The Politics of Delegation

THE POLITICS OF DELEGATION

BOOKS OF RELATED INTEREST

Church and State in Contemporary Europe: The Chimera of Neutrality edited by John T.S.Madeley and Zsolt Enyedi The Enlarged European Union: Diversity and Adaptation edited by Peter Mair and Jan Zielonka The Swiss Labyrinth: Institutions, Outcomes and Redesign edited by Jan-Erik Lane Europeanised Politics? European Integration and National Political Systems edited by Klaus H.Goetz and Simon Hix Recasting European Welfare States edited by Maurizio Ferrera and Martin Rhodes The Changing French Political System edited by Robert Elgie Compounded Representation in West European Federations edited by Joanne B.Brzinski, Thomas D.Lancaster and Christian Tuschhoff Politics and Policy in Democratic Spain: No Longer Different? edited by Paul Hey wood Britain in the Nineties: The Politics of Paradox edited by Hugh Berrington Crisis and Transition in Italian Politics edited by Martin Bull and Martin Rhodes

Southern European Welfare States: Between Crisis and Reform edited by Martin Rhodes The Euro-Mediterranean Partnership: Political and Economic Perspectives edited by Richard Gillespie The State in Western Europe: Retreat or Redefinition? edited by Wolfgang C.Müller and Vincent Wright The Regions and the European Community edited by Robert Leonardi The Regional Dimension of the European Union edited by Charlie Jeffery National Parliaments and the European Union edited by Philip Norton (new in paperback)

THE POLITICS OF DELEGATION Editors

MARK THATCHER ALEC STONE SWEET

FRANK CASS LONDON • PORTLAND, OR

First published in 2003 in Great Britain by FRANK CASS AND COMPANY LIMITED Crown House, 47 Chase Side, Southgate, London N14 5BP, England This edition published in the Taylor & Francis e-Library, 2005. To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to http://www.ebookstore.tandf.co.uk/. and in the United States of America by FRANK CASS c/o International Specialized Book Services, Inc. 920 NE 58th Avenue, Suite 300, Portland, OR 97213–3786 Website http://www.frankcass.com/ Copyright © 2003 Frank Cass & Co. Ltd British Library Cataloguing in Publication Data The politics of delegation 1. Delegation of powers 2. Administrative agencies I. Thatcher, Mark II. Stone Sweet, Alec 320.9 ISBN 0-203-00598-8 Master e-book ISBN

ISBN 0 7146 5561 9 (Print Edition) ISBN 0 7146 8443 0 (paper) Library of Congress Cataloging-in-Publication Data The politics of delegation/editors, Mark Thatcher, Alec Stone Sweet. p. cm. Includes bibliographical references (p.) and index. ISBN 0-7146-5561-9 (cloth)—ISBN 0-7146-8443-0 (paper)) 1. Delegation of powers—Europe. 2. Representative government and representation— Europe. 3. Legitimacy of governments—Europe. 4. Delegation of powers—United States. I. Thatcher, Mark. II. Stone Sweet, Alec. III. Title. JN12.P632 2003 320.44′04–dc211 2003009976 This group of studies first appeared in a Special Issue of West European Politics (ISSN 01402382), Vol.25, No.1 (January 2002). All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form, or by any means, electronic, mechanical, photocopying, recording or otherwise without the prior written permission of Frank Cass and Company Limited.

Contents

Foreword

Theory and Practice of Delegation to Non-Majoritarian Institutions Mark Thatcher and Alec Stone Sweet Delegation to Supranational Institutions: Why, How, and with What Consequences? Jonas Tallberg Rational Fictions: Central Bank Independence and the Social Logic of Delegation Kathleen R.McNamara Constitutional Courts and Parliamentary Democracy Alec Stone Sweet Institutional Choice and Bureaucratic Autonomy in Germany Marian Döhler Delegation to Independent Regulatory Agencies: Pressures, Functions and Contextual Mediation Mark Thatcher The Unanticipated Consequences of Creating Independent Competition Agencies Stephen Wilks with Ian Bartle Judicial Delegation Doctrines: The US, Britain, and France Martin Shapiro Learning from the Americanists (Again): Theory and Method in the Study of Delegation Mark A.Pollack

vii

1 20

40 69 90 111

131 152 174

Abstracts

191

Notes on Contributors

195

Index

196

Foreword

The contributions to the present volume were discussed at meetings at Nuffield College, Oxford (November 1999), and at the Robert Schuman Centre, the European University Institute, San Domenico di Fiesole (March 2001). We are grateful to Nuffield College, STICERD (Suntory and Toyota International Centres for Economics and Related Disciplines) at the London School of Economics, and the Robert Schuman Centre, for their generous funding of this research project. We thank Maria Kampp for assistance with editing. Mark Thatcher also gratefully acknowledges ESRC grant number no. L216 25 2007 within the ESRC Future Governance programme and wider support from the Centre for the Analysis of Risk and Regulation. We also express our thanks to an anonymous reviewer of the draft manuscript and the editors of West European Politics. The origins of the project lie with one of the founding editors of West European Politics, Vincent Wright. As with so many other creative ideas and enjoyable projects, the idea of the book began through a good lunch with him. Sadly, he was not able to see the completed project. Vincent Wright provided remarkable stimulus and inspiration, as well as great generosity. The volume is dedicated to his memory.

Theory and Practice of Delegation to NonMajoritarian Institutions

MARK THATCHER and ALEC STONE SWEET A transformation in governance has swept across Western Europe. During the past halfcentury, states, executives, and parliaments have empowered an increasing number of non-majoritarian institutions (NMIs)1 to make public policy. In the fields of utility regulation, telecommunications, antitrust, and media pluralism, and even in the provision of health and welfare benefits, myriad independent regulatory bodies have been created and become the loci for making new rules, or applying existing ones to new situations, at the national level. At the supranational level, central bankers, insulated from direct political control, set monetary policy. In Brussels, European Commission officials propose legislation and enforce ever wider European Union regulation. In Luxembourg, the Court of Justice controls member state compliance with European law, reviewing the lawfulness of activities of national parliaments, governments, and administrators. The ongoing exercise of authority by non-majoritarian bodies is today central to governance in a growing number of policy domains. This volume seeks to address the sources, consequences, and dynamics of delegation to NMIs in Western Europe. Here we introduce the core themes and issues raised by the project, while each subsequent chapter explores the politics of delegation more specifically in different polities, sectors, and institutional settings. We begin by discussing how the group has collectively chosen to define and conceptualise our topic; we then present theoretical materials used by political scientists to explain delegation to non-majoritarian institutions. These ideas developed in the United States during the 1980s, primarily in research on the relationship between Congress and American regulatory agencies (see Pollack, this volume).2 We also briefly discuss relevant ideas found in ‘new-institutional’ organisational sociology.3 Following from the seminal work of Giandomenico Majone4 and Mark Pollack,5 a literature that uses or otherwise engages American delegation theory to conceptualise or explain aspects of supranational politics in the EU has emerged, and is becoming a standard reference point.6 Nonetheless, comparative research on the sources and consequences of institutional innovation through delegation to NMIs has been, until recently, scarce.7 Thereafter, we examine the politics of delegation to NMIs in Europe the decision to delegate, the institutional design of delegation and the consequences of delegation, in light of the project’s findings.

The politics of delegation

2

CONCEPTS AND THEORY The title of this volume, The Politics of Delegation, would seem to be relatively selfexplanatory, yet commonplace words do not always mean the same things to different people. Our project is concerned with politics, by which we mean public governance. Although explicit acts of delegation can constitute private systems of governance (for example, those of the modern corporation), and although private domains are full of politics, we have chosen to limit our focus to delegation through public authority. We nonetheless conceive of governance in a broad, generic way. By governance, we mean the processes through which the rule systems in place in any human community are adapted, on an ongoing basis, to the needs and purposes of those who live under them.8 When governments draft legislative bills, and when legislators debate, amend, and adopt statutes, they act to (re)construct legal regimes, and thereby to govern. When administrative officials, operating under a grant of statutory authority, interpret in order to apply the law in concrete situations, they too govern. Of course, in a very important sense, all public governance operates by way of delegation in Europe. The standard model of parliamentary democracy, for example, has it that (a) the (sovereign) people bestow authority on legislators by elections, (b) ministers derive their collective powers from parliament, and (c) the extent of an administrator’s discretion is determined by statute, as controlled by courts. In liberal democratic theory, explicit acts of delegation legitimise the exercise of public authority. We have chosen to narrow our focus further, to non-majoritarian institutions, which we define as those governmental entities that (a) possess and exercise some grant of specialised public authority, separate from that of other institutions, but (b) are neither directly elected by the people, nor directly managed by elected officials. We exclude from consideration state powers organised within the bureaucracy, when the exercise of such powers are placed under the direct control of ministers and the civil service. We do not exclude a specialised organ or agency that may be linked to a ministry in certain formal ways, so long as that body is not merely a department or administrative office of a larger bureaucratic entity. As this volume shows, such institutions take a wide diversity of structural forms in Europe, including independent regulatory agencies, specialised courts, standard-setting boards, central banks, the European Commission, and so on. Given our substantive concerns, we define delegation as an authoritative decision, formalised as a matter of public law, that (a) transfers policy making authority away from established, representative organs (those that are directly elected, or are managed directly by elected politicians), to (b) a non-majoritarian institution, whether public or private. The Principal-Agent Approach The ‘Principal-Agent’ (P-A) framework, developed in American political science from materials found in organisational and ‘transaction cost’ economics,9 dominates research on the topic of delegation to non-majoritarian institutions. (Although different terms are sometimes used in the various strains of this research, we ignore these distinctions here, focusing on the main elements of P-A approaches; moreover, because Mark Pollack’s contribution to this volume, among other things, surveys the development of the P-A approach in research on American politics, we discuss only the basics here.) In some

Theory and practice of delegation to non-majoritarian institutions

3

respects, this dominance is hardly surprising, since the framework was explicitly developed to account for delegation. The framework has proved popular, not least because it offers the analyst readymade, appropriate concepts, and focuses attention on some of the important empirical questions about which most political scientists interested in the topic are likely to be concerned. That said, the P-A framework as it presently exists is just that—a framework. As causal theory—that is, as an integrated body of concepts, operationalisable variables, and testable propositions—it remains incomplete. Indeed, as two proponents of the approach have recently noted, ‘scholars are only beginning to scratch the surface of what needs to be done to establish the theoretical validity of the transaction cost approach’10 Further, hypotheses generated from interpretive case studies and formal models of delegation have not always found support in subsequent research designed to test them.11 It is our view that the P-A construct, a general, relatively flexible, and established approach to thinking about delegation, could not be ignored in research on NMIs in Europe. The directors of this project did not impose (for example, by requiring the adoption of) P-A analysis on contributors. Instead, we asked those critical of the approach to indicate to our readers the source of their objections. Further, some contributors sought explicitly to supplement, or replace, P-A ideas with concepts found in other approaches.12 In all cases, we agreed that important empirical aspects of the politics delegation should be prioritised, not just theory. Functional Logics of Delegation For our purposes, principals are those political officials who use their authority to establish non-majoritarian institutions through a public act of delegation. Agents are those who govern by exercising delegated powers. By assumption, principals are initially in control, in the strict sense that the precise terms of the agent’s remit are a matter of institutional design, and the authority to constitute or not to constitute agents falls within the principals’ jurisdiction. Because principals are willing to pay the costs of delegation (for example, time and deployment of resources to establish the new institution and to monitor its activities), they must have found it in their interest to do so. Indeed, it must be that the expected benefits of delegation outweigh the costs. Put simply, delegation is functional for principals. One can be more specific. The more common rationales for delegation from legislators to agencies, and from nation states to international organisations, are well known. Principals constitute agents to help them: • resolve commitment problems (agents are expected to work to enhance the credibility of promises made, either between multiple principals, or vis-à-vis principals and their constituents, given underlying collective action problems); • overcome information asymmetries in technical areas of governance (agents are expected to develop and employ expertise in order to produce, or help principals produce appropriate public policy); • enhance the efficiency of rule making (agents are expected to respond to relatively specific problems and issues that arise, while principals set and then update the more general terms of policy);

