely influenced by modern macroeconomic views that emphasized the role of expectations, credibility, and institutional constraints. Indeed the arguments used to promote a return to greater fixity in the LDCs were very similar to those used to support systems such as the ERM. In spite of the increasing enthusiasm for fixed rates during the early 1990s a number of authors - including Patrick Minford in this volume - continued to argue that exchange rate flexibility allowed countries, both developed and developing, to avoid real exchange rate (RER) overvaluation, and to accommodate shocks to real exchange rate fundamentals without incurring real costs.6 The implementation of major stabilization programs in eastern Europe and the former Soviet Union added interest to the exchange rate debate in the 1990s. Many countries in the region, including Poland, Czechoslovakia, and the former Yugoslavia, adopted a fixed exchange rate as a fundamental component of their anti-inflationary programs. However, some authors have argued that this approach is likely to generate a real exchange rate appreciation, putting the balance-of-payments target of the program in jeopardy.7 The debate on the desirability of alternative exchange rate regimes stems largely from the fact that exchange rates are perceived as playing two different roles. On the one hand, exchange rates, jointly with other policies, play an important role in helping maintain international competitiveness. On the other hand, they help in promoting macroeconomic stability and low inflation. In a way, when making decisions regarding exFor aflavorof the discussion within the IMF, see, for example, Burton and Gillman (1991), Aghevli, Khan, and Montiel (1991), and Flood and Marion (1991). See Nunnenkamp (1992).
8
Sebastian Edwards
change rate action, economic authorities face a classical policy dilemma. The recent literature on the subject has focused on optimal ways of assigning exchange rates and other policy tools to competing objectives, and on determining the optimal degree of exchange rate flexibility. In this volume the contributions by Minford, Guitian, Frankel, and myself deal with different aspects of the selection of the most appropriate exchange rate regime under alternative institutional context. In his chapter in this volume Minford presents simulation results from the Liverpool Model suggesting that a system of flexible exchange rates gives rise to a significantly more stable economy than an ERM-type system. This is especially the case when in the band-type system no attempts are made to coordinate monetary policy across countries. Minford goes further to argue that a monetary union, such as the one envisaged in the European Monetary Union (EMU), will result in a high degree of output instability if supply shocks are dominant. In his contribution to this collection Jeffrey Frankel also deals with the selection of an exchange rate regime, and discusses the choices faced by rapidly growing middle income countries, such as the Asian newly industrialized countries (NICs). He argues that the degree of openness of the country will largely determine whether it is worthwhile to adopt a fixed exchange rate and to give up, in the process, monetary independence. Frankel points out that the availability of alternative means of adjustment should also be an important ingredient in deciding on what type of exchange rate regime to adopt. In the second part of his chapter Frankel deals with the selection of a nominal anchor for monetary policy and argues that, under a set of plausible conditions, nominal GNP dominates the nominal exchange rate, money supply, and the price level. In his analysis of the experiences of Australia and New Zealand, Victor Argy also tackles the important issue of choosing an exchange rate regime. He points out that the appropriate regime for a particular country will depend on the structural characteristics of the economy, and especially on the degree of labor market and wage rate flexibility. Much of the discussion on the desirability of alternative monetary and exchange rate arrangements has been couched in terms of the theory of optimal currency areas pioneered by Mundell (1961). The unification of Germany provides possibly the most dramatic example of a rapid formation of a political and economic union. In his contribution to this collection Nolling deals with the German experience and discusses the considerations behind the introduction of the deutsche mark in East Germany. He also deals with some of the consequences - both real and monetary - of having implemented the unification at a one-to-one exchange rate. In my own chapter I evaluate the functioning of fixed exchange rate
Introduction
9
regimes both in the long and short runs. In the first part of my essay I analyze the long-term - that is, at least fifty years - experiences of a score of South American countries with fixed exchange rates. I argue that while this system worked well during relatively tranquil periods, it largely failed during the turbulent 1970s and 1980s. In the second part I analyze the experiences of Chile and Mexico with exchange rate-based stabilization programs. I find out that in Mexico the adoption of an exchange rate anchor changed the character of inflation and accelerated the convergence of domestic to world inflation. Interestingly enough this was not the case in Chile during the late 1970s and early 1980s. Real exchange rates and capital flows: the sequencing issue An important problem that has been addressed when designing a liberalization strategy refers to the sequencing of reform (McKinnon 1991). This issue was first considered in the 1980s in discussions dealing with the Southern Cone (Argentina, Chile, and Uruguay) experiences, and emphasized the macroeconomic consequences of alternative sequences. Recent discussions on the experiences of the former communist nations have broadened the debate, and have asked how privatization and other basic institutional reforms should be timed in the effort to move toward market orientation. It is now generally accepted that resolving the fiscal imbalance and attaining some degree of macroeconomic reform should be a priority in implementing a structural reform. Most analysts also agree that the trade liberalization reform should precede the liberalization of the capital account, and that financial reform should only be implemented once a modern and efficient supervisory framework is in place.8 The behavior of the real exchange rate is at the heart of this policy prescription. The central issue is that liberalizing the capital account would, under some conditions, result in large capital inflows and in an appreciation of the real exchange rate (McKinnon 1982, Edwards 1984, Harberger 1985).9 The problem with this is that an appreciation of the real exchange rate will send the "wrong" signal to the real sector, frustrating the reallocation of resources called for by the trade reform. The effects of this real exchange rate appreciation will be particularly serious if, as argued by McKinnon (1982) and Edwards (1984), the transitional period is characterized by "abnormally" high capital inflows that result in tempo8 9
Lai (1985) presents a dissenting view. This would be the case if the opening of the capital account is done in the context of an overall liberalization program, where the country becomes attractive for foreign investors and speculators.
10
Sebastian Edwards
rary real appreciations. If, however, the opening of the capital account is postponed, the real sector will be able to adjust and the new allocation of resources will be consolidated. According to this view, only at this time should the capital account be liberalized. Hanson's contribution to this volume deals with a number of analytical arguments related to sequencing and reviews the experiences of a score of Latin American countries. The chapters by Giovannini, Park and Park, and Fukuda expand the analysis in several directions and focus on important experiences in Europe and Asia. As pointed out, recent discussions on the sequencing of reform have enriched the analysis, and have included other markets. An increasing number of authors have argued that the reform of the labor market - and in particular the removal of distortions that discourage labor mobility should precede the trade reform, as well as the relaxation of capital controls. In Edwards (1992a) I argue that it is even possible that the liberalization of trade in the presence of highly distorted labor markets will be counterproductive, generating overall welfare losses in the country in question. Interestingly enough, the discussions on the sequencing of reform have only addressed in detail the order in which the liberalization of various "real" sectors in society should proceed. For instance, only a few studies - such as Krueger (1981) and Edwards (1984) - have dealt with the order of reform of agriculture, industry, government (privatization), financial services, and education. The key question here is the extent to which independent reforms will bear all their potential fruits, or whether the existence of synergism implies that in a broad-based liberalization process the reforms in different sectors reinforce each other.10 As the preceding discussion has suggested, real exchange rate behavior is a key element during a trade liberalization transition. According to traditional manuals on "how to liberalize," a large devaluation should constitute the first step in a trade reform process. Bhagwati (1978) and Krueger (1978) have pointed out that in the presence of quotas and import licenses a (real) exchange rate depreciation will reduce the rents received by importers, shifting relative prices in favor of export-oriented activities and, thus, reducing the extent of the antiexport bias.11 Maintaining a depreciated and competitive real exchange rate during a trade liberalization process is also important in order to avoid an explosion in imports growth and a balance of payments crisis. Under most circumstances a reduction in the extent of protection will tend to generate a 10
11
Of course, this discussion is related to second best analysis of policy measures. See Edwards (1992b) for a formal multisector model to analyze the welfare consequences of alternative reform packages. See Krueger (1978, 1981) and Michaely, Choksi, and Papageorgiou (1991).
Introduction
11
rapid and immediate surge in imports. On the other hand, the expansion of exports usually takes some time. Consequently, there is a danger that a trade liberalization reform will generate a large trade balance disequilibrium in the short run. This, however, will not happen if there is a depreciated real exchange rate that encourages exports and helps keep imports in check. Many countries have historically failed to sustain a depreciated real exchange rate during the transition. This has mainly been the result of expansionary macroeconomic policies that generate speculation, losses of international reserves and, in many cases, a reversal of the reform effort. In the conclusions to the massive World Bank project on trade reform, Michaely et al. (1991) succinctly summarize the key role of the real exchange rate in determining the success of liberalization programs: "The long term performance of the real exchange rate clearly differentiates liberalizers' from 'non-liberalizers'" (p. 119). Edwards (1989) used data on thirty-nine exchange rate crises and found that in almost every case, real exchange rate overvaluation ended up with drastic increases in the degree of protectionism. In their contribution to this volume, Park and Park construct a simulation model to evaluate the real exchange rate impact of opening the capital account on Korea's real exchange rate. Park and Park argue that recent attempts to relax slowly some of the controls to capital mobility in Korea have resulted in a dramatic increase in speculative activities. According to them this has, in turn, generated a rapid and somewhat artificial real-estate boom. Overall, Park and Park conclude that, under most circumstances, a rapid capital market liberalization would have a large effect on real exchange rates and would have a substantial negative effect on Korea's competitiveness. During the late 1980s and early 1990s a large number of Latin American countries implemented major trade liberalization reforms that opened significantly their economies to foreign competition. In almost every country the liberalization effort was supplemented, at least initially, by active efforts at generating significant real exchange rate depreciations. These competitive real exchange rates were at the center of the vigorous performance of most of Latin America's exports sectors during the late 1980s and early 1990s. However, during 1992-93 most Latin countries have experienced significant real exchange rate appreciations and losses in competitiveness. In their contribution to this book Calvo, Leiderman, and Reinhart also deal with the capital inflows problem, and argue that the concomitant real appreciations (and losses in international competitiveness) have generated considerable concern among policy makers and political leaders. In 1991— 92 capital inflows into Latin America increased significantly, after eight years of negative resource transfers. This increased availability of foreign
12
Sebastian Edwards
funds affected the real exchange rate through increased aggregate expenditure. A proportion of the newly available resources has been spent on nontradables - including the real-estate sector - putting pressure on their relative prices and on domestic inflation. An interesting feature of the recent capital movements is that a large proportion corresponds to portfolio investment and relatively little is direct foreign investment. Mexico was the most important recipient of foreign funds in that region during the early 1990s. As Calvo et al. argue in their chapter, real exchange rate appreciations generated by increased capital inflows are not a completely new phenomenon in Latin America. In the late 1970s most countries in the region, but especially the Southern Cone nations, were flooded with foreign resources that led to large real appreciations. The fact that this previous episode ended in the debt crisis has added drama to the current concern about the possible negative effects of these capital flows. Whether these capital movements are temporary - and thus subject to sudden reversals as in 1982 - is particularly important in evaluating their possible consequences. Calvo et al. argue that the most important causes behind the generalized inflow of resources are external. In particular, their empirical analysis suggests that the two main reasons triggering these capital movements are the recession in the industrialized world and the reduction in U.S. interest rates. These authors suggest that once these world economic conditions change, the volume of capital flowing to Latin America will be reduced. This means that at that point the pressure over the real exchange rate will subside and a real exchange rate depreciation will be required. In his chapter James Hanson presents a comprehensive discussion on the costs and benefits of opening the capital account, with especial reference to the sequencing issue in Latin America. In his discussion he makes the important point that, in the final analysis, the consequences of relaxing capital controls will depend on whether the domestic financial system is sound or weak. He argues that the rather disappointing Southern Cone (Argentina, Chile, and Uruguay) experience with capital account liberalization in the late 1970s and early 1980s was largely the result of lax banking regulations. During the late 1980s and early 1990s many Latin American countries tried to cope with the real appreciation pressures in several ways. Colombia, for instance, tried to sterilize the accumulation of reserves by placing domestic bonds (OMAs) on the local market in 1991.12 In order to place 12
An important peculiarity of the Colombian case is that the original inflow of foreign exchange came through the trade account.
Introduction
13
these bonds, however, the local interest rate had to increase, making them relatively more attractive. This generated a widening interest rate differential in favor of Colombia, which attracted new capital flows that, in order to be sterilized, required new bond placements. This process generated a vicious cycle that contributed to a very large accumulation of domestic debt, without significantly affecting the real exchange rate. This experience shows vividly the difficulties faced by the authorities wishing to handle real exchange rate movements. In particular, this case indicates that real shocks - such as an increase in foreign capital inflows - cannot be tackled successfully using monetary policy instruments. Argentina has recently tried to deal with the real appreciation by engineering a "pseudo" devaluation through a simultaneous increase in import tariffs and export subsidies. Although it is too early to know how this measure will affect the degree of competitiveness in the country, preliminary computations suggest that the magnitude of the adjustment obtained via tariffs-cum-subsidies package may be rather small. Mexico has followed a different route and has decided to postpone the adoption of a completely fixed exchange rate. In October 1992 the pace of the daily nominal exchange rate adjustment was doubled to forty cents. As in the case of Argentina, it is too early to evaluate how effective these measures have been in dealing with the real appreciation trend. However, as pointed out earlier, a number of analysts of the Mexican scene have already argued that this measure is clearly not enough. Chile has tackled the real appreciation problem by implementing a broad set of measures, including the management of exchange rate policy relative to a three-currency basket, imposing reserve requirements on capital inflows, allowing the nominal exchange rate to appreciate somewhat, and undertaking significant sterilization operations. In spite of this multifront approach, Chile has not avoided real exchange rate pressures. Between December 1991 and December 1992 the Chilean-US, bilateral real exchange rate appreciated by approximately 10 percent. As a result of this, exporters and agriculture producers have been mounting increasing pressure on the government for special treatment, arguing that an implicit contract had been broken by allowing the real exchange rate to appreciate. This type of political reaction is, in fact, becoming more and more generalized throughout the region, adding a difficult social dimension to the real exchange rate issue. Although there is no easy way to handle the real appreciation pressures, historical experience shows that there are, at least, two possible avenues that the authorities can follow. First, in those countries where the dominant force behind real exchange rate movements is price inertia in the presence of nominal exchange rate anchor policies, the adoption of a prag-
14
Sebastian Edwards
matic crawling peg system will usually help. This means that, to some extent, the inflationary targets will have to be less ambitious as a periodic exchange rate adjustment will result in some inflation.13 However, to the extent that this policy is supplemented by tight overall fiscal policy there should be no concern regarding inflationary explosions. Second, the discrimination between short-term (speculative) capital and longer-term capital should go a long way in helping resolve the preoccupations regarding the effects of capital movements on real exchange rates. To the extent that short-term capital flows are more volatile, and thus capital inflows are genuinely long-term, especially if they help finance investment projects in the tradables sector, the change in the RER will be a "true equilibrium" phenomenon, and should be recognized as such by implementing the required adjustment resource allocation. In practice, however, discriminating between "permanent" and "transitory" capital inflows is difficult; at the end policy makers are forced to make a judgment call. In his chapter Alberto Giovannini discusses the "capital inflows problem" from a European perspective, arguing that there is a remarkable similarity with the Latin American experience. He analyzes the consequences of the capital account liberalization reforms in France, Italy, and Spain and contrasts their experiences to that of the Latin American countries. Giovannini points out that in all three countries large capital inflows were observed after 1987. Moreover, in all three cases the exchange rate was not allowed to move freely to accommodate the changes in the capital account, and the larger inflows were reflected in sizable accumulation of international reserves. This, in turn, resulted in higher nontradables inflation, as has been the case in many Latin American nations. Giovannini also points out that, in spite of the large capital inflows, real interest rates remained rather high in the three European countries - that is, significantly higher than in Germany. References Aghevli, B., M. Khan, and P. Montiel. (1991). "Exchange Rate Policy in Developing Countries: Some Analytical Issues." IMF Occasional Paper no. 78. Washington, D.C.: International Monetary Fund. Bhagwati, J. (1978). Foreign Trade Regimes and Economic Development: Anatomy and Consequences of Exchange Control Regimes. Cambridge, Mass.: NBER. Burton, D., and M. Gillman. (1991). "Exchange Rate Policy and the IMF." Finance and Development 28 (September): 18-21. Edwards, S. (1984). "The Order of Liberalization of the External Sector in Developing Countries." Princeton Essays on International Finance no. 156, Princeton University. (1985). "Stabilization with Liberalization: An Evaluation of Ten Years of Chile's 13
More specifically, with this option the one-digit inflationary goal will be postponed.
Introduction
15
Experiment with Free Market Policies 1973-1983." Economic Development and Cultural Change 33 (January): 223-54. (1988). "Financial Deregulation and Segmented Capital Markets: The Case of Korea." World Development 18 (January): 185-94. (1989). Real Exchange Rates, Devaluation and Adjustment. Cambridge, Mass.: MIT Press. (1992a). "Sequencing and Welfare: Labor Markets and Agriculture." In I. Goldin and A. Winters (eds.), Open Economies: Structural Adjustment and Agriculture. Cambridge: CEPR/Cambridge University Press. (1992b). "The Sequencing of Structural Adjustment and Stabilization." San Francisco: International Center for Economic Growth Occasional Paper no. 34. (1993). "Exchange Rates as Nominal Anchors." Weltwirtschaftliches Archiv 129, no. 1: 1-32. Edwards, S., and M. Khan. (1985). "Interest Rate Determination in Developing Countries: A Conceptual Framework." IMF Staff Papers 32, no. 3 (September): 377-403. Einzig, Paul. (1934). Exchange Control. London: Macmillan. Feldstein, M., and C. Horioka. (1980). "Domestic Saving and International Capital Flows." Economic Journal 90(June): 1-28. Flood, R., and N. Marion. (1991). "The Choice of the Exchange Rate System." IMF Working Paper no. 90. Washington, D.C.: International Monetary Fund. Frankel, J. (1989). "Quantifying International Capital Mobility in the 1980s." Cambridge, Mass.: NBER Working Paper no. 2856, February. Giavazzi, F , and A. Giovannini. (1989). Limited Exchange Rate Flexibility. Cambridge, Mass.: MIT Press. Hanson, J. (1992). "Opening the Capital Account." Washington, D.C.: World Bank Policy Research Working Paper no. 901. Haque, N., and P. Montiel. (1990). "Capital Mobility in Developing Countries Some Empirical Tests." IMF Working Paper no. 117, December. Washington, D.C.: International Monetary Fund. Harberger, A. (1978). "Perspectives on Capital and Technology in Less Developed Countries." In M. Artis and A. Nobay (eds.), Contemporary Economic Analysis. London: Croom Helm, 151-69. (1980). "Vignettes on the World Capital Market." American Economic Review 70, no. 2 (May): 331-37. (1985). "Observations on the Chilean Economy 1973-1983." Economic Development and Cultural Change 33: 451-62. Ito, T. (1993). "U.S. Political Pressure and Economic Liberalization in East Asia." In J. Frankel and M. Kahler (eds.), Regionalism and Rivalry: Japan and the United States in Pacific Asia. Chicago: NBER/University of Chicago Press, 121-35. Krueger, A. (1978). Foreign Trade Regimes and Economic Development: Liberalization Attempts and Consequences. Cambridge, Mass.: NBER. (1981). Trade and Employment in Developing Countries. Chicago: University of Chicago Press. Lai, D. (1985). "The Real Aspects of Stabilization and Structural Adjustment Policies." Washington, D.C.: World Bank Staff Working Paper no. 636. McKinnon, R. (1982). "The Order of Economic Liberalization: Lessons from Chile and Argentina." Carnegie-Rochester Conference Series on Public Policy 17(fall): 476-81.
16
Sebastian Edwards
(1991). The Order of Economic Liberalization: Financial Control in the Transition to a Market Economy. Baltimore: Johns Hopkins University Press. Michaely, M., A. Choksi, and D. Papageorgiou (eds.). (1991). Liberalizing Foreign Trade. Oxford: Basil Blackwell. Mundell, R. (1961). "A Theory of Optimum Currency Areas." American Economic Review 51 (September): 785-99. Nunnenkamp, P. (1992). "Critical Issues of Stabilization in Post-Socialist Countries." In K. Kaczynski (ed.), Reintegration of Poland into the Western Economy. Warsaw: Foreign Trade Research Institute, 123-41. Reisen, H., and H. Yeches. (1991). "Time-Varying Estimates on the Openness of the Capital Account in Korea and Taiwan." Paris: OECD Development Centre Technical Paper no. 42, August. Svensson, Lars E. O. (1993). "Estimating Forward Interest Rates." Quarterly Review (Sveriges Rijksbank, Sweden), no. 3: 32-42.
PART I
Monetary policy and stabilization in open economies
CHAPTER 1
Stabilization and liberalization policies in semi-open economies Robert Mundell
What is a "semi-open" economy and how is it different from an "open" economy? One meaning comes from trade theory: Whereas an open economy produces tradables, a semi-open economy has a substantial sector of nontradables. In an open economy, the main relative price is the terms of trade, but in a semi-open economy, the relative price of domestic (nontraded) and international (traded) goods, often referred to as the real exchange rate, must also be taken into account. Openness may refer to either the current account or the capital account. Another meaning of semi-open refers to artificial restrictions to the trade or the capital accounts. A semi-open economy is one that is partially closed because of trade impediments or restrictions to capital movements. Frequently recurring problems are those of determining how and in what order artificial impediments to trade and capital movements should be lifted, and how each affects the conduct of monetary and exchange rate policy. The liberalization and stabilization problems are in principle separate from one another. The liberalization problem is that of determining how rapidly, and to what degree, an economy should be moved toward its "natural" openness, as determined by geography, population, technology, resource structure, and transport costs. Liberalization policy involves reduction of tariffs, quotas, barriers to investment, and controls on capital movements. A problem associated with liberalization is that of determining its appropriate direction, pace, and sequence: Should liberalization policies be directed first at the trade or the capital account or should they occur simultaneously? The stabilization problem is that of controlling inflation and exchange rate depreciation. There is general consensus that stabilization requires a package of policies that includes fiscal retrenchment, monetary restraint, 19
20
Robert Mundell
and exchange rate stability. But three areas of disagreement include (1) the feasibility of using the exchange rate as a deflationary device; (2) the usefulness of international capital flows during the stabilization period; and (3) the role events in the world economy play in the success of stabilization policies. Although in principle distinct, the liberalization and stabilization problems interact with one another. Liberalization can cause instability and stabilization can make liberalization more difficult. A sudden liberalization of trade can increase unemployment in import-competing industries, and liberalization of capital movements can cause sudden embarrassments for monetary policy directed at price stability. How should liberalization and stabilization policies be coordinated with one another? To an important extent, the liberalization and stabilization problems can be studied independently of a country's level of development or economic system. One reason may be that the "semi-open economy model" is relevant for both developing and more advanced countries. But a more important reason is that there has been a considerable convergence between the institutions of advanced and developing countries. Developing countries have acquired considerable finance structure in the form of capital markets, banks, and other financial institutions; and advanced countries have shown that they are by no means immune to the problems of bank instability. It is now possible to improve policies in the advanced countries, and in the former socialist countries, by study of the solution for problems facing the developing countries. Exchange rate as a disciplinary device The exchange rate approach to disinflation starts from initial conditions of inflation and currency depreciation. A sine qua non of the approach is a balanced budget so that monetary policy can be geared to the balance of payments. By fixing the exchange rate, the prices of international goods stop rising, while domestic goods continue to rise, shifting demand away from domestic goods, and creating a balance-of-payments deficit. In the absence of any domestic credit creation, the reserve sales necessary to prevent currency depreciation lower the money supply (or its rate of growth with controlled credit creation), starting the disinflationary process.1 An objection to the use of the exchange rate as a disinflationary brake is that it can lead to overvaluation. Fixing the exchange rate brakes the rise in the prices of international goods, but wage rates and the price of domestic goods may continue to increase. If the momentum of wage rates continues after exchange rate stabilization, they will overshoot equilib1
For recent discussions of the adjustment process see Mundell (1989a, 1989b, 1991b, 1992c).
Stabilization policies in semi-open economies
21
rium, leading to a decline in competitiveness of (especially nontraditional) exports and the general syndrome of overvaluation. The monetary restraint imposed by the fixed exchange rate regime often leads to capital imports that undermine monetary discipline. History is replete with examples of countries that have, in the process of introducing liberalization and stabilization policies, ended up with overvalued currencies. The problem has been much discussed in the literature. Studied examples include stabilization policies in the Southern Cone (Chile, Argentina, and Uruguay) in the late 1970s and early 1980s; in Mexico (1987-92); in Argentina in 1990-92; in Spain and Italy after 1985; in Canada after 1987; in Britain after 1989; in New Zealand (1984-88); in Peru (1987-92); in Turkey (1977-81 and 1986-91); in Poland (1989-92); and in Hungary (1991-92).2 Exchange rate stabilization and monetary restraint seem to lead to overvaluation of the real exchange rate and eventually a breakdown of the stabilization plan.
The real exchange rate and the terms of trade The semi-open economy models that have been used to analyze stabilization and liberalization problems focus on changes in the real exchange rate, with the assumption that the terms of trade are constant. This corresponds to the model of the semi-open economy in which a country is too small to affect its terms of trade. Nevertheless, there are pitfalls in applying this model to actual situations. The fact that a country has no influence on its terms of trade does not mean that the terms of trade are constant. Stabilization plans can break down because of internal factors or because of changes in world market conditions. Wage increases can appreciate the real exchange rate and impair a country's competitiveness; on the other hand, changes in world market conditions can worsen a country's terms of trade and make the task of stabilization more difficult or politically impossible. It is important to use an economic model that is capable of distinguishing between these two cases. The open economy model focuses on changes in the terms of trade and is not appropriate for dealing with changes in competitiveness due to movements of the real exchange rate. On the other hand, the semi-open economy model focuses on the real exchange rate and conceals possible changes in the terms of trade. A stabilization plan might fail because of changes in the terms of trade, even though the real exchange rate remains 2
See, for example, Edwards and Cox-Edwards (1987); Dornbusch and Park (1987); Dornbusch (1992); McKinnon (1991); Lizondo and Montiel (1989); Montiel and Ostry (1991); and Mundell (1989a, 1991a).
22
Robert Mundell Px TOT
TOT1 price of exports
RER
price of imports
Pm
Figure 1.1. The terms of trade and the real exchange rate. unchanged; or it might fail because of changes in the real exchange rate, the terms of trade remaining constant. There is no reason, in theory, to expect any systematic relation between changes in the terms of trade and changes in the real exchange rate. The money prices of three variables - prices of domestic goods, exports, and imports - cannot, in general, be collapsed into two monetary variables, the prices of domestic and prices of international goods. Nor can two real variables - the terms of trade and the real exchange rate - be collapsed into a single variable called the real exchange rate. The terms of trade can worsen (improve) because of increases (decreases) in the price of imports or decreases (increases) in the price of exports: In the first case the real exchange rate appreciates; in the second, it depreciates. Thus, in Figure 1.1, the real exchange rate corresponds to the RER curve (given the price of domestic goods), whereas the terms of trade is given by the slope of the TOT line. The terms of trade are determined by external factors while the real exchange rate, by internal factors. An externally induced fall in the terms of trade shifts the TOT line to, say, TOT'. At fixed exchange rates, the equilibrium would move to R if the fall in the terms of trade were due to a rise in import prices, and to T if it were due to a fall in export prices. With given prices of domestic goods, the real exchange rate would therefore depreciate in the former and appreciate in the latter case.
Stabilization policies in semi-open economies
23
Under flexible exchange rates and a fixed money supply, the nominal exchange rate would appreciate when faced with an increase in the price of imports, and depreciate when faced with a decrease in the price of exports, in both cases leading to an equilibrium between T and R, such as that at U, where the real exchange rate is unchanged.3 Other external factors The Chilean stabilization plan offers an example where the breakdown of the stabilization plan coincided with a fall in the terms of trade. The rise in the price of oil in 1979 was followed by the collapse of copper prices in 1982 (from $1.00 per pound in 1980 to $0.67 in 1982). The first change depreciated, and the second appreciated the real exchange rate; but both worsened Chile's terms of trade.4 Changes in the terms of trade are not the only external factors affecting success of failure of stabilization plans. Changes in the nominal exchange rate and the interest rate affected many of the developing countries (including Chile) in the early 1980s. The dollar appreciated strongly with the shift in the U.S. policy mix under President Ronald Reagan. Fiscal expansion and monetary tightness led to appreciation of the dollar and soaring real interest rates, with a short but sharp recession in 1982. The appreciation of the dollar helped to overvalue the peso, high real interest rates aggravated the burden of debt service, and the world recession weakened the market for nontraditional exports. Several other countries' attempts at stabilization have failed partly because of external factors. Argentina's and Uruguay's stabilization plans in the late 1970s and early 1980s provide another example. In 1979, the U.S. government imposed an embargo on grain exports to the Soviet Union, in reaction against its invasion of Afghanistan. The Soviet Union turned to Argentina as an alternative source, and the resulting export boom made it possible to introduce Argentina's stabilization plan. Huge spillover effects of the boom were experienced in the form of a massive capital influx to neighboring Uruguay, largely directed at building the resort center of Punte del Este; this was the "Afghanistan effect." When the boom subsided in 1981 and 1982, budgets became unbalanced and the stabilization plans of both countries collapsed. The plans were all undermined by 3
4
The change in the real exchange rate depends on the weights attached to imports and exports in the index of international goods prices. Given a fixed money supply and flexible exchange rates, the real exchange rate would be unchanged when the terms of trade change if the weights in the prices of international goods in the real exchange rate index were the same as the weights of those goods in the demand for money function. See Edwards (1984, 1985); Corbo (1985, 1986); Edwards and Cox-Edwards (1987); and Cuadra and Valdes (1992) for analyses of the Chilean case.
24
Robert Mundell
the combination of a rising dollar, falling terms of trade, high real interest rates, and the world slump. In the cases of both Argentina and Uruguay, however, fault also lay in the inability of the authorities to maintain fiscal solvency. External factors sometimes combine with policy mistakes to leave a country saddled with an overvalued real exchange rate. The North American Free Trade Area Agreement had been signed by President Ronald Reagan of the United States and Prime Minister Brian Mulroney of Canada on January 2, 1988. A year earlier, the Bank of Canada had dedicated monetary policy to the achievement of a zero inflation rate, raising interest rates several percentage points above those in the United States. Both policies led to a massive influx of capital that pushed up the Canadian dollar by 30 percent. The current account, which had been in approximate balance in the first half of the 1980s, went into a deficit of over 4 percent of GDP. By 1992 the restrictive monetary policy was relaxed, but only after the economy had moved into severe depression, with unemployment, above 10 percent, higher than in any other G-7 countries with the exception of the United Kingdom. The policy mistake was in relying on an appreciating exchange rate as the main vehicle for enforcing (without success) price stability. An analogous situation arose in the United Kingdom. The announcement of completion of the common market in the Economic Community led to a massive influx of capital into London and a real estate boom. The pound, which had been as low as U.S.S1.05 in 1985, soared, going above U.S.S1.90 when the United Kingdom made the fateful decision to enter the exchange rate mechanism (ERM) of the European Monetary System (EMS), at the high rate of DM 3 per pound. The capital inflow, combined with the overvalued pound, led to soaring imports and sluggish exports, and led to the largest recession in postwar history, with unemployment exceeding 10 percent. When the German unification shock struck the EMS, the United Kingdom opted out, allowing the pound to depreciate against both the dollar and the mark. The policy mistake was not so much in entering the ERM of the EMS, but in entering it at a greatly overvalued rate, temporarily pushed up by an influx of capital.
The problem of capital imports
Stabilization failures cannot be understood without consideration of the role played by capital movements. But the benign or harmful effect of capital movements in stabilization and liberalization plans has been a subject of considerable dispute. One of the most controversial questions has
Stabilization policies in semi-open economies
25
been whether liberalization of trade should be accompanied by liberalization of the capital account, or whether the latter should be delayed. The literature seems to be against simultaneous liberalization of the capital account. Among the famous cases where capital imports have allegedly upset stabilization plans, one must include Korea in 1966-70, and Chile, Argentina, and Uruguay in 1978-82. The situation of Chile came as close to a controlled experiment as is likely to exist in economics. The combined liberalization and stabilization programs seemed to be upset by the capital imports that inundated Chile in 1978-82. That country's fiscal deficit had reached 25 percent of GDP in 1973, but it was brought under control by 1977 and 1978, making financial stabilization for the first time feasible. At the same time Chile moved toward free trade: Nontariff barriers were first converted to tariff equivalents, and then the tariff rates were systematically lowered. From an average of 94 percent in 1973, average tariff rates fell to 44 percent at the end of 1975, 33 percent at the end of 1976, 18 percent at the end of 1977, 12 percent at the end of 1988, and 10 percent after 1979. Exchange rate stabilization proceeded in steps: First, multiple exchange rates were unified into a single rate; second, beginning in February 1978, the tablita - a depreciating crawl of the peso at preannounced rates - was introduced, which held depreciation to 24 percent in 1978, an additional 14 percent depreciation in 1979 until, in the middle of that year, the peso was fixed at 39 pesos to the dollar and held at that rate until the collapse of 1982. The inflation rate, which had been 92 percent in 1977, came down to 35 percent in 1980, 20 percent in 1981, and 10 percent in 1982. Real GDP rose at a rate of 9.9 percent in 1977, 8.2 percent in 1978, 8.3 percent in 1979, 7.8 percent in 1980, and 5.5 percent in 1981 (only to fall by 14.1 percent in 1982). The spectacular success of the Chilean stabilization policy was more apparent than real, more transitory than permanent. The deceleration and then fixing of the exchange rate was achieved only at high real interest rates, which, in conjunction with the rise in the marginal efficiency of capital due to trade liberalization, led to massive inflows of capital from abroad. Short-term capital inflows more than doubled to $0.6 billion in 1977, tripled to $1.8 billion in 1978, $1.9 billion in 1979, $3 billion in 1980, and $4.3 billion in 1981. The latter figure for 1981 was larger than Chilean exports of $3.6 billion in that year. These capital imports financed an explosion of imports and a contraction of exports: From imports of $2.9 billion in 1978, imports rose to $3.6 billion in 1982. The current account deficit soared from $0.5 billion in 1977 to $4.7 billion in 1981, equal to 14.5 percent of GDP. What are the lessons from the Chilean episode? One, already noted,
26
Robert Mundell
is that the use of the exchange rate policy as a deflationary device was incompatible with independently rising wage rates; it would have been better to put wages under the tablita along with the exchange rate, as a temporary device during the transition period. Even so, the excessive growth of real wages would have been checked by rising unemployment had it not been for excessive capital imports, drawn in by high real interest rates. In effect Chile's borrowing sustained real wages and living standards beyond its means over the stabilization period. In theory, capital imports should be a help rather than a hindrance; the essence of capital imports is that they allow a country to achieve economic targets earlier than would otherwise be possible. Unfortunately, however, there are some negative externalities. One is that the borrowing goes into consumption rather than investment, permitting the capital-importing country to live beyond its means, pay higher real wages, or finance a budget deficit, without any offset in future output with which to service the loans. Even if the liabilities are entirely in private hands, the government may feel compelled to transform the unrepayable debt into sovereign debt rather than allow execution of mortgages and other collateral. Argentina, in 1990-92, represents another case where stabilization policy has fallen victim to overvaluation of labor paid for by capital imports. To build up confidence in the exchange rate, Argentina enacted a convertibility law that allowed Argentines to use dollars in local transactions, required the central bank to hold a dollar in reserve for every peso in circulation, and forbade the central bank from financing budget deficits; in effect Argentina went on a "currency-board" system. This policy was successful in reducing inflation rates from 800 percent in 1990 to 56 percent in 1991 and to less than 20 percent in the first part of 1992. But it was still too high to keep the Argentine economy competitive at fixed exchange rates; wage rates continued to rise faster than productivity. The trade balance, which had grown to $8.6 billion in 1990, fell to $4.7 billion in 1991 and is expected to be -$1.0 billion in 1992, the first deficit in more than a decade. The current account, which was in surplus by $1.9 billion in 1990, was $2.7 billion in deficit in 1991, and is expected to be much worse in 1992. To counter these alarming trends, the minister of finance, Domingo F. Cavallo, announced, on October 28, 1992, tax and subsidy measures to stimulate exports, partly reversing the free-market deregulation plan he had introduced in March 1990. By November 1992, there has been a flight from the peso and interest rates on short-term peso loans were up to more than 50 percent. These examples raise questions about the usefulness of increased capital imports as part of a stabilization program. Capital imports are neither good nor bad in themselves: They finance an excess of domestic spending
Stabilization policies in semi-open economies
27
over income and, in principle, allow a country to achieve economic objectives sooner than would otherwise be possible; on the other hand, they also increase a country's net external debt. Unless capital imports increase productive capacity by enough to yield a net rate of return higher than debt service, countries do not gain in the long run by capital imports. There are two risks to capital imports. First, it may undermine a monetary stabilization policy; second, it may serve to finance wage increases and therefore consumption spending in the labor sector above the marginal productivity of labor; this is a recipe for bankruptcy. Even if the borrowing is entirely incurred by the private sector, it can build up problems for the country as a whole in the future: In the debt crisis of the 1980s, several governments found it necessary to transform private debt into sovereign debt, nationalizing companies and banks rather than acquiescing in their bankruptcy. Korea's successful stabilization Korea, in 1964, introduced trade reforms and monetary stabilization coupled with a substantial devaluation of the won, from 130 to 255, and then (in 1965) to 271 against the dollar. Inflation was brought down to about 10 percent; interest rates were raised to positive real levels; a fiscal reform led to a doubling of tax revenues; and exports and investment grew rapidly. These measures, in conjunction with the trade liberalization, had raised the marginal efficiency of capital and led to substantial short-term capital imports, threatening to appreciate the exchange rate. The authorities had three alternatives: (1) to let the exchange rate appreciate; (2) hold the exchange rate and allow a corresponding monetary expansion; and (3) control capital imports. The first alternative would have immediately reduced the competitiveness of Korea's new industries in export market; the second would have brought on a return to inflation; and the third would have been inconsistent with the liberalization plan. The second option was chosen, the influx of foreign exchange reserves was monetized, domestic spending increased, and the current account deficit soared. Inflation returned but overvaluation was prevented by a slow depreciation of the currency. Successful stabilization in Korea had to await the 1980s. The Korean stabilization policy of 1981-89 stands out as an example of successful stabilization policy. It was successful on several grounds. Prior to 1982, Korean inflation had typically hovered around 20 percent; but from 1982-89, it was always less than 7 percent. Throughout this period GDP at constant prices achieved the remarkable average growth rate
28
Robert Mundell
of 9.7 percent, the most remarkable performance of any economy in the 1980s.5 What was responsible for the Korean success in contrast to the failure of so many other stabilization plans? One answer is that Korea avoided the mistakes of using the exchange rate as the active instrument for disinflation and thus avoided overvaluation. Over the early period, from 1981 to 1984, the exchange rate was managed by a downward crawl at a rate of depreciation that slightly exceeded the inflation rate.6 At the same time interest rates were cut to anticipate the declining inflation rate. The policy of exchange depreciation was especially important in these years, 1981-85, when the dollar was the strongest currency in the world. In the following years, from 1985-89, three additional external factors contributed to the Korean success. First, the price of oil fell drastically, improving Korea's terms of trade and raising the marginal productivity of capital in Korea. Second, interest rates declined from their peak and lowered Korea's debt-service problem. Third, the dollar declined, allowing Korea to maintain its competitiveness despite some pressure from the United States for appreciation. Korea's success was therefore due to sound policy management in combination with fortuitous external circumstances. Policy gets the credit for the early years of success; and external circumstances for the last half of the decade. In the first phase, Korea not only balanced its budget and applied monetary restraint, but it also was able to lower nominal interest rates pari passu with the reduction in the inflation rate; meanwhile the won was allowed to depreciate to prevent overvaluation. The continuation of the growth boom was definitely helped by changes favorable to Korea in the world environment. Recent developments in Korea
Since 1981 the Korean economy has been in what appears to be a perpetual boom, with a high and even rising share of investment in GDP that reached 39 percent in 1991. From 1982 through 1989, this performance was accompanied by price stability in the sense of an inflation rate averaging 5 percent and a currency that appreciated slightly against the dollar, from an average of 731 per dollar in 1982 to 671.5 per dollar in 1989. Over this period, the budget was in essential balance and the current account balance continually improved, moving into a surplus, equal to 4.4 percent of GDP in 1986. The current account surplus rose to 7.5 percent of GDP 5
6
After a dip to 6.2 percent in 1989, the real GDP growth rate recovered to 9.2 and 8.4 in 1990 and 1991. See McKinnon (1991: 78) for a discussion of this point.
Stabilization policies in semi-open economies
29
in 1987, and 8.1 percent of GDP in 1988 (equal to $14.2 billion), but falling to 2.4 percent of GDP in 1989. From 1989, however, two worrisome features appeared in the Korean economy. One was an increase in the rate of inflation to 10.7 percent in 1990 and 10.8 percent in 1991, falling slightly, however, in 1992. The other was the turnaround in the current account. The surpluses in the four years 1986 to 1989 - thought by some to be "structural" - turned into deficits of 0.9 percent of GDP in 1990 and 3.1 percent in 1991; it is expected to be in deficit also in 1992 and 1993. An even larger turnaround occurred in the trade balance: From $11.5 billion in 1988, it fell to $4.6 billion in 1989, to a deficit of $2.0 billion in 1990, and a deficit of $7.0 billion in 1991. In the period 1989 through 1991, the fall in exports to the United States and Japan was more than made up by the rise in exports to the rest of the world, so that Korea has managed to increase its global exports in those years; imports from the United States, Japan, and the rest of the world, however, rose much more rapidly than exports. Due to the current account deficit, Korea's net external debt has been rising. Gross external debt rose from $33.1 billion in 1989 to $38.5 billion in 1991, and is expected to rise to $52.8 billion at the end of 1992. On the other hand, Korea's overseas assets totaled $27.5 billion in March 1992 so that net external debt was only $ 13.5 billion (but expected to rise substantially by the end of the year). In 1991, according to the balance-of-payments accounts, Korea received income from investments overseas of $3.4 billion and paid out $3.8 billion on foreign investments in Korea (according to the national accounts, net factor payments abroad were $1.5 billion). A missing component of the analysis relates to the demographic factor. The position of a country's current account is influenced by its rates of net saving and investment, both of which are affected by changes in the age distribution of the population.7 What is often called "structural" balances are in fact endogenous to the population structure. Tighter money, high interest rates, and a falling stock market greeted the liberalization of the capital market in January 1992. The stock market rallied with the liberalization, but then continued its downward track, the Korean Composite Stock Price Index (KOSPI) falling below 560 in June 1992, from its peak of 918.6 in 1989. Current policy
Current economic policy is directed at containing inflation and reducing the exploding current account deficit. To this end the current plan aims to 7
See Mundell (1990, 1992a) for analysis of how demographics affects the balance of payments; and Kim (1992) for a recent examination of stages in the balance of payments in several countries with some focus on Korea.
30
Robert Mundell
keep wage increases to 5 percent, has put curbs on imported goods, and encouraged the substitution of domestic-made machinery. The Bank of Korea is trying to maintain a restrictive monetary policy in order to cool down the country's overheated economy and to stabilize prices, keeping the growth of the broad measure of the money supply (M2) within 18.5 percent, necessary to achieve the target inflation rate below 9 percent (with real GDP growth of 7 percent). The two main problems with the economy relate to the persistence of moderate inflation and the deterioration of the current account. In both these respects the performance of the economy has been worse than that achieved in the 1980s. A major source of the difficulty is the influx of foreign capital that simultaneously undermines monetary stability and finances the excess of expenditure over income that constitutes the current account deficit. The dilemma is that, on the one hand, more restrictive monetary policies are needed to bring down the inflation rate; but, on the other hand, tight money will attract more capital and worsen the current account. In principle, an open capital market is likely to be conducive to economic efficiency. Yet capital imports have negative side effects, as most of the heavily indebted countries now realize. The private sector has a tendency to incur debt that is often later turned into sovereign debt, as an alternative to bankruptcy. Real GDP in Korea has been growing more rapidly than nearly every other country, fed by investment ratios near 40 percent of GDP. With wage rates, of a skilled labor force of 19 million, in the range of only $4.50 an hour, the return to capital remains extremely high. It may be better for Korea to rely less on capital imports even if it would mean reducing the share of investment in GDP to 35 percent. A small reduction in the rate of growth of GDP might not involve much of a reduction in growth of GNP, if debt service is taken into account. In the late 1980s, Korea seemed to have arrived at a new "stage" in its balance of payments, with cumulative current account surpluses amounting to $33.7 billion in 1986-89; this surplus enabled Korea to reduce its net debt to $13.5 billion, and its net debt service to about $1 billion. Already in 1986 Korea had become a "mature-debtor-lender" with an export surplus, and was fast approaching a position where it was no longer a creditor. But the turnaround in policy in 1989-92 has put the shift to creditor status a long way off. Was Korea's strong external position in 1986-89 merely a lucky confluence of the three favorable external circumstances - low dollar, low interest rates, and low price of oil? The answer is no; the "three lows" are even lower today! The only unfavorable external factor for Korea was the U.S. slump and the emergence of a growth slump in Japan. But this was
Stabilization policies in semi-open economies
31
more than made up by the favorable situation in Europe. The unification of Germany brought about a reversal of the German trade balance; and Korea was able to increase its exports to Europe. At the same time the prospects for expanding trade with China appear better than ever before. On balance, then, the external situation has been favorable to Korea and will probably be increasingly favorable with the progress of the global recovery. The demographic factor should also be taken into account in assessing the stages of Korea's balance of payments. Age distribution affects the balance of payments because different generations have different balances of investment and saving. Dependents - both young and old - are net consumers; whereas the working age population are net producers. But the ratio of dependents to workers - the "dependency ratio" - is only part of the demographic story and not the most important part. More important is the difference between the investment and saving patterns of workers of different generations. While both "junior" and "senior" workers save substantially, junior workers invest more (on houses, consumer durables, etc.) than they save and are thus net borrowers; whereas senior workers save more than they invest and are thus net lenders. Demographic shocks (e.g., a baby dearth or boom) create population bulges in one group or the other that introduce a cyclical pattern to domestic spending relative to income - and therefore the current account - that reverberates over the generations. It would be a mistake to evaluate Korea's prospects for generating future current account surplus without an examination of the two key demographic variables: the dependency ratio and the juniorsenior work ratio. "How does Korea fit into the demographic cycle?" is a question crying out for an answer. Various conditions have to be met for Korea to turn its current account around; they relate to the goods and services market, the capital market, and the money market.8 One way for Korea to turn its current account around is to move toward increased net capital exports, partly by encouraging Korean direct capital exports, partly by withholding taxes on unwelcome short-term capital imports, and partly by the development of a larger foreign exchange reserve invested in interest-bearing dollar assets. This would almost certainly require also a fall in the real exchange rate, which, in practice, could only be accomplished by a fall in the nominal exchange rate, an increase in the number of won per dollar. Currency depreciation always carries the risk of an increase in the rate 8
To improve the current account, it is simultaneously necessary to increase (1) the gap between GNP and domestic expenditure; (2) the capital and/or reserve account; (3) the sum of the domestic excess demands for securities and money; and (4) the rest-of-the-world counterparts for these categories. For a discussion of the sixteen macroeconomic relation-
32
Robert Mundell
of inflation. With respect to exchange rate policy, it is always desirable to steer a balanced course between the Scylla of overvaluation and uncompetitiveness and the Charybdis of undervaluation and inflationary pressure. The Korean won appreciated significantly against the dollar from an average of 870 in 1985 to 671 in 1989; since 1989, it has depreciated. The current won is about 15 percent higher against the dollar than it was in 1985 despite the fact that prices have risen more rapidly than in the United States and wages have soared. A lower won would almost certainly be a necessary concomitant of renewed net capital exports and a turnaround in the current account.9 References Brock, Philip L. (ed.). (1992). If Texas Were Chile: A Primer on Banking Reform. San Francisco: ICS Press. Brock, Philip L., Michael B. Connolly, and Claudio Gonzalez-Vega (eds.). (1989). Latin American Debt and Adjustment: External Shocks and Macroeconomic Policies. New York: Greenwood Press. Corbo, Vittorio. (1985). Reforms with Macroeconomic Adjustment in Chile during 1974-84. World Development 13: 893-916. (1986). The Role of the Real Exchange Rate in Macroeconomic Adjustment: The Case of Chile, 1973-82. Paper presented at the Central Bank of Ecuador Conference on Trade Liberalization, Quito, January 1992. Corbo, Vittorio, and Jaime de Melo (eds.). (1985). Liberalization with Stabilization in the Southern Cone of Latin America. World Development 13: 863-66. Cuadra, Sergio de la, and Salvador Valdes. (1992). Myths and Facts about Financial Liberalization in Chile. In Brock (1992): 11-101. Dornbusch, Rudiger. (1992). Disinflation and Real Exchange Rates. Mimeograph copy of paper presented at the University of Chicago International Economics Workshop Conference, October 24. Dornbusch, Rudiger, and Yung Chul Park. (1987). Korean Growth Policy. Brookings Papers on Economic Activity 2: 389-444. Edwards, Sebastian. (1984). The Order of Liberalization of the External Sector in Developing Countries. Princeton Essays on International Finance, no. 156, Princeton University. (1985). Stabilization and Liberalization: An Evaluation of Ten Years of Chile's Experiment with Free Market Policies, 1973-84. Economic Development and Cultural Change 33 (January): 223-54. Edwards, Sebastian, and Alejandra Cox-Edwards. (1987). Monetarism and Liberalization: The Chilean Experiment. Cambridge, Mass.: Ballinger. Kim, Sang Kyom. (1992). The New Approaches to Stages in Analysis of the Balance of Payments. Ph.D. diss., University of Pennsylvania.
9
ships that must be changed when the current account is altered, see Mundell (1989a, 1989b, 1991a). To the extent that Korea is free to do so, the authorities should allow the won to depreciate against the dollar and perhaps also the yen and the mark even if those currencies depreciate against the dollar.
Stabilization policies in semi-open economies
33
Lizondo, S., and P. Montiel. (1989). Contractionary Devaluation in Developing countries. IMF Staff Papers 36 (March): 182-227. McKinnon, Ronald I. (1991). The Order of Economic Liberalization: Financial Control in the Transition to a Market Economy. Baltimore: Johns Hopkins University Press. Montiel, P., and J. Ostry. (1991). Macroeconomic Implications of Real Exchange Rate Targeting in Developing Countries. IMF Staff Papers 38 (December): 872-900. Mundell, Robert A. (1989a). Trade Balance Patterns as Global General Equilibrium: The Seventeenth Approach to the Balance of Payments. Rivista di Politica Economica 79, no. 10 (September): 9-60. (1989b). The Global Adjustment System. Rivista di Politica Economica 79, no. 12 (December): 351-464. (1989c). Latin American Debt and the Transfer Problem: Introduction. In Brock, Connolly, and Gonzalez-Vega (1989): 1-17. (1990). The International Distribution of Saving: Past, Present and Future. Rivista di Politica Economica 80, no. 10 (September): 5-56. (1991a). The Great Exchange Rate Controversy: Trade Balances and the International Monetary System. In Fred Bergsten (ed.), International Adjustment and Financing: The Lessons of 1985-1991. Washington, D.C.: Institute for International Economics, 187-239. (1991b). The Overvalued Canadian Dollar. In Fakhari Siddiqui (ed.), The Economic Impact and Implications of the Canada-U.S. Free Trade Agreement. Lewiston: Edwin Mellon Press, 75-112. (1992a). Fiscal Policy and the Theory of International Trade. In Herbert Giersch (ed.), Money, Trade and Competition: Essays in Memory ofEgon Sohmen. Berlin: Springer-Verlag, 145-63. (1992b). Stabilization Policies in Less Developed and Socialist Countries. In Emil-Maria Claassen (ed.), Stabilization Policies in Less Developed and Socialist Countries. San Francisco: International Center for Economic Growth, 122-38. (1992c). The Quantity Theory of Money in an Open Economy: Variations on the Hume-Polak Model. In Jacob Frenkel and Morris Goldstein (eds.), International Financial Policy: Essays in Honor of Jacques J. Polak. Washington, D.C.: International Monetary Fund, 167-74.
CHAPTER 2
Monetary regime choice for a semi-open country Jeffrey A. Frankel
It is natural that a country that industrializes will also begin to liberalize its goods and financial markets. As it seeks to move more fully into the international community of industrialized countries, it will be called upon to allow ever more aspects of its economy to be determined in the marketplace rather than by the government. But it does not follow that every aspect, every macroeconomic variable, should be determined by the marketplace. To focus on a clear example, the exchange rate should not necessarily be determined in the marketplace. Letting the exchange rate float makes more sense if the monetary authorities have decided to fix the money supply (or other nominal quantity). But an equally admissible alternative plan is to fix the exchange rate and let the money supply do the adjusting. One must choose among equally plausible regimes. To make this point is not to knock down a "straw man." The U.S. Treasury has in recent years advised newly industrialized countries (NICs) in East Asia that free-market principles necessarily imply free-floating exchange rates.1 Free-marketeers Milton Friedman and Beryl Sprinkel might agree with that choice, but free-marketeers Robert Mundell and Jack Kemp would not. 1
Democracy, discipline and deficits
This chapter reviews choices among regimes facing a relatively small, trade-oriented, liberalizing, industrializing country. We begin by observing that the problem is neither interesting nor realistic unless due allowance is 1
In October 1988 the U.S. Treasury, in its "Report to the Congress on International Economic and Exchange Rate Policy" required by the Omnibus Trade and Competitiveness Act of 1988, concluded that Korea and Taiwan "manipulated" their exchange rates, within the meaning of the legislation. Financial policy talks with Korea followed, in February and November 1990.1 discuss recent liberalization of Korean financial markets and foreign exchange markets, and the role of U.S. pressure, in Frankel (1993b). 35
36
Jeffrey A. Frankel
made for market failures, such as sticky prices, as well as political failures, such as populist spending binges. On the one hand, if there are no sticky prices, frictions, or other market failures, then it follows that everything should indeed be left up to the market.2 If there are no issues of political economy, on the other hand, then the government should retain complete discretion, so as to be free in the future to move all levers in optimal response to the latest developments. Issues of political economy are particularly relevant if a country is undergoing a transition to democracy at the same time as its economic transition. It is not that an authoritarian government is more likely to produce good economic policies than a democracy.3 Authoritarians frequently meet neither of the two criteria one wants from a philosopher king: being well informed and being well intentioned. But democracies are routinely subject to certain pressures in their economic policy making. Awareness that this is so must heavily condition the problem of what regime to choose in advance.4 Indeed the argument for the government precommitting to any regime, rather than retaining short-term discretion, rests on the existence of these pressures and the need for discipline to resist them. To be more specific, there are good political economy reasons for making (1) a precommitment not to inflate, and (2) a precommitment not to overborrow. The first problem, a bias toward excessive monetary expansion, was amply demonstrated in the worldwide inflation of the 1970s. It gave rise to a burgeoning literature on the desirability of time-consistent rules for monetary policy, that is, credible precommitments to a nominal anchor. It also gave rise to a declaration by major central banks of an allegiance to monetarism. But the 1980s left many central bankers disillusioned with monetarism, and the question of the optimal nominal anchor is still an open one. The second problem, overborrowing, was amply demonstrated in the international debt crisis of 1982. In theory, openness to international capital flows offers enormous advantages: the ability to borrow abroad to finance development of a country where the rate of return to investment at home is high, the ability to smooth spending out over recessions and other short-term fluctuations in income, and the ability to diversify risk internationally. In practice, the option to borrow or lend internationally is misused as often as it is used in the optimal way that our theories assume. 2
3
4
The authorities still have to choose a money supply. But in a sufficiently perfect world, the choice makes no difference. Economists have begun to tackle issues like these that they used to leave to political scientists. Barro (1989), e.g., finds in a cross-section of countries that a measure of the extent of political rights is positively correlated with growth. On populism, see Dornbusch and Edwards (1991).
Monetary choice for a semi-open country
37
One has only to observe that countries tend to borrow internationally when they are undergoing temporary booms, and to pay back in downturns, to realize that the theories of intertemporal optimization are missing something. The explanation for such procyclical borrowing probably lies in the nature of the supply of funds from imperfectly informed lenders and in the political economy of the local groups who get to spend the money.5 Other possible sources of imperfection in international capital markets include flows motivated by tax evasion, speculative bubbles, contagion, and the lack of an international enforcement mechanism in the event of default. The point is that a case could be made for keeping controls on capital inflows, in order to avoid the temptation to overborrow. The argument for precommitting not to inflate and the argument for precommitting not to borrow abroad could be seen as two components of a more general precommitment not to run an excessive government budget deficit. Such deficits can be financed either by monetization/inflation or by foreign borrowing. In a country with sufficiently developed domestic financial markets, they can also be financed by borrowing from domestic residents. The United States sought a fiscal commitment mechanism in the Gramm-Rudman-Hollings legislation. The members of the European Community sought a fiscal commitment mechanism in the terms of the Maastricht Treaty. Both experiments have to date been unsuccessful. Taking our cue from the G-7 countries, we will assume in the remainder of this chapter that precommitments against domestic borrowing and international borrowing are not practical, presumably because the advantages of being able to run deficits at times are too great. We take as given that the country in question is opening up its capital markets. There do exist, after all, a few countries like Korea that have tended to exhibit the self-control necessary to avoid overborrowing. (Korea in the 1970s mostly used its international borrowing for high-return investment rather than private or government consumption, and in the 1980s did not wait for international bankers to cut off lending before taking the measures to adjust to higher world borrowing costs.) Korea appears, in any case, to have embarked on a path of financial liberalization. We henceforth focus, rather, on the choice of exchange rate and monetary regimes, taking financial liberalization as given. 2
The choice of fixed versus flexible exchange rate
The debate between adherents of fixed or flexible exchange rates is often phrased as a choice between absolutes. But the optimal currency area literature introduced thirty years ago by Mundell (1961) and McKinnon 5
E.g., for the case of commodity-producing countries, Cardenas (1991).
38
Jeffrey A. Frankel
(1963) demonstrated clearly that one choice cannot be right for all countries. It does not seem sensible for an extremely small open country or province to have an independent currency. This point has been illustrated anew in recent years by plans for European Monetary Union, although Europe has also demonstrated (in the 1992 crisis) the practical difficulty of knowing when a country is in fact sufficiently open to give up its monetary independence. We will review the advantages of flexible exchange rates, and then the advantages of fixed exchange rates. We will take special note of the aspects of a particular country that determine which set of advantages is likely to dominate. 2.1
The advantages of flexible exchange rates
The advantage of flexible rates is that, freed of the obligation to keep the exchange rate fixed, monetary policy can respond independently to disturbances. When a country opens up its financial markets to international capital flows, the point becomes stronger. Monetary policy becomes a powerful instrument. A monetary expansion under floating exchange rates has much of its effect via the international channel - a depreciation of the currency and the resulting stimulus to net foreign demand - supplementing the traditional channel of a lower real interest rate and resulting stimulus to domestic demand. The Mundell-Fleming model originally showed that the more highly mobile is capital, the stronger this effect is. In the limit of perfect capital mobility, the expected rate of return in the domestic country is tied to the world rate of return. If exchange rates are fixed, there is no scope for monetary policy to have independent effects at all. In that case, a flexible exchange rate is a sine qua non of monetary independence. How important is it to have an independent monetary policy, and thus by implication to have a flexible exchange rate? This depends on two questions. (1) How often does the domestic country experience a disturbance that calls for a response that differs from what is occurring among its neighbors? (2) If an independent monetary response (a reduction in interest rates or a devaluation) is not an option, what alternative means of adjustment are there? The first question in turn subdivides into two questions, (la) To what extent does the domestic country experience shocks that are different from those experienced by its neighbors? Here the extent to which the economies are integrated by trade is key6 (lb) When the domestic country expe6
Some economists assume that an increase in trade between neighbors reduces the correlation between their shocks, because it increases specialization, and thus ironically makes the Optimum Currency Area less likely to hold (e.g., recently, Bayoumi and Eichengreen
Monetary choice for a semi-open country
39
riences a shock similar to that of its neighbors, to what extent does it wish to respond independently, for example, because of a different priority placed on fighting inflation relative to sustaining output and employment? This is largely a matter of the extent to which the countries have divergent values. If the answer to these questions is "not to a great extent," then the region should be happy to share the monetary policy of its neighbor. But otherwise, it will often find itself, in the aftermath of a shock, wishing to make some sort of independent response or adjustment. The second question concerns alternative means of adjustment, which, if available, might make a deliberate monetary response unnecessary. This question also has two components. (2a) If a region experiences a negative shock, such as a loss in demand for its products, and there is no effective macroeconomic response, can its workers easily move to other regions? Labor mobility across geographic boundaries is the optimum currency area criterion on which the original Mundell (1961) article focused. It depends both on formal barriers to travel and migration, such as those recently relaxed within Europe, and more broadly on linguistic and cultural compatibility. Recent research has shown that when a region of the United States experiences a negative shock, the major means by which markets eventually adjust is not a gradual reduction in wages, but a gradual movement by workers to other regions of the United States.7 (2b) If all other means of adjustment fail (macroeconomic expansion, devaluation, lower wages, and out-migration), is there a supraregional or supranational federal system that will undertake fiscal transfers to the depressed region or country? A federal fiscal system operates in the United States,8 and France fiscally supports the franc-pegging governments of West Africa. But recent developments in Europe suggest that there is not as much political will in the northern countries to make transfers to other countries as advocates of European Monetary Union had hoped. The foregoing cataloging of the various factors that might make an independent monetary policy unimportant or unnecessary shows a common theme. When a region is highly integrated with its neighbors - sharing common disturbances and values, or with easy movement of labor or transfers across its borders - monetary independence is less necessary. We must now ask what advantage there might be in giving up monetary
7 8
[1992] and Blanchard and Muet [1993]). This would be an excellent subject for future research; the question must depend on the reason for the increased trade, and the nature of subsequent shocks. I believe, however, that for most sources of trade and most shocks, an increase in trade tends to increase the correlation of shocks, and thereby to strengthen the argument for pegging to neighbors. Blanchard and Katz (1992). Sachs and Sala-i-Martin (1989) document the transfers from the U.S. federal government that flow automatically to a U.S. state experiencing a downturn.
40
Jeffrey A. Frankel
independence. Even if the usefulness of monetary expansion and devaluation diminishes when there are alternatives, are these not options that are always of some use to retain? We have only considered the advantages of flexible exchange rates. We must now consider the advantages of fixed exchange rates. 22
The advantages of fixed exchange rates
The advantages of a fixed exchange rate, again, fall into two broad categories. (1) First, stabilizing the currency reduces exchange rate uncertainty facing exporters and importers, as well as international borrowers and lenders. The discouraging effect on international trade and finance that exchange rate risk might have was one of the most important arguments used by those who opposed a general move to floating exchange rates before 1973. (2) Second, a fixed exchange rate can serve as an effective nominal anchor for monetary policy, and thus can assure price stability. We consider each advantage in turn. The hypothesized advantages of exchange rate stability per se constitute too large a subject even for the sort of capsule summary we are pursuing here. Critics of the way floating rates have operated among the G-7 countries over the past twenty years have tended to focus more on longer-term "misalignments" rather than short-term volatility. Misalignments such as the 1984-85 overvaluation of the U.S. dollar are perceived to impose longterm costs in the form of protectionist barriers and a diminished capital stock in tradable goods sectors. Key to evaluating arguments regarding either long-term misalignments or short-term volatility is a means of evaluating whether private financial markets, with their occasional speculative bubbles and other possible defects, are more or less likely to produce unneeded or undesirable exchange rate movements or misalignments than is the political process, with all its defects, under a pegged-rate system. This debate is very much unsettled. 2.3
The effect of exchange rate variability on trade
The danger of misalignments is not the major motivation behind European efforts to stabilize exchange rates among themselves. Promoting intra-European trade has been a more important motivation. Economists reviewing the post-1973 record have tended to be skeptical about the effect of exchange rate uncertainty on trade. They point out that markets in forward exchange and other derivatives allow an importer, say, to hedge the risk of an increase in the price of foreign currency. It would be a mistake, however, to think that all exchange rate risk can be hedged
Monetary choice for a semi-open country
41
in this way, even in theory. Although any given importer can hedge his exposure, someone, somewhere, will have to bear some exchange risk, and that person will demand a price to compensate him for doing so. The empirical record on the effect of exchange rate variability on trade since 1973 is mixed. Notwithstanding the high level of volatility in the twenty years since exchange rates began to float, the international volume of trade has grown rather rapidly. Time series studies such as Hooper and Kohlhagen (1978) found only very limited evidence of effects. Some later studies found relatively more effects, but overall surveys of the subject do not present a strong case for an effect on trade. The problem with the time series studies is that other factors have changed since 1973, at the same time as exchange rate variability. (Some factors leading to greater trade over the past twenty years are economic growth, reduced tariff barriers, and possibly lower costs of transportation and communication.) If a quantum change in the level of uncertainty would have to be sustained for a number of years before it could be reliably perceived, let alone before it could lead to a reallocation of resources between traded goods and nontraded goods, twenty years of time series data are perhaps not the most promising place to look. Together with Shang-jin Wei, I have applied to this problem a crosssection data set of bilateral trade flows between 1,953 pairs of countries. We use the gravity model, to explain the volume of bilateral trade (in logarithmic form) by four basic determinants: the sizes of the two countries, their GNP/capitas, the distance between them, and a dummy variable indicating whether they share a common border. In Frankel (1992, 1993), we see how much of the residual can be explained by regional trade groupings, such as common membership in the proposed East Asian Economic Caucus. In Frankel and Wei (1993), we also see how much can be explained by bilateral exchange rate variability. Volatility is defined to be the standard deviation of the monthly first difference of the logarithmic real exchange rate (sd). The equation estimated is as follows. lo g (7\) = a + p,log (GNPfiNP) + ^logiGNP/popfiNPIpop) + $3\og(DISTANCE) + ( J {C) + 5 ADJ is a dummy variable indicating when two countries share a common border. EC, WH, and EA are dummy variables indicating when both countries are located in the same geographic area (the European Community, Western Hemisphere, or East Asia, respectively). The major advantage of this approach is that it brings data from a wide variety of country experiences to bear on the problem. The major
42
Jeffrey A. Frankel
disadvantage is the likelihood of simultaneous causality: If exchange rate variability shows up with an apparent negative effect on the volume of bilateral trade, it could be due to the government's efforts to stabilize the currency vis-a-vis a valued trading partner as easily as the reverse. With this consideration, we also use the method of instrumental variable estimation to tackle the possible simultaneity bias. Ordinary least squares estimation results are reported in Table 2.1. They show a relatively large and statistically significant effect of real exchange rate variability on trade. Consider the hypothetical experiment of a doubling of exchange rate variability, based on the 1980 equation. To give perspective to this experiment, the standard deviation of bilateral changes in the real exchange rate experienced by the average Western Hemisphere country more than doubled between 1980 and 1990 (it increased by a factor of 2.75), whereas in East Asia and Europe it remained roughly unchanged.9 The 1980 OLS coefficient on the log variability is -.066. The implication is that a doubling of uncertainty reduces the volume of trade by an estimated 4.6 percent (= .066 log(2)). It is likely, however, that much of this apparent effect is due to reverse causality. Instrumental-variables estimation results, reported in Table 2.2, show a smaller (and less significant) effect.10 The 1980 coefficient falls to -.010, which implies that a doubling of the standard deviation reduces trade by an estimated 0.7 percent. This effect seems relatively small. 2.4
The exchange rate as a nominal anchor
The second kind of advantage of a fixed exchange rate is that it provides one possible nominal anchor for monetary policy. There are other possible nominal anchors; we consider the issue at length in the next part of the chapter. But one point regarding openness needs to be made here, because it relates to the optimum currency area question. For much of the analysis of optimum currency areas, the distinction between a fixed exchange rate and a common currency is not an important one. (There are some minor issues of transactions costs and seignorage.) But when we consider the credibility of a commitment not to increase the money supply or not to devalue, the distinction becomes more important. 9
10
The level of the standard deviation was roughly .23 percent in all three parts of the world in 1980. These statistics for each country represent a simple unweighted averaging across sixty-three trading partners, and thus do not reflect the greater importance of the larger countries. The instrumental variable is the standard deviation of the logarithmic change in one country's money supply relative to the other. We hypothesize that this variable is correlated with the variability of the nominal and real exchange rate (and indeed it is, in our sample), and we hope that it is uncorrelated with other determinants of bilateral trade patterns.
Table 2.1. Exchange rate volatility and bilateral trade (OLS estimation) Volat
Adj WH EEC EAEC APEC adj. R2 S.E.E.
GNPs GNP/cap
Dist
.74" .02
.29" .02
-.56" .72" .52" .04 .18 .15
.23 .18
.88" .27
1.51" .17
.71
1.20
Nominal Ex rate
-.046" .76" .023 .02
.26" .02
-.68" .27 .05 .21
.16 .23
.03 .18
1.04" .37
1.35" .20
.73
1.20
Real Ex rate
-.066" .74" .029 .02
.27" .02
-.67" .43' .43' .18 .05 .22 .20
.04 .20
.96" .37
1.30" 1.30" .22
.76
1.14
.76" .02
.25" .02
-.70" .75" .33" .33" .44" .04 .18 .16 .17
.59" .26
11.28" .28" .17
.74
1.17
.015 .021
.77" .02
.24" .02
-.74" .61" .23 .05 .19 .18
.43" .17
.79" .36
1.18" .19
.75
1.16
-.026 .0
.76" .02
.24" .02
-.75" .45" .01 -.75" .05 .22 .20
.26* .17
.72" .36
.12" 11.12" .21
.78
1.12
.75" .02
.09" .02
-.56" .79" .92" .04 .16 .14
.47" .16
.69" .24
1.36" .15
.11
1.07
.076" .77" .014 .02
.09" .02
-.66" .61" .82" .04 .16 .14
.54" .16
.75" .33
1.36" .17
.79
1.04
-.048" .79" .023 .02
.11" .02
-.60" .31* .51" .04 .20 .17
.27* .17
.95" .38
1.06" .28
.83
.97
1980
1985
Nominal Ex rate Real Ex rate 1990
Nominal Ex rate Real Ex rate
Notes: (1) All the variables except the dummies are in logarithm; all the regressions have an intercept for which the estimate is not reported here. (2) Standard errors are below the coefficient estimates. (3) a, b, c, and d denote "statistically significant" at the 99, 95, 90, and 85 percent levels, respectively. The dependent variable is the log of bilateral trade.
Table 2.2. Exchange rate volatility and bilateral trade (instrumental variable estimation) Volat
GNPs GNP/cap
Dist
Adj WH EEC EAEC APEC adj. R2 S.E.E.
1980 Nominal Ex rate
-.008* .73" .005 .02
.27" .02
-.56" .74" .54" .04 .18 .15
.20 .18
.93" .27
1.48" .17
.71
1.20
Real Ex rate
-.010* .73" .005 .02
.26" .02
-.56" .75" .56" .05 .18 .15
.22 .18
.94" .27
1.48" .17
.71
1.20
Nominal Ex rate
-.001 .005
.76" .02
.24" .02
-.70" .76" .34* .04 .18 .16
.43* .17
.59* .26
1.28" .17
.74
1.17
Real Ex rate
-.000 .005
.76" .02
.25" .02
-.70" .75" .33* .04 .18 .16
.43* .17
.59* .26
1.28" .17
.74
1.17
Nominal Ex rate
.029" .77" .005 .02
.15" .02
-.57" .71" .88" .04 .16 .14
.44" .16
.47* .24
1.40" .15
.77
1.06
Real Ex rate
.032" .77" .005 .02
.15" .02
-.57" .71" .87" .04 .16 .14
.43" .16
.45' .24
1.39" .15
.78
1.06
1985
1990
Notes: (1) All the variables except the dummies are in logarithm; all the regressions have an intercept for which the estimate is not reported here. (2) Standard errors are below the coefficient estimates. (3) a, b, c, and d denote "statistically significant" at the 99, 95, 90, and 85 percent levels, respectively. The dependent variable is the log of bilateral trade.
Monetary choice for a semi-open country
45
If a country literally shares a common currency with its neighbors, the commitment not to devalue is close to absolute. The recent examples of the disintegration of the Soviet Union and Czechoslovakia illustrate that reintroducing a distinction between one country's rubles and another's is not absolutely impossible. But it is extremely difficult. If a country merely has a fixed exchange rate, but retains a separate currency, the option of devaluation is always there. (The French-speaking West African countries, for example, which have been more completely tied to the currency of their former colonizer than has any other set of countries, in 1994 devalued for the first time.)11 If the option of devaluation exists, speculators and other private actors will be fully aware of it. When considering the possibility of using the exchange rate as a nominal anchor for monetary policy, the question then arises (as it does even more for the other possible nominal anchors to be considered): What can make the commitment credible? David Romer (1991) has argued that a commitment to fix the exchange rate is more credible the more highly open the country is to international trade. The argument is that the cost of reneging on the commitment and devaluing the adverse impact on the price level, for example - will be higher the more important trade is in the economy. He examines carefully a sample of 114 countries, and finds statistical support for the theory that an exchange rate peg is a more credible anchor for countries that are more open. Thus openness (as defined by trade) is a key parameter determining the importance of the advantages to pegging the exchange rate. In the preceding section we saw that economic integration was also a key parameter determining how easily a country could dispense with the advantages of floating. (Recall that integration there could be defined by a sharing of common economic disturbances or of common values, or by easy movement of labor or transfers across national borders.) In short, the balance between the advantages of a fixed exchange rate and the advantages of a flexible exchange rate depends critically on the degree to which the country in question is economically and culturally integrated with its neighbors. In the case of Europe, integration is increasing, but is still well behind the standard set by the states of the United States, as Eichengreen (1990) has shown. Recent developments, which appear to have derailed European Monetary Union, suggest that the residents of the EC 12 are less 1
' On the rules-versus-discretion aspect of the West African monetary union, see Devarajan and Rodrik (1991).
46
Jeffrey A. Frankel
ready to sacrifice their monetary independence than their leaders had thought. In the case of Asia, economic integration with the outside world is relatively high, when defined by trade, despite the existence of some formidable barriers. Trade is a relatively high percentage of GDP for most East Asian countries. This includes a lot of intraregional trade, especially after adjusting for the fact that the East Asian countries are not as close to each other or as high in GDP/capita as, for example, the members of the European Community.12 The high level of intraregional trade may explain why several recent studies have found that economic disturbances are correlated among East Asian countries. Bayoumi and Eichengreen (1992, 17-18) find supply disturbances to be significantly correlated among Korea, Japan, and Taiwan (and also among Hong Kong, Singapore, Malaysia, and Indonesia; among this latter group demand disturbances are correlated as well). These correlations, like those within northern Europe, are found to be "not dissimilar from those found in regional data for the United States" (p. 23), and to be greater than anything found among Western Hemisphere countries. Goto and Hamada (1993), using a principle-components analysis of macroeconomic variables, also find that "East Asia is a group as homogeneous as the European Community (p. 11)," and that real disturbances to investment are correlated among East Asian countries. By other criteria, such as labor mobility, East Asian countries p.re probably less open than European countries, let alone the regions of the United States. 2.5
Some further arguments for exchange rate flexibility in a country like Korea
A few other factors in choosing an exchange rate regime are particularly relevant to economies in East Asia, countries that are liberalizing financially and rapidly industrializing. We consider in turn some implications of the East Asian geographical location, financial liberalization, and rapid growth. Our discussion of openness left out an important difference between 12
Estimates in Table 2.1 show the extent to which bilateral trade flows can be attributed to special regional factors (which presumably include cultural homogeneity, preferential trading arrangements, and other policy initiatives), as opposed to such readily observed natural determinants as size and proximity. For further explication, see Frankel (1993a) and Frankel and Wei (1994). Other recent studies of the bias toward intraregional trade in Asia (and other parts of the world) include Anderson and Norheim (1993), Drysdale and Garnaut (1992), and Petri (1992). Recent studies using the gravity model are Wang and Winters (1991) and Hamilton and Winters (1992).
Monetary choice for a semi-open country
47
the situations in East Asia and Europe. By many measures, such as bilateral trade biases and financial influences, East Asian countries are less closely tied to Japan, or to each other more generally, than they are tied to a Pacific grouping that includes the United States (along with Canada, Australia, and New Zealand).13 This means that, to the extent that they are judged sufficiently open to merit pegging their currencies, it is not clear whether they should peg to the yen or the dollar. Park and Park (1991b) highlight the conflict that yen-dollar fluctuations create for East Asian exchange rate policies. This issue is missing from the traditional literature on optimum currency areas, which usually makes the simplifying assumption that there is a single large neighbor against which the country in question must simply either peg or float.14 We return to the issue of which large trading partner to peg to, in a discussion at the end of the chapter. Next, there is the simple point that a floating exchange rate is a more viable option if financial markets are well developed and internationalized. The two properties, a free-floating exchange rate and free financial markets, are quite distinct.15 Nevertheless the one is made easier by the other. Finally, there is the point that rapidly growing countries are known to experience a trend real appreciation in their currencies (if the growth comes from supply-side factors, such as rapid increases in productivity). "Equilibrium" theorists who in the abstract attribute every observed short-run fluctuation in exchange rates to fundamental real factors like productivity and consumer tastes usually overreach. Nevertheless, the pattern of real appreciation experienced by industrializing countries is systematic and rooted in real factors. The explanation is that the relative price of nontraded goods is low in poor countries (labor and land are cheap), and rises with the stage of development. Sometimes, as with Korea and Taiwan in the second half of the 1980s, the real appreciation comes in concentrated form, as the result of substantial capital inflows that force the central bank to choose between a nominal appreciation of the currency and a potentially inflationary increase in reserve holdings. Some countries may be able to avoid either nominal appreciation or inflation by sterilizing the reserve inflows. The usual view is that this is more likely to be feasible when domestic financial markets are liberalized and well developed: (1) If the country has liberalized with 13
14 15
The references are those in the preceding footnote. Park and Park (1991a) foresee a continuation of the dependence of the East Asian newly industrialized countries (NICs) on the U.S. market. An exception is Marston (1984). A point on which recent U.S. Treasury reports to Congress on the subject of Korean liberalization are less than lucid. Frankel (1993b).
48
Jeffrey A. Frankel
respect to capital outflows, it can reduce the magnitude of the net inflows. (2) If the country has liberalized with respect to domestic bond markets, there is scope for open market sales by the central bank to sterilize reserve inflows. Korea did some of the right things in 1986-89: paying off external debt, and sterilizing reserve inflows by selling monetary stabilization bonds and raising reserve requirements.16 But the actions were not strong enough to prevent inflationary growth in the money supply, and nominal appreciation of the won as well. The absence of active domestic bond markets in which the Bank of Korea might have been able more fully to sterilize its purchases of dollars in exchange for won (by selling domestic bonds in exchange for won and thereby preventing the supply of won in the hands of the public from expanding) has been attributed to the cessation of financial liberalization in the period 1984-87.17 It would follow that further financial liberalization is indeed a good idea for Korea; facilitating sterilization operations in the future is but one of the reasons, so that the central bank can undertake offsetting open market sales.18 With or without well-developed financial markets (and with or without political pressure from large deficit-prone trading partners), a country experiencing sustained rapid productivity growth will eventually have to allow its currency to appreciate in real terms. The implication for the choice of monetary regime is that, if a country hopes seriously to maintain an inflation rate no higher than that of its major trading partners, fixing the exchange rate cannot be a permanent policy; eventually there will have to be an upward revaluation. Reisen (1991) argues on these grounds that the Asian NICs should avoid fixed exchange rates. If a country does opt for floating over fixed exchange rates, that frees up monetary policy for other objectives. But which other objective? Price stability or output stability? (Either way, fiscal policy under a floating exchange rate might best be assigned to look after the trade balance.)19 We 16 17 18
19
See, e.g., Kwack (1994). Kim (1990, 17). Reisen (1993), however, has argued that some Southeast Asian countries have succeeded in sterilizing reserve inflows by using large state-controlled funds to dominate the market, which would presumably not be possible in unregulated and well-developed financial markets. Boughton (1989). It should be mentioned that fiscal policy and monetary policy are not independent policy instruments, and therefore the issue of what targets to assign them does not arise, until a country undertakes sufficient financial liberalization that trade deficits can be financed by borrowing abroad, and budget deficits can be financed by borrowing both at home and abroad. In a more primitive economy, where deficits can only be financed with money, the distinction between fiscal and monetary policy all but disappears.
Monetary choice for a semi-open country
49
turn now to monetary theory's other classic policy debate, rules versus discretion. 3
A nominal anchor for monetary policy
The past twenty-five years of research on the use of monetary policy to affect output and inflation has followed a distinct logical progression. First, in 1969, Friedman and Phelps introduced expected inflation into the Phillips curve. They pointed out that a monetary expansion to raise output would come at the expense of ever accelerating inflation, so that the increase in output could not persist in the long run. Second, Lucas, Sargent, and Barro made the expectations rational. The implication was that policy makers could not have a systematic effect on output even in the short run. They might as well give up on the idea of affecting output and simply aim for zero inflation. Third, Fischer, Taylor, and others introduced contracts that made wages and prices sticky. The result was to return some effectiveness to monetary policy in responding to disturbances, but again only in the short run. 3.1
Rules versus discretion
In the past ten years of monetary theory, the debate on "rules versus discretion" has moved to the center stage of relevant research. Rational expectations in itself did not imply that the government should abandon all discretionary policy; as already noted, there was still scope for responding to disturbances in the short run, provided policy makers acted with sufficient humility and awareness of the long-run implications. There appeared to be no formal basis to arguments such as Milton Friedman's that the government should completely renounce discretion in favor of rules. How could the country benefit from voluntarily giving up a policy tool? Kydland and Prescott (1977) introduced the notion of time consistency, the need for a precommitment that would bind government policy makers and enter private expectations. In the case of monetary policy, a binding precommitment to slow money growth would cause workers and others to reduce their expectations of inflation; the result would be a lower actual inflation rate for any given level of output. At first it seemed that such a precommitment could only improve welfare if, in its absence, discretionary monetary policy were subject to political pressures that aimed for a point on the short-run output-inflation trade-off that was higher than optimal. Such pressures could result because either those who dominated the political process did not understand the longer-run inflationary effects of
50
Jeffrey A. Frankel
expansion, they put a lower value on price stability than was in the national interest, or they had a higher discount rate than was in the national interest. These arguments for insulating monetary policy from populist pressures have some validity in their own right. But Barro and Gordon (1983) showed that discretion could lead to excessive expansion even when the policy makers sought to maximize the "correct" objective function, that is, the correct quadratic loss function in output and inflation. The key to this result is the assumption that the loss function (shared by the policy makers and the country as a whole) is centered around a level of output that is greater than potential output. This assumption dramatically expanded the boundaries of the existing models. The recognition that any country would like a higher level of output if it could have it sounds obvious. Previous authors had felt bound to rule it out on the grounds that, in the long run, a level of output higher than potential is not attainable. But just because the bliss point is not attainable does not mean that the correct objective function is not centered around it. (Technically it requires the existence of some other distortion, such as the existence of unemployment compensation, that artificially raises the natural rate of unemployment or lowers potential output. But there are plenty of those.) Figure 2.1 illustrates the problem. The objective function is assumed to be centered around the point corresponding to zero inflation and output equal to ky*, where y* represents potential output and k>\, capturing the preference for higher output. Isowelfare curves radiate out from that point. In the long run, the supply relationship is vertical at y*. In the short run, for a given expected inflation rate pe, the supply relationship is an upward-sloping line through the point (y*9 pe). In the absence of a credible precommitment (and in the absence of any disturbances), the optimizing government will set aggregate demand so as to pick out the point (B) on the supply curve where it is tangent to an isowelfare curve. The graph makes clear that this corresponds to a positive expected inflation rate (which in turn becomes the actual inflation rate, in the absence of disturbances). The country can do better by making a binding commitment to aim for zero inflation, at point C. Here the economy is on a higher isowelfare line, because inflation is lower, with no loss in output. The superiority of this framework is shown immediately by its ability to explain the fact that virtually all countries experience average inflation rates above zero, lacking, as most of them do, truly binding commitment mechanisms. In the traditional theory, where potential output and optimal output were assumed to coincide, it followed that positive inflation rates (which followed only in the aftermath of positive shocks) were no more
Monetary choice for a semi-open country
51
Optimal A.D. under discretion
A.D. under credible commitment to zero inflation
Y
*
Figure 2.1. The advantage of rules over discretion (in a model with inflation bias but no shocks). frequent than negative ones (in the aftermath of negative shocks). This implication was clearly at variance with reality. The conclusion from the Barro-Gordon model was that governments should not merely announce their intention to aim for zero inflation and ignore output fluctuations, but should actually be bound in such a way as to prevent themselves from straying from this commitment even if subsequent events seem to call for it. This result was the long-missing formal justification for rules over discretion.
52
Jeffrey A. Frankel
The Barro-Gordon model stacked the deck in favor of rules, by leaving out the possibility of short-run disturbances to which the authorities might usefully respond if they were free to do so (just as earlier Keynesian authors had stacked the deck in favor of discretion by leaving out the possibility of a bias toward inflation). The syntheses of Rogoff (1985, 1986), Fischer (1991), and Persson and Tabellini (1990) included both short-run disturbances and a bias toward inflation. The result was a realistic intermediate case, which called for some intermediate degree of commitment to a nominal anchor (or else, in the last section of Rogoff [1985], appointment of a central banker who placed somewhat more weight on the priority of fighting inflation than did the general population). The optimal degree of commitment depended on such parameters as the slope of the short-run supply relationship, the weight placed on the inflation objective, and so forth. 3.2
A Iternative nominal anchors for monetary policy
There are a variety of possible candidates for the nominal variable to which monetary policy might commit: the exchange rate, money supply, price of gold or other commodities, general price index, and nominal GNP or GDP. In most of the models of credible precommitment, it makes no difference what is the nominal anchor in terms of which the commitment is phrased, or even whether the commitment is phrased in terms of a complicated linear combination of nominal variables. These are models in which everyone can with certainty observe accurately such variables as the price level and output, infer in detail what disturbances have occurred, and recognize instantly if the monetary authorities are deviating from their announced commitment. So long as the authorities choose a monetary rule that genuinely gives zero inflation in the absence of disturbances, the public will be able to perceive the sincerity of the commitment. In practice, it is clear that there is a great deal of uncertainty, that central bankers will typically claim they are standing by their commitment (and will attribute any observed deviation from the announced targets to a large unanticipated disturbance), and that the public will have difficulty monitoring the authorities. It follows that only commitments that are simple and are phrased in terms of an observable variable can be monitored. As soon as an unanticipated disturbance occurs, it makes a great deal of difference which variable was chosen for the commitment. A commitment to the wrong nominal anchor can unnecessarily increase the costs of abiding by the commitment when the disturbance is realized. What are the proper grounds for choosing among candidates for the nominal variable to which the monetary authorities commit? The appendix to this chapter considers the problem formally. It makes no judgment
Monetary choice for a semi-open country
53
on the desirable degree of precommitment to a nominal target. But whatever the degree of precommitment to a nominal target, we argue that nominal GDP (or nominal demand) is likely to make a more suitable target than the three other nominal variables that have been proposed: the money supply, the price level, or the exchange rate. The general argument has been made well by others.20 In the event of disturbances in the banking system, disturbances in the public's demand for money, or other disturbances affecting the demand for goods, a policy of holding nominal GNP steady insulates the economy; neither real income nor the price level need be affected. In the event of disturbances to supply, such as the oil price increases of the 1970s, the change is divided equiproportionately between an increase in the price level and a fall in output. For some countries, this is roughly the split that a discretionary policy would choose anyway. In general, fixing nominal GNP will not give precisely the right answer, depending on the weights on inflation and real growth in the objective function. But if the choice is among the available nominal anchors, nominal GNP gives an outcome characterized by greater stability of output and the price level. The appendix begins by showing that a nominal GNP target strictly dominates a money supply target, in the sense of minimizing a quadratic loss function, regardless how important inflation-fighting credibility is. The point can also be made in terms of Figure 2.2, which illustrates several alternative nominal anchors, all of them set so as to produce zero inflation in the case of zero disturbances. In the case of the nominal GNP rule, any demand-side disturbances are automatically offset, so that the aggregate demand curve is held steady. The range of variation of output is relatively narrow, resulting only from the inevitable aggregate supply shocks. In the case of the money rule, shifts in the aggregate demand relationship gratuitously increase the range of variation of output (and of the price level). We next consider a price level rule. Central banks have long stated that price stability is a central objective. The Bundesbank has "the aim of safeguarding the currency" as the single ultimate goal written into the institution's charter but, as a strategy, puts more emphasis on annual target rates of change of the money supply than of the price level. (Similarly, the Bank of Japan announces "projections" of M2 money growth; but econometric analysis of the bank's behavior suggests that these are not monetarist targets.)21 Canada recently gave legal status to explicit annual price level targeting. New Zealand has in a sense gone the farthest, by writing the cen20
21
Tobin (1980), Bean (1983), Meade (1984), Gordon (1985), Hall (1985), Taylor (1985), and McCallum (1987, 1988), e.g., argue in favor of targeting nominal GNP in the closed economy context. Williamson and Miller (1987, 7-10) propose targeting nominal demand as part of their "blueprint" for exchange rate target zones. Ito (1989) and Hutchison and Judd (1992).
54
Jeffrey A. Frankel AS. with positive shock AS. with no shock
AS. with negative shock
AD. with no shock (or shocks offset automatically)
Y
Range of variation of output:
*
under nominal GNP rule
under pnce level rule under money supply rule
Figure 2.2. A comparison of three rules: money supply, price level, and nominal GNP (in a model with shocks).
tral bank governor's contract so that his salary is tied directly to his success at eliminating inflation. The appendix shows that the price level rule dominates the money rule, so long as any weight at all is placed on the inflation objective. Figure 2.2 suggests why: The price level rule eliminates the effects of demand disturbances.
Monetary choice for a semi-open country
55
The price level rule and nominal GNP rule have this latter attraction in common. Which is better? Figure 2.2 shows that the nominal GNP rule has the advantage of narrowing the range of variation of output. Not surprisingly, the price level rule has the advantage of narrowing the range of variation of the price level. One cannot say for certain which advantage is more important. The appendix derives the condition under which the nominal GNP rule is the superior one. The condition is likely to hold unless the short-run supply relationship is believed to be very steep, or very low relative weight is attached to the output objective. The second half of the appendix introduces an exchange rate target as a candidate for nominal anchor. It shows that the penalty that goes with a regime of stabilizing the exchange rate is to be saddled with a monetary policy that destabilizes the overall price level, relative to a regime of stabilizing nominal GNP The conclusion within this framework is that, to opt for a fixed exchange rate regime, one has to put very high weight on the objective of stabilizing the exchange rate. It is natural for the weight on the exchange rate objective to be higher for a relatively small open economy like Hong Kong, than for a large, relatively self-sufficient economy like the United States. (The reasons are those already given in the discussion of optimum currency areas.) Nevertheless, the model in the appendix gives the result that, for the exchange rate rule to dominate the nominal GNP rule, one has to be prepared to argue that a 10 percent fluctuation in the exchange rate causes greater trouble than a 10 percent fluctuation in the price level. It is unlikely that this condition would be met even in a very open economy. The result seems a little too strong. Perhaps something has been left out of the model? Many things are, of necessity, omitted from a model of this sort. The possibility of speculative bubbles has been left out. One could rescue the exchange rate rule by assuming that much of the disturbances in the exchange rate equation will disappear when the regime changes, rather than having to be accommodated by the money supply.22 Perhaps the most important factor that is left out is the difficulty of monitoring the government's actions to fulfill its commitment. The model assumes that if the commitment takes the simple form of pegging a single nominal variable, the government can succeed in doing this exactly, and the public can instantly observe that it is doing so. These assumptions clearly do apply to a simple exchange rate peg, but do not apply to the other candidates for nominal anchor. 22
Williamson and Miller (1987, 54-55; Miller and Wirliamson 1988) do precisely this: assume that there is a large "fad" component to exchange ratefluctuationsunder the current floating regime, and that it would disappear under their target zone proposal. The idea is not absurd. But it certainly "stacks the deck" in any comparison of the two regimes.
56
Jeffrey A. Frankel
3.3
Monitoring commitment to the nominal anchor
It is easy to rank the variables that are nominal anchor candidates according to how frequently they are reported, and therefore how quickly the public can become aware of a deviation from the promised path: The exchange rate is available virtually continuously, the money supply on a weekly basis (in the United States and some other countries), the price level on a monthly basis, and nominal GDP only on a quarterly basis. (Nominal GDP is furthermore often subject to substantial subsequent revisions.) The ranking according to how well the monetary authorities can control the variable in question is similar, with one exception: They can probably control nominal GDP more directly than the price level, under the assumption that their only influence on the price level comes via their influence on aggregate demand. But, to repeat the argument for a nominal GDP target, there is little point in committing to, say, a money supply target merely because the target can be attained; if the variable is far removed from the ultimate objectives (output and inflation), then future shocks will lead one to regret ex post having chosen that nominal anchor. In practice, proposals that the money supply, nominal GDP, or price level should be chosen as nominal anchor do not intend a claim that the variable can be pegged exactly, but rather that it should be controlled so as to lie within a specified target zone for the year. (The same could be done, of course, for the exchange rate. Such target zones have been the basis of both the European Monetary System and a sizable related academic literature.) One could formalize the notion of credible commitment to a target zone for any of these variables. The authorities announce that they will guide the variable in question so that, despite the known existence of shocks with certain variances, the variable will fall within the specified zone 95 percent of the time. If the authorities do the statistics correctly, setting the width of the band appropriately, and then carry out their commitment faithfully, the public can readily test their performance statistically. To illustrate, if nominal GDP were observed to fall outside the specified band two years in a row, the probability of this happening by chance, that is, because of unusually large disturbances, would be so small as to be negligible: (.05)(.05) = .0025. The public would be justified under these circumstances in concluding that the commitment was not genuine. It follows that the commitment can be made credible in advance, provided the band is set sufficiently wide and the central bank faces a perceived penalty (public embarrassment) for violating the target zone.
Monetary choice for a semi-open country
3.4
57
Basket pegs for East Asian countries
We mentioned in Section 2.3 the problem, facing Korea and other East Asian countries that might contemplate pegging their currencies, that their trade is heavily split between the United States and Japan. Given the large variation in the yen-dollar rate, a peg to the dollar creates substantial variability vis-a-vis the yen, and vice versa. The standard advice to such countries is to peg their currencies to a basket of major currencies, weighted according to shares of trade (or other more sophisticated formulas). A country that follows such a policy will eliminate uncertainty regarding the future value of its effective exchange rate.23 Indeed Malaysia and Thailand have been officially classified by the IMF as pegging to a currency composite. Korea had an announced policy in the 1980s of setting the won with reference to a basket (though an "alpha" term allowed departure from the basket). In theory, a commitment to peg to a basket should be just as effective a nominal anchor as a commitment to peg to a single currency. Also in theory, it makes no difference whether the government announces the weights in the basket. (Most countries in fact do not announce the weights.) In practice however, these differences can be important from the standpoint of the ability of the public to monitor the authorities' faithfulness to their commitment.24 If a country truly followed a precise basket peg, without crawl, realignment, minor variation inside a band, or changes in the weights, it would be easy for an observer to verify the commitment. If there were, say, ten major currencies that might appear in the country's basket, it would take only eleven observations of actual exchange rates to estimate the ten weights (provided the ten currencies in question moved vis-a-vis each other during the sample period), and a twelfth observation to verify that the peg was being precisely maintained. In practice however, few countries follow such a literal basket peg. When one seeks to estimate the implicit weights in an econometric equation for the value of the local currency, one should in theory get a perfect R2. Thailand comes very close for the period January 1991 to May 1992, with an R2 of .99 and estimated weights of about .8 on the dollar, .1 23
24
This does not succeed in eliminating bilateral exchange rate variability with major trading partners of the sort examined in Tables 2.1 and 2.2. The question whether it is variability in effective or bilateral exchange rates that discourages trade has been insufficiently researched. It depends on whether the exchange rate uncertainty that matters comes at the stage when the firm decides to invest resources in tradable goods production, or at the stage when it agrees to contract with a specific customer in a specific currency. See, e.g., Lowell (1992) and Takagi (1988).