The politics of delegation

4

• avoid taking blame for unpopular policies (agents are expected to maximise policy goals that principals know may sometimes be unpopular with important societal groups). Of course, for any specific act of delegation, these and other rationales may overlap, or be one and the same. Principals can realise the benefits of delegation only by granting discretion to the agent, that is, through sharing some of their authority to govern. We will define discretion more carefully shortly. For now, let us agree that principals know that agents are likely to develop their own interests—including an interest in producing the best policies, given their own understanding of the nature of the policy problem and the environment—and that these interests may conflict with those of the principals. Simplifying, the extent to which an agent actually does generate outcomes that are ‘different from the policies preferred by those who [have] delegated power’,13 is registered as an ‘agency loss’. Principals can reduce agency losses in advance by conferring restrictive powers on the agent, or on an ongoing basis by monitoring and reining in the agent. Institutional design is an exercise in choosing from a menu of both ex ante and ex post (that is, ongoing) controls.14 Yet, because the benefits of delegation decline the more the principal limits the agent’s discretion, the trick is to delegate just the amount of power to enable agents to achieve desired outcomes with minimal agency loss. After all, legislators and member states are not required to delegate; they could govern, through statutory, administrative, and judicial means, without creating new agents. Contributors to this volume were asked to confront questions of institutional design. Why, and on what terms, do relevant acts of delegation occur? What systems of oversight are constructed to monitor the activities of agents? Through what mechanisms can decisions taken by agents be quashed? Agency Discretion and Control In the American literature the question of how best to define and operationalise the gap that (inevitably) develops between what principals want from agents and what agents actually do remains an open one. Underlying the debate are differing views of the nature of discretion. In this volume, we conceive of this gap in terms of a theoretical ‘zone of discretion’. This zone is constituted by (a) the sum of delegated powers (policy discretion) granted by the principal to the agent, minus (b) the sum of control instruments, available for use by the principals to shape (constrain) or annul (reverse) policy outcomes that emerge as a result of the agent’s performance of set tasks. The zone of discretion can be defined without reference to the policy preferences of the principal or the agent, as we have just done. But for a number of reasons we expect such preferences to be fundamental to the dynamics of the relationship between discretion and control. First, if principals create an agent in order to realise certain policy objectives, then the distribution of policy preferences among them, at the ex ante moment of delegation, will help to determine how the zone of discretion will be constructed. If principals wish to commit their successors to the goal of low inflation, for example, they may decide to create a central bank that is more or less insulated from interference by future elected officials. If, faced with high levels of uncertainty in a given domain, they hope to perfect the terms of regulatory policy—as problems emerge and evolve—they may seek to

Theory and practice of delegation to non-majoritarian institutions

5

construct effective ex post controls over their agent. If they desire the smooth implementation of product standards, they may require that the agent consult with those who are being regulated. Thus, functional logics can generate relatively straightforward and testable propositions. The more principals seek to pre-commit themselves to specific policy outcomes, for example, the more powers they will delegate to an agent, and the weaker will be ex post mechanisms of control. In contrast, the more principals seek a rich range of policy alternatives from which to select, on an ongoing basis, the stronger will be the ex post mechanisms of control. Our point is not that there is one functional logic of delegation, but diverse ones, and that this diversity is partly generated by the preferences and objectives of principals at the moment of constituting the agent. Second, once an agent begins to exercise decision making powers, the extent to which ex post instruments of control will actually be activated will be partly related to the principal’s preferences over outcomes. The smaller the zone of discretion, the greater the agent’s interest will be in monitoring and anticipating the principal’s reactions to activities, to the extent of the fear, or wish to avoid, having decisions overturned. Third, in many situations, principals are not unified, which increases the complexity of dealing with changes in preferences. Composite principals—that is, a principal comprised of multiple actors whose collective makeup changes periodically through, for example, elections—may not possess stable, coherent preferences over time. Instead, they may be competitive with one another over some or many issues, as when member state governments in the EU disagree on matters of policy that fall within the agents’ mandate. Even more complicated, the initial act of delegation may parcel out—among multiple actors—the functions normally associated with that of the principal. One organ may possess the power to quash a decision taken by the agent, while another has staffing and funding authority. In such cases, the linkages between the principals’ policy preferences, the agent’s performance, and the principals’ capacity to control the agent may be diluted. In the EU, for example, the Court of Justice, the Court of First Instance, and the EU legislator (the Council of Ministers and the EP), have powers to nullify the prospective effects of certain Commission acts taken pursuant to secondary legislation, powers that depend on legal basis and other circumstances. However, when the Commission enforces competition rules, only the EU courts and those entities empowered to revise the treaty may reverse its decisions. Is the Court of Justice a principal when it quashes a Commission decision, or revises a Treaty provision through constitutional interpretation in ways that bind the member states? No, since the founders of the treaty designed the Court as an agent whose tasks include monitoring the activities of other agents, and of the member states themselves. Control over the Commission’s purse strings is mostly supranational, while control over its recruitment is mostly intergovernmental; the EP can dismiss the Commission on its own if it so decides. In situations of this complexity, the analyst cannot assume that principals can control agents; instead, the relationship between principals and agents must be carefully stipulated, in light of the various means of control at the disposal of the former. Finally, institutional arrangements—for example, the relative sizes of ‘zone of discretions’—are continuo us variab les that org anise prin agent relations. Majone has forcefully argued15 that when ‘political property rights’16 have been transferred to agents, more or less completely and in permanence, the standard P-A framework loses much of its relevance and utility. Some members of this group agree (for example, Stone Sweet,

The politics of delegation

6

this volume). Majone proposes a model of fiduciary relations, in which principals, faced with acute commitment problems, delegate broad, open-ended authority to trustees to govern in their stead. In Majone’s terms: ‘Trustees are agents, but not all agents are trustees; a trustee is an agent and something more’.17 A trustee typically wields the power to govern those who have delegated in the first place. In the EU, for example, one of the Court’s tasks is to interpret authoritatively provisions of the Treaty of Rome; such rulings govern all legal persons in the Community, including the member states. The Court, as constitutional court,18 is a trustee that is well insulated from formal controls. However, when the Court interprets a regulation or directive, it acts more as the agent of the EU legislator, and the EU legislator monitors and corrects the Court’s rulings, through subsequent legislative acts, as the legislator see fit.19 Similarly, as Majone points out, when the Commission enforces competition rules it acts as a trustee (the competence is stipulated by the Treaty); when it applies secondary legislation to situations, it acts more as an agent. One proposition that can be derived from such distinctions is that in a trustee situation, that is, where the zone of discretion is all but unlimited, temporal changes in the distribution of the principals’ preferences will have less impact on the agent’s activities, or policy outcomes, than in an agency situation. The problem of knowing how to identify the exact point, along any given spectrum that arrays various forms of delegation, the agent is more properly theorised as a trustee has not been resolved. Alternatives to Principal-Agent Logics This project did not seek to achieve intra-group consensus on methods or theory, but rather to initiate debates about both, in light of our respective empirical interests. Indeed, one of this volume’s objectives is to assess alternatives to P-A approaches to delegation. Let us clear away one false issue immediately. We see no point in seeking to ‘disprove’ the P-A construct, which is a vocabulary and a body of concepts, any more than one could ‘disprove’ rational choice approaches to politics more generally. There is a long tradition, both inside and outside the parameters set by the framework, of criticising existing applications of the framework for having been poorly specified.20 Faced with a challenge to received wisdom, the P-A framework, like most rational choice-based theorising, is almost always flexible enough to absorb new variables and logics, albeit through respecifying concepts in its own peculiar language. That said, purely functional analysis is incapable of explaining some important aspects of the emergence and evolution of institutional arrangements. Stipulating the existence of sufficient functional demand to support and sustain a new non-majoritarian institution tells us next to nothing about how actors came to define problems in ways that pointed to delegation. Functional problems can be addressed in myriad ways. How actors perceive, and then select from, the choices available to them is almost always conditioned by local histories, pre-existing institutional arrangements, and contingent forces and events. Further, assuming that a specific institution does provide some of the benefits of delegation tells us nothing about how agents actually perform, and with what effect on the world. Generally, functional analyses alone are ill-suited to deal with the temporal dynamics of delegation. The P-A framework may be dominated by other approaches on these and other dimensions as well. So-called sociological institutionalists, for example, have explored

Theory and practice of delegation to non-majoritarian institutions

7

the question of why some forms spread as quickly as they do, across social systems, at one time in history, rather than others.21 Historical accounts of ‘institutional isomorphism’—how organisational forms and models develop, standardise, and diffuse22—are typically more causally complete than are functional accounts. Although P-A approaches focus heavily on institutional choice through design, organisational sociologists have been more concerned with the broader processes that advantage some institutional arrangements while disadvantaging others. Further, unlike theories that are specified more generally, standard P-A models operate at very low levels of abstraction, and do not in themselves travel well. After all, a properly specified model purports to represent the particularities of a specific case. Last, P-A approaches have been slow to develop tools to capture the dynamics of the relationship between principals, agents, and society, relationships typically constructed over time by complex feedback loops that are difficult to model rigorously in gametheoretic ways. So-called historical institutionalists claim to have developed better tools, and they may be right. We return to these and related topics in the next section. Depending upon the research question, there may be comparative advantages to using an existing alternative. For some purposes, a P-A analysis might profitably complement a sociological or historical approach, or one may help to explain residual variance left over by the other. But it also may be that explanatory approaches developed by sociologists and organisational theorists simply perform better, across the board, than do P-A approaches. There are fierce difficulties associated with adjudicating between theories that may be a priori incommensurate. In this project, we encouraged contributors to explore these issues as fully as possible, in terms of more specific empirical concerns. DELEGATING TO NON-MAJORITARIAN INSTITUTIONS IN EUROPE At first sight, the idea of elected politicians voluntarily passing powers to unelected bodies such as courts, regulators, independent central banks and the European Commission and Court of Justice runs counter to post-war experiences of expansion in state activities and competencies in Western Europe. Not surprisingly, our studies show that elected officials in Europe have delegated in order to resolve various collective action problems. National governments have transferred powers to EU institutions in order to deal with the negative externalities of market integration, as well as to monitor and enforce EU law (Tallberg, this volume). Constitutional courts have provided a response to the dilemma of political parties who agreed on the benefits of constitutional ‘rules of the game’, but disagreed, sometimes fundamentally, on the precise content of those rules (Stone Sweet, this volume). Independent regulatory authorities (Thatcher, this volume) and competition authorities (Wilks and Bartle, this volume) have responded to pressures to stabilise environments for investors, while enabling the state to respond more effectively to increased technical complexity.23 The rhetoric justifying independent central banks (McNamara, this volume) emphasises the advantages of low inflation and price stability, thus reassuring key business and financial interests. All of these bodies may offer elected politicians scapegoats for hard choices for which they might otherwise be blamed.

The politics of delegation

8

While functional demands for delegation can almost always be identified, most of the contributors to this volume argue that the pressures on elected officials and the functions that NMIs can perform for them do not, alone, explain the expansion of NMIs in Europe. Several lines of attack are developed against purely functionalist models. First, those who delegate are not just responding to functional demands; instead the perception that delegation to NMIs is the best option for dealing with certain problems is socially constructed, and that process is always analytically prior to the decision to delegate. Katherine McNamara (this volume) denies that independent central banks are a natural or necessary response to resolve a commitment problem. She shows that independent central banks do not always produce lower inflation and higher growth, that higher inflation does not always harm economic growth, and that partisan-electoral factors do not always lead to higher rates of inflation. Indeed, central bank independence was introduced in Western Europe in the 1990s in an era of very low inflation. Instead, what is crucial is that elected officials around the globe have come to see the independent central bank as an institution that fits better with more general ideas about how the economy works than does a bank controlled directly by a national treasury. In the field of competition policy, Stephen Wilks and Ian Bartle (this volume) argue that the creation of competition authorities in Britain was initially symbolic. Governments wished to reassure business whilst believing that the authorities would be unimportant (although in fact, those expectations were not fulfilled). A second set of challenges for functionalist explanations arises from variation, whether read across nation-states or policy domains. Most P-A work on delegation has not attempted much cross-national comparison, although Western Europe offers a fertile ground for developing and testing theory.24 European polities, partly because they are similarly developed and integrated into international and regional economies, face similar functional pressures for delegation to NMIs. Yet outcomes vary. Although delegation to NMIs has spread, countries have made quite different choices over when, whether and how to delegate. Thus, Mark Thatcher (this volume) shows that the practice of delegating to national independent regulatory agencies has diffused steadily across various policy sectors, but that the structure and operation of these agencies varies widely within any given domain, despite the existence of similar pressures. There is also important variation across sectors, within countries. Marian Döhler (this volume), for instance, finds that Germany has delegated relatively little beyond long-standing, successful NMIs, such as the Bundesbank, the Federal Cartel Office, and the Constitutional Court. Third, functionalist explanations cannot account for the timing of delegation. Germany established at least a semi-independent central bank and a national competition authority in the 1950s, while other countries only did so in the late 1980s and 1990s. The date of creation of constitutional courts and independent regulatory agencies for the utilities also differs across countries. Many of the governance problems emphasised by proponents of P-A approaches pre-date delegation, credible commitment in monetary policy or regulation, technical complexity or unpopular policies. Yet they were rarely met by delegation to NMIs before the 1980s and 1990s. Two sets of responses to these problems emerged within our group. One was to adapt the P-A approach to the politics of NMIs in Europe by carefully specifying local conditions. In the European context, which, in contrast to the US situation, features strong political parties and parliamentary government, the principals are mostly

Theory and practice of delegation to non-majoritarian institutions

9

governments and political parties. Although parliaments may pass legislation to delegate powers to NMIs, they are largely controlled by the executive through strong party systems and constitutional constraints. Contributors then investigate how and why governments and parties delegate, in the context of new or inherited political and constitutional structures. Several of the chapters in this volume (Stone Sweet, Tallberg, and Pollack) proceed in this way. Others in the group felt that the P-A framework was too limiting, and not sensitive enough to the contingencies of history, agency, ideas, and so on (Döhler, Thatcher, Wilks and Bartle). They wanted richer accounts of why, when, and how delegation occurs, and they explore factors that P-A analyses conceptualise differently, mention only in passing, or ignore altogether. We discuss three of these factors here: the definition of interests, institutional isomorphism, and historical legacies. These are applied to the two key elements of institutional design: the decision to delegate; and the institutional form selected. The Decision to Delegate to an NMI Since delegation is not costless, principals must believe that their interests are better served by delegating to an NMI, relative to an existing situation or an alternative institutional arrangement. Their decisions are influenced by national and international examples of delegation, by pressures from those to be regulated or otherwise concerned, and by existing institutional constraints. Hence we need to investigate the interests of decision makers, policy learning and institutional isomorphism, and institutional inheritances. Interests. Offering a fuller explanation of the spread of delegation of NMIs (including variations in extent and timing), whether within a P-A framework or outside it, must begin by specifying the interests of the key actors involved in the decision to delegate. The question of how, across Europe, delegation to NMIs came to be considered an appropriate response to an increasing number of governance problems is an empirical one; it must be investigated rather than assumed or traced backwards from the existence or lack of delegation. Some members of the group addressed this question through methods commonly associated with ‘historical’ and ‘sociological’ institutionalism,25 showing how preferences for new NMIs were constructed or altered (for example, McNamara, Thatcher, Wilks/Bartle, this volume). They show several routes whereby the interests of elected officials altered. Pressures were exerted on governments by powerful interests such as bankers and business, who sought delegation for their own ends, such as a more predictable, less ‘politicised’ environment. Assumptions of how agreed policy aims could be achieved evolved—for example, politicians came to believe that their interests in low inflation were better served by an independent central banks. In fields such as general competition policy or sectoral regulation, elected officials, confronted by scandals, technical complexity and international pressures, together with new ideas and examples for reform, concluded that having direct control over policy, or at least appearing to have it, was no longer advantageous. Other contributors adopted standard assumptions of rational choice (for example, Shapiro, Stone Sweet, Tallberg, this volume), but then focused attention on how the activities of agents within domains constituted by delegation ‘feeds back’ onto the wider political system. Thus judges, be they in constitutional courts or engaged in judicial

The politics of delegation

10

review, have developed legal doctrines that extend their powers and constrain those of elected politicians. In the EU, the ECJ exploited differences among national governments to enhance the effectiveness of EU law, and thus to increase its owns powers in the polity. Policy Learning and Institutional Isomorphism. Policy learning and institutional isomorphism (transfer of institutional forms) have been crucial to the spread of NMIs in Europe.26 At times, the American experience with delegation has generated templates for reform and norms of appropriateness that have travelled across the Atlantic, particularly to Britain.27 In other situations, European polities have copied one another. General competition authorities spread via an ‘orgy of borrowing’ (Wilks and Bartle, this volume); the creation of an independent regulator in telecommunications in Britain was followed by similar bodies in other European countries (Thatcher, this volume); the Austrian constitutional court ultimately mutated into a pan-European institution, through successive waves of democratisation in this century (Stone Sweet, this volume); independent central banks (McNamara, this volume) became de rigueur in the 1990s.28 Cross-domain isomorphism has probably been less important, although it has taken place across closely related fields, such as utilities regulation.29 Much of the day-to-day politics in domains governed by NMIs is heavily structured by the activities of knowledge-based elites. These include scientists, other technical experts, economists, bankers, and lawyers. Such actors have developed strong corporate or professional interests in the work of NMIs. They help to diffuse NMIs and the models of governance that go with them facilitate institutional isomorphism, not least by working to legitimise some forms of governance while delegitimising others.30 In increasingly Europeanised and internationalised domains such as central banking, utilities regulation or competition policy, transnational communities of professionals have contributed to making delegation to NMIs part of ‘good governance’ norms that have become orthodoxy.31 The process has not been disinterested: these actors increase their own social power and influence. ‘Coercive isomorphism’32 —tha diffusiou of institutional forms and practices through legal obligation backed up by monitoring and enforcement mechanisms—has also played a role in the spread of delegation to NMIs. European integration has favored isomorphic processes.33 The Maastricht Treaty, for example, insisted on independent central banks as a precondition for entry into a single currency. Many important pieces of EC secondary legislation provide pressure for delegation to specialised agencies. New constitutions, the European Convention of Human Rights, and the development of EU law require judges to engage in more judicial review of agencies (Shapiro, this volume). Institutional Legacies. The pace and scope of delegation to NMIs have been strongly mediated by national-level factors.34 State structures influence the possibility, attractiveness, and ease of delegation to NMIs. In countries such as Britain, with relatively few constraints on the executive and no entrenched constitution, delegation to NMIs can be achieved relatively quickly. Thus the Bank of England was made independent in 1997, almost literally overnight, despite no such pledge having been made in Labour’s election manifesto. When constitutions have broken down and are completely rewritten, powerful opportunities exist for delegation—for instance, by creating new constitutional courts or agencies. However, if the creation of a new NMI requires constitutional amendment, with super-majorities or special procedures, reforms will be

Theory and practice of delegation to non-majoritarian institutions

11

more difficult to introduce. Delegation in Germany faces an uphill battle to the extent that it meets with various legal obstacles, such as the constitutionally protected prerogatives of the Länder (see Döhler, this volume). State structures and institutional legacies affect the relative costs and benefits of different organisational forms.35 In Britain, regulatory commissions date back to the nineteenth century. Moreover, governments find it relatively easy to alter the organisational basis and personnel of NMIs such as regulatory agencies which do not enjoy specific constitutional protection. In countries like Italy, such agencies have formal independence and nomination procedures can be complex due to the involvement of the legislature. In Germany, an alternative and apparently successful regulatory model of federal agencies within ministries, a powerful Cartel Office and strong regional traditions, have acted as a barrier to the creation of sectoral independent regulatory agencies.36 Inheritance also affects learning and the search for organisational forms. Thus, long-standing linkages by British policy makers to the United States aided the importation of American examples of delegation; one example is the way in which utility commissions such as the FCC in telecommunications, inspired semi-independent regulators in Britain.37 In contrast, continental states have looked more to each other for institutional developments. Institutional Forms of Delegation to NMIs Delegation to non-majoritarian institutions has taken many forms in Europe, both crossnationally and across domains. Comparison of ‘zones of discretion’, requiring examination of the powers delegated and the controls imposed on NMIs, offers insights into the variety of institutional forms. Constitutional courts and independent central banks operate in the most extensive zones of discretion, although there is significant variation across countries. Their decisions are difficult to reverse, often requiring constitutional amendment, they face relatively few ongoing controls and their organisational position is frequently highly protected by being embedded in constitutions. The European Commission operates in a more restrictive environment when it deals with representatives of the member-states in the Council of Ministers or the European Council, than it does when it enforces the competition rules found in the Rome Treaty. In Majone’s terms, in the former context it looks and behaves like an agent vis-à-vis principals, while in the latter it resembles, as does the constitutional judge, a trustee exercising fiduciary powers (Stone Sweet, Tallberg, this volume). The sum of powers delegated to national regulatory authorities looks modest when compared to the courts and central banks. Regulators’ powers are circumscribed—for instance, to policing specified anti-competitive behaviour, approving certain types of merger, imposing fines and enforcing licences. Their decisions can be reversed relatively easily—by ordinary legislation, judicial ruling, or sometimes simply by ministerial order. They require considerable resources to function effectively, but these are allocated by annual budgetary procedures under the control of their principals, namely governments and legislatures. Legally, elected officials can easily abolish or alter the organisational structure of regulatory authorities since these are rarely protected by constitutional provisions. The volume, nonetheless, shows that the authority of these agents has

The politics of delegation

12

increased over time. Finally, delegation has been weakest towards executive agencies of governments (see Döhler, this volume). Indeed, most were excluded from this volume because they did not meet the minimum requirements to constitute a NMI, particularly that of not being directly managed by elected officials and enjoying powers separate from other institutions, notably ministers and civil servants. One explanation of differing forms across domains is to link institutional design to pressures on principals and the functions performed by NMIs offers. Where delegation takes place in order to secure credible commitment, principals cannot impose many ex post controls over the agent without undermining the very purpose of delegation. The more acute the commitment problem that principals face (or believe they face), the more discretion they will delegate—for example, in enforcing EU law or in monetary policy. Where officials delegate to increase technical efficiency, reduce their workload, or improve their information, extensive ex post controls are often more compatible with objectives. Thus, when principals seek to take advantage of the technical expertise of NMIs in regulatory environments characterised by great uncertainty, they will create agents (and not trustees) whose purpose is to lower the costs of making good policy. Such general functionalist explanations of institutional form are useful but equally are subject to the same criticisms as those applied to analyses of the decision to delegate. They offer only limited predictions about specific choices made in countries or domains that are strongly influenced by history and context. Choices over institutional form are often greatly constrained by cognitive and normative limits, existing state structures, and institutional legacies. Hence contributors enriched their explanations by looking at such factors either within a P-A framework or outside it. State and domain traditions provide templates of institutional forms. Often elected officials copy existing forms, leading to ‘institutional isomorphism’. In Germany, the same ‘genetic code’ has led to an almost automatic copying of the same institutional form for federal agencies (Döhler, this volume). Institutional forms have been ‘read across’ domains within countries for regulatory agencies, leading to a ‘British utility regulatory’ model, or French ‘Autorités Administratives Indépendentes’ (Thatcher, this volume). The spread of the Kelsenian constitutional court, now in place from Portugal to Russia, proceeded first from the rejection of American-style judicial review, given embedded separation of powers doctrines, and then through rote copying (Stone Sweet, this volume). Constitutional and state arrangements strongly structure decisions on institutional form. In countries with legislatures that enjoy greater independence from the executive, nomination to, and controls over, agents tend to be shared between the two bodies (for example, Germany, Italy). Where executive powers are more highly centralised (Britain, France), such controls are typically wielded by cabinets and ministries.38 The extent of judicial review of agencies and its biases towards support for agencies or for challengers to them, are influenced by factors such as national legal doctrines, legal fashions, and the training, recruitment and career patterns of judges (Shapiro, this volume).

Theory and practice of delegation to non-majoritarian institutions

13

POLITICAL DYNAMICS AND CONSEQUENCES The consequences of delegation to NMIs are as important as decisions over initial institutional design for our case studies. NMIs have become powerful participants in policy making and may now constitute a ‘fourth branch of government’ in Europe. Delegation to them has affected the distribution of power, with winners and losers. It has also had impacts on processes of decision making, substantive policy making and further institutional reform. For analytic purposes, we can divide the effects of delegation to NMIs into two groups: the immediate and direct effects of NMI behaviour; feedback effects whereby NMI decisions affect the wider environment of NMIs leading to further rounds of consequences. In addition, delegation to NMIs has raised wider concerns of legitimacy and accountability. The case studies suggest that the greater the initial ‘zone of discretion’, the more significant are the effects (direct and indirect) of delegation and the more likely are unanticipated consequences. NMIs such as constitutional courts, which have great powers and face few controls (that is, are better understood as trustees than agents), routinely generate sweeping outcomes that are frequently unanticipated. Nevertheless, understanding the consequences of delegation requires tracing the effects of NMI behaviour in their context and inclusion of the reactions of other actors: the formal initial zone of discretion is the starting point for analysis, not the end of it. Direct Effects Within the state, delegation has seen the emergence of powerful NMIs. Analysts using PA approaches typically focus on the extent to which agents may escape from the latter’s control, due to information asymmetries, divided or multiple principals, and so on. They then discuss how the design of formal controls can limit such ‘agency losses’.39 Yet determining ‘agency losses’ and ‘control’ poses fierce methodological difficulties, especially of observed equivalence, as Mark Pollack points out. We have therefore examined the relationship between elected officials and NMIs in relation to the initial ‘zone of discretion’ established by formal powers and controls. As a result, case studies are sensitive to how the position of NMIs vis-à-vis their principals may change over time, and in ways unanticipated at the moment of delegation. The British government believed that general competition authorities would be largely symbolic, to placate business interests; in practice, they became bodies exerting considerable powers (Wilks and Bartle, this volume). The ECJ has unexpectedly expanded its authority (and that of national judges) through its rulings (Tallberg, this volume). At times, the power of NMIs has differed sharply from formal institutional arrangements. The German Federal Cartel Office became a great deal more independent of elected politicians than warranted by its initial formal position. Establishing constitutional courts in Europe ultimately led to an important expansion in judicial, relative to legislative, power (Stone Sweet). Delegation has also influenced the power of private interests. It has offered new opportunities to challenge public decisions compared with previous state structures. Constitutional courts and courts reviewing administrative agencies through judicial