58
Jeffrey A. Frankel
on the yen, and . 1 on the mark. But Thailand fails the test for the preceding two-year period. Malaysia also comes close, with an R2 of .94 (when allowing for a statistically significant trend appreciation) and estimated weights of about .8 on the dollar, .1 on the yen, and .1 on the mark, for the period January 1991 to May 1992. But Indonesia does somewhat better than Malaysia throughout the period 1987-92, even though it is classified by the IMF as a managed floater rather than a basket pegger.25 The Korean won shows up as being linked rather simply to the U.S. dollar in the late 1980s, when it was supposedly following a loose basket strategy. (It shows up as being linked just as closely and simply to the U.S. dollar in 1991-92, after it had announced a switch to a market average rate (MAR) system. This is somewhat surprising, as the Korean Ministry of Finance and U.S. Treasury Department have agreed that the MAR system constitutes a move away from a dollar peg toward a market float.) For present purposes, the point is that if one makes allowances for even a small degree of fluctuation within a band, it may take several years to test reliably the central bank's claim to be following a basket peg. The alternative hypothesis is that random or trend variation is sufficiently great that one cannot distinguish the supposed basket pegger from a flexiblerate currency. If this is true of inference by the econometrician, it is also true of inference by the market observer, who seeks to monitor the central bank and form expectations of future inflation. One could rank precise pegging arrangements according to the time it takes to verify the commitment: a simple dollar or yen peg comes first (because the man in the street can verify it instantly), then a special drawing right (or European currency unit) peg, then a peg to an announced basket, and lastly a peg to an unannounced basket. Needless to say, allowance for a trend, substantial band, or changes in weights or level of the parity, all complicate the process. The conclusion is that, to the extent the motivation for pegging is to make a credible commitment to noninflationary monetary policy, a basket peg with secret weights is not the best choice. Even a peg with announced weights is not as effective as a simple peg to a single currency, if one wishes the average citizen to be able to monitor the commitment on a daily basis. 4
Conclusions
This chapter has considered two fundamental sorts of choices regarding the monetary regime that face any country: exchange rate arrangement and whether monetary policy should be governed by a rule. The two choices obviously intersect, in that a fixed exchange rate is one of the pos25
Indonesia has a significant trend depreciation, as did Malaysia in 1987-88. Frankel and Wei (1994). These tests use weekly data.
Monetary choice for a semi-open country
59
sible rules. However we have examined other rules as well, those phrased in terms of the money supply, price level, and nominal GDP. We have considered how certain specific characteristics of Korea and other East Asian NICs affect these choices. First, these countries have either already liberalized financially (Hong Kong and Singapore) or are in the process (Taiwan and Korea). This makes a floating exchange rate a more viable option. It also makes a flexible exchange rate a necessary option, if the country wishes to retain monetary independence. Second, the countries are trade-oriented, even if Korea is not as open as Hong Kong. Four economic and social aspects of openness can reduce the need to have an independent monetary policy and, therefore, flexible rates: common shocks, common objectives, labor mobility, and fiscal transfers. By none of these criteria is Korea as suitable a unit to give up its monetary independence as, for example, the individual regions of the United States, or even the individual countries of Europe. Perhaps more relevant for the East Asian countries are two aspects of openness that increase the attractiveness of fixed exchange rates: the positive, if relatively small, effect that exchange rate stability has on trade (as partially documented here in Tables 2.1 and 2.2), and the credibility of using the exchange rate as a nominal anchor for monetary policy (as documented by Romer 1991). Third, the East Asians' trade is divided between the United States and Japan. This makes the option of a simple peg to either the dollar or the yen less attractive. A basket peg is an alternative, but we have argued that it is not quite as effective a nominal anchor. Fourth, they are undergoing rapid long-term productivity growth, which leads to trend real appreciation. This means that, if they wish to keep the inflation rate down to the level of their major trading partners, they cannot fix the exchange rate indefinitely. Overall, the aggregate of the arguments seems to recommend a degree of exchange rate flexibility. This leaves the question of the optimal choice for a nominal anchor for monetary policy, if it is not to be a fixed exchange rate. Within the context of the model in the appendix, a nominal GDP target is shown to dominate a money supply target and, under certain conditions, a price level target as well. Appendix: a comparison of discretion and four alternative rules
We compare five possible policy regimes: (1) full discretion by national policy makers, (2) a rigid money supply rule, (3) a rigid nominal GDP rule, (4) a rigid price level rule, and (5) a rigid exchange rate rule. (The analysis thus extends that in Frankel and Chinn [1991] by adding a price
60
Jeffrey A. Frankel
level rule to the list of candidates.) In the case of each of the possible nominal anchors, proponents sometimes have in mind a target zone system; the assumption of a rigid rule makes the analysis simpler.26 The approach, incorporating the advantages both to rules and discretion, follows Rogoff (1985b), Fischer (1991), and Persson and Tabellini (1990). Throughout, we assume an aggregate supply relationship: y = y* + b(p-
Pe) + w,
(1) e
where y represents output, y* potential output, p the price level, p the expected price level (or they could be the actual and expected inflation rates, respectively), and u a supply disturbance, with all variables expressed as logs.27 Al
The closed economy objective function
We begin with the case where the objective function includes output and the price level, but not the exchange rate, import prices, or the trade balance; we call this the case of a closed economy. The loss function is simply: L = ap2 + (y-ky*)\
(2)
where a is the weight assigned to the inflation objective, and we assume that the lagged or expected price level relative to which p is measured can be normalized to zero.28 We impose k > 1, which builds in an expansionary bias to discretionary policy making. L = ap2 + [y*(l -k)
+ b(p - p<) + u\\
(3)
(i) Discretionary policy Under full discretion, the policy maker each period chooses aggregate demand so as to minimize that period's L, with pe given. (1/2) dLldp = ap + [y*(l - k) + b(p - pe) + u]b = 0.
(4)
P = [-/"(I - k)b + b2pe - bu] I [a + b2].
(5)
Under rational expectations, p* = Ep = -y* (I - k)bla. 26
27
28
(6)
Rogoff (1985) warns that the welfare ranking among the candidate variables for rigid targeting need not be the same as the welfare ranking among the candidate variables for partial commitment. It should be noted that, if the parameter b is thought to depend on the variance of the price level, then our results could be vulnerable to the famous Lucas critique. Bean (1983b) and West (1986) use a quadratic objective function that includes only output. But clearly inflation must be added to the objective function if one wants to be able to consider the advantages of precommitting to a nominal target or rule.
Monetary choice for a semi-open country
61
So we can solve (5) for the price level: p = -y*(l
- k) [b/a] - u b/[a + b2].
(7)
From (2), the expected loss function then works out to: EL = (1 + b2/a)[y*(l - k)]2 + [a/(a + b2)] var(u).
(8)
The first term represents the inflationary bias in the system, whereas the second represents the effect of the supply disturbance after the authorities have chosen the optimal split between inflation and output. (ii) Money rule To consider alternative regimes, we must be explicit about the money market equilibrium condition. (In case 1, it was implicit that the money supply m was the variable that the authorities were using to control demand.) m=p
+ y - v,
(9)
where v represents velocity shocks. We assume v uncorrelated with u. If the authorities precommit to a fixed money growth rule in order to reduce expected inflation in long-run equilibrium, then they must give up on affecting y. The optimal money growth rate is the one that sets Ep at the target value for p, namely 0. Thus they will set the money supply m at Ey, which in this case is y*. The aggregate demand equation thus becomes p + y = y* + v.
(10)
Combining with the aggregate supply relationship (1), the equilibrium is given by y = y* + (u + bv)l{\ + b),
p = (v - w)/(l 4- b).
(11)
Substituting into (2), the expected loss function is EL = (1 - k)2y*2 + {(1 + a) var(u) + [a + b2]var(v)}/(\ + b)2.
(12)
The first term is smaller than the corresponding term in the discretion case, because the precommitment eliminates expected inflation; but the second term is probably larger, because the authorities have given up the ability to respond to money demand shocks. Which regime is better, discretion or a money rule, depends on how big the shocks are, and how big a weight (a) is placed on inflation fighting. (Hi) Nominal GNP rule In the case of a nominal GNP rule, the authorities vary the money supply in such a way as to accommodate velocity shocks. (10) is replaced by the condition that p + y is constant. The solution is the same as in case 2, but
62
Jeffrey A. Frankel
with the v disturbance dropped. Thus the expected loss collapses from (12) to: EL = (1 - k)2y*2 + [(1 + a)l{\ + b)2]var(u).
(13)
This unambiguously dominates the money rule case. It is still not possible, without knowing var(u) or a, to say that the rule dominates discretion. It is quite likely, especially if the variance of u is substantial, that an absolute commitment to a rule would be unwisely constraining - hence, the argument for a target zone rather than a single number, and for subjecting the central bank chairman to a mere loss of reputation if he misses the target rather than a firing squad. But it seems clear that, to whatever extent the country chooses to commit to a nominal anchor, nominal GNP dominates the money supply as the candidate for anchor. (iv) Price level rule Under a price level rule, the authorities set monetary policy so that the price level is not just zero in expectation, but is zero regardless of later shocks. From equation (3), we have directly, L = [y*(\ -k)
+ u]2.
(14)
EL = [y*(l - k)}2 + var(w).
A comparison with (12) shows that the price level rule dominates the money supply rule so long as a is large relative to b, or velocity shocks are large. A comparison with (13) shows that the price level rule is dominated by the nominal GNP rule if [(1 + a)l{\ + b)2] < 1, that is, so long as alb<2
+ b.
This condition does not automatically hold. But the condition alb < 1 is not a very difficult one to satisfy, from which the necessary condition easily follows. The reader can decide whether he or she believes the condition a < b by conducting a simple thought experiment. If it were possible, hypothetically, to double output permanently at the cost of doubling inflation (starting from the position where optimal discretion produces an inflationary steady state, presumably the norm for most countries), would the output gain be worth the cost? Most economists do not believe that the short-run costs of inflation ("shoe-leather costs" of trips to the bank, confusion of relative price signals, uncertainty, etc.) are very high. The argument against inflating rather consists of the long-run considerations, with which the adoption of any nominal anchor (money supply, nominal GNP, exchange rate, or price level) is explicitly designed to deal. It follows that a in the objective function is thought to be less than b.
Monetary choice for a semi-open country A2
63
The open economy objective function
We reconsider here a likely objection to choosing nominal GNP as the focus of monetary policy in an open economy, that it neglects the exchange rate. The alternative of setting monetary policy so as to stabilize the exchange rate will not look attractive unless the exchange rate enters the objective function, perhaps indirectly via the consumer price index or the trade balance. Here we confront the argument head-on, and include the exchange rate directly in the loss function along with output and the price level; we call this the case of the open economy objective function. Thus we replace (2) with: L = ap2 + (y - ky*)2 + cs\
(15)
where s is the spot exchange rate measured relative to some equilibrium or target value and c is the weight placed on exchange rate stability per se. We are implicitly assuming that policy makers wish to minimize long-term swings of the exchange rate around its average value, rather than shortterm uncertainty in the exchange rate. There is no point in specifying an elaborate model of the exchange rate under free floating. All the empirical results say that most of the variation in the exchange rate cannot be explained (even ex post; to say nothing of prediction) by measurable macroeconomic variables, and thus can only be attributed to an error term that we here call e. But we must include the money supply in the equation; otherwise we do not allow the authorities the possibility of affecting the exchange rate. Our equation is simply: s = m — y + e.
(16)
We assume that e is uncorrelated with the supply disturbance w. From (9), s = p — v + e.
(17)
We assume that the same aggregate supply relationship holds as before, equation (1). So we can write the loss function (15) as: L = ap2 + [(1 - k)y* + b(p - pe) + u]2 + c(p - v 4- e)2.
(18)
We proceed as before to consider possible regimes. (i)
Discretion
{M2)dLldp = ap + [y*{\ - k) + b(p - pe) + u]b + c(p - v + e) (19) = o. p = [-y*(\ - k)b + b2pe -bu + c(v - e)] I [a + b2 + c].
64
Jeffrey A. Frankel
The rationally expected p is given by pe = Ep: pe = - ( i _ k)by*l(a + c).
(20)
Substituting into (20) yields: p = - (1 - %*[&/(? + c)] + [c(v - e) - bu]/[a 4- b2 4- c].
(21)
The loss function is EL = [(1 - k)j*] 2 (a + b2 4- c)/(a + c) 4- {(a 4- c)var(w) 4- c(a 4- 6 2 )[w(v) + var(e)]}l(a 4- i 2 4- c).
(22)
f/ZJ Money rule As when we considered a money rule before, so that expected inflation is zero the authorities set m at y*, and (10) applies. Thus the same solution (11) for y and p also applies. The exchange rate is given by substituting the solution for/? from (11) into (15): s = e-[(u
+ bv)/(\ + b)].
(23)
The additional s term is the only difference from (12) in the expected loss function: EL = [y*(l - k)]2 + [(1 + a + c)l{\ + b2)]var(u) + [(a + b2 (24) + cb2)/(l + b)2]var(v) + [c]var{e). Again the comparison with discretion depends on the various magnitudes. (in) Nominal GNP rule When the monetary authorities are able to vary m so as to keep p + y constant, the velocity shocks v drop out. The expected loss function becomes EL = [y*(l - k)]2 4- [(1 + a + c)/(l + b)2]var(u) 4- c var(e). (25) As before, the nominal GNP rule unambiguously dominates the money rule. In practice the e shocks in the exchange rate equation are very large, and dwarf the u shocks in the aggregate supply equation, as is documented here. (The exchange rate often moves 10 percent in a year, without corresponding movements in the money supply or other observable macroeconomic variables; try to imagine similar movements of real output.)29 If the weight c on the s target is substantial, then the last term in the expected loss equation may be important. 29
These assertions are documented in Frankel and Chinn (1991).
Monetary choice for a semi-open country
65
(iv) Exchange rate rule Again, the authorities cannot affect y in long-run equilibrium. But now it is the exchange rate that they peg in such a way that Ep = 0, which from (17) is s = 0. The ex-post price level is then given by p = v — e.
(26)
From (1), y = y* + b(v - e) + u.
(27)
From (14), EL = (a + b2)Var(v - e) + [y*(l - k)]2 + Var(w).
(28)
Assume that v and e are uncorrelated, so that Var(v - e) can be replaced with Var(y) + Var(e).30 The coefficient on var(e) is (a + b2), as compared with the coefficient c in the expected loss (25) under the nominal GNP rule. We made the point that e shocks in practice dwarf u shocks. If we reason on this basis, even if v shocks are also small and a = c (the objective function puts no greater weight on a 10 percent fluctuation of the price level than on a 10 percent fluctuation of the exchange rate), which is extremely conservative, the expected loss from fixing s is greater than the expected loss from fixing nominal GNP. The reason is that under an exchange rate rule e shocks are allowed to affect the money supply and therefore the overall price level. Once we allow for v shocks (which are in between u and e shocks in magnitude; Frankel and Chinn, 1991), the case for nominal GNP targeting is even stronger. One would have to put extraordinarily high weight on the exchange rate objective to prefer an exchange rate rule. (v) Price level rule Finally we return to the price level rule. From equation (18), with the price level at zero, L = [(1 - k)y* + u]2 + c(-v + e)2.
(29)
A comparison with (28) shows that the price level rule dominates the exchange rate rule if a + b2 > c. The condition a > c, again, merely says that fluctuations in the exchange rate are not as damaging as equal-percentage fluctuations in the price level. So the price level rule appears easily to dominate the exchange rate rule. The comparison with the nominal GNP rule is more difficult to make, however. If we use (25), the nominal GNP rule dominates if the 30
The case where they are correlated, which is likely to be relevant, is considered in the second half of Frankel and Chinn (1991).
66
Jeffrey A. Frankel
condition a + c < b2 holds. (In addition, velocity shocks add to the relative superiority of the nominal GNP rule.) We have already argued that a is small relative to b, and that c is small relative to both of them. Still, the condition is not assured, and one would not want to assert an answer in the absence of more information. References Anderson, Kym, and Hege Norheim. (1993). "History, geography and regional economic integration." In K. Anderson and R. Blackhurt (eds.), Regionalism and the global trading system. London: Harvester Wheatsheaf. Barro, Robert. (1989). "A cross-country study of growth, saving and government." Cambridge, Mass.: NBER Working Paper no. 2855, February. Barro, Robert, and David Gordon. (1983). "A positive theory of monetary policy in a natural rate model." Journal of Political Economy 91, no. 4 (August): 589-610. Bayoumi, Tamim, and Barry Eichengreen. (1992). "One money or many? On analyzing the prospects for monetary unification in Europe and other parts of the world." International Monetary Fund and the University of California, Berkeley, August. Bean, Charles. (1983a). "Targeting nominal income: An appraisal." Economic Journal 91 (December): 806-19. (1983b). "Little bit more evidence on the natural rate hypothesis." London School of Economics and Political Science, Centre for Labour Economics Discussion Paper no. 149: 1-29. February. Blanchard, Olivier, and Lawrence Katz. (1992). "Regional evolutions." Brookings Papers on Economic Activity 1: 1-75. Blanchard, Olivier Jean, and Pierre Alain Muet. (1993). "Competitiveness through disinflation: An assessment of the French macroeconomic strategy." Economic Policy: A European Forum 8 (April): 12-56. Boughton, James. (1989). "Policy assignment strategies with somewhat flexible exchange rates." In M. Miller, B. Eichengreen, and R. Portes (eds.), Blueprints for exchange rate management. San Diego, Calif: Academic Press, for Centre for Economic and Policy Research, 218-41. Cardenas, Mauricio. (1991). "Coffee exports, endogenous state policy and the business cycle." Ph.D. diss., University of California, Berkeley. Devarajan, Shantayanan, and Dani Rodrik. (1991). "Do the benefits of fixed exchange rates outweigh their costs? The franc zone in Africa." Cambridge, Mass.: NBER Working Paper no. 3727, June. Dornbusch, Rudiger, and Sebastian Edwards (eds.). (1991). The macroeconomics of populism in Latin America. Chicago: University of Chicago Press. Drysdale, Peter, and Ross Garnaut. (1992). "The Pacific: An application of a general theory of economic integration." Twentieth Pacific Trade and Development Conference, Washington, D.C., September 10-12. Eichengreen, Barry. (1990). "One money for Europe? Lessons from the US currency union." Economic Policy: A European Forum 5 (April): 118-87. Fischer, Stanley. (1991). "Rules vs. discretion in monetary policy." In B. Friedman and F. Hahn (eds.), Handbook of monetary economics. North Holland: Amsterdam.
Monetary choice for a semi-open country
67
Frankel, Jeffrey A. (1993a). "Is Japan creating a yen bloc in East Asia and the Pacific?" In J. Frankel and M. Kahler (eds.), Regionalism and rivalry: Japan and the U.S. in Pacific Asia. Chicago: University of Chicago Press. (1993b). "Foreign exchange policy, monetary policy and capital market liberalization in Korea." In C. Taylor (ed.), Korean-U.S. financial issues. Washington, D.C.: Korea Economic Institute of America. Frankel, Jeffrey, and Menzie Chinn. (1991). "The stabilizing properties of a nominal GNP rule in an open economy," University of California, Berkeley Economics Working Paper no. 91-166, May. (Forthcoming, JMCB, 1995.) Frankel, Jeffrey A., and Shang-jin Wei. (1994). "Yen bloc or dollar bloc: Exchange rate policies of the East Asian economies." In T. Lto and A. Krueger (eds.), Macroeconomic linkage. Chicago: University of Chicago Press. Gordon, Robert. (1985). "The conduct of domestic monetary policy." In Albert Ando, et al. (eds.), Monetary policy in our times. Cambridge, Mass.: MIT Press, 45-81. Goto, Junichi, and Koichi Hamada. (1993). "Economic preconditions for the Asian regional integration." Yale University Economic Growth Center Discussion Paper no. 685: 1-34, February. Hall, Robert. (1985). "Monetary policy with an elastic price standard." In Price stability and public policy. Federal Reserve Bank of Kansas City, 35-54. Hamilton, Carl, and L. Alan Winters. (1992). "Opening up international trade in eastern Europe." Economic Policy 14 (April): 77-116. Hooper, Peter, and Steven Kohlhagen. (1978). "The effect of exchange rate uncertainty on the prices and volume of international trade." Journal of International Economics 8 (December): 483-511. Hutchison, Michael, and John Judd. (1992). "Central bank secrecy and money surprises: International evidence." Review of Economics and Statistics 74 (February): 135-45. lto, Takatoshi. (1989). "Is the Bank of Japan a closet monetarist? Monetary targeting in Japan, 1978-88." Cambridge, Mass.: NBER Working Paper no. 2879, March. Kim, Kihwan. (1990). "Deregulating the domestic economy: Korea's experience in the 1980s." Senior Policy Seminar, Caracas. Revised December. Kwack, Sung. (1994). "Monetary control, sterilization, and exchange rate policy: Korea during the period of current account surpluses, 1986-1990," in R. Glick and M. Hutchison (eds.), Exchange rate policy and interdependence: Perspectives from the Pacific Basin. New York: Cambridge University Press. Kydland, F , and E. Prescott. (1977). "Rules rather than discretion: The inconsistency of optimal plans." Journal of Political Economy 85 (June): 473-91. Lowell, Julia. (1992). "Do governments do what they say (and do we believe them)?: Two essays on national debt and exchange regime policies." Ph.D. diss., University of California, Berkeley. Marston, Richard. (1984). "Exchange rate unions as an alternative to flexible rates: The effects of real and monetary disturbances." In J. Bilson and R. Marston (eds.), Exchange rate theory and practice. Chicago: University of Chicago Press, 407-37. McCallum, Bennett. (1987). "The case for rules in the conduct of monetary policy: A concrete example, Federal Reserve Bank of Richmond." Economic Review 73 (September-October): 10-18. (1988). "Robustness properties of a rule for monetary policy." IMF Seminar
68
Jeffrey A. Frankel
Series no. 1988-4: 1-35, revised February. Presented at IMF seminar, "Rules for implementing monetary policy," May 5, 1988. McKinnon, Ronald. (1963). "Optimum currency areas." American Economic Review 53 (September): 717-24. Meade, James. (1984). "A new Keynesian Bretton Woods." Three Banks Review 142 (June): 8-25. Miller, Marcus H., and John Williamson. (1988). "International monetary system: an analysis of alternative regimes." CEPR Discussion Paper no. 266 (July): 1-29. With comments by Stanley Fischer and Gilles Oudiz, pp. 1048-54. Also issued in European Economic Review 32 (June): 1031-48. Mundell, Robert. (1961). "The theory of optimal currency areas." American Economic Review 51 (November): 509-17. Park, Yung Chul, and Won-Am Park. (1991a). "Changing Japanese trade patterns and the East Asian NICs." In Paul Krugman (ed.), Trade with Japan: Has the door opened wider?. Chicago: University of Chicago Press, for NBER, 57-89. (1991b). "Exchange rate policies for the East Asian newly industrialized countries." In Emil-Maria Claasen (ed.), Exchange rate policies in developing and post-socialist countries. San Francisco: ICS Press. 122-41. Persson, Torsten, and Guido Tabellini. (1990). Macroeconomic policy, credibility and politics. Chur, Switzerland: Harwood Academic Publishers. Petri, Peter. (1992). "The East Asian trading bloc: An analytical history." NBER conference, Del Mar, California, April 3-5, 1992. In Jeffrey Frankel and Miles Kahler (eds.), Regionalism and rivalry: Japan and the U.S. in Pacific Asia. Chicago: University of Chicago Press, 1993, 310-35. Reisen, Helmut. (1991). "Exchange rate policies for the East Asian newly industrialized countries: Comment." In Emil-Maria Claasen (ed.), Exchange rate policies in developing and post-socialist countries. San Francisco: ICS Press. 93-108. (1993). "Southeast Asia and the 'impossible trinity.'" International Economic Insights 4 (May-June): 21-23. Reisen, Helmut, and Axel van Trotsenburg. (1988). "Should the Asian NICs peg to the yen?" Intereconomics 23 (July/August): 172-77. Rogoff, Kenneth. (1985). "The optimal degree of commitment to an intermediate monetary target." Quarterly Journal of Economics 100 (November): 1169-89. (1986). "Reputational constraints on monetary policy." Cambridge, Mass.: NBER Working Paper no. 1986: 1-61, July. Romer, David. (1991). "Openness and inflation: Theory and evidence." Cambridge, Mass.: NBER Working Paper no. 3936, December. Sachs, Jeffrey, and Xavier D. Sala-i-Martin. (1989). "Federal fiscal policy and optimum currency areas." Unpublished manuscript. Takagi, Shinji. (1988). "A basket policy: Operational issues for developing countries." World Development 16, no. 2: 271-79. Taylor, John. (1985). "What would nominal GNP targeting do to the business cycle?" Carnegie-Rochester Conference Series on Public Policy 22: 61-84. Tobin, James. (1980). "Stabilization policy ten years after." Brookings Papers on Economic Activity 1: 19-72. Wang, Zhen Kun, and L. Alan Winters. (1991). "The Trading potential of Eastern Europe." London: Centre for Economic Policy Research Discussion Paper no. 610, November.
Monetary choice for a semi-open country
69
West, Kenneth D. (1986). "Targeting nominal income: A note." Economic Journal 96 (Dec): 1077-1083. Williamson, John, and Marcus Miller. (1987). "Targets and indicators: A blueprint for the international coordination of economic policy." Policy analyses in international economics, no. 22, September. Washington, D.C.: Institute for International Economics.
CHAPTER 3
Capital account liberalization: bringing policy in line with reality Manuel
1
Guitidn
Introduction
An important feature of the international financial code of conduct established in the Bretton Woods Conference toward the end of World War II was the acceptability of controls and restrictions on international capital flows. In particular, the Articles of Agreement of the International Monetary Fund, which contain such code of conduct, prescribe that: Members may exercise such controls as are necessary to regulate international capital movements, but no member may exercise these controls in a manner which restricts payments for current transactions or which will unduly delay transfers of funds in settlements of commitments, except as provided in Article VIII, Section 3(b) and in Article XIV, Section 2. (Article VI, Section 3). There was logic at the time of the drafting and acceptance of the Articles of Agreement in including this feature into the rules of the game that were to govern international transactions. The flow of trade, current and capital account exchanges among countries, had been disrupted prior to and during the hostilities; capital movements had not become yet a preThe views expressed in the chapter are mine and they should not be attributed to the International Monetary Fund. The general issue of international capital flows and the particular subject of capital account liberalization have been extensively studied in the International Monetary Fund. In the preparation of this essay, I have relied heavily on the most recent studies in the institution; these include International Monetary Fund, Determinants and Systemic Consequences of International Capital Flows: A Study by the Research Department of the International Monetary Fund, IMF Occasional Paper no. 77 (Washington, D.C.: International Monetary Fund, 1991); Donald J. Mathieson and Liliana Rojas-Suarez, "Liberalization of the Capital Account: Experiences and Issues," IMF Working Paper, no. 46 (Washington, D.C.: International Monetary Fund, June 1992); and Guillermo A. Calvo, Leonardo Leiderman, and Carmen M. Reinhart: "Capital Inflows and Real Exchange Rate Appreciation in Latin America: The Role of External Factors," IMF Working Paper no. 62 (Washington, D.C.: International Monetary Fund, August 1992). 71
72
Manuel Guitian
dominant force in the international transaction network; and for purposes of reconstruction and the resumption of normality in the world economy, encouragement of international trade and exchange became essential. Consequently, attention concentrated on the liberalization of trade and other current account flows, an aim that became enshrined in the Articles of Agreement of the International Monetary Fund. 1 But it is also worth mentioning that an additional consideration for capital controls was their ability to preserve the independence of domestic policies. Limited though such ability is likely to be, it warrants bearing in mind that capital controls, by allowing countries to pursue inconsistent policies (at least for some time), may have been an obstacle to the aims of international policy coordination. Since the Bretton Woods era, however, international capital movements have acquired increasing importance in the world economy. Undoubtedly, the growing relevance of capital flows in the international economy during the past two decades, although it can be traced to numerous other factors, owes much to the opening of trade and current accounts that characterized the evolution of the international economy during the period of prevalence of the Bretton Woods order.2 This span of time, which extended through the early 1970s, witnessed a progressive liberalization of current international transactions in the industrial countries as well as the corresponding integration of their national economies and those of an increasing number of developing countries into the international system. Establishment of "current account convertibility" had been broadly attained in the industrial world by the late 1950s or early 1960s.3 With it, the importance and pervasiveness of international trade and other current account flows began to characterize the world economic setting, which became increasingly interdependent as a result. In what might well be termed a revolt against economic interdependence, the Bretton Woods par value system collapsed in the early 1970s, 1
2
3
See Article 8 of the Articles of Agreement in International Monetary Fund (1978). It must be pointed out, in this context, that in its definition of current account transactions, the Articles of Agreement include some capital account items (see Article 30 of the Articles of Agreement). International capital flows are also a subject of interest for the Organization of Economic Cooperation and Development (OECD), as made evident by the adoption of the Code of Liberalization of Current Invisible Transactions and the Code of Liberalization of Capital soon after the organization's establishment: see Fischer and Raisen (1992). The Bretton Woods period encompasses the time during which the international economy operated under a fixed exchange rate regime, the par value system adopted at the Bretton Woods Conference that established the International Monetary Fund and the World Bank. See Horsefield (1969), de Vries (1976), and Guitian (1992a). For an exposition of issues related to currency convertibility, see Polak (1991), Williamson (1991), and Guitian (1992b) as well as Greene and Isard (1991).
Capital account liberalization
73
when countries decided to move away from such a system and adopt instead a regime of flexible exchange rate arrangements. The shift in focus in world financial relations from an international standpoint (which is a central feature of a par value or fixed exchange rate setting) to a national perspective (which is typical of a flexible exchange rate framework) augured the appearance of a tendency toward the relative closing of national economies, that is, toward an international environment composed of isolated and insulated national components. Paradoxically, however, the tendency was to an important extent weakened by the emergence of growing international capital flows, which contributed to keep the fabric of the international economic system closely woven. The reality of developments in the world economy began to surpass the prescriptions of the code of conduct, as capital flows became a predominant aspect of international economic relations, even though the use of capital controls remained acceptable.4 The importance of the capital account continued growing into the 1980s, a period when the subject of liberalization of capital movements began to attract attention, both in the developing and the industrial worlds.5 The purpose of this essay is to examine the reasons why the code of conduct for international economic policy should catch up with the reality of world economic affairs. As such, the standpoint of the analysis will basically be normative, that is, it will deal with what ought to be, rather than positive, a vantage point that would stress instead what it is. The plan of the chapter is as follows. It will first examine the evolution of international capital flows and the consequent divergence between the reality of the international economy and the code of conduct that is supposed to guide it (section 2). Subsequently, the analysis will focus on the logic of capital account liberalization, both from the perspective of its costs and benefits as well as from the viewpoint of its inevitability (section 3). The chapter will then examine the interaction of free capital movements with national macroeconomic policies, in general, and with monetary and exchange rate policies, in particular (section 4). This examination will be followed by a discussion of the systemic implications of the liberalization of international capital flows; in this context, the issues that arise whenever capital account freedom is accompanied by a regime of fixed exchange rates will also be discussed (section 5). The main points, issues for discussion, and conclusions of the analysis, will be gathered in a concluding 4
5
This period has been examined in extenso by de Vries (1985). The points made in the text have been discussed in more detail in Guitian (1992c), where the issues related to the international debt crisis, its origins and resolution are also studied. See, for example, Lenain (1992), Mathieson and Rojas-Suarez (1992), and Guitian (1992d), as well as International Monetary Fund (1991).
74
Manuel Guitian
section (section 6). At the time of writing (mid-September 1992), turmoil developed in European and world currency markets, a turn of events that is of interest in the context of this essay. Therefore, a postscriptum examining them has also been included (section 7). 2
Economic reality and the code of conduct
An essential development in the international economy in the period since the abandonment of the Bretton Woods order has been the expansion in the scale of gross and net capital flows and the resulting integration of national financial markets, particularly those in the industrial countries. A number of features have accompanied this expansion. Not only has it exceeded the growth of international trade flows over the same period but it has led to and reflected the presence of foreign investors in the domestic capital markets of the main industrial countries, a presence that is there to stay, barring major upheavals in the world economy. Then, an important proportion of the scale of expansion of capital flows has been private, rather than official in nature, with bank lending and security flows as predominant modalities. As for the developing countries, heavy borrowing from international banking sources in the late 1970s was followed, as is well known, by the virtual halt in the banking and other related flows in the early 1980s. The process changed the composition of capital movements from private to official as well as their nature from voluntary to involuntary in the period of resolution of the international debt crisis.6 Possibly the most critical feature of the recent evolution of capital movements has been the relaxation of capital controls, the bulk of which took place in the context of a broad liberalization and deregulation of domestic financial markets in industrial countries. These events could not but lead to a growing integration of national and international financial markets, which has been made evident not only by the scale of capital flows but also by developments in yield differentials. Thus, capital movements not only have expanded markedly in magnitude but they have also led to arbitrage transactions that have significantly narrowed the margin for national yield differentials.7 In effect, financial market integration has outstripped that of goods markets, which are often hampered by recurrent protectionist pressures. This is, in itself, somewhat paradoxical when 6
7
On the subject of the international debt crisis, the literature is abundant. A recent source can be found in "The International Debt Crisis: What Have we Learned?" in Guitian (1992c); see also Guitian (1992e). As for the general evolution of international capital movements over the past two decades, see International Monetary Fund (1991). For further discussion of the evaluation of yield differentials, see International Monetary Fund (1991) and Frankel (1991).
Capital account liberalization
75
viewed from the standpoint of the international code of conduct. The bulk of international efforts toward a liberal system has been channeled to the flow of trade in goods and services and to freedom of current international transactions. Not only have the Articles of Agreement of the International Monetary Fund stressed the importance of current account convertibility, but the General Agreement on Tariffs and Trade (GATT) was established to uphold and foster a code of conduct in the specific field of international trade. In contrast, although continuing efforts have been made by the Organization for Economic Cooperation and Development (OECD) toward the acceptance of a code of capital transaction liberalization, the rules of the game espoused by the membership of the International Monetary Fund did not include or extend to capital account convertibility, that is, to freedom of capital transactions. And yet, the evolution of the international economy has been such that capital market interdependence seems to have run well ahead of trade flows integration. A number of factors account for the relatively rapid and deep integration of international capital markets. They include elements like availability of cross-border global investment opportunities, which, to be seized, required intermediation of resources across nations. These trends also allowed for cross-border portfolio and risk diversification and have contributed to the enhancement of the efficiency of investment. In essence, the factors just enumerated reflect the importance of economic fundamentals in the determination of capital flows. But capital movements are also affected by national policies and by the differences therein as well as by the wedges created by the prevalence of economic distortions that vary across countries. Economic policies are pervasive in their effects and therefore, capital flows reflect them directly - as is the case, for example, with tax policies or official guarantees - or indirectly, as illustrated by inappropriate macroeconomic policies. Distortions, in turn, which in the end amount to the existence of structures of differential costs across countries, also influence the direction and scale of capital flows. These flows will tend to go toward the destination where the costs are the lowest, that is, typically, toward havens where the distortions are the least important. The upshot of recent trends in capital accounts in the world economy has been that after the great leap given during Bretton Woods toward integration of national economies through free trade in goods and services, a quantum jump has taken place on capital account, an event well ahead of the expectations embedded in the existing code of international financial conduct. There can be no doubt that the central challenge that confronted the membership of the International Monetary Fund at the time of its inception was to ensure the integration of war-ridden and restricted economies
76
Manuel Guitian
into an orderly international economic system. To this end, emphasis was placed on the attainment of freedom of current international transactions, which, as already noted, became the standard for convertibility. This approach reflected the concern with the notion of competitiveness of an economy and with the retention of national economic policy autonomy (hence, the acceptability of capital controls). To a large extent, the effectiveness of the approach that was then taken to help the development of a relatively integrated world trading system both explains and has provided the grounds for the emergence of capital flows on an increasingly important scale. These flows have materialized in the presence - and despite the acceptability - of capital controls, a curious instance of events overtaking rules intended to govern them. 3
Logic of capital account liberalization
Interferences with international capital transactions have taken a variety of modalities, including exchange restrictions and controls as well as quantitative limitations on the scale of capital movements. Exchange practices typically have taken the form of dual or multiple exchange rates or of differential tax treatment of international financial transactions. Quantitative restrictions, in turn, have constrained the foreign asset-liability positions of banks and other domestic financial institutions or the domestic operations of foreign financial entities or the external portfolios, direct investments or real estate assets of domestic residents, or the direct investments undertaken by foreigners. Just as varied as the types of capital controls have been the grounds on which they have been normally justified; these can be classified into four main categories: containment of balance-of-payments instability or undue exchange rate volatility; retention of domestic savings and prevention of excessive foreign ownership of domestic factors of production; taxation of domestic capital and financial transactions; and reinforcement of domestic stabilization and reform efforts. Much though these lines of reasoning may appear defensible in practice, in theory their rationale is far more debatable. Balance-of-payments crises or unstable exchange rates will hardly be averted on a sustained basis only by the adoption and maintenance of capital controls. Fundamentally, external imbalances have roots in sources other than capital movements and, therefore, the imposition of capital restrictions cannot be expected to replace appropriate action to remove the basic source or sources of the imbalance. Thus, at best, capital controls can only act as palliatives while other essential policy measures yield their results. Similarly, capital restrictions are unlikely to retain domestic savings in the ab-
Capital account liberalization
77
sence of appropriate incentives to keep those savings from spontaneously flowing abroad; here again, it must be stressed that in general, capital flight is caused by factors other than the existence of freedom of capital movements and it is toward those factors - rather than to the confinement of that freedom - that policy ought to be addressed. Taxation of capital transactions, if excessive, is unlikely to be helped much by capital controls. Clearly whenever such taxation exceeds significantly its levels abroad, it will be imperative that a wedge be created in order to prevent the shrinkage of the tax base. The introduction of capital controls can be instrumental to this end, but it must also be acknowledged that their continued effectiveness cannot be assured, as long as the incentives to circumvent them - that is, the taxation differential that made them necessary to begin with - continue to prevail. And with regard to the ability of capital controls to assist in the stabilization and reform efforts undertaken domestically, an argument in this direction can be plausibly made on a temporary basis; but it must also be kept in mind in gauging the usefulness of such controls for these purposes that their very presence may well impair the credibility of those efforts themselves. Arguments in support of capital controls have also often been made on the basis of the costs of capital account openness. Prominent among those arguments is the belief that capital flows, if left unfettered, will increase economic instability in that they contribute to magnify, rather than to offset, external shocks. But this instability argument does not depend only on whether capital movements act to reinforce or to compensate the shock; it also hinges on whether the shock is internal or external, permanent or transitory, and once these other nuances are included, we are confronted again with the real consideration that capital controls cannot replace action aimed directly at the removal of the underlying causes of the instability. Another type of perceived cost of an open capital account is the limitation it can impose on the effectiveness of domestic policies. True though this concern is, it applies to economic openness in general, that is, it encompasses the openness of the trade and the current accounts as well. Therefore, this would be an argument against opening the economy (which would carry hardly any support as a general economic policy proposition) and not against capital account liberalization as such. Yet another cost of open capital account is the fear of losing domestic savings or of a shrinking tax base. As argued previously, these are contestable arguments.8 Traditional benefits of capital account liberalization, in turn, include the welfare increase that is made possible by the availability of foreign 8
An extensive discussion of the costs and benefits of liberalization of the capital account will be found in Hanson (1992).
78
Manuel Guitian
savings to supplement the domestic resource base, which makes it feasible to have a larger capital stock and place the economy in a potentially higher growth path than otherwise. From another perspective, free trade in capital (e.g., through international borrowing and lending flows) can help reduce the costs of the intertemporal misalignments that periodically arise between the patterns of production and consumption. Another benefit to be derived from freedom in capital transactions is that by allowing for trade in risk assets, the sharing and diversification of risks that otherwise would not be available become feasible. In essence, the analysis of the costs and benefits of capital account liberalization does not differ from the traditional tenets derived from the conventional examination of the advantages and disadvantages of free trade, in general. Both of them have to do with the pros and cons of closed versus open systems. Closed systems, other things equal, will neither suffer from (import) external shocks nor will they disseminate (export) the consequences of internal shocks. Open systems, in contrast, will be influenced by (import) external shocks and they will disseminate (export) the effects of their domestic shocks. Capital controls, which help close otherwise open economies, act to insulate them within the system, for better or for worse. As is the case with all types of controls, the effectiveness of those used to contain capital flows will erode over time. Yet, for a time at least, they can inhibit certain types of external financial transactions, limit access to international markets, restrict domestic financial market competition, and, most likely, discourage the repatriation of flight capital. In this manner, capital controls can introduce distortions and give rise to inefficiencies in the domestic financial system that will impair the competitiveness of the economy at large. Thus, although they are often advocated as a protective shield, capital controls can instead enhance the economy's vulnerability to financial and other economic shocks. The presence of capital controls can add further distortions if they encourage the duration of prevailing sources of imbalance and of inappropriate policies in an economy. In these circumstances, only controls of increasing restrictiveness will be able to contain the erosion of their effectiveness to allow for the maintenance of the causes of imbalance and for the delay in appropriate policy action. Apart from the costs that also arise from likely attempts to capture the "rents" to which capital controls give rise, there is also the risk of moral hazard that the efforts to evade capital controls can create for other areas of economic activity; thus, the use of capital controls to tax capital transactions not only can encourage flight of resources abroad but also can provide incentives to undertake other
Capital account liberalization
79
domestic unofficial activities, a moral hazard risk that should not be underestimated. 4
Capital flows, monetary and exchange policies
The possibility, already alluded to, of avoiding taxes when capital controls are lifted points out clearly the immediate impact of capital account openness on the effectiveness of government policies. Again, this is no more and no less than a specific manifestation of the general principle that it is easier to influence a closed system by policy or by other internal means than it is to affect an open environment. Therefore, clearly the ability to tax arbitrarily is very much limited by the abandonment or dismantling of capital controls. This may be seen (and it has often been described) as a reduction in the effectiveness of fiscal policy brought about by the liberalization of capital flows. But there is no such reduction in fiscal policy efficiency, a more important consideration, because efficiency is bound to increase with the elimination of the distortions caused by capital controls. Moreover, it neither needs to hold with regard to fiscal policy effectiveness from two different standpoints: one, that the very tendency (often successful) to circumvent the controls will make the ability to tax in the presence of capital restrictions more apparent than real; another, that the immediate effectiveness of controls (actually, it should be the effectiveness of the introduction and progressive tightening of such controls) should be distinguished from their longer-term effectiveness, which undoubtedly will shrink with the erosion of the ability of controls to insulate the economy. The relationships that are most often discussed are those between capital movements and monetary as well as exchange policies. A well-known argument that links the three together is often brought up in discussions of the advantages and disadvantages of different exchange rate regimes. In this context, the point has been frequently made that, with a freely floating exchange rate, the independence and effectiveness of domestic monetary policy can be maintained even in the presence of an open capital account. Correspondingly, the flip side of this argument is also often brought up: That is, with a fixed exchange rate, the independence and effectiveness of domestic monetary policy can only be preserved by the establishment of capital controls. Actually, this line of reasoning has been invoked many a time as an explanation for the acceptance of capital controls in the Bretton Woods code of conduct. Both arguments, however, are spurious, except in the short run. Under flexible exchange rates, the independence and effectiveness of domestic monetary policy with an open capital account can only be maintained if
80
Manuel Guitian
the monetary policy stance is broadly compatible with that prevailing abroad. Monetary policy, like other domestic policies in general, competes in the world economy so that there will be a tendency for resources to flow toward the areas where the best-quality monetary management is in effect. Actually, the argument is that this tendency is then contained by the freely floating exchange rate since this rate will move (depreciate) to the extent necessary to eliminate the incentive for resources to flow abroad. In essence, then, the argument is equivalent to the case - also often made that with a flexible exchange rate, a national economy can choose its own inflation rate. But this can only be rarely true, as basically resources will tend to flow toward stable environments rather than stay in settings where the efforts to maintain stability are characterized by attempts to try to offset one instability (that in the price level) with another (that in the exchange rate). Similarly, in the case of fixed exchange rates, it is also misleading to argue that through capital controls the independence and effectiveness of monetary policy can be preserved. As was the case with flexible exchange rates, monetary policies compete internationally; the tendency will develop for resources to move toward the domain of the best-quality policies and the situation here is one of containing that tendency by means of quantity restrictions (capital controls), rather than by means of price adjustments (flexible exchange rates). In effect, an analogous argument can be made that the attainment of monetary autonomy by capital controls is equivalent to saying that by the introduction of capital controls, a country can choose its own inflation rate. Other than in the short run, the argument is indeed spurious; controls can only attempt to contain the tendency for resources to flow out, but they will not eliminate it. In the process, their effectiveness will be eroded and, with it, the independence of monetary policy and the possibility of choosing a national inflation rate that differs significantly from the one prevailing abroad. This sequence of events can be slowed down, though not definitely stalled, by progressively tighter controls, of course. In general, however, the force of behavior will be for resources to move toward stable environments and to escape from those where the consequences of price instability are temporarily contained by capital controls. The basic argument to be made is that domestic monetary policies are endogenous to domestic as well as foreign economic market forces. Attempts to escape from the influences of these forces, effective though they may initially be, can only be doomed to failure. All the well-known reasons against the use of quantitative restrictions generally accepted in the context of the market for goods and services and of the consequent inter-
Capital account liberalization
81
national trade flows apply just as much in the context of the capital market and of the similarly consequent international capital movements. With the globalization of international capital markets that has resulted from the dismantling of national capital controls, a phenomenon that has been observed in many economies is that capital flowing from abroad can threaten domestic inflation objectives (another way of expressing the loss of monetary independence), or that it can impair the economy's competitiveness (thus showing the link between capital flows and exchange rate developments). The dilemma that such capital inflows pose to the receiving economy is that action taken to meet inflation objectives (in particular, letting the exchange rate appreciate rather than allowing for the domestic monetary expansion that would result from the inflows of capital in the absence of exchange rate adjustments) will endanger competitiveness; and, on the contrary, protecting competitiveness (by not allowing the exchange rate to appreciate and, thus, accepting the monetary expansion induced by capital inflows) may jeopardize inflation control. In most circumstances, both a measure of inflation control and of competitiveness are typically lost. Traditional means to resolve the conflict have consisted in attempts to sterilize the capital inflows - at the expense of expansions in domestic public debt, which tend to keep interest rates high, thus raising fiscal costs and perpetuating inflows in search of attractive yields; adoption of dual exchange rates, which distort the economy and are often circumvented; effort* to tighten fiscal policy, which is a valid policy option, if the scale of the fiscal adjustment and of capital inflows are commensurate, though not otherwise; and generally the adoption of sound policies, which clearly is an appropriate course of action. Over the longer haul, besides the maintenance of sound domestic policies, the optimal reaction to capital inflows must encompass the enhancement of the flexibility of factor and product markets so that the tendency toward exchange rate appreciation can be compensated for by domestic cost and price decreases, thus keeping competitiveness safe and inflation at bay. In sum, freedom of capital flows does not eliminate macroeconomic policy independence. Openness of the economy does that. And the pursuit of a flexible exchange rate policy does not restore policy independence. Although it is true that exchange rate flexibility helps to insulate and close the economy, it will not be sufficient to bring about complete insulation, particularly in the presence of unsustainable domestic policies that will provide incentives for resources to flow to more stable settings. Thus, the perceptions of domestic policy constraints due to capital account opening can only be real in the short term, as a transitory stage.
82
Manuel Guitian
And such perceptions do not pertain exclusively to capital controls. They are applicable to all types of controls, be they on flow (trade) or stock (capital) variables or of domestic or foreign origin. Much though capital controls can be justified as instruments to deal with specific circumstances or defended on country-case grounds, sight should not be lost of the fundamental fact that controls are inefficient as instruments to guide economic behavior or as a basis for policy effectiveness. 5
Systemic considerations
Many of the observations made in the chapter so far can be placed in a systemic setting in order to provide the rationale for the obvious need of bringing the international code of conduct in line with the reality of the world economic environment. The discussion in this systemic context will be conducted from the standpoint of the International Monetary Fund, in its capacity as guardian of the code of conduct. Much progress has been made in the past half century toward the acceptance of the advantages of trade and current account openness, an endeavor that represented perhaps the central challenge that the world economy confronted in the wake of World War II. Supervision of the acceptance and discharge of the responsibilities of such openness in terms of the observance of constraints on national policies (e.g., by the proscription of use of exchange restrictions or of discriminatory practices) was a main function of the International Monetary Fund. To a large extent, such progress laid natural grounds for the emergence of capital flows on an increasing scale, which came only to reinforce the commitment to an open and globally integrated system. The past two decades witnessed the developments in capital accounts rise in prominence and eventually surpass in importance those on current accounts.9 And yet, despite these broadranging developments and the onset and resolution of an international debt crisis, the international code of conduct has remained unchanged in the domain of the norms governing capital flows. These flows have continued to be treated as they had been in the Bretton Woods period, when their scale was far more limited and when the concern of the international economy focused mainly on the resumption of trade and other current account flows. But at present, the presence and predominance of capital movements have become permanent features of the world economic setting; in effect, it can well be said that they constitute part of its structure. And those 9
A very detailed examination of these capital account developments will be found in Morris Goldstein, Donald J. Mathieson, and Timothy Lane: "Determinants and Systemic Consequences of International Capital Flows," in International Monetary Fund (1991).
Capital account liberalization
83
movements took off in the period of flexible exchange rate arrangements that succeeded the Bretton Woods order. It could have been argued that such flexibility rendered redundant the need for capital (or any other external) controls, just as it was often argued that freely floating exchange rates reduced the need for international reserves. Neither argument conformed to the reality that accompanied exchange rate flexibility: Both international reserves continued to be needed and capital controls remained in place in many countries, even though in a growing number of cases those controls became de facto or de jure less predominant. In the process, a state of world economic affairs that linked exchange rates and exchange rate management closely with trade and current account flows moved to a setting where exchange rates and exchange rate management - because of the influence of capital flows and the capital account - became even more closely related to monetary policies and to monetary conditions in country economies. An important consequence of growing capital movements has been that they helped make clear that the independence of monetary policy, in particular, and of domestic economic management, in general, is indeed more apparent than real. A corollary of this consequence is the need for consistency of policies across countries as a condition for orderly economic developments. These considerations in effect stress the increasing futility of attempts at pursuing national objectives at the expense of international aims. Nowhere is this consideration better illustrated than in the context of developments in the European Monetary System (EMS) which, as I have noted elsewhere, represents actually a quantum jump over the Bretton Woods order in that it is based on fixed exchange rates with freedom of capital movements.10 The intimate linkages among domestic financial policies of EMS members have become a well-known reality often written about; this is particularly the case with regard to the benefits and costs of anchoring one's policies on those of the less inflationary members of the system. This should not be surprising in a regime that ensures the openness of its components both in terms of prices (fixed exchange rates) and Many observers attributed the success of the EMS to the existence of capital controls in many of the participating economies, the argument being that the independence given up by exchange rate fixity was regained by capital restrictions. These arguments carry far less weight at present, when capital flows have been liberalized within the EMS. For further discussion, see Guitian (1988 and 1992c). At the time of writing (mid-September 1992), turmoil characterized European and world money markets and the EMS was confronting perhaps the most serious challenge of its existence. I do not believe that those turbulent events counter the fundamental logic of this essay, but they do call for a commentary that is provided in section 7.
84
Manuel Guitian
of quantities (free capital flows). In these circumstances, any domestic policy imbalance in the anchor economy (say, an improper policy mix) has repercussions on the other economies of the system. And the latter cannot sever this linkage unless the market on its own gives the policies in other countries in the system a measure of credibility that exceeds that typically accorded to the anchor country. Actually, to operate optimally, systems like the EMS should be able to shift the role of anchor between the countries so that the group at large can benefit at all times from the advantage of being linked to the country with the set of policies of best quality. That is, the EMS should import the best policies and the resulting best economic performance from where they prevail and export them to where they are most needed. Looking at the international economy as a whole, there are grounds to argue for the introduction of capital account convertibility in the code of conduct as an aim of equivalent importance to current account convertibility. To this end, capital controls and other restrictions on international capital transactions should receive a treatment similar to that given to current account restrictions. The separate treatment given so far to these two types of restrictions carries no weight in logic. But at present, it neither carries weight in practice, as, in effect, capital movements flow across nations on an increasing scale and with growing freedom. As a result, where they remain, capital controls have become far less effective than they would be in a setting characterized by a network of widespread interferences with movements of assets across national borders. As was the case with current account liberalization during Bretton Woods, the establishment of freedom for capital transactions represents a key challenge for the membership of the International Monetary Fund. But the challenge has already been substantially met by world economic developments, for practical purposes.
6
Concluding remarks
There is a consensus that capital account openness, other things equal, is preferable to the prevalence of capital controls. The case for those controls is typically based on second-best reasoning and it has often been accompanied by discussions of the conditions that an economy should satisfy before considering external financial liberalization. These conditions typically include the establishment of a stable macroeconomic setting at a realistic rate of exchange. An important element of the setting will also of course be a sustainable fiscal position. It will be desirable in addition to have a liberal domestic financial system and one already endowed with a
Capital account liberalization
85
sound safety net to withstand the inroads likely to be caused by foreign competition. But like in many other areas of economics, waiting for conditions such as those just listed to materialize before proceeding to liberalize capital movements may prove the best recipe for the permanence of capital controls. Just as plausible would be the opposite argument, one that would contend that the capital account should be opened with no prior conditions for the following reasons. An open capital account will constrain domestic policies to the extent necessary to bring about balance and stability to the economy. On this reasoning, a stable macroeconomy, rather than a precondition for liberalization of capital transactions, can also be seen as a result of capital account convertibility. Similar arguments can be made with regard to the exchange rate - open capital flows will bring about a realistic exchange rate rather than require it ex ante; and the same reasoning applies to the need for domestic financial liberalization, which can also be seen as an outcome of, rather than a prerequisite for, external financial opening. Clearly, in the extreme, neither of these positions is likely to be applicable to any concrete country economy. Neither should the opening of the capital account wait until all desirable conditions are in place, because such a point in time is unlikely to be reached. Nor should it proceed in the absence of a minimum of favorable domestic economic conditions. Balance between these two extremes will be necessary. Such balance means that capital account liberalization should be undertaken in less than optimal domestic economic conditions, but not under circumstances so far away from optimality that the credibility of the decision to open the economy to international financial transactions is so impaired that it cannot be sustained. Another pragmatic consideration that is often raised in discussions of capital account liberalization concerns the sequence of balance-ofpayments opening; that is, discussions have often focused on whether the current account opening should precede or follow the liberalization of the capital account.11 Many arguments favor the opening of the current account first for a variety of reasons including, in particular, the varying speed of adjustment in the market for goods and services and in that for capital; the former is perceived as slower and, therefore, in need of lead time in the opening to the external environment. But there does not seem to be an a priori reason why the two accounts could not be opened up An excellent discussion of the various issues involved in the decision to liberalize an economy and of the sequence in which it should be undertaken will be found in McKinnon (1991).
86
Manuel Guitian
simultaneously, if only because the very presence of capital controls can be an obstacle to current account liberalization. Although there is no categorical answer to the issue of sequencing - which, in any event, is influenced by the initial conditions of individual economies - it is plausible to argue that simultaneous rather than sequential liberalization, if accompanied by credible, sound domestic economic policies in a stable external environment, would reinforce one another. A related issue to the sequence of liberalization is that of its speed, on which arguments can be grouped as favoring either a gradual or a rapid approach. Here again, no single, categorical answers can be given. But the advantages and disadvantages of each approach warrant gauging. Gradual opening softens the inroads of external competition and provides leeway for domestic preparation to confront that competition. But precisely by giving leeway to adjust, there is no guarantee that the leeway will be, in effect, used to confront external competition as opposed to the tendency to continue exploiting the opportunities of a closed or partially closed economy. If gradual opening encourages delays in adjustment, its costs will not fall below those from a fast liberalization and, yet, it will not benefit from the latter's advantages. These advantages are the transparent signals that rapid opening of the economy convey to economic agents and the consequent total absence of leeway for delays in behavioral adjustments to external competition. It is generally agreed that efficiency criteria argue for completely free exchange systems, with appropriate prudential safeguards, of course. But there are also pragmatic considerations that advocate the establishment of full currency convertibility. These considerations include the desirability of strengthening the role of market forces as opposed to administrative controls, the need to encourage international resource flows, and the need to supplement domestic financial market liberalization and deregulation. In conclusion, economic logic advocates the dismantling of capital controls; developments in the world economy make them undesirable and ineffective; and a strong case can be made in support of rapid and decisive liberalization of capital transactions. All these considerations underwrite strongly a code of conduct that eschews resort to capital controls as an acceptable course of action for economic policy. 7
A post-scriptum on currency market turmoil
At the time when this essay was virtually completed (mid-September 1992), turbulence settled in European and world currency markets and the European Monetary System (EMS) found itself confronting possibly
Capital account liberalization
87
the most serious challenge of its existence. In a nutshell, short-term capital and monetary flows moved against the currencies perceived as vulnerable in the system and the flows were of a magnitude sufficient to prompt the withdrawal of Italy and the United Kingdom from the exchange rate mechanism (ERM) of the EMS. They also brought about a downward exchange rate realignment of Spain's currency and, most important in the context of this chapter, the reintroduction on a temporary basis of capital control measures in Ireland, Portugal, and Spain. Dramatic though these events seemed at the time, their occurrence cannot be considered totally unexpected or unfounded. The EMS over the past five years or so had moved more and more toward a regime of fixed, nonadjustable, exchange rates. Yet, the absence of recourse to the exchange rate realignments that were a feature of the ERM was not accompanied by the progressive convergence of the economic policies of participating countries that would have been required by a fixed exchange rate system. A fundamental inconsistency thus developed that, if not corrected opportunely, would have to have consequences such as those recently experienced. In effect, the imbalances that had accumulated and the resulting divergencies that had developed among the EMS countries over nearly a five-year period of fixity in their exchange rates were such that coping with them would have required responses of a far more joint nature than that envisaged in the rules of operation of the system. Perhaps the most important issue that turmoil in international money markets poses for the arguments made in this chapter is whether the emergence of monetary turbulence is not the most powerful argument in favor of the desirability of capital controls. I think the answer to this proposition is negative. It remains true that capital controls can only contain the market turbulence and, therefore, they represent, at best, temporary expedients to cope with currency turmoil; but they cannot replace actions to remove the underlying sources of the turmoil itself This much is, I believe, generally agreed. Rather than wonder about the need for capital controls as a means of insulating national economies, the focus of policy makers should be on the constraints that market forces place on economic policies. Capital mobility is but one illustration of those market forces. As is the case in other economic areas and markets, controls on capital transactions to eliminate or curtail such mobility can hardly be seen as efficient instruments to deal with economic imbalances. The choice for nation states in the current world environment is not whether they are willing to live with the constraints of interdependence but only on how they will live with them. A possibility is to generalize the regime of flexible exchange arrangements that replaced the Bretton Woods
88
Manuel Guitian
par value system.12 This line of reasoning is based on the desirability of containing the degree of interdependence among national economies. As argued earlier, flexible exchange rate arrangements tend to limit the extent of spillovers across economies. But also, as already pointed out, the ability of flexible exchange rates to provide insulation is more limited than often imagined, so the advantages to be gained from this course of action (as, indeed, the experience since Bretton Woods demonstrates) are likely to be more apparent than real. The other choice is to confront the reality of international interdependence and, with it, the inevitability of the constraints it imposes on national freedom of action and act accordingly; that is, design and adopt domestic policies that recognize the existence of those constraints. This is the approach behind the ERM in particular, and the EMS more generally. However, as made evident by the recent currency market events, the recognition of constraints is easier to state than to implement. The functioning of the ERM allows for essentially complete dissemination of spillovers across the participating economies, a feature that recent events in European currency markets have also highlighted. In this setting, currency turmoil, if limited in scale and duration, can be handled by actions described earlier in the paper - for example, sterilization and the tightening of fiscal policies. But these options presume that the source of the turmoil is transitory. Should it be of a more permanent nature, these steps are unlikely to be sufficient. And yet, these are not grounds to advocate (or resort to) capital controls. The fundamental instrument to handle sustained currency market turmoil must include two critical elements to be both effective and efficient: elimination of domestic policy inconsistencies; and the attainment of flexibility in the economy's domestic factor and product markets, so that domestic costs and prices can provide the valve of adjustment to economic shocks and imbalances and replace exchange rate flexibility. In sum, the only grounds on which to argue for capital controls are their short-term effectiveness and as an expedient means to cope with turbulence while more fundamental actions are taken. But they cannot be seen as substitutes for appropriate domestic policies or for the flexibility 12
This is the approach advocated for a long time by Milton Friedman and which he has reiterated recently in a commentary he made on the turbulence in money markets; see Milton Friedman, "Deja Vu in Currency Markets," Wall Street Journal, September 22, 1992. A formal move in this direction was made on August 2, 1993, when a decision was taken to increase the margins of fluctuation of EMS currencies around their central parities from 2.25 percent to 15 percent. This decision altered the balance between rules and discretion that had characterized the EMS, shifting it toward discretion. But it has not affected the basic arguments made in this chapter.
Capital account liberalization
89
required in all economies to confront certain shocks, be that flexibility provided by the exchange rate regime or, alternatively, by the possibility of upward and downward adjustments in domestic costs and prices.
References Calvo, Guillermo A., Leonardo Leiderman, and Carmen M. Reinhart. (1992). "Capital Inflows and Real Exchange Rate Appreciation in Latin America: The Role of External Factors." IMF Working Paper no. 62, August. Washington, D.C.: International Monetary Fund. de Vries, Margaret Garritsen. (1976). The International Monetary Fund, 19661971: The System under Stress. Washington, D.C.: International Monetary Fund. (1985). The International Monetary Fund, 1972-1978: Cooperation on Trial. Washington, D.C.: International Monetary Fund. Fischer, Bernhard, and Helmut Reisen. (1992). Towards Capital Account Convertibility. OECD Development Centre, Policy Brief no. 4. Paris: Organization for Economic Cooperation and Development. Frankel, Jeffrey A. (1991). "Quantifying International Capital Mobility in the 1980s." In Douglas Bernheim and John B. Shoven (eds.), National Saving and Economic Performance. Chicago: University of Chicago Press, 56-11. Greene, Joshua E., and Peter Isard. (1991). Currency Convertibility and the Transformation of Centrally Planned Economies. IMF Occasional Paper no. 81. Washington, D.C.: International Monetary Fund. Guitian, Manuel. (1988). "The European Monetary System: A Balance between Rules and Discretion." In Policy Coordination in the European Monetary System. IMF Occasional Paper no. 61. Washington, D.C.: International Monetary Fund, 38-54. (1992a). The Unique Nature of the Responsibilities of the International Monetary Fund. IMF Pamphlet Series no. 46. Washington, D.C.: International Monetary Fund. (1992b). "Monetary and Exchange Rate Policies for Stabilization and Reform." Paper presented to a seminar on The Transition and Adjustment at the Joint Vienna Institute, October 6. (1992c). Rules and Discretion in International Economic Policy IMF Occasional Paper no. 97. Washington, D C : International Monetary Fund. (1992d). "An Old-Fashioned Vision of Monetary Policy for an Emerging Europe." Paper prepared for a conference on The Monetary Future of Europe, El Pazo de Marinan, La Coruna, Spain, December 11-12. (1992e). "Remarks on the Debt Crisis." In Paul A. Volcker and Toyoo Gyohten (eds.), Changing Fortunes: The World's Money and the Decline of American Supremacy New York: Times Books, 43-58. Hanson, James A. (1992). "Opening the Capital Account: A Survey of Issues and Results." Policy Research Working Papers no. 901. Washington, D.C.: World Bank, 58-73. Horseneld, J. Keith, ed. (1969). The International Monetary Fund, 1945-65: Twenty Years of International Monetary Cooperation. Washington, D C : International Monetary Fund.
90
Manuel Guitian
International Monetary Fund. (1978). Articles of Agreement. Washington, D.C.: International Monetary Fund. (1991). Determinants and Systemic Consequences of International Capital Flows: A Study by the Research Department of the International Monetary Fund, IMF Occasional Paper no. 77. Washington, D.C.: International Monetary Fund. Lenain, Patrick. (1992). "Issues in External Financial Liberalization." Unpublished paper, Paris. Mathieson, Donald J., and Liliana Rojas-Suarez. (1992). "Liberalization of the Capital Account: Experiences and Issues." IMF Working Paper no. 46. Washington, D.C.: International Monetary Fund, June. McKinnon, Ronald I. (1991). The Order of Economic Liberalization. Baltimore: Johns Hopkins University Press. Polak, Jacques J. (1991). "Convertibility: An Indispensable Element in the Transition Process in Eastern Europe," in J. Williamson (1991), 103-19. Williamson, John (ed.). (1991). Currency Convertibility in Eastern Europe. Washington, D.C.: Institute for International Economics.
PART II
Capital mobility and macroeconomic policy in Europe
CHAPTER 4
The lessons of European monetary and exchange rate experience Patrick Minford
Seldom can the faults of a man-made system have been made as brutally clear as those of the European exchange rate mechanism (ERM) in the fall of 1992. For over a decade the ERM survived in a manner, defying gloomy predictions that its fate would be similar to that of the "Snake" in the 1970s, which disintegrated a few years after the collapse of Bretton Woods. The ERM survived even the abolition of exchange controls in 1990. But finally its luck ran out as it was hit by a series of large shocks the weak U.S. economy producing a falling dollar, German reunification generating prolonged high German interest rates, and the even split of French opinion over the Maastricht draft treaty for a future EEC. In this chapter I will try to explain why this has been a most damaging experiment in European monetary policy and has ultimately ended in predictable failure. I hope that I will be able to persuade monetary authorities to give a wide berth to policies based on fixed exchange rates, so perplexingly attractive to the elites of the non-German European continent; and that they will instead be tempted by the more robust AngloSaxon-German and now Japanese tradition of domestic monetary targeting within floating exchange rates. The European monetary dilemma
The project to stabilize exchange rates in the European Economic Community (EEC) and ultimately to have a single currency has an obvious political motivation behind it. Had it simply been part of a political program for creating a Eurostate, a continental superpower, then it would The work reported here draws heavily on Hughes Hallett, Minford, and Rastogi (1991) and Minford, Rastogi, and Hughes Hallett (1992). I am grateful to my coauthors, Andrew Hughes Hallett and Anupam Rastogi, for their permission to use this material. I am grateful for comments on this chapter from participants in this conference, especially my discussant Fa-Chin Liang, Victor Argy, Michael Dooley, Sebastian Edwards, and Alberto Giovannini.
93
94
Patrick Minford
have gained enthusiastic adherents but probably far more and far deadlier opponents, primarily nationalistic. For this reason the currency project has been put forward to the public as an essentially technical proposal designed to reap purely economic benefits. The basic thrust of these has been that stable, and ultimately immutable, exchange rates would create economic stability and so facilitate the single market. The problem, of course, with such a technical approach is that it flies in the face of much theory and evidence. But the fact that it was masking a deeper, political motivation has endowed the discussion within Europe with an almost religious quality, with those daring to question the technical details being regarded by the pro-unity majority in the same light as the Christian theologian questioning the historical accuracy of the Resurrection. Although there was, as during 1990 and 1991, a strong desire among the elite, and somewhat unreflectingly among the people also, for European unity, it was difficult to get a hearing for a reasoned questioning of the technical case on the continent of Europe: Only in Britain was it possible, though even there the attitude in that period was that of traditional tolerance for tiresome dissent. Only in 1992, as the economic environment soured, was the case properly heard - and, as so often when the veil of religious fervor is removed by some external event, people could not understand why they ever listened to anything else. The history of Euro-fixing started after the collapse of Bretton Woods in 1970. Germany was not unhappy with the new floating regime, indeed had mainly been instrumental in bringing it about through its opposition to the inflationary policies of the United States, the dominant creator of international liquidity under Bretton Woods. However, the other EEC governments, notably France and Italy, with the United Kingdom acquiescing, wanted a return to fixed rates. Their model was Bretton Woods, with the built-in flexibility of exchange rate change in response to "fundamental disequilibrium." Thus "fixed-but-adjustable" exchange rates within a cross-currency grid within the EEC was the basis of the Snake of the 1970s, set up on the recommendation of the Werner Report. Unfortunately for such a system, already by then international capital mobility had become huge; in spite of widespread exchange controls on nationals' transactions in domestic assets, there could be none on their transactions in foreign assets because these lay outside national jurisdiction, nor was there in practice any control over foreign nationals' transactions in domestic assets, because of the profits made from these activities (e.g., in the City of London from the Eurocurrency market). With high capital mobility, a fixed-but-adjustable exchange rate faces speculative attack, with markets betting on a devaluation for a safe one-way option bet (because if there is no devaluation they do not lose).
European monetary and exchange rate experience
95
There is another problem: the inconsistency between allowing a country no freedom to adjust its interest rates in the short term because it must protect its exchange rate, and yet allowing it full freedom in the long term when its exchange rate can freely move. Government interest rates and exchange rates can be destabilized in these circumstances, with consequent instability for the rest of the economy. These problems were only too apparent with the Snake, which soon collapsed under the weight of speculation in the context of the policy dissension set up by the first oil crisis, with some countries wishing to go for a more permissive medium-term policy to dampen the oil recession and others, especially Germany, wanting to lean hard against the inflationary consequences. The ERM itself was set up in 1979, under the impetus of a renewed French quest for fixity. It was buttressed by a new Franco-German entente. The early 1980s ERM was of the same fixed-but-adjustable sort as the Snake, though now parity changes had to be agreed by all members. As before, however, there were severe policy conflicts between the ERM's principles of fixity and the monetary desires of members. The most vivid example was France under Mitterrand's "dash for growth" in the early 1980s; it devalued a total of 45 percent against the deutsche mark. Italy too devalued sharply, by nearly 60 percent. The ERM nearly broke down. It was saved by the decision of Mitterrand to put his policy in reverse, using the popular excuse that it was necessary for the sake of the ERM and so of European unity. Our Liverpool work examined the potential instability for this type of ERM, using the technique of "stochastic simulation" where you pepper an economic model of the EEC with repeated typical shocks and compare the variability of the economy under different monetary regimes. As will be shown in more detail, we found indeed that this sort of ERM was likely to exhibit considerably higher instability than floating, especially if there were no attempts to coordinate monetary policies of the EEC members. In the light of such instability, the EEC was forced to move in one or other direction - back to floating or on to fixed. It moved toward more and more fixity, essentially for the political reasons mentioned previously, since 1987 there were - until the September 16, "Black Wednesday" meltdown - effectively no devaluations within the ERM. Finally the logic of this movement was to drive it to European Monetary Union (EMU) itself. Let us consider this intended process in two stages: the transitional period of the "tight ERM" and the final stage of EMU, irrevocably fixed exchange rates and the single money. The problem with tight ERM is that there is very little flexibility in interest rates as in the exchange rate. Interest rates must be set to defend the exchange rate, which can barely move. The result is that interest
96
Patrick Minford
rates may well be inappropriate for domestic needs, while the currency may get out of line with domestic costs and create international uncompetitiveness. These problems are clearly illustrated by Britain's recent experience. In early 1988, when we were only "shadow" ERM members, we kept the pound down against the deutsche mark and pulled interest rates right down to that end - they reached 7.5 percent in the spring - with the result that we had an uncontrolled boom that year. Then, from mid-1990 as we talked the pound up into joining by October, interest rates were held high to keep it up at our chosen rate of DM 2.95. Yet by then the economy was in free fall. By holding interest rates at 15 percent, and then 14 percent until mid-February, with only grudging falls afterward, the recession was prolonged and truly became a slump. The pound was also overvalued; this was clear from the continued current account deficit even in the depths of the recession. Yet an even more cautionary tale is available from Italy whose failure to devalue since 1987 combined with steady inflation well above Germany's had led, until Black Wednesday, to an overvaluation we estimate at around 30 percent. Italy coped with this by subsidies, both overt and hidden, to its trading industries. But subsidies create inefficiency and a poor allocation of resources. Some economists have argued that these problems would be largely resolved if exchange rates were irrevocably fixed. Then wages and prices would be under maximum pressure to come into line as devaluation could never bail out domestic industries. Interest rates too would be equal at some average European level - neither too high nor too low, provided there was a sound European central bank. This view has carried much weight with the EC Commission, which has accordingly pushed for rapid moves to EMU. EMU now seems likely to join the fate of the monetary union predicted (during the years of the Snake) for 1980 by Georges Pompidou. It too was once a prospect embraced with enthusiasm as an "inevitable" phase in the ongoing process of European Union. In brief, it failed to answer the issues raised in the optimal currency area literature of the 1960s. These amount to whether, in the undeniable absence within the EEC both of a large federal budget and of high interregional labor mobility, the loss of the exchange rate and interest rate stabilizers is seriously damaging to macroeconomic stability. It turns out that it is, roughly doubling the instability experienced under floating - we will give details of this calculation, in the course of our analysis of different exchange rate regimes. Though now unlikely to occur in the foreseeable future, its properties are of considerable relevance as a benchmark of what is achievable through monetary unification.
European monetary and exchange rate experience
97
Pnce Variab ility 225 200 _ /
175.
/-ERM •2
/
150. 125. 1
100_
y^-EMU
75. /* 50.
\
1 - • y~"Floating
25 _ 0 D
25
50
75
100
125
150 Output Variability
Figure 4.1. Minford et al. estimates on Liverpool model (variability measured by standard deviation). Key: 1 = money supply fixed; 2 = noncooperative contingent response; 3 = cooperative contingent response.
Figure 4.1 summarizes the results of our stochastic simulations of floating, the ERM and EMU The method used is discussed in more detail in the next section: Basically we take our estimated Liverpool multilateral world model (briefly described in the appendix) and apply to it repeated shocks typical of recent decades, computing the resulting variability under different regimes. The model assumes rational expectations so that the regime changes are embedded in expectations reactions; though the parameters of the model are not completely immune to regime shifts, they were estimated over a combination of fixed and floating years' data and show apparently little evidence of shift between these regimes. Thus there may in practice be little vulnerability to regime effects on model behavior (Lucas's critique in Lucas 1976). In any case, alternative methods of econometric evaluation are simply unavailable in the case of EMU and weak in the case of the European Monetary System (EMS) because of difficulties in controlling for nuisance factors; therefore, this method seems to be the best we can offer for these purposes.
98
Patrick Minford
Our calculations show price and output variability for the EEC on average under these regimes, with different assumptions about monetary policy. It can be seen that whatever monetary regime is assumed, the ERM is approximately twice as unstable as EMU (both for prices and output) because of the adjustability of the parity assumed in it: This both promotes speculative attack and permits inconsistent monetary policies leading to sharp parity changes from time to time. Given the drive to unity, it indicates just why the ERM was planned to become more rigid and ultimately (some argued very rapidly) converge on EMU. The more rigid the system the closer the instability drops to the EMU level. However EMU is also about twice as unstable as floating, illustrating the scale of the optimal currency area effect in the European case. Again this is so regardless of monetary policies being followed. We looked at a Friedman rule of fixed money, unresponsive to shocks (marked 1 on Figure 4.1), at independently responsive monetary policies (2; this is ruled out under EMU by definition), and finally at monetary cooperation to find national monetary responses most appropriate to the EEC as a whole (3). What is interesting is that this last case under floating is by a fair margin the "best buy" for the EEC. It indicates the direction of possible reconstruction of the ERM, as well as giving a pointer for other countries desiring regional monetary cooperation. The ultimate dilemma of the EEC in monetary matters has been that, on the one hand, there was a political logic demanding monetary unification, with no viable stopping place before full monetary union, against, on the other, an economic logic requiring the restoration of floating exchange rates in order to avoid serious macroeconomic instability. Until Black Wednesday the former, political logic carried the day. After it, the economic damage of the monetary unification process looks too large to bear. I now turn in more detail to the theory and empirical evidence (summarized in Figure 4.1) lying behind the statements made above about macroeconomic instability. Evaluating monetary union
A good starting point for discussing the advantages of a common currency is the report of the European Commission (1990), a group that has made a heroic attempt to measure them. It speculates that the efficiency gain from removing currency uncertainty and exchange costs may be worth as much as 10 percent of EC GNR Virtually all of this comes from the effect of a supposed reduction in the risk premium on the cost of capital. Various other gains are adduced from the common currency: increased price sta-
European monetary and exchange rate experience
99
bility (including through enhanced credibility), more disciplined public finance, and greater macroeconomic stability. To begin with price stability, on which the report places much emphasis, this is attainable with or without EMU, as is credibility. Whether EMU makes them more easily attainable is a matter of political economy, which is not tackled by the EC's report: What will the European central bank's powers and incentives be, and what legitimacy will be conferred on it by the twelve democratic peoples of the EC? These questions are understandably not addressed in the EC report but merely assumed away. But even if they were convincingly answered, there would still be the alternative of pursuing price stability credibly by domestic means. Domestic means of commitment exist and are effectively achieved in many OECD economies such as Germany, the United States, and Japan: Essentially they involve some form of domestic nominal target, whether the money supply or nominal income or simply price behavior, with the commitment penalties supplied by the political process. To put it crudely, governments that fail to deliver acceptable price performance lose votes - ultimately this must be the only effective penalty in a sovereign democracy. It is perfectly feasible for the electorate to penalize the abandonment of counterinflationary commitments, while rewarding appropriate responses to shocks (provided, of course, these can be roughly monitored). The idea of a central bank that is truly independent of the political process - created so to speak by economists as a benevolent tyrant - has no basis in political economy. However, an electorate wishing to implement a penalty system of the sort just described might well endow a central bank with the outward forms of independence, to assist it in the monitoring process (Minford 1992). This domestic commitment process highlights the difficulties inherent in designing a Europe-wide commitment process for monetary policy. As the Germans have persistently pointed out, there is a "democratic deficit" in the European "union." Without democratic political union there would indeed be a lack of democratic legitimacy for a European central bank imposing tough monetary policies on twelve (or more) national regions of Europe, some of which may well be suffering from severe recession in spite of an EC-wide inflation. This lack would also fatally undermine credibility. If we turn to public finance the report envisages greater discipline arising from the inability of a national government to raise taxation through printing money ("seigniorage" or the "inflation tax"); this inability is reinforced by the Maastricht fiscal criteria, designed to limit the risks of country bailouts. Here a virtue is made of what in the literature is regarded as a problem, making optimal taxation more difficult. However, the EC report offers no mechanisms, only pious hopes and peer country pressure,
100
Patrick Minford
for achieving movement to fiscal balance under EMU. No seigniorage and no alternative fiscal mechanism - it hardly seems to constitute a fiscal improvement. The distressing implications are illustrated by the difficulties of Italy, unable to raise taxes through inflation, condemned by the Maastricht criteria to ambitious targets for primary budget surpluses, and yet (judging by long-standing performance) without the political capacity to deliver the necessary taxes or spending cuts. The large efficiency gain speculated on in the EC report comes from removing the transaction costs of currency exchange and the hedging costs of guarding against currency uncertainty. Two arguments suggest its estimate is likely to be well on the high side: 1 Currency risk is diversifiable in a world of many currencies and investment vehicles whose risks are correlated with currency risk. Hence the cost of hedging should tend to the premium on specific risk, namely zero. 2 A transaction involving currency exchange should not, on the face of it (given the negligible cost of keying in electronic orders), cost any extra in credit transactions than an ordinary transaction in home currency, other than the cost of hedging net balances in any one currency between clearings. These two arguments suggest that the only saving comes in the exchange of notes and coin. But this is extremely limited: Notes and coin are generally a small percentage (with one or two exceptions such as Italy), and that part of it which is exchanged for foreign notes and coin is a small percentage of that again. The EC Commission's estimate of the transactions cost saving on its own is 0.4 percent of GDP; this is based on a survey of financial firms' commission charges, and these are applied to estimates of the volume of business attracting these charges. Although this estimate is more believable than the huge aggregate figure, it may still be on the high side. Indeed the report suggests that it is heavily concentrated among the smaller countries with less sophisticated banking systems (and for a country with a sophisticated system the estimate may be as low as 0.1 percent). We have seen so far that the only firm gain from EMU adduced by the EC Commission is that connected with transactions costs savings: and this, at 0.1 percent of GDP for a country like Britain or Germany is extraordinarily small. Indeed, if the gain were as large as the EC estimate, then other pairs of nations enjoying a similar degree of intertrade would have surely actively considered a common currency. Yet the countries of European Free Trade Agreement (EFTA), of North America, of eastern
European monetary and exchange rate experience
101
Europe, to name but a few candidates, have never seemed to put this idea seriously on the treaty agenda. This last point suggests that, whatever the truth in the European Commission's estimate, there must be some key drawback to nations having a common currency. Indeed, the extensive literature on the "optimal common currency area" (e.g., Mundell 1961, McKinnon 1963, Kenen 1969) points to just such a drawback. Its thrust is that exchange rate adjustment is, because more rapid than national wage and price adjustment, more effective in stabilizing an open economy in the face of differential national shocks (an extreme example of which would be German reunification). This thrust is contrary to the commission's last assertion, that EMU would bring greater macroeconomic stability. The optimal currency literature invokes a variety of factors that may improve the stabilizing capacity of monetary union - notably, a high degree of labor mobility and a central budget capable of making large fiscal transfers. But these factors are, as we shall see, absent in the case of the EC. I wish now to examine how serious a drawback this "stabilization" aspect is for EMU, using the method of stochastic simulation. Although this method has its weaknesses, to be discussed, there is no viable alternative: Econometrics cannot disentangle the effects of monetary union from longfunctioning unions such as the United States and there is no data on European union. What econometric work exists merely bears on the extent of integration across European countries: For example, Bayoumi and Eichengreen (1992) show that the United States is more integrated than the EC, while Grauwe and Vanhaverbeke (1991), as well as Hagen and Neumann (1991), show that the EC is less integrated across countries than it is within countries. Stabilization gains and losses of monetary union In evaluating the stabilization aspect for any given common currency proposal, one must make assumptions about the institutional framework. It makes a lot of difference whether there is a high degree of labor mobility and what the fiscal transfer arrangements are. In the case of the EC, it has been noted by Eichengreen (1990) that the fiscal offset to any national or regional decline in GDP is less than 1 percent (as against 30 percent in the United States, for instance); nor are there any plans to raise this offset coefficient to any number remotely comparable with the U.S. one. The Delors Committee (Delors 1989) called for a doubling but even that may well not be agreed by the nations that have to double their fiscal contribution to Brussels (still less the much larger amounts suggested as the minimum necessary by Macdougall 1977).
102
Patrick Minford
Labor mobility is also limited, for the vast bulk of nationals in the richer countries of the EC. The key reason appears to be language and cultural differences, which make for instance a Frenchman pause before resettling in Frankfurt. What significant migration there is is from the poorer countries to the richer, as immigration controls that normally stop this are invalidated. But even this is not great because capital mobility within the free EC market enables workers in poor countries such as Spain, parts of the United Kingdom and of Italy, on the "periphery" of the EC, to attract investment and enjoy improved wages without the cost of moving. We assume that trade is free for our purposes here even if 1992 has not yet fully come through and may not do so until the next century. We also assume that capital is perfectly mobile within the EC and that there are no exchange controls. It is obvious, as the optimum currency area literature stresses, that whatever stabilization task is performed by flexible exchange rates and divergent monetary policy, it can partially be substituted for by either labor mobility or fiscal transfers. To that extent, the EC is handicapped in its bid to be an optimum currency area. Our calculations will reflect this handicap. (If EMU goes ahead regardless, that might well lead to demands for further progress on fiscal policy and labor mobility, but that is another matter.) Before proceeding, we should note a point frequently made by those in favor of EMU: that monetary arrangements as such should not alter the long-run equilibrium state of the real economy, merely national price levels. Real business cycle theorists would go further and argue that even the short-run behavior of real variables should be immune to monetary arrangements, provided that they do not constitute a change in transactions technology. The approach here is straightforward. I agree, and the models I shall rely on explicitly assume, that real long-run equilibriums are immune to monetary arrangements. But these models feature a variety of ways in which short-run behavior differs under different monetary arrangements: So my position is not that of the real business cycle approach. That approach has not yet come up with a representation of European economies that can rival that produced by more conventional modern models with their stress on nominal contracts, monetary surprises and "rational" (informationally efficient) expectations. We still have to answer the question of whether one can properly capture behavioral differences across such different regimes as flexible rates and the totally fixed rates of a common currency. This problem, Lucas's critique (Lucas 1976), was addressed briefly, where I argued that in practi-
European monetary and exchange rate experience
103
cal terms it should not be too serious for our work, given that the key concern originally addressed by Lucas, the imposition of rational expectations, is dealt with explicitly. In principle, however, it is inherent in most pieces of applied work that address issues of policy interest, which usually involve whether there should be a different policy regime. Ultimately, there is no totally satisfactory answer. Either one comes up with a model of stylized consumers and producers whose tastes and technology can be identified and assumed not to change - but such a model will have a poor fit and is short of institutional relevance - or one produces a model of aggregate supplies and demands, as used here, in which the estimates fit reasonably but the estimated coefficients themselves may be found to shift as people change their behavior in changed environments. I make main use of the Liverpool multilateral model linking separate models of the major OECD economies: In addition I summarize some parallel work that has been done on Multimod, the IMF's simulation model of the same broad type. Both are reasonably well known representatives of the class of model described earlier. Multimod (Masson, Symansky, and Meredith 1990) takes parameters from the empirical literature and is a simulation rather than a forecasting model; it assumes fairly long overlapping wage contracts. Liverpool (Minford, Agenor, and Nowell 1986) is fully estimated; it uses a "new classical" framework in which contracts do not overlap but are disturbed by annual monetary surprises. In both models adjustment costs produce long-drawn-out adjustment. The Liverpool model has been estimated over both fixed and floating periods and appears to have reasonable stability across this particular regime change (as noted for the U.K. component in Minford, Ioannidis, and Marwaha 1983). To this extent it should be serviceable for our study here, which can be thought of as comparing floating rates with ERM, which forms part of the sample period, and with stages 2 and 3 of the Delors proposals (Delors 1989), where currencies are "irrevocably" pegged with a trivial margin (1 percent at stage 2) or none at all (stage 3). Such a model cannot be used with as much confidence to evaluate a Europe in which there is the common currency itself - but it may be the nearest we can get with our present econometric technology. One should remember nevertheless that even after a common currency is established, it is in principle possible for any sovereign country to withdraw from it later. It would require the cessation of national sovereignty and, therefore, complete political union to guarantee totally against secession; so perhaps EMU and stages 2 and 3 are not so distinct as might at first appear. Before embarking on this comparison, it is worth asking what one might expect to find. Holding money supplies of relevant countries constant (as in the rule suggested by Friedman 1968), one would observe the
104
Patrick Minford R I
M [FLOATING]
B [FIXED]
Figure 4.2. Framework for monetary and demand shocks (under floating LM operates, under fixed BB operates). relative capacity of the two exchange rate regimes to stabilize shocks "model stability." We have already explained in broad terms how, following the arguments of the optimal currency area literature, we would expect greater stability under floating exchange rates (with the implied flexibility of national interest rates). More specifically, if one thinks of the Liverpool and Multimod models as up-to-date versions of the Mundell-Fleming model (Mundell 1962, Fleming 1962 - the updating consists in rational expectations and wage-price transmission but with some short-term nominal rigidity remaining), then it is a well-known feature of this model that only under domestic monetary shocks will totally fixed exchange rates stabilize output more than floating exchange rates. In this case holding interest rates constant, with the money supply endogenous, as under fixed rates, perfectly stabilizes output. However, with shocks to aggregate demand or to foreign monetary conditions, the variation in interest rates and the exchange rate under floating stabilizes output more, as the IS curve has to move along the LM curve (see Figure 4.2). If one considers supply shocks (Figure 4.3), it remains likely that floating will stabilize output more than fixed rates, since the fixed rate aggregate demand curve (arising from declining competitiveness and rising real interest rates as prices rise) should be shifted rightward under floating by the decline in the exchange rate. The effects on prices, real interest rates, and real exchange rates are complex and ambiguous; nevertheless, if there is substantially greater output instability under fixed rates, one would presume that this would also
European monetary and exchange rate experience
105
Figure 4.3. Negative supply shock illustrated: under fixed AD curve does not shift, under floating it shifts to A 'D' as exchange rate falls.
destabilize these other variables. The overall balance of effects therefore points to greater stability under floating, unless (as has not been the case in the past) the predominant source of shocks is disturbances to money demand (for an essentially similar analysis, and some further empirical backing from deterministic simulations, see Hughes Hallett and Vines 1990). When one turns to the possibilities for active monetary responses to shocks - something that few would wish to rule out, especially when shocks are large and identifiable - there may appear to be a trade-off between the number of independent monetary instruments for dealing with differential national shocks (maximized under floating) and the extent of monetary coordination (which must be total under fully fixed rates). Nevertheless, coordination under fixed rates will not imply the same set of monetary policies as coordination under floating rates; and it must in principle be the case that cooperation using the full scope of independence granted by floating will be optimal. Again, therefore, we would expect floating to have the advantage in this aspect of the competition. The results of the available studies are summarized in Figure 4.1 (our own) and Figures 4.4 and 4.5 (those using Multimod). Full details of these studies are in European Commission (1990, annex E), Minford et al. (1992), and Masson and Symansky (1992). Both the commission and Masson and Symansky use Multimod for their stochastic simulations. However the commission's study has been
106
Patrick Minford Inflation Variability 120 _,
110.
Float
100.
^^\^/
90.
80.
D
70.
EMS (evolving)
EMU (fixed exchange rates)
EMU (single currency)
60
i
80
EMS (of mid - 198C
90
100
i
110
i
120
i
130
140
Output Variability
Figure 4.4. EC Commission estimates (variability measured by standard deviation). criticized in both the other studies for two main aspects of its approach. First, it attributes an implausibly high variability to risk premiums on interest differentials between EC currencies (these risks are, of course, eliminated under EMU). Second, it assumes somewhat unusual monetary policy regimes (rules where interest rates vary in response to inflation and output targets). Using these assumptions, it finds that EMU is superior in stability to floating. When Masson and Symansky allow for plausibly lower risk premium variability and for alternative monetary regimes, they obtain results more in line with the prior expectations described previously. Altering the commission's simulations for risk premiums alone reduces substantially the "advantage" it found for EMU. However it is not the whole story: The monetary policy assumptions made under the differing exchange rate regimes must also be appropriate. If a destabilizing monetary policy is assumed under floating, while a stabilizing one is assumed under EMU, EMU's relative stability will be biased upward.
European monetary and exchange rate experience
107
Price Variability 170^ 160. 150_ 140_ 130. 120_ 110. 100. FLOATING 90. 80_ 70. 80
90
100
110
120 130 Output Variability
Figure 4.5. Masson and Symansky estimates with Multimod (variability measured by standard deviation). Key: R3 = 3% bands, real exchange rate trigger for realignment; E5 = 5% bands, equilibrium realignment; M(E/F) = money target; Y(E/F) = nominal income target; E = EMU: F = floating. In principle one wishes to assume under each exchange rate regime the optimal monetary policy within the feasible set: That reveals the best potential performance of each exchange rate regime. It is also reasonable to ask how each exchange rate system will perform under optimal automatic rules for monetary policy: It may be better for such rules to be embraced by politicians because of the difficulties surrounding successful discretionary policy. Again, though, one would wish to choose the best automatic rule for monetary policy under each exchange rate regime. Masson and Symansky make two main comparisons: Under fixed money supply rules and under rules where interest rates target money GDP. Their work makes it possible to put the Multimod simulations alongside those using the Liverpool model produced by Minford et al. (1992). These authors looked at a different set of monetary regimes, in general richer, in respect to automatic rules not quite so rich. In respect to regimes with nonautomatic responses (discretionary policy, in which
108
Patrick Minford
however there is assumed to be no attempted manipulation of expectations by false promises - "time inconsistency"), they investigated fully optimal monetary policies, both under Nash noncooperative policies and under EC cooperation. Among automatic rules, however, they restricted themselves solely to fixed money supplies. The main point that emerges from the three sets of results shown in Figures 4.1, 4.4, and 4.5 is that, whether one uses the Liverpool or the Multimod models, once shocks are put on a comparable basis, then floating gives scope for substantially better stability, provided monetary policies are suitably chosen, than EMU. This result for Multimod is not as strong as for the Liverpool model; this may well reflect the absence of optimization in the Multimod exercise. However, what emerges overall is essentially the conclusion one would reach from following the logic of the optimal currency area literature in its modern stochastic version. I have avoided any discussion of the political passions raised by the issue of sovereignty. But, in fact, no nation would be abandoning its sovereignty under the proposals for union so far discussed. Effectively these have been for little more than totally fixed exchange rates with a single monetary authority; any country disliking the result could in principle leave it and issue its own currency again, supposing this to have been merged without trace (again not proposed given intentions, e.g., to issue in the United Kingdom notes with the queen's head on one side). It is this very lack of a political dimension (and so lack of a serious central budget as well as limited labor mobility) that loads the dice against macroeconomic stability, besides the problem of legitimacy. Yet, as is widely conceded, it is totally unrealistic to discuss any political dimension. In these circumstances one can perhaps discuss the merits of a move to total exchange fixity in economic terms purely. If the economic gains were great, then a treaty - reversible like all treaties - acquiring them could be worthwhile, even though there would still be inevitable doubts about the practical problems of reversibility and the resulting political hostage to fortune. In this sense political doubts over sovereignty place the burden of proof on EMU. The ERM and possible transition to monetary union The macroeconomic instability caused by EMU at the present stage of economic integration in Europe, when set in the scales beside the small gain in transactions costs discussed earlier, seems to outweigh it decisively. This is not to say that increasing integration - that is EC intratrade in goods, services, and factors of production, nothing whatever to do with
European monetary and exchange rate experience
109
the "convergence" hitherto discussed between governments - will not eventually push the calculation the other way, for a gradually increasing set of nations ("variable geometry" or "n-tier" union). Increasing integration both raises the gain from transactions costs and reduces the loss from instability (as shocks become more similar in nature and impact). So there is a need for a transitional regime that both permits this evolving union to occur at the appropriate moments and works effectively for the united as well as the not-so-united. As noted summarily the entries for the exchange rate mechanism in Figures 4.1, 4.4, and 4.5 are palpably worse than for EMU, illustrating the generally appreciated feature that the ERM creates a degree of fixity without full credibility or ultimate policy commitment. This gives the worst of both worlds, and accounts for the one element of agreement on all sides: that the system should move either to complete fixity (and so EMU) or back to floating. The derivation however of the ERM results is a vexed matter. There are potentially as many sets of results as there are potential rules for exchange rate adjustment. In Figure 4.6 we present a scatter of such potential results, using the Liverpool model. The very large instability in our "default EMS" comes about partly from the absence of national monetary policy optimization especially with cooperation. (It is also noteworthy that capital controls further destabilize the ERM.) Optimize and cooperate, or prevent nations from exploiting the parity change option, and you obtain results closer to EMU, and comparable therefore with those just discussed (there does also seem to be somewhat more volatility, according to Multimod, under our forwardlooking equilibrium exchange rate adjustment as compared with the backward-looking ones examined by Masson and Symansky). Broadly the two models seem to reach comparable conclusions: namely, that ERM is less stable than full EMU, how much so depending on just how rigid (close to EMU) the ERM rules are. On that there is certainly general agreement. What could Europe do?
There is, as we have seen, one regime that gives both good results and achieves the sort of cooperation that has been one of the proclaimed aims of those who set up the ERM. The cooperation that it involves between central banks should also smooth the path to monetary union when appropriate; at such a moment the cooperation would become a lockstep by agreement. The regime in question is illustrated in Figure 4.6 for the Liverpool results (unfortunately, it is not available for Multimod): cooperative
110
Patrick Minford Price Variability 300 _ 275. 250.
ERM with Capital Controls
225. 200.
Default ERM ( As in Fig. 1)
175. 150.
ERM - No Parity Change
125. 100.
'EMU
75 50
75
100
125
150
200 175 Output Variability
Figure 4.6. Some ERM variants-Liverpool Model. Key: 1 = money supplyfixed;2 = noncooperative contingent response; 3 = cooperative contingent response. floating, numbered 3 among the floating regimes. This regime of course achieves the best of both worlds: exploiting the degrees of freedom in floating while enjoying the cooperation intended by ERM. There are those who argue that such cooperation is inconceivable within Europe without the rules of an exchange rate arrangement like the ERM. Yet cooperative equilibrium (which respects individual country interests but arranges mutually improving policy trades) is a dominant one when players can communicate, since they have a joint interest in achieving it. Hence the argument can be turned on its head: It is unlikely that, absent the suboptimal constraints of the ERM, central banks used to communication would abstain from such a cooperative equilibrium. Politically too the arrangement would be attractive, given the concerns mentioned at the start of this essay. In short what I am proposing is a reformed stage 2, in which the European Monetary Institute could take on the role of coordinating forum in
European monetary and exchange rate experience
111
which the twelve governments, represented by their central banks under whatever national controls are desired, could meet regularly and share their monetary plans. All exchange rate bands and intervention targets would be abandoned: The central banks would set monetary targets, both for suitable money supply measures and for operating instruments, especially interest rates. They would also design their own detailed rules and operating procedures for achieving this cooperative equilibrium in everyday practice.