The politics of delegation

14

review have allowed those who lose through the party political and central government process another opportunity to challenge decisions. Independent regulatory authorities have been more open to new entrants than governments, who engaged in closed relationships with national champion firms.40 NMIs have marked substantive policy decisions. General competition authorities and regulatory agencies have vigourously promoted liberalisation and the reduction of crosssubsidies. The European Commission and ECJ have aided increased European integration, economic liberalisation and the rights of individuals (including firms) against states. Courts have sometimes checked government secrecy, protected individual liberties and pushed forward policies across many fields. Independent central banks have promoted monetarist policies that give inflation a higher immediate priority than combating unemployment and seek to limit public expenditure. These are but a few examples of the effects of NMIs drawn from our case studies, for there is barely a policy domain unaffected by NMI action. In analysing the direct impacts of NMIs, the formal powers of NMIs and controls for elected politicians offer a good starting point. The case studies also suggest that the wider political context affects the use of controls for example, the party structure and number of veto players and points. In countries such as Britain with few veto points and players, single party governments with majorities can more easily apply controls than in other countries (see, for example, Shapiro, this volume). However, formal structures offer only limited explanation for the consequences of delegation. NMIs are not passive rather, they can be active, creative actors. They have learned how to exploit their powers in unexpected ways—the legal doctrines created by the ECJ and constitutional courts offer excellent cases (Shapiro, Stone Sweet and Tallberg, this volume). Informal norms grow up that may alter or subvert formal arrangements—for example, independence for the German Cartel Office. NMIs have sought to go beyond their formal roles and developed norms that may differ from their formal objectives and position (for instance, the German Federal Cartel Office and Bundesbank making independence from elected politicians a central norm or sectoral regulators setting themselves the promotion of fair and effective competition as their central objective—see Wilks/Bartle, Döhler and Thatcher contributions, this volume). Feedback Effects The behaviour of NMIs modify the decision making, investments, and strategic interactions of other actors, which may feed back on principals and the NMIs themselves. The activities of courts or regulatory agencies have led interest groups to reorient their activities to them, in order to increase their own influence; in turn, the behaviour of those interest groups has provided new opportunities for NMIs to expand their influence. Earlier choices to delegate have also influenced later choices, for delegation is often a process rather than a one-off event and, moreover, one in which NMIs can be active agents. In the EU, the Commission and ECJ have been major players in treaty making processes, some of which have broadened their powers. At the same time, national governments have learned from previous experience: they excluded some new domains from the first ‘pillar’ of decision making; in other domains, where inter-governmental cooperation failed (for example, immigration or security), they delegated new powers to EU

Theory and practice of delegation to non-majoritarian institutions

15

institutions (Jonas Tallberg, this volume). Similarly, the apparent success of regulatory agencies has led to a widening of their powers in countries such as Britain and France. Explaining feedback effects requires attention to how the environment, especially the zone of discretion, structures the interactions between NMIs, their principals, and other actors. The ECJ’s interpretations of the Treaty of Rome have produced powerful feedback effects, some of which have affected later Treaty amendments, partly because reversal requires unanimity among member states (Tallberg contribution, this volume).41 Another aspect is the presence of informal norms. The initial limited formal autonomy of the German Federal Cartel Office was expanded by strong informal norms; in turn, these hindered the creation of other regulatory bodies (Wilks/Bartle, Döhler, Thatcher, this volume). Feedback effects such as these were unanticipated by principals at the time of the initial delegation. Legitimacy and Accountability Delegation to NMIs poses difficulties for the traditional model of parliamentary accountability in Western Europe. Moreover, the legitimacy of NMIs affects their power and acceptance. ‘Output legitimacy’ is the current standard justification for NMIs.42 The superior performance of NMIs, relative to the result that would be likely if elected politicians were to perform the functions themselves, justifies their existence. Benefits include more ‘efficient’ policy making procedures and better policy outcomes, enhanced rights protection, lower inflation, and so on. In this view, delegation is portrayed as neutral and beneficial. Indeed, the legitimacy of NMIs is only at issue when agents are corrupted, or fail to deliver adequate levels of promised benefits.43 Yet the output model of legitimacy, because it ignores too much of political importance, does not offer a comprehensive response to many important legitimacy concerns. The model typically assumes ideological consensus on the role, function, and benefits to be accrued from delegation, whereas these points are contested by parties and groups. There are winners and losers from delegation to NMIs which form part of political struggles. The liberalisation of markets, part of the logic of creating independent regulatory authorities, have been the subject of vigorous partisan wrangling, precisely because some industries and interests will lose from liberalisation and the re-regulation that comes with it. There is intense controversy about how and whether to widen and deepen European political integration. Central banks may pursue low inflation, but their decisions may induce higher levels of unemployment. A second model of legitimacy is procedural. It relies on the process of decision making by NMIs being better than the insular, often secret, deliberations of cabinets and executives. NMIs typically pursue clearer, more focused mandates than elected politicians, all the more so since delegation usually involves explicitly setting objectives for them. They are often obliged to give reasons for their decisions and to publish information. They are frequently more open than central governments to interested groups—for example, the European Commission is much more accessible compared to national bureaucracies,44 whilst in the utilities independent sectoral regulators have ended closed relationships of ministries and suppliers by creating elaborate consultation procedures. Legislatures have powers over NMIs, including rights over nominations,

The politics of delegation

16

information and calling hearings, and have used them to publish valuable information and to put pressure on NMIs. According to this view, procedural legitimacy is a fair and democratic substitute for electoral accountability, given the broader purposes of delegation. As NMIs have spread in Europe, these two models of legitimacy are increasingly combined: output legitimacy is bolstered by procedural legitimacy. CONCLUSIONS Delegation to Non-Majoritarian Institutions has reconfigured the architecture of the state and the EU, altered public policies, and raised issues of legitimacy and accountability. This volume explores some of the crucial questions at the very core of these politics. Who has delegated and why? What institutions have been established, for what purposes, and how great are their respective zones of discretion? How have NMIs actually governed and with what effects? Although functional rationales can help us to answer some of these questions, purely functionalist analysis suffers from a number of major limitations, since it plays down the historically contingent, contested, and political sources and logics of delegation.45 Moreover, approaches that explain institutional design, without going further, miss much that is vital in the consequences of delegation. These limitations are well exposed by analysis of delegation in Western Europe, especially by cross-national comparison over time and across domains. In response, this volume’s contributors either supplement or replace purely functional accounts of delegation with more fine-grained or historically sensitive analyses of institutional choice. They investigate decisions to delegate in their contexts—country, time, and domain. They acknowledge constraints on the choices of principals—cognitive, normative, political, and constitutional—thus incorporating concern for policy learning and the spread of ideas, as well as the weight of inherited structures, norms, and experience. The volume shows that a broad range of outcomes are impossible to predict from the initial conditions of delegation, but are instead produced by the dynamics of interactions between NMIs and other multiple actors. The politics of delegation in Europe today calls for analyses going well beyond simple functional logics, involving how actors’ interests are defined, policy is made and enforced, and the legitimacy of government is conceived. NOTES 1. We have chosen to use the term institution as a synonym for a political body, organ, or organisation. In the standard vocabulary of social science institutionalism, institutions refer to the constitutive elements of the rule systems found in any human community. We use the terms, ‘rules’, ‘law’, and ‘legal instruments’ to refer to institutions in this second sense. 2. Good recent reviews of this literature include J.Huber and C.Shipan, ‘The Costs of Control: Legislators, Agencies, and Transaction Costs’, Legislative Studies Quarterly 25(2000), pp. 25–42; J.Bendor, A.Glazer and T.Hammond, ‘Theories of Delegation’, Annual Review of Political Science 4 (2001), pp. 235–69.

Theory and practice of delegation to non-majoritarian institutions

17

3. E.g., W.Powell and P.DiMaggio (eds.), The New Institutionalism in Organizational Analysis (Chicago: University of Chicago Press 1991). 4. E.g., G.Majone, ‘The Rise of the Regulatory State in Europe’, West European Politics 17/3 (1994), pp. 77–101; G.Majone, Regulating Europe (London: Routledge 1996). 5. E.g., M.Pollack, ‘Delegation, Agency, and Agenda Setting in the European Community’, International Organization 51/1 (1997), pp. 99–134; M.Pollack, ‘The Engines of Integration? Supranational Autonomy and Influence in the European Union’, in W. Sandholtz and A.Stone Sweet (eds.), European Integration and Supranational Governance (Oxford: Oxford University Press 1998), pp. 217–49. 6. In addition to the work of Pollack and Majone cited above, see J.Tallberg, ‘Making States Comply: The European Commission, the European Court of Justice, and the Enforcement of the Internal Market’ (Ph.D. diss., Lund University 1999); J.Tallberg, ‘The Anatomy of Autonomy: An Institutional Account of Variation in Supranational Influence’, Journal of Common Market Studies 38/5 (2000), pp. 843–64, A.Stone Sweet and J.Caporaso, ‘From Free Trade to Supranational Polity: The European Court and Integration’, in Sandholtz and Stone Sweet (eds.), European Integration and Supranational Governance. 7. See T.Bergman, W.Müller and K.Strøm (eds.), Parliamentary Democracy and the Chain of Delegation, special issue of European Journal of Political Research 37 (2000). 8. A.Stone Sweet, ‘Judicialization and the Construction of Governance’, Comparative Political Studies 31 (1999), p. 147. 9. T.Moe, ‘The New Economics of Organization’, American Journal of Political Science 28 (1985), pp. 739–77; D.Epstein and S.O’Halloran, Delegating Powers: A Transaction Cost Politics Approach to Policy Making under Separation of Powers (Cambridge: Cambridge University Press 1999), ch. 3. 10. Huber and Shipan, ‘The Costs of Control’, p. 41. 11. E.g., S.Balla, ‘Administrative Procedures and Political Control of the Bureaucracy’, American Political Science Review 92/3 (1998), pp. 663–73. 12. Alternative ways of thinking about delegation do not necessarily embody alternative, or rival, explanations. For explanations to compete, they must be specified in a sufficiently clear causal form to be compared with one another, in light of empirical research and findings. It is possible that virtually identical hypotheses, or causal propositions, could be generated from quite different theoretical vocabularies and materials. 13. Epstein and O’Halloran, Delegating Powers, pp. 24–5. 14. K.Bawn, ‘Choosing Strategies to Control the Bureaucracy: Statutory Constraints, Oversight, and the Committee System’, Journal of Law, Economics, and Organization 13 (1997), pp. 101–26. 15. G.Majone, ‘Two Logics of Delegation: Agency and Fiduciary Relations in EU Governance’, European Union Politics 2/1 (2001), pp. 103–22. 16. T.Moe, ‘Political Institutions: The Neglected Side of the Story’, Journal of Law, Economics, and Organisation 6 (1990), pp. 213–53. 17. Majone, ‘Two Logics of Delegation’, p. 113. 18. A.Stone Sweet, Governing with Judge: Constitutional Politics in Europe (Oxford: Oxford University Press 2000). 19. Stone Sweet and Caporaso, ‘From Free Trade to Supranational Polity’. 20. E.g., Balla, ‘Administrative Procedures’; T.Moe, ‘An Assessment of the Positive Theory of “Congressional Dominance”’, Legislative Studies Quarterly 12 (1987), pp. 475–520. 21. J.W.Meyer and R.L.Jepperson, ‘The “Actors” of Modern Society: The Cultural Construction of Social Agency’, Sociological Theory 18 (2000), pp. 100–120; R.L. Jepperson, ‘The Development and Application of Sociological Neoinstitutionalism’, Robert Schuman Centre Working Paper 2001/5.

The politics of delegation

18

22. P.DiMaggio and W.Powell, The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality’, in W.Powell and P.DiMaggio (eds.), The New Institutionalism in Organizational Analysis (Chicago: University of Chicago Press 1991). 23. See also B.Levy and P.Spiller (eds.), Regulation, Institutions and Commitment (Cambridge: Cambridge University Press 1996). 24. In the field of European politics, exceptions include: Bergman et al., Parliamentary Democracy and the Chain of Delegation; J.Huber and A.Lupia, ‘Cabinet Instability and Delegation in Parliamentary Democracies’, American Journal of Political Science 45/1 (2001), pp. 18–33; Stone Sweet, Governing with Judges. 25. Cf. P.Hall and R.Taylor, ‘Political Science and the Three New Institutionalisms’, Political Studies 44/4 (1996), pp. 936–57; K.Thelen, ‘Historical Institutionalism in Comparative Politics’, The Annual Review of Political Science 1999 (Palo Alto: Annual Reviews 1999); E.Immergut, ‘The Theoretical Core of the New Institutionalism’, Politics and Society 25/1 (1998), pp. 5–34. 26. For institutional isomorphism, see DiMaggio and Powell, ‘The Iron Cage Revisited’; for policy transfer and learning, see reviews, notably D.Dolowitz and D.Marsh (1996), ‘Who Learns What From Whom? A Review of the Policy Transfer Literature’, Political Studies 44/2, pp. 343–57; P.A.Hall, ‘Policy Paradigms, Social Learning, and the State’, Comparative Politics 25/3 (1993), pp. 275–96; J.L.Campbell, ‘Institutional Analysis and the Role of Ideas in Political Economy’, Theory and Society 27/4 (1998), pp. 377–409; R.Rose, Lessondrawing in Public Policy (London: Chatham House Publishers 1993). 27. Cf. H.Wolman, ‘Understanding Cross-National Policy Transfers: The Case of Britain and the US’, Governance 5/1 (1992), pp. 27–45; for the example of telecommunications, see M. Thatcher, The Politics of Telecommunications (Oxford: Oxford University Press 1999). 28. See also B.Doern and S.Wilks (eds.), Comparative Competition Policy: National Institutions in a Global Market (Oxford: Oxford University Press 1996); K.McNamera, The Currency of Ideas: Monetary Politics in the European Union (Ithaca: Cornell University Press 1998). 29. D.Coen and M.Thatcher (eds.), Regulating European Utilities, special issue of Current Politics and Economics of Europe 9/4 (2000). 30. Through a process that DiMaggio and Powell call ‘normative isomorphism’. 31. Cf. S.Jacobs, Regulatory Governance: Improving the Basis for Sectoral Regulation (Paris: OECD 2000); J.L.Guesch and P.Spiller, Managing the Regulatory Process: Design, Concepts, Issues and the Latin American and Carribbean Story (World Bank: Washington 1999); B.Levy and P.Spiller, Regulation, Institutions and Commitment (Cambridge: Cambridge University Press 1996). 32. DiMaggio and Powell, ‘The Iron Cage Revisited’. 33. C.Radaelli, ‘Policy Transfer in the European Union’, Governance 13/1 (2000), pp. 25–43. 34. S.Berger and R.Dore (eds.), National Diversity and Global Capitalism (Ithaca: Cornell University Press 1996); V.Schmidt, ‘Still Three Models of Capitalism? The Dynamics of Economic Adjustment in Britain, Germany and France’, in R.Czada and S.Lütz (eds.), Die Politische Konstitution von Märkten (Opladen: Westdeutscher Verlag 2000); S.K.Vogel, Freer Markets, More Rules. Regulatory Reform in Advanced Industrial Countries (Ithaca: Cornell University Press 1996). 35. For a rational choice view of the effects of political context on delegation, see J.Huber, C. Shipan and M.Pfahler, ‘Legislatures and Statutory Control of Bureaucracy’, American Journal of Political Science 45/2 (2001), pp. 330–45; J.Huber and C.Shipan, ‘Legislators and Agencies: A Theoretical Reappraisal’, Legislative Studies Quarterly 25 (2000), pp. 25– 52. 36. Cf. A.Héritier, C.Knill and S.Mingers, Ringing the Changes (Berlin: De Gruyter 1996); for electricity where no sectoral regulator was created in contrast to other EU countries, see B. Eberlein, ‘Institutional Change and Continuity in German Infrastructure Management: The Case of Electricity Reform’, German Politics 9/3 (2000), pp. 81–104.

Theory and practice of delegation to non-majoritarian institutions

19

37. Cf. Wolman, ‘Understanding Cross National Policy Transfers’; Thatcher, The Politics of Telecommunications. 38. See Huber et al., ‘Legislatures and Statutory Control of Bureaucracy’; and Huber and Shipan, ‘Legislators and Agencies’. 39. Such losses arise from agency ‘slippage’ (agents following their own preferences which diverge from those of its principal(s)) and ‘shirking’ (institutional incentives causing the agent to behave contrary to the wishes of its principal(s)). See M. McCubbins and T. Schwartz, ‘Congressional Oversight Overlooked: Police Patrols versus Fire Alarms’, American Journal of Political Science 28/1 (1984), pp. 165–79; McCubbins et al., ‘Administrative Procedures’; T. Moe, ‘An Assessment of the Positive Theory of Congressional Dominance’, Legislative Studies Quarterly 12/4 (1987), pp. 475–520. 40. Thatcher, The Politics of Telecommunications; J.Hayward (ed.), Industrial Enterprise and European Integration. From National to Internationalized Champions: Firms and Governments in the West European Economy (Oxford: Oxford University Press 1995). 41. Explored at length in Stone Sweet, Governing with Judges, ch. 6. 42. Cf. Levy and Spiller, Regulation, Institutions and Commitment; Majone, Regulating Europe. 43. G.Majone, ‘The Crisis of Community Credibility’, Journal of Common Market Studies 38/2 (2000), pp. 273–302. 44. S.Mazey and J.Richardson (eds.), Lobbying in the European Community (Oxford: Oxford University Press 1993). 45. Cf. Moe, ‘Political Institutions’.

Delegation to Supranational Institutions: Why, How, and with What Consequences?

JONAS TALLBERG The growth of governance beyond the nation state is one of the most pronounced political trends in recent decades. To address problems that cannot be dealt with effectively at the national level, governments jointly develop international governance structures. Increasingly, the design of such structures involves the delegation of decision making powers to institutions that are organisationally and politically independent from the founding states, and therefore conceived of as ‘supranational’. In this respect, the evolution of international governance reflects a general development in politics: the delegation of political authority from representative organs to non-majoritarian institutions, which are neither directly elected by the people nor directly managed by elected politicians. Delegation to supranational institutions raises a set of questions of general relevance in political science. Why do elected politicians and sovereign member states freely give up some of their right to govern? Why do governments prefer certain forms of institutional design instead of others in the delegation of political authority? What are the consequences of delegation to supranational institutions for the problem solving capacity and democratic legitimacy of international governance structures? This chapter addresses these themes with specific reference to the European Union (EU).1 The overarching question is why, how, and with what consequences national governments delegate political authority to the supranational institutions of the EU. In an international comparative perspective, one of the distinguishing features of the EU is the extensive degree of delegation to the supranational and non-majoritarian European Commission and European Court of Justice (ECJ). Slightly simplified, the Commission has been entrusted with the tasks of developing proposals for new EU policy, executing EU policy within some specified domains, and securing member state compliance with EU rules, whereas the ECJ has been delegated the power to interpret the EU treaties and to ensure that EC law is correctly applied in the member states. A decade ago, an inventory of existing research on delegation to the EU’s supranational institutions would have yielded a limited number of publications. Today, the field of EU studies boasts a considerable, and steadily growing, number of works that explore various aspects of these processes of delegation.2 Two features distinguish this literature. First, this research is predominantly focused on the degree of discretion enjoyed by the supranational institutions, or, conversely, the degree of control exercised

Delegation to supranational institutions

21

by EU governments. The historical origin of this focus is the debate between neofunctionalism and intergovernmentalism on the driving forces of European integration, where competing conceptions of the role of the supranational institutions are pitted against each other. Second, existing literature shares a common anchoring in rational choice institutionalism in general, and principal-agent (P-A) analysis in particular. Building on the work of rational choice institutionalists on the US Congress, international regimes, and regulation, students of European politics have turned to principal-agent and transaction-cost analysis when attempting to explain patterns of delegation in the EU. The purpose of this chapter is not to denounce the merits of existing research, which has reached a certain cumulative strength. Instead, the ambition is to explore the dynamic linkages between stages of the delegation process that too often are studied in splendid isolation. Why, how, and with what consequences are not three separate questions with three separate answers, as they are currently treated in the literature, but analytical themes that are best understood in relation to each other. The chapter integrates the rationale, design, and consequences of delegation into a coherent and dynamic framework, where institutional function is the central explanatory variable. In the rational institutionalist argument presented here, the three stages of delegation are linked in a four-step causal chain, where (1) the expected consequences of delegation motivate EU governments to delegate certain functions to supranational institutions; (2) the nature of these functions influences the design of mechanisms for controlling the institutions; (3) institutional design shapes the consequences of delegation by facilitating or obstructing attempts by the institutions to implement private agendas; and (4) the consequences of previous rounds of delegation affect future delegation, institutional design, and interaction, through positive and negative feed-back loops. The discussion is structured in four substantive sections, each exploring one of the links in this dynamic framework. THE RATIONALE OF DELEGATION The first step in the process of delegation is the decision to confer political authority on supranational institutions. This decision involves both a choice in favour of delegation as the preferred way of solving a particular problem, and a choice between what powers to delegate. This section demonstrates how the authority conferred on the Commission and the ECJ can be linked to a set of common functional problems, often motivating delegation to political agents. The Functional Basis of Delegation Why do the elected politicians of sovereign member states decide to create, and delegate public authority to, supranational institutions beyond direct democratic control? Rational choice institutionalism offers an answer that is distinctly functional, bordering on tautological. Delegation is explained in terms of the anticipated effects for the delegating party, and is likely to take place when the expected benefits outweigh the expected costs.

The politics of delegation

22

The principal-agent model is the analytical expression of this functional logic, together with the notion of transaction costs. In the seminal article where the principal-agent imagery was first introduced, Stephen Ross describes how this relationship arises ‘between two (or more) parties when one, designated as the agent, acts for, on behalf of, or as a representative for the other, designated the principal, in a particular domain of decision problems’.3 The principal and the agent enter into a contractual arrangement, in which the principal chooses to delegate certain functions to the agent in the expectation that the agent will act in ways that produce outcomes desired by the principal. Delegation to supranational institutions constitutes an active choice between alternative governance structures.4 In effect, all principals face the choice of whether to perform the desired functions ‘in house’, or to ‘out-source’ them, to use modern management terminology. States can either attempt to address pressing problems on their own, in purely intergovernmental co-operation with others, or by delegating political authority to supranational institutions. This suggestive trichotomy illustrates that the decision facing principals is not one of either/or, but rather a choice between a large number of alternative governance structures, where no and full delegation merely mark the outer parameters. Delegation involves both costs and benefits for the delegating party and, in a rational world, the relative attractiveness of alternative governance structures is determined by the balance between the two. In the P-A literature, the disadvantages of delegation are generally referred to as agency costs, which arise from the setting up of agents, the construction of mechanisms to control agents, and the negative effects of agents straying beyond their mandates. The benefits of delegation lie in the reduction of political transaction costs, by providing solutions to collective-action problems that prevent efficient political exchange. The benefits may be grouped in four general categories: (a) facilitating credible policy commitments, as agents allow politicians to jointly tie their hands; (b) reducing information asymmetries, as agents develop and employ policy-relevant expertise; (c) improving decision making efficiency, as agents manage detailed rule making, thus saving politicians’ time and effort for more general policy decisions; and (d) shifting blame for unpopular decisions and policy failures, as agents get to carry the consequences, thus allowing politicians to escape electoral punishment.5 The four sets of benefits translate into hypotheses about what kind of functions states are likely to delegate to supranational institutions. That said, we should keep in mind that any specific case of delegation may involve more than one of these rationales. The Delegation of Supranational Powers in the EU The most stable finding of P-A analysis in the EU context is probably the close fit between the functionalist predictions of rational institutionalism and the actual powers of the EU’s supranational institutions. As shown most comprehensively by Mark Pollack, the Commission and the ECJ have been delegated judicial and political competencies that serve to ameliorate the typical functional problems.6 Below, these observations are analysed according to the four functional purposes isolated above. First, in an act of self-commitment, member states have entrusted both the Commission and the ECJ with enforcement powers that effectively encourage

Delegation to supranational institutions

23

compliance, thus reducing the problem of credible commitments.7 In its role as the ‘guardian of the treaties’, the Commission has the power to initiate infringement proceedings against non-complying states. The ECJ has the final say in such proceedings, and ultimately determines whether a member state is in compliance or not. To the extent that states disregard the ECJ’s infringement judgments, the two institutions jointly possess the power to sanction governments through economic penalties. In addition to these powers at the centralised EU level, the ECJ plays a crucial role in the decentralised enforcement of EC law by citizens in national courts. Second, the Commission’s powers of initiation and execution are both intimately bound up with the development of policy expertise. When producing proposals for new EU legislation, the Commission employs the technical expertise of its specialised directorates-general and its extensive network of expert committees. When the Commission chisels out detailed implementation legislation, its policy expertise is exploited for decisions whose technical character make them unsuitable for the politicians in the Council. Moreover, the Commission’s general level of policy expertise is used on an everyday basis in Council negotiations, when called upon to clarify the likely effects of possible changes to existing proposals. In all cases, the Commission generates and employs expert knowledge, thus mitigating information asymmetries that otherwise could have resulted in less informed policy making. Third, both institutions have been delegated powers that enhance the efficiency of EU decision making.8 Rather than member governments negotiating the minute details of all policy proposals in the Council, they adopt rules and guidelines whose specific content is often worked out through the Commission’s implementation measures. The Commission’s executive authority to take certain administrative decisions independently similarly relieves EU governments of having to negotiate and decide on detailed and often insignificant matters. The Commission’s traditional position as mediator in the Council, and more recent function as broker between the Council and the European Parliament, further enhances the efficiency of bargaining in the EU. The fact that the treaties of the EU are framework treaties, only spelling out the general principles that apply for decision making in the particular policy domains, rather than outlining all applicable substantive rules, grants the ECJ a key function in enhancing decision making efficiency. Through its interpretation of EU treaties and legislation, the ECJ fills in the fine print of these incomplete contracts. Fourth, the Commission’s and the ECJ’s close involvement in the EU’s political process offers extensive scope for blame shifting, to the advantage of EU governments.9 The blame for policy failures, as well as for uncomfortable but necessary decisions, can be shifted onto the supranational institutions, which often lack both the capacity and the interest to shift it back. In view of their preference for ‘more Europe’, and the absence of electoral pressure, the supranational institutions are ideal scapegoats for unpopular policy developments in the EU. By the same token, member governments can claim credit for popular policy developments, with only limited regard to their own involvement in bringing them about. Rather than doing it themselves, the member states of the EU have delegated powers of initiation, execution, interpretation, and enforcement to the Commission and the ECJ in a wide range of policy domains. For each new treaty that has been concluded, the supranational institutions have either been delegated more far-reaching competencies in