Conclusion
The proposal here is a route out of the present monetary impasse in Europe, where there is little enthusiasm, rightly on our view, for EMU and yet substantial and - again justified - dissatisfaction with the current ERM. The proposed route is an evolving structure of central bank cooperation that exploits the full stabilizing potential of floating exchange rates. Recent European monetary experience has shown dramatically the dangers of premature currency linkage. It is no accident that in spite of two "accords" (Plaza and Louvre) intended to stabilize exchange rates, the major world economies have continued to float, with only limited feedback from exchange rates to monetary policy. The few EEC economies that have tried linking to the (floating) deutsche mark have had a rough ride. For the observer the lessons are clear. In a world of large shocks, rapidly transmitted internationally, use the only system with a proven modern track record: domestic monetary control within floating exchange rates and a free international money market.
Appendix: outline of the Liverpool world model and summary of results
The model consists of nine country models (all on floating exchange rates except the two Benelux countries fixed to a "basket") and three trade blocs (export and import prices and volumes for each bloc) to link those other countries not modeled closely. It is discussed now as if it were a twocountry model ("United States" and "Europe," say), with a floating exchange rate between the two: The 'home' Country can be described (in log-linear form) by the following stylized equations:
112
Patrick Minford
y = 8e6 - 8/ - 8ee + 8FyF 4- 8/1 4- kly_l + e8
(IS)(8F < 1)
(A-l)
y = <rp(p- Ep_x) + ae(e - X2e_x)
(PP) = ap
(A-2)
(LM)
(A-3)
(money supply)
(A-4)
(deficit process)
(A-5)
x = -flee - pyy + (3FyF
(current balance)
(A-6)
A0 = c/)(x + ^d - Ty) - qkR - Ap
(balance-sheet constraint)
(A-7)
r = rF-
(efficient market condition)
(A-8)
(nominal-real interestrate identity)
(A-9)
4- Aj>_, + e?
m =
+ ixQ0 - IJLRR + p + \4(m - p)_x + e^
\L^.
Am =
Ee+l - e
R = r + Ep+l - p,
y = output (log) 6 = real value of financial wealth (log) p = consumer price index (log) m = money supply (log) r = real interest rate (fraction per year) R = nominal interest rate (fraction per year) e = real exchange rate (fractional departure from equilibrium) d = government deficit, including interest payments, as a fraction of GDP d = cyclically adjusted d x = current account balance (fraction of GDP) E~x = rational expectation on data through / - i F = subscript denotes "foreign." All coefficients are positive. The ey are error terms, which may be autocorrelated. The constant terms have been set at zero, implying that all real variables (y e, 6, r) can be treated as deviations from equilibrium. Equations (A-l) and (A-3), IS and LM curves, come from interrelated private-sector demand functions for financial assets, physical assets, and nondurable consumption. The main point to note is the role of financial wealth, 9, in demand for both goods and money. Equation (A-2) is the PP, or Phillips, curve that relates output to un-
European monetary and exchange rate experience
113
anticipated inflation (seen from the last year, on the grounds that there is a one-year wage-contracting lag for a significant unionized fraction of the labor force) and, the open economy aspect, the real exchange rate; the latter effect arises because as the terms of trade improve (the real exchange rate rises) the consumption real wage increases relative to the own-product real wage. In equation (A-4), the money supply function, the assumption is that the government pursues monetary targets dictated by its expected equilibrium deficit, except for temporary spells (em, which may be a process with autoregressive and moving-average components) when it attempts exogenously to vary the fiscal-monetary mix. This long-run tendency to go for "balanced" monetary financing is the result of the intertemporal budget constraint on government, which prevents permanent "monetary" financing of a deficit. 6 is the ratio of GDP to financial assets; hence dd is the expected long-run rate of injection of financial assets into the economy arising from the government's deficit, and equation (A-4) states that this will ultimately be matched by the rate of monetary injection. Since both d and such temporary spells are assumed to be exogenous, the level of M at any time is exogenously determined (by the history of d and these spells). Equations (A-3) and (A-4) give rise to the LM curve in the (p,y) domain in Figure 4.7. Equation (A-5) is the postulated process driving the exogenous cyclically adjusted deficit (tax rates and government spending as a fraction of GDP are thus exogenous, while the actual deficit is endogenous); it is treated as a random walk (as are tax rates and government spending by implication). The parameter T reflects the marginal (net) tax rate over the cycle. Equation (A-6) is a standard net exports (current balance) equation. Equation (A-7) then equates changes in financial wealth with this current balance, the change in net foreign assets, plus the deficit, the change in government liabilities, minus the effects on asset values of inflation and interest rate changes. Equation (A-8) is the interest parity condition adjusted for expected exchange rate change, in terms of real interest and exchange rates. The information available is assumed to be contemporaneous in this annual setup. This is identically equivalent with the usual formulation in terms of the nominal interest differential. Note that our definition of real interest rate uses the consumption deflator, p, and that of the real exchange rate uses the two countries' consumption deflators converted to a common currency. These deflators do, of course, include an effect of foreign prices through the prices of imports. Equation (A-9) is the familiar Fisher iden-
114
Patrick Minford HOME
FOREIGN
W
Pfi P Figure 4.7. Full macroeconomic equilibrium in the world economy. tity. The model for the other economy is a mirror image. Notice, however, that xf = —x, and ef = —e, so we can drop the equivalents of equations (A-6) and (A-8), obtaining yF =
8fdF (A-l 1) A > F - pF.x)
AmF = V'dF + <,
(A-12) (A-13) (A-14)
European monetary and exchange rate experience F
A0F = <\> {-x + dFRF=rF+
EpF+l -pP
F
F
T yF) - q ARF - ApF
115
(A-15) (A-16)
Rational expectation models of this type have by now a rather familiar behavior in solution. There is, first, the impact effect of an unanticipated shock or "surprise." Secondly, there is the model's path back to equilibrium from this impact along its "stable manifold" - that is, as determined by the stable roots of the model's characteristic equation - which in a wellbehaved model will be unique (the condition for uniqueness is that the number of nonstable roots be equal to the number of expected future variables). This path will in turn have two elements: a moving average component (in which the economy "jumps" on to the stable manifold) and the effect of the stable roots (the stable manifold). Hence, an example of a final form equation for an endogenous variable, z,, will be (1 - p{L) (1 p2L) zt — 7ry(l - rrijL) st + TT2(\ - m2L) where L is the lag operator, pv p2
the stable roots and en e2 the error terms in the structural equations. It so happens that in the Liverpool world model the final-form equation approximates to a first-order process (1 - pL)zt, = . . . where/? is of magnitude 0.7. Furthermore, for the real exchange rate, et, the net impact of the moving-average terms is very small, so that (1 - 0.7L) et — r\t where ?), is a composite "surprise" effect from all the various structural errors. This allows us to illustrate the model graphically in a rather simple way: Since Ee+l — 0.1 e, r = rF + (1 - 0.7)e, whence we may eliminate r from the model in terms of rf and e. Figure 4.7 shows the IS and PP curves for the home country in (e, y) space in the upper right quadrant and the LM curve in (/?, y) space in the lower right one. The WW curve shows values of e and y for which A6= 0; that is, full stock equilibrium: To its right 6 is falling, to its left 0 is rising. Since the IS curve is shifted leftward by a fall in 6, the system equilibrates by pushing the ISIPP curves' intersection to the WW curve. Because d = Am* = Ap* (monetary equilibrium), y* is determined as that level of supply (along the PP curve) at which there is current account balance. Prices then settle wherever this income intersects the LM curve. For the foreign country, the same applies. (Of course, the foreign price, PF, is determined independently of p because the floating exchange rate interposes to make them consistent with each other and e.) Only the slopes of IS, PP, and WW curves change sign because e - —eF Clearly, e has to settle at the same value for both countries. This is achieved by rp the "world" real interest rate, which moves the IS curves of the two countries until aggregate supply equals aggregate demand in both countries at the common real exchange rate.
116
Patrick Minford References
Bayoumi, Tamim, and Barry Eichengreen. (1992). "Shocking aspects of European monetary unification" Cambridge, Mass.: NBER Working Paper no. 3949:[l]-39, January. Delors, I (1989). "Report on economic and monetary union in the European Community." In European Commission, The Delors Report. Luxembourg: Office for Official Publications of the EC. Eichengreen, B. (1990). "One money for Europe? Lessons from the U.S. currency union." Economic Policy 5 (April): 118-87. European Commission. (1990). "One market, one money - an evaluation of the potential benefits and costs of forming an economic and monetary union." European Economy no. 44 (October): 3-347. Fleming, John Marcus. (1962). "Domestic financial policies under fixed and under floating exchange rates." IMF Staff Papers 9 (November): 369-79. Friedman, M. (1968). "The role of monetary policy." American Economic Review 58:1-17. Grauwe, P. De, and W. Vanhaverbeke. (1991). "Is Europe an optimum currency area? Evidence from regional data." CEPR Working Paper no. 555, May, 1-23. Hagen, J. von, and M. Neumann. (1991). "Real exchange rates within and between currency areas: how far away is EMU?" Discussion Paper no. 62. Bloomington: Center for Global Business, Indiana University. Hughes Hallett, A., P. Minford, and A. Rastogi. (1991). "The European Monetary System: achievements and survival." CEPR Discussion Paper no. 502, January, 1-71. Hughes Hallett, A. X, and D. Vines. (1993). "On the possible costs of European Monetary Union." Manchester School of Economic and Social Studies 61 (March): 35-64. Kenen, P. B. (1969). "The theory of optimum currency areas: an eclectic view." In R. A. Mundell and A. K. Swoboda (eds.), Monetary problems of the international economy. Chicago: University of Chicago Press, 41-63. Lucas, Robert E., Jr. (1976). "Econometric policy evaluation: a critique." Journal of Monetary Economics 1, no. 2 (supplementary series): 19-46. MacDougall, Donald. (1977). Public Finance in European Integration, EC Commission, The MacDougall Report. Luxembourg: Office for Official Publications of the EC. Masson, P., and S. Symansky. (1992). "Evaluating the EMS and EMU using stochastic simulations: some issues." Mimeograph copy. IMF. Masson, P., S. Symansky, and G. Meredith. (1990). "Multimod Mark II: a revised and extended model." IMF Occasional Paper no. 71 (July). Washington, D.C.: International Monetary Fund. McKinnon, R. I. (1963). "Optimum currency areas." American Economic Review 53: 717-25. Minford, P. (1992). "Time inconsistency democracy and optimal contingent rules." Mimeograph copy. University of Liverpool. Minford, P., P. Agenor, and E. Nowell. (1986). "A new classical econometric model of the world economy." Economic Modelling 3:154-74. Minford, P., A. Rastogi, and A. Hughes Hallett. (1992). "The price of EMU revisited." CEPR Working Paper no. 656. Revised and extended version of P. Min-
European monetary and exchange rate experience
117
ford and A. Rastogi (1990), "The price of Emu." In R. Dornbusch and R. Layard (eds.), Britain and EMU. London: Centre for Economic Performance in association with Financial Markets Group, LSE, 173-204. Minford, P., C. Ioannidis, and S. Marwaha. (1983). "Dynamic predictive tests of a model under adaptive and rational expectations." Economic Letters 2: 115-21. Mundell, R. A. (1961). "A theory of optimum currency areas." American Economic Review 51: 657-65. (1962). The appropriate use of monetary and fiscal policy for internal and external stability. IME Staff Papers 9 (March): 70-77.
CHAPTER 5
Experience with controls on international capital movements in OECD countries: solution or problem for monetary policy? Jeffrey R Shafer
1
Introduction
Whether, and to what extent, to remove controls on international capital flows is a key strategic question for Korea and other countries in the Pacific Basin where economic capacities are advancing rapidly and where institutions are being transformed in other spheres. A large and still growing literature has developed around the experiences and situations of the more advanced developing economies.1 With Korea and some other countries in the Pacific Basin coming to resemble more closely the affluent market-oriented countries of the Organization for Economic Cooperation and Development (OECD) in economic productivity and institutions, the experiences of this group of countries and the lessons to be learned from them have become more relevant. During the 1950s and 1960s the OECD countries employed capital controls extensively and as a matter of routine as policy instruments in the conduct of monetary policy. Beginning in the 1970s, governments began to dismantle these controls. Today, few controls remain and nowhere in the OECD do they play a central role in monetary policy. This chapter reviews the experiences of OECD countries, which has culminated in liberalized capital markets and a general recognition that, This chapter was prepared while the author was Deputy Director, Country Studies and Economic Prospects Branch of the OECD Economics Department. The views expressed are those of the author and do not necessarily reflect those of the OECD, its member countries, or the U.S. Treasury Department. 1 The roots of this literature are to be found in the general literatures on both international finance and development economics. Among the recent works that provide a general overview are Fischer and Reisen (1992), Mathieson and Rojas-Suarez (1992). Many excellent papers consider the issues in the context of Korea and other specific countries, but others are better able to provide a guide to these.
119
120
Jeffrey R. Shafer
while not without problems, an open money and capital market is the best institutional setting for sustained monetary stability. Section 2 of the paper provides an overview of the use of capital controls in OECD countries from 1945 to the present. Section 3 identifies some of the effects that capital controls and their removal have had on financial and economic developments at a macroeconomic level. Section 4 distills from the experiences of OECD countries the costs and benefits of using capital controls, linking these to broader issues of the choice of a framework for macroeconomic policy. Section 5 pulls together some concluding thoughts. 2
An overview of capital controls in the OECD countries: 1945 to the present Capital controls unquestioned: 1945-60
At the end of World War II, the economies and financial systems of most of the world lay in ruin. The United States stood almost alone in having both its productive structure and its institutions intact. An accepted basis for the conduct of international trade and finance, never securely reestablished in the interwar period, simply did not exist. A blueprint for construction of a new international monetary order had been agreed, essentially between the United States and Britain, at the Bretton Woods Conference in 1944. But it would be a decade and a half before a core of countries were conducting their international financial affairs as envisaged in the International Monetary Fund (IMF) Articles of Agreement promulgated there - that is, by maintaining a fixed parity with the U.S. dollar or gold, and not imposing restrictions on foreign exchange transactions involving trade. During the immediate postwar period, freedom of capital movements was hardly an issue. Meeting basic human needs in wartorn lands, replacing worthless currencies, dealing with outstanding debts, reconstructing economies, and resuming foreign trade were more pressing concerns. Foreign exchange transactions associated with trade were closely controlled; international flows of capital through markets were unthinkable. Moreover, the mainstream wisdom that had emerged from the financial chaos of the 1920s and 1930s was that strict controls on international capital movements were essential for financial stability. This wisdom was embedded in the IMF articles, and consequently capital market liberalization was not a part of the effort to reestablish what was considered to be normalcy. The last key steps in this process occurred:2 2
Solomon (1977, 23-27) and Kindleberger (1984, 442).
Controls on international capital movements
121
1 in 1958, when most western European countries restored current account convertibility (i.e., again permitted nonresidents who earned currency in the course of trade to exchange it for another currency at exchange rates within established margins); 2 in 1960, when the European Payments Union for regional clearing of trade transactions was dismantled; 3 in 1961, when western European countries renounced their claims as a transitional measure under IMF Article XIV to maintain restrictions on international payments for current transactions; and 4 in 1964, when Japan established current-account convertibility for the yen. During the reconstruction period that preceded the return to convertibility, events did not call seriously into question the Bretton Woods vision of fixed exchange rates, free international flows of goods, and restricted international flow of capital, although there were dissenting voices. A mafor exchange rate realignment took place in 1949, and occasional devaluations followed (including two by France - one in 1957 and another in 1962). The hot international monetary issue of the 1950s was the adequacy and stability of the supply of international liquidity formulated as the "dollar-shortage problem." (International liquidity continued to be a preoccupation in the 1960s, with concerns about its composition - increasingly U.S. dollars - and, toward the end of the decade, about a glut.) No chronic problem or crisis was ascribed to impediments to the free flow of capital. The normalcy that was sought was one free of what were seen as the three failures of economic policy during the interwar period: (1) failure to achieve durable exchange rate stability; (2) failure to preserve open markets for trade; and (3) failure to manage individual economies so as to avoid inflation or depression. Hence the maintenance of parities came to be seen as an end in itself; although the IMF articles envisaged parity changes in the event of fundamental disequilibrium, in practice this came to be considered a last resort and clear evidence of policy failure. The freeing up of trade was also an end in itself, pursued multilaterally in the General Agreement on Tariffs and Trade (GATT) with sustained leadership by the United States, and regionally in Europe in both the Common Market and the European Free Trade Association. Macroeconomic stability was considered not to be an inherent property of a market economy, and opening up to either trade or capital* flows was widely thought to expose an economy to additional shocks, and hence to added instability, more than it provided a buffer or an anchor in the global economy. On
122
Jeffrey R. Shafer
the other hand, stability was thought to be achievable through the active use of both domestic monetary and fiscal policy. These prevailing views implied two objectives for monetary policy - contributing to domestic stabilization and maintaining the exchange rate - requiring at least two monetary instruments. Capital controls were seen as the second monetary instruments.3 Not only were controls seen as required to allow an independent monetary policy, they were also intended to curb capital flows that were considered destabilizing. Without controls, the fundamental objectives could apparently not be achieved. With them, there was no problem at the level of the individual country. (This view will be assessed in Section 4.) Several developments at the end of the 1950s and beginning of the 1960s suggest that the wisdom of the Bretton Woods dichotomy - freedom for trade, controls for finance - was being questioned and tested. For one thing, Canada had refrained from imposing all-encompassing capital controls and, marching to a different drummer, allowed its exchange rate to float from 1952 until 1962. Germany went further than other countries in the 1957 return to convertibility in extending it to capital account, as well as current account, transactions. And the United States refrained from introducing capital controls during the 1950s (retaining, however, the prohibition on private gold holding introduced in 1933), even though there were times when capital flows were a matter of some domestic policy concern.4 German and U.S. policies were presumably motivated by concern for the distortions in financial markets and the competitive advantages as financial centers that Frankfurt and New York might consequently lose at a time when capital controls posed no particular macroeconomic problems for those countries. As we shall see, both countries subsequently imposed controls (the United States on outflows, Germany on inflows) when balance-of-payments developments impinged on the pursuit of domestic macroeconomic objectives. Liberalization on the agenda: 1961-73 Capital account liberalization came clearly onto the international economic agenda with the establishment of the OECD in 1961. Article 2 of the OECD Convention states in part that "Members agree that they will, 3
4
This line of thinking is clear in a 1964 report of the G-10 deputies on the functioning of the International Monetary System as quoted and discussed in Solomon (1977, 58-59). Indeed, this report and other mainstream writing of the time called for an even wider panoply of instruments. Although the dollar shortage concerns of Europeans in the early 1950s had no domestic U.S. counterpart, capital outflows when interest rates were lowered during the 1958 recession were a first muted signal that U.S. monetary policy could potentially be constrained
Controls on international capital movements
123
both individually and jointly, . . . pursue their efforts to reduce or abolish obstacles to the exchange of goods and services and current payments and maintain the liberalization of capital movements" (emphasis added). The case for removing barriers to capital flows "rested on the desirability of facilitating international trade, allowing direct investment with its associated transfers of skills and technology as a stimulus to growth, and providing enterprises and individuals with greater financing and investment opportunities by giving them access to markets abroad" (OECD 1990, 9). Thus, capital account liberalization was seen as a logical extension of trade liberalization and complementary to it. The United States pressed liberalization in the OECD in the 1960s. This policy was advocated for all countries on economic efficiency grounds, but there was undoubtedly an expectation that U.S. financial institutions and multinational enterprises would reap some of the benefits. U.S. officials also believed that the heavily regulated and closed money and capital markets of Europe and Japan provided governments there the means to manage their balance of payments in ways that weakened the external position of the United States.5 This theme has recurred from time to time to the present day. Progress toward liberalization was slow, despite the push that it received from the establishment of the OECD, partly because of the same protectionist pressures that have slowed trade liberalization, but also because capital controls, at least over short-term capital flows, continued to be seen as essential for the conduct of monetary policy. The limited progress that was made in the 1960s was mainly in the area of long-term capital. Restraints on short-term capital flows were tightened or relaxed as a macroeconomic management tool (OECD 1990, 10). Indeed, the United States introduced controls on capital outflows as a balance-of-payments measure with the interest-equalization tax in 1963. A range of controls was imposed by U.S. authorities on both short- and long-term capital flows involving banks, portfolio investors, and multinationals over the subsequent decade. As in other countries, these controls were tightened and sometimes eased as market pressures built up and subsided. Germany imposed special reserve requirements on capital inflows at various times in the 1960s and 1970s, although it refrained from general controls.6 It became evident in the 1960s that two recognized problems with capital controls were serious and becoming more so. First, capital flows in the form of leads and lags in payments for current transactions could not be well controlled once current account convertibility was established. The
5 6
by external considerations, given the U.S. obligation to maintain convertibility of foreign official holdings of dollars into gold. See, e.g., Solomon's (1977) characterization of the U.S.-European dialogue in the OECD. OECD (1973) recounts three episodes in the 1960s and early 1970s.
124
Jeffrey R. Shafer
problem had been diagnosed for the case of the United Kingdom in the early 1950s (Katz 1953). As Fukao (1990) points out, a two-month shift in average payment terms for exports and imports would have exceeded Japan's foreign exchange reserves by 50 percent in the mid-1960s. At the time, the potential size of leads and lags was used by some to argue that the stock of international liquidity was inadequate. The second problem that became evident with capital controls was that they leaked. Some money crossed borders in suitcases, of course. Much more was undoubtedly moved through arrangements with those having access to foreign exchange through trade. By their extralegal nature, leakages are not recorded in official balance-of-payments statistics, but the behavior of errors and omissions in those accounts at times of pressure on a currency are indicative of their importance. Such flows occur chronically, as well as during crises, thus allowing the gradual buildup of an invisible overhang of foreign assets or liabilities. The returns available outside the net of controls reward those who break the rules at the expense of those who follow them. Perhaps the most spectacular visible evidence of leakages from capital flows was the growth of the Eurodollar market in London during the 1960s. Although this market had its beginnings in the late 1950s as a mechanism for the Moscow Norodny Bank and others to hold dollars without putting them in New York,7 the market grew on the flow of dollars out of the United States, despite capital controls. It also provided a way of doing international business in London at a time when sterling transactions were tightly restricted. The dollars were placed in London because they could earn a higher yield there, since dollars outside the net of U.S. capital controls were less encumbered and hence more valuable than in New York. Many of the depositors were placing money in the market that they had obtained outside their own country's network of capital controls, as well as those of the United States. Indeed, the market was not limited to dollars. Either explicitly, or implicitly through combining a dollar deposit with a foreign exchange swap contract for the same maturity, many currencies were available. The Eurocurrency market at the end of the 1960s was a creation of capital controls, and a testimony to their porousness. Having grown to maturity, the market has lived on to fulfill new roles in a world where capital controls are much less important. The period 1967 to 1973 was one of upheaval in international financial markets. The devaluation of the U.K. pound in 1967 represented a failure of multilateral efforts to maintain fixed exchange rates among the major 7
Kindleberger (1984, 450) gives this and two other legends concerning the birth of the Eurodollar market.
Controls on international capital movements
125
currencies. The devaluation of the French franc and transitional floating of the German mark to a new higher parity in 1969 were two more signals that the international monetary system was flawed. Suspension of the convertibility of the U.S. dollar into gold and the Smithsonian realignment of 1971 failed to restore stability; by 1973 the major currencies were floating and smaller countries were operating under a range of ad hoc arrangements. An effort in the Committee of Twenty to design a new system moved slowly.8 During this period, capital controls were deployed actively in crisis management mode. Thus, before much momentum had built up behind liberalization efforts in the OECD, controls were being reinforced. The mid-1970s saw the most extensive use by OECD countries of controls on both inflows and outflows of capital of any time after the early 1960s. This is visible in charts l a - l h (OECD 1980), which show reservations and derogations entered by governments with respect to the OECD Code of Liberalization of capital movements. Most categories of capital flows covered by the codes show a mid-1970s bulge in the extent of controls. The role of controls reconsidered: 1974-1979 Although strong voices had been raised in favor of floating exchange rates, the major countries adopted this strategy in 1973 under intense market pressure rather than as a deliberately chosen course. The regime was given multilateral legitimacy by the Rambouillet, France, Summit in 1976 and subsequently by the Second Amendment of the IMF Articles of Agreement. The shift to floating exchange rates by some countries reopened the issue of capital controls. Floating exchange rates promised and provided an alternative to capital controls as a way of achieving the scope to conduct an independent monetary policy. A change in domestic liquidity conditions9 would induce a capital inflow under a fixed exchange rate (in the absence of controls), neutralizing the initial action. With a floating exchange rate, by contrast, the central bank could control its own balance sheet. The point is simply that with one instrument, or even a set of instruments all of which operate on domestic liquidity, one could in general achieve one but only one of three possible targets: the domestic money (or credit) stock, an interest rate, or an exchange rate. It was also argued by some that floating exchange rates might make an additional instrument, sterilized intervention, 8 9
Solomon (1977, 343-48) provides a chronology. On the tightening side this could be an open-market sale of securities, an increase of the central bank lending rate, an increase in reserve requirements, or a tightening of bank lending rate limits.
126
Jeffrey R. Shafer
effective even without capital controls. This would provide both monetary independence and some scope to influence the exchange rate. The conventional wisdom from experience, however, is that sterilized intervention is of little or no effect except as a signaling device.10 The two countries for whom a desire to retain monetary independence had motivated the imposition of capital controls on what had been liberal regimes at the beginning of the 1960s moved relatively quickly to remove them once it began to be clear that a fixed exchange rate regime would not be reestablished in the IMF any time soon. The Committee of Twenty shelved its drafts in mid-January 1974 and decided to adopt an evolutionary approach for reform of the international monetary system; before the end of the month the United States had removed all of its controls on outflows of capital, and Germany had relaxed its restrictions on inflows (Solomon 1977, 348). It is difficult to assess the immediate effect of these actions. The dollar did depreciate by about 10 percent, and the mark appreciated over the next several months, but this came after a strong dollar rally and in the immediate confused aftermath of the first oil shock (Solomon 1977, 280). Much larger currency movements than this might have been expected if controls had been exercising the power ascribed to them by their advocates. Except for Switzerland, other OECD countries did not immediately follow the United States and Germany down the road of liberalization; indeed, a number shored up their administrative defenses against flows of funds across their borders in the midst of the turmoil of 1974-75, as indicated by Figure 5.1. The majority continued to maintain an exchange rate commitment of one sort or another. For example, six EEC countries sought to maintain a joint float, and Sweden and Norway associated their currencies with this snake. France was in and out of this arrangement during the mid-1970s (Solomon 1977, 348-49). Japan finally floated the 10
Sterilized intervention, that is, changing the composition of Central Bank assets without changing their aggregate size, would have no more effect than changing the mix of one thousand won and five thousand won notes in circulation in Korea if nonmonetary assets denominated in different currencies were perfect substitutes, as they would virtually be if they were not distinguished by the application of capital controls or the possibility of a change in their relative price. But with floating exchange rates, efficient risk management could imply that the relevant assets were imperfect substitutes, thereby providing a second effective monetary instrument. The collective judgment of the monetary authorities of the major countries in 1984 is given in Working Group and Jurgensen (1983). The conclusion of the report was that sterilized intervention is of little or no lasting effectiveness. This is not actively contested today among international monetary practitioners in OECD countries. Intervention has nevertheless been used frequently. It is seen as a short-run tactical instrument rather than as a means to pursue independent domestic and external monetary policy objectives on a sustained basis.
127 O T H E R C A P I T A L M O V E M E N T O P E f U T I O N S (ITEMS XJ - X V ) V////A
Percentage ot (terns covered by limited reservations Percentage of items covered by fuU reservations Percentage of Kerns covered by general derogations
Per cent tOO r~
. ,_
. ,
74
7 S 7 a i O » ? t 4 S f t
C A P I T A L M O V E M E K T O P E R A T I O N S O F I T E M S I T O IX Per cent 100
Per cent 100
V/////A
Percentage ot Items covered by limited reservations Percentage of items covered by lull reservations Percentage ol Hems covered by general derogations
Capital Inflows
-60 -80 -100
Capital Outflows M i l i l i l i l i l i h l i l t l i l i l > l i l i t i l i l i i , i . i , i .
— a 6 M 7 Q 7 2
Figure 5.1. OECD capital flow restrictions. Reproduced from OECD (1990).
P * cent —1 mo
128 DIRECT INVESTMENT (ITEMS I) ADMtSSJON O F SECURFDES T O CAPfTAL M A R K E T S ( I T E M III)
Percentage of Hems covered by United reservations Percentage of Hems covered by full reservations Percentage of Items covered by general derogations
V/////X
Percent
Parcantaga of items covered by limited reservations
BBtyii^D Peroentao* o( items covered by full reservations E::::::::i:!";!:!l Percentage of items covered by generai derogations
-
-80
-
•100 Fotowiog th« anwndmwit of « M Cod* In 1966totmroduo* tha rt^t of ««tabiishnwnt. a l oounirtM which had previously Ubaraisad Mam VA lodoad a limttad rcaarvation.
Figure 5.1. OECD capital flow restrictions, (com.)
hhl ihl
.1
1,1.1,1,1,1
1,1,11
-to
129 BUILDINQ OA PURCHASE OF REAL ESTATE {ITEMS VI / A1 AND B1> Parcantaga of Hams covarad by Mmttao rasarvations Parcantaga of Hams oovarad by M l rasarvrtons Parcantaga ol Hamc oovarad by ganaral daregations Par cant 100 I
BUYING AND SELLING OF SECURITIES (ITEMS IV AND V)
-
Par oani 100
Capital inflows
Par cant i 100
iff) Peroanttge of items oovarad by limited reservations Percantaga of itams covered by M l reservations Percentage of items covered by general derogations
Capital inflows
-40
Capital Outflow*
•100 Tha buying and sailing oi cotactiva investment securities is included from 1973 onwards.
Figure 5.1. OECD capital flow restrictions, (cont.)
' » ' • ' • •
••'
70
72
74
i I • li ti » i l l
-
-ao
LLLLLJIOO
130 COMMERCIAL CREDITS AND LOAMS [ITEM VIN (I)]
FINANCIAL CREDITS AND LOANS [ITEMS VDI (10 AND IX] f///'/77i
Partanttga of tern* oovarad by Hmtod ratarvafions
:|:j:>|:::| Parcantaga of ilaots oovarad by ganaraj darogabons
Parcantaga ot Hams oovtrad by g*Mf*J dtroesfont Pafcant
Par cent
ParcanUgaofHamsoovaradbyGmnadratarvations Paroantaga of Itamt covarad by M l rasarvations
PwwnttQV of Hvitw OOVMVO by luw fM#o^SBOAS
100
100
Par cant 10
°
60
Captal Inflows
-
•40
Capital O u M o w t
•100
li l i l i
I . I . I . I , t i t , I ,
l i l i l . l i l . l i
1 . 1 , 1 ,
UJJ
.100
Figure 5.1. OECD capital flow restrictions, (cont.)
-100
[ i I 11 11 i I 11 11 11 i I 11 11 11 i h I 11 11 i I i I 11 i I i I 11 11 i I 111 .100
Controls on international capital movements
131
yen but "de facto, the management of the market resembled very closely that of the fixed exchange rate period" (Fukao 1990, 27). The ambition to achieve an exchange rate objective while maintaining monetary independence continued to provide a rationale for keeping capital controls in these countries. Two other related considerations weighed more heavily in countries other than the United States and Germany, and undoubtedly influenced decisions to retain, and even reinforce, capital controls. First, although allowing the exchange rate to float did provide the expected monetary policy autonomy, it did not provide "independence" in the sense of insulating countries from developments abroad, as some theoretical models had predicted (notably monetary models embodying the assumption of continuous purchasing power parity). Events abroad most certainly were felt in economies with floating exchange rates; the oil shock and subsequent global recession made this abundantly clear. Capital controls seemed to offer a means of at least mitigating the effects of some of these shocks. Second, there was widespread skepticism among practitioners in central banks and treasuries in the view held increasingly widely in academic circles that unrestricted capital flows would be stabilizing rather than destabilizing. Some evidence began to emerge early on that floating exchange rates did not pass statistical tests for speculative efficiency (e.g., Dooley and Shafer 1976). Such evidence has continued to accumulate, but it is fair to say that there is no consensus as to what it means - a number of assumptions, in addition to speculative efficiency, lie behind the interpretation of the statistical tests. The argument that capital flows are too often capricious, made even when strong capital movements obviously reflect strong incentives or expectations created by policy, continues to be a rearguard defense against liberalization of capital flows. The countries that led the way toward liberalization conducted autonomous monetary policies with floating exchange rates, and they clearly attached less importance to insulating the economy from external disturbances than others. For the United States, at least, this is understandable because the large size of the domestic economy and financial market means that external factors are relatively less important. All three countries that reduced or eliminated controls in the mid-1970s - the United States, Germany, and Switzerland - faced clearer evidence than others of the relative ineffectiveness of controls in moderating capital flows, and they felt particularly acutely some of the costs of using them for this purpose. The dollar was the principal international currency, and the mark and Swiss franc were leading alternatives. The active markets for these currencies in London (where sterling could not be used for international financial transactions) and other international banking centers (Luxem-
132
Jeffrey R. Shafer
bourg was particularly important for marks) showed the severe limits on the control that national authorities could exercise over the use of their currencies. In addition, the growth of offshore markets was coming at the expense of the financial industries in New York, Frankfurt, and Zurich. The cost could be measured in terms of jobs and profits. As we shall see, concern not to disadvantage nationals seeking to compete in the emerging global financial market was an important reason behind liberalization in other OECD countries from 1980 to the present. A tide of liberalization: 1980-92 Economic and financial developments in the 1970s led OECD governments to reconsider almost all aspects of economic policy. The decade had seen inflation ratchet up and government indebtedness mount in many OECD countries as a result of policies adopted to sustain the growth of economies. Growth nevertheless faltered. Exchange rate fluctuations were large relative to previous postwar experience, whether a currency was floating or was in and out of various fixed rate arrangements. Although the oil price shocks of 1974 and 1979 provided an excuse for poor performance, there was a growing feeling that a new approach was needed. New governments came to power promising new approaches. Two pillars of the strategy that emerged, and was enunciated in OECD ministerial and G-7 summit communiques from 1980 onward, cast a new light on capital controls. First, the dismantling of controls and regulations that impeded or distorted private initiative in all spheres of economic activity come to be seen as necessary to restore the dynamism and adaptability of OECD countries. Financial market deregulation, both domestically and at the border, was at the core of this agenda.11 Once the process began, the financial institution competitiveness considerations mentioned earlier provided added reason for others to follow. Second, the basic way of thinking about macroeconomic policy underwent change. The prevailing approach of the 1950s, 1960s, and 1970s could fairly be described as focused on the short term. The objectives were to boost output growth and employment, while keeping inflation from accelerating and reserve loss from becoming a constraint. The instruments were the budget deficit, interest rates, and controls on anything that became bothersome (not only international capital flows, but bank credit, bank lending and deposit rates, wages, prices, imports, the labor force of a firm, and at one time or another almost every other aspect of economic activ11
OECD (1987) provides an assessment of the need for microeconomic reforms across the board, including in financial markets.
Controls on international capital movements
133
ity). Both objectives and instruments were, or operated on, "flow variables." Little attention was paid to how key stock variables in the economy - government indebtedness, the capital stock, foreign indebtedness, productivity - might be evolving in ways that would prove unsustainable or otherwise mean costs in the future from meeting immediate objectives. The new strategy replaced this short-term orientation with a mediumterm one. While the formulation of a medium-term strategy varied from country to country, two elements formed a common core. The first was a medium-term budgetary plan that would put government debt on a nonexplosive path. (This was honored in the breech by the U.S. president and Congress, which allowed large federal government deficits year after year to make a mockery of medium-term plans to reduce them. But many other countries put their public business on a more stable course in the difficult conditions of the first half of the 1980s.) The second core element of a medium-term strategy was a monetary policy that did not accommodate inflationary shocks. Monetary authorities sought to build credibility with financial markets by establishing a record of following an announced course, and by achieving a good inflation record. This reflected the conclusion drawn from the experience of OECD countries in the 1970s that accommodating monetary policy had only a temporary effect on output and employment, at a cost of permanently higher inflation. That a general trend toward liberalization of economic activity would favor the removal of controls of international capital movements is obvious. That the new macroeconomic policy strategy would also do so is less obvious, but controls are no less anomalous in a strategy that focuses on following a sustainable course and building credibility with market participants. The argument will be taken up again in Section 4, but a sketch is needed for this chronology. The two basic points are intuitively easy to grasp even if often overlooked. One is that controls in a macroeconomic context have the objective of making otherwise inconsistent short-run objectives achievable. This almost always involves allowing an imbalance to build up (e.g., a fundamentally misaligned exchange rate), which will ultimately necessitate brutal adjustment when it becomes impossible to sustain it. The second point is that it is fundamentally contradictory to seek to build credibility with markets and at the same time to prevent market participants from giving voice to their assessments by freely managing their financial affairs. It has also proved impractical to do so. The very existence of controls or maintenance of standby authority to impose them administratively creates uncertainty and hence introduces an additional element of volatility into exchange rate expectations and consequently the rates themselves. It also promotes wariness in markets, which tends to
134
Jeffrey R. Shafer
reduce the demand for a currency that is subject to controls on outflows. Consequently, currencies have often strengthened when such controls have been removed. The adherents to the new economic strategy in the OECD included both countries that considered a floating exchange rate as an essential element and countries that saw a commitment to a fixed exchange rate as a cornerstone of their strategies. Among the former were the United States, Japan, Canada, Switzerland, Australia, and eventually New Zealand. Among the latter were the members of the European Monetary System (EMS) other than Germany, the Nordic countries, and Austria. (Germany was a full participant in the EMS, but the Bundesbank was expected to set a standard within that arrangement, not to adapt to monetary conditions set by others.) The United Kingdom experimented with floating, shadowing the mark and joining the exchange rate mechanism (ERM) of the EMS. It has recently withdrawn from the ERM, not yet having found an exchange rate regime that is right for it over the longer term. The choice for each country has been essentially one of weighing the costs and benefits of making one's own monetary policy, and accepting the consequent disturbances in the external sector when it led domestic monetary conditions to differ from those abroad, or to fix the exchange rate and allow foreign monetary disturbance to be felt domestically. The balance of costs and benefits depends on the strength of domestic institutions, the importance of trade, and the reliability of a foreign anchor. It is not surprising that this calculus has come out differently for different governments. The calculus for liberalization, which is taken up in Section 4, appeared to come out more on the side of benefits as the 1980s unfolded: Capital account liberalization emerged as a cornerstone of both variants of the strategy. As we have seen, floating seemed to lead naturally to dismantling capital controls. Indeed, Canada, which had the longest experience with floating, had never introduced all-encompassing controls. It seemed only logical, if markets were to set exchange rates, that they do so without artificial impediments. It is true that Japan maintained controls in the 1970s when it attempted to manage closely its float. But this led to the same sort of unsustainable situation in 1977 that characterized fixed exchange rates, and the yen had to be allowed to appreciate by one-third against the U.S. dollar in a little more than one year. Over time, both management of the yen and capital controls became less tight. Countries that fixed exchange rates no longer saw this as an end in itself, but rather as a means to achieve better inflation performance with less disturbance to trade than otherwise.12 Hence financial markets had an 12
The trade facilitation motive was explicit in the efforts of President Giscard d'Estaing of France and Chancellor Schmidt of Germany to "create a domain of exchange rate stability
Controls on international capital movements
135
important role in transmitting to domestic markets the monetary conditions set in a low-inflation anchor country. Controls would only interfere with this role. Moreover, if they allowed a country's currency to become misaligned by a large margin, the commitment would lose credibility. Short-term fixes would undermine the medium-term strategy. Hence, in the late 1980s, capital account liberalization and a hardening of exchange rate commitments within, and on the perimeter, of the EMS went hand in hand with many European governments' strategies to restore price stability and public confidence in it. It is difficult to gauge the relative importance of the various factors mentioned here in the liberalization of capital flows that swept through OECD countries in the 1980s. The extent of the change is even greater than suggested by Figure 5.1, since that does not take account of the severity of restrictions. Only a few OECD countries now have capital controls of any significance for monetary policy (restrictions on direct investment, imposed for other reasons, are more significant although they, too, have been substantially reduced). Indeed, before September 1992, when Spain, Ireland, and Portugal reimposed controls in the face of speculation that developed over the prospects for European Monetary Union, only Greece, Iceland, and Turkey had controls worth noting in this context. Even these countries have taken major steps toward an open capital account. The tide of liberalization began with the United Kingdom in 1979, when the new Thatcher government summarily removed a system of controls that had been operated and had been elaborated throughout the postwar period. With a floating exchange rate, as well, the intention was to conduct an independent monetary policy using market-based instruments. A year later, Japan adopted a new foreign-exchange law that allowed all international financial transactions unless specifically restricted, replacing the old law that prohibited all foreign transactions unless specifically authorized. Although considerable liberalization had occurred over the years under the old law and many restrictions were retained under the new law, this legislation marked a critical turning point in policy and practice (Fukao 1990, 3). Subsequently, other countries have liberalized, usually step by step, but often once the process has been begun, the implementation has been acin Europe" through the EMS. Whether exchange rate volatility has an important effect on the volume of trade remains a topic of empirical research. The inflation stabilization motive took on increasing prominence within the EMS over the course of 1989, as participating countries used their link to the German mark as a way of anchoring monetary policy in a commitment to low inflation. This required them to forgo frequent recourse to realignment, which had characterized the EMS in its early days and distinguished it from the Bretton Woods system in practice.
136
Jeffrey R. Shafer
celerated as remaining controls have come to seem anomalous. Australia and New Zealand dismantled most controls in 1983 and 1984. The Netherlands was effectively fully liberalized by 1986. Denmark, France, and Finland moved gradually, with the first two completing the process before the end of the 1980s. Sweden and Norway began to move later, but liberalized rapidly in 1989 and 1990. Other EC countries and Austria also progressively liberalized in the late 1980s, and had complied with an EC directive for removal of capital flows before the reimposition of controls by a few countries in September 1992. Indeed, the decision in 1988 by EC countries to phase out capital controls as a part of the Unified Market Program was an important goad to action, not only in member countries of the EC, but also in other European countries that were negotiating the agreement on a European Economic Area.13 The Unified Market became a second process of multilateral peer pressure to reinforce the momentum that had been building for liberalization in the OECD throughout the 1980s. 3
The effects of capital controls and their removal
Although much used until the 1980s, and once widely thought to be an indispensable tool of macroeconomic management, in retrospect capital controls could be considered to have been of great macroeconomic importance on a sustained basis in only a few OECD countries. Whether the capital flows that they impeded in the short run would have been stabilizing or destabilizing is not clear. What is clear is that capital controls on particular occasions allowed governments to defer policy adjustment in times of pressure, but on many of these occasions they were ultimately forced to adjust. In addition to their macroeconomic role, controls were an important factor shaping the development of domestic financial markets - at times providing trade protection for banks that were subjected to costs from domestic regulation, and at other times impairing the international competitiveness of the financial sector. This section reviews some of the evidence for these propositions. Macroeconomic effects It is not possible to estimate with any reliability the effect of capital controls on net capital flows (in the case of a fixed exchange rate) or on the exchange rate (in the case of a floating rate).14 The problem is that suffi13 14
Olsen (1990) makes this clear for the case of Norway. One obvious prediction concerning gross flows is generally borne out, however. Removal of broad controls leads to an internationalization of financial activity, with both claims on and liabilities to foreigners rising substantially. Sweden provides an example. According to Lindeneus (1990), by the end of 1989 when the last capital controls for monetary policy
Controls on international capital movements
137
ciently reliable empirical models of capital flows or of exchange rate determination have not been developed. Without such models, there is no way of knowing what flows would have been in the absence of capital controls. It is sometimes possible to obtain at least rough estimates of the effects of controls on particular gross flows (e.g., domestic purchases of foreign equities) by comparing observed flows with those that have occurred in other countries in the absence of controls, or in the same country at a time when controls were not in effect. In the absence of effective, allencompassing controls, however, the main effect of restrictions of one channel will be to direct flows into other channels. The net effect cannot be inferred from the direct effects, even when the latter can be observed clearly. Although the effects of capital controls on financial quantities of macroeconomic importance are unmeasurable as a practical matter, their effects on financial prices can be observed. In particular, it is possible to observe how much domestic interest rates differ from rates on equivalent instruments outside the reach of capital controls (offshore rates). The argument is one of arbitrage: If, for example, a higher yield is available on a London Eurodollar deposit than on a New York deposit, it must be because those holding deposits in New York are impeded from shifting their balances to London. Other factors, in addition to what come to mind as capital controls, could play a role - for example, taxes or the soundness of banks at home versus an offshore center. The data to follow show at most small differentials in the absence of explicit capital controls, which might be attributable to these factors. This should not be surprising since potential arbitrageurs in a typical OECD country include banks operating in both domestic and offshore markets. It is more important to note that a differential reflects not only the direct effects of capital controls in place, but also the effects of expectations that controls might be tightened or eased in the future, or that controls not currently binding might become a constraint in the future.15 Figure 5.2 shows the differentials between domestic and offshore three-
15
purposes were removed, Swedish enterprises had foreign currency liabilities equivalent to 340 billion krona, up from only 20 billion krona in the early 1980s, and equal to 30 percent of their total liabilities. One-third of this stock was built up in 1989 alone. It was mostly offset by Swedish placements abroad, however, as the net capital inflow was only 43 billion krona in that year. Dooley and Isard (1980) develop a model to split the differential between the domestic German interest rate and the Euromark rate from 1970 to 1974 into two components: "a political risk component," which they define as the probability of controls being imposed in the future, and "a capital controls component," which reflects the direct effects of capital controls. Both are taken to reflect the effects of capital controls in a general sense in the present chapter.
UNITED STATES
:1 73:1 75:1 77:1 79:1 81:1 83:1 85:1 87:1 89:i 91:1 72:1 74:1 76:1 78:1 80:1 82:1 84:1 86:1 88:1 90:1 92:1 Domestic Rate -
Offshore Rate
JAPAN
GERMANY
i;iT3iiY5ii77ii79ii8i;r83;v85;i87!i89;i9i;i 72:1 74:1 76:1 78:1 80:1 82:1 84:1 86:1 88:1 90:1 92:1 Domestic Rate -
Offshore Rate
UNITED KINGDOM
15 10-1
*71:1 73:1 75:1 77:1 79:1 81.1 83:1 85:1 87:1 89:1 91:1 72:1 74:1 76:1 78:1 80:1 82:1 84:1 86:1 88:1 90:1 92:1 Domestic Rate -
• Offshore Rate
71:1 "'73:i 1 ' 1 75:i 1 "77:i" I 79:i" T 8i:i" 83:1 85:1 87:1 '89:1 ""91:i" 72:1 74:1 76:1 78:1 80:1 82:1 84:1 86:1 88:1 90:1 92:1 Domestic Rate -
Offshore Rate
Figure 5.2. Offshore and domestic interest rates. Sources: Federal Reserve, OECD (U.S. domestic rate), and Mitsuhiro Fukao (yen rates).
CANADA
FRANCE 2O 18 1614 12 ia 8 6 4 1:1 73:1 75:1 77:1 79:1 81:1 83:1 85:1 87:1 89:1 91:1 72:1 74:1 76:1 78:1 80:1 82:1 84:1 86:1 88:1 90:1 92:1 Domestic Rate -
- Offshore Rate
ITALY
71:1 73:1 75:1 77:1 79:1 81:1 83:1 85:1 87:1 89:1 91:1 72:1 74:1 76:1 78:1 80:1 82:1 84:1 86:1 88:1 90:1 92:1 Domestic Rate -
• Offshore Rate
NETHERLANDS
3fr 25j 2015 10 :1 73:1 75:1 77:1 79:1 81:1 83:1 85:1 87:1 89:1 91:1 72:1 74:1 76:1 78:1 80:1 82:1 84:1 86:1 88:1 90:1 92:1 • Domestic Rate -
Offshore Rate
Figure 5.2. Offshore and domestic interest rates (cont.)
71:1 73:1 75:1 77:1 79:1 81:1 83:1 85:1 87:1 '89:i" "91:1" 72:1 74:1 76:1 78:1 80:1 82:1 84:1 86:1 88:1 90:1 92:1 • Domestic Rate -
- Offshore Rate
SWITZERLAND
1:1 73:1 75:1 77:1 79:1 81:1 83:1 85:1 87:1 89:1 91:1 72:1 74:1 76:1 78:1 80:1 82:1 84:1 86:1 88:1 90:1 92:1 Domestic Rate -
• Offshore Rate
BELGIUM
SWEDEN
72:1 74:1 76:1
78:1 80:1
Domestic Rate
82:1 84:1
86:1
88:1 90:1
92:1
Offshore Rate
Figure 5.2. Offshore and domestic interest rates (cont.)
i 72:1
74:1 76:1
78:1 80:1 82:1 84:1
Domestic Rate
86:1
Offshore Rate
88:1 90:1
92:1
Controls on international capital movements
141 16
month money-market interest rates for the G-10 countries. These charts show that from the time capital controls were abandoned by each country, the domestic and offshore rates have moved very closely together - in some countries they have been indistinguishable. This is strong evidence that restrictions on capital flows are indeed no longer a major factor allowing domestic interest rates to deviate from those in international markets, although there are factors that have sustained some relatively small differentials.17 By contrast, when capital controls were in effect in the 1970s they sometimes allowed large differentials to emerge, but sustained differentials were generally 200 basis points or much less. In some countries, offshore and domestic interest rates diverged surprisingly little given the extensive system of capital controls that were in effect. It cannot be inferred from this that the controls failed; it may have been that the authorities sought monetary conditions that were little different than might have prevailed without controls. A summary impression of the differential path for each country might be outlined as follows. United States Capital controls appear to have done little to insulate the domestic and offshore markets in the 1971-73 period: The two rates moved closely together. A margin between domestic and offshore dollar rates persisted after the removal of general capital controls in early 1974, reflecting the continuation of Eurodollar reserve requirements. Indeed, the differential widened in the late 1970s and early 1980s as high interest rates raised the implicit tax in reserve requirements. The differential eroded over the 1980s as interest rates came down and as Eurodollar reserve requirements on banks were first reduced and subsequently set to zero. Japan During most of the 1970s, capital controls allowed for somewhat tighter domestic monetary conditions than prevailed on the internal yen market, but in 1974 following the first oil shock, they insulated domestic markets somewhat from the effects of a loss of confidence in the yen, which pushed offshore rates to very high levels. 16
The offshore rates are, in fact, constructed from data sets containing Eurodollar interest rates and the three-month forward discount of the currency against the dollar. The formula is: D is the ninety-day forward discount on the currency in percent R'i - R'%{\ + D/4)
17
where R'i is the offshore interest rate on the /th currency. It may also be that domestic markets are not completely integrated so that the moneymarket rates may not be a relevant indication of risk-free investment returns or borrowing costs available throughout an economy.
142
Jeffrey R. Shafer
Germany The effects of measures to curb capital inflows while pursuing tight monetary policy are significant over the 1971-74 period; the effects of measures taken in the late 1970s were much smaller. The United Kingdom Capital controls were used consistently to maintain easier monetary conditions than otherwise possible until they were removed in 1979, but the size of the differential was on a downward trend over the 1970s. France
Until the commitment to liberalize in the late 1980s, capital controls were used by the authorities to maintain easier domestic monetary conditions than prevailed outside their reach. In the early 1980s, when the franc parity was fixed in the narrow bond of the EMS but continued to be frequently adjusted, controls provided considerable insulation during episodes of upward movement in offshore rates, driven by expectations of a devaluation. Italy Controls were used to maintain easy monetary conditions and offshore rates were high and volatile, reflecting chronically shaky confidence in the currency. Spikes in the offshore rate, reflecting expectations of a devaluation, were typically followed by devaluation. From 1983 until the recent removal of capital controls, domestic interest rates were generally kept above offshore rates, which were progressively less affected by bouts of speculation on a devaluation within the EMS. Domestic considerations inflation and public sector deficits - kept monetary conditions tight. When speculative pressure built up on the lira recently, domestic and offshore rates rose in tandem without capital controls to provide insulation for the domestic market. Canada Small differentials reflect the generally liberal policy toward capital flows. Netherlands Although controls over inflows and outflows were on the books until relatively recently, they have not been a significant factor over the period from the mid-1970s.
Controls on international capital movements
143
Switzerland Consistent with the limited recourse to capital controls, the offshore rate tracks the domestic rate closely. Sweden Controls were used to maintain the domestic interest rate, about 2-3 percent below uncontrolled rates throughout the late 1970s and early 1980s. After the devaluation of 1982, the differential never reopened although capital controls were retained for a number of years. Belgium Although the authorities operated a dual exchange rate regime until 1990, in only two periods (in the mid-1970s when domestic rates were pushed up and in the early 1980s when they were kept down as offshore rates rose) did an important differential open up between domestic and offshore rates. One impression given by Figure 5.2 is that, where capital controls were important, domestic interest rates were less volatile than offshore rates. It is tempting to conclude that capital controls provided for greater stability in monetary conditions, and to speculate that macroeconomic conditions, more generally, might have been stabler as a result of their use. Such conclusions are unwarranted, however. Both the offshore rates and domestic rates are affected by the existence of controls and by the policies that were pursued in conjunction with them. In particular, offshore rates, relatively insulated from domestic rates, reflect mainly expectations of exchange rate change. When policies that made an exchange rate increasingly unsustainable were pursued with the support of capital controls, the offshore interest rate would be pushed up. The higher interest rate was needed to compensate holders of the currency for the expected value of the decline in its value (and perhaps a pure risk factor as well). Thus the volatility of offshore rates cannot be taken as an indicator of how volatile domestic rates would have been in the absence of controls. Indeed, even a conclusion that interest rates would have been higher, on average, does not follow from an observation that an offshore rate was higher than the domestic rate, on average. One can make a stronger argument from these charts that domestic interest rate swings have generally been less in the absence of capital controls, even though they have been damped relative to movements in offshore rates when controls were in place. There does not seem to be a systematic trend in the magnitude of interest rate swings in countries like
144
Jeffrey R. Shafer
Germany, Canada, Switzerland, and the Netherlands, where capital controls were never seen as a regular tool of macroeconomic policy. By contrast, domestic interest rate swings in Sweden, Belgium, Japan, the United Kingdom, Italy, and France appear larger in the earlier period when controls were in force than they have been more recently, in some cases by a substantial margin. This suggests that the insulation from external financial disturbances and the autonomy provided by capital controls to set monetary conditions in the light of short-run domestic objectives did not result in stabler macroeconomic conditions. Market-determined capital flows would appear to have contributed to stability, on average, either directly or indirectly by giving monetary authorities useful signals concerning the need to make timely policy adjustments, with the result that swings in interest rates have been less pronounced. Not only is there a general tendency for domestic interest rates to be stabler in the absence of capital controls, but there is no general tendency for volatility to be high in a transitional period. It is reasonable to assume that this is largely because authorities have first relieved the pressure on capital controls and then chosen a tranquil period for their removal so as to minimize disruption during the transition. The United Kingdom may be an exception in this respect; controls were removed at one fell swoop during a period of rising interest rates worldwide shortly after the formation of the Thatcher government, which was committed to deregulation in all spheres of the economy. The amplitude of U.K. interest rate movements was not reduced in the two or three years after 1979, but rates peaked almost immediately, and the trend was subsequently down. The conclusion that seems warranted on the basis of OECD country experience is that, with reasonable care as to timing, capital controls can be removed without great disruption of domestic financial market conditions, whether the exchange rate is floating or fixed. This conclusion may not carry over to the case of a country that has far from completed a macroeconomic stabilization process. Of the countries shown in Figure 5.2, only the United Kingdom could be argued to illustrate a problem that has become an issue in the literature on capital accovnt liberalization in developing countries.18 The problem can arise in two forms, depending on the policy regime. One form, with a floating exchange rate, is that tight policy to reduce inflation can drive the exchange rate up, thereby concentrating the squeeze on the tradables sector. The second form is that with a credible fixed exchange rate policy, capital will flow back in, leading to monetary expansion and considerable infla18
Fischer and Reisen (1992) and Williamson (1992) are particularly concerned about the problems that may arise if liberalization precedes stabilization.
Controls on international capital movements
145
tion before the policy first bites strongly on the tradables sector, which has meanwhile suffered a loss of competitiveness. Among OECD countries not shown in Figure 5.2, New Zealand exhibits the problem in the first form and Spain in the second. A squeeze that falls disproportionately on the tradables sector seems unavoidable without retaining capital controls, but this overlooks the possibilities for raising national saving and shifting more of the squeeze onto the nontradables sector by improving the government's financial balance. Imposing capital controls to relieve the pressure of capital inflows at best provides an alternative means of balancing the squeeze across sectors. At worst, it makes it easier for a government to put off necessary fiscal consolidation. Experience of OECD countries suggests that there is no painless way of curing inflation. If Figure 5.2 extended to September 1992, after the data for this paper had been assembled, another round of pressure on fixed exchange rates in Europe would be reflected in the interest rates. Without capital controls to insulate domestic markets, several monetary authorities have had to choose between allowing domestic interest rates to rise sufficiently to compensate for the risk that parity might not be maintained. Of the countries shown in Figure 5.2, the United Kingdom and Italy chose to abandon their parities rather than keep domestic interest rates high enough to maintain a fixed rate. The French franc was subjected to intense pressure in September, but this subsided once the resolve of the authorities was demonstrated, given a widely held view in the market that the position of the franc was fundamentally sustainable. In Sweden, pressure on the krona built up in late 1991 and was reflected in an interest rate increase shown in Figure 5.2. The krona came under pressure again in September, and this has not yet entirely subsided. Market doubts are being translated into tighter monetary conditions. The Swedish authorities consider this to be the principal means of responding to these doubts. Reintroduction of controls might provide some insulation of domestic interest rates, but it would undermine confidence in the krona over the medium term. Indeed, it might result in larger sales of krona by foreigners so that, even with a differential between domestic and offshore krona interest rates, domestic rates could be forced higher than otherwise in the effort to maintain the parity of the krona. This underscores an element of irreversibility in a decision to remove capital flows. Once foreigners have come to include a currency in their portfolios and domestic residents have diversified abroad, the reimposition of controls is likely to affect strongly decisions on how to place funds that are already beyond the reach of the authorities. The effect is to undermine the achievement of the objectives of controls. Spain, Portugal, and Ireland, not included in Figure 5.2, reintroduced
146
Jeffrey R. Shafer
some controls in September 1992 in the midst of the turmoil in the EMS. It does not appear to be the intention of these governments to retain controls as a permanent feature of policy, indeed this would run counter to policy agreed in the European Community to which they belong. It is not clear what, if any, relief from pressure controls have allowed. International investors are reexamining their strategies with respect to placements in the peseta, the most important of the three currencies internationally. The likelihood that peseta interest rates will be higher over the longer term will increase the longer controls remain in place. It is natural to attempt to draw lessons from this recent episode concerning the soundness of a policy of fixed exchange rates and free capital movements. The episode shows even more strongly than earlier episodes of exchange market instability that capital flows can develop great intensity very quickly. Some monetary authorities chose to let their currency float in the face of this pressure. But others (France and Sweden) did not at this time, and they were able to resist market pressures by raising interest rates without resort to capital controls. The episode also shows that a sustained policy that has built credibility in an unchanged parity can protect a country from speculative attack. Belgium and the Netherlands experienced no significant pressure on domestic interest rates. The commitment of the authorities in these countries to an unchanged parity with the mark is longer-standing than that of other European countries, and the public in these countries appeared strongly supportive of closer European integration while the verdict of the French public hung in the balance in the summer of 1992. Although not subject to speculative pressures on their interest rates, these countries must accept fluctuations in their interest rates in line with those in the German mark, the de facto pivotal currency in the EMS. This has been a problem in the period since German unification, as the mark anchor has slipped and the German monetary policy required to remaster inflation has been tighter than required in the other countries. This is a cost of an unconditional exchange rate commitment, which must be weighed against the stability that it provides on other occasions, including in September 1992, when less credible commitments were tested.
Structural considerations The focus of this chapter is on the use of capital controls as a macroeconomic management tool. The picture would be incomplete, however, without a brief review of related structural issues in the financial sector. Controls on direct investment flows have also been imposed for industrial
Controls on international capital movements
147
policy, national identity, or security reasons. These issues are not considered here. A financial system that is highly distorted by regulation or by taxes and subsidies on particular financial activities, whether explicit or implicit, is unviable in the absence of capital controls. Financial intermediation will move abroad. Hence, liberalization of capital flows has either gone along with, or quickly forced, a rationalization of domestic financial sector regulation. Indeed, this is seen by many as one of the advantages of external liberalization. But the essential point is to recognize that one cannot liberalize externally and maintain large distortions domestically. Two typical problems make the point. First, policies that seek to maintain nonmarket interest rates on credits or deposits are undermined by international arbitrage. For example, when capital controls are liberalized, opportunities are created for those with access to subsidized credits to borrow and place money abroad. For this reason, OECD countries that had conducted monetary policy partly by setting limits on overall credit extended by banks (e.g., France and the Netherlands) and rationing it at interest rates below market-clearing levels, shifted to more market-oriented means of implementing monetary policy before removing controls entirely. Similarly, deposit interest rate ceilings for below market-clearing rates have not been sustainable without controls. For example, the U.S. Regulation Q interest rate ceilings were a factor shifting banking activities to the Eurodollar market in the 1970s after controls over international flows were removed. Second, taxes on financial intermediation can cause financial activity to move offshore unless backed up by controls. The U.S. banking system suffered from a shift of deposits to the Eurodollar market, in part because domestic deposits were subject to non-interest-paying reserve requirements. Similar shifts away from reservable deposits to domestic alternatives also weakened the effectiveness of domestic monetary instruments, although dollar liquidity was never beyond the control of the Federal Reserve. In an effort to level the playing field somewhat following the removal of capital controls in 1974, reserve requirements on funds raised abroad by domestic banking offices remained subject for many years to so-called Eurodollar reserve requirements under Regulation M of the Federal Reserve System. The result did not entirely remove the competitive disadvantage of domestic banks, however, and they were also subject to competition from domestic nonbank intermediation channels. In the end, the reserve requirements on large time deposits and the Eurodollar reserve requirements were both reduced to zero in 1991. As noted earlier, financial competitiveness considerations were a positive reason for liberalizing capital flows in a number of OECD countries.
148
Jeffrey R. Shafer
Although protection of domestic banks burdened by reserve requirements was a motive for retaining some aspects of controls after they were dispensed with as a macroeconomic policy tool, financial competitiveness considerations weighed more heavily on the side of freeing up capital controls. Again, using the United States as an example, domestic offices of U.S. banks lost international banking business to foreign offices and banks during the period of voluntary foreign credit restraint in the late 1960s and early 1970s. Hence the banking industry was a source of pressure to remove these controls. The competitiveness of the banking sector was also a consideration behind German and Swiss reluctance to employ controls. The United Kingdom sought, with considerable success, to maintain its position as an international financial center by creating a favorable deregulated environment for the conduct of international banking activities that were insulated from the highly regulated sterling market. This was an exception, however. Without the history and infrastructure of the City of London behind them, most other countries found themselves disadvantaged by capital controls in the competition to provide financial services. During the 1980s, this became a critical factor in decisions to liberalize capital flows and to deregulate domestic finance. The results of domestic financial deregulation must, at this stage, be judged as mixed. The range of services available at competitive prices has broadened in financial markets. However, prudential problems have emerged in a number of markets that were deregulated in the 1980s. While factors other than deregulation have been at work, these problems have prompted a review of the adequacy of prudential controls and of distorted incentives in financial markets arising from explicit or implicit government guarantees. There is no trend, however, toward reimposition of direct controls. In the absence of this, financial competitiveness considerations are likely to continue to weigh in favor of liberalized capital flows in the OECD countries. In this environment, prudential concerns and the desire of each country not to disadvantage its own financial institutions in international competition are likely to bring supervisory authorities together in multilateral efforts to harmonize some elements of prudential policy. The Basel capital ratios for banks are an example of this. 4
Pros and cons of the use of capital controls in macroeconomic policy
The question of whether capital controls are a useful tool in the conduct of macroeconomic policy is answered quickly by two groups of people. One group is typified by those at Bretton Woods: doubtful of the selfequilibrating capacity of the real economy, committed to stable exchange
Controls on international capital movements
149
rates, fearful of the disequilibrating potential of capital flows, and confident in the capacity of officials to recognize what is needed to optimize economic performance and to act on it. For this group, capital controls are a valuable tool of macroeconomic management. While not strictly necessary to achieve both a domestic demand and an exchange rate target, given a fiscal as well as a monetary instrument, capital controls provide flexibility and insulation from one potential source of shocks that reduce the burden on the other instruments. The second group is typified by the 1970s monetarists: confident in the self-equilibrating capacity of the real economy, advocates of floating exchange rates, believers in the equilibrating potential of capital flows, concerned that markets be free to play their allocative role efficiently, and mistrustful of the capability of officials to optimize economic performance and of their motivation to do so. For this group, capital controls are not only unnecessary to achieve optimal economic performance, their use is positively damaging. For a third group that would not be wholly comfortable in either of these camps, the issue is one of weighing pros and cons. The balances struck by OECD governments, as inferred from their use of controls, have nearly all changed radically over the past twenty years. Behind this there seem to be changes in the evaluation of several considerations that have been reviewed: threats to economic stability in the short-term, medium-term, and long-term; the role of the exchange rate in macroeconomic policy; and the role of financial markets in promoting an efficient allocation of resources. The debate over whether capital flows are stabilizing or destabilizing in the short-run has not been resolved. As noted earlier, there are features of the behavior of floating exchange rates that give ample room for doubt that speculation is always stabilizing in the sense of being driven by rational expectations. On the other hand, exchange rate movements have never left newspaper columnists without plausible ex-post explanations. On balance, floating exchange rates have been sufficiently understandable in terms of fundamentals for governments to focus on these when exchange rate movements have threatened to pose problems. And while some fixed exchange rate arrangements have been the object of fierce speculative attacks (which may be rational), others do not draw speculative attention and are maintained without the support of controls. One theoretical proposition has been borne out by experience: It is not possible to maintain a stable course for interest rates and conduct an exchange rate policy that
150
Jeffrey R. Shafer
gives rise to expectations of discreet changes - one cannot adjust a fixed rate by sizable amounts on a regular basis. Taking a medium-term perspective of, say, one to three years, the experience of the postwar period has by and large eroded confidence that policies can keep demand on a stable course continuously, while bolstering confidence that it will nevertheless come back on track if monetary policy follows a stable path. The experience of the past two years may have frayed confidence on the latter point somewhat. Demand has been persistently weak in most of the OECD countries, and it has become difficult to say just what is a stable monetary policy. The very widespread nature of the difficulties, however, suggests that a resolution of the problem, if it does not resolve itself, would be better found through cooperation than by countries' insulating their financial markets and pursuing independent solutions. Still, it must be admitted that the past month has seen the first steps in a decade in the direction of increasing controls in OECD countries in response to the strength of speculation against the parity structure of the EMS. An increase in the importance attached to long-run considerations in macroeconomic policy thinking has perhaps altered the rational for controls most significantly. By long run, I mean a period of five to ten years, not an abstract period that is irrelevant to human experience. Macroeconomic policy thinking of the 1950s gave little attention to the long run. (Growth was a concern, but the factors affecting output trends were considered separately from those giving rise to fluctuations around the trend.) In particular, several quantities that have since become a concern got little attention. First was persistent inflation, seen as the cumulative result of efforts to boost output beyond the potential of the economy, sustained by expectations, and costly to unwind. Second was the net foreign asset or foreign debt position of the nation, which could build up relative to the scale of the economy if net exports were persistently in large surplus or deficit. Recognition of the importance of these long-term factors, as a consequence of their arising as problems for individual countries, highlighted the unsustainability of a policy of the persistent use of capital controls. Essentially, controls were a means of maintaining a disequilibrium by cutting off the flows of funds that would keep domestic and foreign price levels in line (either by an exchange rate adjustment or by a monetaryinduced reversal of differential inflation). While this might allow an exchange rate to be maintained while inflating and losing competitiveness, or while keeping cost increases on tradable goods below partner countries and gaining competitiveness, the cumulative current account imbalance would gradually create debt or credit balances, in either official accounts
Controls on international capital movements
151
or private holdings. These would ultimately require adjustment. If policy is not adjusted in a timely fashion, offshore interest rates will begin to reflect the expectation of impeding exchange rate adjustment, requiring even tighter capital controls or allowing domestic interest rates to incorporate this expectation. Thus operating monetary policy with the protection of controls cannot be a way of achieving macroeconomic goals on a lasting basis. It is necessary to do this with the market-based instruments at hand, and with structural policies to strengthen the self-equilibrating forces in the economy. The way the exchange rate is viewed macroeconomically has changed, and this has affected the cost-benefit balance for capital controls. Allowing an exchange rate to float meant that a large part of the rationale for capital controls disappeared. But even countries with fixed exchange rates in the 1980s tended to see them differently than they did earlier. In the course of that decade, fixing the exchange rate to a country with a record of stable, low inflation came to be seen as a way of anchoring monetary policy, and importing stability. This offered an alternative to independent monetary policy as a way of achieving this objective. This option was especially attractive to European countries that had very close trade links with one another. Other OECD countries have preferred to pursue price stability directly, using a range of intermediate indicators to guide policy. Anchoring monetary policy in the exchange rate means more than having the influence of global price competition felt in the tradable goods sector. It means allowing monetary conditions to follow those in the pivot country, with perhaps a premium during a transitional period reflecting the risk as assessed by the market that the policy will not be maintained. In addition to determining explicit monetary conditions, the anchor is expected to operate on expectations as credibility is established, so as to lock in low inflation. Capital controls would impede the functioning of the anchor in both roles. First, it would allow monetary conditions to vary from the pivot currency. Second, it would allow an unsustainable situation to develop, hence removing the basis for establishing credibility. As discussed in Section 3, the strategy of anchoring monetary policies to those of the Bundesbank has led to strains in Europe following the unexpected shock of German unification. Some governments have altered or abandoned their parities. Others are accepting monetary conditions that have been tighter in the short-run than called for on domestic grounds. But the investment in credibility has been seen worth protecting. Doing so means forgoing capital controls. The alternative of not fixing the exchange rate and conducting an independent monetary policy, as Switzerland continues to do in Europe, would also not call for controls. An additional rationale for fixing exchange rates in Europe has been as
152
Jeffrey R. Shafer
one stage in a series leading to a common currency under the Maastricht Treaty. Convergence of economic policies and key aspects of performance are prerequisites for the introduction of a common currency. The removal of capital controls was seen as an early stage, which will ensure that there are no suppressed imbalances that could lead to adjustment problems at the stage of moving to a common currency. Capital controls could only slow convergence. The resource allocation argument for removing capital controls has influenced governments strongly, as indicated in Section 3. Financial sector problems in a large number of OECD countries have tempered the strong claims for the efficiency of allocation decisions made by private market operators. But cross-border flows do not seem to have been a particular source of problems. Bad loans have been overwhelmingly made by local lenders. The access to global markets provided by a liberalized regime for capital flows provides potential backup financial intermediation channels for countries with weakened financial institutions. These are the pros and cons of capital controls that OECD governments have weighed, and continue to weigh. The considerations are not all on one side, but over the past fifteen years or so the balance has clearly come out on the side of net benefits from liberalization. One additional consideration has been international cooperation as promoted within the OECD, and regionally in the European Community. Even the framers of the Bretton Woods agreement, who saw a central role for controls, recognized a potential for countries to use controls to gain an advantage over other countries by using them to manage their exchange rates with a nonmarket-based instrument. All countries would be worse off if each sought to gain such an advantage. The problem has been less serious among OECD countries than feared, because the costs of imposing controls have been higher on the country doing so than was thought earlier. Nevertheless, this has been an issue from time to time. The multilateral review process in the OECD has provided a way of achieving transparency, and bringing peer pressure to bear on countries to move toward more liberal regimes. By doing so it has reduced the potential for tensions to build up. Member countries recognize their responsibility to follow practices that have been developed collectively, and which are frequently reviewed to provide a better financial environment for all. 5
Summary and concluding thoughts
Controls over international capital movements were seen by the framers of the post-World War II international monetary order as an indispensable policy tool. The original Articles of Agreement of the IMF envisaged
Controls on international capital movements
153
their routine use. The articles sought to proscribe some practices out of concern that controls could be abused by one country to obtain an unfair advantage over other countries in the trade domain, not because controls on capital movements were thought to be damaging, per se, to either the country imposing them or to the international financial system. Thus, although the prohibition of most private foreign exchange transactions in the western European countries and Japan gradually gave way to systems of control that allowed considerable scope for capital movements that were not considered "speculative," restrictions continued to be seen by most governments as a necessary macroeconomic policy tool. (There were exceptions - the United States, Germany, Canada, and Switzerland favored free, capital flows as a rule, having recourse to controls as expedients.) Policies toward cross-border financial flows were paralleled in domestic financial policies, which also generally employed controls for monetary management and other purposes. With the establishment of the OECD in 1961, liberalization of capital controls moved formally onto the international agenda, but progress was very slow for over a decade. In the 1970s and 1980s, the view gained ground that controls on financial activity for monetary policy purposes were both unnecessary and harmful. In addition, they became increasingly difficult to maintain as financial markets developed in sophistication. Restrictions on international banking and portfolio flows were dismantled, again in parallel with extensive domestic financial liberalization, to the point where they can be considered unimportant in the OECD countries. While a few countries recently reintroduced some controls in the midst of extremely disturbed conditions in the European Monetary System, this is clearly seen as a backward step, to be reversed as soon as possible. Today OECD governments are overwhelmingly committed to the free flow of funds internationally, and to conducting monetary policies in this context, whether anchored in a commitment to a fixed exchange rate or in the independent pursuit of policies designed to achieve and maintain price stability. The great change in doctrine and practice from close control to relatively free movement of capital reflected the lessons of experience. Capital controls did not allow governments to avoid balance-of-payments crises; when they bought some time, the result was most often the buildup of a larger disequilibrium, and a more difficult subsequent adjustment. If we look back, the theory of monetary policy in the early postwar period focused almost exclusively on flows and short-term considerations, as well as distrust of markets. This meant that the inconsistency between cutting domestic financial markets off from international markets and seeking to anchor monetary policy in a fixed exchange rate was not appreciated. Nor
154
Jeffrey R. Shafer
did capital controls provide a durable solution for monetary authorities that sought monetary independence in order to avoid importing inflation. Germany and Switzerland deployed capital controls in order to loosen their monetary policies from the anchor of the U.S. dollar, which was drifting toward higher and higher inflation in the late 1960s and early 1970s. In the end, however, the authorities in these countries could provide a stabler monetary environment only by cutting loose their exchange rates from the dollar. The dismantling of capital controls was accelerated in some countries by floating of the exchange rate. This was the case when the controls were, as in the case of the United States, an accretion of emergency responses to pressures on a fixed parity, and they no longer had a purpose. But it also became clear that the use of controls could have hard-to-predict effects on expectations and, hence, on the behavior of a floating exchange rate. The trend toward liberalization reflected a recognition that controls were neither necessary nor desirable for the conduct of monetary policy. There was also a growing appreciation among financial authorities of the costs, in terms of the efficient allocation saving, of a command and control approach to monetary policy. This was one force behind the liberalization of domestic financial markets, a shift to market-based instruments of monetary control in place of credit or other balance sheet restrictions, and a parallel liberalization of international transactions. Opening up allowed countries, especially smaller countries, to establish a more competitive financial environment and, thereby, to make a wider range of services available at low cost to domestic residents. The pace of domestic liberalization put some limits on how fast controls over international flows could be dismantled, since regulatory distortions in domestic markets could induce large flows of funds and shift of financial intermediation abroad. But it became difficult to isolate increasingly mature and outward-looking domestic markets from the global market and maintain a strict regulatory regime; international capital flows became a source of pressure for reform of domestic markets. As OECD governments have removed controls on financial transactions, both domestically and internationally, they have also reconsidered the fundamental orientation of monetary policy. It is generally recognized that monetary policy must be directed toward inflation control, indeed toward price stability over the medium term, in order to create a climate in which savings are efficiently allocated with a view to longer-term profits and productivity. How best to achieve this objective is much more controversial. Some monetary authorities have anchored their monetary policy in a commitment to a fixed exchange rate with another currency that has a good record of stability. Others have sought to achieve their inflation
Controls on international capital movements
155
objective independently, relying on a range of indicators and intermediate targets. The results of both approaches have been a reversal of the earlier upward trend of inflation. In many OECD countries inflation is down to, or approaching, levels that meet most pragmatic definitions of price stability. No approach has been free of problems and disturbances, however; indeed, the history of monetary policy is one of a continued need to adapt to new and unexpected developments, something that is never done without having to learn from some mistakes. It has not been possible to smooth out short-run disturbances, or even to avoid occasional miscalculations that may add to them. One lesson that has been learned is that policy must be anchored in a commitment to price stability, with an institutional framework and a record of operations that make this commitment credible. International capital controls have no role in this context. Indeed, they were too often employed to avoid facing up to the policy requirements of sustained monetary stability - to pursue multiple objectives in the short run, at a cost of failing to achieve the fundamental objectives over the medium- to long-term. Deregulated markets force monetary authorities to focus on medium- to longer-term objectives in order to stabilize expectations. Their actions are subject to continuous close scrutiny. This makes the conduct of monetary policy extraordinarily challenging. But authorities that meet the challenge on a sustained basis find that markets are forgiving of temporary miscalculations that are corrected in a timely fashion. Those that deviate more seriously are given timely signals by markets to change course. Controls on international capital flows proved not to be the solution to the problem of conflicting external and internal objectives of monetary policy in the 1950s and 1960s. Indeed, the idea that monetary policy could pursue multiple objectives on a sustained basis was a part of the problem. The removal of capital controls has created a basis for the conduct of monetary policy, whether independently or based on a fixed exchange rate, that is more firmly anchored in medium-term objectives, with timely feedback from markets. It has not been a panacea, but the removal of capital controls is part of the solution.
References Dooley, Michael P., and Peter Isard. (1980). "Capital Controls, Political Risk and Deviations from Interest Rate Parity." Journal of Political Economy 88: 370-84. Dooley, Michael P., and Jeffrey R. Shafer. (1976). "Analysis of Short-Run Exchange Rate Behavior, March 1973 to September 1975." International Finance
156
Jeffrey R. Shafer
Discussion Papers 76. Washington, D.C.: Board of Governors of the Federal Reserve System. Fischer, Bernhard, and Helmut Reisen (1992). Towards Capital Account Convertibility. Paris: OECD Development Centre Policy Brief no. 4. Fukao, Mitsuhiro. (1990). "Liberalization of Japan's Foreign Exchange Controls and Structural Changes in the Balance of Payments." Translated and reprinted from Bank of Japan Monetary and Economic Studies 8, no. 2 (September): 338-61. Katz, Samuel. (1953). "Leads and Lags in Sterling Payments." Review of Economics and Statistics 35, no. 1 (February): 75-81. Kindleberger, Charles P. (1984). A Financial History of Western Europe. London: Allen and Unwin. Lindeneus, Christian. (1990). "Exchange Deregulation - Short and Long-Term Effects." Quarterly Review, (Sveriges Rijksbank, Sweden), no. 3: 44-58. Mathieson, Donald J., and Liliana Rojas-Suarez. (1992). "Liberalization of the Capital Account: Experiences and Issues. IMF Working Paper no. 46. Washington, D.C.: International Monetary Fund. Montiel, Peter J., and Jonathan D. Ostry (1991). "Real Exchange Rate Targeting under Capital Controls: Can Money Provide a Nominal Anchor?" IMF Working Paper no. 68. Washington, D.C.: International Monetary Fund. OECD. (1973). Monetary Policy in Germany Paris. (1980). Controls on International Capital Movements: Experience with Controls on International Portfolio Operations in Shares and Bonds. Paris. (1982). Controls on International Capital Movements: The Experience with Controls on International Financial Credits, Loans and Deposits. Paris. (1987). Structural Adjustment and Economic Performance. Paris. (1990). Liberalization of Capital Movements and Financial Services in the OECD Area. Paris. Olsen, Karl. (1990). "The Dismantling of Norwegian Foreign Exchange Regulations." Norges Bank Economic Bulletins 61 (September): 164-70. Solomon, Robert. (1977). The International Monetary System, 1945-1976: An Insider's View. New York: Harper and Row. Williamson, John. (1992). "A Cost-Benefit Analysis of Capital Account Liberalisation." Paper presented to OECD Development Centre Seminar on Financial Opening: Developing Country Policy Issues and Experience, Paris, July 6 and 7. Working Group on Exchange Market Intervention and Philippe Jurgensen. (1983). "Report of the Working Group on Exchange Market Intervention." Presented at the Working Group on Exchange Market Intervention established at the Versailles Summit of the Heads of State and Government, June 4, 5 and 6, 1982. Revised, March 1983.
CHAPTER 6
Real exchange rates and capital flows: EMS experiences Alberto
1
Giovannini
Introduction
At the end of the 1970s, several Latin American countries pursued a new kind of stabilization programs. These programs, which have been labeled "new style," mixed some variant of exchange rate targeting with monetarist-oriented stabilization policies (see Diaz Alejandro 1981, Dornbusch 1982, Calvo 1983). The interest for these programs, when they were first put in place, came from their stark differences from the traditional International Monetary Fund (IMF)-sponsored plans. The latter relied on nominal exchange rate adjustments to correct relative price distortions and on budget discipline to stem the pressures to inflate. The former, instead, aimed at fighting inflation through nominal exchange rate pegging. In addition, another important pillar of the "new style" programs was an increase in nominal interest rates as part of a package of liberalization of domestic financial markets and international capital flows. The idea of stabilizing inflation through the exchange rate gained increased popularity in the 1980s, with the prominent examples being the Israeli stabilization and the experience of European countries. After 1983, the high-inflation countries that belonged to the exchange rate mechanism of the European Monetary System (EMS) took a stronger resolve on the stability of their exchange rate vis-a-vis the low-inflation country in the system, West Germany. The relative longevity of the EMS, the enthusiasm for the single-market program in the EC and the liberalization of international capital flows all contributed to making the exchange rate mechanism of the EMS a pole of attraction for many countries, both within I am grateful to Veronica De Romanis for excellent research assistance; to Datastream Inc. for providing the data (through a special grant to Columbia Business School); and to Reena Mithal for assistance with the data.
157
158
Alberto Giovannini
and outside the EC. The attraction became irresistible with the start of Economic and Monetary Union. The official philosophy of the hard-currency resolve of European countries is close to the ideas that gained popularity in academic circles in the mid-1980s. These ideas were an extension of the Simons (1936), Calvo (1978), Kydland and Prescott (1977), and Barro and Gordon (1983) propositions on the superiority of rules over discretion in monetary policy. Thus, the story went, pegging the exchange rate credibly binds monetary policy and - as a result of the higher credibility of monetary authorities disinflation is less costly and the maintenance of a low rate of inflation is less difficult. Besides a greater emphasis on credibility, although the issue of credibility was by no means extraneous to the Latin American debate - see in particular Edwards (1985) - the basic characteristics of the exchange rate policies pursued by European countries in the late 1980s match closely those of the Latin American countries in the late 1970s. The discussion and analysis of the Latin American experience identified capital flows as the central issue of concern in these stabilization plans. The concept of the "capital inflow problem" was coined to describe the large inflows of financial capital into the countries that had started the exchange rate stabilization and the capital account liberalization. Among other things, these capital flows were viewed as threatening to monetary stability (in the absence of sterilization, they could give rise to excessive growth of the supply of money) and to the sustainability of the policy packages (the inflows of capital were thought of being the primary cause of real exchange rate appreciations). This chapter reviews the empirical regularities that have been studied in connection with the Latin American experiences, as they apply to European countries. In light of the European experience, it discusses the models that have been applied to explain such empirical regularities. The chapter is closed with a section on the policy questions stemming from the discussion of the data and of the models.
2
The empirical regularities
The European countries I have included in this discussion are three: France, Italy, and Spain. France and Italy entered the exchange rate mechanism of the EMS in March 1979, whereas Spain joined in June 1989. In this exchange rate arrangement, countries effectively peg to the deutsche mark, and attempt to converge to German inflation rates.
Real exchange rates and capital flows
159
Table 6.1. The liberalization of capital account transactions Year
France
Spain
Italy
1987
"Devise titre" abolished.
Direct investment liberalized.
1988
Firms allowed to undertake freely foreign currency loans. Abolition of all restrictions applicable to banks and firms involved in international trade. Nonresidents allowed to borrow in France freely. Acquisition of foreign monetary assets by residents liberalized.
Purchases of foreign medium and long-term financial assets liberalized. Purchases of short-term securities liberalized.
Residents allowed to hold ECU deposits in authorized banks.
Liberalization of all major financial transactions, including deposits, loans, and currency.
1990
Sources: IMF, Annual Report on Exchange Arrangements and Exchange Restrictions, various issues, and national sources.
2.1
Exchange regime and exchange restrictions
The first two "regularities" that define the capital inflow problem regard the policies chosen by government authorities. In the late 1970s, the countries that embarked in the "new style" programs generally accompanied the exchange rate peg with a comprehensive liberalization of international transactions, including both current account transactions and capital account transactions. Of course, goods trade is substantially freer among OECD countries than it was in Latin America prior to the reforms of the late 1970s, and intra-EEC trade is even more liberalized. However, the EMS countries started liberalizing capital account transactions only in the second half of the 1980s. In addition, goods trade liberalization in Spain was almost contemporaneous to the joining of the EMS. Table 6.1 reports the measures taken by our three countries aimed at liberalizing capital account transactions. The table shows that capital account liberalization is not necessarily a function of the exchange rate policy, since Spain undertook it before joining the exchange rate mechanism of the EMS.1 However, to date Spain retains more control on international financial transactions than the other two countries. Indeed, some of the provisions on capital flows reflected a concern of authorities about capital inflows before the time the peseta became a full member of the EMS.
160
Alberto Giovannini
- 8 'i • i i 1 1 i i 1 1 i i
1979
1981
Current Account
1983
1985 date
Portfolio Invest.
1987 1989 1991 ESSS Ann.Change:Roser
Figure 6.1. Balance-of-payments data for France. On the exchange rate regime, it is to be noted that realignments in the EMS have become less frequent over the years. It is generally accepted that the resolve on greater exchange rate stability was manifested more openly since 1987. 2.2
Capital inflows
The defining empirical characteristic of the "new style" Latin American programs was capital inflows. As Diaz Alejandro (1983) and Calvo (1983) point out, capital inflows were regarded in Latin America to be the quickest and clearest sign of success of the stabilization. Figures 6.1, 6.2, and 6.3 report quarterly balance-of-payments data for the three countries, since their inception in the EMS. The data are in constant US. dollars, and the bars are annual changes in foreign exchange reserves. The figures show that the common characteristic of the three countries is a persistent current account deficit, of comparable magnitude. Large capital inflows seem to appear only after 1987. In France, the large increase in foreign exchange reserves in 1990 is associated with large positive balances in portfolio capital. In Italy, the accumulation of foreign exchange reserves is sizable and persistent since 1987, indicating that the equally large current account deficits were more than financed by capital inflows. Finally, in Spain the accumulation of foreign exchange reserves is as large as in Italy,
Real exchange rates and capital flows
161
10000
-SOOO'II.MMIIIH,
1979
1981
Current Account
,,1:111.1,1.
1983 1985 1987 1989 1991 date Portfolio Invest.
{§§3 Ann.Change:Reser
Figure 6.2. Balance-of-payments data for Italy.
1989
1990
1991
1992
date Current Account
Portfolio Invest.
ESSS Ann.Change.Reser
Figure 6.3. Balance-of-payments data for Spain.
162
Alberto Giovannini
1979
1981
1983
1985
1987
1989
1991
Year foreign balance
Italian balance
Figure 6.4. Private loans in Italian balance of payments. and portfolio inflows are more than offsetting the current account deficits (Figure 6.3). Notice in Figure 6.2 that in Italy, despite the large accumulation of foreign exchange reserves at the central bank, portfolio inflows did not match in size the current account deficits. The difference is largely made up by private loans (Figure 6.4), which boomed after 1987. Figure 6.4 shows that the big increase has been in loans from nonresidents taken up by Italian residents. The very large reserve buildups that characterize the experience of countries pegging the nominal exchange rate in a regime of free international capital mobility, which we have observed in Figure 6.1, 6.2, and 6.3, are often blamed as a threat to monetary stability. It is often argued that they can jeopardize the contractionist resolve of monetary authorities, and give rise to unwanted increases in monetary circulation. To assess that conjecture in the case of the countries studied here, I have plotted, in Figures 6.5, 6.6, and 6.7, the rate of growth of high-powered money, as well as the rate of growth of high-powered money accounted for by foreign exchange reserves (i.e., the rate of growth of foreign exchange reserves multiplied by the share of foreign exchange reserves in high-powered money). For France and Italy, Figures 6.5 and 6.6 show that, in the ab-
163
Real exchange rates and capital flows
100
Mar 1979
Mar 1982
Mar 1965 date
Hi-Powered Money
Mar 1986
Mar 1991
Contribution of BoP
Figure 6.5. Balance-of-payments conributions to money growth: France.
Mar 1979
Mar 1982
Mar 1985 date
Hi-Powered Money
1988
Mar 1991
Contribution of BoP
Figure 6.6. Balance-of-payments contributions to money growth: Italy.
164
Alberto Giovannini 25
Jun 1989
Jun 1990
Jun 1991 date
Hi-Powered Money
Contribution of BoP
Figure 6.7. Balance-of-payments contributions to money growth: Spain. sence of offsetting movements in domestic credit, reserve movements would have accounted for major fluctuations in the rate of growth of highpowered money. Such fluctuations are, however, not observed. The figures, of course, do not allow us to conclude that the lack of correlation between the growth of high-powered money growth and the growth of foreign exchange reserves is due to sterilization of reserve movements. Notice that, in the cases of Italy and Spain, the last part of the sample is one where high reserve inflows are associated with a high growth of reserve money. 2.3
Inflation and real exchange rates
The very orthodox monetarists who supported the "new style" stabilization programs in Latin America were relying on purchasing power parity (PPP) to accomplish a freeze in the price level through the freeze in the exchange rate. However, PPP did not materialize in Latin America, and has not materialized in Europe either. Figures 6.8, 6.9, and 6.10 report the nominal exchange rates of France, Italy, and Spain relative to the country that is the anchor of the EMS, Germany, as well as the differentials between the CPI inflation rate in those countries and the corresponding rate in Germany. In the cases of France and Italy, Figures 6.8 and 6.9 show a remarkable process of convergence, whereby the inflation differen-
Real exchange rates and capital flows
Mar 1979
M a r l 982
Mar 1985 Date
165
Mar 1988
Mar 1991
Figure 6.8. Inflation differential and exchange rate: France.
Mar 1979
Mar 1982
Mar 1985 Date
Mar 1988
Mar 1991
Figure 6.9. Inflation differential and exchange rate: Italy.
166
Alberto Giovannini
1989
Jun1990
Jun1991
Jun1992
Date
Figure 6.10. Inflation differential and exchange rate: Spain. tial decreases from peaks of 8 percent in France and 16 percent in Italy to negative and virtually zero in about thirteen years. This convergence process is, interestingly enough, interrupted in France after the EMS realignments of 1983 and 1986, and in Italy after the realignment of 1983. Italy's, as well as to some extent France's, convergence slowed down noticeably in the period starting from early 1988 to the end of 1990, when it seemed that inflation differentials had reached a plateau, dangerously high in the case of Italy (4 percent). In Spain, the behavior of the CPI inflation rate relative to Germany is similar to that in Italy. The differential starts at about 4 percent in 1989 to fall to about 1.5 to 2 percent in 1992. The behavior of inflation, however, conceals what happens to relative prices. The delicate trade-off is between the inflationary impulses of devaluations and the chance they offer to correct relative price distortions. Expost, France appears to have played out that trade-off egregiously. Figure 6.11 plots the inflation differential from Figure 6.8, together with an index of relative CPIs. The figure shows that, despite the accelerations in inflation after the 1983 and 1986 devaluations, relative prices were adjusted sizably in those occasions. In addition, note that the three inflation cycles appearing in Figure 6.11 are such that, after each devaluation, the rate of acceleration of inflation decreases noticeably. As a result of this, the rela-
Real exchange rates and capital flows
167
S
-2 Mar 1979
Mar 1982
Mar 1985 Date
Intl. Diff.
Mar 1968
•1991
Relative CPI
Figure 6.11. Inflation differential and real exchange rate: France. tive consumer price index is in 1992 only 6 percent above where it was at the start of the EMS in March 1979.2 In Italy the experience has been dramatically different. The cumulative increase in the relative CPI index from March 1979 to the summer of 1992 shown in Figure 6.12 is as high as 45 percent. The most striking difference with France is that the EMS realignments fail to affect relative prices significantly or in a lasting way. This remarkable fact might be due to the difference in the size of fluctuation bands the two countries had. With a wide band, a realignment that is of the same order of magnitude of the width of the band can hardly affect the spot exchange rate and relative prices. Hence, as Figure 6.9 seems to suggest, a wide band permits a trend in the nominal exchange rate, and appears to be almost irrelevant for the behavior of the nominal exchange rate. The Spanish experience, covering a much shorter time span, is more difficult to characterize. Indeed, the period in Figure 6.13 does not include any realignment. Despite a noticeable fall in the differential of inflation rates, consumer prices in Spain were 10 percent higher than those in Germany after three years of EMS membership. Notice that the increase in German inflation after unification does improve convergence noticeably, both in France and in Italy.
Alberto Giovannini
168
18
M a r l 979
1982
Mar 1985 Date
Infl. Diff.
Mar 1988
Mar 1991
Relative CPI
Figure 6.12. Inflation differential and real exchange rate: Italy.
112
Jun 1989
Jun1990
Jun 1991
100
Date Infl. Diff.
Relative CPI
Figure 6.13. Inflation differential and real exchange rate: Spain.
Real exchange rates and capital
flows
169
As is to be expected in countries open to international trade, these fluctuations in relative consumer prices hide interesting cross-sectoral differences. Sectors that are more exposed to international competition cannot afford very high relative price increases. Thus, as Diaz Alejandro (1984) points out with reference to Latin American countries, it is often the case that nontradable goods inflation is higher than tradable goods inflation. The three panels in Figure 6.14, taken from De Gregorio, Giovannini, and Krueger (1992), show a remarkable trend for the relative price of nontradables, both in Italy and Spain. There is almost a one-to-one correspondence between the problem of the stubborness of inflation after the exchange rate is pegged and the problem of the increase in the relative price of nontradables,3 which I will discuss more in detail in the following section. 2.4
Real interest rates
Section 2.2 showed the very large inflows of capital, especially into Italy and Spain after 1987, matched by large balance-of-payments surpluses. A major empirical regularity noted by scholars of Latin American countries is that, despite the very large capital inflows, real interest in those countries remained very high. The problem with real interest rates is how to measure them. There is the question of the choice of the appropriate deflator, as well as the question of ex-ante versus ex-post inflation estimates. The data that I reproduce allow reference to different concepts of real interest rates. Consider Figures 6.15, 6.16, and 6.17. They report the upper EMS fluctuation band for the franc, the lira, and the peseta (the maximum allowed depreciation of the three currencies relative to the deutsche mark), the one-year forward exchange rate of the deutsche mark (the price of one deutsche mark in terms of, respectively, French francs, liras, and pesetas, for delivery one year hence), and the one-year forward foreign exchange premium (the difference between the one-year forward exchange rate and the spot exchange rate, relative to the spot exchange rate). In Euromarkets the forward premium is, plus or minus transactions costs, equal to the ratio of (1 plus) the domestic and (1 plus) the foreign (in this case DM) interest rate. The latter is approximately equal to the differential between those currencies' one-year Eurodeposit rates and the corresponding deutsche mark Eurodeposit rates. 3
As De Gregorio et al. (1992) point out, fluctuation in terms of trade as well as fluctuations in the relative price of nontradables in partner countries account for the differences between the behavior of the relative price of nontradables and the behavior of relative inflation rates.
(1980=100)
FRANCE
1975 197619771978 1979 1980198119821983 19841985198619871988 1989 1990
(1980= 100)
ITALY
i i i i i i i i i i i i i i 1975 197619771978 197919801981198219831984198519861987198819891990
(1980= 100) 118
V -\ no " \ iO8 \ fO6 \ 104
U6 14
SPAIN
£ 102 100 98 96 94 92 90 88.
>v
\
\ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 19751976 19771978 1979198019811982198319841985198619871988 19891990
Figure 6.14. Relative price of nontradables: PT (industry)/PN (service).
-10 3.9-
•8
3,8-
6
3.7-
•4
3.6-
2 0 n-
3.4-jJO.O"1
-2 -4
'—
t-31 -86Oct-30-87 Oct-28-88 Oct-27-89Oct-26-90 Oct-25-91 Date
Upper ERM Band
J
1-Year Fwd Rate
-6
1-Year Fwd Prem.
Figure 6.15. One-year forward French franc exchange rate and foreign exchange premium.
-31-86Oct-30-87Oct-28-88Oct-27-89Oct-26-90Oct-25-91 Date —
Upper ERM Band
1-Year Fwd Rate
1-Year Fwd Prem.
Figure 6.16. One-year forward lira exchange rate and foreign exchange premium.
172
Alberto Giovannini
Jun-23-89
Jun-22-90
Upper ERM Band
Jun-21-91 Date
1-Year Fwd Rat©
Jun-19-92
1-Year Fwd Prem.
Figure 6.17. One-year forward peseta exchange rate and foreign exchange premium.
If risk premiums are of second order, the forward exchange rate is a good proxy of the expected future exchange rate. In that case, Figures 6.15, 6.16, and 6.17 show whether or not the upper fluctuation limits of the deutsche mark vis-a-vis the franc, the lira, and the peseta were regarded believable by the markets. Between October 1986 and October 1989, the one-year forward deutsche mark relative to the franc and the lira was consistently above the upper fluctuation limit, an indication that the EMS parity was not believable. Of course, the high forward exchange rate was also reflected in a high forward premium: That is, interest rates in francs and liras were high because the official parities were not believed by the market. The one-year interest rate differentials are reported on the axis of the graphs; they start at about 5 percent for France and 8 percent for Italy. In the case of France they converge almost fully, except for the instabilities associated with the recent EMS crisis, whereas in the case of Italy the oneyear interest rate differential rarely goes below 2 percent, indicating failed convergence. The behavior of the peseta interest rate differential from 1989 on is similar to that of the lira, except that in the case of the peseta the high interest rates reflected to a much larger extent expectations of depreciation of that currency within the EMS band: The forward rate, except for the very recent period, is always within the maximum fluctuation limit.
Real exchange rates and capital flows
C)ct-86
Oct-87
France - Germany
Oct-88
Oct-89 Date
173
Oct-90
Italy - Germany
Oct-91
Spain - Germany
Figure 6.18. Ex-post real interest rate differential. Ex-post real interest rate differentials are shown in Figure 6.18. They are computed subtracting from the forward premium the differential in CPI inflation rates. Hence they are approximately equal to the difference between nominal interest rates and own-currency inflation rates. Figure 6.18 shows that, throughout the whole period from 1986 to early 1992, expost real interest rates have been higher in France, Italy, and Spain than they have been in Germany, except for the period from December 1989 to October 1991, when ex-post Italian rates were lower than their German counterparts. An estimate of long-term interest rate differentials is contained in Figure 6.19. The figure plots indexes of yield to maturity on government bonds. These indexes are computed on bonds maturing from one to five years hence, and so they cannot be used for a precise assessment of longterm expectations. Nevertheless, Figure 6.19 shows the big difference between yields on French versus Spanish and Italian bonds relative to German bonds. The very high yield differentials between Italian and Spanish bonds and German bonds largely reflect exchange rate expectations.4 By contrast, no significant long-term movements in the French franc4
In the case of Italian bonds, uncertainties and inefficiencies on the reimbursement of the withholding tax to entitled nonresidents are such that Repubblica bonds are traded on a net basis: Part of the observed differential is thus due to tax factors.
174
Alberto Giovannini
"fm .—,
7
Jan 1985
Jan 1987
France - Germany
\x\
V\
_.-*
Jan 1989 Date Italy - Germany
Jan 1991
Spain - Germany
Figure 6.19. Government bond yield differentials. deutsche mark exchange rate seem to be reflected in the relative yield differential between French and German government bonds. In summary, the high interest rates that characterized the capital inflow problem in Latin America are - with the necessary proportional adjustments due to the very high initial inflation rates in Latin America - also an issue of primary concern among European countries.
2.5
Economic Activity
The literature on the capital inflow problem in Latin America points out that, especially towards the end of the experiment with the exchange-rate peg, economic activity slows down considerably. Figure 6.20 shows an important fact: Business cycles appear to be highly synchronous among the four European countries studied here. For that reason, it is difficult to argue that, for example, the slowdown in economic activity in France, Italy, and Spain is due to the exchange rate policies followed in these countries relative to Germany. Although Germany's slowdown has been delayed by the effects of the unification with the East, the German cycle of industrial production growth broadly matches the one in the other countries.
Real exchange rates and capital flows
Mar 1979
Mar 1982
Germany
175
Mar 1985 Date France
Mar 1988
• Italy
Mar 1991
Spain
Figure 6.20. Annual growth in industrial production.
3
Interpreting the data
The discussion in the previous section has highlighted the presence of a remarkably common pattern among the Latin American experiences of the late 1970s and the European experiences of the 1980s. The common pattern is both in the policies pursued (exchange rate pegging and financial liberalization) and in the responses of macroeconomic variables (large capital inflows, high real interest rates, real exchange rate appreciations). Given this common pattern, it is tempting - and appropriate as a first pass - to look for a common explanation. Candidate explanations can be divided in two major classes: models that rely on wage or price stickiness, and models that do not. In the first class, which includes - among others - Dornbusch (1982) for the Latin American countries and Giovannini (1990) for the European countries, the problem of exchange rate-based stabilizations is due to a combination of slow responses of inflation and lack of credibility. The freeze of the nominal exchange rate does not bring down immediately the rate of inflation because of inertia in wages and (possibly) prices, thus giving rise to an appreciation of the real exchange rate. The real appreciation causes a worsening of the current account balance and, at the same time, brings
176
Alberto Giovannini
expectations of future exchange rate corrections. These expectations have two effects. On one side, they slow down inflation convergence, since they are incorporated in new wage contracts; on the other side, they exert upward pressure on interest rates. This pressure is accommodated by the central bank, which can only use interest rates to defend the nominal exchange rate parity. The model of sluggish inflation and credibility suggests that, in order for the convergence to be successful, the central bank and government authorities have to endure a prolonged period of very high interest rates and severe exchange rate overvaluation to convince the public that the fixed exchange rate is there to stay. The adjustment of relative prices, however, can only occur if the low-inflation target is overshot, since in the transition the slow adjustment of inflation accumulates relative price distortions. The alternative model, which does not rely on sluggishness, is based on the distinction between traded goods inflation and inflation in the nontradables sector (see Calvo [1983] for the first application to the capital inflow problem, and De Gregorio et al. [1992] for a discussion of it in the European context). As I have argued, one of the regularities of the capital inflow problem is the higher inflation in the nontradables sector. Assume for simplicity that in the tradables sector the law of one price holds exactly. What can account for the increase in the relative price of nontradables? The traditional two-sector model allows for three main classes of shocks: productivity shocks, demand shocks, and real wage shocks. Consider productivity shocks first. An increase in productivity in the tradable sector puts upward pressure on the real wage rate, and causes a real appreciation of the exchange rate, that is an increase in the price of nontradables relative to tradables. The increase in nontradables prices is accompanied by an increase in the tradables sector (increase in employment there, and a trade balance surplus, since spending patterns are likely not to change as much as productivity). Spending shocks could come from two sources: wealth effects and exogenous spending increases, like for example government purchases of goods and services. The former could be associated with the initial confidence booms that often accompany the financial liberalization and exchange rate peg. An increase in spending generates an increase in the real exchange rate, a decrease in the relative size of the tradables sector (since the real exchange rate increase increases labor costs in the tradables sector and gives rise to labor shedding there, to recover average labor productivity) and a current account deficit (since the increase in spending increases the balance of spending and income). Finally, real wage shocks in the two-sector model could originate either from a shift in bargaining attitudes by trade unions if the wage was set in
Real exchange rates and capital
flows
177
a centralized bargaining agreement, or from a shift in labor supply. In both cases, the higher labor costs are passed through into higher relative prices of nontradables, but they are absorbed in the nontradables sector only through a decrease in employment, with the attendant increase in marginal and average labor productivity. Which of the two broad classes of models helps explain the regularities documented in Section 2? A study of the two-sector model as it applies to European countries is carried out by De Gregorio et al. (1992). They find that, in general, while government purchases of goods and services can have some explanatory power in a regression equation of the real exchange rate, in practice the variation in government spending (at constant relative prices) is so small that it cannot explain the very sizable changes in relative prices documented in this chapter. An identification of the two other sources of shocks - productivity and real wages - can be carried out only by paying close attention to the share of tradables and nontradables in value added, as well as the balance of trade in goods and services. De Gregorio et al. (1992) argue that perhaps only in the case of Spain the increase in the relative price of nontradables observed after that country joined the exchange rate mechanism can be ascribed to independent real wage shocks. Of course, the ability of the two-sector model to describe the Spanish experience is not to be taken to mean that sluggish wages and prices and credibility are not important. These real wage shocks could be the result of contracting based on expectations of the exchange rate that were not consistent with the monetary authority's resolve and, ex post, ended up producing too high real wages. In the case of Spain, the share of tradables in value added decreases since 1989. The case of Italy is more puzzling. In that country, the share of nontradables in value added has been virtually constant since 1985. And yet, measures of total factor productivity in the aggregate tradables sector are relatively constant since the first half the 1980s. In conclusion of this section, it is useful to point out that the traditional sources of nontradables inflation, spending and productivity shocks, are not easily identifiable from the data on government purchases, total factor productivity (when available), and shares in value added of tradables and nontradables. The difficulty of finding the simple model mirrored in the data is likely caused by costs of adjustment both in the intersectoral reallocation of capital and in the intersectoral reallocation of labor. 4
Policy questions
As I pointed out in the preceding section, the similarity of the macroeconomic responses to the hard-currency experiments in Latin America
178
Alberto Giovannini
and Europe would seem to call for a common model, but that model is hard to identify. Some of the difficulty is also due to the fact that, when looking at Europe, there are some differences across countries. What accounts for the differences across these three countries, and in particular across France and Italy, which remained in the EMS for the same period of time? Exchange rate policy and debt seem to be deserving the closest attention. Consider exchange rate policy first. A comparison of Figures 6.8 and 6.11 with 6.9 and 6.12 suggests major differences between the two countries. The narrow band allowed France to achieve nominal exchange rate movements that looked more like a step function, whereas in the case of Italy the wide band seems to accommodate an independent trend of the exchange rate. As a result, the nominal exchange rate adjustments in France provoked noticeable improvements in the real exchange rate, whereas that does not seem to be the case of Italy. These figures trivially suggest that a narrow band might be a stronger weapon to fight inflation and currency overvaluation. The second problem is debt. It is well known that much of the recent foreign exchange crisis in Italy was paralleled by increased nervousness in financial markets on the stability of Italian public finances. Could the higher debt and larger deficits in Italy have been the cause of the real exchange rate appreciation, of the high real interest rates, and of the current account deficit? Traditional macroeconomic models suggest that the effects of the stock of debt on the level of interest rates, through some sort of portfolio balance model, should be second order. In addition, the effects of government deficits on national savings are notorious in Italy, which until recently has had one of the highest private savings ratios in Europe. Hence high interest rates, and the current account deficits, cannot be easily ascribed to crowding-out type of phenomena. What, however, has been important in the recent foreign exchange crisis was the knowledge that high interest rates were especially detrimental to sustainability of Italian public finances, and therefore the suspicion that the very high rates necessary to stave off speculative attacks could not be achieved. In this sense, weak public finances make a program of inflation convergence through currency pegging a vulnerable one. References Barro, R. J., and D. B. Gordon. (1983). "A Positive Theory of Monetary Policy in a Natural Rate Model." Journal of Political Economy 91: 589-610. Calvo, G. A. (1978). "On the Time Consistency of Optimal Policy in a Monetary Economy." Econometrica 46, no. 6: 1411-28. (1983). "Trying to Stabilize: Some Theoretical Reflections Based on the Case of
Real exchange rates and capital
flows
179
Argentina." In P. Aspe Armella, R. Dornbusch, and M. Obstfeld (eds.), Financial Policies and the World Capital Market: The Problem of Latin American Countries. Chicago: University of Chicago Press for NBER. Calvo, G. A., L. Leiderman, and C. M. Reinhart. (1993). "Capital Inflows and Real Exchange Rate Appreciation in Latin America: The Role of External Factors." IMF Staff Papers 40 (March): 108-51. De Gregorio, J., A. Giovannini, and T. Krueger (1992). "The Boom of NonTradeable Goods Prices in Europe: Evidence and Interpretation." Mimeograph copy. International Monetary Fund (October). Diaz-Alejandro, C. F. (1981). "Southern Cone Stabilization Plans." In W. R. Cline and S. Weintraub (eds.), Economic Stabilization in Developing Countries. Washington, D.C.: Brookings Institution. (1984). "Latin American Debt: I Don't Think We Are in Kansas Anymore." Brookings Papers on Economic Activity, 335-89. Dornbusch, R. (1982). "Stabilization Policies in Developing Countries: What Have We Learned?" World Development 10, no. 9: 701-8. Edwards, S. (1985). "Stabilization with Liberalization: An Evaluation of Ten Years of Chile's Experiment with Free Market Policies, 1973-1983." In Economic Development and Cultural Change 132 (January): 223-54. Giovannini A. (1990). "European Currency Reform: Progress and Prospects." Brookings Papers on Economic Activity 2: 217-92. International Monetary Fund. Annual report on exchange arrangements and exchange restrictions. Washington, D.C.: Various issues. Kydland, F , and E. C. Prescott. (1977). "Rules Rather Than Discretion: The Inconsistency of Optimal Plans." Journal of Political Economy 85: 473-91. Simons, H. (1936). "Rules Versus Authorities in Monetary Policy" Journal of Political Economy 44, no. 1: 1-30.
CHAPTER 7
Monetary policy after German unification Wilhelm Nolling
1
The basic problem
It has been known for a long time that the centralized planned economy in its various forms - forms that differ only in degree and not in kind - is not in a position, either now or in the future, to fulfill the demands placed on a modern national economic system. These demands include adequate economic growth, rising living standards, a high level of employment, price stability, social equality, and the conservation of the environment and natural resources. All the Comecon countries had one decisive feature in common. Without exception they have to come to terms with the Stalinist heritage of a centralized planned economy. The main defect of this system is that it was not able to react with sufficient flexibility to external influences or to initiate internal changes. Of course, there were national differences between the various planned economies - for example, between Hungary and Poland. However, these differences are of secondary importance in comparison with the structural problems, which are our prime focus of interest in this chapter.1 For decades, the dominant theme in academic and political discussions has been the transition from capitalism to socialism - and not vice versa. Marx and Schumpeter1 belong to the leading exponents of this school of thought. Of course, everyone will remember the discussions about the "third way" - in other words, the synthesis of the two antagonistic economic orders. In view of the prevailing political power structures, however, a transition in the opposite direction seemed unthinkable and was thus never discussed. This situation has changed virtually overnight. We discover that there are no examples to follow, no textbook solutions on which This chapter was written from the perspective of Germany's economic and political situation as of December 1992.
181
182
Wilhelm Nolling
to base political decisions. The present challenge is akin to open-heart surgery. Since the reforms began, there has been much discussion about the conflicting approaches of "gradual evolution"2 on the one hand and "shock therapy" on the other. Leaving aside the violent upheavals in eastern Europe after World War II, the West German currency reform of 1948 is the only true example of shock therapy in the postwar era. The introduction of the free-market system to the GDR was a clean sweep, a radical break with the past. Two years and four months ago, the social market economy system was introduced into the GDR at one go, after a very short period of preparation. The deutsche mark became the only legal tender there, and the Deutsche Bundesbank assumed responsibility for internal and external monetary policy in the still formally independent East Germany.3 Monetary, economic, and social union was the decisive step toward the political reunification of the two parts of Germany three months later. The clarification of "external aspects" was followed by the signing of the Unification Treaty on August 31. Through the accession of the GDR to the Federal Republic on October 3, 1990, it was fully integrated into the legal system in force in western Germany. Despite all the difficulties the Bundesbank has faced since the 1st of July 1990, the international reputation of the mark is beyond dispute today. This is evidenced in the great trust placed by "economic subjects" in Germany and abroad in the Bundesbank's overriding commitment to monetary stability. Section 3 of the Bundesbank Act states: "The Deutsche Bundesbank regulates the quantity of money in circulation and of credit supplied to the economy, using the money powers conferred on it by this Act, with the aim of safeguarding the currency." It is hard to imagine a more unambiguous commitment on the part of a national central bank regarding the objective of price stability. 2
Reasons for applying shock therapy in the GDR
In analyzing our experience one should remember four important facts. First, the economic fundamentals of the West German economy were in rather good shape. Price increases were low. Public debt was about 50 2
3
It is worth mentioning Alfred Marshall's basic attitude to this topic (1961, XII): "Economic evolution is gradual. Its progress is sometimes arrested or reversed by political catastrophes: but its forward movements are never sudden. . . . The mecca of the economist lies in economic biology rather than in economic dynamics." For details of the Monetary Economic and Social Union, see Deutsche Bundesbank 1990a, 40-45; 1990b, 13-28.
Monetary policy after German unification
183
percent higher than seven years ago, when the Kohl government took over from Helmut Schmidt; however, in 1989 the increase had slowed down and was counterbalanced by a surplus in the social security accounts. Our external status showed an impressive net worth of more than DM 500 billion. Unemployment, although very high in historical comparison, had come down slowly; over the years there had been an impressive increase in employment. Second, the West German society was by and large ignorant of the state of the East German economy and of many important aspects of daily life there - especially the psychology of a population under more than forty years of dictatorial stress and a much lower standard of living, which was not achievement-oriented at all. Third, in the early days of 1990 there was a lot of distrust that the much desired political changes might not last, that the still governing communists and the occupying Russian army might put an end to the newly gained freedom of movement and expression. Fourth, the East German citizens were impatient and demanded freedom as well as the deutsche mark in place of their inferior currency. They wanted to put an end to their perceived economic and social inferiority and to be in a position to buy high-quality Western goods. These psychological and economic elements were intensified by the flood of emigrants from East to West (175,000 in the first four months of 1990) and by party political rivalries. The coalition government in Bonn, in particular, raised the expectations of the East German population. The success of the coalition parties in the election on March 18,1990, is attributable to their willingness to comply with the wishes of the majority of the East German population. In all these respects there were major factors operating beyond the control or influence of politics. Thus, the political options available were rather limited: To formulate and conduct a somewhat rational policy was not easy. The attempt was blunted by political pressures, election-result-oriented powerplays, unrealistic expectations (wishful thinking), and a lot of market economy ideology. In general a gradual, evolutionary approach depends on certain preconditions. In order to achieve a successful and orderly transition, political decision makers must be aware of the effects of a given measure at a given time. In other words, it is important to know how the old system functions and to recognize what changes are necessary for a gradual transition to the free-market system. It is clear for all to see that these conditions were not fulfilled and that the evolutionary approach is possible only on the basis of trial and error. In order to avoid a lot of these political difficulties, the proponents of
184
Wilhelm Nolling
shock therapy believed in a radical approach and aimed to turn the entire "organism" of the national economy on its head, to revolutionize the entire economic environment. An automatic consequence of this approach is that production capacities become obsolete, leading to unemployment and a decline in the gross national product. Success depends on the extent to which living standards decline and the way in which this decline is distributed among the various groups in society. It also depends on the speed with which the benefits of shock therapy make themselves felt. However, shock therapy has one decisive disadvantage: Its positive effects on the organism of the national economy are entirely unpredictable. 3
Reasons for the introduction of the deutsche mark to the GDR
July 1, 1990, marked the beginning of the greatest economic and social experiment that the Western world has ever seen: the fusion of two states with diametrically opposed economic systems. Before the final political decision was taken on February 7, 1990, there had been a growing discussion about the various options.4 The general opinion was that, from a purely economic viewpoint, a course of shock therapy was inadvisable indeed. However, the special situation of the GDR - that is, the enormous political pressure exerted by mass emigration, the incipient collapse of the GDR economy,5 and the inability to think of any other, less damaging alternatives - led to a paniclike reaction. An evolutionary approach was rejected in favor of shock treatment. 4
The basic political response
The decision to expose the GDR to shock therapy as a way of rapidly restructuring its economic system was based on four key strategic elements.6 The first strategic element was the introduction of the deutsche mark to the territory of the GDR. 7 The second element was the hope that sufficient 4 5
6
7
For details of the discussion see Nolling 1991, 12-16, 32-41. For the last comprehensive analysis of the GDR economy before 1990, see Deutscher Bundestag Bundestags-Drucksache 11/11, 1987. This interpretation of an integrated strategy with originally three key elements was first developed in a speech for an international conference at Szirak, Hungary, in August 1990 and published in Nolling 1991, 42-43. For the state of discussion in the month of the decision to introduce the deutsche mark, see Nolling 1990.
Monetary policy after German unification
185
private capital and experienced management would be available to offset the destruction of production capacity wrought by the shock therapy. The third element entailed the transfer of public budget funds from west to east. On the one hand, public funds were needed to maintain the social welfare of East German citizens until such time as productivity in the eastern part of Germany reached West German levels. On the other hand, public funds needed to be spent on improving and modernizing East Germany's infrastructure. The fourth element involved the need to build an entirely new set of institutions in East Germany. Internal revenue service, labor, and social security offices, local communal authorities, and democratic bodies at the various levels of government were all required. 5
The German Monetary Union and its effects Introduction of the deutsche mark
The introduction of the deutsche mark had obvious advantages. The mark is one of the world's hardest currencies and, as such, is in high demand and accepted everywhere. It was assumed that the desire to acquire marks would provide incentives for extra effort. Good money for good products was seen as a stimulus for greater productivity. In macroeconomic terms the introduction of the mark meant that a united Germany would assume the burden of the GDR's external debts. What is more, the discriminatory divisions between those who earned and owned deutsche marks and those who earned and owned East German marks were sure to disappear. These benefits must be set against certain serious disadvantages. In effect, the introduction of the deutsche mark at an official exchange rate of 1:1 led to an appreciation of the East German mark. Whatever method of calculation is used, one arrives at an appreciation of between 200 and 300 percent;8 from now on, the former GDR economy was fully exposed to international competition. The consequences of this were catastrophic; virtually no company in the former GDR has remained unscathed. Just imagine what would happen if the currency of a fully functioning economy such as Spain, the United Kingdom, or West Germany were to appreciate overnight by the same percentage? The effects on exports would be disastrous - as eastern German companies are now experiencing to their cost. The daily reports of company failures thus come as no surprise. The de8
In government-controlled trade with the Federal Republic, for instance, a conversion rate of M 4.40 for DM 1 was generally applied until the end of 1989. From January 1, 1990, onward the official exchange rate between the deutsche mark and the GDR mark individually was 1:3. See Deutsche Bundesbank 1990b, 16.
186
Wilhelm Ndlling
struction of production capacity and sharp rises in unemployment are the logical consequences of the prescribed course of shock therapy. Wage policies played a crucial role in this context. Encouraged by the federal government's promise that no one would be worse off as a result of the German Monetary Union, the trade unions have pushed through significant wage rises, which unfortunately do not reflect similar increases in productivity. In eastern Germany the hopes of a speedy upswing have not been fulfilled so far. In particular, the manufacturing sector still has to contend with serious adjustment problems, with the result that a broadly based, self-sustaining upswing has not been set in motion yet. In addition to the general difficulties presented by adjusting to production under market conditions and by wage and salary rises far in excess of the improvement in productivity,9 this has owed a lot to the fact that the markets in eastern and central Europe almost ceased to exist following the collapse of the old socialist system in the former Soviet Union. The resulting massive demand losses led to a rapid decline in industrial production in the new federal states. Since the middle of 1992, production has stabilized, though, at a level which is less than one-third compared with that during the first half of 1990.10 Many old jobs have now disappeared without a sufficient number of new employment opportunities having emerged. Actually deindustrialization has taken place at a frightening pace. In December 1992 the job losses amounted to 3.5 million. They have been distributed as follows: 1.1 million people are unemployed; 355,000 people are employed in job creation schemes; 835,000 are attending further and advanced training courses; and 578,000 are drawing transitional benefits for early retirement.11 So, overall 2.9 million people are directly or indirectly supported by the state. In addition, more than half a million East Germans commute to the West. Bringing down the exceptionally high level of unemployment continues to be the most urgent problem facing eastern and western Germany alike. The role of private investment Highly subsidized private investment is seen as a key element in the reform strategy. The intention is to give the private sector priority over state inter9
10 11
For a recent overview of the productivity level in East Germany, see Deutsches Institut fur Wirtschaftsforschung 1992, 646-49. Source: Deutsche Bundesbank, data bank. For the figures, see Monthly Report of the Deutsche Bundesbank, January 1993, 15.
Monetary policy after German unification
187
vention. The Bundesbank has also spoken out in favor of this approach. At a press conference given in Bonn on February 9, 1990, the president of the Bundesbank was quoted as follows: "For me the most important point of all - more important than the introduction of the deutsche mark - is that the necessary transfer payments decline to the same extent as private capital flows into the GDR." 12 However, we are not satisfied with the amounts provided for so far. There was and still exists the problem of privatization. To solve this task, the eastern German economy needs to be completely restructured. The gigantic and inefficient former state-owned combines must be dismembered and privatized. We have developed a special rather successful solution in the form of the "Treuhandanstalt" or "Trusteeship Corporation," serving as an "intensive care unit."13 As from the autumn of 1990 the fate of the companies in the Treuhandanstalt's intensive care unit has had one of three possible outcomes: Either the "patient" was given up for dead due to the impossibility of restructuring; or it was released for further "treatment" in a privatized form; or else it remained under the administration of the Treuhandanstalt while the possibility of restructuring was investigated. The company was then either sold off or finally closed down. Between July 1990 and June 1992 the Treuhandanstalt took over a total of more than 10,000 enterprises. The privatization balance at the end of June 1992 stands at 8,175 enterprises.14 Up to the end of June, 1475 management buyouts had been realized. The Treuhandanstalt has registered increasing interest from foreign investors: By June of this year, 412 contracts have been concluded, thereby securing 111,000 jobs and promising an investment of DM 12 billion. France tops the list of foreign investors at DM 2.7 billion, followed by the United States with DM 1.65 billion. As a total result of the Treuhandanstalt activities job guarantees have been given to 1.23 million workers; investment guarantees amounting to DM 140 billion have been provided. So far the total privatization proceeds amount to DM 31 billion (far short of the Utopian figure of DM 600 billion quoted by the former Treuhandanstalt Chairman Rohwedder). The role of public transfers In the context of German economic and monetary union, the transfers of public funds from western to eastern Germany was intended to have a 12 13 14
Pohl 1990, 1. This approach is explained in Nolling 1992a. Recent data concerning the Treuhandanstalt and its privatization performance are published in "Treuhandanstalt, Informationen," which can be ordered from: VVCO, Am Treptower Park 28-30, O-l 193 Berlin, Germany.
188
Wilhelm Ndlling
"pump priming" function and support the introduction of the deutsche mark to the economic system of the former GDR. Other supporting measures have included the favorable conversion of savings accounts and company bank deposits. Public transfers were considered necessary during the first phase of economic and monetary union, as it must be assumed that other sources of revenue to finance public spending requirements would remain inadequate for a long time to come. We now realize that these transfer payments will turn out to be much higher than expected (Table 7.1).15 The eastern German economy is being drip-fed by the "old" Federal Republic. This year alone (1992), transfer payments amounting to DM 180 billion will be channeled into the "new" federal states in the east.16 This is equivalent to more than 6 percent of western Germany's gross national product. The bulk of these funds is being spent on consumption. We realize, however, that in the long run the standard of living in eastern Germany cannot be "borrowed" in the form of income transfers from western Germany. The revitalization of the eastern Germany economy must be achieved primarily through private and public investment. 6
Repercussions of German Monetary Union on the Federal Republic
At first economic and monetary union worked like a massive economic promotion program in the Keynesian sense. The only difference was that in the beginning it boosted an economy that did not need additional stimulation due to the already high level of capacity utilization. This increased the danger of price rises and also resulted in very large public deficits. After strengthening the West German economy and the economies of most of Germany's major trading partners over a period of two years, the stimuli of GMU have subsided considerably during 1992. As a result of the unification process, monetary conditions have changed totally. We aimed to achieve two objectives at the same time: exchange rate stability and keeping inflation low. Massive government transfers and inflationary pressures make monetary management more difficult.17 As I mentioned before the conversion rates of the East German mark were chosen on political grounds and resulted in an excessive expansion of the money supply.18 In December 1992 the money supply M3 ex15 16 17
18
For a closer examination, see Deutsche Bundesbank, 1992, 15-22. For a good first estimation of the transfers from 1991 to 1994, see Dohnanyi 1990, 261. For one of the best analyses in 1990, see G. Schinasi, et al., "Monetary and Financial Issues in German Unification," in Lipschitz and McDonald 1990, 144-54. For details, see Nolling 1992b.
Table 7.1. Western German public sector payments to eastern Germany in 1991 and 1992 (DM billion) Payments Gross payments by the federal government* Payments by the western German Lander governments and local authorities' Expenditure of the "German Unity" Fund financed on credit Loans of the ERP Special Fund and specialized banks which are interest-subsidized by means of public funds Gross payments from the EC budget Deficit of the Federal Labor Office in eastern Germany Deficit of the staturoty pension insurance funds in eastern Germany* Gross payments, total of which Expenditure Shortfalls in tax revenue in western Germany Waiver of share in turnover tax less Tax revenue of the federal government in eastern Germany* Tax revenue of the EC in eastern Germany Net payments, total of which Federal government Western German Lander governments and local authorities "German Unity" Fund ERP Special Fund, specialized banks EC Federal labor office Pension insurance funds 0
1991 pe"
1992 pe*
81
109
8
12
31
24
21 4
25 4
25
30 14
— 170
218
(164)
(210)
(2) (4)
(4) (4)
-28 -3
-35 -3
139
180
53
74
8 31 21 1 25 —
12 24 25 1 30 14
pe = partly estimated. The (mathematically somewhat imprecise) figures are subject to considerable margins of uncertainty. ^Expenditure due to unification (1991 outturn, 1992 target), excluding expenditure on the CIS, and after deduction of the Lander governments' refunds to the "German Unity" Fund in respect of debt service payments and of payments from the EC budget, plus shortfalls in tax revenue owing to tax concessions. In 1992 including the transfer to finance the 1992 deficit of the federal labor office, which was included in the 1991 budget, including the waiver of structural assistance funds, including the advance pension payment made in December 1991. 'Breakdown of indirect taxes on the basis of the final consumption rather than the regional provenance. Source: Deutsche Bundesbank 1992, 16.
190
Wilhelm Nolling
ceeded its average level in the fourth quarter of 1991 by 8.8 percent. Monetary growth is thus still running well above the target corridor of 3.5 to 5.5 percent set for 1992. In 1992 the cost of living in West Germany was 4 percent higher than a year before, after the year-on-year inflation rate had peaked at 4.8 percent in March 1992. Adjustment to this new situation is evidently still incomplete, although it is certainly encouraging that the upward price trend in western Germany has slowed down of late. I feel that our inflation record, compared with historical levels, is not so bad, taking into consideration that in the past we have often experienced price increases around this 4 percent level and that unification proved to be much more costly than anticipated. So far German unification has not led to any serious disruptions in the capital and foreign exchange markets. The effective exchange rate of the deutsche mark (against eighteen of the main industrial nations) has improved by 4.8 percent by the end of 1992 compared with the rate in December 31, 1989. Long-term interest rates have fallen back to their preunification level (Table 7.2). Since September 1992 it has been encouraging to see that the Bundesbank is in a position to lower interest rates mostly in open market transactions in securities. I am certain we will pursue this policy in order to stimulate our economy and remove the burden of "excessively high" interest rates of the economies around us.19 The cost of living in the new federal states has continued to increase faster than in western Germany. In 1992 consumer prices were 12 percent up on the previous year. This is nothing compared with the price increases in other transforming eastern societies. The price increases in East Germany were mainly attributable to the first step - taken last October toward adjusting rents to market levels. In addition, on account of sharp wage increases, the prices of many other goods not exposed to supraregional competition, especially in the services sector, have risen at aboveaverage rates. As I mentioned before, the public sector deficit causes great concern. It is likely to go up from DM 110 billion last year to about DM 120-30 billion in 1992. It is estimated that Germany's total indebtedness will rise to DM 2.3 trillion by the end of 1995 compared with DM 929 billion in 1989. Beside these huge additional deficits, there has been a substantial increase in taxation and social security taxes (solidarity surcharge, indirect taxes). Despite these measures and a one percentage point increase of the value added tax at the beginning of 1993, the government is urgently 19
For the position of Germany in the world economy, see Nolling 1992c.
Table 7.2. Trend in interest rates since 1988 Interest rate
Overdraft loan Mortgage loan Bank discount Overnight money Three-month money Discount rate Lombard rate Repurchase rate Current yield on public sector bonds
Lowest rate
Highest rate
8.02 6.03 4.08 3.13 3.32 2.50 4.50 5.50
14.08 (Aug. '92) 11.09 10.42 (Aug. '92) 8.63 11.08 (Aug. '92) 8.16 9.72 (Aug. '92) 7.58 10.07 (Aug. '92) 8.25 8.75 (Jul.-Sep. '92) 6.00 9.75 (Dec. '91-Jul. '92) 8.00 9.72 (Aug. '92) 7.75
(Apr. '88) (Apr. '88) (Apr. '88) (Jan.'88) (Feb. '88) (until July '88) (until July '88) (Jan. '88)
5.70 (March'88)
9.10 (Jan.'91)
Jan. '90 Feb. '90* Dec. '90 Dec. '91 Dec. '92
7.90
11.24 9.30 8.27 7.77 8.26 6.00 8.00 7.85
11.97 9.94 8.65 8.43 9.20 6.00 8.50 8.60
12.95 10.21 10.02 9.23 9.67 8.00 9.75 9.28
13.66 9.71 10.41 8.91 9.03 8.25 9.50 8.79
8.60
9.00
8.50
7.30
Notes: Interest rates in %. Data collection completed January 29, 1993. Average monthly rates. Overdraft loans under DM 1 million. Mortgage loans for residential properties at floating interest rate. Bank discount loans: bills, rediscountable at the Bundesbank, up to DM 100,000. Repurchase rate: average monthly interest rate for security repurchase agreements with a term of one month. "Announcement of monetary union. Source: Deutsche Bundesbank, data bank.
192
Wilhelm Nolling
pleading for more revenues in order to be able to carry the financial burden of the former GDR. Needless to say, the economic scene in Germany continues to be characterized by big regional differences. After a strong growth surge at the beginning of the year, the pace of economic activity in western Germany has slowed down since the spring. In 1992 the real growth in gross domestic product will be only 1.5 percent, following 3.1 percent in 1991, and 4.5 percent in 1990.20 In the course of unification, Germany's current account showed an unprecedented deficit of about DM 40 billion in 1992, following a DM 33 billion deficit in 1991, whereas there had been a surplus of DM 108 billion in 1989. The swing between 1989 and 1992 amounts to approximately DM 150 billion. 7
Conclusions and prospects
In hindsight it is rather easy to identify political shortcomings and costly mistakes. There were a few economists and leading politicians around who realized from the very beginning that the task at hand was gigantic and most difficult to master. But it should also be mentioned that there is a lot of success clearly visible. The years ahead will continue to be difficult for the German economy and for monetary policy. We would be closer to experiencing the oftquoted "flourishing landscapes" (Chancellor Kohl) in eastern Germany if the following political mistakes and omissions had been avoided (although unification would have been a difficult task under any political circumstances). First, the time needed to accomplish the process of real economic adjustment was seriously underestimated, especially in 1990. Second, the government's economic policy makers were not willing or able to take full stock of the economic situation in eastern Germany. And such a critical "stock check" still has not been carried out up to the present day. And there was too much reliance on the working of the market forces, certainly not enough planning done to dismantle the command economy. Third, the government failed to use the mood of euphoria in Germany at the time of unification to implement significant income sacrifices in the West in order to promote an early upswing in the eastern part of Germany. This has inevitably resulted in unsound fiscal policies and a wrong policy 20
The forecast for 1993 predicts a real decline in gross domestic product of 0.5 percent. "Sachverstandigenrat zur Begutachtung der gesamtwirtschaftlichen Entwicklung" 1992, 17*.
Monetary policy after German unification
193
mix. German unification now has to be financed largely out of government borrowing. Fourth, the average conversion rate adopted at the time of the German Monetary Union was too generous, as the newly created money stock exceeded the supply capabilities of the German economy. Due to the excessive pressures on monetary policy, German interest rates are higher than they would have been if a balanced policy mix had been applied from the outset; however, our long-term interest rate is now below preunification levels. Fifth, the unwillingness to make required sacrifices in the West has contributed to exaggerated expectations in the East. Psychological factors ("the mental dividing walls") have led to unrealistic income expectations and to wage rises far beyond any gains in productivity. For this reason unemployment in eastern Germany is higher than it otherwise would have been. We are only now getting together to develop an income policy compatible with the economics of unification. Sixth, one of the major barriers to an economic upswing in eastern Germany can be traced back directly to the Unity Treaty and its provisions relating to property ownership. The principle of "restitution before compensation" has turned out to be an obstacle to investment. This list is by no means exhaustive. Due to the errors and omissions I have outlined - in particular those in the field of income policy - we are now (1993) facing the most difficult year yet in the process of German unification. According to the latest forecasts, unemployment in eastern and western Germany will rise significantly. The total public sector deficit will increase by a further DM 5 billion to DM 125 billion, or 4.3 percent of gross national product in 1993. Although experience shows that there have been shortcomings and policy mistakes, the fact remains that unification proved to be peaceful, has resulted in a remarkable increase of the standard of living for all parts of the East German population, and created hundreds of thousands of new jobs, especially in the service and construction industries. Thousands of firms have been successfully privatized. Due to the losses of jobs and uncertainty about the future, anxiety, frustration, and a sense of hopelessness exist in parts of East German society. I am certain that the policy measures already enacted and contemplated will turn the tide. However, it is even more difficult now to achieve better results because the world economy is deteriorating, which is above all also bad for our export-oriented society in West Germany. I understand that Korean interest in the German experience is great and very topical. My chapter can be read as a blueprint of how not to carry out unification. In addition, I stress the point that the "big brother"
194
Wilhelm Nolling
in Korea is not as big and wealthy as West Germany still is, and that its weaker brother in the north seems to be more debilitated than East Germany was and still is. In spite of all the economic burdens facing Germany during the unification process, we do not forget that - in historical terms - we are paying a small price for German unity. Who knows how much we would have promised to pay a mere three years ago in order to free 16 million Germans from their intellectual, physical, and monetary serfdom? At present Europe is deeply enmeshed in getting the Maastricht Treaty enacted. Whether it will become law is not certain. The outcomes of political developments in Denmark and Great Britain especially are difficult to forecast. In my view it is hard to understand, even irresponsible to aim for European Monetary Union in the face of the problems and the lessons to be learned from German unification.21
References Deutsche Bundesbank. (1990a). "Terms of the currency conversion in the German Democratic Republic on July 1, 1990." Monthly Report of the Deutsche Bundesbank (June). (1990b). "The monetary union with the German Democratic Republic." Monthly Report of the Deutsche Bundesbank (July). (1992). "Public financial transfers to eastern Germany in 1991 and 1992." Monthly Report of the Deutsche Bundesbank (March). Monthly Report of the Deutsche Bundesbank. Frankfurt am Main. Various issues. Deutscher Bundestag Bundestags-Drucksache 11/11. (1987). Materialien zum Bericht zur Lage der Nation im geteilten Deutschland. February. Deutsches Institut fur Wirtschafts-forschung (DIW). (1992). Wochenbericht 48/ 92. November 26. Dohnanyi, K. (1990). "Das Deutsche Wagnis." Munich. Lipschitz, L., McDonald, D. (eds.). (1990). "German unification - economic." IMF, Occasional Paper no. 75, December. Washington, D.C.: International Monetary Fund. Marshall, A. (1961). Principles of Economics. 8 ed. London: Macmillan for the Royal Economic Society. Nolling, W. (1990). "Wiedervereinigung - Chancen ohne Ende?" Hamburger Beitrdge zur Wirtschafts- und Wdhrungspolitik (February 28). (1991). "Geld und die deutsche Vereinigung." In Hamburger Beitrdge zur Wirtschafts- und Wdhrungspolitik in Europa 8 (July). (1992a). "Privatization and restructuring according to the Treuhandanstalt." Discussion Paper presented at a conference on Industrial Restructuring in Eastern Europe, organized by National Bureau of Economic Research, Inc., Cambridge, Massachusetts, February 29. 21
See Nolling 1993.
Monetary policy after German unification
195
(1992b). "Geld und die deutsche Vereinigung." In Deutsche Bundesbank, Ausztige aus Presseartikeln no. 7. (1992c). "Bundesbank monetary policy and the overall economic direction of Germany." In Deutsche Bundesbank, Auszuge aus Presseartikeln no. 44. (1993). Monetary Policy in Europe after Maastricht. Translated by Brian Rasmussen; foreword by Rudiger Dornbusch. New York: St. Martin's Press. Pohl, K. O. (1990). "Erklarung vor der Bundespressekonferenz, Bonn, February 9, 1990." In Deutsche Bundesbank, Auszuge aus Presseartikeln no. 1, February 12. Sachverstandigenrat zur Begutachtung der gesamtwirtschaftlichen Entwicklung. (1992). Jahresgutachten 1992-93. (October). Schumpeter, J. A. (1942). Capitalism, Socialism and Democracy. New York:
Harper.
PART III
Capital controls and macroeconomic policy in the Asia-Pacific region
CHAPTER
Capital movements, real asset speculation, and macroeconomic adjustment in Korea Yung Chul Park and Won-Am Park
1
Introduction
Domestic financial deregulation and capital account liberalization in advanced countries have contributed to a greater mobility of capital and financial integration of these economies for the past two decades. In contrast, however, developing countries have been slow and reluctant to remove various restrictions on international capital flows for fear that such a policy could undermine macroeconomic stability and weaken autonomy in the conduct of monetary and exchange rate policy. In this respect Korea is not an exception. Although Korea has been under strong pressures to liberalize its financial markets and capital movements by its major trading partners, including the United States, the Korean monetary authorities have maintained the position that capital market liberalization should be sequenced to stable macroeconomic conditions characterized by a strong or balanced current account, stable prices, and domestic interest rates. But the freemarket advocates see it differently, arguing that if capital market deregulation is not carried out, the preconditions for macroeconomic stability will never be met. The debate between the two sides continues to this day. Korea had maintained a system of extensive capital control until the early 1980s. Since then it has implemented an overall plan for economic liberalization. As a result, many argue that Korea has already achieved a great deal in loosening up the restrictions on capital movements (see Nam 1992). Recent studies, however, challenge this view by showing that the capital account remains as closed as it has ever been (Reisen and Yeches 1991). Why has it been so difficult for the Korean monetary authorities to deregulate capital account transactions? Could some of the adjustment difficulties following the capital account opening be identified and gauged empirically? Is there a less painful way of liberalizing the capital account? These are some of the questions this chapter addresses. 199
200
Yung Chul Park and Won-Am Park
To obtain a proper perspective on the progress Korea has made in opening up the capital account in recent years, Section 2 attempts to "measure" in terms of several indicators the openness of the capital account. Not surprisingly perhaps, we have not been able to establish whether Korea's capital account has become more open in recent years. Section 3 reviews recent economic developments and the role of real asset speculation in Korea. Section 4 conducts a number of counterfactual exercises that estimate the effects of capital account liberalization using a Keynesian open economy macroeconomic model (Korea Development Institute Model). This is followed in Section 5 by the development of another model where the markets for real assets such as land and nontradable goods play an important role in assessing the effects of capital inflow on growth, inflation, and the current account. Our discussion in these latter sections supports the view that such stable macroeconomic conditions as stable prices and current account balance are indeed prerequisites to a successful capital account liberalization. 2
Developments in the openness of Korea's capital account
With a sharp deterioration in the current account and the massive foreign debt in the early 1980s, the Korean monetary authorities tightened their controls over capital movement. During the first half of the 1980s, the control structure encouraged capital inflow but restricted outflow. But the emergence of a large surplus in the current account in the latter half of the 1980s made it easier for the monetary authorities to remove some of the restrictions on capital movement: As the large surplus in the current account complicated money supply management and created pressure for the appreciation of the won, the Korean government took measures to induce capital outflow, such as allowing individual acquisition of real estate abroad. Table 8.1 shows developments in Korea's capital account in the 1980s. On the liability side, changes in the long-term capital balance were caused primarily by changes in public loans: Public borrowings were the major source of capital inflow in the first half of the 1980s; they became the major source of capital outflows in the second half of the 1980s. Commercial borrowings also followed the same pattern. On the asset side, overseas direct investment has been the major source of capital outflow, amounting to almost 90 percent of the total net asset accumulation in 1991. Table 8.1 also presents two indicators of capital mobility. One is the value of capital flow as a percentage of total trade. This figure declined in the second half of the 1980s and was quite small in magnitude compared with those of other countries such as Taiwan and Indonesia. The second
Macroeconomic adjustment in Korea
201
Table 8.1. Koreas capital account (U.S. $ million) andmeasures of capital mobility (%)
Current account Capital account Long-term Liabilities Public loan Commercial loan Assets (increase, - ) Short-term Liabilities Assets (increase, - ) Capital flow/trade Long-term Short-term Two-way capital flow0 Long-term Short-term
1982-85
1986-90
$-2,232 1,596 1,702 2,250 981 -81 -540 -106 -133 27
$6,301 -1,807 -2,673 -1,771 -907 -741 -902 866
5.6% 5.8 1.0 0.34 0.35 0.28
1,043
-176 3.9% 3.4 1.2 0.07 0.15 0.28
| net capital balance | | capital inflow | + | capital outflow | Source: Bank of Korea, Balance of Payments Statistics, various issues.
indicator is an index of two-way trade in financial assets, which is characterized by trade in risky assets. This index for Korea was also low compared with the similar indexes for other countries including Taiwan (Jwa 1992) and declined in the second half of the 1980s. According to these indicators, capital mobility between Korea and other countries declined in the second half of the 1980s, supporting the view that there was little change in capital account control in Korea. However, it should be noted that the two indicators are hardly a reliable measure of capital mobility. In fact, they suffer from a crucial defect in that they grossly underestimate the actual value of capital movement. These indicators measure capital flow by taking the difference in the stocks at two points in time instead of recording all transactions taking place during a given period. To avoid this problem, we turn to more traditional measures of the openness of the capital account. Capital mobility or the openness of the capital account is often measured either by the degree of linkage between domestic saving and domestic investment or the degree of linkage between the domestic and international interest rates (the interest rate parity). Since the savings-
202
Yung Chul Park and Won-Am Park
(1985=100)
(%p.a)
130
35 real effective exchange rates 1>
real interest differentials (3-year moving average)2)
80'
•
••
•
M M , . M M , M M ,
'()
74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 YEAR 1) Using trade weights of four major trading partners, U.S, Japen Germany, and U.K. Higher values mean real appreciation. 2) (Curb market rate - PPI inflation) - (3-month Eurodollar rates - four countries' WPI inflation)
Figure 8.1. Real interest parity in Korea. investment linkage requires the satisfaction of auxiliary conditions in addition to the covered or uncovered interest rate parity (see Frankel 1989), we have concentrated on the interest rate parity. The theory of uncovered interest parity predicts that if capital flows freely and risks can be ignored, then capital flows equalize the expected rates of return on different assets denominated in different currencies. In order to examine the extent to which the uncovered interest parity holds in Korea, we have estimated time series for the real interest differentials and changes in the real exchange rate as shown in Figure 8.1. The quarterly data of the curb market interest rates in Korea, three-month Eurodollar rate, and domestic and foreign wholesale price index have been used. To concentrate on the longterm relationship, the real interest rates are moving-averaged across the previous three years. Figure 8.1 clearly shows that the real interest differential and real exchange rate move in opposite directions in the long run. That is, the high real interest rate in the Korean financial market relative to that in the international financial market causes a real depreciation. Thus, the Korean experience contradicts the theoretical implication that if capital flows are perfectly mobile, there would be a strong positive correlation between the long-run real interest rate differential and the real exchange rate. In-
Macroeconomic adjustment in Korea
203
stead, it supports the view that capital controls have been effective in Korea. However, the second half of the 1980s saw a closer positive relationship between the two, more so if the period between the last quarter of 1987 and the second quarter of 1989 is excluded. This could be interpreted as evidence showing that some progress has been made in liberalizing the capital account during the period. Rather than testing whether the uncovered interest rate parity hypothesis holds, Edwards and Khan (1985) and Haque and Montiel (1990) directly estimated the degree of openness of the capital account. Their approach is to assume that the domestic interest rates can be represented as a weighted average of the foreign interest rates adjusted to expected devaluation and the domestic interest rates under capital control and then estimate the weighing coefficients. Applying this approach, Reisen and Yeches (1991) show that the degree of capital mobility in Korea is low and declined gradually toward the end of the 1980s. This finding on Korea's openness is contrasted with the casual observation in Figure 8.1, which is based upon the uncovered real interest parity. More important, the results of Reisen and Yeches could change, depending upon how one specifies the counterfactual interest rate that would prevail in the absence of capital movement. Jwa (1992) followed the same approach as Reisen and Yeches and found that Korea's openness in the capital account has gradually increased from the early 1980s. 3
Recent economic developments and the role of asset speculation
3.1
Macroeconomy
Although Korea has made some substantial efforts to open its capital account in recent years, it is true that Korea still maintains extensive control over the capital movements. Only since 1992 have foreign investors been allowed to buy individual Korean stocks in the domestic market, subject to a ceiling. This somewhat delayed opening of the capital account could be attributable to Korea's economic development since 1986. During the 1986-88 period, Korea's economic growth was exceptional (see Table 8.2 and Table 8.3). With the change of some macroeconomic environments such as falling petroleum prices, appreciating Japanese currency, low interest rates in international financial markets, and the world economic boom, exports grew rapidly and GNP increased over 12 percent on average for the three consecutive years. In addition, the inflation rate was kept below 4 percent and current account surplus soared to U.S. $28.6 billion. In 1989, however, the Korean economy began to cool off and its growth
Table 8.2. Growth rates of major macroeconomic variables (% Final consumption expenditure Year GNP Total Household Government 81 82 83 84 85 86 87 88 89 90 91 92
5.9 7.2 12.6 9.3 7.0 12.9 13.0 12.4 6.8 9.3 8.4 4.7
4.9 5.6 8.2 6.6 6.3 8.4 8.1 9.7 10.7 10.1 9.3 6.8
4.8 6.5 9.1 7.6 6.4 8.0 8.3 9.8 11.0 10.4 9.3 6.4
5.7 1.0 3.4 1.5 5.6 10.8 6.9 9.4 9.7 8.9 9.4 8.8
Gross capital formation Total
Construction
Equipment
Exports
Imports
-4.1 10.4 17.8 10.9 4.7 12.0 16.5 13.4 16.9 18.3 16.1 -1.8
-7.6 18.4 24.8 7.4 4.9 3.1 14.0 13.8 18.5 29.1 11.0 -2.6
0.0 1.6 8.9 16.0 4.5 23.9 19.4 13.0 15.2 18.4 12.8 -0.8
15.1 4.5 19.2 7.9 4.5 26.1 21.6 12.5 -3.8 4.2 9.8 9.7
5.9 3.0 12.0 7.4 -0.6 17.8 19.4 12.8 16.3 14.4 17.5 1.9
Source: Bank of Korea, National Income of Korea, 1990. Bank of Korea, Economic Statistics Yearbook, various issues.
Macroeconomic adjustment in Korea
205
Table 8.3. Balance of payments Current account Total (mil $) 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992
-5,320.7 -4,646.0 -2,649.6 -1,606.0 -1,372.6 -887.4 4,617.0 9,853.9 14,160.7 5,054.6 -2,179.4 -8,727.7 -4,604.9
Exports (mil $)
Imports (won/$)
Capital account" (mil $)
17,214.0 20,670.8 20,879.2 23,203.9 26,334.6 26,441.5 33,913.2 46,243.8 59,648.2 61,408.7 63,123.6 69,581.5 75,103.7
21,598.1 24,299.1 23,473.6 24,967.4 27,370.5 26,460.5 29,707.3 38,584.8 48,202.8 56,811.5 65,127.2 76,561.3 77,302.1
3,801.0 2,759.6 1,233.9 2,163.9 1,309.5 513.3 -2,374.0 -5,842.8 -1,396.5 -3,302.2 3,881.2 4,227.0 7,760.2
Overall balance (mil $) -1,889.6 -2,297.0 -2,711.2 -384.4 -957.5 -1,254.5 1,699.5 5,202.1 12,175.2 2,453.1 -273.9 -3,740.8 4,754.5
Exchange rate (won/$) 659.9 700.5 748.8 795.5 827.4 890.2 861.4 792.3 684.1 679.6 716.4 760.8 788.4
"Sum of long-term and short-term capital balance. Source: Bank of Korea, Balance of Payment Statistics, various issues.
rate eased to 6.8 percent. Much of the deceleration could be attributed to declining export earnings. Seen in terms of total trade volume, Korea's exports fell as much as 4 percent. Slow economic growth in advanced industrial countries, which had been major markets for Korean exports, was one of the major causes of the slump. At the same time, Korean exporters were pricing themselves out of these markets as they were unable to control cost increases. Nominal wages in the manufacturing sector went up while labor productivity fell. The manufacturing unit labor cost jumped by more than 25 percent in 1989. Korean exporters were losing out to other competitors in Asia in many products, especially in labor-intensive ones. Despite these cost-push effects, it is interesting to note that consumer prices remained relatively stable and rose only by 5.7 percent for the year. The softening of crude oil markets and stable import prifces together with the declining export contributed to the moderating price increases. Toward the latter part of 1989, there emerged a consensus that without expansionary measures the decline in output growth might not be arrested. Price movements did not indicate any sign of acceleration, though they were somewhat erratic, and the current account position gave some room for domestic expansion. With this background of macroeconomic developments, Korean policy makers shifted their macroeconomic policy
206
Yung Chul Park and Won-Am Park
stance from maintaining price stability to advocating moderate expansion of internal demand. The internal demand expansion was expedient and it helped meet the target of housing construction pledged earlier during the 1987 presidential campaign. The zeal with which Korean policy makers promoted housing construction in 1990 was unsurpassed in Korea's recent history. Investment in residential building in real terms shot up 62 percent in 1990, on top of an 18.4 percent increase in fixed investment. With moderate increases in the unit labor cost and depreciating currency, Korea's exports began to surge again and recorded a 3 percent real growth in 1990. Export growth and internal demand expansion were strong enough to move the economy out of the mild slowdown and back onto the rapid growth path once again. Subsequently, the economy grew 9.0 percent in 1990 and 8.4 percent in 1991. In retrospect, this internal demand-led expansion was costly and became one of the major causes of Korea's economic problems in recent years. The upsurge in domestic demand in excess of the supply capacity resulted in a large current account deficit and inevitably generated inflationary forces from both the cost and demand sides. The current account deficit which was a little less than 1 percent of GNP in 1990 more than tripled to 3 percent in 1991 (or $8.73 billion). Consumer prices rose 8.6 percent in 1990, and almost 10 percent in 1991. 3.2
Asset markets speculation
As seen in Section 3.1, the Korean economy maintained robust growth throughout the latter half of 1980s. However, the high growth coupled with huge current account surplus from 1986 to 1988 and internal demand-led expansion in 1989-90 took a toll on the Korean economy and contributed to Korea's current economic problem. Throughout the latter half of 1980s, the asset markets in Korea were very unstable. With a record amount of current account surplus and double-digit growth rate in 1986-88, the savings rate rose from 29.1 percent in 1986 to 35.1 percent in 1989 and real interest rates began to fall. People looked for a way to invest their savings and the stock market became a favorite place. At first glance, economic fundamentals seemed to boost the stock price. Increasing exports and double-digit economic growth lifted the revenues and profits of major Korean firms and helped raise the stock price. Since returns in other financial instruments did not change significantly, investment in stocks gave relatively high yields and it became attractive to many investors (see Table 8.4 and Table 8.5). As more and more money was put into the stock market, investments in stocks became increasingly speculative and stock prices skyrocketed. As shown
Table 8.4. Interest rates
Year
Saving rate"
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992
23.1 22.7 24.2 27.6 29.4 29.1 32.8 36.2 38.1 35.3 36.0 36.3 34.9
Corporate bond yield* Nominal
Real'
Change of exchange rate*
Eurodollar rate'
U.S. prime rate
30.1 24.4 16.8 14.2 14.1 14.2 13.2 13.8 18.1 15.2 16.6 18.3 18.6
1.3 2.9 9.7 10.8 11.8 11.8 10.5 10.8 11.0 9.5 8.0 9.0 12.4
36.23 6.15 6.90 6.24 4.01 7.59 -3.24 -8.02 -13.66 -0.66 5.41 6.20 3.63
14.19 16.87 13.29 9.72 10.94 8.40 6.86 7.18 7.98 9.28 8.31 5.99 3.74
15.27 18.87 14.86 10.79 12.04 9.93 8.35 8.21 9.32 10.92 10.01 8.46 6.25
"Savings/GNP. *Less than 1 year. c Nominal rate - inflation rate (CPI). 'Won against U.S. dollar, end of period. 'London Interbank Offer Rates on U.S. dollar deposits, three-month. Source: IMF, International Financial Statistics, various issues. Bank of Korea, Monthly Bulletin, various issues.
Table 8.5. Capital gains in stock market and land markets Nominal GNP (A) Year (bil won) 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992
Consumer Stock price price index Wage index (%) (%) (%)
36,749.7 28.8 45,528.1 21.5 52,182.3 7.1 61,722.3 3.4 70,083.9 2.3 78,088.4 2.4 90,598.7 2.7 106,024.4 3.0 126,230.5 7.1 141,794.4 5.7 166,437.8 8.6 206,681.2 9.3 229,938.5 6.2
_ 20.2 14.6 11.9 7.7 10.2 9.2 11.3 19.1 24.6 20.0 16.9 —
-9.8 16.1 3.4 4.7 3.3 5.3 64.0 83.3 65.9 32.5 -18.7 -12.0 -10.6
Market value of listed stocks (bil won)
Capital gains from stocks B/A (%) (B)
2,526.6 2,959.1 3,000.5 3,489.7 5,148.5 6,570.4 11,994.2 26,172.2 64,543.7 95,476.8 79,019.7 73,117.8 84,712.0
406.8 100.0 9.0 293.1 272.9 4,205.1 9,991.2 17,273.6 20,976.7 -17,854.2 -8,850.2 -7,750.5
0.89 0.19 0.01 0.42 0.35 4.64 9.42 13.68 14.79 -10.73 -4.28 -3.37
Land price index (%) 11.68 7.51 5.40 18.50 13.20 7.00 7.30 14.67 27.47 31.97 20.58 12.78 -1.27
Market value of land (bil won)
Capital gains from land (C)
C/A (%)
92,138.6 99,868.5 106,013.3 128,361.2 147,471.5 159,270.1 172,610.5 201,955.4 268,183.3 372,323.6 461,370.0 520,374.5 513,765.7
7,729.9 6,144.8 22,347.9 19,110.3 11,798.6 13,340.4 29,344.9 66,227.9 104,140.3 89,046.4 59,004.5 -6,608.8
16.98 11.78 36.21 27.27 15.11 14.72 27.68 52.47 73.44 53.50 28.55 -2.87
Source: Bank of Korea, Economic Statistics Yearbook, various issues. Korea Research Institute for Human Settlements, Quarterly Construction Economic Review, various issues. Hak and Pyo (1992).
Macroeconomic adjustment in Korea
209
in Table 8.5, the stock price index quadrupled from 1985 to 1990 and capital gains in the stock market alone in 1988 amounted to 15 percent ofGNP. Another, more disturbing development since 1987 has been the skyrocketing land prices throughout the country (see Table 8.5). Over the four-year period beginning in 1987, officially recorded prices of land, on average, more than doubled. The capital gain from this land market in 1989 alone amounted to 70 percent of total GNR This massive increase in the cost of land and other real estate has added to inflationary pressures and magnified the economic disparities among different income classes and sectors of the economy. One could point to a number of developments, all of domestic origin, that have led to this rampant land speculation. During the presidential campaign in 1987, all candidates, including the incumbent, committed themselves to a large number of construction projects, ranging from local roads to large port facilities and massive investment for housing development. Therefore, regardless of the election outcome, it was clear that the new government would invest heavily in residential and other construction projects. This expectation generated a huge demand for all kinds of land sites and contributed speculative bubbles to many land price movements. In addition, the large increase in the current account surplus, much of which was converted into domestic liquidity, provided the fuel that set off the rampant land speculation in 1987. The effect of speculations in these two markets on the Korean economy and, in particular, on macroeconomic variables was enormous. Wealth effect on consumption began to appear and construction investment grew beyond its nominal rate. Inflation rate rose to 9.3 percent in 1990, from 3.0 percent in 1987, with an increasing domestic demand and current account deficit problems showed up in five years again. However, with a declining economic growth in 1989 and 1990, speculation in the stock market began to disappear. Especially, during the 199092 period, stock prices fell by almost 40 percent. Even though the Korean government attempted several times to keep the stock market prices from falling, it fell continuously through the period. It took higher cost to get rid of the speculation in the land market. The Korean government took a strong position to ease the land market speculation and, finally, in 1992, the land price fell by 1.3 percent for the first time in ten years. Under these circumstance, it was obvious that any infusion of foreign capital during the period would make the Korean asset market including stock market and real estate market more unstable and even boost more speculation. Korean monetary authorities then believed that, unless these two unstable Korean asset markets were settled down and the financial
210
Yung Chul Park and Won-Am Park
system was liberalized, any opening of the capital market should be delayed. An increase in capital inflows was believed to cause a monetary expansion, which would, in turn, strengthen inflationary pressures. 4
The effects of capital account liberalization in a Keynesian open macroeconomic model
This section conducts several counterfactual exercises designed to assess the effects of capital account liberalization in a Keynesian open macroeconomic model (see the appendix for the structure of the model). This model captures some of the salient features of more advanced industrializing countries and then adjudicates some anomalies found in the Southern Cone experiences in the opening of the capital account in the late 1970s and early 1980s. 4.1
The open economy macro model and the Southern Cone experiences
Dismantling capital controls involves both benefits and costs. The traditional welfare analysis of capital account liberalization focuses on the benefits from increased availability of foreign savings that is likely to be realized when the capital account is opened. It is similar to the analysis of welfare gains from trade liberalization. The more recent analysis stresses, however, the benefits from international portfolio diversification of risky assets. The costs range from the loss of policy autonomy to macroeconomic instability caused by capital flight and speculative inflow of foreign capital.1 Although it is widely recognized that developing countries would benefit a great deal from relatively free capital account transactions, many studies have also shown that macroeconomic stability is a precondition for any successful capital account liberalization. The relaxation of capital controls causes initially a real appreciation of domestic currency, which will be followed by a real depreciation in the long run. The real appreciation associated with the increased capital inflow results in a lower interest rate, higher growth, and a deficit in the current account. Keynesian models of inflation and external capital flows can easily explain the postliberalization phenomena of steep increases in the prices of nontraded goods relative to traded goods with a current-account deterioration and massive foreign reserve accumulation. The reasoning for this type of adjustment associated with capital inflow runs as follows (Khan 1
See Hanson (1992) for a recent survey.
Macroeconomic adjustment in Korea
211
and Zahler 1983, Obstfeld 1986). When a domestic interest rate in the period of preliberalization exceeds the depreciation-adjusted world rate, the domestic interest rate falls with the capital account opening. The fall in the domestic interest rate causes domestic demand to rise. The real exchange rate then starts to appreciate in the face of an excess demand, and as a result the current account deteriorates. As the deterioration in the current account leads to decreases in expenditure and wealth, the real appreciation may cease and revert to real depreciation in the medium to long term. At an early stage of capital account liberalization, the adjustment process is likely to be accompanied by a period of economic distress that may be characterized by real appreciation with a current-account deficit, higher inflation, and higher growth until the deflationary impact kicks in. Macroeconomic instability at the time of capital account opening will in all likelihood exacerbate and prolong the difficult adjustment process. In the asset markets, the fall in the domestic interest rate raises the demand for money. The private sector reaches a portfolio equilibrium by borrowing foreign exchange from abroad and selling it for domestic money. If the central bank intends to hold the exchange rate fixed, it must purchase foreign exchange and sell domestic money. Thus, the private borrowing results in an instantaneous rise in the domestic money supply. The experiences of the Southern Cone liberalization in the late 1970s and early 1980s are claimed to be at variance with the increased capital inflow predicted by standard open macroeconomy models. They in fact raise more fundamental questions as to whether domestic interest rates will fall in the wake of liberalization and whether capital flow will respond to interest differentials between home and foreign capital markets. The Southern Cone experiences show that the domestic interest rate can rise even after the capital account is opened, resulting in falling output and massive unemployment. To cite the Chilean case, the peso interest rates were extremely high in relation to those on dollar deposits and the London interbank offer rates (LIBOR) in the 1979-81 period. That is to say, although the Chilean authorities removed most of the controls on capital flows during the same period, the spread between the peso interest rates and LIBOR declined only slightly. A number of explanations have been given for the sustained interest differentials in Chile. One is that the devaluation risks associated with the real appreciation immediately after liberalization were reflected in the high domestic interest rates. In contrast, Sjaastad (1983) argues that the spread reflected the cost of arbitrage between domestic and foreign assets. Arbitrage required a number of costly transactions, because commercial banks
212
Yung Chul Park and Won-Am Park
could not arbitrage directly between domestic and foreign assets by, for instance, taking a position in foreign currency, due to the remaining controls on financial institutions. Another explanation proposed by Harberger (1985) was that the high (real) interest rates in Chile stemmed from the decapitalized state of the Chilean economy and inadequate supervision of banking operations. The decapitalization reflects an extremely low rate of saving and investment. The lack of supervision reflects a substantial expansion in "false demand" for credit, which is essentially the rolling over of bad loans. Edwards (1986) takes a more balanced view, that Chile's high interest rates might be explained by several factors such as the high world interest rates prevailing at that time, the risks related to expected devaluation and foreign debt repayments, and the significant reduction in the real liquidity in the monetary adjustment process. The interest rate differentials that were maintained after the liberalization reform in Chile did not induce as much capital inflow as they could have under different circumstances. This raises another question: What explains the massive inflow of foreign capital immediately after the capital account deregulation? Edwards (1986) points out two factors: One is the foreign investors' perception of the increased profitability of domestic investment after liberalization, and the other is the domestic investors' excess demand for foreign funds before liberalization. Our counterfactual exercises do not incorporate the channels that produced higher domestic interest rates after the capital account opening as in the Southern Cone countries. Our exercises only incorporate some of the more standard features of a model that may be relevant to the advanced industrializing countries.2 However, they could replicate aspects of the Southern Cone experiences. 4.2
Impact of the capital account liberalization in Korea
We now examine the effects of capital account liberalization in an open economy macro model in which an increased capital inflow will initially raise the domestic demand and be associated with real appreciation and current account deterioration. For the counterfactual exercise, this chapter uses an extension of the 2
Many studies on the impact of capital account liberalization suffer from the same problems as ours have. Condon, Corbo, and de Melo (1990) assessed the impact of capital inflows for Chile, but they obtained the standard results rather than the results unique to Chile, such as higher domestic interest rates and falling output. Khan and Zahler (1983) standardized the experiments on the impact of capital flows and considered the risk premiums from foreign debts. But they simply assumed that capital flows respond to the interest rate differentials.
Macroeconomic adjustment in Korea
213
Table 8.6. Impact of capital inflow under managed exchanged rate system {Vo point change from base run)
Capital inflow (US $ bil) GNP Consumption Investment Export Import CPI Land price index Foreign exchange rate Current account (US $ bil)
1987
1988
1989
1990
4.00 0.64 0.60 4.04 -1.11 1.47 1.72 5.15 -3.32 -0.39
4.00 0.16 0.57 3.80 -1.28 3.04 1.32 3.76 -6.87 -2.07
4.00 0.00 0.09 1.42 -0.46 1.45 0.10 0.19 -8.38 -3.35
4.00 -0.32 -0.30 -1.52 0.00 0.18 -0.85 -2.78 -8.76 -4.36
Note: The counterfactual exercise assumes that capital inflow increases by $4 billion each year under the managed floating system, i.e., capital movement is controlled and foreign exchange rate adjusts only partially according to the overall balance.
Korean Development Institute (KDI) Quarterly Macroeconomic model that captures the channels described in Section 4.1. The basic features of the model are summarized in the appendix. To focus attention on the role of real assets, we incorporate a market fundamentals equation for the land price. And we also include real land value and the rate of change in land price as explanatory variables in the private consumption and fixed investment equation, respectively. Since the counterfactual exercises on capital inflows or capital account liberalization differ according to the extent to which capital account is liberalized and how the exchange rate regime is managed, we carry out the following two exercises. In the first exercise we assume that the Korean foreign exchange rate regime remains unchanged with capital inflow. The Korean government still adopts a managed floating system under which the amount of capital inflow is controlled and the foreign exchange rate is only partially adjusted depending on the direction of change in current account or overall account. In the second case we assume the perfect capital mobility and a flexible exchange rate regime. In the first exercise it is assumed that U.S. $4 billion of foreign capital flow into Korea every year from 1987 to 1990 and the simulation results are given in Table 8.6. The capital inflow appreciates the won-dollar exchange rate and increases domestic money supply. Appreciation of the Korean currency reduces the exports and raises the imports, resulting in a deterioration of the current account and a fall in GNP growth. This
214
Yung Chul Park and Won-Am Park
deflationary effect is, however, substantially offset in earlier years by the increase in the money supply, which expands internal demand. This is the reason that consumer price index (CPI) rises before falling in 1990.3 As seen in Table 8.6, the won-dollar exchange rate appreciates by 3.4 percentage points from the base run and GNP rises by 0.64 percentage point in the first year. CPI and land price index rise by 1.7 and 5.5 percentage points, respectively, and the current account deteriorates by $0.4 billion in the first year. In the four years of medium term, the foreign exchange rate appreciates approximately by 7 percentage points, on average, and the current account worsens as much as by $2.5 billion. GNP growth becomes negligible and CPI and land price index rise by 0.5 and 2 percentage points, respectively. In summary, this exercise shows that sustained capital inflows promote economic growth only in the short run at the cost of higher inflation and current account deterioration and there is no gain in GNP from capital inflows in the long run. All of these developments are expected in the standard Keynesian model with a managed floating system. According to the model, capital inflows appreciate the nominal exchange rate (won-dollar) at first. However, because the foreign exchange rate adjusts only partially, domestic money supply increases, resulting in an expansion of aggregate demand in the end. With the appreciating foreign exchange rate and rising domestic price level, terms of trade deteriorate. As a result, the current account surplus (deficit) begins to decrease (increase) with falling exports and rising imports and aggregate demand also falls. The second exercise shows the simulation result of the case of the complete decontrol of capital flows and the switch to the flexible exchange rate regime in the first quarter of 1987. Under the flexible exchange rate regime, the overall balance is in equilibrium by definition so that the net foreign assets of domestic banks are assumed to stay constant and the domestic money supply does not change. Therefore, after the regime changes, the foreign exchange rate adjusts rapidly and the current account changes. To make this simulation more general, we consider the case in which extra capital flows into Korea as the Korean capital market opens. As indicated previously, it is reasonable to expect that foreign capital flows into Korea with the capital market opening because the domestic interest rate in Korea is higher than the world interest rate in 1987 (see Table 8.4). The more 3
In our model, it is assumed that the won-dollar exchange rate (ER) is adjusted on the basis of the exchange rate a period earlier and the size of overall balance (OB). For our simulation, we estimated the following equation. ERt = ao + a] ERti
+ a2 OB,
Macroeconomic adjustment in Korea
215
Table 8.7. Perfect capital mobility and flexible exchange rate regime (% point change from base run)
Capital inflow (US $ bil) GNP Consumption Investment Export Import CPI Foreign exchange rate Land price index Wage Interest rate Current account (US $ bil)
Casel
Case 2
Case 3
Case 4
0.00 -5.38 -0.90 -2.80 -9.50 6.01 -0.38 -18.40 -1.12 -1.73 -0.16 -0.54
1.00 -11.25 -1.82 -2.74 -18.30 16.41 -0.38 -32.78 -1.12 -2.84 -0.39 -1.54
2.00 -16.54 -2.72 -3.63 -25.90 26.64 -0.38 -43.89 -1.12 -3.92 -0.61 -2.54
3.00 -21.51 -3.59 -4.49 -32.55 36.70 -0.38 -52.76 -1.12 -4.96 -0.82 -3.54
Note: The counterfactual exercise assumes a switch to the flexible exchange rate system under perfect capital mobility.
capital flows into Korea, the more the foreign exchange rate will appreciate, and, finally, both the current account and GNP growth will worsen. Table 8.7 shows simulated values of key macroeconomic variables after the capital market opening. The first column is the simulation result after the regime changes without any new capital inflows and the second through the fourth columns are those with extra capital inflows. As seen in the Table 8.7, without extra capital inflows, the Korean currency sharply appreciates by 18.0 percentage points and terms of trade significantly deteriorate with the opening of the Korean capital market. Current account deteriorates by $0.5 billion and GNP falls by 5.4 percentage points with falling exports and growing imports. In short, the Korean economy falls into a serious recession. Moreover, if there are extra capital inflows, the recession becomes more severe, as seen in the second through fourth columns of Table 8.7. In our model, a land price equation is introduced as a function of money stock (m2) and inflation rate in accordance with a market fundamentals approach. Under the managed floating system, land prices rise with the increases in capital inflow or money supply and the rising land price brings inflation. As the regime changes, the economy experiences a serious recession and land price declines. Since Korea is a small country that heavily depends on external trade,
216
Yung Chul Park and Won-Am Park
a sudden movement to a flexible exchange rate system or an opening of its capital market would have an enormous effect on its economy. In general, capital inflows would cause a sharp currency appreciation and seriously disturb the current account and GNP growth. The capital inflows, if it is directed to real asset markets as it was during the 1987-89 period in Korea, could easily set off speculation on the real asset market. Unfortunately, we have not been able to incorporate any real asset speculation in our simulation because of the difficulty of formulating a process of expectation formation of future land prices. With real asset speculation, however, the real asset price is likely to jump with capital inflow. The jump may in turn trigger speculation on real assets including land that will have a serious repercussion throughout the economy, as it will increase other prices and also add to inflationary expectation. We explain this possible development in a simple general equilibrium model in the next section. 5
Models for capital movements and real asset speculation
The counterfactual exercises in terms of a Keynesian open macroeconomic model in the preceding section indicate that an exogenous increase in capital inflow could deepen some of the maladies that the Korean economy has suffered from in recent years. However, the model cannot capture the effects of an increased capital inflow on key macroeconomic variables vis-a-vis its effects on the real asset markets, in particular the speculation on real estate, which has played a pivotal role in determining the duration and extent of cyclical fluctuations in Korea. In what follows, it will be shown that within a general equilibrium model a nominal exchange rate appreciation associated with an increase in capital inflow could push up the prices of real estate - land sites and commercial and residential buildings - and nontradable goods. In a country where real asset speculation has been an intermittent but chronic feature of the economy, the price increase could easily spark off a new round of real asset speculation, which could be expected to last on average three or four years as it did during the 1988-91 period. The real asset boom, once it generates the expectation that real asset prices will continue to rise for a period ranging from three to four years, could offset the effects of capital inflow on domestic interest rates: Real interest rates will remain high because real asset speculation increases the demand for credit and money. In response to the increase in the prices of nontradables and real estate, resources will shift to this sector from tradable ones. Nominal wages in the nontradable and construction industries will climb rapidly, pushing up the wages also in the tradable goods sector. Because of the expansion in
Macroeconomic adjustment in Korea
217
the nontradable sector, GNP growth will be higher than otherwise, but prices of both nontradables and tradables will continue to increase as long as the real asset boom continues. After a while, the Keynesian adjustment process described in the preceding section will set in, and eventually the economy will slow down and the real asset price increase will also decelerate. During this process of adjustment, it is possible that continuing inflow of foreign capital in response to real interest rate differential could feed on the speculative bubble in the real asset market and as a result prolong the adjustment process and in the end intensify the deflationary effect when the bubble in the real asset market bursts. We will first sketch a simple fundamental valuation model for land prices that will help explain the adjustment process and then discuss a simple two-sector general equilibrium portfolio balance model focusing on stabilizing real asset speculation. Of course, the land price movements in Korea exhibited properties of speculation to some extent. However, the current theories on the speculative bubble do not yet provide convincing explanations for how the speculative bubbles get started and why they collapse. To make the bubble story more convincing, one must explain not only short-term movements that may exhibit speculative properties but also medium- to long-term movements to the steady state. For these reasons, we focus on stabilizing speculation in the following. This discussion may explain in part why the Korean monetary authorities have been reluctant to open up the capital account in recent years, despite the mounting foreign pressures for financial liberalization. 5.7
Fundamentals valuation equation
We intend to explain the high prices of Korean land and stocks on the basis of the fundamentals valuation equation. As early as in 1987 the change in the fundamentals of the land price determinants indicated that land prices would skyrocket in the coming years. Although the danger signs were clearly visible, Korean planners went ahead with their ambitious housing construction plan, which, as expected, sparked off land speculation. As a first approximation, the land price-rental ratio R may be defined as R = —*—, (1) r - gr where r is the real interest rate of the economy and gr is the expected rate of growth of rental earnings. During 1985-88, the ratio increased almost three times, largely because of the double-digit growth. The sharp increase
218
Yung Chul Park and Won-Am Park
in R foretold the impending surge in the real asset prices, but it was largely ignored. Once the speculation started, monetary and fiscal policies were hardly effective in curbing the real asset boom. Although this formula can explain sharp changes in the price of land, the fundamentals valuation equation does not seem to explain the sustained rise in Korean land prices since 1989. As GNP growth slowed down sharply in 1989, expectations for future economic growth became lower but real interest rates did not fall. Thus the developments in fundamentals since 1989 do not support the continued surge in land prices even after 1989. The same problem can be found in the application of the fundamentals valuation equation to Japanese land prices. Although Frankel (1991) applied the fundamentals formula to explain the recent rise in Japanese land prices, he could not depend too much on this formula because expectations for future economic growth have become lower and real interest rates higher in recent years. 5.2
Two-sector general equilibrium portfolio balance model
We now present a very simple model for Korea's land prices and current account that captures the linkages among capital inflow, real asset speculation, and the current account. For the sake of simplicity, the model assumes full employment, a fixed exchange rate (or a crawling peg), and purchasing power parity. We start with the case that the public holds only two assets: domestic money and real estate. This is the case in Korea where foreign exchange is concentrated in the vault of the Central Bank. The economy is assumed to produce both traded and nontraded (home) goods. Given factor endowments and technology, the production of both goods depends only on their relative price, q:
where E = nominal exchange rate, P1* = foreign price of traded goods, and PH = price of home goods. The small country faces a given foreign price of traded goods P7*, which is assumed to be unity (thus foreign inflation is ignored as usual). The output of both goods is specified as dY IT = rfo); — > 0, (3a) dq dYH Y» = Y»{q)' — < 0, (3b) dq where Y7 and YH denote the output of traded and nontraded goods, re-
Macroeconomic adjustment in Korea
219
spectively. The demand for two goods is assumed to depend on the relative price of traded goods, q, and the real wealth of the public, a: CT = CT(q, a);^<0,~-> dq da
0,
(4a)
C" = C ' % a); ^ - > 0, — > 0, (4b) oq da where C7 and CH denote the demand for traded and nontraded goods, respectively. The real wealth of the public measured in units of tradable goods is composed of domestic money (A/), and real estate (Z) whose physical stock is fixed and whose price is denoted by Pz: E
E
We note that we measure the real value of wealth in terms of traded goods in order to provide a simple analysis with the qualitative conclusions unaffected. Equilibrium in the market for nontraded goods is Y» = C"(q, a), from which the following holds: dYH _ dCH a
=
V(gy9
v' =
dq
dq dCH
< 0.
(6)
Equation (6) exhibits a negative relationship between the real wealth of the public and the relative price of traded goods. An increase in the relative price of traded goods induces an incipient excess demand for nontraded goods, requiring a decline in the real value of wealth to attain an equilibrium in the home goods market. It is assumed that domestic money stock changes only with changes in the stock of the international reserves or changes in the domestic currency value of an excess demand or supply of traded goods: M=E(YT-C7)
= EMlfa
< 0>
(7)
where a dot denotes the rate of change in a variable with respect to time. From (6) and (7) we get:
220
Yung Chul Park and Won-Am Park
where V~l denotes the inverse function of V, if one assumes its monotonicity. Thus, an increase in the real wealth of the public measured in units of tradable goods reduces the rate of domestic money accumulation. We assume that the desired ratio of the irreproducible real estate to domestic money depends on the expected difference between the rates of return on the two assets. Since the expected real rates of return on real estate and domestic money are Pz — P and -P, respectively, where P denotes general inflation rate (a circumflex denotes the percentage change in a variable), the difference is Pz. Notice that E is fixed so that PZ7 — = H(P*); H' > 0, M
(9)
or 9
— = (Up); G' > 0, (9') m where p = PZIE and m = MIE. The system can be represented by a set of state variables m and p. Figure 8.2 illustrated the phase diagram for two state variables of real balance of domestic money and real estate premium. When m = 0, equation (7) determines the unique steady-state value of real wealth at a. We know from (5) that a = m + pz. Thus, the m = 0 locus is downward sloping in the (m, p) plane. The p = 0 locus should be upward sloping from (9'). Therefore, the saddle path is also upward sloping in Figure 8.2. According to Figure 8.2, the capital inflow occasioned by interest rate differentials immediately increases the real balance of the public from mQ to mv The premium on real estate must rise immediately. The incipient increase in the value of real wealth caused by the foreign capital inflow brings about real appreciation and current account deterioration according to equation (6). Afterward, the economy moves to the steady-state equilibrium with real depreciation and decreases in real estate premiums. We can also explain why currency appreciation bids up land prices. In Figure 8.2, the impact of the exogenous increases in money is the same as that of won appreciation, as both induce increases in real balance. It seems to us that the recent rise in Korea's land prices could be attributable mostly to currency appreciation and excess liquidity due to the current account surplus rather than the capital inflow: The real appreciation has caused the prices of the domestic commodity and the real estate. As shown in Table 8.5, the CPI inflation rate continued to increase and land prices shot up along with real appreciation since 1987. We now turn to a more complicated case where the public holds domestic and foreign money as well as real estate. The currency substitution is
Macroeconomic adjustment in Korea
221
a 2
Figure 8.2. Impact of capital inflow and currency appreciation on the real exchange rate and the price of real estate without currency substitution.
certainly not the case in Korea but it could be the case for countries where the capital account is to a larger extent liberalized. In this model of currency substitution the real wealth of the public is M
(10)
The stock of foreign money changes also according to an excess demand or supply of tradable goods:
F= yr-c7=M'Ja
(ii)
Our assumption about currency substitution is that the ratio of the domestic currency value of foreign to domestic money depends on the difference between the expected rates of return on the two moneys, which reads E, as these are E - P and -P, respectively:
222
Yung Chul Park and Won-Am Park EF — = L(E); U > 0. M
(12)
From (9) and (12), it follows that PZZ _ HjP2) EF L(E) Finally, to keep the model tractable, we specify the functional form of H and L in (9) and (12). This specification comes for the purpose that the desired ratio of real estate to foreign money also depends on the expected difference between the rates of return on the assets. Let the functional form be: z
, L{E) = C2e\
where both Cx and C2 are positive constants. Then, PZ7 ^ EF
r
-
C C2
hi ePZ-E
and so that PZZ = J(PZ - E); f > 0 EF
(13)
Since E is assumed to be exogenously controlled, the desired ratio of domestic money to foreign money is also exogenously fixed, which we denote as a. Therefore, the real wealth in this model can be transformed into: a
M = _ + d
F
+
PZ7
= (1 + a ) F + A,
(10')
Hi
where h = PZZIE = P Z The system can be represented by either a set of state variables of F and q or a set of F and h. Figure 8.3 illustrates the phase diagram for two sets of state variables. In the (F, q) space, the F = 0 locus should be horizontal at q according to (6) and (11). The slope of the q = 0 locus is not determined, but it does not affect the negativity of slope of the saddle path to the equilibrium. In the (F, h) space, the F = 0 locus should be downward sloping, since q and a are fixed in the steady state in equation (6) and (11). The h = 0 locus is upward sloping from (13). Thus the saddle path of (F, h) must be upward sloping.4 The increased capital inflow is associated with the increase in the foreign money stock from Fo to F{ (thereby increasing the domestic money 4
For the derivation in detail, see Park (1987a).
\
A
cx \
B
B' ^
F,
(a)
fi-0
F, (b) Figure 8.3. Impact of capital inflow and appreciation on the real exchange rate and price of real estate with currency substitution.
224 Yung Chul Park and Won-Am Park supply as long as exchange rates are controlled) in Figure 8.3. With increased capital inflow the economy shifts immediately to the point B in which real exchange rates appreciate and real estate premiums are raised. A decline in the rate of home currency depreciation through disinflation (or nominal appreciation) will shift the q = 0 locus downward in Figure 8.3(a) and will rotate the F = 0 locus downward in Figure 8.3(b), resulting in real appreciation and a decline in the real value of real estate. When the rate of domestic currency depreciation is reduced, people want to hold more real domestic money by substituting out of foreign money and real estate in real terms. This is possible with real appreciation and a decline in the real value of real estate. Since the nominal exchange rate will not change instantly, however, real appreciation will be achieved by increases in both the domestic commodity price and the price of real estate. The major difference between the case with currency substitution and without currency substitution is that the real estate premium undershoots the steady-state level with currency substitution immediately after a decline in the rate of home currency depreciation (or nominal appreciation).
6
Concluding remarks
This chapter has examined two major topics: the developments in the openness of Korea's capital account in the 1980s and the effects of an increase in foreign capital inflow on domestic economy. As for the first topic, we have been unable to decide conclusively whether Korea's capital account has become more open in recent years since the assessment differs according to the approach one takes. Our experiments of the effects of foreign capital inflow on the Korean economy through the simulation and theoretical models, if one keeps in mind the Southern Cone experiences in the very late 1970s, show that an increased capital inflow would produce a higher growth at the cost of higher inflation and current account deterioration with real appreciation of domestic currency. The exercises also show that the switch to the flexible exchange rate system could invite the vicious spiral among exchange rates, prices, and market interest rates, resulting in economic instabilities. An interesting thing is that the market interest rates would not fall rapidly to the international level with capital account opening. With regard to the real asset speculation in Korea, it was shown that the disequilibriums in the domestic and external markets could be more severe with speculative capital inflows and the consequent jump in the real estate premiums. It seems to us that the real asset boom in the second half
Macroeconomic adjustment in Korea
225
of 1980s was mainly caused by won appreciation and the huge trade surplus. In sum, our study suggests that one cannot be overcautious about the opening of capital account, especially when the Korean domestic economic conditions are not stable, characterized, for example, by real asset speculation and a massive deficit in the current account. Appendix The KDI Quarterly Macroeconomic Model was developed by Won-Am Park (1987b) and later modified and updated by Yoon-ha Yoo (1990). A major focus of the model is to interrelate the real and financial sectors of the Korean economy. Thus, the model incorporates credit availability to firms for investment, includes money as a determinant of consumption, and emphasizes links between the monetary sector and the balance of payments. The model was estimated using quarterly data from the first quarter of 1973 to the last quarter of 1991. Seasonal dummies were included and, where appropriate, the estimation was corrected for serial correlation. The model consists of six blocks of equations: the GNP, government sector, labor market, wages and prices, the balance of payments, and financial sector. Real gross national expenditure is composed of private consumption expenditure, private fixed investment, inventory investment, government expenditure, exports and imports of commodities and nonfactor services, and net factor income from abroad. Real GNP is divided into two components: production from agriculture, forestry, and fisheries, and other sectors. The supply and demand for money are determined in the financial block, where interest rates such as corporate bond yield adjust to equilibrate the market. The overall balance of payments and the government budget deficit are both linked to the money supply. Prices are subject to both demand-pull and cost-push factors. Wholesale prices are determined by firm's production costs. The unit value index for exports in dollar terms is assumed to be influenced by the won's exchange rate vis-a-vis the U.S. dollar as well as by export production costs (wages and intermediate input costs). Import unit values are determined by the import prices of capital goods and raw materials, including oil. The wage equation is an expectations-augmented Phillips curve. Finally, the unemployment rate is determined by the gap between potential and actual output, a variant of the Okun's law. To draw attention on the role of real asset speculation, the market fundamentals equation for the land price is formulated and real land value
226
Yung Chul Park and Won-Am Park
and land price inflation are incorporated in the private consumption and fixed investment equation, respectively. The consumer price equation also includes the land price. References Bank of Korea. (1990). National Income of Korea. Seoul. Balance of Payment Statistics. Seoul. Various issues. Economic Statistics Yearbook. Seoul. Various issues. Monthly bulletin. Seoul. Various issues. Condon, T , V. Corbo, and J. de Melo. (1990). "Exchange Rate-Based Disinflation, Wage Rigidity, and Capital Inflows: Tradeoffs for Chile 1977-81." Journal of Development Economics 32: 113-31. Edwards, S. (1986). "Monetarism in Chile, 1973-1983: Some Economic Puzzles." Economic Development and Cultural Change 34 (April): 423-671. Edwards, S., and M. Khan. (1985). "Interest Rate Determination in Developing Countries: A Conceptual Framework." IMF Staff Papers, 32, no. 3: 377-403. Frankel, J. (1989). "Quantifying International Capital Mobility in the 1980s." Cambridge, Mass.: NBER Working Paper no. 2856, February. (1991). "Japanese Finance in the 1980's: A Survey." In P. Krugman (ed.), Trade with Japan: Has the Door Opened Wider? Chicago: University of Chicago Press, 225-68. Hak, K., and K. Pyo. (1992). "A Synthetic Estimate of the National Wealth of Korea, 1953-1990." Unpublished manuscript. Seoul: Korean Development Institute, February. Hanson, J. A. (1992). "An Open Capital Account: A Brief Survey of the Issues and the Results." Paper presented at The Impact of Financial Reform at the World Bank, Washington, D.C., April 2-3. Haque, N., and P. Montiel. (1990). "Capital Mobility in Developing Countries: Some Empirical Tests." IMF Working Paper no. 117. Washington, D.C.: International Monetary Fund. Harberger, A. C. (1985). "Observations on the Chilean Economy, 1973-1983." Economic Development and Cultural Change 33 (April): 451-62. International Monetary Fund. International Financial Statistics. Washington, D.C.: Various issues. Jwa, S. (1992). "Capital Mobility in Korea since the Early 1980s: Comparison with Japan, Taiwan and Indonesia." Unpublished manuscript, Korea Development Institute. Khan, M. S., and R. Zahler. (1983). "The Macroeconomic Effects of Changes in Barriers to Trade and Capital Flows: A Simulation Analysis." IMF Staff Papers 30 (March): 223-82. Korean Research Institute for Human Settlements. Quarterly Construction Economic Review. Seoul. Various issues. McKibbin, W. J., and J. Sachs. (1989). "The McKibbin-Sachs Global Model: Theory and Application." Cambridge, Mass.: NBER Working Paper no. 3100, September. Nam, S. (1992). "Korea's Financial Reform since the Early 1980s." Seoul: KDI Working Paper no. 9207. Obstfeld, M. (1986). "Capital Inflows, the Current Account, and the Real Ex-
Macroeconomic adjustment in Korea
227
change Rate: The Consequences of Stabilization and Liberalization." In S. Edwards and L. Ahamed (eds.), Economic Adjustment and Exchange Rates in Developing Countries. Chicago: University of Chicago Press, 201-29. Park, W. (1987a). "Crawling Peg, Inflation Hedges, and Exchange Rate Dynamics." Journal of International Economics 23: 131-50. (1987b). "A Quarterly Macroeconometric Model for the Korean Economy." Seoul: KDI Working Paper no. 8716, December. Reisen, H., and H. Yeches. (1991). "Time-Varying Estimates on the Openness of the Capital Account in Korea and Taiwan." Paris: OECD Development Centre Technical Paper no. 42, August. Sjaastad, L. A. (1983). "Failure of Economic Liberalism in the Cone of Latin America." World Economy 6 (March): 5-26. Yoo, Y. (1990). "The KDI Quarterly Model of the Korean Economy." Seoul: KDI Working Paper no. 9104, August.
CHAPTER 9
The determinants of capital controls and their effects on trade balance during the period of capital market liberalization in Japan Shin-ichi Fukuda
1
Introduction
The purpose of this essay is to investigate the determinants of capital controls and their effects on trade balance during the period of capital market liberalization in Japan. As is well known, the capital market in Japan was gradually liberalized in the 1970s. However, the liberalization was sometimes accompanied by the introduction of other capital controls. This chapter examines how these capital controls were liberalized and what distributional effects they had on the trade balance during the period of capital market liberalization in Japan. In the analysis, we measure the degree of capital liberalization by deviations from covered interest rate parity. If the capital market is perfectly liberalized, this covered interest rate parity must always hold. In fact, after December 1980 (the year of the enactment of the new Foreign Exchange and Foreign Trade Control Law in Japan), deviations from covered interest rate parity became negligible in Japan. However, deviations were very significant for several periods in the 1970s. Thus, we can approximately measure the degree of Japanese capital controls by using these deviations in the 1970s. Several empirical studies have investigated the determinants of Japanese capital controls in the 1970s (e.g., Ueda and Fujii 1986, Fukao 1990, Komiya and Suda 1991). This essay first reconfirms the results of these The author is grateful for helpful comments by Eui-Gak Hwang, Hiroo Taguchi, Sebastian Edwards (editor), and other participants of the conference. The author also thanks the seminar participants at Bank of Japan and Hitotsubashi University for their comments to the earlier version of this essay. This research is supported by the Ministry of Education of Japan.