The politics of delegation

24

existing issue areas, or had their functional tasks extended to new issue areas. In the section on the dynamics of delegation, the analysis focuses on how the experiences of previous rounds of interaction are integrated into later decisions about delegation, thus producing complex patterns of variation in delegation over time and across issue areas. THE INSTITUTIONAL DESIGN OF DELEGATION The decision to delegate political authority to supranational institutions is intimately linked to the question of institutional design, which constitutes a second analytical stage in the delegation process. The key question is how national governments go about ensuring that the supranational institutions act in ways that produce the outcomes motivating the delegation of powers. This section argues that the kind of functions delegated to the EU’s supranational institutions heavily condition the control mechanisms that national governments can design and have designed. The Principal’s Control Problem Every decision to delegate essentially involves two choices—what powers to delegate and what institutional control mechanisms to craft. The need for principals to establish control mechanisms flows from the central assumption in P-A analysis that all acts of delegation are inherently problematic, because of the simultaneous presence of conflicting preferences and information asymmetries at an analytical ex ante stage. Not only are agents likely to have preferences that diverge from those of their principals, but they also tend to know more about their own interests and actions than their principals do, granting them an informational advantage. The combined effect is a strategic setting offering both an incentive and an opportunity for agents to pursue their own preferences at the expense of principals’—to ‘shirk’, in the P-A vocabulary. The essence of the principal’s problem is the design of institutional control mechanisms that will induce the desired behaviour on the part of the agent. The P-A literature stresses monitoring and sanctions as the two main components of such an institutional design. Monitoring affects agent behaviour by making it less likely that shirking will go unnoticed. Sanctions, in turn, encourage the agent to fulfil its functions faithfully, by raising the costs of non-compliance. What truly makes delegation a dilemma is the fact that its very rationale may prevent government principals from establishing effective control mechanisms. Certain functions commonly delegated require that agents enjoy substantive levels of discretion in the execution of their powers; that supranational agents to some degree are truly autonomous from their member state principals. In this sense, the design of institutional control mechanisms is shaped by functional concerns, since the initial powers delegated to an agent heavily condition the oversight mechanisms that principals may employ.10 In relation to the four functions isolated in the previous section, the main dividing line is between powers that serve to reduce the problem of credible commitments, and powers whose purpose it is to provide technical expertise, enhance decision making efficiency, and shift blame for unpopular decisions. In the first case, the very rationale of delegation is for member states to collectively tie their hands by conferring authority on actors that

Delegation to supranational institutions

25

cannot be directly controlled. The act of self-commitment is only meaningful to the extent that supranational agents enjoy extensive discretion in the execution of their functions and do not face the immediate threat of having their decisions overturned by government principals. To exemplify with one common form of political power that falls within this category, monitoring of compliance is of limited value if those whose compliance is being monitored in turn control the monitor. In the case of the other three reasons for delegation, control mechanisms do not necessarily compromise the agent’s capacity to fulfil the functions it has been entrusted. Supranational institutions may build policy expertise, engage in detailed rule making, and attract blame even in the presence of some form of institutional control. For these reasons, we would expect member state principals to allow a greater degree of discretion when delegating powers that reduce problems of credible commitment, while tightening control mechanisms in all other cases. The Functional Basis of EU Control Mechanisms As predicted by the rational institutionalist perspective on delegation, EU governments have designed control mechanisms to ensure that the supranational institutions do not exploit the discretion they have been delegated. But more important than confirming the existence of control mechanisms is noting the distinct patterns of variation in the oversight instruments that apply to the Commission and the ECJ, even though both institutions were set up at the same time, by the same member states, and in the same spirit of supranationalism. These differences reflect the nature of the particular functions delegated to the two institutions. The control mechanisms designed by member governments vary systematically depending on what functional problems the Commission and the ECJ have been set to mitigate or resolve. When enhancing the credibility of member state commitments, by enforcing compliance and safeguarding the basic constitutional treaty, the supranational institutions are typically relieved of intrusive control mechanisms. When providing technical expertise or enhancing the efficiency of decision making by engaging in regulation, implementation, and policy interpretation, the Commission and the ECJ are subject to closer monitoring and more credible threats of sanctions. Below, a stylised account is provided of the primary means of control that pertain to the Commission’s powers of policy initiation, execution, and enforcement, as well as the ECJ’s powers of rule interpretation and enforcement.11 The monopoly granted to the Commission in the initiation of new EU legislation is carefully matched by the control possibilities flowing from the requirement of member state approval in the Council. The need to mobilise support forces the Commission to seek informal contacts with member states in the preparation of legislation. This grants governments insight into internal Commission policy formulation and an ability to shape the outcome by signalling their preferences. The final presentation of a Commission proposal to the Council both decreases informational asymmetries and offers governments the possibility to sanction the institution by dismissing the initiative, if the Commission has not been sufficiently attentive to their preferences. Should member governments agree to a proposal whose long-term effects they do not understand at the time of adoption, they can always re-legislate, which in most cases only requires the

The politics of delegation

26

support of a qualified majority. By any measure, the Commission is closely watched in its initiative function, and quite susceptible to ex post corrections by member governments. When the Commission executes EU policy by taking decisions that implement Council legislation, the primary instrument of control is the comitology system of oversight committees.12 The system was explicitly established to allow EU governments a voice in Commission implementation, and entails that the Commission must present its draft decisions for approval in the committees before adopting them. The degree of control differs depending on whether the policy area in question is subject to advisory committees, management committees, or regulatory committees. Whereas advisory committees can only counsel the Commission, management committees can block implementation measures by a qualified majority and thereby have them referred to the Council for decision. Regulatory committees, finally, must approve Commission acts by a qualified majority, or else they are referred to the Council. Slightly simplified, the more sensitive the policy area, the more likely it is to fall under a restrictive control procedure. This variation notwithstanding, the comitology committees offer an everyday system for monitoring and sanctioning the Commission in the execution of its implementation powers. In contrast to its position in policy initiation and execution, the Commission is freed of control mechanisms when fulfilling its function as guardian of the treaties. No means or procedures exist by which EU governments are allowed a role in the Commission’s monitoring of member state compliance. Acting on complaints from citizens or on the results of its own inquiries, the Commission independently decides whether to initiate infringement proceedings against states suspected of violating EU rules. Similarly, no form of member state approval is required when the Commission fixes the penalties that EU governments must pay if found guilty of not having implemented the ECJ’s infringement judgments. By all accounts, member states have taken a hands-off approach in the institutional design pertaining to the Commission’s enforcement function. Governments’ means for controlling the ECJ in its interpretation of EU rules closely reflect the alternative logics of credible commitments and efficient decision making. Depending on whether the ECJ interprets constitutional rules or EU legislation, it is subject to more or less intrusive control mechanisms. When interpreting the treaties and functioning as a constitutional court, the ECJ is relieved of control mechanisms. This absence of oversight instruments is typical of constitutional courts, which themselves interpret their procedural domain.13 By contrast, the ECJ is subject to a form of control when interpreting EU legislation and functioning as a statute-reviewing court. To the extent that member governments do not share the ECJ’s reading of directives and regulations, they can always correct its interpretation by re-legislating, using the ordinary procedures and decision rules. In this function, the ECJ shares its relatively vulnerable position with other courts engaged in the judicial review of legislation, since the rewriting of statutes is one form of control instrument typically available to national legislatures.14 Similarly to the Commission, the ECJ can conduct its part in the enforcement of state compliance without any form of government control. Regardless of whether the ECJ assesses compliance with treaty rules or EU legislation, governments lack instruments for reversing its decisions, short of treaty revision. With the entry into force of the Maastricht Treaty in 1993, this act of self-commitment was further reinforced, as the ECJ was

Delegation to supranational institutions

27

granted the power, without the need for government approval, to financially sanction states that neglect its infringement judgments. In this rational institutionalist account, function constitutes the antecedent or underlying factor, shaping EU governments’ design of control mechanisms. Variation in control mechanisms across the functions of the same institution is explained by the degree of discretion that is required to solve the functional problem in question, as is correspondence in control mechanisms between two institutions that differ in constitutive terms but fulfil the same function. This basic logic carries further implications for our understanding of variation in control across individual institutions, such as the Commission and the ECJ. Since many institutions are vested with more than one function, they should be thought of as functional composites, whose specific combination of functions shapes the overall level of control. This may explain why member governments are sometimes considered to enjoy more intrusive forms of control in relation to the Commission than with regard to the ECJ. The Commission is comparatively more engaged in providing policy expertise and facilitating decision making than in securing credible commitments. For the ECJ, the reverse is true. Though its contribution to the enhancement of efficiency should not be neglected, the functional emphasis is clearly on providing effective forms of self-commitment. THE CONSEQUENCES OF DELEGATION The third stage in the process of delegation is its consequences. Does the delegation of public authority to supranational institutions generate the political benefits that first motivated member states to take this step, or does it result in effects that were not foreseen? Just as function conditions the use of control mechanisms, the institutional design in turn shapes the consequences of delegation. In this section, it is contended that the control instruments established by EU governments have been insufficient to prevent the supranational institutions, especially the ECJ, from exploiting their discretion, with constitutional, substantive, and legitimacy effects. Anticipated Benefits and Unwanted Shirking Delegation is likely to result in both desired and undesired effects. The benefit of delegation is the value to EU governments of the functions performed by the supranational institutions on their behalf. These are the anticipated consequences that motivated the delegation of power in the first place. The undesired effects are those agency costs or losses that result when an agent pursues its own preferences at the expense of the principal’s. Rather than encouraging an assessment of the benefits of delegation—which are assumed to be positive—P-A theory directs out attention towards the problem of agents exploiting their discretion. Slightly simplified, any given decision by an agent may be of three kinds: in line with its principal’s policy goals; different from the principal’s policy goals; or different from the principal’s policy goals, but reversed by the principal’s control mechanisms.15 It is the second category of actions, where supranational agents

The politics of delegation

28

succeed in pursuing their own agendas without being reined in by member governments, that result in unwelcome consequences of delegation. What these consequences consist of in concrete terms depends on the political context and the agent’s policy preferences. The likelihood of negative consequences of delegation is shaped by the institutional design. Whether by economic logic, accident or design, member governments may have set up control mechanisms that are incapable of preventing or correcting shirking by their supranational agents. The analytically most interesting case is the form of deliberate abstention from intrusive control instruments that was discussed in the previous section. To facilitate credible commitments, government principals allow their agents an unusual degree of discretion, which in turn may be exploited by the supranational institutions for the purpose of pursuing their own political objectives. At a more general level, and regardless of what functions agents have been delegated, it also tends to be prohibitively costly for principals to operate control mechanisms that eliminate all forms of shirking. Indeed, in economic terms, it is only rational for principals to invest in control to the point where the marginal benefit of better agent compliance equals the marginal cost of running the mechanisms. Supranational Discretion and Effects on Integration Large parts of the existing research on European integration can be read as an analysis of the effects of sovereign states having delegated powers to joint supranational institutions. In fact, neofunctionalism and intergovernmentalism—the two dominating theoretical perspectives in the study of European integration—may be interpreted as two alternative answers to the question of delegation effects. Neofunctionalists generally suggest that EU governments, with the initial delegation of power, created a set of supranational institutions whose later actions have contributed to unanticipated integration effects.16 In positivistic jargon, the supranational institutions have exerted independent causal influence on the process of integration. Intergovernmentalists, by contrast, tend to conceive of the supranational institutions as essentially passive devices facilitating intergovernmental bargaining, by offering member states a possibility to secure credible commitments.17 The institutions simply fulfil the functions delegated to them by member governments, which remain firmly in control of the process of integration. That the delegation of political power to the supranational institutions has been beneficial for co-operation in the EU is contested by neither neofunctionalists nor intergovernmentalists. Expressed in P-A terms, the point of contention is the magnitude of the agency costs. To what extent has delegation generated effects that are unanticipated and undesirable from the point of view of EU governments? One of the primary merits of importing P-A analysis into the study of European integration has been its capacity to provide conditional answers to this question, by way of identifying and explaining variation in supranational influence.18 Owing partly to these efforts, few would disagree today with the assertion that the institutions at a number of historical occasions have exploited their discretion for purposes of advancing private interests. Among the best researched cases are the ECJ’s transformation of the European legal system in the early 1960s, and the Commission’s strategic launching of the internal market initiative in the mid-1980s.19