229
230
Shin-ichi Fukuda
previous studies, which propose that the degree of capital controls in Japan was determined by the excess deficit and surplus in trade balance in the 1970s. The chapter also empirically shows that the degree of capital controls significantly changed the effects of interest rate differentials on trade balance and current account. We then investigate what distributional effects these capital controls had on trade balance. In our investigation, we pay special attention to the role of Japan's Multinational Trading Companies (MTCs), which are called the Sogo Shosha in Japan. We show that the tightening of capital controls raised the trade volume of MTCs and changed the weight of exports to Southeast Asian countries in the 1970s. In Japan, there was the principle of "actual demand" in the foreign exchange market before the capital market liberalization. Under this principle, participants in the forward exchange market were strongly restricted. However, trading companies could access the forward exchange market even when capital controls were tightened. In addition, they could circumvent various capital controls through leads and lags, that is, by the actions of traders to adjust the length of payment periods (see Komiya and Suda 1991). Hence, the trading companies could exploit a big arbitrage opportunity when there were significant deviations from covered interest parity. In particular, MTCs had an advantage in exploiting the profit because they are not only trading companies but also financial intermediaries among companies (see Drucker 1975). In addition, MTCs could handle large amounts of exports and imports at the same time (see, e.g., Yamazawa and Kohama 1985), which gave them more chances to use leads and lags. Therefore, it is very important to note the role of MTCs under capital controls and its effects on the Japanese trading structure during the period of capital market liberalization in Japan. The chapter proceeds as follows. The next section briefly reviews history of capital controls in Japan. Section 3 measures the degree of capital controls by using deviations from covered interest parity. Section 4 examines the determinants of capital controls in the 1970s. Sections 5 investigates the effects of these capital controls on trade balance in Japan. Section 6 explores the role of MTCs under capital controls. Section 7 summarizes our main results and refers to their implications. 2
A brief history of Japanese capital controls
Deregulations of Japanese capital controls began in the first half of the 1970s and were basically completed by the new Foreign Exchange and Foreign Trade Control Law of December 1980 (see Table 9.1 for a more detailed chronology of Japanese capital controls). Before the early 1970s,
Table 9.1. A selective history of Japanese capital controls, 1970-present April 1970 January 1971 July 1971 February 1972 March 1972
May 1972 October 1972 November 1973 December 1973 January 1974
August 1974 June 1976 November 1976 April 1977 June 1977 March 1978
July 1978 October 1980 December 1980
(D) investment trusts permitted to purchase foreign securities (with an upper limit of $100 million) (D) insurance companies permitted to purchase listed foreign securities (up to $100 million) (D) securities companies permitted to purchase and sell listed foreign securities on behalf of individual investors (D) liberalization of trust banks to purchase and sell listed foreign securities (D) major foreign exchange banks permitted to purchase and sell listed foreign securities (R) foreign investors prohibited from acquiring short-term government bills (D) purchases of unlisted foreign securities by residents liberalized (R) medium- and long-term loans subject to blanket approval (D) regulation on purchases of Japanese stocks by nonresidents abolished (D) regulation on purchases of Japanese secondary bonds by nonresidents abolished (R) residents' purchases of short-term foreign currency securities prohibited (R) foreign exchange banks, securities companies, insurance companies, and investment trusts requested to refrain voluntarily from increasing the net value of foreign securities investments (relaxed in June 1975) (D) purchases of Japanese primary bonds by nonresidents liberalized (D) purchases of foreign securities by residents subject to automatic approval (D) regulation on medium- and long-term loans relaxed (D) purchases of unlisted bonds liberalized (D) purchases of Japanese primary bonds by nonresidents subject to automatic approval (R) purchases, by foreign investors, of yen-denominated domestic bonds with remaining maturities of less than five years and one month prohibited (until February 1979) (R) foreign currency-denominated medium- and long-term loans by major foreign exchange banks subject to blanket approval (D) yen-denominated medium- and long-term loans liberalized the enactment of the New Foreign Exchange and Foreign Trade Control Law (D) Japanese companies and individuals can invest in foreign securities without security firm's intermediation (D) liberalization of foreign loans by Japanese (D) nonresidents can purchase and sell Japanese securities without any licensing (D) nonresidents can issue bonds in Japan with only prior reporting (continued on following page)
232
Shin-ichi Fukuda
Table 9.1. (cont.) December 1980
April 1982
May 1983 November 1983 April 1984 July 1984 April 1985 February 1986 March 1986
April 1986
June 1987 May 1989
(D) Japanese residents can open accounts denominated in foreign currencies with market interest rates (D) foreign direct investment in Japan now subject only to notification (R) life insurance companies requested to limit voluntarily the net purchase of foreign bonds to about 10 percent of the net increase in assets (D) The Postal Life Insurance System permitted to purchase foreign securities up to 10 percent of total assets (R) controls introduced on increases in the foreign currency assets of pension trusts (D) yen-denominated foreign loans by banks liberalized (D) the principle of real demand for forward contracts abolished (D) the system of designated securities companies abolished (D) medium- and long-term Euroyen lending to nonresidents liberalized by overseas branches of Japanese banks (D) trust banks permitted to invest in foreign bonds up to 1 percent of total assets (for loan trust accounts) (D) regulations on foreign securities investments by life and nonlife insurance companies relaxed from 10 to 25 percent of total assets (from 25 to 30 percent in August 1986) (D) regulations on foreign securities investments by trust banks (for pension trust accounts) relaxed from 10 to 25 percent of total assets (from 25 to 30 percent in August) (D) the investment limit on foreign securities for the Postal Life Insurance System increased from 10 to 20 percent of total assets (D) the voluntary restraint on medium- and long-term Euroyen loans to residents abolished
Note: This chronology is based on Fukao (1990) and Takagi (1991). In the table, (D) means the deregulation of capital controls and (R) means the regulation of capital controls.
most capital flows were generally restricted. No Japanese security companies could buy foreign securities and bonds, and no foreign companies could buy Japanese securities. Throughout the 1970s, deregulations of Japanese capital controls came in several steps. Most deregulations generally occurred when the trade imbalance was substantial in Japan. When the Japanese trade deficit was large and the yen was depreciating rapidly, deregulations to encourage capital inflows took place; and when the Japanese trade surplus was significant and the yen was appreciating quickly, deregulations to encourage capital outflows were introduced. However, these deregulations were sometimes accompanied by regulations of reverse capital flows.
Determinants of capital controls in Japan
233
For example, the first oil crisis in 1973 and 1974 caused large trade deficit and substantial depreciation of the yen in Japan. At that period, deregulations of capital inflows began; short-term government securities became available to nonresidents to encourage capital inflows in August 1974. However, regulations of capital outflows were tightened at the same time, and after January 1974 authorities placed restrictions on the purchase of foreign securities by Japanese institutional investors and foreign currency deposits by residents. The opposite deregulations and regulations took place in 1977 when the trade balance turned into significant surplus and the yen appreciated substantially. At that period, capital outflows were deregulated, and in June 1977 Japanese security firms and others were allowed to acquire foreign securities. However, at the same time, capital inflows were strictly regulated, and authorities imposed several restrictions on capital inflows (e.g., the restrictions on the purchase of Japanese securities by foreigners in March 1978). In December 1980, a new law became effective allowing free flows of capital in and out of Japan with a few exceptions. Basic deregulations of Japanese capital controls were completed by this new law. Major outcomes of this law are summarized as follows: Japanese companies and individuals are now allowed to invest in foreign securities without securityfirm'sintermediation; a foreign loan by Japanese no longer needs permits; nonresidents can purchase and sell Japanese securities without any licensing; nonresidents can issue bonds in Japan only with prior reporting; and Japanese residents can open accounts denominated in foreign currencies with market interest rates. Even after this new law, there remained some minor restrictions on capital movements, especially before 1984. However, the restrictions were much less significant than those in the 1970s (see, e.g., Fukao and Okina 1989).
3
Measures of the degree of effective capital controls
In the 1970s, a series of deregulations reduced the degree of capital controls in Japan. However, it is not easy to calculate the effectiveness of those capital controls that continued in the 1970s. In the following analysis, we measure the degree of effective capital controls by using deviations from covered interest rate parity. In an environment with perfect capital mobility, covered interest rate parity is expected to hold continuously. Thus, if there existed various regulations of capital flows before December 1980 (the year of enactment of the new Foreign Exchange and Foreign Trade Control Law in Japan), we could show apparent unexploited gains from covered interest arbitrage
234
Shin-ichi Fukuda
72.1
73.1
74.1
75.1
76.1
77.1
78.1
79.1
80.1
81.1
82.1
83.1
84.1
Year/Month
Figure 9.1. Deviations from covered interest parity. in the 1970s. In particular, the unexploited gains would be larger when regulations on Japanese capital controls were very tight. In previous literature, several studies have examined the yen-dollar covered interest parity. For example, Otani and Tiwari (1981) and Otani (1983) examined covered interest parity from 1978 to March 1981. Ito (1986) and Fukao (1990) extended these studies and investigated the yendollar covered interest parity from January 1972 to December 1984. All of them found that there were significant deviations from the parity in the 1970s. They also found that the deviations became negligible and converged to zero after December 1980. Since a longer time series of arbitrage measures is available in Ito (1986), this chapter uses one of Ito's arbitrage measures. Formally, the measure is defined as follows DC = [(1 + EURO$/400)(FIS) - (1 4- RJA/400)]*400
(1)
where EUROS is the three-month Eurodollar deposit rate in London, RJA is the three-month repurchase agreement (Gensaki) rate in Tokyo, Fis the three-month forward exchange rate measured in yen per dollar, and S is the spot exchange rate measured in yen per dollar. Multiplied by 400, equation (1) expressed the per-period gain measure in the annual percentage yield. Figure 9.1 depicts the measure of effective capital controls that Ito (1986) calculated by equation (1) and shows several significant deviations
Determinants of capital controls in Japan
235
from covered interest parity in the 1970s. We can see the following four characteristics on the Japanese capital controls during the period of capital market liberalization: (1) strong restrictions in the Japanese capital market, which caused most wild fluctuations in the deviations before 1974, especially the very strong restrictions of capital outflows in the end of 1973 and in the beginning of 1974; (2) restrictions on capital inflows, which caused apparent arbitrage gains toward yen-dominated assets in the beginning of 1975 and in all of 1978; (3) minor restrictions on capital outflows between 1979 and 1980; and (4) no significant capital control after December 1980. These characteristics clearly correspond to the historical evidence of Japanese capital controls described in section 2 and Table 9.1. Although these characteristics are those of short-term capital movements, the timing of the restrictions is also closely related to that of long-term capital movements. For example, Ueda and Fujii (1986) categorized the periods of Japanese long-term capital controls into the following five periods: Period 1 (1970-early 1973), the beginning of capital market liberalization; Period 2 (November 1973-75), restrictions on capital outflows and deregulations of capital inflows; Period 3 (1977-78), restrictions on capital inflows, encouraging capital outflows; Period 4 (1979-80), deregulations of capital inflows; Period 5 (after December 1980), the completion of capital market liberalization. These categorized periods are qualitatively similar to those of the short-term capital controls in Figure 9.1, implying that the following results are less sensitive to our specific measure of capital controls. 4
The determinants of Japanese capital controls in the 1970s
This section investigates the determinants of Japanese capital controls by using our arbitrage measure discussed in the preceding section. The section first explains the method of estimation (i.e., the method of models of frictions) and then presents two types of empirical results focusing on the movements of Japanese trade balance. Estimation method There are many instances in which the dependent variable responds to only large values of the exogenous variables. Such models are called models of friction in econometrics (see Maddala 1983). Since the capital controls seem to have responded only to large exogenous shocks, the determinants of capital controls can be well described by one of these models. Define DCt as the desired level of capital controls in period t, which may depend on variables like trade balance, exchange rate, and so on.
236
Shin-ichi Fukuda
Because of transactions costs (e.g., administration costs), small changes in trade balance or exchange rate will have no effect on the changes in the actual level of capital controls. Thus, if we define the actual level of capital controls (i.e., deviations from covered interest parity) in period t by DCt9 DCt can be written as:
DCt =a{ + d + DCt if DCt < -SL{ DCt=vti{-ax
(2)
+ d + DCt if DCt > a2
Here v, is stochastic deviations from covered interest parity without capital controls. In the following analysis, we suppose that random variable v, lies between — d and d. We also set parameter "d" to be 0.5 because except for April 1981 the absolute value of DCt was always below 0.5 after December 1980. If we assume that the desired level of capital controls in period t is described by DCt = $'Xt + w,, the likelihood function for this model is L =ni(llaM(DCt -ax-djB'JQ/cr} • n2(\/a)[{(a2 - d - P'X)/*} - <\>{{-ax - d - P'Xt)/a}] • n 3 (l/o)
(3)
where refers to the standard normal density function and <2> refers to the cumulative normal. The product IT is over the sets 1,2, and 3 of observations for which DCt < - a , , -ax < DCt ^ a2, and DCt > a2, respectively. Maximizing this likelihood function with respect to av a2, J3, and a, we can estimate the determinants of the desired level of capital controls during the period of capital market liberalization in Japan. Estimation results I By using the above maximum likelihood method, we first estimate the determinants of capital controls in Japan when the desired level of capital controls in period / is described by either trade balance, current account, or the growth rate of exchange rate. That is, we maximize the likelihood function (3), assuming that the desired level of capital controls is described by DCt = b TBt.x + u,
(4a)
DCt = b CAt_x + ut,
(4b)
DC, = b DEt_x + ul9
(4c)
where TBt = trade balance, CAt = current account, and DEt = the growth rate of the yen-dollar exchange rate in terms of dollars.
Determinants of capital controls in Japan
237
Table 9.2 shows our estimation results of equations (4a)-(4c) by the maximum likelihood method, and also shows the estimation results by ordinary least squares. In order to avoid the simultaneous bias, lagged variables were used for independent variables. In the table, when each variable was regressed independently, the estimated coefficients were significantly negative in all cases. This implies that capital outflows were tightened when trade deficit and current account deficit were large and the yen depreciated substantially, whereas capital inflows were regulated when trade surplus or current account surplus became large and the yen appreciated substantially. However, when we regressed DC* on trade balance and exchange rate together, the coefficient of trade balance was the only significant one. Therefore, in controlling capital movements, the Japanese government was sensitive to trade balance and current account, but not to the exchange rate movements, which may temporarily deviate from its fundamental value. Estimation results II In general, the changes of trade balance were affected by various components. Thus, it is very important to see which components of trade balance affected the degree of capital controls in Japan. Here we divide the components of trade balance into two alternative directions. The first direction is to divide trade balance into unanticipated and anticipated ones. We examine what effects these unanticipated and anticipated parts of trade balance had on the degree of capital controls. In the estimation, we suppose that the desired level of capital controls in period t is described by DCt = bxUEXt + b2UIMt + b3EEXt + \EIMt + w,,
(5)
where UEXt = the amount of unanticipated exports at time t, UIMt = the amount of unanticipated imports at time t, EEXt = the amount of anticipated exports at time /, and EIMt = the amount of anticipated imports at time t. Data of anticipated exports and imports were taken from survey data in "The Prospects of Exports and Imports by Main Shosha [Economic Planning Agency]." The survey asks thirty-three of Japan's major trading companies about the prospects for their exports and imports for the next two quarters, in addition to their realized values. The survey covers about 60 percent of total exports and imports on a clearance basis. Table 9.3 provides the estimation results. In the case of import shocks, the degree of capital controls was responsive to both anticipated and unanticipated shocks. In the case of export shocks, the degree of capital controls was responsive only to anticipated shocks. These results imply that
Table 9.2. The determinants of capital controls in Japan OLS Constant
1.64 (2.83)
a\ al TB
ML
-0.017 (-4.17)
OLS
ML
0.25 (0.59) 4.75 (4.84) 0.48 (0.50) -0.003 (-3.23)
CA
Log of likelihood function
5.12 (21.7)
4.72 (4.82) 0.51 (0.52) -0.003 (-3.27)
-0.003 (-2.90)
5.23 (21.0)
-311
ML
1.58 (2.72)
-0.002 (-3.76)
-16.16 (-1.12)
-286
OLS
1.75 (2.09) 2.55 (2.50)
2.59 (3.24) 2.82 (2.58)
-0.001 (-3.80)
-310
ML
0.017 (0.04)
DE <x2
OLS
-287
-318
Note: /-values are in parentheses, a2 is the estimated variance of ut.
-22.96 (-1.57) 5.52 (20.1)
-299
-16.38 (-1.17)
-309
-15.84 (-0.92) 5.14 (17.97)
-286
Determinants of capital controls in Japan
239
Table 9.3. The effects of anticipated and unanticipated changes of trade balance Constant OLS
a\
al
-3.20 (-0.07)
-6.71 (-0.89) 13.71 -11.67 (0.28) (-0.60)
18.42 (0.81)
ML
VEX
a2
VIM
EEX
EIM
11.60 (2.87) 26.52 (2.37)
-11.19 (-1.95) -25.36 (-2.01)
9.92 (1.83) 25.79 4.45 (1.99) (5.65)
LLF -94.2 -76.0
Note: f-values are in parentheses. LLF = log of likelihood function, a 2 is the estimated variance of ur
Table 9.4. The effects of exports to alternative areas Constant OLS ML
a\
al
EXVS
EXA
IM
4.15 (0.82)
-2.47 (-2.54) -2.75 (-2.06)
1.49 (1.12) 1.12 (0.57)
0.25 (1.32) 0.49 (1.52)
0.226 (0.08) 0.50 (0.10)
a2
LLF -311
5.25 (19.4)
-291
Note: /-values are in parentheses. LLF = log of likelihood function, a 2 is the estimated variance of ur
Japanese government could correctly control capital movements caused by import shocks and anticipated export shocks. However, the latter result also indicates that capital controls were less effective in reducing the effects of unanticipated export shocks, implying that there existed some significant capital movements that may circumvent the government controls during the period of capital market liberalization in Japan. The second direction in dividing the trade balance is to sort out exports depending on export areas. We estimate whether the degree of capital controls was related to export areas. In the estimation, we suppose that the desired level of capital controls in period t is described by DCt = bxUSEXt + b2AEXt + b3IMt + w,,
(6)
where USEXt = the amount of exports to the United States and AEXt = the amount of exports to Southeast Asian countries. Since DC* is negatively correlated with trade balance, the expected results are that b] <0,b2< 0, and b3 > 0. Table 9.4 summarizes our estimation results. In the cases of exports to the United States, the estimated coefficient was correctly signed and significant. However, in the cases of exports to Southeast Asian countries, the estimated coefficient was not
240
Shin-ichi Fukuda
significant. This implies that the degree of capital controls was more responsive to exports to the United States and not responsive to exports to Southeast Asian countries. The latter implication is quite noteworthy because it indicates that exports to Southeast Asian countries might have caused some significant amount of circumventive capital flows during the period of capital market liberalization in Japan. 5
The effects of capital controls on trade balance
In the preceding section, we showed that Japanese capital controls were tightened when trade account (or current account) was imbalanced in the 1970s. The purpose of this section is to investigate whether these capital controls actually had significant effects on trade balance and current account in the 1970s. In investigation, we compared the effects of interest rate differentials between yen and dollar on trade balance and current account before and after December 1980. Since the deregulations of Japanese capital controls were basically completed after December 1980, the effects of interest rate differentials on trade balance and current account would be significantly different before and after December 1980 if capital controls had some significant effects in Japan. Thus, for two sample periods before and after December 1980, we first estimated the following equations: TBt = const + a TBt_x + 2 /=1 6. Ratet_p
(7a)
CAt = const + a CAt_x + 2iSBl bt Ratet_p
(7b)
where Ratet is call rate minus the three-month Eurodollar deposit rate in London at time t. In order to avoid the simultaneous bias, lagged variables were used for independent variables. In estimating the distributed lag, we used the method of the Almon distributed lag. The first four columns in Table 9.5 are the estimation results. If we look at the sum of coefficients of interest rate differentials, we can see that it is positive in all cases. Although it is not significant in the 1970s, it is very significant in the 1980s. This result is not consistent with the MundellFleming type story; in the Mundell-Fleming model, the rise of domestic interest rate appreciates the exchange rate and deteriorates trade balance. However, it is consistent with the traditional absorption approach and the income-savings balance approach because the rise of domestic interest rate has a negative effect on the domestic investment and output. The significance levels in the table clearly show that the effects of interest rate differentials on trade balance and current account were larger in the 1980s than in the 1970s. This implies that capital controls in the 1970s might have reduced the effects of interest rate differentials on trade bal-
Determinants of capital controls in Japan
241
Table 9.5. The effects of interest rate differentials on trade balance (TB) and current account (CA) 1972.1-1980. 12 TB Constant a
2,b, adR2 Durban A-statistic
1.397 (7.16) 0.807 (13.34) 0.024 (1.34) 0.76 -3.72
CA 1.237 (6.26) 0.808 (13.62) 0.035 (1.80) 0.77 -4.22
1981.1-1991 .8
1972.1-1980. 12
TB
CA
3.608 (7.26) 0.876 (22.44) 0.128 (2.65) 0.88 -4.03
3.293 (5.94) 0.843 (18.42) 0.128 (2.32) 0.83 -4.93
TB 1.462 (7.30) 0.749 (10.15) 0.063 (1.86) 0.80 -3.35
CA 1.270 (6.41) 0.746 (10.27) 0.081 (2.18) 0.77 -3.76
Note: /-values are in parentheses.
ance and current account. Thus, we next investigate whether the effects of interest rate differentials on trade balance and current account were amplified when the effects of capital controls were eliminated. We estimated the following equations: TBt = const + a TBt_x + X(b{ + c*DCt_) Ratet_r
(8a)
CAt = const + a CAt_x + 2(6. + c* DCt_) Ratet_r
(8b)
In order to avoid the simultaneous bias, lagged variables were used for independent variables in (8a) and (8b). The last two columns in Table 9.5 are their estimation results. They show that the sum of coefficients of bi in the 1970s is still smaller than that in the 1980s. However, they also show that when we eliminate the effects of capital controls, it becomes larger and more significant. This implies that in terms of policy implications, capital controls in the 1970s were almost successful in reducing the excess capital movements through suppressing the effects of interest rate differentials on trade balance. 6
The role of Japan's Multinational Trading Companies (MTCs) under capital controls
This section investigates what distributional effects Japanese capital controls had on trade balance during the period of capital market liberalization. In the analysis, we pay special attention to the role of Japan's Multinational Trading Companies (MTCs). This is because, as we explained in the introduction, MTCs might have been able to have big arbitrage
242
Shin-ichi Fukuda
opportunities by accessing the foreward exchange market or by using leads and lags (i.e., adjusting the length of payment periods) under capital controls. We first examine whether MTCs increased their weight in total exports and imports when capital controls were tightened. We estimated the following equations: WEXt = const + (a + &Dt)* DCt + (y + 8 Dt) * (time trend),
(9a)
WIMt = const + ( + Tj D) * DCt + (
(9b)
where WEXt - the weight of major trading companies in total exports, WIMt = the weight of major trading companies in total imports, and Dt = dummy variable, which is one only after 1977. As in equation (5), data of exports and imports by major trading companies were taken from survey data in "The Prospects of Exports and Imports by Main Shosha [Economic Planning Agency]." The survey asks Japan's major trading companies about the realized values of their exports and imports. Since MTCs cover most of the exports and imports of the surveyed companies, variables WEX and WIM in equation (9) are good proxies for the weight of MTCs in total exports and imports respectively. Because the number of surveyed companies changed after 1977, we added dummy variables after 1977 in estimating equations (9a) and (9b). In addition, in order to avoid the simultaneous bias, we estimated equations either by instrumental variable method or by ordinary least squares with lagged dependent variables. Table 9.6 is the estimation results. From the table, we can easily see that the estimate of a is significantly negative and that the estimate of <(> is significantly positive. That is, the weight of MTCs in total exports went up when trade surplus was large and capital inflows were strictly regulated, whereas the weight of MTCs in total imports rose when trade deficit was large and the controls of capital outflows were tightened. Since the increased exports (imports) mean the rise of capital inflows (outflows), these results imply that MTCs could circumvent the government capital controls during the period of capital market liberalization in Japan. Since the weight of exports of MTCs is biased to specific areas, the above results may have a special implication for the distributional effects of capital controls on trade balance. That is, since the exports of MTCs have more weight to Southeast Asian countries and less to the United States, the changes in the weight of MTCs in total exports may have shifted the weights of exports to Southeast Asian countries under capital controls. Thus, we estimated the following two types of equations.
243
Determinants of capital controls in Japan
Table 9.6. The weight of'Sogo Shosha" in total exports and imports under capital controls Dependent variable
Estimation method
WEX
Instrument OLS
WIM
Instrument OLS
Constant
DC
Time
adR2
DW
3.89 (106.7) 3.87 (121.1) 4.15 (121.7) 4.16 (118.2)
-0.0067 (-1.89) -0.0042 (-2.48) 0.0073 (2.69) 0.0018 (0.99)
0.0068 (2.62) 0.0070 (3.14) -0.0089 (-3.80) -0.0091 (-3.74)
0.92
1.60
0.92
1.38
0.92
1.54
0.89
1.19
Note: /-values are in parentheses. Lagged dependent and independent variables were used as instruments. The estimated coefficients of dummy variables are not shown here.
DUSXt = const + 2,A/ DCt_p
(10a)
DAEXt = const +
(10b)
DCt
WUSXt = const + 2/irf. DCt_p
(lla)
WAEXt = const + Xpf DCt_n
(lib)
where DUSXt = the growth rate of exports to the United States, DAEXt = the growth rate of exports to Southeast Asian countries, WUSXt = the weight of exports to the United States in total exports, and WAEXt = the weight of exports to Southeast Asian countries in total exports. If this conjecture is correct, we can expect that the sum of coefficients of |x, and p, are significantly negative but that of A., and IT . are not. In order to avoid the simultaneous bias, lagged variables were used for independent variables. Table 9.7 shows the estimation results. In the estimates of equations (10a) and (lla), the degree of capital controls had little explanatory power for the growth rate of exports to the United States. However, in the estimates of equations (10b) and (lib), DC had significantly negative effects on the growth rate of exports to Southeast Asia. That is, when capital inflows were regulated, there was a shift of exports to Southeast Asian countries. This implies that capital controls had significant distributional effects on Japanese trade patterns. It also indicates that MTCs might have played an important role in the process.
244
Shin-ichi Fukuda
Table 9.7. The effects of capital controls on export areas Dependent variable
DUSX
Constant
0.0044 (0.48)
DAEX 0.0053 (0.72)
Lagged dependent variable
DC(-\)
-0.0005 (-0.22) -0.0010 (-0.75) -0.008 (-0.53) -0.0023 (-0.78) -0.042 0.81
DC(-2) DC(-3) X,Z)C(-i) ad/? 2 DW Durban /i-statistic
Note: r-values are in parentheses. XpC(-i)
7
-0.0021 (-1.18) -0.0036 (-3.46) -0.0029 (-2.36) -0.0085 (-3.69) 0.260 2.01
wusx
WAEX
0.0092 (0.36)
0.0325 (1.29)
0.9674 (9.17) -0.0001 (-0.23) -0.0002 (-0.64) -0.0002 (-0.43) -0.0005 (-0.69) 0.710
0.8651 (7.95) -0.0002 (-0.46) -0.0008 (-3.43) -0.0007 (-2.59) -0.0018 (-3.16) 0.710
4.86
-0.047
denotes sum of lag coefficients.
Conclusions
This essay empirically investigated the determinants of capital controls and their effects on trade balance during the period of capital market liberalization in Japan. Main results can be summarized as follows. First, trade balance was one of the main determinants of capital controls in the 1970s. In particular, anticipated export shocks and export shocks to the United States were significant in explaining the degree of capital controls in the 1970s. However, neither unanticipated export shocks nor export shocks to Southeast Asian countries explained the degree of capital controls. Second, capital controls were mostly successful in suppressing the effects of interest rate differentials on trade balance in the 1970s. However, the capital controls also had significant distributional effects on trade balance through increasing the weight of MTCs in total exports and imports. In particular, the tightening of the regulations on capital inflows raised the weight of MTCs in total exports and increased the weight of exports to Southeast Asian countries. These distributional effects of capital controls were neglected in previous literature. However, if capital controls changed the distributional trade
Determinants of capital controls in Japan
245
patterns, they have very important welfare implications for Japanese capital controls. Furthermore, if the changed distributional trade patterns are unanticipated and are biased to some specific areas, they may also explain why capital controls were not responsive to unanticipated export shocks and export shocks to Southeast Asian countries. In terms of policy implications, Japanese capital controls were mostly successful in reducing the excess capital movements caused by trade imbalances during the period of capital market liberalization. However, in terms of welfare implications, they were accompanied by social costs distorting resource allocations. In distorting resource allocations, MTCs exploited arbitrage gains and might have caused some significant shifts in the Japanese trade patterns. References Drucker, Peter F. (1975). "Economic Realities and Economic Enterprise Strategy." In Ezra F. Vogel (ed.), Modern Japanese Organization and Decision Making. University of Berkeley, California Press, 238-39. Fukao, Mitsuhiro. (1990). "Liberalization of Japan's Foreign Exchange Controls and Structural Changes in the Balance of Payments." Monetary and Economic Studies (Bank of Japan) 8: 101-65. Fukao, Mitsuhiro, and Kunio Okina. (1989). "Internationalization of Financial Markets and Balance of Payments Imbalances: A Japanese Perspective." Carnegie-Rochester Conference Series on Public Policy 30: 167-220. Ito, Takatoshi. (1986). "Capital Controls and Covered Interest Parity between the Yen and the Dollar." Economic Studies Quarterly 37: 223-41. Komiya, Ryuraro, and Miyako Suda. (1991). Japans Foreign Exchange Policy, 1971-1982. Sydney: Allen and Unwin. Maddala, G. S. (1983). Limited Dependent and Qualitative Variables in Econometrics. Cambridge: Cambridge University Press. Otani, Ichiro. (1983). "Exchange Rate Instability and Capital Controls: The Japanese Experience, 1978-81." In D. Bigman and Teizo Taya (eds.), Exchange Rate and Trade Instability: Causes, Consequences and Remedies. Cambridge, Mass.: Ballinger, 311-37. Otani, Ichiro, and Siddhart Tiwari. (1981). "Capital Controls and Interest Rate Parity: The Japanese Experience 1978-81." IMF Staff Papers 28: 793-815. Takagi, Shinji. (1991). "Foreign Exchange Market Intervention and Domestic Monetary Control in Japan, 1973-89." Japan and the World Economy 3: 147-80. Ueda, Kazuo, and Mariko Fujii. (1986). "Saikin niokeru Wagakuni no Sihon Ryusitu nituite (On the recent capital outflows in Japan)" (in Japanese). Financial Review (Ministry of Finance) 3: 9-53. Yamazawa, Ippei, and Hirohisa Kohama. (1985). "Trading Companies and the Expansion of Foreign Trade: Japan, Korea, and Thailand." In K. Ohkawa and G. Ranis (eds.), Japan and the Developing Countries. Oxford: Basil Blackwell, 426-46.
CHAPTER 10
Capital mobility and economic policy Michael Dooley
Various countries are currently engaged in reform programs designed to increase the role of market forces in guiding resource allocation. The incentives behind these programs, and their scope, vary but are related to the view that, as economies develop, the complexity of economic activity quickly outstrips the administrative ability of government decision makers. Korea is an important example of an economy in transition in that the government intends over time to allow residents to participate freely in international capital markets. Such a policy initiative has far-reaching implications for other important economic institutions and policies. In fact, it is probably not an exaggeration to claim that all the ways in which the government interacts with the economy will be affected by this single development. One of the standard results from open economy theoretical models is that the degree of capital mobility is an important determinant of the effects of monetary and fiscal policies. In this chapter we argue that an equally important consideration for policy makers is that capital mobility increases the speed with which the private sector reacts to policy decisions. A related concern is that, once under way, the pace of liberalization of capital markets is difficult to control and even difficult to measure. Thus, during the transition, policy makers need information about the degree of integration of international and domestic capital markets. A review of recent attempts to directly measure capital mobility in transition economies is provided here. One approach to evaluating the effects of capital market integration is to identify historical episodes that seem to fit the circumstances of the country in question. In this chapter the historical experience of Germany in the early 1970s is reviewed because in several superficial respects, at least, the problems faced by the authorities of that time are similar to the problems faced by economies in transition today. 247
248
Michael Dooley
A factor that is difficult to quantify but that may be important is that, for Germany in the 1970s and Korea in recent years, important trading partners had increasingly adopted open financial markets. In the case of Germany, the move to current account convertibility in western Europe and the growth of Eurocurrency markets in the 1960s meant that German capital control programs faced increased capital market integration outside their borders. In the case of Korea, a major deregulation of Japan's financial markets in the mid-1980s and the development of offshore markets in several Asian countries mean that residents now face an expanded menu of financial instruments offshore. Thus, in both cases, the authorities' capital control regulations may have interacted with an increasing pull from offshore markets. Both countries experienced current account surpluses that generated political conflict with their trading partners and that generated large increases in the net foreign asset position of the central bank. Both governments have relied on extensive and apparently effective legal restrictions on capital flows to help manage the exchange rate. In both cases intervention was routinely sterilized so that the central bank apparently maintained independent domestic monetary targets. Finally, in both economies exports had been an important source of economic growth and this may have constrained exchange rate policy in that there was strong resistance to real exchange rate appreciation. The potential advantage of studying the German experience is that we know the outcome. A complex set of controls over capital inflows was abandoned and the exchange rate permitted to float - with important amounts of intervention. The conventional interpretation of the exchange market crisis leading up to the decisive confrontation in early 1973 is that the industrial countries were forced to abandon a system of fixed or heavily managed exchange rates by private capital flows. In spite of what was considered an effective capital control program, the Bundesbank was forced to purchase $2.7 million in a single morning in March 1973 before abandoning an exchange rate parity that had been set less than a month earlier. The recent speculative attack on the ERM should serve as a reminder that the industrial countries still do not choose to allow their currencies to float because of academic arguments about the superiority of a floating rate regime. The decision to open capital markets in the industrial countries in the early 1970s and in the economic community in the 1980s meant that the floating exchange rate regime had chosen them. Does a similar opening in Asia mean that managed exchange rate systems will quickly pass into history?
Capital mobility and economic policy
249
Policy issues In an informal way one can appreciate the importance of access to international capital markets by observing that such access allows residents to act on their expectations about economic policy and economic developments in a way that is not available with a closed capital account. In this chapter we emphasize the related idea that international capital mobility substantially compresses the time the authorities have to react to a change in the external environment. The argument follows a line that would have been very familiar to economists and policy makers in the industrial countries during the Bretton Woods system but which is often neglected in the current debate about liberalization. With a closed capital account the government can enjoy the apparent stability associated with a fixed exchange rate and can, at least for short time periods, maintain a domestic monetary target such as the price level. Once the capital account is opened, the government can choose to maintain the fixed exchange rate or the domestic price level but not both. One reason that this argument tends to be submerged in theoretical models of transition economies is that in the majority of such models the assumption is that purchasing power parity holds. Suppose this is the case so that P = ep*
(1)
where p is the domestic price level, e is the nominal exchange rate, and p* is the foreign price level. If arbitrage in goods markets or the "law of one price" maintains this condition, then with or without capital mobility there is never a trade-off between the exchange rate target and a domestic price level. For an economy very open to trade in goods and services this may not be a bad assumption. Suppose, for example, that all domestic product is exported and all domestic consumption is imported. Then the "domestic" price level is e p* and there is no meaning to a nominal target other than the exchange rate. But economies with a large nontraded goods sector are probably not well described by the assumption that goods market arbitrage maintains the relative price of traded and nontraded goods. In these economies without capital mobility the domestic price level is a weighted average of traded and nontraded goods prices. Even with a fixed nominal exchange rate the domestic money supply is a policy variable for a long time since residents can only acquire foreign exchange through the trade account. Lags between relative price changes and changes in the trade balance are generally estimated at a year or more. It follows that the feedback from
250
Michael Dooley
the current account to the monetary base may appear to be quite large, but it also appears to be quite slow. A reduced form model that illustrates this point would include the following relationships: (2)
MIP = M(i, Y) Y= YT+ Y"
(3) T
r= TTig) Y" = Wiq)
dY l8q > 0 8Y l8tq<0
(5)
Cr=CT(i,q)
8C7l8q < 0
(6)
SCISq > 0
(7)
N
C = C (i, q)
r/
(4)
(8)
TB= V-
(7
(9)
where: M - money supply / = domestic interest rate C1*, Y7 = consumption, output of traded goods CN, Y™ = consumption, output of nontraded goods q = relative price of traded and nontraded goods TB = trade balance The impact effect of an increase in M is a fall in the domestic interest rate and a rise in the price of nontraded goods. The much slower reaction in the goods market is an increased output of nontraded goods, as q rises, and a reduced output of traded goods (if one assumes full employment and slowly shifting factors of production). Output of traded goods falls and the difference between C and Y is satisfied by a trade deficit. The trade deficit will lower the monetary base as the monetary authority purchases domestic currency with reserves and eventually the system will return to its initial position. If the government sterilizes the trade deficit, the monetary base remains the same and the government is, in effect, borrowing in order to maintain the distorted relative prices and consumption choices of the private sector. The limit of the government's ability to finance the distortion is probably related to its ability to tax and there is probably an interesting maximizing model behind the descriptions model presented here. Our conjecture is that the government has a considerable time response to the trade flows generated by its monetary initiative.
Capital mobility and economic policy
251
As will be documented, in the cases of Korea and Germany the sterilization of trade imbalances seems to have been quite high. Thus, the domestic price level did not rise as fast as the usual analysis of a fixed-rate regime would suggest. Moreover, it appears in the case of Korea that the exchange rate was indirectly revalued through relaxation of trade restrictions in addition to managed changes in the nominal exchange rate. With no capital mobility, expectations about the eventual need for a change in the nominal exchange rate, or increases in the domestic currency price of nontraded goods, had provided incentives for residents and nonresidents to accelerate exports and wait on imports but the response seems to have been quite limited. Finally, in the case of Korea, the government's ability to manipulate commercial policy or even aggregate demand may have blunted the expected value of getting around the capital account restrictions. Thus, the stability of the fixed exchange rate system should be seen as relying on not only capital controls but the fact that sluggish responses to expectations about the fundamentals gives the government time to react with commercial policies, demand management policies, and managed changes in the exchange rate. The apparent difference between the historical cases seems to be that Korea has been successful in adjusting the current account whereas Germany was unable to do so. It seems to follow that in giving up capital account restrictions the government is also giving up alternatives to exchange rate flexibility in dealing with external balance. Henry Wallich (1984) summarized this view in observing that the adoption of floating rates by the industrial countries in 1973 "was a defensive maneuver on the part of national authorities by which they avoided what would very likely have become a system of widespread exchange controls. In a regulatory context, floating rates are not so much the opposite of fixed rates, but of a congeries of controls over international trade and payments."
Capital mobility and real exchange rate variability A related lesson from the experience of the industrial countries is not just that opening the capital account provides the means for betting on the future but also that the real exchange rate will travel over a wide enough field to make betting on it or hedging against it a prominent part of the economic environment. The so-called volatility of real exchange rates under floating is well documented. It is also clear that our structural models of how real exchange rates are related to fundamentals are of almost no value as predictors. In practical terms, this means that any nominal ex-
252
Michael Dooley
change rate picked the day before the capital account is opened is, given an initial price level for nontraded goods, likely to look "wrong" to private investors not many days later. In the time period over which we find out if this judgment is a good reading of the fundamentals, the central bank may be faced with private capital flows that are very large relative to the government's resources. A promising new approach to understanding the variability of real exchange rates exploits the fact that the equilibrium price of traded and nontraded goods might be sensitive to changes in investment in different countries. As Krugman (1989) points out, economists have resisted the idea that equilibrium relative prices of nontraded goods in industrial countries could change rapidly or by "large" amounts because we believe these countries are "all alike." In contrast, Peter Isard and I have developed a number of models based on the idea that the "transfer problem" linked with changing investment returns could explain large discrete jumps in real exchange rates (Dooley and Isard 1987, 1991; Dooley, Isard, and Taylor 1993). A recent paper (Backus, Kehoe, and Kydland 1992) provides an interesting theoretical basis for such a view. A credible plan for economic liberalization could generate changes in the expected steady-state capital stock, a new pattern of net capital flows, and an immediate change in the relative price of traded and nontraded goods and the real exchange rate. The problem of the unknown equilibrium real exchange rate is particularly interesting for countries in the midst of historically important structural changes in their economies. Once capital markets are completely liberalized, the choice of a fixed nominal exchange rate implies that changes in equilibrium real exchange rates can only be accomplished by changes in the domestic price level. Thus, it seems to follow that uncertainty about equilibrium real exchange rates during the transition is translated into uncertainty about and volatility for the domestic price level if the government fixes the nominal exchange rate. Liberalization of domestic financial markets may make the choice between exchange rate and price stability an important aspect of economic management during the transition. The implications of all this is it is of first-order importance to know from where we start. Economists with experience in Latin America or industrial countries tend to believe that capital controls do not work well. If so, liberalization is not a big shock. If not, it is an important step for the government. The case of Korea is interesting because opinion is widely distributed on this basic fact. In the next two sections we turn to the empirical evidence concerning the degree of capital mobility in Korea.
Capital mobility and economic policy
253
Capital mobility: direct evidence The main difficulty in identifying changes in capital mobility is that familiar measures of capital market integration among industrial countries involve comparisons of domestic and international interest rates (Frankel 1989, Chinn and Frankel 1993, Glick and Hutchinson 1990). In general, data for forward exchange rates, offshore deposit rates, and marketdetermined domestic interest rates are only available after a high degree of capital mobility has been well established. For countries moving away from systems of strict controls on interest rates, it is difficult to determine marginal yields actually faced by savers and investors. In some cases curb interest rates have been used as a measure of domestic interest rates (Edwards 1988, Reisen and Yeches 1991, Maloney 1992). Although these interest rates are available at times when there is a low level of integration, and provide some information, it is not possible to determine the volume of transactions in these markets or the extent to which these rates are influenced by the risk associated with borrowers that are excluded from the controlled credit market. Glick and Hutchinson (1990) find, for example, that for the one country in their sample for which the only interest available was a curb rate, econometric analysis of open interest parity condition yielded implausible results. Moreover, in making international comparisons domestic currency interest rates must be adjusted for exchange rate expectations. In practice this is a serious obstacle for countries in transition since forward exchange rates are typically not available to test covered interest parity conditions. Because the preliberalization exchange rate regime is usually an adjustable peg, the well-known "peso problem" limits the usefulness of ex-post realizations as measures of expected exchange rate changes. Another possible measure of capital mobility would be a qualitative evaluation of legal restrictions over capital movements (Dooley and Isard 1980). The difficulty with this approach is that changes in the intensity of legal restrictions are often associated with private initiatives designed to circumvent existing controls. Thus, a tightening of legal restrictions might be associated with an increase in capital market integration as the authorities respond to private initiatives with a lag and with decreasing effectiveness. Moreover, restrictions tend to stay on the books long after their effectiveness has been diminished by private avoidance. In a recent essay Dooley and Mathieson (1993) attempt to provide a measure of capital mobility for a number of Pacific Basin countries. They employed a technique suggested by Edwards and Khan (1985) and updated by Haque and Montiel (1990) that attempts to explain economic behavior in a country using an international interest rate and an interest
254
Michael Dooley
rate that is consistent with a closed capital account. This approach is appealing because it avoids the need to measure a domestic interest rate, a difficult variable to measure in most transition economies. To our surprise the test indicated that Korea was best characterized as having completely open capital markets. The key assumption in this approach is that it is possible to identify what the monetary base would have been in a given time period if the capital account was completely closed. The Haque and Montiel (HM) model assumes that the domestic market clearing interest rate (/) can be expressed as a weighted average of the uncovered interest parity interest rate (/*) and the domestic market-clearing interest rate that would be observed if the private capital account were completely closed (/') or i = 0 f + (1 - # v ; 0 < < 1
(1)
In order to identify the determinants of /', HM examine the rate of interest that would clear the domestic money market in the absence of private capital flows. The stock of money can be written as M = R + D = R_x + D + AR = R_, + D + CA + KAG + KAP
(2)
where M is the domestic money stock, R is the domestic currency value of foreign exchange reserves, D is the stock of domestic credit, A is the first difference operator, and CA, KA& and KAP are the domestic currency values of the current account, public sector capital account, and private sector capital account. The money stock that would have existed in the absence of private capital flows, which HM denote by M', would equal M' = M - KAP
(3)
HM then determined i' from the money market equilibrium condition that ln{M'IP) = ln(MD/P)
(4)
where P is the domestic price level. In their analysis, HM specified desired money holdings as a negative function of / and a positive function of income (y) and the lagged money supply (M/P)_vl The interest rate (/') that would then clear the money market was then solved for by using (4). Note that the strong assumption is that there is no contemporaneous interaction between private capital flows and the other components of (2). It turns out that this may be a particularly inappropriate assumption in the case of Korea. 1
This reduced form equation could be derived from a stock adjustment model of the form ln(M/P) - ln(M/P)_l = p[ln(MDIP) - ln(MIP)_x] with ln(MD/P) = a0 + a,/ + a2tny, a, < 0, a2 < 0, 0 < j3 < 1.
Capital mobility and economic policy
255
In our analysis, we use a broader definition of the real demand for money to allow for the effects of anticipated inflation and the yield on other close substitutes for narrow money. Thus ln(MDIP) = ot0 - a,(i - if) + a2lny - aAif - as{rD - if)
(5a)
where if is the expected rate of inflation, y is the level of income, rD is the rate of interest paid on time deposits, and In is the natural logarithm. This formulation assumes that the real demand for narrow money (currency plus demand deposits) is influenced by the expected real returns on currency (whose real return is given by the [negative of] the expected rate of inflation [77*]), on time deposits (rD — ir3e), and other (implicit) money market instruments (/ - if). This implies that, while the demand for money will be negatively related to the money market and time deposit interest rates, it could be negatively or positively related to expected inflation (depending on the size of a4 relative to the sum of ax and a5). If wealth holders adjust actual holdings of money to desired holdings gradually, then in(MIP) - ln(MIP)_x = a3(a0 - ax(i - if) + ajny (5b) ln{MIP)_x). - aAif - as{rD -if)This implies that, if /' is the interest rate that would equate ln{M'IP) = ln(MDIP), then it would be given by
(6) can be substituted into (1) to obtain an expression for the unobservable /, and the resulting expression can then be substituted into (5) to obtain ln{MIP) = aoa^ + a2a^lny 4- i/
OL5OL^D
(7)
Estimation of this regression hypothesis for Korea and six other Pacific Rim countries suggests that the value of ifj for Korea is indistinguishable from 1.0 and did not change over the 1970-89 period. In light of the discussion presented here, this result seems beyond the reasonable range of outcomes. A recent essay by Kwack (1993) suggests one reason for the failure of the test outlined here, at least for Korea. Kwack argues that virtually all the recorded private capital transactions appearing in the balance of payments were directed by the government. And, in fact, the practice of providing dollar-denominated loans to private
256
Michael Dooley
M'
M
Ms, M
Figure 10.1. Money market equilibrium.
parties in order to retire expensive foreign debt is a case in which private capital flows are clearly not reacting to foreign interest rates. Kwack argues that the only component of the balance of payments that could react endogenously to foreign variables is the trade account. Thus he proposes a model in which the central bank attempts to sterilize the increase in the monetary base that results from a trade imbalance assuming that the private capital account is a predetermined variable. He estimates a reaction or sterilization function, which suggests that the central bank sterilized about 80 percent of the changes in the monetary base associated with the trade imbalance. If private capital flows are determined by the government, the results reported in Dooley and Mathieson are seriously flawed. Note that the hypothesis embodied in (3) is that all the components of M except private capital flows are set exogenously by the central bank. This allows the interpretation of i' as the domestic interest rate that would have prevailed if the capital account was completely closed. This assumption is illustrated in Figure 10.1. The demand for money is a conventional function of a nominal interest rate. For a completely open economy the observed money demand is equal to the money supply and is consistent with income and the other arguments in (5) and the adjusted international interest rate. The money supply that would have existed in a completely closed economy is measured as the actual money supply less private capital flows. If there was a private capital inflow in this period, the closed economy money supply would have been smaller as illustrated by M' in Figure 10.1.
Capital mobility and economic policy
257
Thus, /' is estimated to be somewhat higher than the adjusted international rate. A potential problem with the interpretation of the estimated value for i' is that the authorities might have directed private capital flows in order to offset the current account imbalance. In this case our estimated value for V would be higher and - given conventional estimates of the interest elasticity of money - demand much higher than the "true" closed economy value. Thus, to identify the exogenous or closed economy portion of the monetary base it would be necessary to identify an endogenous component of the private capital flow. An alternative approach followed by Reisen and Yeches (1992) is to utilize a curb market interest rate as a proxy for /' in (3). Using this approach, he concludes that capital mobility has been quite limited and has evolved slowly, if at all, in recent years. In general, the quality of interest rate data makes direct comparisons of rates of return uncertain measures of capital mobility. The need to identify the complex interactions between private and official capital flows appears to make the first approach described earlier equally uncertain. In any case, there is some evidence that relaxing capital controls would be a genuine regime change in the case of Korea. If this is the case, a variety of models might be employed to evaluate the effects of this change. But given considerable uncertainty about the initial conditions and the stability of the models in the face of important structural changes, it seems prudent to explore alternative approaches to evaluating the likely problems associated with deregulation of financial markets. Capital mobility: a historical comparison In seeking a historical precedent for the policy choices facing the government of Korea, the experience of Germany in the late 1960s and early 1970s seems a likely candidate. The main attraction is that Germany also had an apparently effective program for controlling capital inflows. Figure 10.2 reproduced from Dooley and Isard (1980) shows a qualitative measure of the capital control program in Germany from 1970-74 and the differential between domestic and international interest rates. In this case, domestic interbank deposits and Euro deutsche mark deposits of offshore banks are very good proxies for i* and /'. The effectiveness of the program is measured in an informal way by the 600 basis point spread recorded during the first eight months of 1973.2 A similar calculation using curb rates for Korea is shown in Figure 10.3. For Korea the very large differen2
See the appendix for a chronology of capital controls utilized by the German authorities. There are a number of striking similarities with policies of the Korean government.
Percent per year
10
_ —
Observed interest differential. GDMEDM
~
Interest differential attributed to existing capital controls: Equation 1
- —
Interest differential attributed to
A
existing capital controls: Equation 2
8
6
-
4
"
r .A//
2 0
1970
1971
1972
1973
1974
Figure 10.2. German capital control program, 1970-74.
o>
o>
o>
o»
Figure 10.3. Korean domestic interest rate minus three-month LIBOR.
Capital mobility and economic policy
259
- Korea Percent " Germany
86/68
87/69
88/70
89/71 90/72
91/73 92/74
Figure 10.4. Current account as a percentage of reserve money. Data are for Korea 1986-92, Germany 1968-74. tials through 1988 seem to indicate that domestic borrowers without access to credits from state-controlled banks paid much higher rates than those available on international markets. This suggests a very restrictive capital control program. The dramatic fall in the differential in recent years also suggests a rapid convergence in rates of return in domestic and offshore markets. This, in turn, suggests that the domestic market is being opened to international transactions. Another element of the similarity is the prominence of trade surpluses in shaping monetary conditions over an extended time period. Figure 10.4 shows the current account surplus as a share of the monetary base for Korea and Germany over the time periods of interest. Note that the surplus relative to base money was much larger for Korea early in the period but that the surplus was eliminated and approximate balance achieved by 1990. In contrast, the surplus in Germany never exceeded about 25 percent of the base in any year but the trend toward the end of the period was for increasing surpluses. A possible interpretation of this evidence is that the government of Korea used the time given by capital controls to encourage adjustment in the current balance while the government of Germany was unwilling or unable to do so. Figure 10.5 shows private capital flows relative to base money over the period. The experience in terms of the trend and scale of capital flows is quite similar. In both cases we see a rising trend toward capital outflows and in both cases capital outflows are fairly large relative to the stock of reserve money. The big difference is that for Korea the direction of capital flows tends to offset the current account balance, whereas for Germany the current and private capital accounts have the same sign. This is reflected in Figure 10.6, which shows that throughout the time period growth in net foreign assets plays the major role in expansion of the monetary base in Germany. In contrast for Korea very rapid growth in
260
Michael Dooley 1 0.5 -
Percent
Korea
0 ~ Germany -0.5 -1
86/68
87/69
88/70
89/71
90/72
91/73
92/74
Figure 10.5. Private capital flows as a percentage of reserve money. Data are for Korea 1986-92, Germany 1968-74. 2.5 2 1.5 Percent 1 0.5
86/68
87/69
88/70
89/71
90/72
91/73
92/74
Figure 10.6. Net foreign assets as a percentage of reserve money. Data are for Korea 1986-92, Germany 1968-74. net foreign assets through 1988 is more than accounted for by the current balance with private capital flows tending to reduce the sterilization needed to offset the current account. Moreover, after 1988 the very rapid decline in the current account surplus generates a fall in net foreign assets even though private capital outflows increase. Conclusions
The one clear advantage of the historical pattern approach is that we know what happened in Germany in March 1973. With little warning the volume of capital flows proved larger than the government was willing to offset through exchange market intervention. Is there a similar speculative attack waiting to force Korea and other transition economies toward floating exchange rates? A preliminary assessment is that the situation is fundamentally different in one important respect. In both cases it might be possible to argue that the apparent effectiveness of capital controls has been more apparent than real. In particular, private capital movements in Germany were not large in the months leading up to the crisis so the apparent effectiveness
Capital mobility and economic policy
261
of Korean controls might also prove an illusion. In contrast the experience of the two countries in dealing with current account surpluses was much different. The Korean experience has been that alternative adjustment policies have worked and that private capital flows have tended to reduce the need to sterilize the monetary consequences of the current account imbalance. In the case of Germany, private capital flows tended to reinforce the sterilization required by the current account balance. Moreover, a forecast for the current account for 1974 and beyond without further exchange rate adjustment was for increasing surpluses. All this suggests that the outlook for a controlled relaxation of capital controls in Korea seems more feasible than is suggested by an uncritical review of experiences in the industrial countries. Eventually a liberalized capital account will constrain exchange rate policy but this could be years rather than months away. Appendix: A chronology of German capital controls3 On April 1, 1970, the Bundesbank reintroduced a special reserve ratio on the growth of banks' liabilities to nonresidents. With the exception of the four-month period from September through December 1971, when liabilities of both residents and nonresidents carried equal special reserve ratios, bank liabilities to nonresidents were subject to higher reserve requirements than bank liabilities to residents. This program served two purposes. First, it induced German banks to pay lower deposit rates to nonresidents than to residents. (This effect of the program probably was less important after May 1971, when controls were tightened to make payment of interest on deposits held by nonresidents subject to prior approval by the Bundesbank.) Second, it absorbed reserves and thereby "sterilized" the increase in the monetary base resulting from bank-reported capital inflows. On May 10, 1971, interest payments on nonresident bank deposits exceeding DM 50,000 were made subject to the prior approval of the Bundesbank, which was not normally granted. On March 1, 1972, the federal government introduced a cash deposit requirement (Bardepot) of 40 percent on most types of new credits by nonresidents to German nonbanks that exceeded DM 2 million per individual. The cash deposit, held by the bundesbank, did not bear interest. The deposit was increased to 50 percent effective on July 1, 1972, and the exempt amount was simultaneously reduced to DM 0.5 million. The exemption was further reduced to DM 0.05 million on January 1, 1973. On January 30, 1974, the cash deposit requirement was reduced to 20 3
Based on various issues of the Monthly Report of the German Bundesbank.
262
Michael Dooley
percent and the exemption raised to DM 0.1 million. In mid-September 1974, the cash requirement was eliminated retroactively from August 1, 1974. On June 29, 1972, the federal government decreed that purchases of fixed-interest securities by nonresidents would require prior authorization. Fixed-interest securities included all maturities of bonds: for example, all bank bonds, mortgage bonds, communal bonds, industrial bonds, and public authority bonds. In practice, the authorization requirement was equivalent to the prohibition of such purchases. The authorization requirement for all but short-term securities (with fewer than four years to maturity) was terminated on January 30, 1974. On February 5, 1972, the federal government extended its prior authorization requirement to the acquisition of domestic shares and mutual funds by nonresidents and to the raising of loans abroad by residents, including trade credits. Controls now applied to almost all capital transactions with nonresidents and no longer just to transactions in fixed-interest securities. These additional measures were terminated on January 30, 1974.
References Backus, David K., Patrick J. Kehoe, and Finn Kydland. (1992). "Relative Price Movements in Dynamic General Equlibrium Models of International Trade." Unpublished manuscript, August. Chinn, Menzie David, and Jeffrey A. Frankel. (1993). "Financial Links about the Pacific Rim: 1982-1992." University of California, Berkeley, CIDER Working Paper no. C93-023: 1-46, October. Deutsche Bundesbank. Monthly Report of the Deutsche Bundesbank. Frankfurt am Main. Various issues. Dooley, Michael P., and Peter Isard. (1980). "Capital Controls, Political Risk, and Deviations from Interest Parity." Journal of Political Economy 8 (April): 328-45. (1987). "Country Preferences, Currency Values and Policy Issues." Journal of Policy Modeling 9 (Spring): 65-81. (1991). "Note on Fiscal Policy, Investment Location Decisions, and Exchange Rates." Journal of International Money and Finance 10 (March): 161-68. Dooley, Michael P., Peter Isard, and Mark P. Taylor. (1993). "Exchange Rates, Country-Specific Shocks, and Gold." Group for International and Comparative Studies, University of California, Santa Cruz, Working Paper no. 225, January. Dooley, Michael P., and Donald Mathieson. (1993). "Capital Mobility in the Pacific Rim." In R. Glick and M. Hutchison (eds.), Exchange Rate and Interest Rate Policy: Perspectives from the Pacific Rim. Forthcoming. Edwards, Sebastian. (1988). "Financial Deregulation and Segmented Capital Markets: The Case of Korea." World Development 16, no. 2: 185-94.
Capital mobility and economic policy
263
Edwards, Sebastian, and Mohsin S. Khan. (1985). "Interest Rate Determination in Developing Countries: A Conceptual Framework." IMF Staff Papers 32, no. 3: 377-403. Frankel, Jeffrey. (1989). "Qualifying International Capital Mobility in the 1980V Cambridge, Mass.: NBER Working Paper no. 2856. Glick, Reuven, and Michael Hutchison. (1990). "Financial Liberalization in the Pacific Basin: Implications for Real Interest Rate Linkages." Journal of Japanese and International Economics 4: 30-48. Haque, Nadeem, and Peter Montiel. (1990). "Capital Mobility in Developing Countries - Some Empirical Tests." IMF Working Paper no. 117. Washington, D.C.: International Monetary Fund. Krugman, Paul. (1989). "Differences in Income Elasticities and Trends in Real Exchange Rates." European Economic Review 33: 1031-54. Kwack, Sung. (1993). "Monetary Control, Sterilization, and Exchange Rate Policy: Korea in the Period of Current Account Surpluses, 1982-1992." In R. Glick and M. Hutchison (eds.), Exchange Rate and Interest Rate Policy: Perspectives from the Pacific Rim. Forthcoming. Malony, William F. (1992). "Testing Capital Account Liberalization without Forward Rates: Why Was Japan So Open and Chile So Closed." Unpublished manuscript, August. Reisen, Helmut, and Helene Yeches. (1992). "Time-Varying Estimates on the Openness of the Capital Account in Korea and Taiwan." Paris: OECD Development Centre Technical Papers no. 42, August.
CHAPTER 11
Monetary and exchange rate policies 1973-1991: the Australian and New Zealand experience Victor Argy
This essay reviews the Australian and New Zealand experiences with monetary and exchange rate policies over the period 1973 to 1991. By way of background, the essay begins by presenting some basic facts about the two economies. Section 2 deals with the regulatory environment, and the evolution in that environment, in the two economies. This is addressed under several broad headings: Domestic financial regulation Controls over capital movements The exchange rate regime The regulation of labor markets The status of the central banks Section 3 reviews the key theoretical issues that bear on the design of monetary and exchange rate policies. The ground covered includes: The choice of exchange rate regime The design of monetary policy The impacts of domestic financial deregulation Section 4 addresses the question of how two similar countries like Australia and New Zealand go about choosing the exchange rate regime that best suits their characteristics. Section 5 looks at evidence, for Australia, of the impacts of domestic financial deregulation. Sections 6-8 review in more detail the evolution of monetary and exchange rate policies in Australia and New Zealand. Section 9 takes as a starting point the fact that the two economies, at about the same time, floated, deregulated their domestic financial system, I am grateful to G. Wells, P. Ip, and to participants at the conference for useful comments.
265
266
Victor Argy
and abandoned all exchange controls, and then asks how these developments might have affected the capacity of the economy to absorb shocks. In the past decade or so, the two economies have undergone radical change. They have both liberalized their financial systems and both are moving toward deregulating their labor markets (New Zealand having moved further in that direction). New Zealand has also changed its central bank constitution, now making inflation the sole objective of its monetary policy. Australia has not moved in that direction. If the opposition were to win government in Australia next year, they are already committed to adopting New Zealand's labor market reforms and to changing the Australian Reserve Bank constitution along New Zealand lines. For a general discussion of the Australian experience, see OECD Annual Surveys and Argy (1992a), and the latter for an extensive bibliography. For New Zealand, see Reserve Bank of New Zealand (1986, 1992), OECD Annual Surveys, Wells (1990), and Reserve Bank of New Zealand Quarterly Bulletins. 1
Some broad facts about the two economies
On the inflation front, from about 1975, the two countries started performing worse than the Organization of Economic Cooperation and Development (OECD) average but Australia's performance has been relatively superior to New Zealand's. It is only in recent years that New Zealand's inflation started converging toward Australia's and in 1991 their performance was better than that of the smaller OECD countries and the OECD as a whole (Table 11.1). Australia's unemployment has tracked fairly closely the OECD average (but it has been below that of the smaller OECD countries); in the past two years, however, its relative position has deteriorated dramatically (the current unemployment rate is about 11 percent). New Zealand's performance on unemployment was excellent, at least till the early 1980s. There has been a steady increase in the rate since, reaching now over 10 percent of the labor force (Table 11.1). Both countries have, over the years, agonized over large current account deficits (Table 11.1); New Zealand's has been brought down most dramatically in recent years. The persistently large current account deficits in the two economies have left a legacy of a relatively large foreign debt. In both countries the foreign debt to GDP ratio has accelerated sharply in the 1980s reaching some 42 percent in Australia and some 70 percent in New Zealand by 1990-91 (all in gross terms). Australia managed to convert years of persistent public sector deficits into surpluses by the second half of the 1980s (at least to 1991). New
Table 11.1. Macroeconomic performance (1972-91): Australia (A) and New Zealand (NZ)
Inflation
1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991
Unemployment OECD OECD smaller A
A
NZ
6.0 8.4 16.2 17.9 14.6 10.3 8.4 9.7 10.5 9.3 10.4 9.3 6.4 7.1 8.7 8.0 7.0 6.7 6.2 3.5
7.6 5.4 4.9 8.1 6.0 13.9 14.9 10.9 18.5 8.8 16.2 8.6 12.2 7.5 14.0 8.6 17.8 11.3 14.7 9.4 15.2 7.3 7.1 5.8 6.6 5.1 16.5 4.5 12.2 2.9 13.4 3.3 8.1 3.5 6.4 4.5 6.4 4.7 2.8 4.5
7.4 9.8 14.5 12.9 11.1 11.6 10.9 11.8 15.3 12.6 11.0 9.4 8.9 8.0 5.7 5.7 6.6 7.5 7.8 7.7
2.6 2.3 2.7 4.9 4.7 5.6 6.3 6.2 6.0 5.7 7.1 9.9 8.9 8.2 8.0 8.0 7.1 6.1 6.9 9.6
NZ
OECD OECD smaller A
0.5 3.6 0.2 3.3 0.1 3.6 0.3 5.1 0.4 5.2 0.6 5.2 1.7 5.1 1.9 5.2 2.7 5.9 3.5 6.7 3.7 8.0 5.4 8.6 4.6 8.1 3.6 8.0 4.0 7.9 4.1 7.4 5.6 6.9 7.2 6.4 7.8 6.2 10.3 7.1
Govt financial Current A/C balance
Growth of real GDP
3.3 3.2 3.4 4.4 4.6 4.9 5.6 6.3 7.2 8.2 9.4 10.5 10.7 10.5 10.1 9.7 9.4 8.8 8.6 9.6
3.4 5.6 1.9 1.8 4.0 0.9 3.5 3.6 2.3 3.7 -0.2 0.6 7.4 4.8 2.3 4.4 3.5 4.4 1.7 -1.9
NZ
OECD OECD smaller A
1.5 5.2 9.0 5.5 11.2 3.4 -5.1 0.3 3.2 3.7 -1.2 2.0 -6.1 2.1 1.6 2.8 0.5 2.2 4.4 0.7 2.9 0.9 1.3 1.6 8.6 3.4 1.2 3.1 1.0 2.8 -2.4 2.9 2.9 3.4 -0.7 3.6 0.5 2.9 -2.1 0.6
5.3 6.0 0.7 0.1 4.6 3.7 4.1 3.6 1.2 1.6 0.1 2.7 4.4 3.4 2.8 3.2 4.4 3.3 2.5 1.0
NZ
A
1.9 0.7 +0.9 1.6 1.5 +2.0 -3.3 -11.4 -0.8 -1.1 -10.1 -3.1 -1.9 -7.0 -0.3 -3.1 -5.7 -1.6 -3.8 -2.9 -4.2 -2.0 -3.6 -2.1 -2.8 -4.4 -0.2 -4.9 -5.0 +0.1 -5.0 -7.4 -1.0 -3.7 -4.5 -3.0 -4.8 -8.7 -2.9 -5.6 -7.5 -2.9 -5.6 -6.4 -1.8 -4.1 -5.5 -0.2 -4.1 -1.4 +1.5 -6.1 -3.1 +1.6 4.8 -3.2 +1.4 -3.5 -1.1 +0.3
NZ central govt
— — +1.0 +0.4 -1.5 +0.8 0.0 -4.1 -2.1 -3.9 -3.7 -4.6 -6.6 -5.2 -2.6 -3.5 -2.2 -1.5 -1.4 -1.7
Table 11.2. Real interest rates: Australia and New Zealand 72
73
A -0 1 -2 3 1 4 -2.4 NZ
74
75
-6 0 -5 3 -5.0 -8.3
76
77
78
79
-3 2 -2 0 1 2 07 -9.3 -5.1 - 2 . 0 -1.7
80
81
16 8
4 5
82
83
84
85
86
87
88
89
90
91
43 -3.3
4?
98 6.4
74 2.3
45 3.3
50 0
5 1 6.7
60 7.1
59 6.4
75 7.2
88
89
90
6.5 4.9
-3.0 -1.2
48
Note: Real interest rates = the government bond yield less the percent change in the consumer price index. Source: IMF, International Financial Statistics.
Table 11 .3. Percent change in terms of trade: Australia and New Zealand 72 A 10.0 NZ 15.7
73
74
75
28.6 22.0
-15.9 -22.8
-17.0 -0.5 -25.3 7.1
76
77
78
-7.2 4.5
- 2 . 7 -0.1 2.6 7.6
Source: IMF, International Financial Statistics.
79
82
83
84
80
81
-9.5 -10.9
-0.1 - 2 .1 - 1 . 0 -2.1 - 0 . 6 - 1 .1 - 2 . 5 -0.2
85
86
87
-5.3 -1.8
-7.4
- 1 . 9 14.5 11.0 6.8
Monetary policies: Australia and New Zealand
269
Zealand also adopted a policy of fiscal consolidation from the mid-1980s and succeeded in sharply reducing the deficits by the end of the 1980s (Table 11.1). The gross public debt as a percent of GDP is currently only some 16 percent in Australia whereas in New Zealand it would be around 55 percent. On this front, Australia compares very favorably with the average OECD performance whereas New Zealand is somewhat worse. Real interest rates have tended over the long haul to follow the pattern of other industrial countries; real interest rates were negative in the middle-to-late 1970s but became substantially positive in the 1980s (Table 11.2). Of the two economies New Zealand's is the more open. Average exports-imports as a percent of GNP is some 30 percent in New Zealand but only 23 percent in Australia. As small industrial economies, neither, however, could be described as being very open. The proportion of rural exports as a percent of total exports is higher in New Zealand, being over half. In this respect, Australia's export commodity structure is more diversified, with exports roughly divided into three equal parts: rural, minerals, and manufactures. Australia is also two-thirds self-contained in oil but New Zealand has to import all its oil. In both economies, the principal trading partner is Japan. In Australia, the United States and the United Kingdom are next in that order. In New Zealand Australia is the second trading partner with the United States and the United Kingdom next in that order. As major producers of rural products and commodities, both countries are exposed to large swings in their terms of trade, more so than in most industrial countries (Table 11.3). The fluctuations are roughly similar but not identical. The differences reflect primarily the differences in the export commodity base and the sensitivity of the prices of the exports to fluctuations in the level of world activity. 2
The regulatory environment Domestic financial regulation
In both economies until the early 1980s there were extensive interest rate controls and prescribed asset ratios on banks and other financial institutions. There were also lending guidelines for the banks. Entry into the banking system was limited. In both economies some steps toward liberalization had been taken in the mid to second half of the 1970s. In New Zealand progress was very slow and by the early 1980s remained more heavily regulated than in Australia. In Australia, by mid-1982 quantitative lending guidelines to banks were abandoned; by 1984-85 all significant
270
Victor Argy
controls had been removed. In New Zealand controls were removed between mid-1984 and the end of that year. External financial regulation-capital controls In both economies capital movements were, in principle, regulated, till December 1983 in Australia and December 1984 in New Zealand. The evolution of controls was, however, different in the two economies. In New Zealand capital flows were strictly monitored to the end, with only minor variations in the regulations. Australia's experience is richer in this respect. Australia varied the regulations depending on the state of the balance of payments. Until late 1971 the general principle was that capital flows were prohibited, unless specifically authorized. In reality most capital inflows were allowed entry and indeed encouraged. In sharp contrast, capital outflows were rigidly controlled. Between 1971 and 1978 capital flows were monitored as a method of managing the Australian dollar. When the dollar was strong, as in 1972-73, and from December 1976 to June 1977, capital inflows were discouraged. Two methods were used here: embargoes on borrowings of a certain maturity and a cost penalty on certain types of borrowing (the VDR). In sharp contrast, when the dollar was weak, as in 1974 to 1976, these regulations were reversed, with inflows now encouraged. By June 1978, although technically controls over inflows were still in place, these were not enforced so inflows had by then become largely free. When the dollar was very strong again in 1980-81 and in March-December 1983, controls over inflows were not reimposed. The exchange rate regime The evolution of exchange rate policy is similar in the two economies. The Australian dollar was pegged (but adjustable) to the U.S. dollar between 1973 and September 1974. Between September 1974 and November 1976 the Australian dollar was pegged (but adjustable) to a currency basket. In the years to November 1976 there were several discrete exchange rate adjustments: February 1973 (11+ percent), July 1973 (25 percent tariff cut, 5+ percent), September 1984 (12— percent), November 1976 (17.5- percent). Between November 1976 and March-December 1983, Australia adopted a discretionary version of the crawling peg. A small committee of four adjusted the Australian dollar daily. The basis for the crawl varied. To early 1981 the principal determinant was the overall balance of payments; after that, increased attention came to be paid to the current account. In December 1983 the Australian dollar floated. It has, however, never been a completely clean float.
Monetary policies: Australia and New Zealand
271
The New Zealand dollar was pegged (but adjustable) to the U.S. dollar to July 1973. Between July 1973 and June 1979 it was pegged (but adjustable) against a basket. Between June 1979 and June 1982, New Zealand adopted a purchasing power parity (real exchange rate) variant of the crawling peg. The exchange rate was adjusted to offset differences in inflation rates between New Zealand and its trading partners. New Zealand operated an ex-ante scheme based on consumer prices. If the realized inflation rates were different from the anticipated ones, some correction would be made. In June 1982 New Zealand reverted to an adjustable peg against a basket. As in Australia, there were several discrete exchange rate adjustments: September 1974 (6.2- percent), August 1975 ( 1 5 - percent), November 1976 (2.7- percent), June 1979 ( 5 - percent), March 1983 ( 6 - percent), July 1984 (20— percent). In the three years of the crawl the New Zealand dollar devalued by some 6 percent each year. In March 1985 the New Zealand dollar floated, again not cleanly. Regulation of labor markets Labor markets play a key role in the adjustment to a change in monetary policy and to a real disturbance. The more flexible the labor markets the quicker is the adjustment to equilibrium, with correspondingly weaker effects on employment and output. The two countries have, over the past decades, sought to find an appropriate wages policy that would give them that flexibility. To anticipate, we can summarize the experience by saying that the two economies have evolved from a centralized interventionist regime toward a decentralized enterprise-based bargaining regime. New Zealand has gone further than Australia in adopting a noninterventionist stance. The New Zealand experience To simplify slightly, we can say that until 1987 the system of wage determination was three-tiered: 1
Unions and employers negotiated national award wages for particular occupations. 2 In the event the parties did not agree, reference would be made first to conciliation, then to arbitration. 3 There was periodic intervention by the government in the determination of national wage outcomes. Intervention took several forms: guidelines, direct controls, the imposition of minimum periods of currency for new agreements and constraints on the re-
272
Victor Argy
negotiating of old agreements, wage pauses and freezes, partial indexation directives. A new phase began in 1987. Legislation was passed in 1987 and again in 1991. The legislation aimed ultimately at decentralizing wage bargaining, weakening the unions, and increasing labor market flexibility. The 1987 act had only limited objectives. Unions but not employers were allowed to opt out of the award system, but compulsory unionism was retained. The 1991 legislation went much further down the road. Compulsory unionism was abolished; unions lost their exclusive right to negotiate on behalf of workers; awards and agreements were replaced with employment contracts; parties were free to negotiate the type of contract they wanted (Kerr 1991). The Australian experience The Australian wage determination environment has also undergone radical change and if the opposition party wins government next year they have already pledged to adopt something like the New Zealand 1991 legislation, which goes beyond anything accomplished to date. The traditional Australian model of wage setting is one where you have an independent Arbitration Commission which periodically hears submissions from interested parties, including, most important, the government, employer groups and the ACTU (the Australian Confederation of Trade Unions). The government's input into the proceedings is very important but is not necessarily decisive. After hearing all the evidence the commission made a ruling on national wages. In this sense it was a highly centralized system (but unlike the corporatist systems, say in Austria, Sweden, and Norway). At times, however, collective bargaining between unions and employers has been encouraged and has assumed some importance in wage outcomes. In turn such collective bargaining agreements would be registered as awards by the commission. After the Hawke Labour government took over in early 1983, they negotiated an accord with the unions, which has passed through three stages. The first stage was one where full wage indexation was endorsed. In the second stage this was modified to discount for the effects of the devaluation of 1985-86. In the third stage the accord became less centralized, with increased emphasis being placed on enterprise agreements, within certain ground rules. An important feature of the accord was the wage-tax trade-off*: The government would offer tax concessions in return for wage moderation. The system is undergoing radical change and is currently the subject of a very lively debate. The key difference between the opposition party and
Monetary policies: Australia and New Zealand
273
the present government on this front is that the opposition wants a "free for all" market-oriented enterprise-based system of collective bargaining while the government wants to retain some degree of centralized guideline. Conclusions on Australia and New Zealand Would enterprise bargaining in the two economies improve the labor market adjustment? There is a large literature that tries to rank countries in terms of the degree of centralization in wage setting and macro performance (primarily, unemployment and real wage flexibility - see Calmfors and Driffill 1988). It turns out that the best performers are the centralized and the decentralized economies (i.e., those with enterprise bargaining), whereas the worst performers are the intermediate economies (with industry bargaining). Even if Australia and New Zealand were both "centralist" (a question that could be debated) and we took this framework seriously, it is not evident that the switch to enterprise bargaining will improve macro performance. Moreover, enterprise bargaining has been in place in France, the United Kingdom, Italy, Switzerland, the United States, Canada, and Japan, a group of countries with very divergent macro economic performance. Comparative studies of real wage flexibility suggest Australia is ranked relatively low (Andersen 1989). A new study (Andersen 1992) suggests that Australia had a somewhat higher degree of real wage flexibility than New Zealand. Central bank status There is now a large literature concerned with the potential implications, for the equilibrium inflation rate, of allowing governments complete discretion with respect to the use of monetary policy (Barro and Gordon 1983, Fischer 1988, Rogoff 1987, Argy 1992a). This literature tries to demonstrate that with government discretion and a short policy horizon inflation is likely to end up being excessive without any ultimate improvement in the unemployment rate. The prescriptions that emerge are that in the conduct of monetary policy: 1 Governments should attach overriding importance to inflation, or 2 The central bank, assumed to have a longer time horizon, should be made independent, or 3 More directly, the central bank constitution should be changed to assign inflation absolute priority.
274
Victor Argy
According to the thesis, any one of these changes will achieve the same result of a better inflation performance. There is some empirical evidence that tries to demonstrate that independent central banks have in fact tended to perform better on the inflation front and no worse on the unemployment front (Alesina and Summers 1990). Who has had responsibility for monetary policy in Australia and New Zealand? The Reserve Bank Act of Australia gives the government ultimate responsibility for monetary policy. The Governing Board is required by the 1959 Reserve Bank Act to conduct its monetary policy in a way that "will best contribute to the stability of the currency . . . the maintenance of full employment... and the economic prosperity and welfare of . . . Australia." The current government's position is that the Reserve Bank should "have a broad range of objectives. . . . It is not there to pursue one objective (inflation) singlemindedly." (Dawkins, the treasurer, reported in the Australian, March 27, p. 3). The opposition, however, is now committed to changing the charter of the bank to one that focuses on fighting inflation only, with formal targets in the range of 0 to 2 percent. Under the New Zealand Act the Reserve Bank was required to carry out "the monetary policy of the Government, which shall be directed to the maintenance and promotion of economic and social welfare in New Zealand having regard to the desirability of promoting the highest degree of production, trade and employment and of maintaining a stable internal price level." The Act was changed in 1989. Section 8 now states: "The primary function of the Bank is to formulate and implement monetary policy directed to the economic objective of achieving and maintaining stability in the general level of prices." The then minister of finance and the governor agreed on an inflation target in the range of 0 to 2 percent to be achieved by the end of 1992. The new conservative government extended the deadline to the end of 1993. There are only four smaller OECD member countries with flexible rates that, in principle at any rate, can enjoy an independent monetary policy: Australia, Canada, New Zealand, and Switzerland. Switzerland's Central Bank is independent and gives overriding priority to inflation. Canada in recent years has chosen to do likewise and so now has New Zealand. Australia is thus the odd one out in this group. 3
Monetary and exchange rate policy - theoretical issues
By way of background, before addressing in more detail the Australian and New Zealand experience, it is useful to summarize the key theoretical issues.
Monetary policies: Australia and New Zealand
275
These issues fall conveniently into three categories. The first is concerned with the choice of exchange rate regime. The second is concerned with the design of monetary policy. The third is concerned with the degree of regulation of the domestic financial system. The choice of exchange rate regime What exchange rate regime should a small country choose? There are four key criteria one might apply (Argy 1990). The first is micro-based and turns on efficiency considerations. To begin, an exchange rate regime may deliver efficiency gains from the removal of exchange rate uncertainty. Potential gains include the elimination of or the reduction in the costs of conversion, of portfolio revaluations, of forward cover, and of speculation. An exchange rate regime may also, in principle, improve resource allocation in several ways: by mitigating distortions to trade and to capital flows, by maintaining the exchange rate closer to its longer run "equilibrium" path, or by alleviating the costs of resource shifts. A second criterion focuses on inflation discipline. In essence, we ask which exchange rate regime best succeeds in securing a steady-state rate of inflation closest to its "optimal" level. A third criterion focuses now on the insulation properties of exchange rate regimes. Given the policy stance, we ask here how each regime, in the short run, insulates the economy from disturbances of domestic and of foreign origin. A fourth criterion has to do with policy effectiveness, that is, the capacity of each regime to exploit the policy instruments available so as to achieve the key macro targets of policy. We have, to sum up, four criteria to apply to a large number of potential exchange rate regimes. This is clearly a very tall order. Moreover, the criteria themselves are not always clearly defined or easy to apply. What we do in the sections that follow is provide a general discursive discussion of individual regimes, with closer attention paid to some than to others. The permanent peg versus the float Pegging to a single currency or to a basket will achieve inflation discipline if the country has trading partners whose inflation performance is good, and there are no parallel disciplinary constraints at home. Evidently if the trading partners have a poor inflation record or if discipline is achievable at home (by any of the three routes noted previously under central bank status) the country will not benefit on this front at least from pegging (as against floating). Indeed, if inflation discipline can be achieved at home, flexible rates could have an advantage in that it allows
276
Victor Argy
the country to achieve its own "optimal" inflation rate rather than have it imposed from the outside. What about the insulation criterion? At the analytical level the choice between say a peg and a float will come down to how the economy is modeled, where the disturbances come from, and the target variables explicitly taken into account. Assuming asset substitution is perfect and wage indexation imperfect, for domestic disturbances, fixed rates are more stabilizing for money demand shocks while flexible rates are more stabilizing for expenditure shocks. For supply shocks the outcomes are ambiguous, in a word because the interest rate effect is ambiguous. For external disturbances, relative outcomes are strictly ambiguous (the idea that flexible rates shelter the economy from external disturbances was rebutted long ago). If indexation is perfect, we have the familiar result that there are no real differences between fixed and flexible rate regimes - that is, the impact on all real variables (output, the real exchange rate, the real interest rate) is identical in the two regimes, whatever the disturbance. What can we conclude on the insulation criterion? Since countries are exposed to different mixes of disturbances both at home and abroad, and since some disturbances favor exchange rate stability while others favor exchange rate flexibility, we have here a possible basis for discriminating between countries. But even at this level there are difficulties. First, it is difficult to determine empirically the weight attaching to particular disturbances in individual countries. Second, a measure of expected future, not past, disturbances is needed. For example, if the monetary environment has been unstable in the past - potentially favoring fixed rates - will it remain so in the future? Third, since the importance of different disturbances changes over time, it might be inefficient to change the regime in accordance with changes in these weights (see Flood, Bhandari, and Home 1989). On the policy effectiveness criterion we want to know how efficient policy will be in each regime in achieving targets. The issues here are extremely complex. Monetary policy has to be ruled out as an independent policy instrument in a peg with a high degree of asset substitution. But this may not be a loss depending on the use to which monetary policy is put under flexible rates. If discretion is used unwisely, as it often is, this may count as a disadvantage of a float. Moreover, if under a float monetary policy is directed at inflation, as noted previously this will come down to domestic versus external discipline. On Mundell-Fleming type short-term models fiscal policy is more effective under fixed rates but again much depends on how wisely it is used with fixed rates. Moreover, this only holds for the short run and for purely stabilization objectives.
Monetary policies: Australia and New Zealand
277
The adjustable peg (AP) (with a very small band) 1. In the short run (i.e., in between realignments) the AP regime operates similarly to a permanent peg. In both regimes if capital mobility is very high the authorities lose control over the money supply; on the other hand, if capital mobility is sufficiently restricted, sterilization is feasible and the authorities are able, in principle, to target both the money supply and the exchange rate. 2. It is difficult to operate an AP regime if the going rate of inflation is persistently and significantly different from that of the trading partners. Such differences would bring continuing pressures for exchange rate adjustment, creating tensions and uncertainty. In such conditions a crawling peg or a float would be the more appropriate regime. One precondition, therefore, for the effective functioning of an AP regime is that the country maintain its rate of inflation within manageable reach of its trading partners. Exchange rate adjustments could then be used, in principle, to correct for ex-post unanticipated price disturbances or for fundamental imbalances. 3. The fact that the exchange rate can be used periodically as a policy instrument can itself be very disruptive to the economy. In most cases such exchange rate changes are anticipated, provoking large outflows or inflows of capital. Frequently therefore such movements of capital turn out to be one-way bets at the expense of central banks. A related point is that rumors of changes or suspicions of change may trigger large flows of capital, occasionally forcing the authorities into an unnecessary exchange rate adjustment. These comments suggest that an AP regime is likely to function more efficiently if such flows of capital can be monitored - hence the perceived need to support the AP regime with some control over capital movements. Without restrictions over capital movements the authorities might at times lose control over the money supply or over the exchange rate. For example, an expected devaluation, leading to huge outflows might place the authorities in a very serious dilemma: to sharply restrict the money supply and raise interest rates or to go along with market sentiment, which may be misguided, and devalue. It is significant that, in general, in the countries where an AP has been in operation restrictions on the movement of capital have been in force, (e.g., Sweden, Norway, Finland, Australia, New Zealand). 4. If wages are fully indexed, a devaluation to correct a fundamental imbalance in the current account may be self-defeating, necessitating, ultimately, a new devaluation. This creates the potential for a "vicious" devaluation cycle. (To some extent this represents the earlier Swedish and Finn-
278
Victor Argy
ish experience.) This highlights the importance of accompanying a devaluation with a wage policy. The difficulty is that the exchange rate adjustment has to occur first, so the authorities have to gamble that any subsequent wage negotiation will be successful. As a matter of fact, nearly all countries that have devalued have tried to implement a wage policy that will minimize the spillover of a devaluation into wages. 5. In the AP regime the authorities are required to make the very difficult judgment when to adjust the currency and also by how much. If the market disagrees with their judgment on the size of the adjustment, further uncertainty is created. Even if the market went along with the decision, there is still a question of, say, whether the amount of the exchange rate adjustment will be adequate, taking account of relevant elasticities and spillovers into wages, in due course to correct the imbalance. 6. The role of real wage flexibility is critical. Assume that there is symmetrical real wage flexibility in the sense that any change in the real wage rate (given unemployment) secured under, say, an adjustable peg regime can as readily and as speedily be secured under a peg. The wage rate now replaces the exchange rate as a policy instrument. It is evident that an adjustable peg regime offers no real advantage. Indeed, since an exchange rate change is more disturbing than a wage change, this would count as a disadvantage of the adjustable peg. Friedman (1953), making his classic case for flexible rates, based one argument in favor of flexible rates on an asymmetry. Keynes also thought that real wages could be reduced by raising prices, but not so by lowering wages. Asymmetry could come from straight money illusion (particularly if a wage cut is required) or from the fact that wage policies are harder to negotiate and take longer than an adjustment to the exchange rate. To sum up, if there is asymmetrical real wage flexibility and a fundamental imbalance in the current account emerges, the adjustable peg regime has a clear advantage over the peg. With a peg, output will have to be sacrificed to secure current account objectives. The crawling peg We limit our comments here to the PPP crawl and the discretionary crawl. For a discussion of the PPP crawl, see Dornbusch (1982) and Adams and Gros (1986). New Zealand's experience with thefirstand Australia's experience with the second will be discussed later. Consider first the PPP crawl. 1. A PPP crawl, by definition, roughly maintains a country's competitive position. At the same time, exchange rate adjustments are relatively small and fairly predictable. These are the main advantages of a PPP crawl. 2. A PPP crawl may be giving the green light to a potentially high inflation country to maintain its relatively high rate of inflation. From the
Monetary policies: Australia and New Zealand
279
standpoint of "discipline," therefore, the PPP crawl may be inferior to the AP. 3. With relative inflation rates largely known, the crawl may be fairly predictable. If there are no restrictions on capital flows, the country crawling downward (upward) will have to maintain an interest rate advantage (disadvantage) that will exactly offset the anticipated exchange rate change. Thus, as Williamson (1981) put it, "monetary policy could not be used for internal stabilisation" (but, in principle, of course fiscal policy could be). In this respect, at any rate, the PPP crawl operates in a way that is similar to a peg with perfect asset substitution. 4. There will be occasions when a change in the real exchange rate will be needed. For example, although competitiveness may be maintained, a current account imbalance may emerge, which may require a change in the real exchange rate. It is, however, difficult to accommodate periodic changes in real exchange rates within the system without exposing the economy to speculative bursts. At this point the system will resemble the AP. 5. Trying to maintain the real exchange rate in the face of a real shock may be destabilizing at times. Suppose we had full wage indexation. A fiscal expansion will lead in the short run to an expansion in both output and prices. The authorities will now devalue to maintain the real exchange rate but this may ultimately prove self-defeating. 6. A relatively minor point is that there is a wide range of potential relative price indexes that might be used as a basis for the PPP crawl: export prices, consumer prices, unit labor costs, wholesale prices. We turn now to the discretionary crawl. The principal difference between the discretionary crawl and the PPP crawl is that the basis of the crawl is now more flexible. This is both an advantage and a disadvantage. It is an advantage because, if necessary, the authorities can fine-tune the small adjustments to the ongoing conditions. This also allows somewhat more scope for an independent monetary policy. It is a disadvantage because the crawl is now less predictable and so the system becomes more exposed to speculation, albeit on a modest scale. In other respects it is similar to the PPP crawl. Inflation "discipline" is lacking; at the same time if large changes in fundamentals occur, the public will not believe that small changes are sustainable and the system will lose credibility. The design of monetary policy Monetary policy is predetermined with fixed rates. Under flexible rates, however, there is a wide choice among alternative strategies. The authori-
280
Victor Argy
ties could, in principle, target inflation directly or target nominal income or target the growth in the money stock periodically (with varying bands) or adopt a discretionary "checklist" approach (where a number of targets are taken into account) or follow a simple Friedman type money growth rule (or Meltzer's [1988] variant which explicitly takes account of recent trends in velocity) or adopt an econometrically derived monetary policy reaction function. Domestic financial deregulation The case for domestic financial deregulation is usually put in terms of the microefficiency gains that will accrue. These can take many forms: the greater variety of financial services; the removal of discriminations stemming from a regulated system of credit rationing; the removal of interest rate distortions, allowing financial flows to move into the most efficient sectors; welfare-augmenting switches from the payment of implicit interest (e.g., subsiding transactions and deposit accounts and offering an extended system of branches) to the payment of explicit interest; and greater competitiveness among financial institutions with benefits to consumers in terms of reduced profit margins. There is, however, another side to the ledger. One other important aspect of deregulation is its potential for creating "financial fragility" by exposing the financial sector to greater risk (see Blundell-Wignall and Browne 1991). There are several potential links between deregulation and financial fragility. First, greater competition among financial institutions may encourage greater risk-taking, in an effort to earn a larger return on assets. Second, the relaxation of asset controls may free institutions to make riskier loans. Third, the liberalized environment may generate excessive speculation, financed by institutional borrowing, which drives up asset prices to unsustainable levels. In the end, there will be a day of reckoning when financial institutions might find themselves dangerously exposed to bad debts. Fourth, we will also see that deregulation would be associated with greater interest rate volatility. At times when interest rates are rising, institutions with relatively inflexible loan rates may become particularly vulnerable. What about the macro implications of domestic deregulation? Here it is important to distinguish steady-state from transitional effects. (For a fuller discussion, see Argy 1992b.) First, what does domestic deregulation do in steady state to policy effectiveness and to the capacity of the economy to absorb shocks: That is, is deregulation a good or a bad insulator?
Monetary policies: Australia and New Zealand
281
If banks are now free to adjust the interest rates on their deposits in response to changes in market interest rates, the interest sensitivity of money demand is likely to fall. Suppose the market interest rate rose; without deregulation there would be a shift out of deposits into "bonds"; with deregulation, the interest rate on deposits increases, thus limiting the shift out of deposits. At the limit, if interest is earned on bank cash, the interest sensitivity approaches zero. There are several important implications flowing from this. First, monetary policy becomes more effective,1 but on the other hand the effect on income of a money demand shock is amplified. Second, it may become harder to control the money aggregates. At the limit, when the interest elasticity of money approaches zero, open market operations are no longer viable (but on the other hand interest rate policy is still feasible). In any event, to achieve a given interest rate outcome, more pressure will now need to be placed on the money stock. Third, fiscal policy is weakened but, on the other hand, the economy is better insulated from real shocks. Fourth, for any given disturbance the interest rate will now become more volatile. More volatile interest rates, in turn, may also mean higher average interest rates (Argy 1982). What happens in the transition phases? A most important initial effect is the reintermediation process. Deregulation of interest rates allows banks to offer more competitive interest rates on their deposits. If we assume the money stock is fixed, the upward shift in money demand will raise market interest rates, reducing aggregate expenditure. In principle, however, the cash base could be expanded to absorb the increase in money demand but there will be a good deal of uncertainty about the magnitude of the money demand shift, so accommodation will be difficult in practice. During the transition stages the effects of monetary policy will be very uncertain and thus for a while at least it will be more difficult to conduct a monetary policy. Money aggregates will become distorted as signals of monetary policy. However, when the economy has fully adjusted to the deregulation, money aggregates will come into their own again. The initial stages of deregulation will also be associated with substantial changes to the whole financial scene. Competition among financial institutions will be fiercer. Nonbanks will respond to greater competition from the banks by providing new services, perhaps trimming their profit margins and competing more aggressively on interest rates. They may also compete by offering checkable deposits. Thus one undoubted consequence 1
If the original financial system is relatively primitive (in the sense that there are few sources of funds outside the banking system) and credit rationing is in place, deregulation will weaken, not strengthen, monetary policy. This argument may have had some force in the New Zealand case but much less so in Australia.
282
Victor Argy
of deregulation is that banks and nonbanks become more alike. Portfolio restrictions on banks and nonbanks are also likely to be relaxed or abandoned with interest rate deregulation, creating additional ripples on the financial scene. As is well known, financial regulation tends to breed innovations that aim at side-stepping the regulations; so deregulation may set in motion some reversal of previous innovations (although here ratchet effects may be at work) while creating new opportunities and challenges for innovation in the liberalized environment. 4
The choice of exchange rate regime for Australia and New Zealand
Dealing as we are with small economies with similar economic structures and experience, it seems reasonable to apply a common framework in addressing this question concerning the exchange rate. Most of the smaller industrial countries have chosen to either peg or float. Norway, Finland, and perhaps Sweden are still caught in a halfway house between a permanent and an adjustable basket peg (but their heart is set on a peg). Drawing on these tendencies and experiences the serious choices for Australia and New Zealand come down to a peg or to a continuation of the float. They can peg to a single country or to a basket. Japan is the biggest trading partner. It also has a good inflation record. Both Australia and New Zealand also now have an inflation rate (below 2 percent) that parallels Japan's. Would pegging to the yen make sense? In our discussion of this question we draw in part on the optimal currency literature. To begin, efficiency gains (reduction in uncertainty and in the costs of conversion) can be expected to be small (at most perhaps some 0.3-0.4 percent of GDP, going on studies of the EMS). The argument that turns on inflation discipline is weak since as we have seen Australia and New Zealand can achieve that on their own (and indeed the latter now has).2 Any loss of the exchange rate as an adjustment tool between AustraliaNew Zealand and Japan could in principle be counteracted, in part at least, if labor were very mobile across the regions. There is, however, virtually no labor mobility between Australia-New Zealand on the one hand and Japan on the other. 2
Openness is also very widely thought to be a characteristic favoring union (but the theoretical argument here is strictly ambiguous; moreover, it is not an easy concept to apply in this context). Australia and New Zealand are not very open economies so the case on this ground would not be strong.