Delegation to supranational institutions

29

The ECJ’s greater capacity to implement a private agenda is the most important finding in comparisons between the two supranational institutions.20 This observation accords well with the predictions generated by P-A analysis on the likelihood of shirking under alternative institutional designs. The relative absence of intrusive control mechanisms grants the ECJ a degree of discretion which exceeds that of the more tightly controlled Commission, thus facilitating a comparatively bolder interpretation of the delegated mandate. It is equally symptomatic that the most widely cited cases of ECJ activism took place in the interpretation of constitutional treaty provisions, where member states’ means of control are weaker than in the interpretation of secondary law. The ECJ has learned to exploit diverging member state positions for the purpose of implementing agendas it knows governments cannot undo. To reverse the ECJ’s treaty interpretations, all EU governments must agree on the suggested revisions of the Treaty at an intergovernmental conference (IGC), and all member states must ratify the new Treaty. Historically, no attempt to rewrite the ECJ’s treaty interpretations, or to sanction the institution directly, has ever succeeded, though not for lack of trying.21 In essence, the ECJ has learned to count on the support of at least one member state, and, on this basis, developed techniques for introducing controversial doctrines that secure the institution against sanctions.22 The Commission, by contrast, operates in a world with more immediate and credible threats of sanctions. Rationally anticipating the risk of having its legislative proposals dismissed by the Council, or of having its executive decisions referred to the Council from the comitology committees, the Commission tends to present only proposals with a high likelihood of adoption.23 Interestingly, but hardly surprisingly, existing research shows that the Commission enjoys an unusual capacity to pursue private interests when drawing on the support of the less constrained ECJ.24 When the supranational institutions succeed in pursuing their own interests, these private agendas not only diverge from member state preferences, but also translate particular political objectives into concrete political effects. In an attempt to contextualise the unanticipated consequences of delegation, a distinction is made between constitutional effects, substantive effects, and effects on legitimacy. All three are direct consequences of political outcomes being biased in the orientation of the supranational institutions. Constitutional effects are fundamental and lasting changes to relations of power and accountability in the EU. Acting on visions of a federal Europe, the supranational institutions have worked to promote developments in the direction of a European constitution, a European constitutional court, a European catalogue of rights, and a federal-type division of competencies between European, national, and regional levels. The most prominent case is the ECJ’s transformation of the European legal system, and the associated promotion of itself to the position of a de facto constitutional court. By introducing the doctrines of direct effect and EC law supremacy in the early 1960s, the ECJ turned the EU’s preliminary ruling system from a mechanism that allowed individuals to challenge EC law in national courts into a means for contesting national law and enforcing EC law in national courts. EC law not only trumped national law when in conflict, but also created legally enforceable rights for individuals. National courts became the linchpins of the European legal system and entered into a symbiotic relationship with the ECJ, where the European court interpreted EC law, while national

The politics of delegation

30

courts referred cases to the ECJ and later applied its interpretation to the facts of these cases. The introduction of these doctrines by the ECJ turned the relationship between the ‘international organisation’ and the ‘sovereign member state’ on its head. But, in addition, it created the basis for future activism on the part of the ECJ, by ensuring a continuous flow of critical cases through the transformed preliminary ruling procedure. Substantive effects are stable biases in policy output along salient dimensions of political contestation. In the EU case, supranational exploitation of discretion has produced discernible effects on at least three such dimensions: European integration vs. state autonomy, individual rights vs. state authority, and left vs. right. On the first dimension, independent influence by the supranational institutions has decidedly favoured policy solutions that place the locus of interest mediation at the European level, and involve binding EU rules rather than intergovernmental co-ordination. Existing research on the supranational institutions suggests that this may be the most pronounced substantive policy bias.25 On the second dimension, supranational influence has tended to enhance the legal rights and opportunities of individuals and firms in relation to the state. In fact, much of the ECJ’s case law can be read as a crusade to protect individuals’ EU rights, for instance, by way of creating legal remedies that can be used against member governments in national courts.26 On the third dimension, it is considerably more difficult to isolate a stable policy bias, not least because of internal ideological conflict within the institutions.27 Rather than promoting either regulation and positive integration or deregulation and negative integration, the EU’s supranational institutions have espoused a form of regulated competition, which combines market liberalisation with re-regulation at the European level.28 Legitimacy effects are changes in the popular perception of the legitimacy of delegation from national majoritarian organs to supranational non-majoritarian institutions. Though prominent in the discourse on democracy in the EU, legitimacy questions have so far taken a back seat in the scholarly debate on supranational agency. But to the extent that legitimacy effects constitute agency losses for national governments, they are relevant even in positive P-A analysis.29 There are few indications, however, that the supranational institutions’ exploitation of discretion has produced stable effects that diverge from underlying popular perceptions of the legitimacy of the EU as such. One interpretation in existing research stresses the lack of political attention paid to the ECJ, safely ‘[t]ucked away in the fairyland Duchy of Luxembourg and blessed, until recently, with benign neglect by the powers that be and the mass media’.30 The competing nature of input- and output-oriented legitimacy standards offers an alternative explanation.31 In simplified terms, the same act of supranational shirking may be interpreted in both positive and negative legitimacy terms, depending on the standards of evaluation. In an input-oriented perspective, delegation becomes less legitimate if supranational institutions, with no independent source of democratic legitimacy, stray beyond the mandates set by elected national governments. In an output-oriented perspective, by contrast, delegation may actually gain in legitimacy from supranational shirking, to the extent that the effects are more favourable—for individuals’ EU rights for instance—than in the absence of independent supranational influence.

Delegation to supranational institutions

31

THE DYNAMICS OF DELEGATION The final link in the causal chain joining the rationale, institutional design, and consequences of delegation is the feed-back loop from effects to new instances of delegation. While often portrayed as such, P-A relationships are seldom static. Rather, delegation tends to be an ongoing process, in general, as well as in the EU case. This section develops the notion of dynamic delegation and demonstrates how EU governments have integrated the lessons learned in previous interaction into future delegation, institutional design, and behaviour. Feed-back Loops in P-A Relationships In a functional world where the anticipated effects of delegation constitute the reason for conferring authority in the first place, the experiences of previous rounds of delegation invariably affect future rounds. In reality, P-A relationships are seldom a one-shot operation, where all relevant contracting action is concentrated to an ex ante stage, and where the agents’ ex post behaviour is a direct consequence of how successful the principals were in providing adequate control mechanisms. Rather, principal-agent relationships tend to be dynamic and interactive, subject to bargaining and revision by the parties.32 At the heart of this dynamism is the notion of strategic adaptation to unintended consequences. From the point of view of political principals, delegation may be flawed because it generates unexpectedly large agency costs, or because political agents, on the contrary, do not possess the powers or the discretion required to produce the desired effects. In this sense, the more general rational institutionalist argument that unintended and sub-optimal consequences of delegation can be corrected at later stages applies also to dynamic P-A relationships. As emphasised by Oliver Williamson: ‘Once the unanticipated consequences are understood, those effects will thereafter be anticipated and the ramifications can be folded back into the organisational design. Unwanted costs will then be mitigated and unanticipated benefits will be enhanced. Better…performance will ordinarily result.’33 If principals and agents are rational, how is it that the consequences of delegation are not fully anticipated and prevented in the initial contract? In essence, why do they need to learn? In a rational institutionalist reading of dynamic delegation, principals and agents are conceived of as boundedly rational, acting on the basis of available rather than full information. Even if government principals realise that the supranational institutions they set up and empower are likely to develop their own private agendas, they are unable to foresee the exact shape of the strategic setting. Moreover, structural changes, such as increases or decreases in the number of principals, may cause shifts in the strategic environment over time. The integration of previous experiences may both affect future behaviour in the same area where they were generated, and feed into strategic judgements about new instances of delegation. Moreover, the experiences may pertain to both the appropriateness of

The politics of delegation

32

delegating powers or the effectiveness of existing control mechanisms in producing the desired effects. Dynamic Delegation in the EU In view of the EU’s half-century-long history of delegation, involving a number of renegotiations of the original treaties, this empirical context ought to be rife with examples of feed-back loops, to the extent that the propositions about dynamic interaction are valid. A closer look at historical junctures in European integration supports such an interpretation. Below, three kinds of feed-back loops that are particularly prominent in the EU are distinguished: (1) member governments integrating the experience of unwelcome supranational exploitation of discretion into the delegation and institutional design in new policy areas; (2) EU governments boosting the powers of the supranational institutions in existing policy areas, in response to the incapacity of previous delegation to achieve the desired effects; and (3) national governments adapting to a strategic political context in which the basic rules of the game have been dislocated by supranational shirking. Given the EU’s history of occasional supranational shirking, it is not surprising that one of the identifiable feed-back effects is the search by national governments for means of preventing this from happening again. Because of the joint-decision trap identified by Fritz Scharpf, these efforts tend to influence delegation and institutional design in new areas of co-operation, rather than in the areas afflicted by supranational shirking.34 When EU governments decide to co-operate in a new policy area, those in favour of less farreaching delegation and stricter control mechanisms are privileged by the requirement of unanimous approval, whereas the situation is the opposite in existing areas of cooperation. Two excellent examples of this kind of feed-back loop are the institutional arrangements established in the area of justice and home affairs in 1991, and the recent creation of the ‘open method of co-ordination’ for co-operation in employment, social affairs, and education. In both cases, existing research tends to explain the adoption of these particular institutional arrangements with the reluctance among member states to compromise, and risk further violations of, national sovereignty.35 Despite the fact that these policy areas fall under the general rubric of regulation, which otherwise is synonymous with the full delegation of powers to the institutions, EU governments settled for less ambitious and more constraining arrangements. When justice and home affairs was first brought into the EU framework as an area of common interest in the Maastricht Treaty, the Commission was given no right of initiative in the most sensitive areas, and only shared initiative in the others. The ECJ, for its part, was given no role whatsoever in this area. Similarly, the open method of co-ordination, first introduced in the employment chapter of the 1997 Amsterdam Treaty, grants but a limited role of initiation and supervision to the Commission, and no authority at all to the ECJ. In both cases, the capacity to prevent unwanted effects was further enhanced by the requirement of unanimous approval of the Commission’s proposals and the non-binding character of the legal instruments in these areas. A second and opposite form of learning effect consists of national governments realising that they have been too restrictive in the delegation of powers. Rather than

Delegation to supranational institutions

33

limiting the institutions’ powers, EU governments extend them, after having experienced that the desired effects cannot be achieved unless more far-reaching powers are delegated. Two particularly revealing cases are the subsequent changes to the institutional arrangements in justice and home affairs in 1997, and the up-grading of the Commission’s and the ECJ’s enforcement powers in the early 1990s. The second chapter in the story about delegation in justice and home affairs is EU governments’ correction in the Amsterdam Treaty of mistakes committed when first setting up the institutional arrangements in Maastricht. To use an analogy from the world of monetary policy, EU governments had managed to overshoot the target when taking measures to limit supranational shirking, thereby stifling meaningful policy progress in the area. Through an exceedingly elaborate arrangement, the Amsterdam Treaty moved in the direction of greater supranational involvement, by allowing the Commission a more prominent position in the initiation of proposals and the ECJ a restricted role in the interpretation of rules. At the same time, the up-grading of powers was coupled with new and intricate control mechanisms.36 EU governments’ decision in Maastricht to boost the supranational institutions’ enforcement powers constituted a similar reaction to the inadequacy of existing arrangements.37 In this case, the enforcement powers originally delegated to the institutions in 1957 had gradually become insufficient to fulfil the function of securing credible commitments. With the completion of the internal market, and the increasing incentives for member states to free-ride, non-compliance with EU rules had spiralled. To ameliorate these problems, EU governments agreed to delegate the power to financially sanction non-complying member states, despite the obvious encroachments on national sovereignty. The third form of feed-back effect consists of member governments adapting to a strategic context in which the basic rules of the game have been dislocated by supranational shirking.38 Supranational actions with constitutional dimensions fundamentally reconstruct the political environment in which principals and agents operate and interact. In simple terms, the political world will never become the same again, and this is a reality to which governments adjust, for instance, by ratifying the moves of the supranational institutions and transforming national regimes. The ECJ’s constitutionalisation of the treaty is the primary source of such dynamic effects, which alter the formal relationship between government principals and supranational agents. This feed-back loop may take a number of forms, as evidenced by existing research on the European constitutionalisation process. The ECJ can reinterpret treaty provisions so as to enhance its own powers and those of its allies, be they national courts or the Commission, henceforth reducing the relative position of member governments in the EU political system. The most prominent example is the previously mentioned transformation of the preliminary ruling procedure into a means for contesting national law and enforcing EC law in national courts. Whereas, initially, many national governments and higher national courts opposed the ECJ’s interpretation, the recalcitrance gradually gave away to the realisation that this new judicial and political order had already become institutionalised, could not be ndone, and essentially constituted a fact of life.39 A related form of feed-back effect results when the ECJ constitutionalises the provisions of an EU directive, that is, interprets these provisions as being inherent in the