Monetary policies: Australia and New Zealand
283
Yet another relevant consideration is the degree of real wage flexibility. Countries with a high degree of real wage flexibility are better equipped to cope with real shocks and so will miss the loss of the exchange rate instrument less. At least until recently Australia and New Zealand did not have a high degree of real wage flexibility but as we have seen the two governments are working hard at this. If real wage flexibility can be achieved as well by wage adjustment as by an exchange rate change, the case for a union is much stronger. On the other hand, if there is real wage rigidity, neither a peg nor a flexible rate will function well. Another relevant consideration is the structure of disturbances. If disturbances are common to the region, the case for the union is strengthened. As it happens, on this criterion the case against such a union is strong. Given that Australia and New Zealand are predominantly commodity exporters and Japan is a commodity importer, their terms of trade will often move in opposite directions, which means that at a time commodity prices are depressed, Australia and New Zealand currencies will be appreciating in trade-weighted terms, reinforcing any adverse effects on the current account. We have already noted the loss of the monetary instrument and the potential enhanced effectiveness in the short run of the fiscal instrument. To conclude, the case for pegging to the yen appears on the face of it to be weak (but, on the other hand, many of the preceding considerations would argue for a monetary union between Australia and New Zealand). Argy, McKibbin, and Siegloff (1989) use a three-country model to evaluate the insulation properties of alternative exchange rate regimes, for different assumptions about wage indexation. The three-country world comprises two large countries and one small one. The small country can peg to either of the larger countries or to a basket or it can float. Real or monetary disturbances can originate in either of the larger economies or in the small economy. Using the MSG 2 model and applying the framework to Australia (the small country) and Japan and the United States (the two large countries), they reached the conclusion that "the floating rate regime for Australia performs well . . . except for a shock to money demand in Australia. . . . of the fixed exchange rate regimes it is better to peg to a basket of currencies, especially for foreign shocks in a world of globally floating exchange rates." We can focus here on the choice between a fixed rate of the basket peg variety and a float. Blundell-Wignall and Gregory (1990) also undertake an analysis for Australia of the insulating properties of fixed and flexible rate regimes. They take as a starting point that the two dominant disturbances in an economy such as Australia's (and we add New Zealand's) have come from
284
Victor Argy
money demand and terms of trade shocks. The terms of trade appear to dominate, so their conclusion is that "a floating exchange rate regime is essential in a commodity-exporting country subject to terms of trade shocks if an inflation target is to be achieved." Jonson (1987) using the reserve bank model (RB 11) explores the insulating properties of a fixed and flexible rate regime faced with a money demand and a terms of trade shock. He focuses on output, employment, and inflation outcomes. A fixed rate regime is better on all counts for a money demand shock but worse on all counts for a terms of trade shock. So again if shocks to terms of trade dominate there is a case for adopting flexible rates. To conclude, pegging to a basket does not appear as a viable option. Pegging to a basket also effectively means accepting the weighted inflation rate of your trading partners. Moreover, with such a peg all bilateral rates become unstable and exposed to speculation. 5
Domestic financial deregulation - the Australian experience
In Section 3,1 tried to summarize the principal transitional as well as the steady-state effects of domestic financial deregulation. In this section we ask briefly how domestic financial deregulation impacted on the Australian economy (there is almost no documentation for New Zealand on this). Here I draw on MacFarlane (1991), Harper (1991), Valentine (1992), Hawtrey, Favotto, and Gilchrist (1991), and Twrdy (1992). Efficiency Have there been microefficiency gains, as claimed by the proponents of deregulation? What at first sight appears a relatively straightforward question turns out to be immensely complicated and indeed is still the subject of a very heated debate. It is impossible to do full justice to all the issues here but we can try to summarize the spirit of the debate. It is universally agreed that the range of financial services offered is now much more comprehensive and this is perceived to be a net benefit. One other expectation was that with the entry of new banks and with the lowering of barriers between banks and nonbanks, competition would become fiercer. This is almost certainly true at the wholesale level (for large business customers) but less evident at the retail level, where four major banks continue to dominate the scene. Deregulation was also expected to reduce the profitability of the banks. In general, this has probably happened but the pretax return for Australian banks is high by international standard. However, when adjustment
Monetary policies: Australia and New Zealand
285
is made for international differences in bond rates, this difference disappears. On the other hand, there is some evidence that profitability of retail banking may have increased slightly while that of wholesale banking may have fallen, reflecting perhaps the differences in the level of competition experienced in the two sectors. Much attention has focused on the impacts of deregulation on banks' interest rate margins (i.e., the difference between their borrowing and their lending rates). The original expectation was that deregulation would reduce margins because prior to deregulation the interest rates were set by the Reserve Bank at levels that were highly profitable (compared with those of other OECD countries); banks would become more competitive and this would reduce their operating costs; banks would switch from "noninterest" to interest methods of pricing (i.e., there would be a replacement of implicit interest with explicit interest). Here the signals are very conflicting and indeed confusing. Hawtrey et al. (1991) argue that there has not been a fall in the average net interest margin earned by the established banks. Moreover, they try to demonstrate that there is a disparate sectoral impact of deregulation on margins and charges: interest margins applying to small to medium-sized businesses and households (i.e., the difference between the housing mortgage rate and the return on savings-investment accounts) may have widened later while that for larger firms may have narrowed slightly. This again may reflect the differences in competition on those two fronts. Hawtrey et al. (1991) also argue that there were special structural changes in 1988 (the reduction in the prime asset ratio and the removal of the obligatory reserve requirement) that should have acted to reduce the interest margin. Valentine (1992), however, also argues that there were factors (e.g., the Bank of International Settlements (BIS) capital adequacy controls and high interest rates that prevailed - these last determine the return required on bank capital) working in the opposite direction. To complicate matters further, Twrdy (1992) has recalculated margins and finds that they may actually have fallen in the last part of the 1980s. Finally, we note that there is evidence to suggest that operating costs as a proportion of total assets may have been reduced (Twrdy 1992). Risk In Section 3 we listed the potential links between deregulation and financial fragility. Some of these were clearly manifest in Australia between 1983 and 1988. Bank credit and credit in general grew excessively in those years (on average by half as much). The deregulatory environment contributed in a variety of ways to these developments (Argy 1992a). The
286
Victor Argy
excess credit in turn went not into spending but into asset purchases, which in turn unleashed the asset inflation in those years. Almost certainly, too, there was some reduction in credit standards. Coinciding with the asset inflation was an overheated economy; in fact this combination prompted the reserve bank to tighten monetary policy in the course of 1988. The "bubble" has since burst leaving some banks exposed to bad loans and lower property prices. Macro consequences We argued in Section 3 that there were several potential macro consequences of domestic deregulation. These, however, are difficult to verify directly. It is almost self-evident that the interest elasticity of money demand has fallen (Swamy and Tavlas 1989). Although difficult to verify the hypothesis that interest rates have been destabilized by deregulation, this too is highly likely to have occurred. Many of the transitional "stresses" noted (the distortion to the money aggregates, the difficulties with monetary policy) were clearly also manifest in Australia (and as well in New Zealand) as we have seen. Perhaps one of the most important revelations from any analysis of the impacts of deregulation is that it appears to take considerably longer for its effects to be absorbed into the system than had been suspected. 6
Monetary and exchange rate policies: the Australian experience (1973-84)
We try to summarize the experience with fixed and crawling rates by addressing three key questions. First was there monetary independence (i.e., did the monetary authorities have control over their chosen money aggregate or the domestic interest rates)? Second, how effective were the exchange controls? (Evidently, if they were ineffective, the money stock could not be independent.) Third, if there was some monetary independence, how apt was the policy? The Labour government (1973-75) Money growth had been substantial prior to the change in government at the end of 1972. The terms of trade were favoring Australia, the Australian dollar was undervalued, and there were large inflows which had proved difficult to sterilize (Porter 1974). The controls over inflows that finally came in September-December 1972 had come too late. One of the first tasks of the new government was to rein in money
Monetary policies: Australia and New Zealand
287
growth; this was achieved by tight monetary policies and by a succession of appreciations and a 25 percent cut in tariffs. In the event, money growth (particularly real money balances) fell too sharply creating financial panic at the time. By mid 1974, in the wake of the appreciations-tariff cuts and the recession abroad, the Australian dollar started to weaken and steps were taken to reverse the stance of capital controls. On September 25, after strong speculative outflows, the Australian dollar was devalued. (See Table 11.4 for the trends in the real effective exchange rate.) The Conservative government (1976-82) The new government opted for a "monetarist" strategy, which was now becoming fashionable abroad. Beginning in the financial year 1976-77 money growth targets (officially called projections) were announced. A policy of reducing money growth was adopted with the objective of gradually bringing inflation down. In the budget speech for 1978-79 we find the contention that "within this monetary projection the outlook for employment will be affected by the outcome of the wage determination process both inside the arbitral tribunals and outside them." The authorities did succeed in the first two years in reducing money growth (indeed money growth fell from some 15 percent in 1974-75 to 11 percent in 1975-76 and then to 8 percent in 1976-77). In turn this did reduce the rate of inflation: By 1978 the rate of inflation had more than halved from its heights of nearly 18 percent in 1975 (Table 11.5). In the subsequent three years, however, money growth significantly overshot its target (projected) rate. Although in the two years that followed (1981-82 and 1982-83) money growth was close to target, the policy had in the meantime lost some credibility. Despite the failures here it is well worth noting that although the objectives of the policy failed, money growth over a period of six years was remarkably stable (in sharp contrast to the volatility under the previous government). On the external front in the course of 1976 the Australian dollar was again weak (Table 11.4) and by the third quarter there were large speculative outflows. In November there was another devaluation, after which the authorities switched to a crawling peg. Late in 1977 and 1978 the current account took a further turn for the worse (Table 11.1). The authorities responded in two ways. First, they reversed the stance of capital controls, progressively dismantling all restrictions on inflows. Second, they began to adjust the effective rate downward on a gradual basis. The distinctive feature of the period mid-1980 to 1981 was the short-
Table 11.4. Australia: index of competitiveness (unit labour costs 1984-85 = 100) Mar 72 Sep72 Dec 72 Mar 73 Jun73 Sep73 Dec 73 Mar 74 Jun74 Sep74 Dec 74 Mar 75 Jun75 Sep75 Dec 75 Mar 76 Jun76 Sep76 Dec 76 Mar 77 Jun77 Sep77 Dec 77 Mar 78 Jun78 Sep78 Dec 78 Mar 79 Jun79 Sep79 Dec 79 Mar 80 Jun80 Sep80 Dec 80 Mar 81 Jun81
98.9 96.3 96.0 104.8 109.4 108.3 114.7 123.1 120.6 131.1 117.8 118.6 117.6 119.8 123.3 121.8 122.3 123.4 116.2 109.7 109.1 108.7 108.0 105.2 101.0 96.4 93.7 93.5 94.4 95.0 95.3 97.1 96.9 99.3 98.8 102.5 106.6
Sep81 Dec 81 Mar 82 Jun82 Sep82 Dec 82 Mar 83 Jun83 Sep83 Dec 83 Mar 84 Jun84 Sep84 Dec 84 Mar 85 Jun85 Sep85 Dec 85 Mar 86 Jun86 Sep86 Dec 86 Mar 87 Jun87 Sep87 Dec 87 Mar 88 Jun88 Sep88 Dec 88 Mar 89 Jun89 Sep89 Dec 89 Mar 90 Jun90 Sep90
Note: An increase means a loss in competitiveness. Source: (Australian) Economic Planning and Advisory Council (EPAC).
113.1 112.1 113.9 116.0 115.7 113.3 109.0 100.0 102.9 106.8 108.7 108.3 105.2 107.7 100.5 86.2 86.4 82.5 81.5 79.9 68.7 71.8 72.3 74.1 76.6 72.4 73.4 79.4 86.0 88.5 91.1 88.5 88.6 92.2 90.4 93.9 95.8
Monetary policies: Australia and New Zealand
289
Table 11.5. Australia: Money growth targets and outcomes
Financial year
M3 money growth target
M3 outcome
1975-76 1976-77 1977-78 1978-79 1979-80 1980-81 1981-82 1982-83 1983-84 1984-85
10-12 8-10 6-8 max 10 9-11 10-11 9-11 9-11° 8-10*
(13.3) 11.0 8.0 11.8 12.3 12.7 11.3 11.1 11.4 (17.5)
"Revised to 10-12 at midyear. ^Suspended January 1985. Source: Jonson (1987).
lived (but nevertheless important in terms of its impacts) resources boom, which came in the wake of the second oil price shock. Australia suddenly became the recipient of huge amounts of capital seeking to take advantage of what looked like very attractive investment prospects, particularly in coal, mineral resources, and allied industries. Capital inflows, some of which were also speculative, more than offset the current account deficit, which now again deteriorated sharply. This led to a potentially very large overall surplus. The authorities responded in part by allowing the currency to appreciate; by about August 1981 the real effective rate had risen by some 12 percent (Table 11.4). At the same time there were large official purchases of U.S. dollars, notably in the first half of 1981. The authorities, however, found it difficult to sterilize these purchases and money supply grew rapidly, exceeding again its target for the year 1980-81. Thus the authorities tried to moderate the appreciation but at the cost of some loss of control over monetary policy. It is interesting to note that this time the authorities did not try to reimpose controls over inflows, although on the face of it this might have helped in part to resolve the dilemmas. On the other hand, some modest measures were taken at the time to encourage more outflows. With the price of oil beginning to fall back by 1982 and the resources boom exhausting itself, increased attention came to be placed on the current account, which in 1982 continued to deteriorate rapidly. Capital in-
290
Victor Argy
flows, however, continued to be large, more than offsetting the current account deficit. Nevertheless, now the authorities, with an eye on the current account, chose to gradually lower the Australian dollar and buy U.S. dollars in the foreign exchange market. From late 1982 to mid 1983 the dollar in real terms had fallen by roughly as much as the gains made in 1980-81 (Table 11.4). The Labour government - March 1983 on In February 1983 with an election campaign under way, fears that the Labour Party might win government led initially to some cautionary outflows of capital. Those outflows gathered momentum and, from late February, reached massive proportions. On March 8, with the new Labour government installed, the dollar was devalued by 10 percent. With this devaluation the "crawling peg" system, under which only small daily adjustments would be made, had broken down. The current account was beginning to show some improvement in 1983; at the same time, the interest rate differential moved more in favor of Australia; moreover, some confidence in the new government was beginning to surface. These forces combined to produce a dramatic reversal in capital flows. In the months that followed inflows became massive; indeed, for a while they were almost double the current account deficit. As in 1980-81 the authorities responded, partly by allowing the dollar to appreciate and partly by intervening in the foreign exchange market. As in 1980-81 they found it impossible to sterilize the inflow. During those months money growth was overshooting its projected range (revised upward by one point in December). Finally, on December 12 the Australian dollar was allowed to float; at the same time most exchange controls were removed. The float enabled the authorities to regain control over money growth in the second half and this allowed money growth over the whole financial year 1983-84 to remain close to its target range. The Australian experience summarized We return to the three original questions put. (For details, see Argy 1987.) There was clearly no monetary independence in 1972, in 1980-81, and again in the course of 1983. These were all periods of intense speculation in favor of the Australian dollar. It needs to be emphasized, importantly, that on all these occasions there were no effective controls over inflows and so the economy was left exposed. It is more difficult to judge whether there was monetary independence in other years. Undoubtedly the domestic interest rate could not diverge
Monetary policies: Australia and New Zealand
291
too sharply from foreign interest rates; so in this respect there was some constraint. Nevertheless, with the help of discrete devaluations and in less turbulent periods, one judges that there was probably some degree of monetary independence. Were the capital controls effective either in stemming inflows when the restrictions were applied or in mitigating the outflows? Argy (1987) reviews all the evidence from several quarters and concludes that there was probably some effectiveness; however, their effectiveness weakened in the longer run or by continued application. Testimony to the fact that the controls over outflows were not completely effective comes from the fact that there were large and increasing speculative bouts (in 1974, in 1976, and in early 1983). Leakages came primarily in the form of the activities of the multinationals and of the traders (leads and lags); in due course the authorities did move to regulate the timing of payments and to increase surveillance over multinational accounts but, at best, these would only have had a marginal impact. Argy also suggests that if the controls were effective in some degree the timing of their imposition, removal, or reimposition did leave something to be desired. In short, they were not applied judiciously. We saw that in 1972 the controls came too late. The reversal in 1974 was again late in coming. The six-month embargo on inflows was retained from November 1974 to January 1977, during which the currency was weak. Nor was there any attempt made to reintroduce controls over inflows either in 1980 or in 1983 when those inflows assumed massive proportions. Again at a time when net outflows ought to have been encouraged there was very little attempt, except for a weak one in 1980, to ease up on controls over outflows. The third question addressed was the aptness of the policies. This is very difficult to answer. We just indicated that there could have been more judicious use made of capital controls. We also indicated that there were times when monetary policy was too volatile. Discrete devaluations in an environment where wages are predominantly indexed are ultimately selfdefeating and they probably were so in Australia. If one assumes some monetary discretion, the only way to judge the policy is to compare it with an alternative feasible strategy, which could potentially include a float or a different monetary policy rule. Only econometric simulations of alternative strategies over the period 1972-82 would throw light on this question and there are none. Crudely, too, judging the monetary policy in terms of inflationary outcomes, we have already seen that Australia's relative performance was a mixed one: It was worse than the OECD in nearly every year between 1972 and 1982 (but, at least between 1979 and 1983, better than the smaller countries).
292
Victor Argy
7
Monetary and exchange rate policies: the New Zealand experience (1972-84)
The New Zealand experience with fixed and crawling rates is much simpler to summarize and can be described in very broad terms. The consensus view appears to be that there was in general more monetary independence in New Zealand but that this independence was being increasingly threatened with time. Money growth in New Zealand also fluctuated wildly at least till 1976, after which until 1982 the fluctuations were more moderate; however, the average growth rate was probably excessive, which in turn explains New Zealand's particularly poor performance on the inflation front. New Zealand never did adopt a policy of announcing money growth targets (Spencer 1990). In the course of the three years, 1979-82, when the crawl was in operation, the current account steadily deteriorated. In principle, the adoption of the crawl did not rule out changes in the real exchange rate but these would have had to be implemented in small doses (a policy that might not have been credible). In the event in 1982, confronted with such an imbalance, the authorities reverted to the adjustable peg. In one sense the timing of the adoption of the crawling peg was perhaps unfortunate. The terms of trade in those years moved significantly against New Zealand (Table 11.3). The breakdown parallels the Australian experience in 1983 when the huge outflows were no longer consistent with small changes in the currency. Also in Australia in 1980-81 during the resources boom when the imbalances were very large, this placed great strains on the operation of the crawling peg. Exchange controls were probably more effective in New Zealand (Nicholl 1977), but became increasingly less effective (Carey and Duggan 1986). It is now widely conceded that by 1984 the exchange controls were no longer functioning well. Capital flows in that year were responding massively to perceived arbitrage or speculative gains. To cite Carey and Duggan "with wide scope for currency flows within the current account it made little sense to focus control on the capital account. Apart from the relative ineffectiveness in controlling foreign exchange flows, the regulations imposed inordinately heavy costs on compliance and enforcement of the rules." 8
The float in Australia and New Zealand Australia
Soon after the Labour government took over in March 1983 the policy of announcing money growth targets was abandoned. The principal reason
Monetary policies: Australia and New Zealand
293
was that the domestic financial deregulation was severely distorting the money aggregates. A secondary reason was that the policy had also lost some credibility in Australia and, by then, overseas. The authorities then adopted a "checklist" approach to monetary policy. In determining the setting of monetary policy the authorities looked at all major economic and financial factors, present and prospective. These included the state of the economy, the balance of payments, prices, other policies, interest rates, the exchange rate, and the money aggregates. These were periodically reviewed and a judgment was then made about the appropriate monetary policy. In retrospect it is possible to be critical of the conduct of monetary policy. One view widely held was that policy was too easy in 1986-88 and then too tight for too long in the face of an overheated economy. The tight monetary policy was primarily responsible for the sharp downturn in the economy in 1990. New Zealand As in Australia's case an important aim of the float in New Zealand was to regain and secure control over monetary aggregates and policy. Spencer (1992) notes: Following the float of the New Zealand dollar in March 1985, the Reserve Bank achieved discretionary control over the monetary base of the New Zealandfinancialsystem. The Bank was therefore in a position, for the first time, to effectively pursue targets for the money and/or credit aggregates; a strategy which had been continually frustrated under the "adjustable-peg" exchange rate regime. New Zealand's monetary policy during the float followed a course roughly similar to Australia's. Policy eased in 1985-87, then became very tight. The principal difference is that New Zealand in due course adopted an inflation target and its policies were intended while in Australia the outcomes were unintended. As in the Australian case, too, the money aggregates became seriously distorted after domestic deregulation (Spencer 1992). New Zealand adopted its own version of the checklist. Australia used a whole range of variables to determine the stance of its monetary policy. New Zealand has an inflation target; it also monitors the settlement cash balances held by banks at the reserve bank to achieve its monetary stance. In determining this stance it has been guided since 1985 by a set of indicators that include the exchange rate, the level and structure of interest rates, the growth of money and credit, inflation expectations, and trends in the real economy. As Spencer (1990) notes:
294
Victor Argy The relative emphasis between interest rates and the exchange rate has shifted over time, but these two principal indicators together have played a far greater role in guiding policy over 1985-90 than the money and credit aggregates . . . the major distorting effects of reintermediation occurred over 1985-6.
The unemployment cost of the recent disinflations in Australia and New Zealand In Australia between 1988 and 1991-92 the inflation rate fell by some 7 points. At the same time the unemployment rate rose by some 5 points over some three years (a cumulative excess over the 1988-89 figure of 9.5 points). In New Zealand inflation fell by some 12 points between 1987 and 1992 while the unemployment rate rose by 6.5 points. The cumulative excess unemployment is some 14 points. On these very crude calculations the cumulative disinflation cost is slightly less in New Zealand. The Australian experience reported here appears to be a little higher than that of disinflation costs reported for previous episodes of disinflation (see Carmichael 1990). For more sophisticated measures of the disinflation costs using econometric simulations, see Murphy (1991) and Andersen (1992). If one accepted the view that labor market deregulation reduces the costs of disinflation, the argument can be made that both governments put the cart before the horse, in the sense that they should have deregulated the labor markets before embarking on a disinflation policy. 9
The 1983-85 deregulation in Australia and New Zealand
Between 1983 and 1985 the two countries deregulated the exchange rate system; the domestic financial system; and the exchange control regime. As a matter of speculation it might be interesting to ask how these developments might have affected the capacity of the economy to absorb shocks. Table 11.6 presents, in summary form, the results of just such an analysis using a modified fixed wage Mundell-Fleming model (for details, see Argy 1991). There are three types of disturbances: expenditure, money demand, and productivity (supply). Each of these three can originate at home or abroad; there are thus six disturbances in toto. Table 11.6 is divided into three parts, corresponding to each of the three deregulatory changes. In each case we ask whether, for each new regime, the disturbance has stabilized or destabilized output, home prices, the interest rate, and the exchange rate. The first part of Table 11.6 summarizes
Table 11.6. Implications of three deregulatory changes for insulation - Australia 1983-85
Shift from fixed to flexible rates
External deregulation (lifting exchange controls)"
Domestic financial deregulation (money demand impact)"
Disturbance
y
pd
r
e
y
pd
r
e
y
pd
r
e
At home Expenditure Money demand Productivity
4 T
4 T
T T
4 T
4 T
T
7
7
7
t t
4 T
7
4 4 4
4 T
7
T T T
T
?
?
T
t t
T
T* 4*
T* 4*
T T T
T T T
4 4 4
4 4 4
T
7
T T T
t
7
4 4 4
t
4 4 4
Expenditure Money demand Productivity
T T
t
T T
t
T
Note: y = output, pd = domestic prices, r = interest rate, e = exchange rate. T(4) means the particular variable is destabilized (stabilized) by the change in regime. The results assume capital mobility is relatively high at point of departure. a Assumes flexible rates in operation. * Probable outcome. Source: Argy (1991).
296
Victor Argy
the insulating properties of the shift toflexiblerates; the second part, the insulating properties of external deregulation; and the third part, the insulating properties of domestic deregulation (in particular the implications of a lower interest rate sensitivity of money demand). It is impossible to explain in detail how each one of the results is arrived at but some explanation of the adjustment mechanism underlying the results is in order. We focus on a domestic expenditure disturbance and a money demand disturbance abroad. We dealfirstwith the effects of a shift from a relativelyfixedto a flexible rate regime. Thefixedrate regime is one where monetary policy is assumed to be used to stabilize the currency. Finally, the macro effects of domestic deregulation are mixed, but perhaps on balance marginally favorable. The interest rate is invariably destabilized; however, for many disturbances price and output are stabilized. At the same time the exchange rate is stabilized for external but, in general, destabilized for domestic disturbances. References Adams, Charles, and Daniel Gros. (1986). The Consequences of Real Exchange Rate Rules for Inflation: Some Illustrative Examples. IMF Staff Papers 33, no. 3 (September): 439-76. Alesina, AJberto, and Lawrence H. Summers. (1990). Central Bank Independence and Macroeconomic Performance: Some Comparative Evidence. Unpublished manuscript. Cambridge, Mass.: Harvard University, June. Anderson, P. S. (1989). Inflation and Output: A Review of the Wage Price Mechanism. Bank for International Settlements (BIS) Economic Papers 24 (January). (1992). OECD Country Experiences with Disinflation. Paper presented at conference organized by Reserve Bank of Australia, July. Argy, Victor. (1982). Exchange-rate Management in Theory and Practice. Princeton: International Finance Section, Department of Economics, Princeton University. (1987). International Financial Liberalisation - The Australian and Japanese Experience Compared. Bank of Japan Monetary and Economic Studies 5, no. 1 (May): 105-67. (1990). Choice of Exchange Rate Regime for a Smaller Economy: A Survey of Some Key Issues. In Victor Argy and Paul De Grauwe (eds.), Choosing an Exchange Rate Regime: The Challenge for Smaller Industrial Countries. Washington, D.C.: IMF, 55-78. (1991). How Does Deregulation Affect the Capacity of an Economy to Absorb Shocks? Manuscript, October. (1992a). Australian Macroeconomic Policy in a Changing World Environment (1973-90). Sydney: Allen and Unwin. (1992b). Equilibrium Rate of Inflation with Discretion and Some Reputation. Quarterly Review Banca Nazionale del Lavoro, Italy, no. 182 (September): 329-47. Argy, Victor, Warwick McKibbin, and Eric Siegloff. (1989). Exchange Rate Re-
Monetary policies: Australia and New Zealand
297
gimes for a Small Economy in a Multi-Country World. Princeton Studies in International Finance 67 (December): 123-47. Barro, Robert I , and D. B. Gordon. (1983). Rules, Discretion and Reputation in a Model of Monetary Policy. Journal of Monetary Economics 12: 101-21. Blundell-Wignall, Adrian, and Frank Browne. (1991). Increasing Financial Market Integration, Real Exchange Rates and Macro Economic Adjustment. Paris: OECD Working Paper no. 96, February. Blundell-Wignall, Adrian, and Robert G. Gregory. (1990). Exchange Rate Policy in Advanced Commodity - Exporting Countries: Australia and New Zealand. In V. Argy and Paul De Grauwe (eds.), Choosing an Exchange Rate Regime: The Challenge for Smaller Industrial Countries. Washington, D.C.: IMF, 224-71. Calmfors, Lars, and John Driffill. (1988). Bargaining Structure, Corporatism and Macroeconomic Performance. Economic Policy 6 (April): 13-62. Carey, D. A., and K. G. Duggan. (1986). The Abolition of Exchange Controls: Financial Policy Reform. New Zealand: Reserve Bank of New Zealand. Carmichael, Jeffrey. (1990). Inflation: Performance and Policy. In S. Grenville (ed.), The Australian Macro-Economy in the 1980s. Sydney: Reserve Bank of Australia. Dornbusch, Rudiger. (1982). PPP Exchange-Rate Rules and Macroeconomic Stability. Journal of Political Economy 90, no. 1 (February): 158-65. Fischer, S. F. (1988). Rules Versus Discretion in Monetary Policy. Cambridge, Mass.: NBER Working Paper no. 2518, February. Flood, Robert P., Jagdeep S. Bhandari, and Jocelyn Home. (1989). Evolution of Exchange Rate Regimes. IMF Staff Papers 36, no. 4 (December): 810-35. Friedman, Milton. (1953). The Case for Flexible Exchange Rates. In Milton Friedman (ed.), Essays in Positive Economics. Chicago: University of Chicago Press, 53-89. Harper, Ian R. (1991). Bank Deregulation in Australia: Choice and Diversity, Gainers and Losers. Paper prepared for the Reserve Bank of Australia Conference on Deregulation of Financial Intermediaries, Sydney, May. Hawtrey, Kim, Ivo Favotto, and David Gilchrist. (1991). The Impact of Bank Deregulation on Australian Manufacturers. Economic Papers (Economic Society of Australia) 10, no. 4 (December): 10-29. International Monetary Fund. International Financial Statistics. Washington, D.C. Various issues. Jonson, P. D. (1987). Monetary Indicators and the Economy. Reserve Bank Bulletin of Australia 50 (December): 241-65. Kerr, Roger. (1991). Freedom of Contract - NZ's Answer to Industrial Constriction. Institute of Public Affairs (IPA) Review 8 (autumn): 11-13. MacFarlane, I. J. (1991). The Lessons for Monetary Policy. Paper prepared for the Reserve Bank of Australia Conference on Deregulation of Financial Intermediaries, Sydney, June. Meltzer, Allan H. (1988). On Monetary Stability and Monetary Reform. In Yoshio Suzuki and Mitsuaki Okabe (eds.), Toward a World of Economic Stability. Tokyo: University of Tokyo Press, 41-63. Murphy, Chris W. (1991). The Transitional Costs of Reducing Inflation Using Monetary Policy in Australia's Inflation Problem. Economic Planning Advisory Council Background Paper no. 11, April. Nicholl, P. W. E. (1977). New Zealand Monetary Policy in the 1970s: Analysis and
298
Victor Argy
Perspective. Reserve Bank of New Zealand Research Paper no. 23, December, 7-19. Organisation for Economic Cooperation and Development (OECD). Annual Surveys of Australia and New Zealand. Porter, M. G. (1974). The Interdependence of Monetary Policy and Capital Flows in Australia. Economic Record 50: 1-20. Reserve Bank of New Zealand. (1986). Financial Policy Reform. Auckland. (1992). Monetary Policy and the New Zealand Financial System. 3rd ed. Auckland. Reserve Bank of New Zealand Quarterly Bulletins. Rogoff, K. (1987). Reputational Constraints on Monetary Policy. CarnegieRochester Conference Series on Public Policy 26: 141-82. Spencer, G. H. (1990). Monetary Policy: The New Zealand Experience 1985-1990. Reserve Bank Bulletin 53, no. 3: 252-66. Spencer, Grant. (1992). Monetary Policy: The New Zealand Experience 19851990. In Monetary Policy and the New Zealand Financial System. 3rd ed. Auckland: Reserve Bank of New Zealand, 21-38. Swamy, P. A. V. B., and George S. Tavlas. (1989). Financial Deregulation, the Demand for Money, and Monetary Policy in Australia. IMF Staff Papers 36, no. 1 (March): 63-101. Twrdy, Karen. (1992). Has Financial Deregulation Improved the Efficiency and Competitiveness of the Banking Sector? Honours thesis, Macquarie University, June. Valentine, Tom J. (1992). The Impact of Banking Deregulation: Further Thoughts. Economic Papers (Economic Society of Australia) 11, no. 2 (February): 87-93. Wells, Graeme. (1990). Economic Reform and Macroeconomic Policy in New Zealand. Australian Economic Review, 4th quarter, 92: 45-60. Williamson, John. (1981). The Crawling Peg in Historical Perspective. In John Williamson (ed.), Exchange Rate Rules: The Theory, Performance and Prospects of the Crawling Peg. New York: Macmillan, 63-93.
PART IV
Capital mobility and exchange rates in Latin America
CHAPTER 12
Exchange rates, inflation, and disinflation: Latin American experiences Sebastian Edwards
1
Introduction
In spite of the collapse of the Bretton Woods system in 1973, most of the developing countries, including many Latin American nations, continued to rely heavily onfixedexchange rates throughout the 1970s. The December 1979 issue of International Financial Statistics (IFS) reports that thirteen Latin American countries had a fixed exchange rate system at that time.1 Many of these countries, in fact, had had a fixed exchange rate for a very long period of time - in some cases, such as Guatemala, since the 1920s. Recently, however, a large number of developing countries have adopted more flexible exchange rate regimes. For example, according to the December 1990 issue of the IFS only three Latin American countries Dominican Republic, Haiti, and Panama - had afixedexchange rate system at that time.2 This abandonment of fixed exchange rates was, to a considerable extent, associated with the debt crisis unleashed in 1982. In order to make major resource transfers to their creditors, the vast majority of the Latin American countries adopted adjustment packages that included as a key component very large nominal devaluations. These devaluations differed in an important way from the traditional norm in Latin America. Most historical exchange rate adjustments in the region were I am grateful to Abraham Vela and Fernando Losada for excellent assistance. I thank my discussant Yoon-ye Cho for useful comments. 1 This count excludes non-Latin countries in the Caribbean, many of which also had fixed rates. For a list of the countries see Table 12.1. 2 The IFS distinguishes several categories of fixed exchange rate countries, including those pegged to the U.S. dollar, to the French franc, and to a "composite of currencies." It is unclear, however, to what extent the countries in this latter group have indeed followed a policy of pegging their currency to a basket. For all practical purposes, if a country alters continuously the composition of the basket the resulting policy will not be one of a pegged exchange rate, but rather a form of exchange rate management.
301
302
Sebastian Edwards
followed by the establishment of a new peg; however, in the 1980s almost every one of these large devaluations was followed by the adoption of some type of managed (adjustable) exchange rate regime. The nominal devaluations, and the subsequent active exchange rate policy adopted by much of Latin America during the 1980s, were strongly encouraged by the International Monetary Fund (IMF). In fact, after the debt crisis of 1982, the fraction of IMF programs that included exchange rate adjustments was significantly higher than in the precrisis period. For instance, whereas all of the high conditionality programs enacted in 1983 included a devaluation condition, only 50 percent of the programs enacted during 1977-80 contemplated exchange rate action.3 A number of authors have recently argued that by allowing countries to adopt administered exchange rate systems characterized by frequent small devaluations, IMF programs have become excessively inflationary. In particular, it has been argued that crawling peg regimes introduce substantial inertia into the inflationary system. According to this view exchange rate policy in the developing countries should move toward greater rigidity - and even complete fixity - as a way to introduce financial discipline and credibility, provide a nominal anchor, and reduce inflation.4 In the 1990s the debate on exchange rates as nominal anchors acquired a new dimension, as the countries of eastern Europe and the former Soviet Union began to reform their economies. Many countries in the region, including Poland, Czechoslovakia, and Yugoslavia, adopted a fixed exchange rate as a fundamental component of their anti-inflationary programs. In the words of the IMF, this policy of "pegged exchange rate was designed to provide an effective anchor for stabilizing prices" (IMF 1992, 109). However, as in the developing countries, the adoption of fixed exchange rates as a way to bring down inflation has generated important debates in eastern Europe. Some authors have argued that this approach is likely to generate a real exchange rate appreciation, putting the balanceof-payment targets in jeopardy.5 The discussion on the desirability of fixed exchange rates has also acquired prominence in East Asia, where some observers - including IMF staff - have argued that the adoption of managed exchange rates has piayed an important role in the inability to reduce inflation rates, in most countries in the region, to international levels. 3
4
5
More specifically, all 1983 programs where a devaluation was institutionally possible - that is, where it was not ruled out by international arrangements, as in the Franc Zone in Africa - included that condition. See Edwards (1989b). For aflavorof the discussion within the IMF see, e.g., Burton and Gillman (1991), Aghevli, Khan, and Montiel (1991), Flood and Marion (1991), and Aghevli and Montiel (1991). In Edwards (1992b) I deal with some of these issues. See Nunnenkamp (1992).
Exchange rates, inflation, and disinflation
303
The purpose of this essay is to analyze some aspects of the relationship between exchange rates, inflation, and disinflation in Latin America. In particular, I deal with two central issues. First, in Section 2, I examine Latin America's historical experiences with fixed exchange rates. I show that in many countries this system worked relatively well until the mid1970s. In fact, the evidence provides some support to the view that in a number of cases fixed exchange rates acted for long periods of time as a constraint on central banks' behavior. However, the analysis also shows that the mere existence of a fixed exchange rate was not a sufficient condition for avoiding inflation outbursts and balance of payments crises. More specifically, the evidence suggests that having maintained a fixed exchange rate during a period of large negative terms of trade shocks in the late 1970s and early 1980s had major negative effects on these countries. The second issue addressed in this chapter is the relationship between exchange rate policy and inflationary inertia. In Section 3,1 argue that one of the costs of abandoning fixed exchange rates has been, in many countries, a substantial increase in the degree of inflationary inertia. I then investigate whether the (re)adoption of a fixed exchange rate helps break inertia and achieve disinflation in a smoother fashion. 2
Latin America under fixed exchange rates
Contrary to popular mythology, not all countries of Latin America have had long inflationary histories. In fact, for many years a number of countries in the region had extremely low inflation rates - lower, even, than those in the industrial countries. Invariably the historically low-inflation Latin American nations - many of which are in Central America - have had a fixed exchange rate regime. However, and as a result of a number of circumstances including the debt crisis, by 1990 no country in the region had a genuine fixed rate system.6 In determining the desirability of returning to fixed rates it is useful to evaluate how this system worked in the region in the past. In particular, it is important to analyze empirically whether the existence of a fixed exchange rate system constrained the ability to conduct monetary policy, and whether domestic inflation was kept in line with world inflation. In this section I first deal with some analytical aspects of the connection between fixed exchange rates, credibility, and inflation. I then discuss the historical record on the performance under fixed exchange rates in four Latin American countries. 6
The exception to this was Panama, which, as is well known, uses the U.S. dollar as its currency.
304
Sebastian Edwards
2.1
Fixed exchange rates and credibility: analytical aspects
Much of the recent enthusiasm for fixed nominal exchange rates is intellectually rooted on the modern credibility and time consistency literature.7 According to this approach, which was pioneered by Calvo (1978) and Kydland and Prescott (1977), governments that have the discretion to alter the nominal exchange rate will tend to abuse their power, introducing an inflationary bias into the economy. The reason for this is that under a set of plausible conditions, such as the existence of labor market rigidities that preclude the economy from reaching full employment, it will be optimal for the government to "surprise" the private sector through unexpected devaluations.8 By engineering these unexpected devaluations, the government hopes to induce a reduction in real wages and, thus, an increase in employment and a boost in output. Naturally, in equilibrium the public will be aware of this incentive faced by the authorities, and will react to it by anticipating the devaluation surprises and hence rendering them ineffective. As a consequence of this strategic interaction between the government and the private sector, the economy will reach a high inflation plateau. What is particularly interesting about this result is that this inflationary bias will be present even if it is explicitly assumed that the government has an aversion for inflation. This is because the government perceives that the marginal benefits of higher inflation - associated with the increase in employment once nominal wages have been set - outweigh its marginal costs.9 An important feature of the credibility literature is that, under most circumstances, policy commitment is welfare superior to discretionary policy. If the government can credibly commit itself to low (or no) inflation, society will be better off: Employment will be the same as in the discretionary policy case, but inflation will be lower. The problem, however, is that governments have a hard time making credible commitments. In the absence of effective constraints that will tie the government's hands, any promise of low inflationary policy will not be credible and, thus, will be self-defeating. A key policy implication of this literature is that defining (and implementing) constraints that will make government precommitments credible, will result in an improvement in society's welfare. It is here where fixed 7
8
9
The new impetus for fixed rates has strongly emerged in the International Monetary Fund. See Aghevlietal. (1991). This assumes that wages are set before the government implements the exchange rate policy, but after it has been announced. See Persson and Tabellini (1990).
Exchange rates, inflation, and disinflation
305
exchange rates come into the picture. It has been argued that the adoption of a fixed exchange rate will constrain governments ability to surprise the private sector through unexpected devaluations. Promises of fiscal discipline will become credible and private sector actions will not elicit successive rounds of inflationary actions.10 In particular, it has been argued that fixed exchange rates provide a reputational constraint on government behavior. The authorities know that if they undertake overly expansive credit policy they will be forced to abandon the parity and devalue. As the recent (mid-1992) crisis of the ERM has shown, exchange rate crises can indeed shatter the reputation of politicians. In spite of its elegant appeal, this view has, in its simplest incarnation, some serious problems. First, in these simple settings exchange rate policy has a very limited role. In fact, in most of these models its only effect is to alter the domestic rate of inflation and, through it, the government perceives it as altering real wages. However, in most modern exchange rate models, nominal devaluations can also help accommodate shocks to real exchange rate (RER) fundamentals - including shocks to the terms of trade - helping to avoid RER misalignment.11 Second, in economies with stochastic shocks, contingent exchange rate rules can, at least in principle, be superior to fixed rates (Flood and Isard 1988). Third, it is not clear why a country that can credibly commit itself to unilaterally fixing the exchange rate cannot commit itself to maintaining a fixed stock of domestic money.12 In recent essays, Devarajan and Rodrik (1992) and Kamin (1991), among others, have addressed the question of the desirability of fixed exchange rates from a more general perspective. For example, in Devarajan and Rodrik (1992), policy makers face a trade-off regarding exchange rate policy: While exchange rate flexibility has an inflationary bias, it also allows the country to reduce output variability. This is accomplished by smoothing, via exchange rate adjustments, the consequences of terms of trade shocks on output. In this model it is not possible to rank a priori fixed and flexible (or active) exchange rate regimes. For large terms-oftrade shocks it is more likely that flexible exchange rates will be superior. Likewise, the more vulnerable the real economy is to these terms of trade shocks, the more desirable will flexible arrangements become. On the 10 11 12
Aghevli et al. (1991). See, e.g., Edwards (1988). By "unilaterally" I mean that the fixed rate is not enforced by a multilateral institutional arrangement such as the European Monetary System or the West African Monetary Union. The recent crisis in the ERM of the EMS has shed serious doubts on the ability of such a system to actually "anchor" exchange rates.
306
Sebastian Edwards
other hand, the greater is the government's built-in inflationary bias, the greater will be its temptation to abuse devaluations, and the less desirable will a flexible arrangement become. John Williamson (1991) has recently dealt with the issues of exchange rate regimes in developing nations. He has argued that a fixed exchange rate would be advisable as long as four conditions are met: (1) the country in question is "small" relative to the rest of the world; (2) the bulk of its international trade is undertaken with the country (or countries) with respect to which it plans to peg its currency; (3) the country wishes to have a rate of inflation similar to that of the country it is pegging its currency to; and (4) there are institutional arrangements that assure that the commitment to a fixed rate is credible. Williamson goes on to argue that, once these four conditions are met, the only remaining argument in favor of flexible exchange rates refers to terms-of-trade shocks. Although Williamson's list of requirements for a successful fixed exchange rate is eminently plausible, it leaves a number of unanswered questions. In particular, his analysis is silent on whether the adoption of fixed exchange rates provides, by itself, an institutional constraint on fiscal policy. More specifically, Williamson's discussion does not address the issue of whether the existence of a fixed exchange rate imposes reputational or other type of constraints on politicians. In the rest of this section I use data on a group of Latin American countries to evaluate the behavior of fixed exchange rates in the 1950s through 1980s period. 2.2
Latin American experiences with fixed exchange rates
Table 12.1 contains a list of the Latin American countries that had a fixed exchange rates regime in the period 1979-91. As can be seen, the number shrunk considerably during this period. Panama was, in fact, the only country that maintained a fixed nominal exchange rate regime uninterruptedly.13 Many of the nations that faced exchange rate crises during the 1980s had a fixed parity with respect to the U.S. dollar since, at least, the inception of the Bretton Woods regime in the 1940s. Their long experience provides a unique opportunity to study the functioning of fixed nominal exchange rates in small open economies. In this subsection I evaluate the experience with fixed exchange rates in the four Latin American countries that had a fixed exchange rate with respect to the U.S. dollar (at least) between 1950 and 1983 - Dominican Republic, Guatemala, Honduras, and El Salvador. I deal with two issues: 13
This, of course, is not much of a surprise and is related to the fact that Panama does not have a central bank and uses the U.S. dollar as its currency.
Exchange rates, inflation, and disinflation
307
Table 12.1. Fixed exchange rate countries in Latin America: 1979-91 1979
1982
1985
1990
1991
Bolivia Chile Costa Rica Dom. Rep. Ecuador El Salvador Guatemala Haiti Honduras Nicaragua Panama Paraguay Venezuela
Dom. Rep. Ecuador El Salvador Guatemala Haiti Honduras Mexico Nicaragua Panama Paraguay Venezuela
Guatemala Haiti Honduras Nicaragua Panama Paraguay Peru Venezuela
Dom. Rep. Haiti Panama
Argentina Nicaragua Panama
Note: With the exception of Haiti all these countries fixed their exchange rate to the U.S. dollar. Source: International Monetary Fund.
First, I analyze the inflationary experiences in these countries between the 1950s and the early 1980s, focusing on whether these countries' rates of inflation differed from those in the United States. Second, I investigate the extent to which fixed exchange rates precluded these countries from undertaking independent monetary policy. In Section 2.3, I expand the discussion to include the exchange rate crises that affected these four countries in the 1980s. An important characteristic of our four countries during the 1950s, 1960s, and 1970s is that they were extremely open economies, with relatively low import tariffs and, for all practical purposes, nonexisting capital controls.14 In a way these four countries provide an almost ideal real world counterpart to the textbook notion of "small open economy with fixed exchange rates." Table 12.2 contains data on quarterly rates of inflation for the Dominican Republic, Guatemala, Honduras, and El Salvador for the period 1957-83, and for three subperiods: 1957-72, 1973-78, 1979-83. Also, data on inflation in the United States are provided for comparison purposes. The first subperiod goes from 1957, the first year for which we have quar14
In I960 the Central American Common Market was created. However, its common external tariff policy was moderate, maintaining a fairly high degree of trade openness in the region.
308
Sebastian Edwards
Table 12.2. Inflation in fixed exchange Latin American countries: quarterly average, 1957.1-1983.4 (percent) Dominican Republic
1957-83 1.31 1957-72 0.44 1973-78 2.57 1979-83 2.34
Guatemala
Honduras
El Salvador
United States
1.15 0.21 3.19 1.81
1.24 0.50 1.84 2.80
1.49 0.23 3.12 3.45
1.22 0.67 1.96 2.05
Source: Constructed from raw data obtained from the IFS. Inflation is defined as the percentage change of the consumer price index.
terly data for all countries, to 1972, the year prior to the first oil shock. The second subperiod covers the years between the oil shocks of the 1970s; the third subperiod goes from 1979 to 1983, the last year when all of our four countries had fixed exchange rates. Two impbrtant facts emerge from Table 12.2. First, for the complete period (1957—83) there is a remarkable degree of similarity between the rates of inflation of the four Latin American countries and the United States. In fact, these data show that Guatemala had an average rate of inflation lower than the United States during these years. It is interesting to note, however, that during this period the rate of inflation was significantly more volatile in our four Latin countries than in the United States. For this complete period the standard deviation of the quarterly rate of inflation was, in these countries, between two and four times higher than the U.S. rate.15 Second, the data show some important differences in inflationary behavior across subperiods. During 1957-72 the average quarterly rates of inflation in all four of these countries were, in fact, below the U.S. rate. Moreover, in the cases of Guatemala and El Salvador inflation was one-third that of the United States! This situation is reversed during 1973— 78. In three of the countries - Honduras being the exception - the domestic rates of inflation exceeded the U.S. rate. Likewise, during 1979-83 in three of our countries domestic inflation exceeded U.S. inflation. These differences in the inflationary record across subperiods suggests that the international environment, and in particular terms-of-trade shocks, may have had something to do with the international transmission mechanism of inflation. In Section 2.3,1 investigate in some detail the way 15
These data are available from the author.
Exchange rates, inflation, and disinflation
309
Table 12.3. Augmented Dickey-Fuller and Ljung-Box tests of inflationary differentials Augmented Dickey-Fuller3
Ljung-Box Q(8)b
Dominican Republic 57.2-78.4 57.2-82.2
-3.3 -4.1
14.9 17.2
Guatemala 57.2-78.4 57.2-82.2
-3.6 -3.9
7.0 7.4
Honduras 57.2-78.4 57.2-82.2
-3.8 -4.4
16.8 19.7
El Salvador 57.2-78.4 57.2-82.2
-2.4 -2.6
25.1 35.6
Note: Time series Dinft = irj — Tif5 Computed with four lags. The critical value of the ADF at the 5 percent level is -2.9. * The critical values of the Q with 8 degrees of freedom at 10 and 5 percent levels are 13.4 and 15.5.
a
in which our four countries reacted to changing international conditions during the 1970s. According to the simplest monetary theories of the international transmission of inflation, under a fixed nominal exchange rate system the rate of domestic inflation in a small open economy does not differ from "the" world rate of inflation (Swoboda 1978). Under this setting, international price disturbances will be rapidly and fully transmitted into the domestic economy. Moreover, local central banks would have no room for engaging in monetary policy: A domestic excess supply (demand) for money will result in an equiproportional loss (gain) of international reserves, leaving the quantity of money and domestic prices unchanged in the economy. If this simple model of the international transmission of inflation holds, we would expect that the time series of inflationary differentials between the domestic country and "the" world will be characterized by white noise. Table 12.3 summarizes the results obtained from an analysis of the time series properties of inflationary differentials between each of our four
310
Sebastian Edwards
countries and the United States. Column one presents Augmented DickeyFuller (ADF) unit-root tests, and column two contains Ljung-Box statistics for white noise series. In every case inflation differentials were defined as follows: Dinft =ift-
(1)
where itt and irf5 are quarterly inflation rates for the country in question and the United States. Inflation, in turn, is defined as the percentage change in the consumer price index. As can be seen from Table 12.3, in the cases of Dominican Republic, Guatemala, and Honduras the hypothesis that the inflationary differential has a unit root is rejected at conventional levels. However, in the case of El Salvador the ADF statistic is slightly below the critical value. The Ljung-Box Q statistics reported in the second column show, on the other hand, that the hypothesis that inflationary differentials is white noise is rejected at the 1 percent level in Dominican Republic, Honduras, and El Salvador. The analysis of the time series properties of Dinf reported in Table 12.3 provides preliminary evidence suggesting that at least in the short run, variables other than "world" inflation affected domestic inflationary dynamics. A particularly important question refers to the ability of the central bank to undertake independent monetary policy.16 Recently, Stockman (1992) has analyzed this issue for a group of industrial countries during the Bretton Woods era. He estimated a series of vector autoregressions for inflationary differentials, and tested whether the coefficients of money growth (as a group) were significantly different to zero. He found that for the majority of the countries the null hypothesis of zero money growth coefficients was rejected, and interpreted these results as evidence in favor of the hypothesis that industrial countries had had some room to undertake independent monetary policy during the Bretton Woods era. In order to deal with the issue of monetary autonomy I estimated equations of the following type using quarterly data for the four Latin countries in the sample: Dinft = a + 2/3, Dinft-i + 25, Gmont_p
(2)
where, as before, Dinf refers to inflationary differentials; Gmont is the rate of growth of narrowly defined money for the country in question; i = 1, . . . , 3; andy = 0,. . . 3.17 If these countries enjoyed, and practiced, mone16
17
This discussion is somewhat related to traditional debates on sterilization and the "offset" coefficient. See Obstfeld (1982). The data were obtained from the IFS. The monetary data were "centered" in order to make them compatible with the inflation figures, which refer to period averages.
Exchange rates, inflation, and disinflation
311
Table 12.4. Likelihood ratio test on significance of coefficients of money and domestic credit growth Rate of growth of money
Rate of growth of domestic credit
Dominican Republic
8.6 (0.071)
10.5 (0.032)
Guatemala
13.4 (0.009)
16.5 (0.007)
Honduras
10.1 (0.038)
4.9 (0.299)
El Salvador
24.9 (0.000)
20.7 (0.000)
Note: Time periods are 1958.2 to 1983.4 for all countries except Honduras, where it is 1962.2 to 1983.4. Numbers in parentheses are significance levels of F-statistics for the null hypothesis that the growth coefficients are jointly zero.
tary independence, we would expect that the coefficients of Gmon would be different from zero as a group. In order to look at this issue from different angles, I also estimated a series of equations of the type of (2), where I replaced Gmon for the rate of growth of domestic credit (Gcredit).19 Table 12.4 presents the likelihood ratio tests obtained from testing the hypothesis that the coefficients of Gmon, or those of Gcredit depending on the equation, were jointly zero. As can be seen, in all cases but one - Honduras when domestic credit growth is used - the null hypothesis can be rejected at conventional levels. Equations of the type of (2) were also estimated for the four countries simultaneously using seemingly unrelated regressions techniques. In that case, the results obtained were similar, and the hypothesis of zero coefficients for the monetary policy variables as a group was rejected.19 Table 12.5 contains data on the sum of the estimated coefficients of money growth and domestic credit in equations of the type of (2) for the 18
19
Since Gmon may not be completely exogenous, using Gcredit helps deal with the endogeneity issue. The problem of potential exogeneity was also tackled by computing variance decompositions for a series of estimated VARs. In all cases, and under all possible orderings of the variables, the results provide strong indications that Gmon was, at least up to six quarters, exogenous. The rationale for estimating the system using seemingly unrelated regression stems from the fact that during the period under study these countries were subject to a series of common shocks, mostly originating from abroad.
312
Sebastian Edwards
Table 12.5. Sum of money growth and domestic growth coefficients in inflationary differential equations Sum of money growth coefficients Dominican Republic
Sum of domestic credit coefficients
0.160 (1.942)
0.170 (1.410)
Guatemala
0.358 (11.861)
0.116 (1.850)
Honduras
0.060 (3.364)
0.070 (0.708)
0.170 (11.851)
0.268 (11.146)
El Salvador
Note: Numbers in parentheses are F-statistics for the null hypothesis that the sum of the coefficients is equal to zero.
four countries under study. As can be seen in all cases but one (Dominican Republic), the null hypothesis that the sum of the rate of growth of money (Gmon) is equal to zero is rejected at conventional levels. The results, however, are slightly different when Gcredit is used. Only in the case of El Salvador is the sum of the coefficients significantly different from zero.20 The evidence presented here suggests that during more than twentyfive years with fixed exchange rates (1957-83), and contrary to the implications of the simple models of inflationary transmission, central banks were able to engage in some independent monetary policy in these four Latin American countries. However, an important question that these regressions cannot answer refers to whether, in spite of their technical ability to engage in independent monetary policy, these countries' economic authorities still chose to act prudently in response to the existence of a fixed exchange rate. In particular, this regression analysis does not address the issue of whether the existence of a fixed exchange rate acted as a constraint on fiscal and monetary policy. I try to deal with this issue in the next Section 2.3 where I investigate the conduct of monetary and fiscal policies and the circumstances surrounding the eventual abandonment of the fixed exchange rate regime in these four countries during the mid and late 1980s.
20
In the case of the three other countries, at least one of the quarterly coefficients of Gcredit was significantly positive, indicating that credit policy was effective at least in the (very) short run.
Exchange rates, inflation, and disinflation
313
15.0
0.0 1960
1965
1970
1975
1980
1985
— GUAEER HONEER
1990 SALEER DOMAER
Figure 12.1. Nominal exchange rates in selected Latin American countries.
2.3
Financial constraints, exchange rates, and crises
In the mid and late 1980s the long experience with fixed nominal exchange rates came to an end in all four countries. As Figure 12.1 shows, El Salvador was the first to abandon its fixed rate, officially devaluing its currency by 100 percent with respect to the dollar in 1986. Honduras was the last country to abandon the official parity in early 1990. The purpose of this section is to analyze in some detail the behavior of fiscal and monetary policy variables, and to investigate the circumstances that led to the abandonment of the fixed parities in the 1980s. In particular I try to shed some light on the question of whether these crises responded mostly to foreign shocks - including terms-of-trade disturbances and the debt crisis - or whether they were rooted in fiscal undiscipline. In analyzing the circumstances surrounding these four large devaluations, I compare the experiences of our four countries to that of Costa Rica. The importance of these comparisons resides in the fact that whereas the Dominican Republic, Guatemala, Honduras, and El Salvador maintained fixed rates during the 1970s and the early 1980s, Costa Rica reacted to both oil shocks with large devaluations (in 1974 and 1981). We
314 Sebastian Edwards can argue that while Costa Rica opted for abandoning the fixed exchange rate in order to achieve a smoother external sector adjustment - even at the cost of higher inflation - the other four nations decided to favor price stability at the cost of a more expensive adjustment process. The fundamental feature of a fixed exchange rate system is that it imposesfinancialdiscipline on the central bank and other government agencies. The authorities are aware that if they systematically violate certain macroeconomic constraints, the country will deplete its international reserves to a minimum level and will be forced to give up the parity. Consequently, a fixed exchange rate imposes upper bounds to domestic credit creation and fiscal imbalances. These constraints, however, operate in the medium to long runs, and do not have to be observed at every moment. It is, in fact, perfectly possible that in a particular period credit creation exceeds this "threshold." However, when this happens, the authorities will rapidly try to change directions, bringing macro policy back to a consistent course.21 If the country can increase its foreign indebtedness, the macro constraints can be side-stepped for a longer period of time. This, however, cannot be a permanent solution. As the Latin countries found out the hard way in the early 1980s, at some point the international community will stop providing funds, and a crisis will erupt. Fixed rates impose two basic financial constraints: First, domestic credit cannot grow, on average, at a rate faster than that of the demand for money. This, in turn, is determined by "world" inflation, real income growth, and the income elasticity of the demand for money.22 Second, the fiscal deficit (as a proportion of GDP)financedby money creation cannot exceed a certain bound determined by the increase in the demand for money and the degree of monetization of the economy.23 Figures 12.2 through 12.5 display estimated upper bounds or thresholds as well as actual rates of domestic credit creation and thefiscaldeficit for our four countries.24 In these figures the solid lines represent the 21
22
23
24
That is, under fixed exchange rate constraints it is assumed that macro authorities follow the "rules of the game." This assumes several things: First, the initial ratio of domestic credit to money is equal to the desired ratio. This is equivalent to saying that initially the central bank holds the desired ratio of reserves. Second, it assumes that the opportunity cost of holding money doesn't vary excessively, and can be ignored. Third, it assumes that the external debt-GDP ratio is in equilibrium. Formally this constraint is derived from the public finance approach to inflation. The deficit to GDP ratio (
Exchange rates, inflation, and disinflation
315
EL SALVADOR
DOMINICAN REPUBLIC 0.36
I960
1965
1970
1975
1980
D I O Q Domestic Credit —— Threshold Level
1985
1990
1960
1965
1970
1980
1985
1990
D log Domestic Credit i-r — Threshold Level
1970 Fiscal Deficit/GDP — —Threshold Level
1975
1975 Fiscal Deficit / G D P Threshold Level
Figure 12.2. Domestic credit creation and Figure 12.3. Domestic credit creation and fiscal deficit: Dominican Republic. fiscal deficit: El Salvador.
actually observed series, whereas the broken lines are the "upper bounds" imposed by the fixed nominal exchange rates regime. These figures, then, capture the periods during which the bounds were exceeded. During the early years (up to the late 1970s or early 1980s) the bounds seemed to have been very effective. Every time the actual rate of credit creation, or the deficit ratio, exceeded their respective thresholds, the authorities rapidly implemented corrective policies, bringing the policy variable below the for each country independently, using annual data for 1950-80. The following long-run elasticities were estimated: Dominican Republic, 1.0; El Salvador, 1.3; Guatemala, 1.1; Honduras, 1.8. Additionally I estimated a demand for money for Costa Rica. The estimated income elasticity is 1.5.
Sebastian Edwards
316
GUATEMALA
1970
- -
1960
1965
1970
HONDURAS
1975 D log Domestic Credit Threshold Level
1975
1980
- -
1985
1990
Fiscal Deficit /GDP - — Threshold Level
Figure 12.4. Domestic credit creation and fiscal deficit: Guatemala. fiscal
1960
1965
1970
D log Domestic Credit Threshold Level
1975
1980
1985
1990
Fiscal Deficit /GDP - - Threshold Level
Figure 12.5. Domestic credit creation and deficit: Honduras.
bound. These data strongly suggest that during these years our four countries were indeed following the "rules of the game," and acted as if the nominal exchange rate regime imposed financial constraints. The data for El Salvador are particularly interesting, showing that until 1979 the deficit ratio had never surpassed the threshold imposed by the fixed rates. In the late 1970s and early 1980s, however, this behavior breaks down; in every country the two policy indicators exceed the estimated thresholds without returning, in the short or medium runs, to a level consistent with fixed exchange rates. It is interesting to note, however, that all four countries were able to maintain their fixed rates for a number of years after the credit and deficit bounds had been violated. The official devaluation dates
Exchange rates, inflation, and disinflation
317
COSTA RICA
I960
1965
1970
- -
1975
1980
1985
1990
D log Domestic Credit Threshold Level
Fiscal Deficit/ GDP - — Threshold Level
Figure 12.6. Domestic credit creation and fiscal deficit: Costa Rica.
are: Dominican Republic, 1986; Guatemala, 1986; Honduras, 1990; and El Salvador, 1986. Figure 12.6 presents data on credit growth and deficit ratio for Costa Rica, a country that implemented major devaluations in 1974 and 1981. As can be seen, the Costa Rican case is very different from that of the other four countries. Since 1974, credit creation and the fiscal ratio exceeded in a systematic way the respective thresholds. This suggests that, even though after the 1974 devaluation a fixed rate was again adopted, this did not constrain government behavior. Figures 12.2 through 12.6 raise a fundamental question: How did our four countries manage to maintain afixedrate for a number of years after
318
Sebastian Edwards
having violated, in a systematic way, their macro constraints? The answer to this question is related to three fundamental factors: terms-of-trade shocks (mainly oil and coffee prices), the availability of international reserves, and foreign indebtedness. The Central American nations were hit hard by the first oil shock; their terms of trade declined between 14 and 20 percent during 1973-75. As a result they developed serious trade account and balance-of-payments problems, and experienced a decline in real income. In every country the crisis was faced with an anticyclical fiscal policy, where the authorities expanded public expenditure in an effort to reduce the real impact of the world recession. These actions, however, had further negative effects on the balance of payments and international reserves. The situation was particularly serious for Costa Rica, which entered the crisis with a very low level of reserves and a considerably appreciated real exchange rate. In late 1974, when the ratio of net foreign assets to monetary base dipped below the 10 percent mark, the Costa Rican colon was devalued by 28.8 percent. With hindsight, however, it may be argued that Costa Rica devalued too early. The problem is that the authorities could not foresee that the world coffee market would go into a frenzy, with prices increasing by 120 percent between 1975 and 1978. Suddenly, the coffee exporting Central American countries were experiencing a trade boom and were subject to the consequences of a Dutch (coffee) disease. International reserves were accumulated, and corrective fiscal policies were undertaken as a way to once again respect the fixed exchange rate constraints. This situation did not last for too long. The coffee market peaked in 1978, and by 1980 these countries were affected by the double shock of increasing oil prices and declining coffee prices. Their terms of trade worsened by 30 to 40 percent. The countries in the region reacted as in 1974 and implemented a countercyclical fiscal policy. The deficit increased and the rate of growth of domestic credit accelerated, as was documented in Figures 12.2 through 12.5. In Costa Rica reserves declined rapidly and the colon was again devalued in 1981. This time the four fixed rate countries departed from their historical behavior and, instead of introducing corrective policies, engaged in a process of rapid foreign borrowing. The increased indebtedness allowed them to temporarily ignore the financial constraints imposed by the fixed exchange rates. In this way the required adjustment following the terms-oftrade shock was somewhat eased in 1979-82. As can be seen in Table 12.6, in every country the foreign debt almost doubled between 1978 and 1982. In late 1982, when the Mexican debt crisis erupted, these countries' access to the international financial market was seriously curtailed and
319
Exchange rates, inflation, and disinflation
Table 12.6. Basic indicators in 1978-86 Dominican Republic
Guatemala
A. Terms of trade (1980 = 100) 121.9 1978 82.8 1982 77 71 1986 95 108 B. Fiscal deficit/GDP (%) 1978 1.5 1982 3.1 1986 NA
1.2 4.7 2.6
Honduras
El Salvador
107.9 79 90.7
127.8 89 89
0.8 6.6 6.5
1.6 6.8 -0.2
C. Total external debt (millions U.S.$) 1,136.2 1,334 1978 813 2,256.2 1,537 1982 2,519 3,141.1 1986 3,667 2,766
910 1,443 1,850
Costa Rica
115 94 106 5.0 0.9 4.5 1,679 3,641 4,575
D. Nominal exchange rate (units of local currency per U.S.$) 2.0 2.5 1978 1.0 1.0 2.0 1982 1.0 2.5 1.0 2.5 2.0 5.0 1986 3.1
8.57 40.25 58.88
Source: Constructed with raw data from CEPAL, World Bank, and IFS.
severe balance-of-payments crises ensued.25 An effort was made to return to the basic policies they had traditionally pursued, tightening credit creation and public expenditures. But the gap between actual and "admissible" policies had become too large. However, given the long attachment to fixed rates, the authorities of all four countries refused to devalue. In a way, the reputational constraint began to operate in a perverse way, delaying an unavoidable realignment. For a number of years the authorities were unwilling to accept the fact that the historical rates were unsustainable, and faced the crises with half-baked measures including the adoption of multiple rates, import licensing, and other trade impediments. The imposition of these distortions imposed significant costs in these economies, delaying the recovery. In fact, Costa Rica - which did devalue early on experienced a much faster recovery of output than our four historical fixers. Whereas two years after the crisis Costa Rica's growth rate returned 25
However, because of the Central American armed conflict, all of these countries were able to receive substantial U.S. assistance. This allowed them to survive a bit longer, but did not help avoid the crises.
320
Sebastian Edwards
to the historical level, this has not happened yet in the other Central American nations. Finally, one by one these countries recognized the unavoidable, and devalued their currencies. Honduras was the last one to accept its fate. In 1990, after the Sandinistas lost the Nicaraguan election, and Honduras's strategic value for the United States suddenly declined, President Callejas realized that he couldn't obtain massive aid and decided to devalue the lempira by 125 percent. More than seven decades of fixed rates and national pride were history. 3
Exchange rate policy, inflationary inertia, and anchors
As was shown in the preceding section, in the long run it is not possible to maintain a fixed nominal exchange rate under conditions of major fiscal imbalances financed by domestic credit creation. If domestic inflationary pressures exceed "the" international rate of inflation, international reserves will decline, overvaluation will take over, and a speculative attack on the central bank foreign exchange holdings will eventually take place.26 Some countries that have suffered from high inflation have dealt with this situation by adopting a passive, or crawling peg, exchange rate regime, where the nominal exchange rate is periodically adjusted according to inflation rate differentials. The motivation behind this type of system is the recognition that, given the country's fiscal stance, it is not possible to maintain a competitive real exchange rate with a fixed nominal exchange rate. In this scenario the economy typically learns how to live with relative high inflation, and generalized indexation becomes institutionalized (Williamson 1991). In the 1960s and 1970s there was considerable enthusiasm regarding the potential role of indexation as a way to isolate the real sector from the effects of inflation. In the last few years, however, it has become clear that the benefits of indexation had been greatly overestimated. The experiences of a large number of countries - especially in Latin America - showed that a generalized indexed system introduces serious rigidities into the economy, precluding required relative price adjustments. Additionally, it has been found that once indexation becomes ingrained, inflationary forces exhibit a remarkable degree of inertia, almost acquiring a life of their own (Edwards 1992b). In the late 1970s and 1980s a number of countries have tried to break these inertial inflationary forces by implementing stabilization programs 26
There is now an extensive literature on speculative attacks under fixed exchange rates. The pioneering piece is Krugman (1979). See Edwards (1989a) for a number of case studies on the subject.
Exchange rates, inflation, and disinflation
321
based on replacing their crawling peg regimes with a fixed exchange rate or nominal exchange rate anchor. The rationale for this policy follows the arguments presented in Section 2.1: In an open economy the nominal exchange rate will provide an efficient and credible anchor for prices, helping reduce inflation more rapidly and with lower real costs than if alternative anchors - such as the nominal quantity of money - are used.27 Naturally, in order for these programs to succeed in moving the economy to lower inflation, the fixing of the nominal exchange rate has to be accompanied by restrictive credit and fiscal policies. Recent experiences with fixed exchange rate-based stabilization programs have had mixed results and have been somewhat controversial. For instance, the Southern Cone (Argentina, Chile, and Uruguay) episodes in the late 1970s have attracted considerable attention and have been largely considered a failure.28 In these countries the adoption of a fixed exchange rate led to significant overvaluation, losses in competitiveness, and eventual crises that included the abandonment of the fixed rate. The fact that these three countries experienced very different fiscal histories during this period adds considerable interest to these episodes. Whereas in Argentina, and to some extent in Uruguay, the adoption of the nominal anchor was not accompanied by corrective fiscal policies, in Chile the exchange rate was fixed with respect to the dollar at a point where an almost perennial fiscal deficit had been transformed into a surplus.29 The Argentinian and Brazilian heterodox programs of the 1980s - the austral and cruzado plans - also relied on nominal exchange rate anchors. However, the inability (or unwillingness) to control other macroeconomic variables, including the fiscal deficit, suggested early on that these attempts would end up in failure and frustration. The experiences of Israel and Mexico, however, have indicated that under some circumstances it is indeed possible to use a nominal exchange rate anchor to advantage, within a broad and consistent stabilization package.30 More recently, the reliance on fixed nominal exchange rates in eastern Europe has elicited considerable interest. In this section I analyze the mechanics of exchange rate indexation and inflationary inertia, and I discuss under what circumstances the adoption of nominal exchange rate anchors are expected to be successful in a stabili27
28
29 30
For a discussion on the use of alternative anchors in stabilization programs see, e.g., Bruno (1991). See, also, Kiguel and Liviatan (1992) and Calvo and Vegh (1990). See Corbo and de Melo (1986). However, Chile's subsequent recovery and solid economic performance has been considered by most authors, a remarkable achievement. Edwards and Cox-Edwards (1991). On Argentina, Brazil, Israel, and Mexico, see Bruno, Fischer, and Helpman (1991). Dornbusch and Simonsen (1983) deal with general issues of indexation.
322
Sebastian Edwards
zation program. In doing this, I deal with the issue of credibility and I stress the need to implement a broad deindexation program that goes beyond the exchange rate. I also provide an empirical analysis of three recent historical experiences with nominal exchange rate anchors. 3.1
Indexation, inertia, and anchors
Consider the case of an economy that produces two types of goods: tradables and nontradables. Tradables prices are assumed to be linked to international prices, whereas nontradables prices are determined by the condition that this market clears at all times. In order to focus on inflationary issues I abstract from problems related to changes in real exchange rate fundamentals, such as terms of trade, the degree of protection, and capital flows.311 assume that initially the country follows a crawling peg exchange rate system, where the exchange rate rule consists of adjusting the nominal exchange rate by a proportion c|> (4> < 1) of lagged inflation differentials. I also assume that wages are adjusted according to a rule that includes lagged inflation as well as expected future inflation. Monetary policy is assumed to be passive and to accommodate inertial inflationary forces. This stylized economy, which captures, for example, some of the most salient aspects of many Latin American countries in the 1980s, can be depicted by the simple set of equations: 7T, = onrTt + (1 - a)TrNt 7TTt = £,_,(
S
N (PJPV Z,) = N (W/PN)
X
(3) (4) (5) (6)
(7)
where the following notation has been used: TTt = rate of change of the domestic price level; TTTt = rate of change of the price of tradables in domestic currency in period t; TTNt = rate of change of nontradable prices in period t; dt = rate of devaluation in period t; 31
On these issues, see Edwards (1989a). The discussion that follows draws partially on Edwards (1992b).
Exchange rates, inflation, and disinflation
323
Tr*Tt = rate of world inflation in period t; £*,_, = expectations operator, where expectations are assumed to be formed in period t — 1; w = rate of change in nominal wages; Z, = index of aggregate macroeconomic policies, which includes monetary expansion beyond passive accommodation of past inflation; ND, Ns = demand and supply for nontradables; ift = expected inflation in period t; , yk = parameters that measure the degree of indexation in this economy; a = parameter that determines the importance of expected inflation in the wage rule. Equation (3) says that the domestic rate of inflation is a weighted average of tradables and nontradables inflation. Equation (4) states that the law of one price holds ex ante, and that the change in the domestic price of tradables is equal to the expected change in the exchange rate plus the expected rate of world inflation.32 Equation (5) is the devaluation rule, and states that the exchange rate is adjusted in a proportion <j> of inflation rate differentials. If = 1 we have a typical purchasing-power parity rule, where the rate of devaluation is equal to (lagged) inflation rate differentials. This type of policy has sometimes been referred to as a "real target" approach.33 Equation (6) is the market clearing condition for nontradables. The demand for nontradables is assumed to depend on relative prices {PJPj) and on aggregate demand (Z,); the supply of nontradables is assumed to be a function of real product wages. Finally, equation (7) is the wage adjustment rule, and states that wage increases depend on two factors: lagged inflation up to K periods and expected future inflation. It is further assumed that (2yk + a) < 1. The special case when a = 0 and Xyk = 1 corresponds to a situation where there is 100 percent backward looking wage indexation. In equation (7) the value of K will determine the degree of inflationary "memory" of this economy. Although in equation (7) w stands for the rate of change of nominal wages, it is perhaps more useful to think this variable captures a 32
33
The presence of the expectations operator reflects the assumption that the domestic price of tradables is set before the rate of devaluation or world inflation is observed. This, however, is a somewhat misleading name since it is used to denote two quite different policies. Whereas some authors refer to a strict PPP rule as a "real targets" policy, others define this "real target" as a policy aimed at accommodating changes in RER fundamentals. See Corden (1991).
324
Sebastian Edwards
broader category of "other" costs. In that sense, then, the coefficients yk can be interpreted as summarizing the degree of indexation of nonexchange rate contracts in the economy.34 This model can be solved in order to find an expression for the dynamics of inflation. In order to simplify the discussion, I first consider the case when wages are adjusted according to inflation in the last period only (K = 1). Further, assume that the wage adjustment rule (7) is a strict weighted average of past inflation and expected inflation. That is, in equation (7) a = 1 - y. Before solving the model, one must make an assumption regarding inflationary expectations. I consider the case of rational expectations where actual realizations of inflation in period t differ from the expectations formed at the beginning of that period by a random term |x: ?r, = 7ft + iir
After manipulating equations (3) through (7) the dynamics of domestic inflation can be written as the following first order difference equation:35
where: 1
_ (" + ae)
a2 = <" + " f - * \
(9)
(10)
(" + sa) + s{\ — a)y (" 4- ae) + s{\ -
(ID
and where " is the demand elasticity of nontradables with respect to relative prices (" < 0), e is the supply elasticity of nontradables with respect to the real product wage (e < 0), 8 is the demand elasticity of nontradables with respect to aggregate demand pressures, and /x' is an error term related tO jLL. In equation (8) coefficient ax provides a measure of the degree of inertia of domestic inflation. The closer is ax to unity the more persistent will inflation be, and the higher the degree of inertia. As can be seen from the definition of ax in equation (8), the degree of inertia in the economy will 34
Notice that in this model no explicit expression has been included for the rate of growth of domestic credit. This responds to the assumption that the monetary authorities follow a passive credit policy that accommodates the inertial inflation.
35
In obtaining (8) I also assumed that Et_}(TT*t) = TT*_V
Exchange rates, inflation, and disinflation
325
depend on the different elasticities involved and, what is particularly important for this discussion, on the indexation parameters and y 36 From equation (8) a number of important features of the dynamics of inflation emerge. First, if there is full lagged indexation of the exchange rate - that is, is smaller than one, the autoregressive term ax will also be smaller than one, and inflation will be characterized by a stationary process. In this case domestic inflation will converge to the world rate of inflation.37 The speed at which this convergence process takes place will depend on the degree of backward wage rate indexation. Third, and related to the previous point, a reduction in the rate of exchange rate indexation will result in a decline in the value of ax and, thus, in a reduction in the degree of inflationary inertia in the economy. This, of course, has been the rationale for adopting nominal exchange rate anchor policies in a number of countries. A fourth interesting feature of equation (8) is that if indexation is totally eliminated, and both and y become simultaneously equal to zero, domestic inflation will immediately converge to world inflation.38 This situation corresponds to a Poincare-type stabilization program.39 Notice, however, that if after the nominal exchange rate is fixed ( = 0) some degree of wage indexation remains (y > 0), there will still be some inertia and the real exchange rate will be subject to consistent appreciation during the transition process. In a sense, then, the authorities face a trade-off where, on the one hand, the exchange rate anchor will reduce inertia, and, on the other hand, it will generate a loss of international competitiveness. Whether the net benefits of this package will be positive will depend on a number of factors, including the initial level of the real exchange rate - an initial condition of undervaluation being preferred - and the extent to which the degree of inertia (coefficient ax in our representation) is actually reduced. Under some conditions, however, it is possible to face situations where, due to the lack of credibility in the anchor policy, the country ends 36
37 38 39
Recently a number of authors have discussed alternative ways of measuring the degree of persistence in the time series of GNP (Cochrane 1988). Much of this recent discussion has centered on measuring persistence in nonstationary series. If one assumes that in the steady state z = 0. This again assumes that aggregate demand measures have been eliminated (z, = 0). See Sargent (1983).
326
Sebastian Edwards
up in the worst of worlds, with no significant reduction in the degree of inflationary inertia and a substantial loss in competitiveness. It is easy to extend the analysis to the more general case when the system has a longer inflationary memory, and wage rate adjustments depend on lagged inflation in more than one period. In this case the dynamics of inflation will be represented by a ^ h order difference equation, and it will still be true that full backward indexation of the rate of devaluation will result in domestic inflation having a unit root. An interesting special case emerges when the wage adjustment rule depends exclusively on past inflation. If only one period lagged inflation is taken into account for wage increases, equation (7) becomes:
In this case the dynamics of inflation are given by: (" + ey) + as{(j> 7T,
y)
7-^—
77,,, T/
'
l
(12)
" + e
'
If y and are equal to one (full backward indexation), domestic inflation will have a unit root. On the other hand, if either <j> or y are smaller than one, the coefficient of irt_x in (12) will be smaller than one, and inflation will be characterized by a stationary process. Although the discussion presented here has been carried out in terms of ways to reduce indexation and inertia, the model is perfectly symmetric for the case when a fixed rate system is replaced by a crawling peg. Under these circumstances the economy will pass from a regime where = 0 to one where > 0, and the degree of inflationary inertia will increase. Whether this process will end up in a complete loss of the nominal anchor will depend on a number of factors, including the degree of (implicit or explicit) indexation of the rest of the economy. An important implicit assumption in the nominal exchange rate anchors approach to disinflation is that the adoption of a fixed nominal exchange rate is a credible policy and that the public believes that, from the date of the new policy announcement, the coefficient will remain lower.40 In fact, one of the most commonly used arguments for favoring nominal exchange rate anchors over monetary anchors has to do with credibility. It has been argued that since nominal exchange rates are more visible, 40
See Agenor and Taylor (1992) for a survey on alternative ways to empirically test for credibility effects.
Exchange rates, inflation, and disinflation
327
they provide a more credible policy than if a constant level of monetary base is announced.41 In terms of the model presented, it follows from equation (8) that if the nominal exchange rate anchors policy is credible, we would empirically observe a structural break in the dynamic properties of inflation. This structural break would indeed take place at - or around - the moment the nominal anchor is implemented. From that point in time onward, the coefficient of lagged inflation in an equation of the type of (8) should decline, reflecting the reduction in the degree of persistence in the inflationary process. Of course, this assumes that the structural roots of inflation - fiscal imbalance and monetary creation - have been controlled by the economic authorities. If, however, the nominal anchor policy lacks credibility, and the public has doubts regarding the extent to which the government will stick to the new policy, the estimated degree of inertia in equations of the type of (8) will not be significantly affected by the adoption of the nominal exchange rate anchor.42 Empirically, there are a number of possible ways to investigate whether the adoption of exchange ratebased stabilization programs have changed inertia. Two approaches used above are: (1) the use of interactive dummy variables to test for structural breaks in coefficient ax in equation (8) at the time (or around the time) of the policy change; and (2) the estimation of equation (8) using varying coefficient techniques.43 3.2
Exchange rates and inflationary inertia
In this section I use data on Chile, Mexico, and Venezuela to investigate the relationship between exchange rate policy and inflationary inertia. These three countries provide interesting and contrasting lessons. Chile and Mexico used the nominal exchange rate as an anchor in their disinflation programs of the 1970s and 1980s. Venezuela, on the other hand, ended a long experience with fixed exchange rates in 1989, when it devalued its currency and adopted a managed exchange rate system. Chile In the late 1970s, and after having eliminated a stubborn fiscal deficit, Chile adopted an exchange rate-based stabilization program. Initially from February 1978 to June 1979 - the program consisted of a preannounced declining rate of devaluation of the domestic currency. This 41 42
43
See Bruno (1991) for related discussions. Edwards and Sturzenegger (1992) provide a model with endogenous credibility of the nominal exchange rate anchor. See Agenor and Taylor (1992).
328
Sebastian Edwards
system, popularly known as the tablita, deliberately set the starting declining rate of devaluation at a lower rate than ongoing inflation. With a trade liberalization reform having virtually eliminated most import barriers, it was expected that this system of preannounced devaluations would have two important effects on inflation. First, it would introduce price discipline through external competition and, second, it would reduce inflationary expectations.44 In terms of the model presented previously, this policy amounted to reducing parameter c() in equation (8). In June 1979, with inflation standing at an annual rate of 34 percent, the government put an end to the system of a preannounced declining rate of devaluation and fixed the nominal exchange rate at thirty-nine pesos per dollar. It was expected that this move to a fixed rate would reinforce and accelerate the convergence of domestic to world inflation.45 When the tablita was adopted in early 1978, and again when the peso was pegged to the dollar in June 1979, it was decided not to alter the wage indexation mechanism, which at that time was characterized by 100 percent adjustment to lagged price increases. Paradoxically, although the authorities expected price setters and other agents to form forwardlooking expectations, they maintained a crucial market linked to a rigidly backward indexation regime. Contrary to what was expected by the architects of the Chilean exchange rate-based stabilization plan, after the exchange rate was fixed in mid-1979, the domestic rate of inflation did not rapidly converge to its world counterpart. In fact, the use of the exchange rate as a stabilization tool helped generate a steady real appreciation of the peso, which, among other things, negatively affected the degree of competitiveness of firms producing goods in the tradable sector, including nontraditional exports.46 In 1982, under considerable pressure and increasing capital flight, the fixed exchange rate was abandoned, as Chile's experiment with a nominal exchange rate anchor came to an end. After suffering a remarkably deep recession, by 1985 the Chilean economy began to recover. A new administered (crawling peg) exchange rate system was put in place, and the backward-looking wage indexation was 44 45 46
On the Chilean experience, see Edwards and Cox-Edwards (1991). T h e s e measures consisted o f setting <J> equal t o zero. Besides the adoption of a fixed exchange rate regime, another important development took place during 1979. Steps toward the liberalization of capital flows were taken, when in June of that year commercial banks were allowed to greatly increase their ratio of foreign liabilities to equity. This relaxation of capital inflows results in massive borrowing from abroad and paved the way to Chile's debt crisis. I have argued in Edwards (1985) that the massive inflow of foreign capital was one of the fundamental causes of real exchange rate overvaluation in Chile.
Exchange rates, inflation, and disinflation
329
Table 12.7. Exchange rates, indexation, and inflationary inertia in Chile (Eq. 13.1)
(Eq. 13.2)
Time period
74.1-82.1
74.1-82.1
Constant
-0.041 (-1.344)
-0.049 (-1.397)
0.750 (12.993)
0.756 (10.317)
DC\*-nt_t DC2**_t
0.019 (0.688) -0.025 (0.430) 0.236 (1.415)
0.477 (1.378)
0.288 (4.812)
0.285 (4.737)
R2
0.970
0.969
DW
2.042
2.036
Z,-i
Note: /-statistics in parentheses.
replaced by a wage-setting mechanism based on bilateral bargaining between unions and firms. An important characteristic of the new exchange rate system was that, instead of following a rigid "real targets" approach, the authorities considered the evolution of real exchange rate "fundamentals" in determining the rate of nominal devaluation in any given period. In order to investigate empirically the way in which the adoption of a nominal exchange rate anchor affected the degree of inflationary inertia in Chile, I estimated equations of the following type using quarterly data:
where the variable D is a dummy that takes the value of one for the period when the nominal exchange anchor is in place and zero otherwise. If the anchors program is effective and credible, the estimated coefficient of b2 should be significantly negative, indicating that this policy successfully reduced the degree of inertia in the system. Moreover, in the extreme case of a Poincare style disinflation, inertia should disappear at the time the new policy is put in place, and (bx + b2) should not be significantly different from zero. In the case of Chile, two dummies were used: The first one (DCl) has
330
Sebastian Edwards
a value of one between the second quarter of 1978, when the preannouncement of the declining rate of devaluation was first adopted and the first quarter of 1982, the last quarter when the nominal anchors approach was in effect. The second dummy (DCl) takes a value of one between the third quarter of 1979 and the first quarter of 1982. That is, D2 covers the period when the nominal exchange rate was strictly fixed. In the estimation of equation (13) 7r* was defined as the quarterly rate of U.S. inflation and z,_, as the rate of growth of domestic credit. The raw data were taken from the IFS tape. Table 12.7 contains the results obtained using OLS. The results obtained are quite interesting. In equations (13.1) and (13.2), the interactive dummies are not significant and, in addition, have a positive sign. This indicates that the implementation of an exchange rate rule in 1978-79 did not alter the degree of inflationary inertia in Chile. These results could be the consequence of a combination of factors, including the fact that the nominal anchor was not credible and that wage rate indexation was left intact during this period. In order to further test whether the adoption of the fixed exchange rate in June 1979 had an effect on the inflation process, a number of tests on the structural stability of the inflation equations were computed. If, indeed the shift from an accommodating adjustable exchange rate regime to a rigidly predetermined one is credible, it would be expected that the inflation equation would capture a change in regime. These stability tests were supportive of the dummy variable results reported previously, and showed no structural break in the inflation equation. For example, in the case of equation (13), the chi-square statistic for structural stability had a value °f X2(6) — 2.03, indicating that there is no evidence of a change in the inflationary regime in mid-1979. These results strongly suggest that the adoption of a predetermined exchange rate in Chile in 1978-79 was not associated with a change in the nature of the inflationary process that one expects from a credible nominal exchange rate anchor policy. In particular, expectations, backward wage rate adjustment practices, and other contract practices (e.g., indexation) do not seem to have been affected in a significant way by the reform in the exchange rate system. As a consequence, the degree of inflationary inertia remained unchanged after the adoption of the exchange rate-based program of the late 1970s. Mexico In 1986-87 the Mexican government embarked on an ambitious stabilization and reform program aimed at regaining price stability, deregulating the economy, and opening foreign trade to international competition. As in the case of Chile in the late 1970s, the manipulation of the nominal
Exchange rates, inflation, and disinflation
331
exchange rate became an important component of the stabilization plan. During 1988 the exchange rate was fixed relative to the United States, and from 1989 onward the rate of devaluation of the peso was preannounced. As in the case of Chile, and in an effort to guide expectations downward, the rate of devaluation was deliberately set below the rate of ongoing inflation. In successive revisions of the program, the preannounced rate of adjustment of the nominal rate was reduced downward, to the point that there was consensus that the policy would eventually end in the adoption of a fixed nominal exchange rate with respect to the U.S. dollar. These successive reductions in the preannounced rate of change of the nominal exchange rate are equivalent to reductions in the value of coefficient in equation (5). The Mexican program differed from the Chilean plan in three important respects. First, in order to shake expectations, the peso was temporarily fixed for one year at the beginning of the program (1988). Once expectations were reduced, a preannounced sliding parity system was adopted. Second, Mexico had a considerably longer transition with a positive (although declining) predetermined and preannounced rate of devaluation. Third, at the outset of the program the real exchange rate was considerably undervalued. Thus, the system had a "built-in cushion" that was able to absorb the process of real exchange rate appreciation that accompanied the "sliding peg." And fourth, whereas in Chile exchange rate deindexation was the only component of the package, in Mexico incomes policies became a central element of the anti-inflationary package, supplementing the exchange rate rule and the fiscal adjustment. Indeed, in late 1987 with the establishment of the Pacto de Solidaridad, unions, entrepreneurs, and the government worked out a political and economic plan for defeating inflation: Price and wage guidelines became important elements of this program.47 In the estimation for Mexico the dummy variable DM was defined with a value of one between the second quarter of 1988, when the Pacto de Solidaridad was enacted, and the second quarter of 1990. The rate of growth of narrowly defined money was used as a measure of aggregate demand pressures. In every equation estimated for Mexico the coefficient of (DMTrt_x) was significantly negative, indicating that the adoption of the preannounced exchange rate system and the other policies in the Pacto were credible, 47
It is crucially important to point out, however, that the adoption of incomes policies in Mexico took place two years after the fiscal accounts had been balanced. See Beristain and Trigueros (1990) for a useful description. This is, in fact, a very important difference between the Mexican program and the heterodox stabilization programs of Argentina, Brazil, and Peru.
332
Sebastian Edwards
significantly changing the dynamics of inflation. The following result was obtained for 1979.1 through 1990.2: 77, = -0.060 + 0.896 ir,.! - 0.194 {DM ir,.,) + 0.698 it*_x (-3.546) (15.119) (-4.559) (1.726)
(14)
+ 0.179 Z, + 0.220 Z,_, + 0.144 Z,_2 R2 = 0.945 (2.789) (2.476) (1.764) Z W = 1.828 Period: 79Q1-90Q2 Formal tests for the stability of the regression as a whole show that the dynamics of inflation in Mexico experienced a structural break in the first quarter of 1988, when the exchange rate-based program and the Pacto de Solidaridad were enacted. The #2(6) statistic for structural stability turned out to be equal to 44.2, rejecting the null hypothesis that there was no structural break in the first quarter of 1988. The contrasting results between Chile and Mexico strongly suggest that it is not enough to adopt a nominal exchange rate anchor to alter the inertial nature of the inflationary process. As the Chilean results show, it is possible to have such a system in place for a considerable period of time without inflicting a serious dent in the dynamics of inflation. To the extent that the public does not perceive the new policy as credible, pricing behavior and contract clauses will not be altered in any significant way, and the ingrained aspects of inertial inflation will continue.48 Although it is not possible to extract from these data the exact underlying macroeconomic reasons for the differences in effectiveness of these two programs, it is possible to speculate that the incomes policies implemented in Mexico alongside the pegged nominal exchange rate provided a broad sense of coherence to the stabilization program. On the contrary, the continuation - and even reinforcement - of the lagged wage indexation rule gave contradictory signals to the private sector in Chile. Venezuela The Venezuelan experience is significantly different from that of Chile and Mexico. Whereas these two countries decided to adopt a predetermined exchange rate system as a way to reduce inflation, Venezuela was forced to abandon a fixed exchange rate regime in the early 1980s. After having maintained a fixed parity with respect to the U.S. dollar since 1964, Vene48
One way to rationalize this is to think that the public interprets the adoption of the fixed rate as a weak commitment, which can be abandoned under certain contingencies. If the private sector perceives that these contingencies are very permissive, the degree of commitment associated with the change in exchange rate regime will be very low, or nonexistent.
Exchange rates, inflation, and disinflation
333
49
zuela devalued its currency in 1983. This devaluation was the result of a combination of factors, including the debt crisis and serious macroeconomic mismanagement. The new exchange rate regime adopted in 1983 was characterized by a dirty float where, for all practical purposes, the central bank sets the daily nominal exchange rate. In reality, this new exchange rate system did not differ in any way from a crawling peg regime. The central bank adjusted the nominal exchange rate according to a number of factors including (or especially) inflationary differentials.50 Since the early 1980s, and especially after the abandonment of the fixed exchange rate, inflation in Venezuela has remained at a stubbornly high level - in the order of 30 to 35 percent per year - in spite of repeated attempts to bring it down through contractionary aggregate demand policies. Some observers have argued that the existing managed exchange rate system is part of the problem, since it has introduced a significant degree of inertia in the inflationary system. From a policy perspective, an important implication of this view is that the adoption of a fixed (or more rigid) exchange rate regime would help reduce inertia. In this section I present some regression results obtained from the estimation of equations of the type of (8) on inflationary inertia for Venezuela. Quarterly data were used for the period 1970 through 1990. The main purpose of this analysis is to investigate whether the adoption of a more flexible exchange rate regime in the 1980s has affected the degree of inertia in the Venezuelan economy. In the initial analysis a dummy variable for regime change (DV) was defined as taking a value of one starting in the first quarter of 1983, until the end of the period (1990.3). Table 12.8 contains the estimates from our basic regression, where, as before, TT is inflation, IT* is U.S. inflation, and GROCC is the rate of growth of domestic credit. The results are quite interesting, suggesting that the country virtually had no inflationary inertia until 1983 - the coefficient of 7r,_, is insignificantly different from zero - a point at which the inertial properties of inflation made a forceful appearance into the country. In fact, these results suggest that after 1983 the Venezuelan inflation process has flirted with loosing its anchor - the coefficient of lagged inflation after 1983 being extremely high (0.83). As a way to inquire further into the nature of the inflationary process the inflation dynamic equation was estimated using a recursive coefficients technique. Naturally, in this estimation the dummy term was excluded. 49 50
At that time a highly inefficient multiple exchange rates system was adopted. In 1989 the new administration of President Carlos Andres Perez unified the exchange rate. The structural characteristics of the Venezuelan economy give the central bank remarkable power to control the foreign exchange market. The state-owned oil company PDVSA is forced to sell all of its foreign exchange to the central bank.
334
Sebastian Edwards Table 12.8. Exchange rates and inflationary inertia in Venezuela (quarterly data 1970-90) (Eq. 13.3) Constant
-0.004 (-0.067)
ir,_,
-0.190 (-0.597)
DV**,^
0.828 (2.702)
<
1.410 (2.437)
Z,
0.240 (0.398)
R2
0.436
DW
2.089
Note: f-statistics in parentheses.
The estimated coefficients of lagged inflation, our measure of inertia according to the model developed here, are depicted in Figure 12.7. As can be seen, until 1979 this coefficient was remarkably low, strongly suggesting that during this period the Venezuelan inflation process lacked any considerable inertial component. In mid-1979 there appears to be a clear structural break with some incipient inertial components showing up. However, between 1979 and 1986 the degree of persistence of inflation was, for all practical purposes, still rather low, with the coefficient of lagged inflation fluctuating around 0.27 and 0.35. The diagram shows, however, that around 1987 a new structural break took place, with inflationary inertia exhibiting an important increase: After having gone through an important jump, the coefficient of lagged inflation seems to have stabilized itself at the 0.8 level. Although this number is still far away from a complete loss of anchor - a situation that occurs when the coefficient of lagged inflation becomes unity - it is significantly higher than the traditional Venezuelan situation, where inflation did not show any significant inflationary component. Although mandatory backward-looking wage indexation has not become a part of the Venezuelan macroeconomic structure, the existing evidence suggests that the indexation of all sorts of contracts has become increasingly important. As these practices become more and more popu-
Exchange rates, inflation, and disinflation 1.25
335
T
0.25 0.00 -0.25 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 .INERTIA] Figure 12.7. Coefficient of lagged inflation.
lar, the inertial component of the inflationary process will gain in importance. In thinking about possible policy scenarios for the future it is important to keep in mind that, as the discussion on the Chilean and Mexican cases presented suggests, stabilization programs based exclusively on the adoption of a fixed exchange rate may fail to reduce inertia and can generate a significant degree of real exchange rate overvaluation. 4
Concluding remarks
The purpose of this essay has been to investigate some aspects of Latin America's experience with fixed exchange rates. The analysis has dealt with both long-term and short-run characteristics of this system in Latin America. First, I have dealt with longer-term issues, evaluating the longterm experiences of four Latin American nations with fixed exchange rates. The evidence analyzed suggests that the existence of a long tradition of exchange rate stability provided some constraints on central bank behavior in these countries. This evidence, however, also indicates that these constraints were limited, and would not survive the combination of populist political pressures and severely negative terms of trade shocks in the late 1970s and early 1980s. In fact, our analysis suggests that by trying to avoid exchange rate adjustment after the debt crisis, these nations incurred severe costs.
336
Sebastian Edwards
The second issue addressed in this essay refers to the use of fixed (or predetermined) nominal exchange rates during a disinflationary process. I develop a simple model of a two-shock economy to analyze the way in which a crawling peg regime increases the degree of inflationary inertia in these countries. The empirical analysis presented in Section 3 indicates that, even when the fiscal side is balanced, the adoption of a fixed nominal exchange rate may not reduce the degree of persistence or inertia in the system. The contrasting experiences of Chile and Mexico suggest that the combination of a fixed exchange rate anchor with income policies (as in Mexico) may be a particularly effective way of reducing inertia.51 References Agenor, P. R., and M. P. Taylor. (1992). "Testing For Credibility Effects" IMF Staff Papers 39, no. 3 (September): 545-71. Aghevli, B., M. Khan, and P. S. Montiel. (1991). "Exchange Rate Policy in Developing Countries: Some Analytical Issues." IMF Occasional Paper no. 78. Washington, D.C.: International Monetary Fund. Aghevli, B., and P. S. Montiel. (1991). "Exchange Rate Policies in Developing Countries." In E. M. Claassen (ed.), Exchange Rate Policies in Developing and Post-Socialist Countries. San Francisco: ICG, 205-42. Beristain, 1, and I. Trigueros. (1990). "Mexico." In J. Williamson (ed.), Latin American Adjustment. Washington, D.C.: HE. Bruno, M. (1991). "High Inflation and the Nominal Anchors of an Open Economy." Princeton Essays on International Finance. Princeton University. Bruno, M., S. Fischer, and C. Helpman (eds.). (1991). Lessons of Economic Stabilization and Its Aftermath. Cambridge, Mass.: MIT Press. Burton, D., and M. Gilman. (1991). "Exchange Rate Policy and the IMF." Finance and Development 28 (September): 12-21. Calvo, G. (1978). "On the Time Consistency of Optimal Policy in a Monetary Economy." Econometrica 46, no. 6 (November): 1411-25. Calvo, G , and C. Vegh. (1990). "Credibility and the Dynamics of Stabilization Policy: A Basic Framework." Washington, D.C.: IMF. Cochrane, J. (1988). "How Big Is the Random Walk in GNP?" Journal of Political Economy 96, no. 5 (October): 893-920. Corbo, V., and J. de Melo. (1986). "What Went Wrong in the Southern Cone?" Economic Development and Cultural Change 43, no. 3: 23-38. Corden, M. (1991). "Exchange Rate Policy in Developing Countries." In J. de Melo and A. Sapir (eds.), Trade Theory and Economic Reform. Oxford: Basil Blackwell, 93-104. Coricelli, F , and R. Rocha. (1990). "Stabilization Programs in Eastern Europe: Poland and Yugoslavia." Washington, D.C.: World Bank. Cukierman, A., S. Edwards, and G. Tabellini. (1992). "Seignorage and Political Instability." American Economic Review 52 (June): 224-61. 51
Notice, however, that this does not mean that exchange rate appreciation will be avoided. Mexico's real exchange rate has, in fact, suffered a significant degree of appreciation. This, however, has been the result of a combination of forces, including the recent massive inflows of foreign funds.
Exchange rates, inflation, and disinflation
337
Devarajan, S., and D. Rodrik. (1992). "Do the Benefits of Fixed Exchange Rates Outweigh Their Costs? The Franc Zone in Africa." In I. Goldin and L. A. Winters (eds.), Open Economy Structural Adjustment and Agriculture. Cambridge: Cambridge University Press, 31-53. Dornbusch, R., and M. H. Simonsen. (1983). Inflation, Debt and Indexation. Cambridge, Mass.: MIT Press. Edwards, S. (1985). "Stabilization with Liberalization: The Chilean Experiment." Economic Development and Cultural Change. 33 (January): 8-23. (1988). Exchange Rate Misalignment in Developing Countries. Baltimore: Johns Hopkins University Press. (1989a). Real Exchange Rates, Devaluation and Adjustment. Cambridge, Mass.: MIT Press. (1989b). "The International Monetary Fund and the Developing Countries." Carnegie-Rochester Conference Series. 31 (autumn): 7-68. (1992a). "Stabilization and Liberalization Policies in Central and Eastern Europe: Lessons From Latin America." In C. Clague and G. Rausser (eds.), The Emergence of Market Economies in Eastern Europe. Oxford: Basil Blackwell, 135-54. (1992b). "Exchange Rates as Nominal Anchors." Cambridge, Mass.: NBER Working Paper, no. 4246. Edwards, S., and A. Cox-Edwards. (1991). Monetarism and Liberalization: The Chilean Experiment. Chicago: University of Chicago Press. Edwards, S., and F. Sturzenegger. (1992). "Inflationary Inertia and Endogenous Indexation." Unpublished manuscript. UCLA. Edwards, S., and G. Tabellini. (1991). "Explaining Inflation and Fiscal Deficits in Developing Countries." Journal of International Money and Finance 10: 281-302. Flood, R., and P. Isard. (1988). "Monetary Policy Strategies." Cambridge, Mass.: NBER Working Paper no. 2770. Flood, R., and N. Marion. (1991). "The Choice of the Exchange Rate System." IMF Working Paper no. 90. Washington, D.C.: International Monetary Fund. Grilli, Vittorio, Donato Masciandaro, and Guido Tabellini. (1991). "Political and Monetary Institutions and Public Financial Policies in the Industrial Countries." Economic Policy: A European Forum 6, no. 2 (October): 341-92. International Monetary Fund. (1992). "Czechoslovakia Provides Case Study for Economic Transformation." IMF Survey, March. International Financial Statistics. Washington, D.C. Various issues. Kamin, S. B. (1991). "Exchange Rate Rules in Support of Disinflation Programs in Developing Countries." Washington, D.C: Board of Governors of the Federal Reserve Board, June. Kiguel, M., and N. Liviatan. (1990). "The Business Cycle Associated with Exchange Rate Based Stabilizations." Washington, D.C.: World Bank. Krugman, P. (1979). "A Model of Balance of Payments Crisis." Journal of Money, Credit and Banking 11, no. 3: 58-77.
Kydland, F , and E. Prescott. (1977). "Rules Rather Than Discretion: The Inconsistency of Optimal Plans." Journal of Political Economy 48, no. 3 (June): 473-91. Nunnenkamp, P. (1992). "Critical Issues of Macroeconomic Stabilization in PostSocialist Countries." Paper presented at conference Reintegration of Poland into the Western Economy, Warsaw.
338
Sebastian Edwards
Obstfeld, U. (1982). "Can We Sterilize: Theory and Evidence." American Economic Review 72, no. 2 (May): 41-50. Persson, T., and G. Tabellini. (1990). Macroeconomic Policy, Credibility and Politics. New York: Harwood. Sargent, T. (1983). "Stopping Moderate Inflations: Poincare and Thatcher." In R. Dornbusch and M. H. Simonsen (eds.), Inflation, Debt, and Indexation. Cambridge, Mass.: MIT Press, 28-43. Stockman, A. (1992). "International Transmission under Bretton Woods." Cambridge, Mass.: NBER Working Paper no. 4127, July. Swoboda, A. (1978). "Gold, Dollars, Euro-Dollars and the World Money Stock under Fixed Exchange Rates." American Economic Review 68, no. 4 (September): 625-42. Williamson, J. (ed.). (1990). Latin American Adjustment: How Much Has Happened?, Washington, D.C.: Institute for International Economics. (1991). "Advice on the Choice of an Exchange Rate Policy." In E. M. Claassen (ed.), Exchange Rate Policies in Developing and Post-Socialist Countries. San Francisco: ICG, 395-408.
CHAPTER 13
Capital inflows to Latin America with reference to the Asian experience Guillermo A. Calvo, Leonardo Leiderman, Carmen M. Reinhart
1
and
Introduction
The revival of substantial international capital inflows to Latin America is perhaps the most visible change in the economic situation of the region during the past two years. Whereas capital inflows to Latin America averaged about $8 billion a year in the second half of the 1980s, they surged to $24 billion in 1990 and to $40 billion by 1991. Of the latter amount, 45 percent went to Mexico, and most of the remainder went to Argentina, Brazil, Chile, Colombia, and Venezuela. Interestingly, capital is returning to most Latin American countries despite the wide differences in macroeconomic policies and economic performance between them. In most countries, the increased capital inflows have been accompanied by an appreciation in the real exchange rate, booming stock and real estate markets, faster economic growth, an accumulation of international reserves, and a strong recovery of secondary market prices for foreign loans. Without doubt, an important part of this phenomenon is explained by the fundamental economic and political reforms that have recently taken place in a number of these countries, including the restructuring of their external debts. Indeed, it would have been difficult to attract foreign capital in the magnitudes mentioned here without these reforms. Nevertheless, while domestic reform is a necessary ingredient for capital inflows, it only partially explains Latin America's forceful reentry in international capital markets. Domestic reforms alone cannot explain why capital inflows have occurred in countries that have not undertaken reforms or why they did The authors thank Sebastian Edwards and Jeffrey Frankel for their helpful comments on an earlier version of this chapter. This report is part of a research project by the authors on international capital flows, at the Research Department at the International Monetary Fund. The views expressed are those of the authors and do not necessarily reflect those of the International Monetary Fund.
339
340
Calvo, Leiderman, and Reinhart
not occur, until only recently, in countries where reforms where introduced well before 1990. This chapter maintains that some of the renewal of capital inflows to Latin America is due to external factors, and can be considered as an external shock common to the region. We argue that falling interest rates, a continuing recession, and balance-of-payments developments in the United States have encouraged investors to shift their resources to Latin America to take advantage of renewed investment opportunities and the increased solvency of that region.1 Taking into account economic developments outside the region helps to explain the universality of these inflows. The present episode may well be an additional case of financial shocks in the center that affect the periphery, of the type stressed by Diaz Alejandro in several of his contributions.2 International capital inflows affect the Latin American economies in at least four dimensions.3 First, they increase the availability of capital in the individual economies and allow domestic agents to smooth out their consumption over time and investors to react to expected changes in profitability. Second, capital inflows have been associated with a marked appreciation of the real exchange rate in most of the sample countries. The larger transfer from abroad has been accompanied by an increase in domestic absorption. If some of the increase in spending falls on the nontraded good, its relative price will increase - the real exchange rate appreciates. Third, capital inflows have an impact on domestic policy making. The desire by some central banks to attenuate the degree of real exchange rate appreciation in the short run frequently leads them to intervene actively and purchase from the private sector part of the inward flow of foreign exchange. Moreover, the attempt to avoid domestic monetization of these purchases has often led the monetary authorities to sterilize some of the inflows. The extent to which the inflows are sustainable is also of concern to the authorities. The history of Latin America gives reason for such concern: The major episodes of capital inflows, during the 1920s and 1978-81, were followed by major economic crises and capital outflows, such as in the 1930s and the debt crisis in the mid-1980s.4 Fourth, Latin America is not the only region that has experienced increased capital inflows in 1991. In fact, similar developments have occurred in Asia and the Middle East. At the same time, there had been a marked rise in capital outflows from the United States and Japan. See, e.g., Diaz-Alejandro (1983) and (1984). For a recent study of the effects of capital movements, see International Monetary Fund (1991). On the role of reforms and capital account liberalization, see Mathieson and RojasSuarez (1993). For a comparison of the current episode to the late 1970s, see Calvo, Leiderman, and Reinhart (1992).
Capital inflows to Latin America
341
capital inflows can provide important - yet ambiguous - signals to participants in world financial markets. An increase in the inflows can be interpreted as reflecting more favorable medium- and long-term investment opportunities in the receiving country. But capital may also pour in for purely short-term speculative purposes, so-called hot money, to a country where lack of credibility of government policies leads to high nominal returns on domestic financial assets. Although it remains to be seen which of these two motives is dominant in the present episode, the strong recovery in secondary market prices of bank claims on several of these countries (see Calvo, Leiderman, and Reinhart 1993) and various other indicators of country risk provide at least partial signals in support of the first, more favorable, motive.5 This essay has four objectives. The first is to document the current episode of capital inflows to Latin America based on data for ten Latin American countries.6 The second is to compare the Latin American experience with that of a number of Asian countries that have also been the recipients of sizable capital inflows and examine to what extent the nature of the capital inflows and macroeconomic consequences are similar in the two cases. The third is to assess the role of external factors in accounting for the observed capital inflows and the real exchange rate appreciation in Latin America. Last, the chapter discusses the implications of capital inflows for economic policy. The chapter is organized as follows. Section 2 deals with the basic concepts and the relationship between capital inflows, the accumulation of reserves, and the gap between national saving and investment. The stylized facts about capital inflows to the region are documented in Section 3, which includes a comparison with several Asian countries.7 Section 4 provides a quantitative assessment of the role of external factors on the accumulation of reserves and on the real exchange rate appreciation in the ten countries considered. The implications of capital inflows for domestic economic policy are discussed in Section 5. 2
The accounting of capital flows
International capital flows are recorded in the nonreserve capital account of the balance of payments (BOP). This account includes all international transactions with assets other than official reserves, such as transactions 5
6
7
For tracing on the evolution over time of individual country ratings, see e.g., LDC Debt Report. The countries included in our sample are Argentina, Bolivia, Brazil, Chile, Colombia, Ecuador, Mexico, Peru, Uruguay, and Venezuela. The countries that make up the Asian sample are Indonesia, Korea, Malaysia, Philippines, Singapore, Sri Lanka, Taiwan, and Thailand.
342
Calvo, Leiderman, and Reinhart
in money, stocks, government bonds, land, factories, and so on. When a national agent sells an asset to someone abroad, the transaction enters his country's balance of payments as a credit on the capital account and is regarded as a capital inflow. Accordingly, net borrowing abroad by domestic agents or a purchase of domestic stocks by foreigners is considered as a capital inflow, which respectively represents debt and equity finance. The simple rules of double-entry accounting ensure that, up to statistical discrepancies, the capital account surplus or net capital inflow (denoted by KA) is related to the current account surplus (denoted by CA) and to the official reserves account (denoted by RA) of the BOP through the identity: CA + KA + RA = 0.8 A property of the current account is that it measures the change in the economy's net foreign wealth. A country that runs a current account deficit must finance this deficit either by a private capital inflow or by a reduction in its official reserves. In both cases the country is running down its net foreign wealth. Another characteristic of the current account is that national income accounting implies that its surplus is equal to the difference between national saving and national investment {CA = S — I). Accordingly, an increase in the current account deficit can be traced to either an increase in /, a decline in S, or any combination of these variables that results in an increased investment-savings gap. Finally, the official reserves account records purchases or sales of official reserve assets by central banks. Thus, this account measures the extent of official foreign exchange intervention by the authorities, and is often referred to as the official settlements balance or the overall balance of payments. The foregoing discussion indicates that there are two polar cases of central bank response to increased capital inflows. If there is no intervention, the increased net exports of assets in the capital account are financing an increase in net imports of goods and services - capital inflows would not be associated with changes in central banks' holdings of official reserves. At the other extreme, if the domestic authorities actively intervene and purchase the foreign exchange brought in by the capital inflow, the increase in KA is matched, one-to-one, by an increase in official reserves. In this case, there is no change in the gap between national saving and national investment, nor is there any change in the net foreign wealth of the economy. The capital inflow would be perfectly correlated with changes in reserves. 8
Notice that RA < 0 implies accumulation of reserves by the monetary authority.
Capital inflows to Latin America 3
343
Stylized facts
In this section we quantify some of the key aspects of the current episode of capital inflows to Latin America and the related underlying macroeconomic developments.9 To document the regional aspects of this phenomenon we aggregate annual data and focus on Latin America as a whole.10 Monthly data for individual countries provide greater detail and are also discussed here and in the section that follows. We compare some of the stylized facts that have characterized the current Latin American capital inflow episode with the recent experience of a number of Asian countries. Last, we elaborate on the role of external developments, especially those in the United States. Anatomy of capital inflows Table 13.1 presents a breakdown of Latin America's balance of payments into its three main accounts. The capital inflows under consideration appear in the form of surpluses in the capital account, of about $24 billion in 1990 and about $40 billion in 1991. It can be seen that a substantial fraction of the inflows have been channeled to reserves, which increased by about $33 billion in 1990-91. During 1990-91 as a whole, the net capital inflow was equally split into a widening in the current account deficit and an increase in official reserves. The former suggests that capital inflows have been associated with an increase in the gap between national investment and national saving. In countries like Chile and Mexico, an important part of the inflows hasfinancedincreases in private investment; yet, in countries like Argentina and Brazil there has been a marked rise in private consumption.11 The sharp increase in official reserves, in turn, indicates that the capital inflow was met with a rather heavy degree of foreign exchange market intervention by the various monetary authorities. Part of the increased capital inflows represent repatriation of previous flight capital, but there are also new investors in Latin America.12 As Table 13.2 reports, the increase in net external borrowing accounts for 70 per9
10
11
12
See also "Latin American Finance and Investment," Financial Times, April 6, 1992, Kuczynski (1992), and Salomon Brothers (1992). For the purposes of the present section, Latin America includes the same set of countries included under Western Hemisphere in IMF's World Economic Outlook and International Financial Statistics. These figures, which are available from the authors, express investment and consumption as shares of GDP and rely on preliminary national income accounts data for 1991. On the role of various policy measures to reverse capital flight—such as amnesties, capital account liberalization, and introduction of foreign-currency denominated domestic instruments—see Collyns et al. (1992) and Mathieson and Rojas-Suarez (1993).
Table 13.1. Latin America: balance ofpayments, 1973-91
Balance on goods, services, and private transfers"
Balances on capital account"
Balance on capital account plus net errors and omissions"
Overall balance*
Year
$ billion (1)
% of GDP (2)
$ billion (3)
% of GDP (4)
$ billion (5)
% of GDP (6)
$ billion (7)
% of GDP (8)
1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991
-4.7 -13.5 -16.3 -11.8 -11.6 -19.4 -21.7 -30.3 -43.5 -42.2 -11.6 -3.2 -4.4 -18.9 -12.0 -12.4 -10.0 -8.8 -22.3
-2.4 -5.3 -6.1 -3.8 -2.7 -4.0 -3.8 -4.3 -5.5 -5.5 -1.7 -0.5 -0.6 -2.6 -1.6 -1.5 -1.1 -0.8 -2.1
_ — — — 19.8 30.5 35.0 47.0 59.4 45.1 22.4 15.5 6.7 14.2 14.5 8.2 15.7 24.1 38.1
_ — — —
8.5 13.3 14.7 16.9 16.4 27.4 32.9 34.0 41.9 23.0 13.6 12.5
4.4 5.2 5.5 5.4 3.8 5.6 5.8 4.9 5.3 3.0 1.9 1.8 0.8 1.7 2.0 0.6 1.3 2.3 3.9
3.8 -0.2 -1.6 5.1 4.8 8.0 11.2 3.7 -1.6 -19.2 2.0 9.3 1.1 -6.6 3.3 -7.7 2.1 15.1 17.5
-0.1 -0.6 1.6 1.1 1.6 2.0 0.5 -0.2 -2.5 0.3 1.4 0.2 -1.0 0.4 -0.9 0.2 1.4 1.7
4.6 6.2 6.2 6.7 7.4 5.9 3.2 2.3 0.9 1.9 1.9 1.0 1.7 2.3 3.8
5.5 12.3 15.3 4.7 12.1 23.9 39.8
2.0
"A minus sign indicates a deficit in the pertinent account. Balance on goods, services, and private transfers is equal to the current account balance less official transfers. The latter is treated in this table as external financing and is included in the capital account. *Column (7) equals the sum of columns (I) and (5). A positive sign in column (7) indicates accumulation of international reserves by the monetary authorities. Source: IMF, World Economic Outlook, various issues.
346
Calvo, Leiderman, and Reinhart
Table 13.2. Latin America: items in the capital account (in billions of U.S. dollars)
Year
Net external borrowing
Nondebt creating flows
Asset transactions (net)"
Errors and omissions"
Total
1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991
6.0 11.1 11.4 14.2 19.4 28.0 30.2 43.1 61.0 45.7 18.7 14.1 6.2 11.3 10.0 3.8 10.9 28.0 17.3
2.5 2.2 3.3 2.7 2.8 4.9 7.2 6.8 8.2 7.2 4.6 4.5 6.1 4.3 6.0 8.8 6.9 8.6 14.1
— — — -2.5 -2.5 -2.4 -3.0 -8.9 -7.7 -0.9 -3.1 -5.4 -1.3 -1.2 -4.3 -2.1 -12.5 6.7
— — — -3.4 -3.1 -2.1 -13.0 -17.5 -22.1 -8.8 -3.0 -1.4 -1.9 0.5 -3.5 -3.6 -0.2 1.7
8.5 13.3 14.7 16.9 16.4 27.4 32.9 34.0 41.9 23.0 13.6 12.5 5.5 12.3 15.3 4.7 12.1 23.9 39.8
"These two categories are included in net external borrowing and nondebt creating flows from 1973-76. Source: Data for Western Hemisphere, IMF, World Economic Outlook, various issues.
cent of the capital inflow in 1990-91. This is primarily due to borrowing by the private sector from foreign private banks.13 Increased external borrowing reflects the restoration of access to voluntary capital market financing after the debt crisis.14 There were also increases in portfolio investment and foreign direct investment. The latter amounted to about $12 billion, $4 billion of which was the result of privatizations.15 Since there has been a substantial degree of central bank intervention in the face of capital inflows, there is an important degree of comovement between official reserves and capital inflows. In fact, if one is interested in monthly developments, for which direct data on capital inflows are not 13 14 15
Some of this increased borrowing may represent hidden repatriation of flight capital. See, e.g., El-Erian (1992) and Collyns et al. (1992), chap. 3. For a comprehensive discussion of the composition of the inflows in the current episode and how it compares with the inflows of the late 1970s, see Collyns et al. (1992).
Capital inflows to Latin America
347
available, changes in reserves are a reasonable proxy for these inflows. Figure 13.1, which depicts monthly data on official international reserves for the countries in our sample, shows that, for most of the countries, there is a pronounced upward trend in the stock of official reserves starting from about the first half of 1990. In 1991, reserves accumulation accelerated as the monetary authorities in most countries reacted to the capital inflows by actively increasing their purchases of foreign assets constituting international reserves.16 Figure 13.2 provides evidence on the behavior of the real effective exchange rates during this period.17 With the exception of Brazil, all countries in our sample are experiencing a real exchange rate appreciation since January of 1991. In half of the cases the real appreciation of the domestic currency began before January 1991. Thus, the increase in capital inflows has been accompanied by a real exchange rate appreciation. This important link between capital flows and the real exchange rate in small open economies is already documented in the empirical literature (see Edwards 1989). Combining the evidence from Figures 13.1 and 13.2 indicates that there is an important degree of comovement in reserves and real exchange rates across countries, despite the wide differences in policies and institutions among them. Rates of return differentials and other macroeconomic developments Expected rates of return on available assets across countries play a key role in investors' decisions on whether to move capital internationally. Since data for expected returns are not readily available, and depend on how one models expectations, we first look at the stylized facts in the form of ex-post returns. As shown in Figure 13.3, there was a large increase in the U.S. dollar stock prices of major Latin American markets in 1991.18 Argentina exhibits the biggest single annual return of almost 400 percent, while Chile and Mexico registered returns of about 100 percent each.19 According to Salomon Brothers, $850 billion of foreign investment en16
17
18
19
Brazil and Uruguay are exceptions to this pattern, as in both countries capital inflows were not accompanied by an increase in reserves. The IMF indexes of the real effective exchange rate are used, hence an appreciation is represented by an increase in the index. The surge in stock prices during 1991 has been followed by more moderate declines in 1992. The price-earnings ratio in Argentina increased from 3.1 in 1990.4 to 38.9 in 1991.4; in Chile it increased from 8.9 in 1990.4 to 17.4 in 1991.4; and in Mexico it moved from 13.2 in 1990.4 to 14.6 in 1991.4. These figures are from Emerging Markets Data Base, International Finance Corporation.
348 ARGENTINA
BOLIVIA
BRAZIL
CHILE
JaivM
COLOMBIA
ECUADOR
Jwv*t
JaivM
349 MEXICO
PERU
URUGUAY
VENEZUELA
Figure 13.1. Latin America: total reserves minus gold, January 1988-July 1992 (billions of U.S. dollars). Source: IMF, International Financial Statistics.
350 ARGENTINA
Jtavto
Jbo-»1
BOLIVIA
Jtn-tt
BRAZIL
CHILE
COLOMBIA
ECUADOR
351 MEXICO
PERU
URUGUAY
VENEZUELA
Figure 13.2. Latin America: real effective exchange rate, January 1988-July 1992. Note: An increase in the index denotes a real exchange rate appreciation. Source: IMF, Information Notice System.
352
Calvo, Leiderman, and Reinhart 12 11 109
Argentina
8" 7 6" 54 32"
1 0 Jan-88
Jan-89
Jan-90
Jan-91
Jan-92
Jan-88
Jan-89
Jan-90
Jan-91
Jan-92
Figure 13.3. Stock market performance, January 1988-August 1992 (stock price indexes in U.S. dollars, January 1988 = 100). Note: The S & P 500 index was used for the United States. Sources: Standard & Poor's and International Finance Corporation, Quarterly Review of Emerging Stock Markets.
tered Brazil's stock market in the last four months of 1991, and about $600 million was invested by foreigners in the Argentine equity market in 1991.20 However, as the figures indicate and Figure 13.3 confirms, the stock market booms and the attendant high returns appear to materialize after capital has begun to flow into the region. It would thus be difficult to argue that high stock market return differentials were responsible for attracting the first wave of capital inflows. 20
See Salomon Brothers (1992).
Capital inflows to Latin America
353
Figure 13.4 provides evidence on the lending and deposit interest rate spreads between U.S. dollar equivalent domestic interest rates and interest rates in the United States. Because in some of these countries interest rates are regulated, and capital mobility is imperfect, spreads across the various countries cannot be compared in a straightforward manner. In addition, domestic interest rates vary markedly from country to country, as shown in Figure 13.4, with Argentina and Peru having the broadest ranges and Bolivia and Colombia the narrowest. With these caveats in mind, the dominant impression from Figure 13.4 is that of relatively high interest differentials in Latin America in the 1990-91 period. It is also evident from Figure 13.4 that the pattern of spreads varies considerably across countries. In effect, this is not surprising since the monetary authorities in these countries have not reacted in a uniform manner to the capital inflows, and the timing of regulatory changes has also varied considerably across the sample countries. Although the relatively high differential rate of return on Latin American assets has been associated with a marked rise in capital inflows to the region, the inflows have not arbitraged away the large differentials. In some countries, such as Argentina, the interest rate differential decreased sharply as capital poured in; yet in others, such as Chile, there was a less pronounced response of the interest rate differential to the inflows (see Figure 13.4). As argued in Section 5, these different patterns may reflect cross-country differences in the authorities' choices between sterilized and nonsterilized intervention. In sum, three main stylized facts emerge with regard to interest rate differentials. First, there is little comovement in domestic interest rates (in U.S. dollars), and hence in spreads, across the countries in our sample. Second, as illustrated in Calvo, Leiderman, and Reinhart (1993), the "noise-to-signal ratio" of the domestic dollar rates varies substantially across countries. Countries offering the highest returns also had the greatest volatility of returns.21 Third, despite the capital inflows the positive differentials have not been fully arbitraged away. Consider how developments in 1991, the year when capital inflows grew to about $40 billion, differ from those in earlier years. First, as Table 13.3 shows, there was a renewal of economic growth. After three years of stagnation, real GDP increased by almost 3 percent in 1991. However, gross capital formation as percent of GDP remained constant at about the same level of the second half of the 1980s, suggesting a more efficient utilization of resources. At the same time, there was a marked drop in the rate of 21
An implication of this discussion is that from the investor's perspective, the information content of a drop in U.S. interest rates is different from that of an equal rise in the domestic interest rate—while in both these cases the interest rate differential would change by the same amount.
354 ARGENTINA
BOLIVIA
BRAZIL
CHILE
« M I i M I Jtottt JUM .tot* JkMI Ljndto SwMd it <Mrwd M ttw Bw* mmtornUA Tiwwwv B« m
COLOMBIA
M l
i M I JMVW JUUft Jwt-aO JkjMO JwwM
ECUADOR
355 MEXICO
PERU
URUGUAY
VENEZUELA
Figure 13.4. Interest rate spreads, January 1988-March 1992 (dollar equivalent of domestic rate less U.S. rate, annual rates). Note: Deposit spreads are based on interest rates on certificates of deposit, while lending spreads are rates charged by banks less the interest rate on U.S. commercial paper. Source: IMF, International Financial Statistics, and various central bank bulletins.
Table 13.3. Latin America: macroeconomic indicators Gross Growth of capital Consumption" Inflation real GDP formation Year (% change) (%ofGDP) (%ofGDP) (%)
Central External govt. fiscal Commodity Terms of prices trade debt balance (%ofGDP) (% change) (% change) ($ billion)
8.4 1973 1974 6.9 3.1 1975 1976 5.5 5.3 1977 1978 4.1 1979 6.1 1980 5.3 1981 1.0 1982 -0.9 1983 - 3 . 2 1984 3.6 1985 3.4 1986 4.3 1987 2.2 0.4 1988 1.0 1989 1990 -0.1 1991 2.9
— — — -2.2 -2.0 -0.7 -0.6 -3.0 -4.0 -3.7 -4.1 -4.0 -5.2 -7.0 -5.8 -6.3 -0.3 -1.0
29.3 24.4 24.7 23.5 25.1 24.8 23.2 23.7 23.0 20.9 17.9 17.2 18.4 18.2 19.9 20.9 19.6 19.6 20.7
74.5 75.8 77.7 79.6 79.0 78.6 79.6 79.7 79.7 80.1 81.2 79.0 76.2 78.9 75.1 72.7 72.6 77.0
32.1 37.5 52.0 66.1 49.9 41.9 46.5 53.7 58.2 64.6 98.6 124.2 128.2 79.4 117.8 243.2 434.2 647.8 162.5
47.4 20.9 -12.5 23.0 27.9 -12.6 14.0 11.8 -15.3 -11.0 6.8 -0.8 -8.3 5.5 -6.8 21.2 -2.3 -7.2 -5.6
6.5 -7.0 -7.5 12.2 8.0 -9.4 5.0 7.9 -5.2 -5.0 -2.7 4.2 -5.4 -10.2 -5.4 -0.6 0.2 -0.1 -4.9
44.4 58.2 68.6 82.0 124.6 154.9 187.2 229.4 285.6 325.5 340.2 360.3 368.2 381.9 419.1 409.3 406.9 422.1 440.7
External debt to exports (ratio)
Reserves to Debt service Reserves imports (ratio) ($ billion) (ratio)
176.2 163.4 195.8 204.1 192.6 215.7 196.8 181.8 207.7 264.7 286.0 275.1 293.7 347.9 341.4 294.7 262.2 251.6 264.8
29.3 27.9 32.2 31.4 28.2 37.1 38.8 33.0 40.6 50.4 40.7 40.5 42.1 46.1 38.5 42.7 30.2 26.9 32.8
"This column includes private and government consumption. Source: Data for Western Hemisphere, IMF, World Economic Outlook and International Financial Statistics, various issues.
12.0 11.9 10.0 15.2 28.5 35.5 42.7 40.3 39.6 28.1 29.3 40.5 41.2 33.3 38.0 30.8 33.0 47.8 65.3
35.1 21.7 17.5 25.8 28.5 35.5 42.7 40.3 39.6 28.1 29.3 30.2 31.7 25.6 27.7 20.0 19.4 26.3 33.5
Capital inflows to Latin America
357
inflation (which nevertheless remained at a three-digit level for the region), and a significant reduction in central government fiscal deficits. The changing economic conditions in Latin America are also reflected in the region's debt and solvency indicators. At $441 billion, external debt amounts to 2.6 times exports of goods and services. Although still high, this ratio has decreased markedly from the 3.5 figure in 1986. Since most of Latin America's external debt to commercial banks is still in terms of floating rates, the drop in short-term U.S. interest rates and the drop in the debt to exports ratio has translated into a rapid decline in the external debt service ratio over the past two years. In fact, the level of the debt service ratio in 1991 is of the same order of magnitude as the levels observed before the debt crisis. These developments represent only part of the changing environment in Latin America of the early 1990s. In addition to these, the move toward privatization and deregulation, the introduction of financial reforms, and the restructuring of existing external debt have all contributed to bringing Latin America back on the list of viable investment locations in world financial markets. A comparison with the Asian experience Latin America has not been the only region receiving sizable capital inflows in recent years. In effect, capital began to flow to Korea in 1988 and to a broader number of Asian countries sometime in 1989. Developments outside the region are frequently credited for the flow of capital to a group of countries that are, by and large, pursuing very diverse policies and have considerable differences in their macroeconomic environment. Specifically, it is argued that declining profit margins in Japan and in the United States have induced Japanese, and, to a lesser extent, American firms to reallocate to areas where lower wages prevail. In addition, during 1988-89 a number of the emerging stock markets in Asia outperformed U.S. and Japanese stock markets by considerable margins.22 Several interesting empirical regularities emerge from comparing the Latin American and Asian experience. First, as Table 13.4 illustrates, the swing in the balance on the capital account (as a percent of GDP) is of the same order of magnitude for the two regions. For the Latin American countries in our sample the change in the capital account amounts to 2.5 percent of GDP; for the Asian countries the capital account surplus widens by 2.3 percent. Second, as is the case for most of the Latin American countries, there is a marked accumulation of international reserves during 22
For instance, the International Finance Corporation (IFC) Asia composite, which includes Korea, Malaysia, Taiwan, and Thailand, registered total returns (in dollars) of 83 percent and 57 percent in 1988 and 1989, respectively.
Table 13.4. Key indicators for selected Latin American and Asian countries (As a percent of GDP) Latin America Bolivia
Brazil
Chile
Colombia
Ecuador
Mexico
Peru
Uruguay
Venezuela
Average for 10 countries
Capital account 1984-89 - 1 . 6 1990-91 0.0
0.6 3.7
-2.3 -1.1
-1.7 4.1
2.0 0.9
-6.3 -5.0
-0.4 5.4
-5.3 -0.7
-2.5 0.3
-3.1 -3.4
-2.1 0.4
Direct investment 1984-89 0.9 1990-91 1.9
0.5 0.9
0.5 0.1
0.5 1.9
1.5 1.1
0.6 0.8
0.8 1.4
0.0 0.1
0.0 0.0
0.1 2.2
0.5 1.1
Investment 1984-89 11.1 1990-91 8.7
5.4 5.6
17.2 16.0
16.0 18.5 19.5 15.8
19.2 17.8
18.6 19.1
19.6 13.7
11.0 12.7
16.5 16.1
15.3 14.5
11.3 11.8
11.1 12.9
10.9 8.9
10.3 10.7
11.6 8.4
11.1 9.9
9.6 9.0
13.7 13.3
10.4 11.3
11.2 11.0
Argentina
Public consumption 1984-89 12.4 1990-91 13.6
Asia Taiwan
Thailand
Average for 8 countries
4.6 4.8
0.1 -6.3
4.2 11.7
1.2 3.5
9.4 12.9
0.6 0.4
-0.3 -2.2
0.8 1.7
1.8 2.7
18.3 20.8
38.9 38.7
23.0 20.1
19.3 22.3
21.8 27.9
25.0 27.8
6.8 7.9
12.8 10.8
9.6 10.1
15.0 16.0
13.6 10.1
11.3 10.5
Korea
Malaysia
Philippines
Capital account 1984-88 2.2 1989-91 4.6
-2.0 0.8
-0.4 6.1
-3.8 2.0
5.0 3.9
Direct investment 1984-88 0.5 1989-91 1.1
0.3 0.1
2.7 5.9
0.8 1.5
28.9 36.2
26.0 31.7
9.8 9.6
12.5 9.0
Indonesia
Investment 1984-88 1989-91
23.8 24.8
Public consumption 1984-88 10.0 1989-91 10.4
Singapore
Sri Lanka
Source: IMF, International Financial Statistics; Pfefferman and Madarasay; (1993); and IMF, World Economic Outlook.
360
Calvo, Leiderman, and Reinhart
the capital inflow period of 1989-91 (see Figure 13.5). The sharp buildup in international reserves in the eight Asian countries considered suggests that, as in Latin America, the capital inflow was met with a heavy degree of intervention on the part of the various monetary authorities. Third, as mentioned previously, the various regional stock markets posted strong gains during the early stages of the capital inflow period.23 There are, however, marked differences between Asia and Latin America in the macroeconomic impact of the capital inflows. As Figure 13.2 illustrates, in Latin America the capital inflows have been accompanied by a real exchange rate appreciation (the exception is Brazil); in Asia such an appreciation is not the norm (Figure 13.6). The real exchange rate appreciated markedly in Korea and more modestly in Singapore but for the remaining countries no sustained appreciation is evident. As previously argued, the larger transfer from abroad is accompanied by an increase in domestic absorption. If some of the increase in spending falls on the nontraded good, its relative price will increase - the real exchange rate appreciates. Although the reasons why the real exchange rate responds differently to the inward flow of capital in the two regions are likely to be numerous, important differences in the composition of aggregate demand may play a key role in determining whether the real exchange rate appreciates or not. As Table 13.4 summarizes, for the Asian countries investment as a share of GDP increases by nearly 3 percentage points during the capital inflows period. By contrast, for the Latin American countries investment falls slightly - the inflows primarily finance higher consumption. It has often been the case for these countries that the increase in investment falls primarily on imported capital goods. On the other hand, the increase in consumption is less tilted toward the traded good. Other things equal, this observation would suggest that a real exchange rate appreciation is more likely when capital inflows finance consumption than when these finance investment.24 Another element influencing the real exchange rate by affecting both the level and composition of aggregate demand is the behavior of public consumption. Some of the Asian countries, most notably Malaysia and Thailand, reacted to the capital inflows by sharply contracting fiscal expenditure.25 These expenditure cuts may reduce or eliminate the real exchange rate pressures through two channels: First, the fiscal 23
24 25
As Figure 13.3 illustrates, most of the Latin American stock markets have recently given up some of the earlier gains. Similarly, the Asian markets weakened during 1990 and 1991 after the earlier surge. This, of course, does not explain Chile and Korea, where there has been a real exchange rate appreciation alongside a sharp rise in investment. For each of those two countries the decline amounted to 3.5 percent of GDP.
Capital inflows to Latin America
361
contraction tends to reduce aggregate demand; second, public consumption may be more biased toward the nontraded good than private sector consumption.26 Yet another reason that may explain why a real exchange rate appreciation failed to materialize was the conduct of monetary policy. While sterilization policies had limited success in reigning in monetary growth in a number of Latin American countries (see Section 5), these policies, which were often conducted by managing public sector savings, were more successful in achieving their objectives in several Asian countries (for a discussion of the experience of Indonesia, Malaysia, Singapore, and Taiwan, see Reisen 1993). Their relative success in limiting the acceleration in the growth of the monetary aggregates had a dampening effect on aggregate demand and limited pressures on the price of nontraded goods.27 Another marked difference between Asia and Latin America is in the composition of capital inflows. Whereas in the Asian countries 40 percent of the increase in capital inflows came in the way of foreign direct investment, for the Latin American countries direct investment accounted for only 20 percent of the increase in inflows. This difference may help explain why concerns over "hot money" and a sudden reversal are more prevalent among Latin American policy circles than among their Asian counterparts. It may also, in part, explain why the increase in investment is much greater for most of the Asian countries. External factors Although it is difficult to point to a single dominant external factor that would account for the present capital inflows to Latin America, several external developments have converged to stimulate such inflows. First, there is the impact of the sharp drop in U.S. short-term interest rates, which are now at about half their level of two years ago and at their lowest levels since the early 1960s. By reducing the external debt service on floating rate debts, this decline in U.S. interest rates has improved the solvency of Latin American debtors. For a given level of interest rates in Latin America, these developments provide incentives for repatriation of capital held in the United States and for increases in borrowing by Latin American agents from capital markets in the United States.28 26 27
28
This does not suggest that fiscal adjustment has not taken place in a number of Latin American countries, but rather that its timing did not coincide with the capital inflows. See Edwards (1989) for a comprehensive discussion of how these and other economic "fundamentals" affect the real exchange rate. Beyond short-term interest rates, returns from other investments in the United States have decreased recently as well—e.g., in the real-estate market.
362 INDONESIA
KOREA
MALAYSIA
PHILIPPINES
363 SINQAPORE
TAIWAN
SRI LANKA
THAILAND
Figure 13.5. Asia: official reserves minus gold, January 1987-July 1992 (billions of U.S. dollars). Source: IMF, International Financial Statistics.
364
INDONESIA
1987
1
KOREA
1987
1989
1990
1991
1992
PHILIPPINES
MALAYSIA
1067
1988
1987
1989
1990
1991
1992
364
INDONESIA
1987
1
KOREA
1987
1989
1990
1991
1992
PHILIPPINES
MALAYSIA
1067
1988
1987
1989
1990
1991
1992
366 Calvo, Leiderman, and Reinhart Second, several external factors probably contributed to the increase in Latin America's current account deficit, and to the need to finance this deficit by increased capital inflows. Two such factors are the continuing recession in the United States and in other industrialized countries, and the continuation of the process of decline in Latin America's terms of trade throughout the past decade - which reflects mainly a decrease in the prices of petroleum and of other commodities. In principle, a decline in a given country's terms of trade can be expected to result in a larger current account deficit (the Harberger-Laursen-Metzler effect) and, in the absence of major intervention by the national authorities, in a larger capital inflow tofinancethis deficit. However, the changes in the terms of trade in 199091 are too small to account for the sharp increase in capital inflows. Thus it seems financial market shocks have played a larger role than terms of trade shocks in accounting for the capital inflows in the present episode. Third, it is seen that during both recent episodes of capital inflows to Latin America - in 1978-82 and 1990-91 - there were sharp swings in the private capital account of the U.S. balance of payments in the form of increased outflows and reduced inflows (Table 13.5). In fact, 1990 and especially 1991 mark thefirstyears of net capital outflows from the United States29 after eight consecutive years of net inflows!30 As Table 13.6 documents, about 60 percent of the increased capital inflows in 1991 are directly associated with increased private capital outflows from the United States to Latin America, as recorded in the U.S. BOP accounts. Similarly, the relatively large capital inflow of 1978-81 to Latin America was matched by increased private capital outflows from the United States.31 Thus the data appear to support the notion that swings in private capital outflows from the United States play a key role as external impulses that affect the size of capital inflows into Latin America. 29
30
31
S o m e examples of this development are as follows: (1) There has been an increase in the amount o f investments in foreign securities by mutual funds in the United States. A s of May 1992, the assets o f stock funds that invest largely outside the United States stood at $41.8 billion, more than twice the level at the end o f 1988, and assets of global funds have soared to $28.5 billion from just $3 billion in 1988. (2) In 1991, the sale of foreign shares in public and private deals doubled, to a record $9.78 billion. Bond deals rose 48 percent to $55.33 billion. (3) N e w foreign investment in U.S. companies and real estate plummeted 66 percent in 1991. See the New York Times, July 5, 1992. A s indicated earlier, private capital outflows from Japan also increased sharply, by $36 billion, in 1991. It is useful to recall how sizable these inflows to the United States were in the mid-1980s (Table 13.4). From net capital outflows of about $20 billion a year in the late 1970s, the private capital account turned around into surpluses (capital inflow), which peaked at $128 billion in 1985. This inflow, which mainly took the form of increased borrowing from abroad, was mostly used to finance high and increasing current account deficits that were well above $100 billion in the second half o f the 1980s.
Capital inflows to Latin America
367
Table 13.5. U.S. balance of payments (in billions of U.S. dollars)
Year
Current account
Capital account
Capital account plus net errors and omissions
Overall balance
1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991
7.07 1.94 18.06 4.18 -14.49 -15.40 0.20 1.20 7.26 -5.86 -40.18 -98.99 -122.25 -145.42 -162.22 -128.99 -106.41 -92.16 -8.66
-9.71 -9.25 -28.67 -25.24 -18.46 -30.63 -14.53 -35.91 -28.07 -28.79 24.72 72.52 108.18 95.78 98.68 101.05 104.91 -4.60 -18.20
-12.30 -10.75 -22.71 -14.68 -20.55 -18.08 9.75 -10.26 -8.50 7.89 36.13 99.71 128.05 111.64 105.36 92.72 123.34 58.90 -21.30
-5.23 -8.81 -4.65 -10.50 -35.04 -33.48 9.95 -9.06 -1.24 2.03 -4.05 0.75 5.80 -33.78 -56.86 -36.27 16.93 -33.26 -29.96
Source: IMF, International Financial Statistics, and U.S. Bureau of Current Business, Survey of Current Business, various issues.
Table 13.6 Changes in capital accounts (in billions of U. S. dollars)
Periods compared
Private capital account of Western Hemisphere
Private capital account of US. with Western Hemisphere
1978-81 against 1976-77 1983-89 against 1978-81 1991 against 1983-89
17.4 -24.4 30.1
-9.9 30.1 -17.5
Note: Positive entries in column 1 indicate an increase in net private capital inflow to Western Hemisphere. A negative entry in column 2 indicates an increase in the net private capital outflow from the United States to the Western Hemisphere.
368
Calvo, Leiderman, and Reinhart
Fourth, in 1990 there were important regulatory changes in capital markets of industrial countries that reduced the transactions costs for agents accessing international capital markets from developing countries.32 4
Role of external factors: econometric analysis
In this section, monthly data for ten Latin American countries covering the period January 1988 to December 1991 are used to analyze in more detail key features of the current episode of capital inflow. The analysis begins by establishing the extent of comovement of official reserves and real exchange rates between these countries, as these proxy for capital inflow. We conduct a similar exercise for the Asian countries and compare the results. We then develop and estimate a model designed to provide a quantitative assessment of the relative importance of external shocks in the recent episode of reserves accumulation and real exchange rate appreciation. Comovement of reserves and the real exchange rate Given the lack of monthly data (and for a number of the countries in the sample, quarterly data) on capital inflows, we examine here the joint behavior of international reserves and the real exchange rate, two variables in the present episode that are closely associated with the inflows. The previous section revealed that there is an important degree of comovement in reserves and real exchange rates across countries, which could be interpreted as reflecting the effects of a common external shock to Latin American countries (Figures 13.1 and 13.2). Accordingly, a first task in this section is to quantitatively examine this issue by using principal component analysis. Principal component analysis provides a way of describing the comovement in data series.33 We begin with ten time series, reserves for each country, and construct a smaller set of series, the principal components, which explain as much of the variance of the original series as possible.34 The higher the degree of comovement that exists among the original ten series, the fewer the number of principal components that will be 32 33
34
See El-Erian (1992) for a comprehensive discussion. For an exposition of principal components analysis, see, e.g., Dhrymes (1970). Swoboda (1983), in an application that is close in spirit to ours, used this approach to examine economic interdependence across different exchange rate regimes for six of the G-7 countries. All the analysis that follows uses the logs of reserves and of the real exchange rate.
Capital inflows to Latin America
369
needed to explain a large portion of the variance of the original series.35 The procedure begins by standardizing the variables, so that each series has a zero mean and a unit standard deviation; this ensures that all series receive uniform treatment and that the construction of the principal component indexes is not influenced disproportionately by the series exhibiting the largest variation. For the Latin American countries we constructed the principal component indexes for the period from January 1988 to July 1992. In addition, for comparative purposes two subperiods are considered: 1988-89 and the capital inflows episode of 1990-92. As Figure 13.1 shows and the top panel of Table 13.7 confirms, the extent of comovement in reserves during the capital inflow period of 1990-92 is considerable and higher than in the preceding two years. The first principal component explains 71 percent of the variation in reserves, while the second principal component explains an additional 16 percent of the variation. Hence, 87 percent of the variance of the ten reserves series is captured by two indexes, thus indicating a sizable degree of comovement. More formally, we tested the null hypothesis that the ten reserve series are linearly independent and found that we could reject this hypothesis at standard significance levels.36 Applying the same procedure to the real exchange rate indicates that the degree of comovement across countries in the region also has increased in the recent capital inflows episode. The fraction of real exchange rate variance explained by the first principal component during 1990-92 is 68 percent. The first two principal components explain a sizable 83 percent of the variance of the real effective exchange rate. As far as the increased covariation of reserves and the real exchange rate in the recent period is concerned, it may well reflect the effects of an external shock, common to the region, in the past two years. Interestingly, when we examined the principal components of the domestic inflation rate, a variable less obviously linked to external factors, we found that the extent of covariation among the inflation rates of these ten countries had diminished rather than increased in the recent period.37 The correlations between the first principal component of reserves and 35
36
37
If the ten series were identical (perfectly collinear), the first principal component would explain 100 percent of the variation of the original series. Alternatively, if all ten series were perfectly uncorrelated, it would take ten principal components to explain all of the variance in the original series; no advantage would be gained by looking at common factors, since none exist. The test statistics, which are distributed as a x1 with 45 degrees of freedom, and the attendant probability values are presented at the bottom of Table 13.6. Applying a different methodology Engle and Issler (1992) find significant comovement in the per capita GDP of several Latin American countries, as these countries share common trends and common cycles.
Table 13.7. Establishing the comovement in macroeconomic series Latin America
Asia
1988.1 to 1992.7 1988.1 to 1989.12 1990.1 to 1992.7 1987.1 to 1992.7 1987.1 to 1988.12 1989.1 to 1992.7 Cumulative R2 Cumulative R2 Cumulative R2 Cumulative R2 Cumulative R2 Cumulative R2 Real exchange rate First principal component Second principal component Chi-squared Probability value
0.44
0.41
0.68
0.43
0.72
0.50
0.75 623.544 (0.0000)
0.78 302.01 (0.0000)
0.83 401.65 (0.0000)
0.71 561.24 (0.0000)
0.88 296.25 (0.0000)
0.75 379.10 (0.0000)
0.66
0.48
0.71
0.67
0.63
0.76
0.79 603.73 (0.0000)
0.69 204.97 (0.0000)
0.87 395.39 (0.0000)
0.88 897.44 (0.0000)
0.83 268.65 (0.0000)
0.87 591.67 (0.0000)
0.38
0.60
0.45
0.62
0.60
0.39
0.67 589.67 (0.0000)
0.88 475.94 (0.0000)
0.64 306.4 (0.0000)
0.78 498.4 (0.0000)
0.77 327.44 (0.0000)
0.67 204.67 (0.0000)
Reserves First principal component Second principal component Chi-squared Probability value Domestic inflation rate 12-month percent change First principal component Second principal component Chi-squared Probability value
Notes: The cumulative R2 gives the percentage of the variance of the original series explained by the first principal component, the first two principal components, and so on. Chi-squared is 45 for Latin America, 28 for Asia.
Capital inflows to Latin America
371
the individual country reserve series tend to confirm the evidence in Figure 13.1. The regional index does quite well in accounting for reserve fluctuations in eight of the ten countries. For the real exchange rate, the results are also anticipated in Figure 13.2.38 The first principal components (plotted in the top panel of Figure 13.7) could be interpreted as regional exchange rate and reserves indexes, and as shown, the upward trend in the two series reflects the common regional experience of a real exchange rate appreciation and accumulation of reserves. Purged of country-specific idiosyncracies, they could reflect the influence of unobservable external factors common to the region as well as any coordinated internal developments in the region.39 Applying the same methodology to the Asian data highlights some of the differences as well as some of the similarities between the experience of the two regions. As the bottom panel of Figure 13.7 shows, the pattern of comovement in reserves is similar to that found for the Latin American countries, with share of the total variation explained by thefirstprincipal component increasing during the capital inflow period. Similarly, the bottom panel of Figure 13.7 traces the "regional" reserve index and points to a sustained accumulation of reserves. By contrast to the Latin American experience, the degree of comovement in the real exchange rate during the capital inflow period diminishes. Most important, as the bottom panel of Figure 13.7 illustrates, the regional real exchange rate index captures the pattern most prevalent in Figure 13.6; namely, it highlights that the "regional" real exchange rate remained fairly stable in the face of capital inflows. Quantifying the role of external factors In this section, the analysis proceeds in two stages: We first construct indexes of the unobserved external factors (or impulses), which are then incorporated in a structural vector autoregression. Second, we perform tests of exclusion restrictions on the foreign factors to determine their sta38
39
Notice that, as shown in Figure 13.3, Brazil's real exchange rate depreciated through most of the sample period and its upturn came fairly late in the sample. Thus, it is not surprising to find that the regional exchange rate index, the first principal component, does poorly in capturing its fluctuations. In effect, their correlation is negative. These details are available upon request. Calvo, Leiderman, and Reinhart (1993) explore the possible role of external factors by examining the simple pairwise correlation coefficients between the principal components indexes for reserves and the real exchange rate and a set of variables from the United States. It is hypothesized that a fall in US. interest rates, stock market returns, real-estate returns, and economic activity would be associated with an increase in the capital inflow to Latin America which would be at least partly reflected in an increase in the regional indexes for reserves and the real exchange rate (the latter indicating a real exchange rate appreciation). As shown there, most of the evidence is indeed in this direction.
372
Calvo, Leiderman, and Reinhart
Latin A m • rica January 1988 - July 1992 2.5
Regional Index of Official Reserves
Regional Index of tha Real Exchange Rate
Jan-88 Jul-88 Jan-89 Jul-89 Jan-90 Jul-90 Jan-91 Jul-91 Jan-92 Jul-92
Asia January 1987 - July 1992 2.5 Regional Index of Official Reserves
2" 1.5-
\
1 0.5"
Regional Index of the Real Exchange Rate Jan-87
Jan-88
Jan-89
Jan-90
Jan-91
Jan-92
Figure 13.7. First principal components. Note: An increase in the real exchange rate index denotes an appreciation. Principal components indexes are constructed to have zero mean and unit variance. tistical significance. In modeling the external impulses, one could consider a whole vector of variables that could have an impact on Latin American economies. Here we opted for an unobserved index model, where the constructed index is correlated with the observed time series on a set of U.S. variables, which includes the nominal rates of return on real estate, stock and bond markets, short-term deposit and lending rates of interest, and detrended real disposable income. Specifically, we constructed and used
Capital inflows to Latin America
373
the first and second principal components of these series. The first principal component captures the joint movement of the various interest rates and economic activity in the United States. The second principal component captures swings in returns on the equity and real estate markets. Having now a measure of external impulses, we embedded them in a structural vector autoregression. Defining PCI, and PClt as the first and second principal components of the U.S. variables and denoting the logs of reserves and the real exchange rate by RESt and REXr respectively, the reduced form of the system is given by:
PCI, = a, + V + X ftiTClr_, + X /= i
/= i
PCI, = a2 + y2t + £ ft^Cl,., + £ fUJCl,., + «f« /= 1
/= 1
RES, = «3 + y3t + £ 0,,PC1,_,. + £ PuPCl,., + £ SJiES,.,
(1)
+£ , =a 4 + y
+£ S As equation (1) illustrates, we allow for dynamic interaction between the foreign factors but impose their temporal exogeneity by not including lagged values of the endogenous variables, reserves and the real exchange rate, in their respective equations (i.e., 8 h = 82/ = S'If. = S'2j = 0); hence, we impose structure on the temporal relationships between these variables.40 Each equation in the system includes a constant and a time trend. Since the tests could be affected by the number of lags included in the right-hand side of each equation, and given that we had no strong priors on this issue, we used the Akaike and Schwarz criteria to select among one-, three-, six-, nine-, and twelve-month lag profiles.41 Both criteria, unless otherwise noted, yielded three lags as optimal. The reduced-form residuals, the M,'S depend on the structural errors, et, and the contemporaneous relationships between the endogenous variables, specifically, ut = etA. So next, we consider the structure of the matrix A, which describes the contemporaneous relationships between the variables. Here we follow the methodology of Bernanke (1986) and Blanchard 40
41
Our procedure is similar to the DYMIMIC models associated with Watson and Engle (1983), and Stock and Watson (1989). One key difference in the approaches is that here we adopt a two-step procedure by first constructing the unobserved factor index (indexes) and then incorporating that factor(s) in a dynamic model. For simulation evidence on the efficacy of these criteria, see Lutkepohl (1985).
374 Calvo, Leiderman, and Reinhart (1989), in that a priori (structural) restrictions are imposed on the identifying matrix. Specifically, since there is a presumption that the foreign factors are exogenous, we do not allow for feedback from shocks to the domestic variables to the reduced form error of thefirstand second principal components of the foreign variables. In addition, we impose the restriction that the principal component indexes are orthogonal by construction, so that they depend on their own shocks, as in equations (2) and (3): PCI, = e™,
(2)
PClt = er,
(3)
while reserves are affected by the structural shocks to the foreign variables and by its own shock, RESt = alxPC\t + anPClt + e™.
(4) EX
REXt = a4lPC\t + a42PC2t + a43RESt + e? .
(5)
The real exchange rate is allowed to respond to all of the shocks.42 After the system was estimated using monthly data from January 1988 to November 1991, we tested for the significance of the foreign factors. Table 13.8 summarizes the results of the tests for exclusion restrictions, tests that involve the temporal relationships. The null hypothesis being tested is that the foreign variables do not affect reserves and the real exchange rate. The high x2 statistics and low probability values indicate that in eight of the ten countries, one can reject the null hypothesis at the 75 percent level of confidence or higher.43 Only in half of the sample countries is there any evidence of a significant contemporaneous relationship between the foreign factors and reserves and/or the real exchange rate. Calvo, Leiderman, and Reinhart (1993) examine in greater detail the relative importance of the foreign factors as well as the impulse response functions of reserves and the real exchange rate. By examining the variance decompositions of the real exchange rate and official reserves, those results indicate that for most countries a sizable fraction (about 50 percent) of the monthly forecast error variance in the real exchange rate and reserves is accounted for by foreign factors with foreign factors explaining the greatest share in countries that experienced no major changes in domestic policies in the period under consideration, 1988-91. 42
43
Alternative orderings are explored. One alternative imposes that there be no contemporaneous relationship between reserves and the real exchange rate, while another treats reserves as the most "endogenous" variable in the system. The results do not differ appreciably from those presented here. Evidence suggesting the importance of U.S. economic developments on the Latin American business cycle is presented in Engle and Issler (1992).
Table 13.8. Tests for the significance of the foreign factors: 1988.1 to 1991.11 Test for exclusion restrictions Country
Chi-squared statistic
Contemporaneous relationships "31
"32
"41
"42
Argentina
14.981 (0.242)
0.091 (0.243)
Bolivia
16.167 (0.184)
-0.092 (0.170)
Brazil
23.224 (0.026)
-0.045 (0.011)
Chile
29.527 (0.003)
-0.031 (0.041)
Colombia"
31.548 (0.002)
-0.014 (0.157)
Ecuador
17.285 (0.139)
-0.230 (0.139)
Mexico
23.203 (0.026)
-0.136 (0.216)
Peru
26.058 (0.015)
0.121 (0.061)
0.15 (0.017)
0.022 (0.128)
0.203 n.a.
Uruguay
11.275 (0.505)
-0.042 (0.042)
0.197 (0.012)
-0.05 (0.153)
0.076 n.a.
Venezuela
9.342 (0.673)
-0.045 (0.266)
0.003 (0.054)
0.743 n.a.
-0.451 n.a. -0.5331 (0.045) 0.481 n.a. -0.246 (0.026) -0.048 n.a. 0.668 (0.082) -0.324 n.a.
-0.28 n.a.
"According to the Akaike and Schwarz criteria the optimal lag length was six months.
-0.225 (0.405)
-0.14 n.a.
-0.011 (0.030)
-0.041 n.a.
0.043 (0.327)
0.323 n.a.
-0.018 (0.152)
0.545 n.a.
0.009 (0.176)
0.024 n.a.
-0.07 (0.376)
1.359 n.a.
-0.056 n.a.
-0.063 (0.627)
376
Calvo, Leiderman, and Reinhart
5
Policy implications
The empirical analysis suggests that external factors have played a role in recent developments in Latin America. These capital flows, in turn, have contributed to the accumulation of reserves and appreciation of the real exchange rate.44 With these stylized facts as background, and taking into account the possibility that external factors may reverse their course in the future, the next key issue concerns the form and timing of the appropriate policy response. Given that the 1980s has been a period of capital shortage for Latin America, the first question in discussing policy responses is: What is the rationale for policy to interfere with present capital inflows? Several countries in the region are in the process of concluding successful negotiations with their creditors, and effectively coming to grips with their fiscal imbalances. Thus, why would capital inflows - which in countries like Chile and Mexico have financed larger private investment - be undesirable? There are at least three types of concerns that policy makers tend to voice about capital inflows: (1) Since capital inflows are typically associated with real exchange rate appreciation and with increased exchange rate volatility, it is feared these may adversely effect the export sector; (2) capital inflows - particularly when massive - may not be properly intermediated and, therefore, may lead to a misallocation of resources; (3) capital inflows - especially when of a "hot money" variety - could be reversed on short notice, possibly leading to a domestic financial crisis. These concerns are not new. Actually, it has been argued that the depth of the debt crisis in the 1980s had a lot to do with the magnitude and sudden reversal of international capital flows. Consequently, these concerns have often led the authorities to react to the capital inflows by implementing a broad variety of policy measures. The remainder of this section examines the relative merits of some of those policies.45 We consider five intervention policies: (1) a tax on capital imports; (2) trade policy; (3) fiscal tightening; (4) central bank sterilized and nonsterilized intervention of capital inflows; (5) a rise in marginal reserve requirements on bank deposits and more regulated bank investments in equity and real-estate markets. Taxes on short-term borrowing abroad were imposed in some coun44
45
In terms of economic agents in Latin America, it is also possible to interpret these developments as originating in a portfolio shift away from foreign (dollar-denominated) and toward domestic financial and physical assets. For a model in which such a portfolio shift leads to a temporary appreciation of the real exchange rate and to accumulation of reserves by the central bank, see Calvo (1983). For a discussion of these issues from the perspective of Chilean monetary and exchange rate policies, see Zahler (1992).
Capital inflows to Latin America 377 tries - Israel in 1978 and Chile 1991. Although this policy is effective in the short run, experience suggests that the private sector is quick in finding ways to dodge those taxes through over- and underinvoicing of imports and exports and increased reliance on parallel financial and foreign exchange markets. Trade policy measures can help to insulate the export sector from real exchange rate appreciation. A possibility is higher export subsidies. However, this policy distorts resource allocation between exportables and importables and the fiscal cost could be substantial. For example, to offset a 20 percent overvaluation of the real exchange rate through export subsidies would increase fiscal expenditures by about 4 percent of GDP, given that the average export-GDP ratio for Latin America hovers around 20 percent. Alternatively, the authorities could increase both export subsidies and import tariffs in the same proportion - so as to avoid creating further relative discrepancies between internal to external terms of trade - and announce that those subsidies and/or tariffs will be phased out in the future. Indeed, if the private sector perceives these measures as transitory, agents are likely to substitute future for present expenditure, contributing to cool off the economy and to attenuate the real exchange rate appreciation. The fiscal cost of this package need not be large, particularly if the trade deficit is small. Furthermore, static distortions are not increased, since such trade policy does not change initial relative price distortions between exports and imports. However, this policy can be criticized on two different grounds. First, its effectiveness depends on the private sector believing that those subsidies and/or tariffs will be phased out in the future; otherwise, there is no reason for individuals to lower present expenditure. Thus, the effectiveness of the policy depends very strongly on credibility - both the credibility of policy, and the credibility of price forecasts. Second, this policy - as the previous one involving only subsidies - deviates from the general worldwide trend toward commercial opening and free-trade agreements. Another policy reaction to greater capital inflows could be to tighten fiscal policy - policy (3) - through higher taxes or lower government expenditure. While this policy is not likely to stop the capital inflow, it may lower aggregate demand and curb the inflationary impact of capital inflows.46 In that context, higher taxes may be less effective than lower government expenditure. Often when credit is widely available - as is the case when the country is subject to massive capital inflows - individuals' expenditures can be largely independent of their tax liability. This is espe46
In addition, to the extent that it reduces the government's need to issue debt, a tighter fiscal stance is also likely to lower domestic interest rates.
378
Calvo, Leiderman, and Reinhart
daily true if higher taxes are expected to be transitory - a somewhat plausible expectation since higher taxes would be associated with transitory capital inflows. In contrast, lower government expenditure - particularly when this expenditure is directed to the purchase of nontraded goods and services - has a direct impact on aggregate demand, which is unlikely to be offset by an expansion of private sector demand. However, contraction of government expenditure is always a sensitive political issue. Overall, it is hard to provide a strong case for adjusting fiscal policy - which is usually set on the basis of medium- or long-term considerations - in response to short-term fluctuations in international capital flows. However, if the authorities had envisioned a tightening of the fiscal stance, the presence of capital inflow may call for earlier action in this respect. Sterilized intervention has been the most popular policy response to the present episode of capital inflows in Latin America. Leading examples of this policy are provided by Chile in 1990-91 and Colombia in 1991. Under capital inflows, this type of intervention amounts to a central bank sale of government bonds in exchange for foreign currencies and securities.47 This policy does not necessarily stop private agents from engaging in international loan transactions. However, if successful, it insulates the stock of domestic money from variations associated with capital mobility. If effective, sterilization will tend to increase domestic nominal and real interest rates, lower aggregate demand, and mitigate the appreciation in the real exchange rate.48 There are, however, two main difficulties with sterilized intervention. First, sterilization leads to an increase in the differential between the interest rate on domestic government debt and international reserves, thus creating a fiscal (or quasi-fiscal) deficit. Second, by preventing a fall in the domestic-foreign interest rate differential, sterilization tends to perpetuate the capital inflow thus exacerbating any problems caused by this inflow. The impact of sterilization on the interest differential can be seen in Figure 13.4, by comparing sterilizing cases, such as Chile and Colombia, against the nonsterilizing case of Argentina. It is seen that in the current capital-inflows episode, the domestic interest rate exhibits a much smaller decline (or an actual increase) in sterilizing than 47 48
For a more detailed discussion of the role of central bank (sterilized and nonsterilized) intervention, see Mussa (1981) and Obstfeld (1991). A necessary condition for these outcomes, and for the effectiveness of sterilized intervention, is that domestic and foreign bonds are imperfect substitutes in agents' portfolios. Casual observation suggests that this seems to be the case in Latin America. Cumby and Obstfeld (1983) produced econometric results for Mexico in the 1970s in support of imperfect substitutability between peso-denominated assets and foreign assets. For industrial countries, Obstfeld (1991) concludes that sterilized intervention is a weak instrument of exchange rate policy, and that monetary and fiscal policies, and not intervention per se, have been the main policy determinants of exchange rates in recent years.
Capital inflows to Latin America
379
in nonsterilizing countries. The evidence from the recent experience of Chile and Colombia indicates that sterilized intervention has not reduced capital inflows. Yet, the increase in the fiscal deficit may be quite substantial; for example, Rodriguez (1991) estimates the fiscal burden of sterilized intervention in Colombia during 1991 at about 0.5 percent of GDP. Consequently, serious doubts can be cast on the desirability of sterilized intervention in the cases where countries are still attempting to reduce domestic debt, and their public sector budgets require further trimming.49 Alternatively, the central bank could opt for nonsterilized intervention, whereby the central bank purchases the foreign exchange brought in by the capital inflow in exchange for domestic money - as, for example, under a fixed exchange rate. This policy can help avoid nominal exchange rate appreciation, and is likely to narrow the domestic-foreign interest rate differential; however, it is likely to generate an increase in the domestic monetary base beyond the central bank's target. The latter, in turn, could fuel inflationary pressures and contribute to the real exchange rate appreciation. It is at this point that credibility considerations about maintaining a fixed nominal exchange rate come into play. In this connection, floating exchange rates have an edge, because the required real exchange rate appreciation does not necessarily call for inflation to accelerate. Furthermore, floating rates allow the domestic central bank to operate as a "lender of last resort." In contrast, under fixed rates and fractional-reserve banking, preventing liquidity-type financial crises - particularly, when capital starts flowing out - may call for the central bank to hold a large stock of international reserves, a costly if not unfeasible undertaking.50 Therefore, these credibility-related considerations give some support to a regime of floating exchange rates when the economy is subject to substantial capital flows.51 As discussed earlier, attempting to insulate the banking system from short-term capital flows is an attractive goal in cases where most of the inflows take the form of increased short-term bank deposits. In these circumstances, a sudden reversal of capital inflows may quickly result in bank failures. Under policy (5), marginal reserve requirements could be sharply raised such that they become higher as the maturity of deposits shortens; 49
50
51
See also Calvo (1991), which provides an example in which social welfare always declines with sterilization, and in which the effectiveness of sterilization relies on its worsening the credibility of an undergoing stabilization program. The problem is exacerbated when, as in most Latin American countries, the liabilities of the banking system are heavily biased toward short-term deposits, enhancing the chances of a run against the domestic banking system. When the system is not subject to big swings of international capital, the opposite conclusion can be reached: Fixed rates may dominate. See Calvo and Vegh (1992).
380
Calvo, Leiderman, and Reinhart
in fact, a 100 percent required-reserve ratio could be imposed on deposits with the shortest maturity. Although this scheme would impose a burden on the banking system, and could result in some disintermediation of the capital inflows, it has the advantage of decreasing banks' exposure to the risks of capital flow reversals. In addition, regulation that limits the exposure of banks to the volatility in equity and real-estate markets would further insulate the banking system from the bubbles associated with sizable capital inflows. To summarize, there are grounds to support a policy intervention mix based on the imposition of a tax on short-term capital imports, on enhancing the flexibility of exchange rates, and on raising marginal reserve requirements on short-term bank deposits. Given the likelyfiscalcosts it is hard to make a strong case in favor of sterilized intervention, unless countries exhibit a strong fiscal stance, and capital inflows are expected to be short-lived. In any case, we believe that none of these policies will drastically change the behavior of real exchange rates or interest rates for an extended period of time. The choice of appropriate policies, however, could decidedly attenuate the detrimental effects of sudden and substantial future capital outflows. References Bernanke, B. S. (1986). "Alternative Explanations of the Money-Income Correlation." Carnegie-Rochester Conference Series on Public Policy 2 (autumn): 49-100. Blanchard, O. J. (1989). "A Traditional Interpretation of Macroeconomic Fluctuations." American Economic Review 79: 1146-64. Calvo, G. A. (1983). "Trying to Stabilize: Some Theoretical Reflections Based on the Case of Argentina." In P. Aspe Armella, R. Dornbusch, and M. Obstfeld (eds.), Financial Policies and the World Capital Market: The Problem of Latin American Countries. Chicago: University of Chicago Press for NBER, 199-220. (1989). "Incredible Reforms." In G. A. Calvo et al. (eds.), Debt, Stabilization and Development. Cambridge, Mass.: Basil Blackwell. (1991). "The Perils of Sterilization." IMF Staff Papers 38 (March): 921-26. Calvo, G. A., and C. A. Vegh. (1992). "Currency Substitution in Developing Countries: An Introduction." Revista de Andlisis Economico 7: 3-27. Calvo, G , L. Leiderman, and C. Reinhart. (1993). "Capital Inflows to Latin America: The Role of External Factors." IMF Staff Papers 40: 108-51. (forthcoming). "Capital Inflows to Latin America: The 1970's and the 1990's." IMF Working Paper no. 85. Forthcoming in E. Bacha (ed.), Development, Trade, and the Environment. London: Macmillan. Collyns, C , J. Clark, R. Feldman, A. Mansur, P. Mylonas, K. Parker, R. Renhack, P. Szymczak, and L. Pauly. (1992). Private Market Financing for Developing Countries. World Economic and Financial Survey Series. Washington, D.C.: International Monetary Fund, December.
Capital inflows to Latin America
381
Cumby, R. E., and M. Obstfeld. (1983). "Capital Mobility and the Scope for Sterilization: Mexico in the 1970s." In P. Aspe Armella, R. Dornbusch, and M. Obstfeld (eds.), Financial Policies and the World Capital Market: The Problem of Latin American Countries. Chicago: University of Chicago Press for NBER, 245-76. Dhrymes, P. J. (1970). Econometrics: Statistical Foundations and Applications. New York: Springer-Verlag, 42-83. Diaz-Alejandro, C. F. (1983). "Stories of the 1930s for the 1980s." In P. Aspe Armella, R. Dornbusch, and M. Obstfeld (eds.), Financial Policies and the World Capital Market: The Problem of Latin American Countries. Chicago: University of Chicago Press for NBER, 5-35. (1984). "Latin American Debt: I Don't Think We Are in Kansas Anymore." Brookings Papers on Economic Activity 2: 335-89. Edwards, S. (1989). Real Exchange Rates, Devaluation, and Adjustment: Exchange Rate Policy in Developing Countries. Cambridge, Mass.: MIT Press, 87-160. El-Erian, M. A. (1992). "Restoration of Access to Voluntary Capital Market Financing." IMF Staff Papers 39: 175-94. Engle, R., and J. Issler. (1992). "Common Trends and Common Cycles in Latin America." Mimeograph copy. International Monetary Fund. (1991). "Determinants and Systemic Consequences of International Capital Flows." IMF Occasional Paper no. 77, March. Washington, D.C.: International Monetary Fund. Information Notice System (Database). Washington, D C . International Financial Statistics. Washington, D.C. Various issues. World Economic Outlook. A survey by the Staff of the International Monetary Fund. Washington, D.C. Various issues. Kuczynski, P. P. (1992). "International Capital Flows into Latin America: What Is the Promise?" Presented at the World Bank Annual Conference on Development Economics, Washington, D.C, April 30. LDC Debt Report. New York: American Banker. Various issues. Lutkepohl, H. (1985). "Comparison of Criteria for Estimating the Order of a Vector Autoregressive Process." Journal of Time Series Analysis 6: 35-52. Mathieson, D. J., and L. Rojas-Suarez. (1993). "Liberalization of the Capital Account: Experiences and Issues." IMF Occasional Paper no. 103, March. Washington, D.C: International Monetary Fund. Mussa, M. (1981). The Role of Official Intervention. Occasional Paper no. 6. New York: Group of 30. Reprinted in L. Melamed (ed.), The Merits of Flexible Exchange Rates: An Anthology. Fairfax, Va.: George Mason University Press, 1988, 78-93. Obstfeld, M. (1991). "The Effectiveness of Foreign-Exchange Intervention: Recent Experience 1985-1988." Working Paper, Harvard University. Pfeffermann, Guy, and Andrea Madarassy. (1993). "Trends in Private Investment in Developing Countries 1993: Statistics for 1970-91." Discussion Paper no. 16. Washington, D.C: International Finance Corporation. Reisen, H. (1993). "The Impossible Trinity in South-East Asia." International Economic Insights. Washington, D.C: Institute for International Economics. Rodriguez, C (1991). "Situacion Monetaria y Cambiaria en Colombia." Unpublished manuscript. Buenos Aires: CEMA, November. Salomon Brothers. (1992). Private Capital Flows to Latin America. New York: Sovereign Assessment Group, February.
382
Calvo, Leiderman, and Reinhart
Stiglitz, J. E. (1990). "Symposium on Bubbles." Journal of Economic Perspectives 4: 13-18. Stock, J. H., and M. W. Watson. (1989). "New Indexes of Coincident and Leading Economic Indicators." NBER Macroeconomics Annual 1989. Cambridge, Mass.: MIT Press, 351-93. Swoboda, A. K. (1983). "Exchange Rate Regimes and European-U.S. Policy Interdependence." IMF Staff Papers 30 (March): 75-102. United States Bureau of Economic Analysis. Survey of Current Business. Washington, D.C.: Various issues. Watson, M. W, and R. F. Engle. (1983). "Alternative Algorithms for the Estimation of Dynamic Factor, Mimic and Varying Coefficient Regression Models." Journal of Econometrics 23: 385-400. Zahler, M. R. (1992). "Politica Monetaria en un Contexto de Apertura de Cuenta de Capitales." Boletin Mensual (Banco Central de Chile) 771 (May): 1169-80.
CHAPTER 14
Opening the capital account: costs, benefits, and sequencing James A. Hanson
1
Introduction
The capital account has been opened, de facto, in the past twenty years by the increase in trade, the internationalization of production, the improvements in communications, and the legalization of foreign currency instruments in a growing number of countries. In line with this de facto opening of the capital account, and the greater reliance on open goods markets, developing country governments naturally are raising questions about fully opening the capital account in a de jure sense. As a background to answering such questions, this chapter surveys the costs and benefits of opening up domestic capital markets and discusses the preconditions to capital account opening and issues of sequencing and pace, including a reexamination of the cases of Uruguay and Chile, which opened their capital accounts fairly early. For purposes of the chapter, a fully liberalized or open capital market is defined as one in which individuals and firms access international financial markets freely, not just one in which the government intermediates capital flows to balance differences in private saving and investment. 2
The costs and benefits of capital account liberalization
2.1
The traditional analysis
The traditional welfare analysis of capital account liberalization is a real theory that focuses on the benefits of allowing foreigners to own more domestic capital (MacDougall 1968). This analysis begins from a situation of autarchy, analogous to the traditional analysis of the welfare implications of free trade, but with only one good. In autarchy, the rate of return The opinions expressed in this chapter are those of the author alone, acting in a personal capacity, and should not be taken to reflect those of the World Bank or its shareholders. 383
384
James Hanson
in the domestic market is assumed to be higher than in the rest of the world. Once the capital account is opened up, this differential generates a capital inflow and a larger capital stock in the home country. In the final equilibrium, the rate of return falls and GDP is higher, because of the larger capital stock. Labor gains at the expense of both domestic and foreign owners of capital, so GNP also is higher. A similar analysis applies in the growth theory context - foreigners will support a higher level of capital in country at lower cost in terms of foregone consumption than domestic saving alone. (See Hanson 1992.) This analysis ignores the substitutability that exists between an open current account and an open capital account. As Samuelson (1948) and Mundell (1968a) point out, commodity trade is a substitute for factor movements. Within the standard Hecksher-Ohlin two-good world, opening the current account lowers the rental rate on capital - and the interest rate (see Samuelson 1965) - in the country with more abundant capital, while raising it in the other. Capital account liberalization and current account liberalization thus are alternatives in the standard trade models. These models imply that the capital account need not be opened to equalize rates of return if the trade account is fully open. Of course the process of equalizing factor returns takes time. Even in the United States, where internal capital flows were free, there were significant, persistent per capita income differentials (see, e.g., Borts and Stein 1964), probably associated with differential returns on real capital and differences in technology. This suggests the desirability of opening the capital account sufficiently to obtain direct foreign investment, which would bring with it technology that cannot be transferred effectively through licenses or trade.1 According to the traditional analysis, financial flows would raise welfare only to the extent that they lead to a higher capital stock. A balanced capital account would provide no benefits. Moreover, since foreign saving, in theory, could be intermediated by the government's borrowing externally, the analysis has no implication for opening the capital account to all participants. 2.2
Risk
The traditional, "real" arguments for opening the capital account have little to do with modern theories of finance. More recent analysis of the capital account switches the emphasis to the role of risk bearing and risk sharing in financial markets in international financial markets. Following the well-known Arrow-Debreu model, if the prices for bearing certain risks differ across countries, then there would be gains from trade in inter1
See Moran 1974 for a discussion of a theory of direct foreign investment along these lines.
Opening the capital account
385
national financial assets embodying these risks that are analogous to the gains from international trade in commodities (Svensson 1988, Persson and Svensson 1985). Conversely, "All of the arguments against restricting international trade in goods also apply to restricting international trade in financial assets, whether these restrictions occur in the forms of direct controls, taxes or regulation of financial intermediaries" (Stockman 1988, 536). Of course, all the caveats regarding the benefits of free trade in goods also would apply to the benefits of free trade in risky assets. In addition, there are caveats related to default and bankruptcy in the intertemporal exchange of risky claims. Bringing in the element of risk is a major shift from the traditional analysis of capital flows. The recognition that international trades in intertemporal claims are motivated by considerations of risk delinks the welfare implications of an open capital account from its effect on investment. Even if saving and investment are unaffected by allowing capital flows, that is, even if the private capital account were exactly balanced by inflows and outflows of capital, the individual agents of the economy would benefit from trade in risky assets. This is, of course, a major difference from the traditional theory. Moreover, the argument that individuals should be allowed to trade assets internationally, based on differences in preferences, production, and evaluation of risk, is, analogous to the trading of commodities, perhaps the strongest argument for open capital markets in the sense defined in the introduction. 2.3
Taxation, and the risk of taxation
Developing country analysts, when asked to rationalize a closed capital account, often answer with some variant of "otherwise we would lose all our savings." In terms of the traditional analysis, this view suggests that capital in the country would fall if the capital account were opened. If this view were true, then it would suggest that the rate of return in the economy is less than the rate of return in the rest of the world. However, developing countries generally are thought to have lower levels of capital per worker than developed countries. This should imply higher rates of return in developing countries in the absence of capital movements. Underlying country risk, unrelated to country policy, is one possible explanation for outflows of capital from developing countries, despite domestic rates of return that exceed world rates. For example, borrowers from a country that is a mono-exporter would face a risk premium in international markets related to the price risk on the mono-export. If the capital account were opened, then residents would try to diversify their assets into instruments with different risks. However, an open capital ac-
386
James Hanson
count might also attract investors interested in diversifying their portfolios by purchasing assets with a different risk-return trade-off than prevails in their own countries. Gross capital outflows would certainly occur, but the capital account balance would depend on the net of the inflows and the outflows, and the extent to which the mono-exporter developed attractive instruments for foreign investors. Taxation and the risk of taxation probably are a more important explanation for the apparent paradox of capital outflow from what should be high return countries than underlying country risk. To investors, high taxes and potential taxes on capital in developing countries, including the possibility of expropriation, easily could offset underlying riskless rates of return to capital and to financial assets that are higher than international rates. The tax issue is compounded by the tax preferences offered foreigners in many industrial countries and banking centers. Offshore banking centers, and many industrial countries, do not tax interest earned on savings accounts or, at least, interest earned by foreign holders of savings accounts. For example, the United States has not taxed interest income on savings accounts belonging to foreigners since the 1970s. Moreover, competition for deposits among financial centers has meant that data on foreign owners of deposits usually is not available to income tax authorities in the country of origin - if one center were to provide the information, then tax considerations would cause capital to shift to other centers. Such differential tax treatment of foreign and domestically owned deposits could, in theory, given rise to two-way capital flows based on tax avoidance, so-called round-tripping (Tanzi and Blejer 1982). Taxation of financial instruments takes many forms beside simple income taxation. Of particular importance in developing countries is the socalled inflation tax. Inflation represents a tax on all financial instruments that have zero or interest rates fixed below market levels. Issuers of these instruments offer them in the market to buy goods and services, bidding up prices; the holders of the existing stock of such instruments find themselves forced to trade goods and services for additional assets in order to maintain the real value of their asset stocks that is desired to carry out transactions or to maintain the purchasing power of their wealth - holders of these instruments are thus effectively taxed.2 Developing country governments are the main beneficiaries of the in2
Reserve requirements that are unremunerated or carry fixed remunerations are another way of imposing the inflation tax. If banks were allowed to issue deposits freely, then they would compete away the government's monopoly. Hence governments in inflationary economies typically impose large, unremunerated reserve requirements on financial intermediaries during times of inflation. This limits competition and, in effect, subjects those depos-
Opening the capital account
387
flation tax. Government liabilities bearing zero or low interest rates such as currency and reserve requirements account for much more of the financial system in developing countries than in industrialized countries (see Table 14.1 for some examples); and developing countries tend to issue these liabilities at a faster rate than developed countries. This in turn has led to higher average inflation rates in developing countries, particularly in Latin America, than in industrialized countries. For example, in 1990 consumer price inflation in the industrialized countries was 5 percent, whereas in the developing countries it was 99 percent; even among the developing countries of Asia, where average inflation is the lowest of the developing country regions, inflation averaged nearly 7 percent (IMF, International Financial Statistics, 1992). The amounts raised through this tax are relatively large, as shown in Table 14.1. Developing country governments resort to the inflation tax, despite its well-known defects, because it is easy to apply - it applies to a broad base, including agriculture, which is notoriously difficult to tax, and does not depend on a dedicated tax administration. It also can be imposed administratively, and without the need to seek legislative approval. The ability to raise these "tax revenues" is dependent on the degree to which the government can make the assets they issue acceptable - the lack of competition for "legal tender." The government's monopoly power to issue legal tender would be decreased, to the extent the capital account is open. If the capital account were open, and the government resorted to the inflation tax, then the public would begin to use foreign exchange denominated assets - currency substitution. The base of the inflation tax the real stock of non-interest-bearing assets denominated in domestic currency - correspondingly would shrink, with a corresponding decline in the ability to issue new liabilities or implying a higher inflation rate (and a smaller inflation tax) for the same issue of new liabilities. To illustrate, suppose, for example, that domestic and foreign currency its that are created to the inflation tax, by forcing a large percentage of them to be deposited, without remuneration, in the central bank. The incidence of the implicit tax from reserve requirements may fall on depositors or borrowers. This means that, with an open capital account, an increase in the inflation tax on reserve requirements may motivate capital inflows as well as outflows. For example, with an open capital account, depositors would be able to demand deposit rates equal to international rates. This would mean that higher-than-international costs of reserve requirements would affect only lending rates, where differences in information reduce competition between foreign and domestic banks more than in deposit taking. The higher lending rates, in turn, would induce foreign banks to increase their direct lending from home offices, thereby generating a capital inflow. Firms that could borrow externally would do so, increasing capital inflows. Obviously, a less than perfectly open capital account could generate outflows by depositors and inflows to borrowers.
388
James Hanson
Table 14.1. Estimated inflation tax in selected developing countries (percent per year or percent of GDP) Argentina (1988)
Chile (1990)
Indonesia (1989)
Korea (1990)
354.0
22.4
10.0
10.8
2.9
24.5
2. Reserve money* (%ofGDP)
2.5
3.0
5.9
7.3
13.7
4.1
3. Inflation taxc (%ofGDP)
9.0
0.7
0.6
0.8
0.4
1.0
4. Seigniorage^ (%ofGDP)
4.3
0.8
1.4
0.7
2.9
0.4
1. Inflation" (% per year)
Malaysia (1989)
Mexico (1989)
fl
GDP Deflator. ^Geometric average, year-end; for Argentina reflects only currency outside deposit money banks, as interest was paid on reserves. c Line 1 times line 2. ''Increase in reserve money over the year/GDP.
were legally permitted to circulate and be used in settlement of contracts (labor and taxes as well as financial). It would not pay anyone to hold or use domestic currency if prices, measured in domestic currency, rose faster than international prices - the base of the inflation tax would fall to zero unless the government restrained its issue of domestic currency to levels that were demanded.3 A less extreme example is the case in which domestic and foreign currency can be used legally to settle some but not all contracts; then the base of the inflation tax falls to levels that depend on the transactions' cost of switching between the two currencies and the frequency and cost, net of interest, of the required transactions.4 A still less extreme example is the case in which residents are allowed to hold deposits, but not legally allowed to settle contracts in foreign currency. Here In the case of Panama, U.S. dollars and coins circulate freely and the government issues coins in local currency; the government is restrained in its issue of such coins by the need to exchange them for U.S. dollars. In Argentina, the 1991 Convertibility Law restrained the issue of local currency by requiring that the Central Bank's holdings of international reserves be at least equal to the money base, in effect requiring demand for money to manifest itself by importing capital. For example, this was the case in Argentina between 1989 and 1990, when accounts denominated in foreign currency were freely available but wages and taxes had to be settled in local currency.
Opening the capital account
389
there would be still more demand for local currency than in the previous case. The risks of potential taxation, as well as actual differences in taxes, are a rationale for "capital flight." The potential for higher future taxation means that domestic financial assets must carry a risk premium to be attractive, in the minds of potential asset holders, to assets in the rest of the world.5 There also exists a risk that a country may default on its international obligations, over and above the risk associated with individual projects (see Dooley and Isard 1980, Hanson 1974). Thus there may be a number of risk premiums, in addition to the risk premiums associated with the various forms of taxation, and differences in the variability in inflation, the exchange rate, and domestic interest rates. These risk premiums, and changes in them, are one explanation of the imperfect substitutability between the financial assets of different countries. They also explain why interest rates may be higher than international rates in countries that are presumed to have low capital-labor ratios. In the traditional model, such risk premiums have a real cost, in terms of reducing the domestic capital-labor ratio and domestic consumption below what it otherwise would be. Hence, policies that can reduce this risk premium would tend to raise GNP. To summarize, "country policy risk" explains why an open capital account might lead to capital outflows from what are potentially high return countries. Capital account controls are needed in order to tax domestic capital at higher-than-international rates. To the extent that the capital account can be kept closed, preventing capital outflow that otherwise would occur because of higher rates of actual and potential taxation, and to the extent that legal arrangements (e.g., government control of the flows) can be made to offset the potential negative impact of these capital controls on capital inflows, then the negative effects of capital account controls can be partially offset. In these circumstance, there may, indeed, be costs of opening the capital account. This is particularly true taking into account the distributional implications of being able to tax capital more heavily than would be possible with an open capital account. There are two critical questions here, related to whether the de facto internationalization of capital already makes it difficult to tax capital. First, capital controls and high taxation are likely to motivate the development of channels through which capital can be transferred internationally. The longer capital controls have been imposed, the more porous they are See, e.g., Dooley 1988. The existence of such risk premiums, and the possibility that foreigners will be treated more favorably than domestic assets holders, can lead to simultaneous borrowing abroad and capital outflows (Dooley 1988, and Khan and Haque 1985).
390
James Hanson
likely to be. To the extent that capital can be moved fairly freely internationally - for example, by overinvoicing imports and underinvoicing exports - then laws to keep the capital account closed simply represent a hindrance that encourages corruption and discriminates against lawabiding citizens. Maintenance of such ineffective laws is not a valid reason for not opening the capital account.6 Moreover, since capital controls are only partially effective, they change the incidence of taxation of capital. Some types of capital can be moved less easily internationally than others, some savers and investors have less access to international markets than others. These types of capital and these agents thus are more subject to capital taxation than those that are de facto internationally mobile. If capital that is de facto internationally mobile also is owned largely by large capitalists, then any favorable distributional implications of capital taxation cum capital controls may be reduced or even reversed. Indeed, one argument for an open capital account is that it allows all citizens, not just those with easy access to foreign exchange, to reduce the burden of taxes on savings, such as the inflation tax. 2.4
Open capital markets as limitations on the effectiveness of policies
The possibility of avoiding taxes when the capital account is open is one example of the general point that an open capital account limits the impact of government policies. An open capital account limits governments' ability to tax capital or financial assets, to the extent that economic agents can easily switch their portfolios internationally to escape taxes. Although in a closed economy the impact of taxes also can be reduced, by substitution in production and consumption into goods with lower rates of taxation, in an open economy even more options are available. Another well-known example of how an open capital account may reduce policy effectiveness is Mundell's (1968) analysis, showing that monetary policy is ineffective in a small economy with a fixed exchange rate, open capital markets, and perfect substitutability between domestic and foreign assets. In Mundell's model, the domestic interest rate is fixed by international flows of capital and cannot be affected by variations in the growth of domestic credit. Thus monetary policy becomes ineffective. These and other examples of policy ineffectiveness often are cited as costs 6
This is even more so to the extent that governments must make special legal arrangements to encourage certain capital inflows by eliminating the capital controls in particular cases. Since these arrangements typically would require government controls over the inflows, they also might encourage misallocation of resources and corruption. This is another cost of capital controls.
Opening the capital account
391
of opening the capital account. However, the importance or even the correctness of this argument is not clear. In general, an open capital account does not eliminate the effectiveness of policy instruments, it only reduces it. In variants of the Mundell model, price stickiness and wealth effects mean money policy does have a role in determining output, even if interest rates are determined by free international capital flows (Dornbusch 1976). In addition, imperfect substitutability between assets means monetary policy can affect the differential between foreign and domestic interest rates (the "risk premium") and thus change the domestic interest rate, even in a completely open capital market. Targeting the interest rate rather than the money stock is another way to take advantage of the imperfect substitutability between domestic and foreign assets, although it is likely to lead to a loss of international reserves. To lower domestic interest rates, a government may offer to lend below international interest rates or roll over government bonds at less than world rates. This will lead to capital flight and loss of international reserves. The policy also may decrease domestic saving, and thus investment.7 Raising the interest rate, by offering interest rates above international rates, may be more successful in affecting (reducing) investment. Surprisingly, however, this policy also is likely to lead to a decline in net international assets. For example, the central bank or the treasury could maintain domestic interest rates above international rates by selling highinterest-rate bonds (or sterilize capital inflows, which is the same as not letting capital inflows drive down the domestic interest rate). This policy will lead to higher capital inflows than otherwise would be the case. Domestic borrowers would face higher interest rates. However, the policy also will generate net obligations for the country, since the inflows of capital can only be invested at the lower, international rate. The country's net external obligations will grow, and eventually raise the risk premium in world markets. Thus, the eventual result of trying to target the interest rate below or above the world rate is a loss in net reserves. Since declining net reserves often are viewtd as an indicator of country risk, either policy is likely to lead to a rise in the interest rate premium facing the country in international markets and, eventually, to higher domestic interest rates. Finally, if only the degree of effectiveness of a policy variable - the impact multiplier in the language of econometrics - is changed when capital markets are opened up, then the ability to make policy is not affected. A Although a government undoubtedly can lower interest rates to certain preferred borrowers and to savers without access to international markets, the resulting reduction in the availability of savings will raise interest rates to the marginal borrower.
392 James Hanson larger dose of the same instrument will achieve the desired effect. If the application of a policy instrument has no cost, then changes in the impact multiplier are not important. However, if the cost of using a policy instrument does rise with the size of the intervention, then there may be some costs of capital account liberalization, in terms of policy effectiveness, but these must be set against the benefits of opening the capital account. One option to overcome any loss of policy effectiveness would be to use an alternative instrument to achieve or to enhance the effectiveness of a policy instrument. For example, again referring to the Mundell model, the use of a floating exchange rate, as opposed to a fixed exchange rate, restores the effectiveness of domestic monetary policy despite an open capital account.8 Thus, monetary independence can be restored at the cost of the government's giving up control over the nominal exchange rate and allowing it to float freely. The free float would increase the variability of the exchange rate and correspondingly imply greater costs for exporters and importers; under certain types of shocks, such as capital inflows, they would bear more of the costs of adjustment than under afixedrate system, where these costs are more spread out over the economy.9 Viewed in this light, capital account controls, to the degree to which they can be made effective, are an alternative tofloatingexchange rates in making monetary policy effective. The country effectively tries to use a free float (or a different exchange rate) for capital account transactions in effect trying to create another policy instrument - in order to obtain monetary independence while maintaining control over the nominal exchange rate applicable to trade. The difference between the multiple exchange rates, and thus the degree of monetary independence, clearly is limited by the ease of arbitrage between the markets and the effectiveness of the capital controls.10 It also should be noted that the effectiveness of some policies may be enhanced by an open capital account. For example, in Mundell's analysis, fiscal policy becomes more effective with an open capital account because 8
9
10
In Mundell's model of a small open economy with a floating exchange rate, a one-time increase in domestic credit depreciates the exchange rate, changing either national output or prices, but not necessarily the interest rate. In other versions of the model, with more complicated policy changes, the domestic interest rate also may change. Conversely, under other shocks, the floating exchange rate would cushion the shock for traders. See, e.g., Fleming 1971, Lanyi 1975, Gros 1988, and Dornbusch 1976. As Lanyi and Gros point out, if a permanent condition (shock) leads to a difference between the two exchange markets, then arbitrage will develop that eventually will eliminate the gap between the two markets or will require intervention in both markets—in effect a tax-subsidy arrangement—to maintain the gap.
Opening the capital account
393
11
there is less crowding out. Fiscal policy thus represents not only an alternative instrument to monetary policy for the control of short-run fluctuations in aggregate demand, but one that is enhanced by an open capital account. Any cost of a decreased effectiveness of monetary policy thus must be weighted against the benefit of an increased effectiveness of fiscal policy. A more fundamental question is whether the changes in the effectiveness of policy instruments represent a cost to society. If an open capital account permits residents to escape the inflation tax, or other distortionary taxes, at low cost, then an open capital account does not a priori represent a cost to society defined as the sum of individuals; it may even be a benefit. Some authors have extended this argument: An open capital account, by placing a limitation on the effectiveness of certain government policies, reduces the incentive for their enactment. Thus a country's credibility improves and the risk premium may decline. However, this argument should not be overstated. Even the most credulous investors recognize that "a government intent on extracting an inflation tax from its own residents . . . has substantial incentives to deviate from a regime of flexible exchange rates and capital mobility" (Sargent 1983, 103). An open capital market can easily be closed if the government wishes to engage in distortionary policies. Perhaps, the argument is better stated by noting that once a capital account is opened, and maintained open for some time, then the incentives to close it again are reduced, because many savers and investors will have diversified their portfolios internationally in the interim, and this will reduce the effectiveness of policy and thus the incentive to engage in such policies. In sum, it is difficult to conclude whether the impact of opening the capital account on the effectiveness of policy instruments represents a cost or a benefit to society. 2.5
Instability
Another common, related argument against open capital accounts is that they can lead to greater instability. Flows of "hot money" not only can offset monetary policy, as discussed in the previous section, they can cause substantial variations in the nominal and real exchange rate, interest rates, and output. An open capital account also can exaggerate or offset terms1
' Of course, a fully open capital account, as defined in this chapter, is not necessary to achieve this result—the government could simply borrow or lend internationally as needed to carry out the desired fiscal policy.
394 James Hanson of-trade shocks, depending on how international creditors react to such shocks. An open capital account could, in theory, cushion temporary fluctuations in tradables output and prices, to the extent it allowed countries to borrow or lend abroad. To some extent that may even have happened in the mid-1970s and at the beginning of the 1980s in some petroleum importing and exporting countries. Direct foreign investment also may act as a shock absorber, since it acts like equity rather than debt. For example, Reynolds (1965) shows that in Chile theflowof resources from the copper companies to the country were stabler than the terms of trade. Although contracyclical capital flows are possible, the traditional model and the empirical results suggest that capital inflows are likely to move sympathetically to a terms-of-trade shock and thus magnify its effect. This means that at best capitalflowscan ease adjustment to a permanent shock, not substitute for it. Moreover, the experience of a number of countries during the debt crisis suggests that the problem of importing world fluctuations through the capital account is only partly a matter of "hot money." The debt problems of the severely indebted, middle-income countries reflect a period in which private medium- and long-term inflows, as well as short-term inflows, first grew to unsustainable levels, and then were cut off at the same time as the terms of trade collapsed and real interest rates were high. The initial inflows reflected unsustainable and contradictory macroeconomic policies in the borrowing countries, but also rapid monetary growth in the industrialized countries and, in the case of oil exporters, favorable terms of trade. For the foreseeable future, such lending excesses are unlikely to recur. Nonetheless, the experience of this period suggests that the capitalflowsmay well magnify shocks. The validity of the instability argument is not, however, solely a question of whether capital flows move pro- or contracyclically to variations in international prices. Rather, it also depends on whether the main source of domesticfluctuationsis external or internal, much like the classic arguments for and against the flexible exchange rates. If the main source of fluctuations is variations in saving and investment rates in the rest of the world or variations in the rest of the world's evaluation of country risk, and if the economy is operating under afixedexchange rate regime, then an open capital account might increase fluctuations (leaving aside the shock absorber argument) because these shocks would enter through the capital as well as the current account. However, if the main source of fluctuations is variations in the balance between domestic saving and investment, then international capitalflowscould reducefluctuationsin output. Also, an open capital account makes it difficult to apply the inflation tax or conduct an independent monetary policy, as noted in the preceding
Opening the capital account
395
section. Moreover, if a government had a tendency toward the erratic use of the inflation tax, but could be convinced to keep an open capital account, then that capital account policy would induce a stabler policy framework by reducing the incentives to use the inflation tax. Thus, the key issue in deciding whether the loss of policy effectiveness is a cost is whether domestic fiscal and monetary policies tend to offset fluctuations in the domestic economy or cause them. 3
The sequencing and speed of capital and current account liberalization
3.1
The preconditions for capital account liberalization
The first precondition to international financial liberalization is a stable macroeconomy. Opening the capital account increases the opportunity for currency substitution, which increases the rate of inflation (in domestic currency) needed to mobilize a given volume of resources through the inflation tax. However, it is important to realize, as noted earlier, that under high and variable inflation the capital account may be substantially open for many citizens. In that case, leaving the capital market legally closed increases the burden of the inflation tax on those without easy access to foreign exchange. Allowing easier access to international capital markets during conditions of high inflation can increase the variability of the economy substantially. In high inflation, the domestic currency money base falls as a percentage of GDP. Every nominal and real shock, and every shift in expectations, then has a proportionately greater impact on domestic financial variables. The resulting variability will be further magnified if shifts into foreign exchange denominated assets are made less costly by capital account liberalization - the money base will become even smaller and the shifts in response to a given shock will become even greater. Although a "stable macroeconomy" cannot be defined exactly, it seems clear that a sustainable fiscal deficit, which requires only a minimal inflation tax, is a particularly important precondition to international, as well as domestic financial liberalization. Of course, external borrowing, directly or indirectly,12 could be used to reduce the need for the inflation tax and the public sector's "crowding out" of domestic investment. (Notice that such increased external public sector borrowing does not mean international financial liberalization as defined in the introduction to this es12
The public sector can borrow abroad directly, or allow domestic banks and firms to borrow externally (perhaps with a government guarantee), either way reducing the pressure on the domestic financial system from a given volume of public sector borrowing.
396
James Hanson
say.) However, this policy option is not a long run solution to an unsustainable public sector deficit. Indeed, government access to foreign borrowing may even slow adjustment, as McKinnon pointed out, and as recent experience suggests. In the late 1970s and early 1980s, many developing countries used substantial external borrowing to avoid reductions in public sector deficits and to sustain overvalued real exchange rates that were inconsistent with aggregate demand policy. When the initial inflow of funds slowed, real interest rates increased, and commodity prices fell, the debt crisis occurred, which depressed growth rates for many years in the more severely indebted countries. It is sometimes argued that opening the capital account will force the government to reduce its reliance on the inflation tax, by lowering the base of the tax significantly. This argument is sometimes extended by saying that opening the capital account will endow anti-inflation programs with greater credibility. To some degree, these arguments are based on the idea that the inflation tax results from the government's rational reflection on the optimal combination of taxes. However, analyses of hyperinflation suggest that these episodes often reflect a desperate attempt to continue spending in excess of tax revenues, rather than a rational decision about the optimum combination of taxes. Moreover, as noted earlier, governments that intend to use the inflation tax typically close the capital account first. Overall, it would seem that merely opening the capital account will not increase the credibility of a stabilization program very much. Indeed, as noted, opening the capital account can increase the public sector's access to foreign loans, thereby reducing pressure for adjustment and building up external debt, while allowing private citizens to engage in capital flight. Hence, it seems unwise to try to use capital account liberalization as a means of strengthening a less-than-stable macroeconomy. The second precondition to international financial liberalization is a sound, liberalized domesticfinancialsystem. If domestic citizens are to be given easy access to international financial markets, through foreign banks, then the domestic banking system must face similar regulations and controls on interest rates-portfolio composition as foreign banks in their home markets to remain competitive. Moreover, the domestic banks must have sound portfolios. Unless the domestic banking system is ready to meet the competitive challenge from a liberalized capital market, then opening the capital account will lead to disintermediation away from it, induce illiquidity and insolvency in domestic banks, and increase claims on the government, in its role as explicit or implicit insurer of the banks' deposits. In a financially closed economy, the domestic banking system often suffers from severe regulation, which benefits the government and other
Opening the capital account 397 preferred borrowers at a cost to domestic savers and nonpreferred borrowers. Perhaps the most important form of such regulation is the control of interest rates-portfolio composition, that manifests itself in high, nonremunerated reserves requirements and below-market, directed credit. Such regulations lead to large spreads between rates on deposits and rates on nonpreferential loans (Hanson and Rocha 1986). Once depositors and nonpreferred borrowers obtain access to international financial markets, these spreads make domesticfinancialintermediaries uncompetitive - because of the regulations, they are unable to set interest rates that will prevent the loss of deposits and their nonpreferential loan customers. Such regulations can be applied to foreign banks operating in the domestic market, but they are likely to be less effective than on domestic banks because foreign banks and associated nonbank intermediaries can often facilitate transactions with home offices that avoid such regulations or they can avoid such regulations within the country using modern technology.13 Hence, such regulations reduce the competitiveness of the domestic banking industry, and are likely to lead to a loss of business that may imperil some domestic banks. The domestic banking sector may be charging high spreads because of high costs and the need to provision against poor loan performance. High costs may reflect a number of factors in addition to poor technology and management, such as the sharing of monopolistic profits with members of bank workers and extensive branch networks that were useful for capturing low cost deposits in a period when interest rate competition was limited. Poor portfolio quality may reflect not only poor loan decisions and unexpected shifts in relative prices but high-risk, directed credits and favorable terms to non-arm's length borrowers. Competition from foreign banks and an open capital market is useful in reducing high spreads that result from poor technology and management, and extensive branch networks. Competition also may force better portfolio selection. However, it probably is desirable for the government to take action to reduce the domestic banks' costs associated with the stock of outstanding bad loans (by fully recognizing such loans) and with high labor costs before allowing foreign banks to compete fully with domestic banks. Otherwise the playing field will not be level for domestic banks and some of them could be imperiled, for example by the burden of poor portfolio quality, eventually leading to the government's being forced to assume responsibility for an even greater volume of deposits. 13
For example, reserve requirements must be computed on some deposit base. Computers can be used to "sweep" large depositors just before the deposit base is calculated and place the funds in accounts with lower reserve requirements—higher interest rates.
398
James Hanson
3.2
The sequence of liberalization: capital or current account
A substantial literature focuses on the question of sequencing external liberalization - whether the capital or current account should be opened first. Interest in this question was aroused by the different paths of liberalization in the Southern Cone of South America at the end of the 1970s. Chile is generally characterized as liberalizing the current account first, Uruguay and Argentina the capital account first. In an influential essay, McKinnon (1982) argues that much of Chile's success, relative to Argentina, was due to the different sequencing of their reforms (and to Chile's fiscal surplus as opposed to Argentina's fiscal deficit). It should be noted, however, that there is a substantial debate on the timing of Chile's capital account liberalization and its role in the crisis of the early 1980s. Of course, a number of other factors were at work, and eventually, all three countries' attempts at liberalization suffered reverses in the debt crisis (see, e.g., Corbo and de Melo 1985, Corbo, de Melo, and Tybout 1986, Edwards and Cox-Edwards 1987, Hanson 1986, and Harberger 1982). The cases of Uruguay and Chile are discussed in the next section. The initial view on sequencing (McKinnon 1973) was that capital account restrictions should be relaxed only after trade reform, and policy makers should limit any unusual influx of foreign capital. The rationale is as follows: Current account liberalization typically requires a real depreciation of the exchange rate, to offset the negative effect on the balance of payments of cuts in protection. In contrast, capital account liberalization tends to produce a real appreciation of the exchange rate. If capital account liberalization were to produce a stable net transfer (capital inflow, less interest payments), then opening both accounts simultaneously might yield something like the final real exchange rate and resources could then be reallocated in accordance with that objective (Edwards 1989). However, capital inflows must be paid for, so the initial net transfer following the opening of the capital account is likely to be larger than the final net transfer. Moreover, the responsiveness of capital to the pent-up demand for local assets is likely to be much faster than the responsiveness of trade flows to the opening of the current account, leading to an unsustainable appreciation of the real exchange rate (Frenkel 1982). Hence, opening the capital account before the current account produces incentives for resource allocation that will be reversed in the final equilibrium (Krueger 1986), and could even retard the adjustment to the reduction in protection. To avoid "unnecessary" resource shifts, a number of authors (e.g., Edwards 1984, Frenkel 1982, Khan and Zahler 1983, McKinnon 1973, 1982) have concluded that it is preferable to open the current account before the capital account.
Opening the capital account
399
There are, however, some difficulties with this argument. First, the volume of capital resources over time is not necessarily the same under the two different sequences of liberalization. Whenever the capital account liberalization occurs, it tends to cause an appreciation; the point is that once protection is reduced in a capital-scarce country, the rate of return is reduced, as discussed in Section 2, decreasing the incentive for capital inflows. Thus, there are less "unnecessary" resource shifts in the current account-capital account sequence in part because there is a smaller stock of capital - the costs of this lower capital stock have to be evaluated relative to the benefits of lower costs of adjustment. This type of argument has led some authors (Little, Scitovsky, and Scott 1970, Michaely 1986, Krueger 1981, 1984) to argue for simultaneous liberalization of the current and capital accounts. In their view, short-run adjustment costs and the opposition to reforms that they generate are an important problem in liberalization. Foreign funds can be used to reduce or offset the costs of these frictions. Thus, these authors generally argue that the capital account should be liberalized at the same time as or even prior to the current account. Of course, this argument is for larger capital inflows; these need not be provided through greater individual access to foreign capital (capital account liberalization) but could come through government intermediation of foreign capital. These points are examples of a more general questioning of much of the sequencing literature. A strict economic analysis of sequencing requires an intertemporal analysis, allowing for different volumes of capital inflow as well as different degrees of misallocation. Moreover, the analysis should not be limited only to distortions between domestic and foreign prices but should allow for domestic distortions as well. Some attempts to carry out this type of analysis have already begun (see, e.g., Khan and Zahler 1983, Rodrick 1987, Edwards 1989, and Edwards and van Wijnbergen 1986); not surprisingly, the argument for the current account-capital account sequence seems to depend on the type and degree of the initial distortions. Moreover, the whole sequencing argument understates the benefits of capital account liberalization because of its neglect of risk. 3.3
Interactions between the speed of current and capital account liberalization
The speed of current account liberalization also is an issue because of the positive relation that exists between expectations of devaluation and the domestic interest rate. In the limit of an open capital account, the domestic interest rate might fully reflect the expectations of devaluation - the well-known uncovered interest parity theorem (Fisher 1930). But even in
400
James Hanson
economies with nominally closed capital accounts, there appears to be some positive link.14 As just noted, a reduction of trade barriers typically needs to be compensated by aparipassu depreciation of the real exchange rate to avoid an increase in the current account deficit. Generally it is thought that this real depreciation can be achieved with lower costs of adjustment by depreciating the nominal exchange rate, rather than through a reduction in domestic wages and prices. However, depreciation in the nominal exchange rate tends to raise the real interest rate in the economy, because of the increased gains to be made from holding financial assets denominated in foreign exchange. The higher real interest rate tends to depress the rate of investment. This argument suggests that the pace of trade liberalization and the rate of compensatory depreciation could be important determinants of the rate of investment and growth, even in a fully credible process of liberalization. The relationship is not linear - very slow trade liberalization requires very little compensatory real depreciation and hence has only minimal effect on the real interest rate and investment. Very rapid trade liberalization means a high real interest rate only for a short period; thereafter, no real devaluation is needed and the expected rate of devaluation and, correspondingly, the real interest rate decrease. It is only gradual trade liberalization, spread over say three to five years, that is likely to have an important effect on the real interest rate and investment. Even in such programs, the effect could be limited, by initially overshooting the exchange rate necessary to maintain a constant trade balance. An example may clarify these points. Consider a program of trade liberalization intended to reduce the average rate of protection of imports by 50 percent. If done over five years at a steady rate, this might require a real depreciation of 3-5 percent per annum to match the growth of imports to the higher growth of exports. However, this depreciation would require an increase of 3-5 percentage points in the real interest rate in order to maintain the attractiveness of financial assets denominated in domestic currency, relative to those denominated in foreign currency. This higher real interest rate would have a negative effect on investment. If the trade liberalization were spread over ten years, the required real depreciation and corresponding increase in the domestic real interest rate would be small; if the full trade liberalization were done at once, or if the exchange rate 14
See, e.g., Edwards and Khan 1985. They show, not surprisingly, that in the relatively open economy of Singapore the interest rate is almost wholly determined by international considerations, including the expectation of devaluation. More surprisingly, they show that in Colombia, which had a legally closed capital account in the period studied, the world interest rate, adjusted for devaluation, was highly significant in the determination of the domestic interest rate. See also World Bank 1984.
Opening the capital account
401
were devalued sharply early in the program of protection reduction, then there would be a much smaller effect on the real interest rate. Clearly the importance of this effect varies from country to country, depending on the degree to which the rate of depreciation affects the domestic interest rate and the effect of the domestic interest rate on investment. However, this argument suggests that it should not be surprising that programs of gradual trade liberalization show little impact on the rate of growth - such programs tend to depress investment by raising the real interest rate and thus reduce the adjustment to the new incentive framework. 4
Liberalizing the capital account: experiences in Latin America
4.1
Background
The theoretical literature on sequencing typically treats capital account liberalization as instantaneous, in contrast to the well-known debate on gradual versus rapid trade reform. (See, e.g., Krueger 1986, Michael 1986, Little, Skitovsky, and Scott 1970, Edwards and van Wijnbergen 1986.) In fact, the capital account may be opened gradually in a variety of ways, analogously to the methods used in opening the current account gradually. For example, taxes on capital inflow (including differential reserve requirements on banks' use of foreign capital) can be varied over time, or limitations on various financial intermediaries' use of foreign capital can be varied over time. Given its assumption of instantaneous liberalization, the theoretical literature provides little guidance on the costs and benefits of these different sequences. Hence, it may be useful to examine the results in practice in some countries. This chapter concentrates on the rich vein of Latin American experience, particularly Uruguay and Chile in the 1970s through the early 1990s. Generally speaking, Latin America had a relatively open capital account for the period between its independence and the Great Depression. External financing went much beyond trade credits: The first government bonds were floated in London in the 1820s, soon after independence (Dawson 1990); British and then US. bondholders financed railroads, public utilities, and mines (Jenks 1927, Rippy 1959, Stallings 1987). In the twentieth century, direct foreign investment financed oil wells in Mexico and Venezuela and copper mines in Chile (Reynolds 1965 and Moran 1974). Many countries had European and U.S. bank branches. In the 1930s and 1940s export prices and foreign investment collapsed during the Great Depression. As in some industrial countries, exchange controls were instituted, requiring surrender of foreign exchange receipts and allocating foreign exchange to certain users - mainly importers as
402
James Hanson
Table 14.2. Average inflation in Western Hemisphere developing countries (decade averages, percent per year)
Average inflation (Standard deviation)
1950s*
1960s
1970s
1980s
11.8 n.a.
18.0 (14.0)
32.3 (15.5)
131.8 (89.0)
a
1958-59; data on Brazil not shown for previous years. Source: International Financial Statistics, 1985, 1992.
debt service was suspended on bonds (see, e.g., F.I.E.L. 1989 on Argentina and Valdes-Prieto 1994 on Chile). During the period of import-substituting industrialization, roughly from the 1950s to the mid-1980s, capital account controls were relaxed and tightened intermittently, but the general trend was toward greater control. Typically, use of foreign exchange was restricted and it often became illegal to own foreign currency without authorization. Controls on foreign banks were tightened in parallel; in some cases (e.g., Argentina, Chile, Colombia, Peru) foreign banks sold off their offices. Although direct foreign investment often was allowed in the newly protected industries, a wave of nationalizations of traditional direct foreign investment in public utilities and raw materials began in the 1960s and crested in the late 1970s (see Sigmund 1980, esp. 36-39).15 Restrictions also gradually were placed on foreign investment in manufacturing. An example was the (1970) Decision 24 of the Andean Pact, which excluded foreign investors from public utilities and banking, limited profits to 14 percent for repatriation and 5 percent for reinvestment, and provided for transfer of majority interest within fifteen to twenty years. Increasing nationalism and a growing faith in state-directed development were the intellectual factors behind these developments; in the 1970s an additional factor was the easy availability of low-cost external bank loans to the public sector. At the same time, in order to levy an increased inflation tax (see Table 14.2) that would provide resources for directed credit and growth of the state, it was critical to limit the holdings and use of foreign exchange. The first major reversals of these trends began in 1974, when Chile and Uruguay began their well-known attempts to liberalize their financial sectors, both externally as well as internally. In 1978, both countries began 15
Two major nationalizations occurred earlier; in 1938 Mexico nationalized foreign oil companies and in 1948 Argentina purchased British-owned railways with funds that had been blocked in World War II.
Opening the capital account
403
to follow a policy of preannounced, declining devaluations - the tablita in an attempt to reduce inflation. In December 1978, Argentina also opened its financial sector and adopted a tablita. All three countries suffered severe reversals when the debt crisis began and commodity prices collapsed; these problems appear to have been magnified by the interaction between the exchange rate system and the financial liberalization. In Argentina these problems were compounded by the failure to close the public sector deficit, which actually rose after 1978 to over 8 percent of GDP in 1981 (Calvo 1983, Fernandez 1985), and the South Atlantic Conflict with Great Britain in early 1982. Not surprisingly, Argentina's financial sector liberalizations not only proved a failure, but led to destabilizing capital outflow - estimated at more than U.S.$20 billion from the end of 1979 to the end of 1982.16 However, the experiences of Uruguay and Chile, which closed their fiscal deficits during the liberalization and thus met the first precondition of international financial liberalization, shed some light on issues of speed and sequencing of the capital account and the problems that may arise. 4.2
The case of Uruguay: 1974-92
Uruguay's experience provides an indication of the benefits of rapidly opening the capital account, within the context of a bank-dominated domestic financial sector. Uruguay undertook the first, and deepest capital account liberalization of the three Southern Cone countries. Its capital account liberalization was only later followed by an incomplete trade liberalization. The positive results of this experience suggest that there are some benefits to opening the capital account even if the current account is fairly closed. Uruguay suffered a long period of slow growth from the 1950s onward. This reflected both a distortionary policy of import-substituting industrialization, made more costly by a market too small to support economies of scale, and a steady expansion of the public sector (see Hanson and de Melo, 1983, 1985). During this period, Uruguay followed the trend of increasing capital account controls described earlier. Uruguay also ranked among the more inflationary Latin American countries during this period. This reflected not only its own use of the inflation tax to expand and sup16
The estimate is based on summing the identified capital inflow from official and private sources, adding the reserve loss and then deducting the current account deficit. (See World Bank 1985 for this methodology.) This estimated capital outflow was only slightly smaller than the U.S.S23 billion of medium- and long-term debt incurred over the same period. Analyses of the period suggest that the public borrowings effectively were poured into the foreign exchange market to sustain the exchange rates of the tablita.
404 James Hanson port a large public sector and distribute subsidies, but also the impact of its two large inflationary neighbors, Argentina and Brazil. With substantial trade, tourist, and capital inflows from countries prone to engage in monetary and exchange rate experiments, Uruguay found it difficult to defend itself from imported inflation. Although it too engaged in various exchange rate experiments, generally speaking its policy was (and continues to be) one of "leaning against the wind" coming from its neighbors. In 1974 the new military government reversed the trend of increasing capital account controls. It allowed purchase and sale of foreign exchange in a parallel market and the establishment of contracts in foreign currency, including dollar deposits. The government began itself to issue dollar denominated debt, which eased itsfinancingof the (declining) public sector deficit. At the same time, the domestic banking system was liberalized through increases in the ceilings on domestic currency deposits, and a reduction in lines of subsidized credits and reserve requirements (to reduce the cost of reserve requirements, interest also was paid on them during this period.) In 1977, barriers to entry in banking were reduced, permitting "banking houses" (casas bancarias) to compete with banks and easing entry of foreign banks, in line with the strategy of turning Uruguay into a financial center. In 1978 the government unified the exchange markets and established a tablita. Thefiscaldeficit was closed between 1974 and 1979, and remained closed through 1981. The results of the internal and external financial liberalization were favorable. By 1981, the ratio of the banking system's deposit liabilities to GDP reached 43 percent, compared with 19 percent in 1974. Foreign currency liabilities grew somewhat faster, from 4 percent of GDP in 1974 to 18 percent in 1981. At least 48 percent of these deposits were held by foreigners, versus 17 percent in 1974.17 Private commercial banks expanded their share of credit - to 65 percent of the total versus 36 percent in 1974 - at the expense of the public Mortgage Bank and Bank of the Republic.18 Of the twenty-one private commercial banks, thirteen were owned outright by foreigners and five of the remaining eight were largely owned by foreigners in 1981. The liberalization of the financial sector, along with increased external borrowing, contributed to a rise in both domestic and foreign saving rates. This made possible higher investment rates, despite the rising real interest rates in pesos and as would be predicted by the advocates of liberalization 17
18
The deposits are considered foreign deposits if the depositor registers with a foreign address. However, Argentines and Brazilians also make deposits in Uruguay through lawyers domiciled in Uruguay and these are typically recorded as domestic deposits. This was in spite of the significant advantages enjoyed by the Bank of the Republic, including no reserve requirements and compulsory deposits of public enterprises.
Opening the capital account
405
(McKinnon 1973). The liberalized financial sector also tended to retain the funds that were attracted; except in 1982, Uruguay was more successful than Argentina or Mexico in avoiding capital flight. Uruguay, in contrast to the rest of Latin America, grew faster over the period 1974-81 than in the period 1955-73; per capita GDP growth increased from -0.1 to 0.88 percent annually. The rise in the real interest rate, despite the opening of the capital account, is a paradox that requires some explanation. The deposit rate in local currency appears to have been closely linked with the deposit rate in dollars, differing from the rate on dollars by a constant (Hanson and de Melo, 1985, 925). The (declining) rate of devaluation does not appear to be a statistically significant determinant of the domestic interest rates in pesos. Since the rate of devaluation was steadily declining under the tablita, this implies that the premium over the actual devaluation was steadily increasing.19 One explanation of these results is that the effect of preannounced slower devaluation was offset by increased risk of a break in the tablita with large step-devaluation, requiring an increasing risk premium over time. This, of course, would imply that uncovered interest parity, using the actual devaluation as a proxy for expected devaluation, did not hold. Finally, with the interest rate apparently being determined by the (slightly declining) dollar interest rate plus a constant, and inflation falling as a result of the downward pressure exerted by the tablita, the real interest rate in pesos increased sharply. The situation changed dramatically in 1981—82. The fall in primary prices, the debt crisis, and the collapse of the country's two main markets (Argentina and Brazil) hit Uruguay hard. The cut in voluntary external commercial loans forced the country to turn a trade deficit averaging 4.8 percent of GDP in 1979-81 into a surplus averaging 5.3 percent of GDP in 1983-85, in order to meet interest obligations. This in turn required massive depreciations and a sharp cut in aggregate demand. Beginning in 1982, the economy suffered a severe recession, that lasted until 1984. However, after 1984, growth picked up substantially. Even including the 1982-84 recession, growth in 1974-91 was higher than in 1955-73, the only country in Latin America where this was the case (except for Ecuador, where oil production came on stream in 1973). Despite the massive external shocks in the early 1980s, Uruguay's capital market remained opened and regulation of the domestic banking system remained limited. However, the collapse in commodity prices in the early 1980s severely strained the banking system. Initially, deposits and loans in pesos grew sharply to take advantage of the high real interest 19
Lizondo appears to obtain a similar result for Mexico in the period 1977-80.
406
James Hanson
rates. In addition, in 1979 and 1980, the banks began to on-lend some of their dollar deposits locally, in pesos, at the increasingly high real interest rates. With the collapse in commodity prices, borrowers found it difficult to repay, especially their peso loans. In 1981-82, the Central Bank offered exchange guarantees, at a small cost, and this did lower the interest rate in pesos somewhat (Hanson and de Melo, 1985). The Central Bank also encouraged commercial banks to refinance the loans, particularly agricultural credits, on preferential terms and provided funds (Central Bank, 1981, 1982). In 1982, as the risk of devaluation increased, the commercial banks sought to convert peso loans into dollar loans as they came due. Dollar deposits grew sharply in speculation against the tablita, and, for the only time during the period, relatively large amounts of funds flowed out of the country (in part this may reflect the concurrent financial problems of Argentina). In 1982-83, the government intervened to purchase the loan portfolios of failing banks and nonperforming loans of other banks. In effect, it provided deposit insurance through monetary emission, which contributed to the large depreciations. Also in 1982, the government accepted the proposal of the foreign banks' offer to sell their loan portfolios in exchange for new loans from the banks equal to 2.5 times the amount of purchase, that is, in exchange for substantial new money, depositors were bailed out, and the private sector's debt to foreign banks was converted into external debt of the government. The three locally owned banks were unable to take advantage of this offer, because they lacked funds to provide the new money necessary to participate in the scheme. In 1985, the new, democratic government decided that these measures had not resolved the economy's overindebtedness and that continuation of the case-by-case individual bankruptcy approach would be timeconsuming and unproductive. It therefore forced creditors to accept a general refinancing, on favorable terms, of qualified internal debt. Over the period 1985-87, the three domestic banks were taken over by the government, leaving the financial system composed of the Central Bank, the public Bank of the Republic and the Mortgage Bank, three intervened banks, nineteen commercial banks owned by foreigners (eleven of which were incorporated nationally), and sixteen "banking houses." The numerous interventions by the Central Bank over the period, and the difficulties in liquidating the three intervened banks generated quasi-fiscal deficits equivalent to about 3 percent of GDP and 0.5 percent of GDP respectively, in 1991. This deficit is, of course, a complicating factor in stabilizing the economy and conducting monetary policy. The government also decided that the previous approach to bank regulation - largely depending on the foreign owners to monitor their own behavior and avoid insolvency because of the cost of negative publicity -
Opening the capital account
407
was insufficient. Minimum reserve requirements were reimposed. In 1989, the government tightened bank regulations, placing stricter control over new banks, requiring adequate capital (by 1992 the banks were expected to attain the Basel standards for capital adequacy), limiting exposure to individual debtors, and requiring more extensive information from the banks. In 1991, the Central Bank also extended its monetary control more firmly over the public Bank of the Republic, by imposing reserve requirements and reducing the benefits of its preferential access to public sector deposits. In common with most of Latin America, Uruguay experienced significant capital inflow in 1991-92. These flows reflected the expansionary monetary policy and lowering of interest rates in most industrial countries. Additional causes of the boom were substantial spillover from the boom in Argentina and the regularization of Uruguay's external debt in the 1991 Brady Plan deal. Indeed, it seems likely that the freezing of deposits in Argentina and Brazil in late 1989 probably increased their citizens' demand for deposits in Uruguay. Initially, Uruguay attempted to cope with these flows by sterilizing them - selling Central Bank debt to mop up liquidity. This policy was unsuccessful as it kept the interest rate high and, with the attempt to maintain a fixed real exchange rate, sustained the interest of foreign savers; it also increased the Central Bank's quasi-fiscal deficit as the interest rates on the Central Bank debt had a negative spread compared with the interest rate on the corresponding accumulation of reserves. Later the Central Bank's policy shifted to one of allowing the domestic interest rates to decline (retiring the Central Bank debt that had been issued) and the real exchange rate to appreciate, as well as widening the band of its managed float. In addition the government used part of its reserve accumulation to finance its debt reduction deal.20 This policy offset the shock only partially. The higher capital inflow, and the boom from higher Argentine imports and tourism, contributed to continued high inflation despite Uruguay's substantial fiscal adjustment. At the same time, the shock was not wholly negative - higher demand raised real growth in 1992 to more than 3.5 percent. A more aggressive exchange rate policy - for example, a full float - probably would have reduced inflation faster but cut off some of the growth, especially in the export industries. Tightening the already strong fiscal contraction would have offset the increase in aggregate demand. However, this was not an easy option to undertake. On the revenue side, Uruguay already has a Uruguay initially financed its debt reduction deal almost wholly out of reserves. In 1992, the World Bank disbursed its support for debt reduction, rebuilding part of the lost reserves.
408
James Hanson
fairly high tax burden and high public sector prices that limit its international competitiveness; on the expenditure side the administration has found it difficult to secure the necessary legislation to reduce excessive public employment, reduce high social security benefits, and privatize major public enterprises from a congress in which it does not enjoy a majority. 4.3
The case of Chile: 1974-92
Chile's capital account liberalization was slower than Uruguay's and took place in the context of a more diversified financial system, a more open current account, and a tradition of larger direct foreign investment. Like Uruguay, Chile traditionally was one of the higher inflation countries in Latin America. Over the 1960s, Chile followed the secular trend toward tighter exchange controls and increased domestic financial repression. Allende's government (1971-73) intensified these trends sharply, as part of its attempt to socialize the economy. The foreign copper companies and the phone company were nationalized in 1971, without compensation. Numerous domestic firms, including most of the banks also were nationalized or "intervened."21 Tight ceilings were placed on interest rates, and credit was allocated by the Central Bank. Exchange control was tightened legally, to support a complex and distortionary system of multiple exchange rates. However, illegal capital flight, through the rapidly depreciating parallel market, became rampant. After the violent coup in 1973, the new military government began in 1974 to reverse many of these measures. The exchange rate and the low public sector prices were increased substantially. The companies "intervened" during the Allende government were returned to their original owners. The government also began negotiating with the expropriated copper companies. Tariffs were lowered sharply, beginning a process that would culminate in 1979 with a maximum tariff rate of 10 percent (Edwards and Cox-Edwards 1987, 112). In the financial sector, the government concentrated on liberalizing the domestic system substantially during the period 1973-76. In March 1974, finance companies were allowed to pay unregulated rates and, by policy, were unregulated - by 1978, twenty-one "financieras" existed accounting for about 6 percent of financial assets (Larrain 1989). The ceilings on commercial bank rates also were increased and, in October 1975, eliminated (although reserve requirements remained high, contributing to a large spread and high real lending rates; see Arellano 1983). Also in 1974, the 21
Domestic nationalized firms received notional compensation; intervened firms were taken over after labor disputes without compensation.
Opening the capital account
409
government began privatizing the "nationalized" banks (and other nationalized firms). Rather than returning them to their original owners, the government sold them. Regulations limited the concentration of bank ownership, and no foreign ownership was permitted. The government received only U.S.S192 million for the shares in the 18 banks sold, part of which was in debt for two years at highly subsidized rates; the 118 nationalized industrial firms were sold off*for only U.S.$290 million on even more favorable terms, with debt of up to ten years maturity (Herrera and Morales 1992, 147). In fact, the regulations were largely circumvented, and at a low cost, industrial-financial ownership became concentrated in a few groups that were willing to accept the large risks of the period, with consequences to be discussed. During its first few years, the government encountered two difficulties in the domestic financial sector. First, the housing finance system (representing nearly 50 percent of deposits) became overindebted to depositors and in 1975 the government froze withdrawals, significantly reducing confidence in the financial system. Second, in late 1976 and early 1977, a financial crisis in the Banco Osorno group led to the government's intervention - in effect bailing out depositors despite the government's earlier declarations that there was no deposit insurance. This bailout set a precedent that had significant implications in the financial crisis of the early 1980s. The capital account was liberalized for individuals in the period 1974— 76, but the opening was much slower for financial intermediaries. In 1974, the Foreign Exchange Control Law was liberalized. Penalties for holding foreign exchange were significantly reduced. A tax and exchange control holiday also was decreed but little foreign exchange was surrendered. In 1975 a separate foreign exchange purchase window was opened, allowing sales of foreign exchange without identification. Article 14 of the Foreign Exchange Control Law was modified to permit the Central Bank to register incoming funds, including funds for stock market investments, and to commit to repatriation of interest and capital (after a holding period that varied from six to twenty-four months; see Herrera and Morales 1992); however, banks were initially not allowed to intermediate funds under Article 14. A second pillar of the international financial liberalization was Decree 600 of 1974, which regulated direct foreign investment. Under the decree, foreign investors could sign a contract with the government that established the tax regime, access to the official capital market for remittances (after a preestablished holding period of three years for capital), and equal treatment with domestic investors. (Given the conflicts between this decree and Decision 24 of the Andean Pact and the differences in tariff policy
410
James Hanson
between Chile and other members of the pact, Chile formally left the pact in 1976.) The net effect of Article 14 and Decree 600 was, by 1976, to provide a relatively open financial market for nonbanks. However, financial intermediaries remained relatively restricted; indeed, they were discriminated against in their ability to intermediate foreign funds. There remains a controversy over whether this distinction was important and whether it affected the later outcomes. Clearly little money flowed into the economy - the capital account balance averaged only about U.S.S300 million per year in 1974-76. A total of less than US.S300 million entered under Article 14, 75 percent in 1976.22 In evaluating the apparent limited impact of the liberalization, it is important to note the lack of official lending to the country for political reasons (which would have offset any increase in private capital), and the perceptions of high risk, related to the political situation (including a potential conflict with Argentina over the Straits of Magellan), the uncertain state of the financial market, and the poor state of the economy in 1974-76. The military government had begun with a gradual stabilization program to reduce inflation. However, the 1974 collapse of copper prices, the high oil prices, and the lack of official external finance left the economy with a growing, uncovered fiscal deficit. In April 1975, in the response to a rising deficit and inflation, the government switched to a shock treatment - across the board cuts in ministerial spending, a temporary 10 percentage point increase in income taxes, and tight monetary policy. The fiscal deficit was cut from 10.5 percent of GDP in 1974 to 2.6 percent in 1975 (see Edwards and Cox-Edwards 1987, chap. 2). However, the collapse in the terms of trade, the fiscal tightening, and the high real interest rates (brought about by the tight money and lack of capital inflow) reduced aggregate demand substantially. GDP fell 13 percent in 1975, unemployment increased sharply, and inflation for the year was 343 percent (although the rate dropped sharply after the first quarter). In this context large private capital inflows were unlikely, no matter what the regulations were on the capital account, and any inflows would have sought high returns to compensate for high risk. The recession was very sharp, but the rebound also was sharp. By 1977 GDP recovered its 1974 level and growth averaged 8 percent annually in 1977-81. (However, in per capita terms, real GDP exceeded its 1974 level only in 1979.) The main problem on the real side was the stubbornness of unemployment despite the high growth. In 1976 the balance of payments 22
Recall, however, that the total cost of the privatized industrial firms and banks was only U.S.$482 million and part of that was in debt of the purchasers. Thus, inflows through Article 14 would have been sufficient to purchase most of the nationalized firms.
Opening the capital account
411
went into surplus, the first of a string of six annual surpluses. The fiscal deficit was turned into a surplus by 1979, the first of five annual surpluses. Annual inflation was roughly halved each year between 1976 and 1978, declining to the 30-40 percent range. In February 1978, in an attempt to lower inflation further, the government began a policy of preannounced, declining devaluations (the tablita) and in June 1979 fixed the exchange rate (after a step-devaluation). However, this proved ineffective, as inflation remained in the 30 percent per annum range in 1978-80, only in 1981 falling to 9.5 percent, as a crisis began to develop.23 The period 1977-81 was characterized by massive capital inflows. These inflows financed current account deficits that averaged 7.7 percent of GDP in 1977-81, plus annual increases in reserves that averaged 2.6 percent of GDP. The government's policy was that the private sector should bear the risks of external indebtedness, and only U.S. $1.6 billion of the U.S.S10.8 billion increase in external debt accumulated between 1976 and 1981 was public. Unlike Argentina, little of Chile's capital inflow seems to have gone into capital flight during this period. On the downside, however, the investment rate did not increase as much as foreign saving (the current account deficit) - the domestic saving rate did not increase very much and actually fell in 1981-82. (See Schmidt-Hebbel 1988 for further discussion.) International financial liberalization proceeded fairly rapidly in 197781. The government felt substantial political pressure to lower the high domestic real interest rates and tried to do so by opening the capital account further.24 First, foreign banks were allowed to open offices. By 1978 the five foreign banks represented about 2.4 percent of credits extended by the financial system; by the end of 1980 their numbers had increased to fourteen and they accounted for nearly 5 percent of bank lending (Valdes-Prieto forthcoming). Second, in mid-1979, the Central Bank opened a foreign exchange "window" for sales of up to US.$10,000; this involved a small outflow in 1980 (U.S.S450 million), but became a vehicle for some capital flight (US.$950 million in 1981 and U.S.S1230 million in 1982) until suspended in October 1982. Third, after September 1977 banks were allowed to intermediate inflows under Article 14, albeit under 23
24
A number of authors have attributed the stubbornness of inflation, despite a fiscal surplus and a relatively open current account, and the steepness of the 1983 recession, to the backward-looking indexation of wages. See, e.g., Corbo 1985 and Edwards and CoxEdwards 1987. However, by the same token, at least part of the growth prior to 1983 must also be attributed to the impact of backward-looking indexation on wages (Hanson 1986). On the domestic side, attempts also were made to lower rates by paying interest on the high reserve requirements (a policy begun in 1976 and then eliminated over 1979-80) and, then, lowering the reserve requirements (1976-1980) (see, e.g., Arellano 1983). In 1981, the rights of creditors in foreign currency were clarified significantly (Valdes-Prieto forthcoming).
412
James Hanson
tight exposure regulations in addition to the minimum term restrictions (twenty-four months after July 1976; see Herrera and Morales 1992). The exposure rules were relaxed in 1979, and finally dropped in April 1980, but reserve requirements were imposed, under an ascending scale that reached 25 percent for loans of two to three years and 100 percent for loans under twenty-four months, which effectively forbid short-term external credits (Valdes-Prieto 1994). This disincentive to short-term borrowing meant that even in 1981 the average term of bank lending under Article 14 was about sixty months (Edwards and Cox-Edwards 1987). Banks also were prevented, by other regulations, from taking a net position in foreign exchange - this of course reduced exchange risk but increased credit risk. Not surprisingly, in light of the restrictions on banks' intermediation of external funds, the main vehicles for capital inflows in 1977-80 were direct private borrowing and window sales of foreign exchange (which could then be converted into domestic bank deposits with an implicit government guarantee). Although bank intermediation increased dramatically after 1978, in 1979 bank intermediation was still less than nonbank inflows under Article 14, not to mention the still larger window purchases of foreign exchange. In 1980, although bank intermediation increased dramatically under Article 14 tripled, to reach U.S.S1.6 billion, nonbank Article 14 inflows, plus window purchases equaled U.S.S1.8 billion. In 1981, the totals of nonbank flows and domestic bank intermediation under Article 14 were about equal, each about U.S.S3 billion (see Valdes-Prieto forthcoming). These figures cast some doubt on whether the source of Chile's debt crisis was a "rapid" opening of the capital market in 1980. The capital account was, as discussed earlier, essentially open to nonbank intermediaries prior to 1977; it is quite likely that if Article 14 had not been extended to banks, then even larger amounts of funds would have been intermediated by nonbanks or been exchanged in window sales and then deposited in banks to earn high, implicitly insured rates of return.25 Even with the opening of the capital account, the growth in debt to external financial markets was not much faster than for developing countries as a whole.26 On the domestic side, perhaps a better explanation of the 25
26
This hypothesis also has some econometric support. Both Corbo 1985 (using the Edwards and Khan model) and Schmidt-Hebbel 1988 (using a Kalman filter technique) find little support for the increasing opening of the capital account between 1977 and 1982. Medium- and long-term lending from financial markets to developing countries rose from US.$11.2 billion at the end of 1976 to U.S.$22.0 billion at the end of 1979 and to US.$36.2 billion at the end of 1982. Chile's debt to private creditors rose from US.$2.6 billion at the end of 1976 to U.S.S6.4 billion at the end of 1979 and US.$12.8 billion at the end of 1982, which was not much faster than the changes in developing country debt. However,
Opening the capital account
413
massive capital inflow was the improvement in economic performance, the growing reputation that Chile enjoyed in international financial markets and foreign investors' perceptions that the political situation, including the potential conflict with Argentina, had stabilized. When these local factors were coupled with the easy money policy in industrial countries and the recycling of petrodollars, the result was the massive increase in commercial bank lending. Despite the massive capital inflow, real interest rates in domestic currency remained fairly high and the premium over LIBOR fairly large (see Edwards and Cox-Edwards 1987, Valdes-Prieto 1994). Although a number of models have been advanced to explain these phenomenon, in general the "fit" is not good. The data suggest a sharp drop in the premium over LIBOR in early 1980, and then a fairly constant large premium (an average of 21 percentage points with a standard error of 3 percentage points) until the first quarter of 1982, despite the fixed exchange rate. This suggests that, as was the case in Uruguay, the exchange rate policy was not really credible and thus a substantial risk premium was demanded for deposits and loans in local currency. Chile's growth declined to 5.5 percent in 1981. With the 1979 increase in oil prices, Chile's terms of trade had collapsed. Then, at the end of 1980, dollar interest rates rose sharply. These increased dollar rates, plus the continued high-risk premium in conjunction with the drop in inflation to below 10 percent, generated a rise in already high real interest rates in domestic currency. Much of the widening of the current account deficit (to 14.5 percent of GDP) and the more than 50 percent increase in debt to external banks went not into investment but into consumption, which increased by 4 percentage points of GDP, perhaps in speculation that the open current account and the fixed exchange rate could not last. The high interest rates also appear to have induced distress borrowing in 1981; in addition there may have been some "round-tripping" to take advantage of increased expectations of devaluation and the implicit guarantee of deposits.27 Bank failures began to occur - the government intervened in eleven intermediaries between November 1981 and December 1982 and subsequently liquidated all of them. Belatedly, the government began to worry about fraud and the fiscal cost of the implicit deposit guarantee. Restrictions were tightened on loans to affiliated parties in August 1981. In 1982 Chile went into a second steep recession. GDP fell 14.5 per-
27
there was a substantial increase, over 50 percent, in Chile's debt in 1981 (see World Debt Tables, 1988-89 and World Bank 1990). "Round tripping" involves a capital inflow into Chile, perhaps guaranteed under Article 14 or through the window to make a deposit, then a conversion of these funds back into dollars through a loan. The evidence for this is implicit—capital inflows through Article
414
James Hanson
cent, and unemployment rose to over 25 percent. Speculation and the continued large current account deficit produced a U.S.S1.4 billion reserve loss over the year. This forced the government to abandon its fixed exchange rate in June 1982, and then to float in August, which led to a 43 percent devaluation. However, the government protected debtors in foreign currency by allowing them to purchase dollars at a preferential rate (a 15-20 percent discount from the official rate), an arrangement that prevailed through mid-1985 and involved serious cost to the government. Further pressure was put on the economy by Mexico's August 1982 suspension of external debt payments, which raised doubts about the future of even the reduced capital inflows to Chile. In late September, the Chilean government switched to a crawling peg with the objective of roughly maintaining the real exchange rate, reinstituted capital account controls, and closed the window for foreign exchange sales. A legal parallel foreign exchange market developed, which has continued through 1992. The combination of the high real interest rates in domestic currency and then the devaluation and the sharp slowdown in new external funds bankrupted many of the banks' borrowers, who in many cases had become the owners of the banks in the mid-1970s and had been extended nonarm's length loans much in excess of the exposure rules (see Arellano 1983, 24, for some data on the extent to which the exposure rules were violated). In January 1983, the government intervened in five more banks, including the two largest, and began to liquidate three others. However, the resolution of these cases dragged on nearly two years. The government guaranteed the deposits of the intervened banks and offered a 75 percent settlement to creditors (including depositors) of the liquidated banks, which domestic depositors accepted. The government also imposed four domestic debt reschedulings on preferential terms, and purchased risky loans from the banks (with a repurchase agreement by the banks' shareholders). Finally, in 1985 the Central Bank recapitalized the intervened banks by offering shares in the market and providing subsidies to small buyers of shares. The recapitalization program was highly successful in diversifying ownership; the two largest banks now have about 39,000 and 16,000 shareholders respectively.28 The government also attempted to impose losses on the foreign creditors of the intervened banks. However, this provoked additional capital outflow from the nonintervened banks and other foreign investors. The government finally was forced to freeze the amortization payments to the
28
14 and the window increased dramatically in 1981, but capital outflows through the window more than doubled (see Valdes-Prieto 1994). For further discussion of this period see Larrain 1989 and Valdes-Prieto 1994.
Opening the capital account
415
foreign banks and then agree to recognize fully all foreign debt claims (including those of the liquidated banks) in the July 1983 rescheduling.29 Since 1983, Chile reduced its external debt sharply through the world's most extensive debt-equity swap program, so that by 1991 it was no longer considered severely indebted; it also rescheduled amortization payments a number of times, most recently in September 1990 when it received favorable terms and some "quasi-new money" (World Bank, 1990). In September 1992, Chile became the first of the formerly severely indebted countries to receive an investment grade rating (BBB) from Standard and Poor, facilitating public (and private) access to international capital markets. The net effect of the 1981-83 financial sector interventions was to leave the Central Bank with a massive debt. The service on this debt, plus Central Bank costs, less interest earned,30 was equivalent to about 3 percent of GDP in 1991. The Central Bank must finance this quasi-fiscal deficit by emission or new debt issues, which complicates its ability to undertake monetary policy for contracyclical reasons and creates significant pressures simply to manage the debt. On the real side, the Chilean economy rebounded less quickly from the 1981-83 crisis than from the 1973-75 downturn. It was not until 1987 that 1981 GDP was exceeded. In part this may reflect the uncertainties created by the lengthy process of financial sector restructuring, including the large debt overhang. However, helped by an improvement in the terms of trade after the mid-1980s, GDP growth averaged over 6 percent annually in 1986-1992. For the period 1974-1991, GDP growth was about the same as in 1955-73, even including the two recessions, but the level of GDP appears not to have recovered from the decline in the 1972-74 period. After 1985 the government began again to liberalize the financial markets, but much more cautiously than in the mid-1970s. Bank supervision was improved substantially. At the same time, the most noteworthy development has been the growing importance of the capitalized pension system, begun in 1981. Under this system, individual contributors choose a mutual fund and invest in personal accounts out of which pensions eventually will be paid (the government guarantees a minimum pension). These mutual funds have become a dominant factor in what has become the deepest capital market in the continent, accounting for 20 percent of bank deposits, 38 percent of central bank securities, 100 percent of treasury securities, 56 percent of mortgage and corporate bonds, and 9 percent of equities (Vittas and Iglesias 1992). The government tightly regu29 30
Foreign banks also suffered large losses, and in many cases stockholders lost their equity. The amount of interest earned depends on payments by the Treasury on its debt to the Central Bank, much of which was incurred as a result of the financial crisis of the early 1980s.
416
James Hanson
lates the investments of the mutual funds, reflecting its caution after the financial crisis of the early 1980s but also the need to roll over its own debt. In 1992, the mutual funds were allowed to invest up to 1.5 percent of their funds abroad. Individuals and nonfinancial firms remain largely free to undertake capital inflows (through either the official market, under restrictions on repayment maturities, or the legal parallel market). In 1991, firms specifically were allowed to invest overseas using the parallel market, and Chilean direct foreign investment in Argentina in 1992 was some U.S.S300 million. Nonfinancial firms also have been allowed to offer shares (American depository receipts, or ADRs) in equity markets and, in 1992, external bonds. Foreign investment, including project financing, has grown dramatically, to take advantage of Chile's many comparative advantages. In contrast, financial firms remain regulated in their ability to borrow and lend abroad, and reserve requirements exist on foreign exchange borrowings to equalize their attractiveness with domestic credits. An exception was the April 1991 liberalization, which allowed banks to invest overseas up to 25 percent of their liabilities in foreign currency. Macroeconomic policy continues to be limited by the open capital account and the overhang of the past. In 1990, the democratically elected government took over in the midst of an unsustainable boom. Committed to undertaking a number of social programs, financed with a tax increase, the government felt compelled to use monetary policy to cool the economy. The government raised the interest rate by offering high-interest-rate Central Bank debt to all comers. The policy successfully slowed inflation, but, in the context of the policy of maintaining a constant real exchange rate to sustain exports, induced a substantial inflow of foreign exchange.31 Since the interest rate on the Central Bank's new debt exceeded the interest rate earned on reserves, the Central Bank had a negative spread. Its deficit increased by the equivalent of a further 0.2 percent of GDP. In 1991 and 1992, the government eased its high interest rate policy and, at the same time, eased its targets on the exchange rate by setting the official rate within a much wider band (+ or — 10 percent of the central parity), allowing greater play for market forces to absorb capital inflows. 4.4
Lessons from the experiences of Uruguay and Chile
The experiences of Uruguay and Chile suggest five lessons. First, the experience of Uruguay suggests that opening the capital ac31
Part came through the parallel market and the capital account, appreciating that rate, but it also was possible to import funds to take advantage of the official rate for current account by overinvoicing exports.
Opening the capital account
417
count is beneficial, even if the current account remains relatively closed that is, import-competing goods enjoy high protection. Second, Uruguay and Chile, as well as many other Latin American countries, ended up nationalizing private external debts; if this is likely, then the government must take action to protect itself against excessive liability. A large percentage of the debts that were assumed reflected external liabilities (including foreign currency denominated deposits) of financial intermediaries that were intermediated domestically. The governments assumed the debts in response to the combination of implicit deposit insurance guarantees, domestic and external pressures, and desires to sustain at least a minimum of capital inflow. This experience suggests that supervision is needed to limit intermediaries' exposure to internationally intermediated funds, broadly defined to include foreign currency - denominated deposits and off-balance sheet items.32 However, supervision is unlikely to be fully effective and in any case the issue is in part a political one - there were strong domestic pressures in Uruguay and Chile to lower domestic borrowing costs by allowing greater intermediation of foreign funds. It is difficult to resist such pressures when growth can be stimulated by capital inflows bearing negative real interest rates. Third, there appear to be a number of policies within the financial sector to limit the intermediation of foreign funds, if the political will exists to limit the boom. These policies include limits on exposure to individual borrowers such as higher required capital, required offshore placement of a certain proportion of foreign currency-denominated liabilities (which in turn requires a fairly open capital account), and equalization of reserve requirements and taxes on foreign and domestic currency-denominated liabilities. Fourth, outside the financial sector there are the more fundamental issues of the exchange regime and the mix of monetary and fiscal policy. The experiences of Uruguay and Chile suggest that a fixed parity, with relatively narrow bands, may have encouraged excessive capital inflows and economic volatility, compared with either a floating regime, or formal dollarization.33 This is particularly true if the country also attempts to fix the interest rate and sterilize the capital inflow - the country will then experience even larger capital inflows and accumulate international reserves. A simplified version of the problem is that the fixed parity regime attracts external lenders by high real interest rates in foreign currency 32
33
In other countries—e.g., Ecuador—the intermediaries' external debts mounted sharply when borrowers who had received off-balance sheet guarantees defaulted on the external debts. Dollarization is no guarantee against excessive foreign indebtedness, however. Panama, despite its formal dollarization, ranks among the world's most indebted countries.
418
James Hanson
terms. This is especially true since the lenders expect to be able to take out their funds in advance of a change in the fixed parity (or, in the case of deposits, be insured against loss). The capital inflow temporarily lowers the real internal interest rate and encourages growth while, in the context of the fixed exchange rate, depressing inflation by financing additional imports and creating a real appreciation. However, the current account deficit widens, especially once interest earnings begin to be repatriated. Eventually, external lenders lose confidence and withdraw their funds. Real interest rates shoot up sharply, and a credit crunch develops, bankrupting borrowers and making the government responsible for deposits. The creation of money to bail out the depositors adds to the pressure on the exchange rate, forcing a devaluation and bankrupting borrowers in foreign currency.34 In contrast to this scenario, in 1991-92, Uruguay and Chile (after a successful but costly episode of using high interest rates to slow inflation) absorbed part of the capital inflow by allowing greater flexibility in the exchange rate (by widening the bands around the central rate and, in Chile, by the use of a legal parallel market)35 and in the interest rate.36 Fifth, the experiences of Uruguay and Chile, which have bankdominated financial systems, do not shed a great deal of light on the appropriate sequencing of liberalization within the capital account. Chile pursued a policy of liberalizing the capital account for individuals while retaining substantial control over financial intermediaries, and the experience when that policy changed, though conditioned by external circumstances, suggests it is a risky one, because of the possibility that the government will be forced to assume the intermediaries' external debts. In the 1990s, Chile has allowed its firms to operate fairly freely in foreign equity markets, but has proceeded slowly on international financial liberalization for financial intermediaries. In particular, pension funds have been restrained from overseas investment. This probably has increased the volatility of the domestic stock market (because of the limited number of eligible 34
35
36
These pressures were exacerbated by the external circumstances at the early 1980s, with the rise in petroleum prices and the worldwide recession brought on by tight money in the industrial countries, which both increased interest rates and reduced commodity prices. The recession at the beginning of the 1990s, in contrast, is characterized by low interest rates, easing the pressures on debtor countries. In contrast to the usual situation the rate applicable to capital account transactions appreciated relative to the current account rate. The difference between the external conditions in the early 1980s and the early 1990s also has made the response easier. In the early 1980s commodity prices were low and industrial countries' monetary policy switched from loose to tight during the recession, as they loosened fiscal policy. In the early 1990s, it appears that loose monetary policy in the industrial countries will accompany low commodity prices (and Chile continues to enjoy high copper prices).
Opening the capital account
419
issues), but the policy also reflects both the government's need to finance its liabilities accumulated during the bank crisis of the 1981-83 and the quest for financial stability that was dealt a rude blow in the early 1980s. 5
Summary and conclusions
The question of the costs and benefits of opening the capital account has become moot to some degree. De facto, capital has become internationalized by the growing integration of the world economy. Compared with the situation in, say, the 1960s, economic agents have much greater access to foreign currency assets. This reflects the freer international mobility of goods, people, and information, as well as the greater legal availability of such assets in many countries. Often, for individuals and nonfinancial firms, it is not access but the menu of assets that is limited. This limitation largely reflects the remaining restrictions on financial firms. It nonetheless seems useful to examine the costs and benefits of opening the capital account, the empirical results of opening the capital account, and the timing and sequencing of opening the capital account, in order to assist countries in the process of capital account liberalization. Theoretically, perhaps the main benefit in opening the capital account - in the sense of allowing individuals and firms full access to international capital markets - seems to come from the greater scope for exercising individual choice in diversifying risk. If domestic and international prices of risky assets differ, then there are benefits from allowing individuals to trade assets or goods and assets. These benefits are analogous to the benefits of free trade in goods, and, correspondingly, are subject to many of the same caveats as the arguments for free trade, plus those of intertemporal trades. The traditional argument for opening the capital account in developing countries is somewhat different - that capital inflows will finance increased investment more cheaply, in terms of foregone consumption, than domestic saving. This argument is perhaps less forceful in theoretical terms than the risk diversification argument. Under this argument, the benefit of capital account liberalization reflects only on the difference between capital inflows and outflows. Ignoring the benefit from balanced trading in risk is roughly analogous to saying that the benefit from free trade is based on the difference between exports and imports. Also, if the only rationale for capital flows were an excess of riskless investment opportunities over domestic saving at the world interest rate, then, in theory, a credible, maximizing government could borrow funds internationally and either reduce domestic "crowding out" or intermediate these funds through the domestic financial system. There would be no
420 James Hanson need for individual economic agents to access international capital markets. The traditional argument also does not imply a need to open both sides of the capital account. From the standpoint of attracting foreign investors, who basically are interested in assuring that remittances from earnings can be transferred, the question is one of the current account, not the capital account. Guarantees of convertibility for debt service and profits, in theory, could be given. In practice, of course, many governments have found it difficult to carry out such intermediation effectively, or to offer credible guarantees. Governments often have overborrowed to support excessive deficits and overvalued exchange rates, have invested in numerous "white elephants," and have been lax in recovering on-lending when international funds have been intermediated through the domestic financial system. Governments also have suspended guarantees of remittances during exchange crises. Experience thus suggests that, in practice, opening the capital account may be an effective way to mobilize and allocate international funds. This argument for opening the capital account is particularly true for foreign direct investment, which is often motivated less by considerations offinancinginvestment and more by considerations of technology transfer, management improvement, risk diversification, and the like. The standard argument against an open capital account in a developing country is that it would cost the country much of its saving. To some extent, such capital outflow may represent diversification of underlying country risk, not related to economic policy. Inflows also may occur as foreigners attempt to diversify country risk. The net flow resulting from the diversification of underlying country risk thus will depend on the country's underlying characteristics and the degree to which it develops instruments to allow the diversification of risk. To a greater degree the loss-of-saving argument probably reflects the diversification of risk arising from economic policy and is a reflection of the limitations that an open capital account places on the taxation of income from capital and financial assets. To the extent the capital account is closed de facto, developing countries are able to taxfinancialassets and, to a lesser extent, capital. If the capital account were open, then capital flight would occur unless such taxation was reduced. This is particularly true given the tax shelters for external funds in many industrial countries, as well as in the tax havens, which in fact are being used by many citizens of developing countries. Countries using the inflation tax, or in which there is a future risk of the inflation tax, are particularly vulnerable to a loss of saving. Inflation, or the threat of inflation, leads to growing use of foreign exchange in transactions and a reduction in the yield of the inflation tax, the more so as the
Opening the capital account
421
capital account is opened. This limitation on the use of the inflation tax should not, however, be considered solely a cost, since this tax is one that does not require public approval; falls heaviest on the poor, who have a large share of their portfolio in local currency assets with zero or fixed interest rates; and has very distortionary effects, particularly in the financial system. Indeed, opening the capital account might even be considered to improve the income distribution, because it equalizes access to international assets at the same time as it reduces the base of the inflation tax. Since opening the capital account reduces the incentive for reliance on the inflation tax, it sometimes is argued that an open capital account represents a signal that the government will refrain from using the inflation tax. However, this signal is certainly a weak one, since thefirstact of a government intending to use the inflation tax is to close the capital market. An open capital account also limits a country's ability to conduct monetary-exchange rate policy and increases exposure to international monetary shocks. With an open capital account, attempts at targeting the monetary stock through open market operations tend to be offset by economic agents undertaking capital flows (at the government-set exchange rate) in order to obtain their desired stocks of financial assets at the prevailing international interest rate. The loss of monetary independence implied by open capital markets also is felt in the greater exposure to international monetary shocks, through the capital as well as the current account. In an open economy, interest rate targeting may temporarily enjoy greater success than attempts to target the money stock, but carries with it some problems. A government may offer bonds above or lend below international interest rates and thereby affect the domestic interest rate, but only so long as it is willing to run down its net international asset position. Offers of loans or rollovers of bonds at less than world rates will lead to capital flight and loss of international reserves. However, offers of interest rates above international rates also lead to a decline in net international assets, because the influx of foreign funds takes place at a negative spread on the bonds that are sold. In either case the ultimate result will be a rise in the interest rate premium facing the country in international markets. Monetary independence can be restored at the cost of the government's giving up control over the nominal exchange rate and allowing it to float freely. However, a free float clearly entails an increase in the variability of the exchange rate and correspondingly greater risks for exporters and importers than a fixed rate system. Countries occasionally have tried to obtain monetary independence and maintain control over the exchange rate for trade by applying a free float to capital account transactions - in
422
James Hanson
effect trying to create another policy instrument. However, the difference between exchange rates in the two markets, and thus the degree of monetary independence, clearly is limited by arbitrage possibilities. Most of the evidence on the impact of opening the capital account relates to industrialized countries.37 There has been no analysis of the benefits of risk diversification or even the size of such diversification, except to point out that the volume of gross capital flows has grown far faster than the rest of the balance of payments - this increase in gross flows is usually hypothesized to represent increased risk diversification. Empirical analyses concentrate on two issues: the degree to which foreign capital inflows finance domestic investment and the degree to which uncovered interest rate parity typically is achieved. In general, the results for industrial countries suggest: 1
Foreign capital flows (net) finance on average, only 10 to 15 percent of investment in industrialized countries. Various authors have argued that this result reflects similarities in savings and investment functions in these countries, similarities in the business cycle across developed countries, policy related shifts within domestic saving or investment that do not affect external finance, country size or risk premium effects as well as imperfect capital mobility. 2 Uncovered interest rate parity typically is not achieved even between industrial countries. It is not clear if this reflects a market failure, econometric difficulties with the model (the estimates typically are a joint test of the hypothesized pattern of expectations and market equilibrium), or variations in the risk premium. Regarding studies on the developing countries, there are no specific studies of the financing of investment. Some of the studies of foreign capital inflows include developing countries; this inclusion (and studies including smaller industrial economies) suggest that foreign saving represents a larger fraction of investment in these countries than in the larger industrial countries. (However, the developing country results may reflect large aid flows to countries with low incomes and/or statistical difficulties, rather than larger private flows.) Data on the gross stock of external debt suggest a high ratio of external debt to domestic capital stock in many developing countries, which is also indicative of large flows and de facto capital mobility. Regarding interest rates, many of the studies in developing countries also analyze the extent of uncovered interest parity, usually allowing for a 37
See Hanson 1994 for a more extended summary of the empirical results of opening the capital account.
Opening the capital account
423
constant risk premium. A number of these studies are of Latin American countries during periods when they followed a policy of using the exchange rate (and in some cases the interest rate) to slow inflation. This suggests the possibility of a "peso" problem in the estimates - that expectations of devaluation may well have been high and even increasing over time but standard, distributed lag approaches to estimating expectations suggest a long period in which uncovered interest parity, even allowing for a constant risk premium, does not hold. In recent periods the "risk premium" in these countries may have risen because of their external debt overhang - this reflects not only risks of higher taxes (including inflation) to service debt, but also the alternative of investing in external debt, an option that is increasingly feasible given the growing breadth and sophistication of that market. Recent studies of a more diversified sample of countries with stabler macroeconomic environments suggest that the domestic interest rates tend to follow international interest rates to a fair degree, after appropriate allowance for a risk premium. These analyses suggest that capital mobility, although imperfect, may have reduced domestic interest rates by as much as 5 percentage points compared with the autarchic situation. This analysis suggests that developing countries would obtain some benefits from opening up the capital account, even leaving aside the benefits from trading in risky assets. The financing of the 10 to 15 percent of investment, suggested by the developed country studies, would be welcome in many developing countries, especially if accompanied by improved technology and management. The potential reduction in the interest rate also seems fairly sizable. The sequencing and speed of capital account liberalization remain important issues. The preconditions for capital account liberalization are a sound fiscal or monetary situation and a reasonably sound and liberalized domestic financial system. The need for fiscal control is perhaps best illustrated by the experience of some countries in Latin America. Large public sector deficits were financed increasingly with external funds. When exchange rate policy became clearly inconsistent with the fiscal deficit and external borrowing became clearly unsustainable, capital flight ensued and additional borrowing was done to sustain the exchange rate. Given the magnitude of the policy errors, capital flight would have developed in any case. However, the degree of openness of the capital account certainly contributed to the capital flight and excess borrowing. The rationale for a reasonably liberalized and sound financial system has no corresponding illustration. However, it is important to adjust financial regulations such as reserve requirements, capital requirements, portfolio composition (including limits on offshore investments), and in-
424 James Hanson terest rate limitations, as well as labor restrictions, so that domestic institutions do not face unfair competition from foreign institutions and capital inflows. If the regulatory framework favors external institutions significantly, then it eventually may lead to bankruptcy of domestic institutions and require costly government support for depositors. Similarly, unless the domestic institutions are reasonably sound, then allowing foreign institutions into the domestic market is likely to put excessive pressure on domestic institutions. Either the government must clean up the portfolios of domestic institutions, or the institutions must be allowed to earn high enough profits to recuperate themselves, before liberalization. The sequencing of capital account and current account liberalization has been analyzed at length. Conventional wisdom seems to be that the current account should be liberalized before the capital account. The argument is that the responsiveness offinancialflowsto capital account liberalization typically would lead to an unwarranted real appreciation of the real exchange rate and an incorrect allocation of investment between tradables and nontradables. However, this argument neglects the points that whenever the capital account liberalization occurs, it would tend to cause a real appreciation; and the amounts of resources available to an economy would be greater if the capital account were liberalized earlier. The positive impact of capital account liberalization on resource availability has led some authors to argue for simultaneous current and capital account liberalization. Other authors have pointed out that the optimum sequencing of liberalization depends on the type and size of distortions prior to liberalization. The experience of Uruguay suggests that significant benefits can be obtained from capital account liberalization, even in the presence of significant protection for import competing goods. The traditional literature on sequencing current and capital account liberalization seems to consider capital account liberalization as occurring all at once, whereas current account liberalization can take place at various rates - witness the debate regarding the appropriate pace of cutting protection. In fact, the experiences of Uruguay and Chile suggest that capital account liberalization also can take place over time - different instruments and institutions can be legalized at different times and the size of operations can be limited by methods such as taxes and reserve requirements on inflows and exposure limitations. A transitory parallel exchange market may develop in such circumstances, and this can help cushion the inflows.38 However, it should be recognized that such policies are much 38
The difficulty with dual foreign exchange markets is, of course, the problem of separating the two markets and the incentives set up to avoid the separation. In the usual case, there is some desire to maintain an overvalued rate for goods transactions to avoid the inflationary consequences of a devaluation. Capital flight occurs, depreciating the rate applicable to
Opening the capital account 425 39 less effective than similar trade limitations. Perhaps the only policy that can truly cushion capital inflows is a combination of a flexible domestic interest rate and a floating exchange rate, which would increase risks to exporters and import competing industries. The recognition that current account liberalization can take place at varying speeds also suggests that some undesired interactions can occur between the current and capital account. In particular, a slow current account liberalization may require a compensating real devaluation over a number of years. The implied upward pressure on the real interest rate via the capital account may discourage domestic investment, particularly in nontraded goods. This suggests that the response to slow trade liberalizations may be less than to rapid trade liberalizations and that exchange rate policy during the period of trade liberalization should take into account the capital as well as the current account. References Arellano, Jose Pedro. (1983). "De la Liberalizacion a la Intervention: El Mercado de Capitales en Chile 1974-83." Estudios CIEPLAN,MDINno. 11 (December): 5-49. Borts, George, and Jerome Stein. (1964). Economic Growth in a Free Market. New York: Colombia University Press. Calvo, Guillermo. (1983). "Trying to Stabilize: The Case of Argentina." In Pedro Aspe, Rudiger Dornbusch, and Maurice Obstfeld, (eds.), Financial Problems of the World Capital Market: The Problem of Latin American Countries. Chicago: University of Chicago, 199-216. Central Bank of Uruguay. (1981). Circular N o . 1110. Montevideo. (1982). Circular N o . 1125. Montevideo. Corbo, Vittorio, and Jaime de Melo (eds.) (1985). "Liberalization with Stabilization in the Southern Cone of Latin America." World Development 13 (August): 863-1015.
39
capital account transactions and creating incentives to undervalue exports and overvalue imports, in order to bring back foreign exchange through the capital account. In contrast, cushioning a capital inflow temporarily appreciates the exchange rate. This would create incentives to overvalue exports and undervalue imports to escape the effective tax on capital inflows, in much the same way as export incentives in Colombia and Turkey led to "fictional" exports at various times. Since any separation of the markets for foreign exchange should be short lived, there probably would not be time to evolve sophisticated mechanisms to escape the separation of markets. The chief difficulty is likely to be the government's unwillingness to unify the markets when the capital account rate begins to drop below the current account rate. For example, it is often suggested that taxes be placed on short-term inflows, to discourage volatile flows that could prove destabilizing. However, such taxes are easily escaped by sophisticated financial intermediaries, which could offer a combination of long-term deposits and off-balance sheet options ("repos") to buy foreign exchange at shorter intervals.
426
James Hanson
Corbo, Vittorio, Jaime de Melo, and James Tybout. (1986). "What Went Wrong with Recent Reforms in the Southern Cone." Economic Development and Cultural Change 34, no. 3: 607-40. Dawson, Frank G. (1990). The First Latin American Debt Crisis. New Haven: Yale University Press. Diaz Alejandro, Carlos. (1970). "Direct Foreign Investment in Latin America." In C. Kindleberger (eds.), The International Corporation. Cambridge, Mass.: MIT, 309-44. Dooley, Michael. (1988). "Capital Flight: A Response to Differences in Financial Risks." IMF Staff Papers 35 (September): 422-36. Dooley, Michael, and Peter Isard. (1980). "Capital Controls, Political Risk, and Deviations from Interest Parity." Journal of Political Economy 88: 370-84. Dornbusch, Rudiger. (1976). "Expectations and Exchange Rate Dynamics." Journal of Political Economy 84: 1161-76. Edwards, Sebastian. (1984). "The Order of Liberalization of the External Sector in Developing Countries." Princeton Essays on International Finance no. 156. Princeton University. (1985). "Money, the Rate of Devaluation, and Interest Rates in a Semi-Open Economy: Colombia 1968-1972." Journal of Money, Credit and Banking 17: 59-68. (1989). "On the Sequencing of Structural Reform." Cambridge, Mass.: NBER Working Paper no. 3138. Edwards, Sebastian, and Alejandra Cox-Edwards. (1987). Monetarism and Liberalization: The Chilean Experiment. Cambridge, Mass.: Ballinger. Edwards, Sebastian, and Moshin Khan. (1985). "Interest Rate Determination in Developing Countries: A Conceptual Framework." IMF Staff Papers 32 (September): 377-403. Edwards, Sebastian, and Sweder van Wijnbergen. (1986). "The Welfare Effects of Trade and Capital Account Liberalization." International Economic Review 27: 141-48. Fernandez, Roque B. (1985). "The Expectations Management Approach to Stabilization in Argentina during 1976-82." In Corbo and de Melo (eds.), World Development 13: 871-892. F.I.E.L. (Fundacion de Investigaciones Economicas Latinoamericanas). (1989). El Control de Cambios en la Argentina. Buenos Aires: Ed. Manantial. Fisher, Irving. (1930). The Theory of Interest. New York: Macmillan. Fleming, J. M. (1971). Essays in International Economics. London: Allen and Unwin, chap. 12. Frenkel, Jacob. (1982). "The Order of Economic Liberalization: A Comment." In K. Brunner and A. Meltzer (eds.), Economic Policy in a World of Change. Carnegie-Rochester Conference Series on Public Policy. Amsterdam: North Holland, 1982, 99-102. Gros, Daniel. (1988). "Dual Exchange Rates in the Presence of Incomplete Market Separation, Long Run Effectiveness and Policy Implications." IMF Staff Papers 35 (September): 437-60. Hanson, James A. (1974). "Optimal International Borrowing and Lending." American Economic Review 64: 616-30. (1986). "What Went Wrong in Chile." In A. Choksi and D. Papageorgiou (eds.), Economic Liberalization in Developing Countries. London: Basil Blackwell, 227-32.
Opening the capital account
427
(1992). "Opening the Capital Account." Working Paper no. 901, August. Washington, D.C.: World Bank. (1994). "Opening the Capital Account: A Survey of Issues and Results." In Gerard Caprio, Itzak Atyias, James Hanson, et al. (eds.), Financial Reform: Theory and Experience. Cambridge: Cambridge University Press. Hanson, James A., and Jamie de Melo. (1983). "The Uruguayan Experience with Stabilization and Liberalization: 1974-81." Journal of Interamerican Studies and World Affairs 25, no. 4: 563-81. (1985). "External Shocks, Financial Reforms, and Stabilization Attempts in Uruguay 1974-1983." World Development 13 (August): 917-40. Hanson, James A., and Roberto Rocha. (1986). High Interest Rates, Spreads, and the Costs of Intermediation. Washington, D.C.: World Bank, Industry and Finance Series no. 18. Harberger, Arnold. (1982). "The Chilean Economy in the 1970s: Crisis, Stabilization, Liberalization, Reform." In K. Brunner and A. Meltzer (eds.), Economic Policy in a World of Change. Carnegie Rochester Conference Series on Public Policy. Amsterdam: North Holland. Herrera, Juan E., and Juan Morales. (1992). "La Inversion Financiera Externa: El Caso de Chile 1974-1978." Estudios CIEPLAN, no. 35 (September): 103-50. International Monetary Fund. (1985, 1992). International Financial Statistics. Washington, D.C. Jenks, Leland H. (1927). The Migration of British Capital to 1875. London: A. A. Knopf. Khan, Moshin, and Nadeem Haque. (1985). "Foreign Borrowing and Capital Flight." IMF Staff Papers 32 (December): 606-28. Khan, Moshin, and Roberto Zahler. (1983). "The Macroeconomic Effects of Changes in Barriers to Trade and Capital Flows: A Simulation Analysis." IMF Staff Papers 30 (June): 223-82. Krueger, Anne. (1981). "Interactions between Inflation and Trade-Regime Objectives in Stabilization Programs." In W. Cline and S. Weintraub (eds.), Economic Stabilization in Developing Countries. Washington, D.C: Brookings Institute. (1984). "Problems of Liberalization." In Arnold Harberger (ed.), World Economic Growth. San Francisco: ICS Press. (1986). "Problems of Liberalization." In A. Choksi and D. Papageorgiou (eds.), Economic Liberalization in Developing Countries. London: Basil Blackwell, 15-31. Krugman, Paul. (1989). "Market Based Debt Reduction Schemes." In J. Frenkel, Michael P. Dooley, and Peter Wickham (eds.), Analytical Issues in Debt. Washington, D.C: International Monetary Fund. Lanyi, Anthony. (1975). "Separate Exchange Markets for Capital and Current Transactions." IMF Staff Papers 22 (September): 714-49. Larrain, Mauricio. (1989). "How the 1981-83 Chilean Banking Crisis Was Handled." World Bank PPR Working Paper no. 300, December. Little, Ian, Tibor Scitovsky, and Maurice Scott. (1970). Industry and Trade in Some Developing Countries: A Comparative Study. Oxford: Oxford University Press. Lizondo, Jose Saul. (1983). "Interest Differential and Covered Interest Arbitrage." In Pedro Aspe, Rudiger Dornbusch, and Maurice Obstfeld (eds.), Financial Problems of the World Capital Market: The Problem of Latin American Countries. Chicago: University of Chicago Press, 221-40.
428
James Hanson
MacDougall, Ian. (1968). "The Benefits and Costs of Private Investment from Abroad: A Theoretical Approach." Reprinted in R. Caves and H. Johnson (eds.), Readings in International Economics. London: Allen and Unwin, 172-94. McKinnon, Ronald. (1973). Money and Capital in Economic Development. Washington, D.C.: Brookings Institute. (1982). "The Order of Economic Liberalization: Lessons from Chile and Argentina." In K. Brunner and A. Meltzer (eds.), Economic Policy in a World of Change. Carnegie-Rochester Conference Series on Public Policy. Amsterdam: North Holland, 159-86. Michaely, Michael. (1986). "The Timing and Sequencing of a Trade Liberalization Policy." In A. Choksi and D. Papageorgiou (eds.), Economic Liberalization in Developing Countries. London: Basil Blackwell, 41-59. Moran, Theodore. (1974). Multinational Corporations and the Politics of Dependence: Copper in Chile. Princeton, N.J.: Princeton University Press. Mundell, Robert. (1968a). "International Trade and Factor Mobility." Reprinted in R. Mundell, International Economics. New York: Macmillan, 85-99. (1968b). "Capital Mobility and Stabilization Policy under Fixed and Flexible Exchange Rates." Reprinted in R. Mundell, International Economics. New York: Macmillan, 250-71. Ortiz, Guillermo. (1983). "Dollarization in Mexico: Causes and Consequences." In Pedro Aspe, Rudiger Dornbusch, and Maurice Obstfeld (eds.), Financial Problems of the World Capital Market: The Problem of Latin American Countries. Chicago: University of Chicago Press, 71-95. Persson, T , and Lars Svensson. (1985). "Current Account Dynamics and the Terms of Trade: Harberger-Laursen-Meltzer Two Generations Later." Journal of Political Economy 75: 43-65. Reynolds, Clark. (1965). "Development Problems of an Export Economy: The Case of Chile and Copper." In Clark Reynolds and Markos Mamalakis, Essays on the Chilean Economy. Homewood, 111.: Irwin, 203-357. Rippy, J. Fred. (1959). British Foreign Investment in Latin America 1822-1949. Minneapolis: University of Minnesota Press. Rodrick, Daniel. (1987). "Trade and Capital Account Liberalization in a Keynesian Economy." Journal of International Economics 124: 113-29. Samuelson, Paul. (1948). "International Trade and the Equalization of Factor Prices." Economic Journal, 58: 163-84. (1965). "Equalization by Trade of the Interest Rate along with the Real Wage." In Robert E. Baldwin et al. (eds.), Trade, Growth and the Balance of Payments: Essays in Honor of Gottfried Haberler. Chicago and Amsterdam: Rand McNally and North-Holland, 35-52. Sargent, Thomas. (1983). Comment on Guillermo Ortiz, "Dollarization in Mexico: Causes and Consequences." In Pedro Aspe, Rudiger Dornbusch, and Maurice Obstfeld (eds.), Financial Problems of the World Capital Market: The Problem of Latin American Countries. Chicago: University of Chicago Press, 95-106. Schmidt-Hebbel, Klaus. (1988). "Consumo e Inversion en Chile (1974-1982): Una Interpretacion Real del Boom." In F. Morande and K. Schmidt-Hebbel (eds.), Del Auge a la Crisis de 1982. Santiago: ILADES, 135-58. Sigmund, Paul. (1980). Multinationals in Latin America: The Politics of Nationalization. Madison: University of Wisconsin Press.
Opening the capital account
429
Stallings, Barbara. (1987). Banker to the Third World: U.S. Portfolio Investment in Latin America 1900-1986. Berkeley: University of California Press. Stockman, Alan. (1988). "On the Roles of International Financial Markets and Their Relevance for Economic Policy." Journal of Money, Credit and Banking 20: 531-49. Svensson, Lars. (1988). "Trade in Risky Assets." American Economic Review 78: 375-94. Tanzi, Vito, and Mario Blejer. (1982). "Inflation, Interest Rate Policy, and Currency Substitution in Developing Economies: A Discussion of Some Major Issues." World Development 10 (September): 781-89. Valdes-Prieto, Salvadore. (1994). "Financial Liberalization and the Capital Account in Chile 1974-1984." In Gerard Caprio, Itzak Atyias, James Hanson, et al. (eds.), Financial Reform: Theory and Experience. Cambridge: Cambridge University Press. Vittas, Demetri, and Augusto Iglesias. (1992). "The Rationale and Performance of Personal Pension Plans in Chile." Washington, D.C.: World Bank PPR Working Paper no. 867. World Bank. (1984). Colombia: Economic Development and Policy under Changing Conditions. Washington, D.C. (1985). World Development Report, 1985, Washington, D.C. (1990). World Debt Tables, 1990-91, Washington, D.C.
Index
adjustable peg (AP), 277-78 arbitrage transactions, 74 Argentina: capital inflows and increase in private consumption, 343; sequencing of reforms in, 9; stabilization plan of, 2324, 25, 26; stock market performance in, 347, 352; tariffs-cum-subsidies and, 13 Asia: lessons for Southeast, 111; trade reforms in, 1; see also names of individual countries assets held by public, 218 Australia: capital control effectiveness in, 291; capital flow and, 136, 270; choice of exchange rate regime for, 282-84; deregulation in (1983-85), 294, 296; domestic financial deregulation in, 284-86; economy of (compared to New Zealand), 266, 269; floating exchange rate and, 134; monetary and exchange rate policy of, 274-82, 286-87, 289-91, 293; regulatory environment in, 269-74; responsibility for monetary policy in, 274 bank failures from sudden reversal of capital inflows, 379 Barro, R., 50-52 basket pegs for currencies of East Asian countries, 57-58 Belgium, exchange rates and, 143 Blundell-Wignall, A., 283-84 bonds, yields on, 173-74 Brazil, capital inflows and increase in private consumption, 343 Bretton Woods Conference, 71, 72, 120 budget deficits, financing of excessive, 37 business cycles, 174 Canada: capital controls and, 122, 142; floating exchange rate and, 134; policy mistake of, 24
431
capital account: advantage to government of a closed, 249; countries moving toward open, 135; growing importance of, 73; open, 392-93 capital account controls, to tax domestic capital at higher-than-interaational rates, 389 capital account convertibility, argument for, 84 capital account openness and liberalization, 84, 159, 210-16; assessing, 5; conditions for undertaking, 85; impact of (for industrialized countries), 422; income distribution and, 421; instability and, 393-95; logic of, 76-79; main benefit of, 419; measures of, 201-2; in OECD countries (1961-73), 122-25; policy effectiveness and, 390, 392-93; preconditions for, 395-97; risk and, 384-85; sequence problems and, 398-401, 423, 424; taxation and, 385-90; traditional analysis of costs and benefits of, 383-84 capital controls, 76; ability to tax and dismantling of, 79; arguments for, 77, 88; categories of, 76; in crisis management mode (1967-73), 125; dismantling of national, 81; EC countries' phasing out of, 136; economy's vulnerability to shocks and, 78; effectiveness and purpose of, 3; floating exchange rates as alternative to, 125-26; ineffectiveness of, 131; inefficiencies and undesirability of, 82, 86; macroeconomics and, 136-37, 141, 148-52; for monetary policy, 135; OECD and, 119, 125-26, 131-32, 132-36; relaxation of, 74; risk of moral hazard and, 78-79; structural considerations and, 146^8 capital flows and movements, 1-2, 5, 71, 216-24, 253-57; from abroad, 81; ac-
432
Index
capital flows and movements (cont.) counting of, 341-43; authorities' time to react to external environment and, 249; common European and Latin American experiences with, 175; economic instability and, 77; effects of unrestricted, 131; expansion in scale of, 74; fixed-butadjustable exchange rate and high volume of, 94; historical comparison of German and Korean, 257, 259-60; indicators of, 200-2; ineffectiveness of controls on, 131; irreversibility of decision to remove, 145; into Latin America, 11-12; monetary and exchange policies and, 79-82; OECD controls on, 119-20, 152-55; real exchange rates and, 9-14, 251-53; speed of private sector's reaction to policy decisions and, 247 capital imports, 24-27 capital inflow, 160, 162, 164; anatomy of Latin American, 343, 346-47; concerns about, 376; external factors in South American, 368-69, 371-74; problem of, 158, 159; reform and, 339; sudden reversals of, 379; tightened fiscal policy to curb inflationary impact of, 377-78 capital markets: globalization of, 81; importance of access to international, 249; integration of, 4; Korean open, 254; liberalized, 119; sources of imperfection in international, 37 capital transactions, challenge of establishing freedom for, 84 central banks, 99, 111; independent monitary policies of Central American, 312; response to increased capital inflows by, 342 Chile: capital account liberalization in, 408-16; capital inflows and increase in private investment in, 343; capitalized pension system in, 415; capital mobility and, 5, 211; debt-equity swap program in, 415; exchange rates and inflationary inertia in, 327-30; exchange rate-based stabilization and, 9, 25; illegal capital flight from, 408; inflation in, 408, 410, 411, 416; real appreciation problem and, 13; recession in, 410-11, 413-15; response to capital inflows and, 378-79; sequence of reform in, 9; stabilization plan of, 23, 25; stock market performance in, 347, 352; sustained interest differentials in, 211-12; vehicles for capital inflows to, 412 code of conduct, 71, 74-76 Colombia: capital mobility and, 5; response
to capital inflows and, 378-79; use of domestic bonds by, 12-13 Comecon countries, 181 Costa Rica: credit growth and deficit ratio for, 317; devaluation in reaction to oil shocks and, 313-14, 318; recovery from crisis, 319-20 credibility, 158, 176; disinflation policies and, 326-27; fixed exchange rates and, 304-6; policy effectiveness and, 377 currency: dangers of prematurely linked, 111; drawback to nations having a common, 101; overvalued, 21; permanent peg vs. the float, 275-76 currency market turmoil, 86-88 currency risk, 100 currency substitution, 387 current account, convertibility of, 74, 84 current account deficits, 266 current account surpluses, political conflict with trading partners and, 248 debt, Italian, 178; see also budget deficits debt crisis (1982), 36 De Gregoria, I, 176-77 devaluation: expectations of, 399-400; in Latin America, 301, 313-14, 316-17, 319-20 Devarajan, S., 305 developing countries, exchange rates of, 6-7 disinflation, 20 domestic financial deregulation, 74, 280-82; Australian, 284-86; macro consequences of, 286 domestic policies that recognize constraints on national freedom of action, 88 Dominican Republic: devaluation in, 31617, 320; end of fixed rates in, 320; increase in foreign debt (1978-82), 318; independent monetary policy of, 312; inflation in, 307 Dooley, M. P., 253-54, 256, 257 East Asian countries, basket pegs for currencies of, 57-58 economic policy, failures of, 121 economic reality, international financial code of conduct and, 74-76 El Salvador: deficit ratio of, 316; devaluation in, 316-17, 320; end of fixed rates in, 320; increase in foreign debt (1978-82), 318; independent monetary policy of, 312; inflation in, 307, 308 European Economic Community (EEC), 93-94, 98
Index European Monetary System (EMS), 83-84, 109; exchange rate mechanism of, 6, 24 European Monetary Union (EMU), 98101, 106, 108-9, 111; German unification and,194 exchange controls, abolition of (1990), 93 exchange rate action, policy dilemmas and, 6-7 exchange rate adjustment to stabilize an open economy, 101 exchange rate-based stabilization, 9; problem with, 175-76 exchange rate mechanism (ERM), 93, 108-10; crisis of (1993), 6; EMS and, 6, 24, 93, 95-96, 108-9; fixed-but-adjustable (within EEC), 94, 95; withdrawal of Italy and U.K. from, 87 exchange rate policy, 178 exchange rate regime: adjustable peg, 211-18', choosing an, 275; crawling peg, 278-79; insulating properties of, 276, 283-84 exchange rates: choice between price stability and stability of, 252; of developing countries, 6-7; as disciplinary devices, 20-21; in the EEC, 94; effect on trade of variability in, 40-42; fixed-but-adjustable, 94; fixed vs. flexible, 37-42, 45-49; fixed nominal, 249; to insulate against shocks, 276; legal restrictions on capital flows to help manage, 248; as nominal anchors, 42, 45-46; pegged, 302; stabilization and, 111; stabilizing inflation through, 157; volatility of, 43t, 44t; see also specific kinds of exchange rates, e.g., fixed exchange rates exchange regime, exchange restrictions and, 159-60 exchange systems, efficiency of completely free, 86 financial markets: deregulation of, 132; floating exchange rate and, 47; growing integration of, 74 fiscal balance under EMU, 99-100 fiscal policy, 79; Central American countries' countercyclical, 318 fiscal solvency, 24 fiscal transfers to depressed regions, 39 fixed exchange rate regime, 72n fixed exchange rates, 8-9, 20, 104, 151-52, 301; advantage of, 40, 42, 45^6; antiinflationary programs and, 7; categories of countries with, 6n; credibility and, 304-6; in developing countries, 6; finan-
433 cial constraints imposed by, 314; low inflation and, 303; stability of, 251 fixed parity regime, 417-18 Fleming, J. M., 104 flexible exchange rates, advantages of, 38-40 floating exchange rates, 104, 118, 131; as alternative to capital controls, 125-26; effectiveness of domestic monetary policy and, 392 foreign economic market forces, 80-81 France: devaluation of the franc (1969), 124; ERM and, 111, 158; capital inflows and, 160, 164; inflation and real exchange rates in, 164, 166-67; increase in foreign exchange reserves of, 160, 161; real interest rates and, 172-74 Frankel, J. A., 41 free-floating exchange rates, 35, 47 friction, models of, 235 Friedman, M., 103 Fukas, M., 231-32t, 233 fundamentals valuation equation, 217-18 General Agreement on Tariffs and Trade (GATT), 74; free trade and, 121; Uruguay round of, 1,2 German Monetary Union: introduction of deutsche mark into GDR and, 184, 185-86; repercussions on Federal Republic of, 188, 190, 192; role of public transfers from western to eastern Germany, 187-88; role (in reform) of subsidized private investment and, 186-87 Germany, 94; Bundesbank commitment to monetary stability, 182; capital controls in, 5, 122, 126, 134, 260-62; capital mobility and, 257, 259-60; currency reform of (1984), 182; current account adjustment by, 251; introduction of free-market system into GDR, 182; Korean interest in unification of, 193-94; learning from the experience of, 247-48; the mark (1969), 125; political mistakes and, 192-93; reasons for shock therapy for GDR, 182-84; sterilization of trade imbalances by, 251; strategic elements in restructuring economic system in, 184-85 Gordon, D , 50-52 Gramm-Rudman-Hollings legislation, 37 Gregory, R. G, 283-84 Guatemala: devaluation in, 316-17, 320; end of fixed rates in, 320; increase in foreign debt of (1978-82), 318; independent monetary policy of, 312; inflation in, 307, 308
434
Index
Haque, N., 253-54 Hawtrey, K., 285 Honduras: devaluation in, 316-17, 320; end of fixed rates in, 320; increase in foreign debt of (1978-82), 318; independent monetary policy of, 312; inflation in, 307, 308 IMF, see International Monetary Fund impulse response functions of reserves, 374 indexation, 320, 322-27 inflation, 275-76; Australia/New Zealand comparison, 266; control of capital inflows from abroad and, 81; distinction between traded goods and non-tradables types of, 176; dynamics of domestic, 324-27; exchange rate to stabilize, 157, 320; indexation and, 320, 325; inertia of domestic, 324-25; international transmission of, 308-9; in Korea, 206, 209, 210; in Latin America, 307-10, 320; real exchange rates and, 164, 166-69; sources of nontradeables, 177 inflation tax, 386-88, 420-21 interest rate parity, 201-3 interest rates, 391; problems with manipulating, 421 International Monetary Fund (IMF), 2, 71, 82, 120, 152-53; devaluations encouraged by, 302; stabilization programs sponsored by, 157 international trade, efforts to liberalize, 1 investment ratio, savings ratio and, 4 Ioannidis, C , 103 Ireland, 111 Italy, 177; capital controls and, 142; capital inflows and, 163, 164; debt of, 178; ERM and, 111, 158; foreign exchange reserves of, 162; inflation and real exchange rates, 164, 166; private loans in balance of payments of, 162; real interest rates and, 172-74 Ito, T., 234 Japan: effects of interest rate differentials on, 240-41; floating exchange rate and, 134; more liberal foreign exchange law (1980), 135; multinational trading companies (MTC) in, 6, 230, 241-44 Japanese capital controls, 5, 134, 141, 22930, 230-33, 233-35; deregulation of, 233; determinants of (1970s), 235-37, 239-40, 244; effects on trade balance of, 240-41, 241-45; success in reducing capital movements caused by trade imbalances, 245; trade balance components that affect, 237, 239-40; trade patterns and, 243
Keynesian open macroeconomic model, 210-16 Korea, 119, 199-200, 224-25; avoidance of overborrowing by, 37; capital controls in, 260-61; capital mobility and, 5, 252, 254-56, 257, 259-60; current account adjustment by, 251; demographics of, 31; economic deceleration of, 204-5; economic growth of, 203; as an economy in transition, 247-48; effective capital controls in, 203; effects of foreign capital inflow on, 224; further financial liberalization for, 48; housing construction in, 206; inflation in, 206, 209, 210; instability of domestic economic conditions in, 225; land prices in, 209, 217-18; open capital markets of, 254; private capital flows directed by government in, 255-56; real asset speculation in, 224-25; real estate boom in, 11; recent developments in, 28-32; sterilization of trade imbalances by, 251, 256; stock prices in, 206, 209, 217; successful stabilization by (1980s), 27-28; upset stabilization plan of, 25; yields from investment in stocks in, 206, 209 Korea's capital account, 200-3; impact of liberalization of, 212-16 Kwack, S., 255-56 labor mobility, 39, 102, 282 lags between relative price changes and changes in trade balance, 249 Latin America: balance of payments accounts in, 343; capital account liberalization in, 401-3, 416-25; capital flight from, 3-4; capital flow into, 11-12, 339, 357, 360-61, 407; current account deficits in, 366; debt crisis in (1982), 2; devaluations in, 301, 313-14, 316-17, 319-20; domestic interest rates in, 211; economic growth in (1991), 353; effects of capital inflows on economies of, 340-41; exchange rate policies in, 320-21; fixed exchange rates in, 306-12, 320; inflation in, 307-10, 320, 353, 357; interest rate spreads in, 353; low inflation nations of, 303; net capital flows to and from United States, 366; new kind of stabilization programs in (1970s), 157, 160; official international reserves for countries of, 347; open economy macro model and Southern Cone experience in, 210-12; rates of return differential in, 347, 352; real effective exchange rates in countries of, 347; real exchange rates and international reserves
Index in, 368-69, 371; sterilization policies in, 361, 376, 378-79; trade reforms in, 1; see also names of individual countries liberalization, interactions between speed of current and capital accounts and, 399-401, 425 Liverpool multilateral world model, 97, 103, 108, 109, 111-16 Maastricht fiscal criteria, 99, 100 Maastricht Treaty, 37, 194 McKinnon, R., 37-38 macroeconomic policy, 81, 132-33 macroeconomic stability: loss of exchange and interest rate stabilizers in EEC and, 96; EMU to achieve, 101-8 Malaysia, currency pegging by, 57 market average rate (MAR) system, 58 Marwaha, S., 103 Masson, P., 105-7 Mathieson, D., 253-54, 256, 257 Mexico: capital inflows and increase in private investment in, 343; exchange rate adjustment by, 13; exchange rates and inflationary inertia in, 330-32; inflation in, 9; stock market performance in, 347, 352 Minford, P., 103, 105, 107 monetarism, 36 monetary independence, obtaining, 421-22 monetary policies: capital flows and exchange policies and, 79-82; design of, 279-80; discretion and alternative rules for, 56-66; floating interest rates and, 392; holding GNP steady and, 53; importance of independent, 38; insulating from populist pressures, 50; monitoring commitment to nominal anchor for, 56; need for consistency across countries, 83; nominal anchors for, 45, 48-49, 52-56 monetary policy variables that are candidates for nominal anchors, 56 monetary regimes: choices for a semi-open country, 35, 58-66; fixed and flexible exchange rates and, 37-42, 45-49 monetary stability, 120 monetary unions, 8 money growth: impact of, 250; precommitment to slow, 49 Montiel, P., 253-54 multimod model, 104, 105-9 Mundell, R. A., 37-39, 104 Netherlands, capital contiols and, 142 newly industrialized countries (NICs), exchange rate regimes in, 8 New Zealand: capital flow and, 136, 270;
435 choice of exchange rate regime for, 282-84; deregulation in (1983-85), 294, 296; economy of (compared to Australia), 266, 269; floating exchange rate and, 134; inflation target in, 293; monetary and exchange rate policy, 274-82, 292, 293-94; openness of economy, 269; regulatory environment in, 269-74; responsibility for monetary policy in, 274 nominal GNP rule, 55 North American Free Trade Area Agreement, 24 Norway, capital flow liberalization and, 136 optimal currency areas, 8 Organization for Economic Cooperation (OECD), 119-22,152-55 overborrowing, 36-37 Panama, fixed nominal exchange rate regime of, 306 Park, W, 225 par value system, 72n, 73 policy decisions, capital mobility and speed of private sector's reaction to, 247 policy rules vs. discretion, 49-52 Portugal, 111 price level rule, 54-55 prices, 166-7, 169 price stability, 53; choice between exchange rate stability and, 252; pursued by domestic means, 99 productivity growth, 48 real asset speculation models, 216-24 real exchange rate (RER): capital mobility and variability of, 251-53; capital flows and, 9-14; common European and Latin American experiences with, 175; inflation and, 157, 164, 166-69; real interest differential and, 202; resistance to appreciation of, 248; and terms of trade, 21-23 real interest differential, 202 real interest rates, 169, 172-73, 202 relative consumer prices, 166-67, 169 resources, monetary policies and movement of, 80 Rodrik, D., 305 savings ratio, investment ratio and, 4 semi-open economies, 4, 19-20; capital imports and, 25-27; real exchange rates and terms of trade, 21-23 sequencing of reform, 9-14, 85-86 shocks, 109; capital controls and, 78; ex-
436
Index
shocks (font.) change rates to insulate against, 276, 283-84; export, 244; monetary responses to, 105; nominal devaluations and, 305; oil, 313-14, 318; real wage, 177; spending, 176-77; supply, 104; terms of trade, 318 Singapore, capital mobility and, 5 Spain, 111; balance of payments data for, 160, 161f; capital account liberalization and, 159-60; exchange rate mechanism of EMS and, 158; foreign exchange reserves of, 162; goods trade liberalization and, 159; inflation and real exchange rates, 164, 166-68; real interest rates and, 173-74; real wage shocks in, 177 stabilization, 111; exchange rate-based, 9; failures of, 24-27; fixed exchange-rate based programs of, 321; semi-open economies and, 19-20 Sweden, capital flow liberalization and, 136, 143 Switzerland: capital controls and, 126, 143; floating exchange rate and, 134 Symansky, S., 105-7 Takagi, S., 231-32t, 233 taxation: of capital, 386, 389-90; of income from capital and financial assets, 420 terms of trade, real exchange rate and, 21-23 Thailand, currency pegging by, 57 trade balance, lags between relative price changes and changes in, 249 trade deficit, sterilization of, 250
trade liberalization: impact of pace of, 400; real exchange rate behavior and, 9-11; semi-open economies and, 19, 20 two-sector general equilibrium portfolio balance model, 218-24 unemployment, in Australia and New Zealand, 266 United Kingdom: capital controls and, 135, 142; devaluation of the pound (1967), 124; ERM and, 111; exchange rate regime of, 134; policy mistake of, 24 United States: capital controls and, 122, 126, 141; capital flows with Latin America, 366; the dollar and gold in (1971), 125; floating exchange rate and, 134; inflation in, 308 Uruguay: capital account liberalization in, 403-8; inflation in, 407; sequencing of reform in, 9; stabilization plan of, 23-24, 25 variance decomposition, 374 Venzuela, exchange rates and inflationary inertia in, 332-35 wealth of the public, measurement of, 219 Wei, S., 41 Werner Report, 94 Williamson, 1, 306 yen-dollar fluctuations, 47 yield differentials, margin for national,74 Yoo, Y, 225