The politics of delegation

34

treaty. The Council is thereby displaced as the site of future rule reversal in this policy domain, since only governments reassembled as a constituent assembly at an intergovernmental conference can change the ECJ’s treaty interpretations. Instead, member governments frequently amend EU directives and national law to conform to the ECJ’s rulings. As Alec Stone Sweet and James Caporaso note in a study tracing the evolution of EU social policy: the Court has supplanted the Council as the locus of lawmaking on more than one occasion, enacting legislative provisions that had stalled in the Council under unanimity voting. Lacking the unanimity necessary to reverse the Court in this area, the member-state governments have been forced to adjust to the Court’s case law, by ratifying the ECJ’s policy choices in Council directives and by revising national legal regimes.40 Finally, the ECJ alters the conditions for future interaction when interpreting Treaty provisions so as to give new rights to individuals, since this recasts the enforcement environment and EU governments’ capacity to get away with non-compliance. When conferring enforceable rights on individuals, the ECJ opens up the possibility of decentralised enforcement through individuals securing their EU rights in national courts, next to centralised enforcement through Commission infringement proceedings. The ECJ’s establishment of the principle of state liability in 1991 is a case in point.41 By granting individuals the right to financial compensation for damage suffered as a result of state non-compliance, the ECJ introduced a decentralised form of sanctions that force governments to think twice about violating EC law and about agreeing to legislative proposals they actually oppose. CONCLUSION This chapter has explored the dynamic character of delegation by developing the functional logic that links the different stages of the delegation process to each other. In four steps, it has demonstrated (1) how the promise of effective solutions to collective action problems motivated the delegation of political authority to supranational institutions; (2) how the particular functions delegated influenced EU governments’ design of the mechanisms for controlling the Commission and the ECJ; (3) how the design of control mechanisms conditioned the effects of delegation by leaving comparatively greater scope for independent influence on the part of the ECJ; and (4) how the consequences of delegation through complex feed-back loops fed into new rounds of delegation, institutional design, and principal-agent interaction. In conclusion, two points related to the general explanatory power of this rational institutionalist perspective can be raised. They both suggest areas for further research on delegation. First, is the P-A approach capable of addressing processes of international institutional design outside the EU context? While the predominant focus of the recent wave of P-A analysis has been the EU, the theoretical contribution of this literature may reach beyond this particular empirical context. The EU is not the only case where governments have set up and delegated functions to supranational institutions,

Delegation to supranational institutions

35

simultaneously creating a potential control problem. Secretariats and dispute-settlement bodies of international organisations, such as the WTO, the IMF, the UN, and the IAEA, generally enjoy certain powers of representation, initiation, execution, and supervision.42 What makes the EU institutions unique in this comparative perspective is the range of the powers delegated, not the act of delegation itself. Though serious gaps in control are less likely to be found elsewhere, there is nothing inherent in P-A analysis that limits the use of this theoretical approach to the study of the EU. Rather, it may be employed to explain delegation decisions, design, and consequences whenever governments confer powers to international institutions. But the rational institutionalist perspective on delegation may also generate important insights in cases where delegation does not take place, or only takes place gradually. As Andrew Moravcsik notes, ‘governments often refuse to assume the political risk of delegation, preferring instead imperfect enforcement and inefficient decision-making, to the surrender of sovereignty’.43 The relationship between functional demands for delegation and concerns for national sovereignty is an area that begs further exploration. Process tracing of shifts over time in the underlying costs and benefits of alternative governance structures is likely to be particularly rewarding. Equally relevant are cases where delegation never occurs despite strong functional demands; cases where entirely new supranational institutions are set up, such as the International Criminal Court; and cases that involve gradual extensions of delegation, such as the transformation of the GATT/WTO dispute-settlement mechanism. Second, to what extent can processes of delegation—in general, as well as in the EU— be explained by competing theoretical perspectives? P-A analysis has so far been largely unchallenged as an explanation of delegation, which neither lends credibility to the claims of P-A theorists, nor encourages theoretical refinement. Given that delegation constitutes a form of institutional choice and design, sociological institutionalism may be conceived of as an alternative to the functional perspective. In a sociological institutionalist reading, the spread of delegation to non-majoritarian institutions would be conceptualised as a case of isomorphism, resulting from social processes of emulation and diffusion, where politicians replicate organisational models collectively sanctioned as appropriate and legitimate.44 While full assessments of the explanatory power of sociological institutionalism in the EU case should be a topic for future research, preliminary observations yield mixed results. The original delegation of powers to the EU’s supranational institutions took place at a time when strictly intergovernmental co-operation was the standard organisational model, resulting in post-war institutions such as the Council of Europe, the Organisation for European Economic Co-operation (OEEC), Nato, and Efta. Rather than an additional example of generally embraced forms of institutional design, the EU was an odd bird, whose supranational character even succeeded in deterring a range of natural partners from joining. If little speaks in favour of isomorphism as an explanation of the original delegation of powers in the EU, there are greater reasons to examine its role in the growing international popularity of delegation today, which may or may not be linked to the experiences of regional integration in Europe. Over the years, the EU has clearly become a less unique case of delegation. Is this pattern best explained by the functional demands of governance beyond the nation state, or by social processes of emulation? The design of

The politics of delegation

36

control mechanisms constitutes yet another area where sociological institutionalism might offer interpretations that diverge from those privileged in this chapter. While, again, predicting the same outcome, sociological institutionalists would explain the observed similarity between the ECJ on the one hand, and constitutional and statutereviewing courts on the other, by the diffusion of models of institutional design, rather than functional concerns. These are areas that should claim the attention of future work, and where contributions based on alternative theoretical perspectives could add to the expanding body of research on delegation. NOTES For comments on earlier versions of this chapter, I am grateful to Marian Döhler, Kathleen McNamara, Mark Pollack, Alec Stone Sweet, Mark Thatcher, Stephen Wilks, an anonymous reviewer, and members of the International Relations seminar at Lund University. 1. I refer to the EU by its current name, except where I relate explicitly to its law, which is referred to as EC (European Community) law. 2. See, e.g., A.Moravcsik, ‘Preferences and Power in the European Community: A Liberal Intergovernmentalist Perspective’, Journal of Common Market Studies 31/4 (1993), pp. 473– 524; A.Moravcsik, The Choice for Europe: Social Purpose and State Power from Messina to Maastricht (Ithaca: Cornell University Press 1998); M.Pollack, ‘Delegation, Agency, and Agenda-Setting in the European Community’, International Organization 51/1 (1997), pp. 99–134; M.Pollack, ‘The Engines of Integration? Supranational Autonomy and Influence in the European Union’, in W.Sandholtz and A.Stone Sweet (eds.), European Integration and Supranational Governance (Oxford: Oxford University Press 1998); M.Pollack, The Engines of European Integration—Delegation, Agency, and Agenda Setting in the EU (Oxford: Oxford University Press 2002); K.Alter, ‘Who are the “Masters of the Treaty”?: European Governments and the European Court of Justice’, International Organization 52/1 (1998), pp. 121–47; A.Stone Sweet and J.Caporaso, ‘From Free Trade to Supranational Polity: The European Court and Integration’, in Sandholtz and Stone Sweet (eds.), European Integration and Supranational Governance; J.Tallberg, ‘Supranational Influence in EU Enforcement: The ECJ and the Principle of State Liability’, Journal of European Public Policy 7/1 (2000), pp. 104–21; J.Tallberg, ‘The Anatomy of Autonomy: An Institutional Account of Variation in Supranational Influence’, Journal of Common Market Studies 38/5 (2000), pp. 843–64; J.Tallberg, European Governance and Supranational Institutions: Making States Comply (London: Routledge 2003); T.Doleys, ‘Member States and the European Commission: Theoretical Insights from the New Economics of Organization’, Journal of European Public Policy 7/4 (2000), pp. 532–53; S.Stetter, ‘Regulating Migration: Authority Delegation in Justice and Home Affairs’, Journal of European Public Policy 7/1 (2000), pp. 80–103; S.K.Schmidt, ‘Only an Agenda Setter? The European Commission’s Power over the Council of Ministers, European Union Politics 1/1 (2000), pp. 37–61; F.Franchino, ‘Control of the Commission’s Executive Functions: Uncertainty, Conflict, and Decision Rules’, European Union Politics 1/1 (2000), pp. 63–92. 3. S.Ross, ‘The Economic Theory of Agency: The Principal’s Problem’, American Economic Review 63/2 (1973), p. 134. 4. For an excellent conceptualisation of delegation as a choice between alternative governance structures, see D.Epstein and S.O’Halloran, Delegating Powers: A Transaction Costs Politics Approach to Policy Making under Separate Powers (Cambridge: Cambridge University Press 1999).

Delegation to supranational institutions

37

5. See the introduction to this volume by M.Thatcher and A.Stone Sweet, but also Epstein and O’Halloran, Delegating Powers; G.Majone, ‘The Regulatory State and its Legitimacy Problems’, West European Politics 22/1 (1999), pp. 1–24. 6. Pollack, ‘Delegation, Agency’; M.Pollack, ‘Delegation, Agency, and Agenda Setting in the Treaty of Amsterdam’, European Integration Online Papers 3/6 (1999). 7. On the delegation of enforcement powers to solve problems of credible commitments, see also G.Garrett, ‘International Cooperation and Institutional Choice: The European Community’s Internal Market’, International Organization 46/2 (1992), pp. 533–60; Moravcsik, The Choice for Europe; Tallberg, ‘Supranational Influence’. 8. On delegation to the ECJ for purposes of enhancing the efficiency of EU decision making, see also Garrett, ‘International Cooperation’. 9. On delegation as a way of facilitating blame shifting, see also M.Smith, ‘The Commission Made Me Do It: The European Commission as a Strategic Asset in Domestic Politics’, in N. Nugent (ed.), At the Heart of the Union: Studies of the European Commission (Basingstoke: Macmillan 1997); Majone, ‘The Regulatory State’; Moravcsik, The Choice for Europe. 10. On institutional function as a factor shaping the design of control mechanisms, see also the introduction by Thatcher and Stone Sweet, and Tallberg, ‘The Anatomy of Autonomy’. 11. The comparison is stylised because it isolates the general instruments of control (or absence thereof) pertaining to each of the institutions’ functions, without due regard to the specific arrangements in certain policy sectors, or to means of control that apply to all EU institutions all the time, such as the threat of Treaty revision. For discussions of EU governments’ control mechanisms, see Pollack, ‘Delegation, Agency’; Alter, ‘Who are the Masters’; Stone Sweet and Caporaso, ‘From Free Trade’; Tallberg, European Governance. 12. For a good analysis, see Franchino, ‘Control of the Commission’. Note, however, that the comitology system also has been interpreted as deliberation rather than control. See C. Joerges and J.Neyer, ‘From Intergovernmental Bargaining to Deliberative Political Process: The Constitutionalization of Comitology’, European Law Journal 3 (1997), pp. 273–99. 13. See A.Stone Sweet’s contribution in this volume. 14. See M.Shapiro’s contribution in this volume. 15. See the introduction to this volume for a more elaborate discussion. 16. E.g., W.Sandholtz and J.Zysman, ‘1992: Recasting the European Bargain’, World Politics 42/1 (1989), pp. 95–128; A.-M.Burley and W.Mattli, ‘Europe before the Court: A Political Theory of Legal Integration’, International Organization 47/1 (1993), pp. 41–76; Sandholtz and Stone Sweet, European Integration. 17. E.g., Garrett, ‘International Cooperation’; Moravcsik, The Choice for Europe. 18. For discussions about methodology and P-A analysis, see Pollack’s contribution in this volume and Pollack, ‘The Engines of Integration?’. For discussions about the merits of P-A analysis in the debate between neofunctionalists and intergovernmentalists, see Pollack, ‘Delegation, Agency’; Stone Sweet and Caporaso, ‘From Free Trade’; Tallberg, European Governance; Doleys, ‘Member States’. 19. On the ECJ’s transformation of the European legal system, see, e.g., E.Stein, ‘Lawyers, Judges, and the Making of a Transnational Constitution’, American Journal of International Law 75 (1981), pp. 1–27; J.Weiler, ‘The Transformation of Europe’, Yale Law Journal 100 (1991), pp. 2403–83; Alter, ‘Who are the Masters’. On the Commission’s part in the internal market initiative, see, e.g., Sandholtz and Zysman, ‘1992: Recasting’; D.Cameron, ‘The 1992 Initiative: Causes and Consequences’, in A.Sbragia (ed.), Euro-Politics: Institutions and Policymaking in the ‘New’ European Community (Washington, DC: Brookings 1992). 20. This theme is further developed in Tallberg, ‘The Anatomy of Autonomy’. 21. The most serious attempt to sanction the ECJ was the attack on the institution at the 1996– 97 intergovernmental conference, when the UK, followed by Germany and France, sought to reduce the ECJ’s powers and to reverse, or limit the effects of, a number of controversial judgments. See Tallberg, ‘Supranational Influence’.

The politics of delegation

38

22. For discussions of such techniques, see, e.g., T.Hartley, The Foundations of European Community Law (Oxford: Clarendon Press 1994); Alter, ‘Who are the Masters?’. 23. Pollack, ‘Delegation, Agency’; Pollack, ‘The Commission as an Agent’, in N.Nugent (ed.), At the Heart of the Union: Studies of the European Commission (Basingstoke: Macmillan 1997); Tallberg, European Governance. 24. Cf. Schmidt, ‘Only an Agenda Setter?’; Tallberg, European Governance. 25. The literature is close to unanimous on the pro-integration nature of the supranational institutions’ preferences. See, e.g., Hartley, ‘The Foundations’; G.Majone, Regulating Europe (London: Routledge 1996); L.Cram, Policy-Making in the EU: Conceptual Lenses and the Integration Process (London: Routledge 1997); Pollack, ‘Delegation, Agency’. 26. See J.Weiler and N.Lockhart, ‘“Taking Rights Seriously” Seriously: The European Court and its Fundamental Rights Jurisprudence—Part I’, Common Market Law Review 32/1 (1995), pp. 51–94; J.Weiler and N.Lockhart, ‘“Taking Rights Seriously” Seriously: The European Court and its Fundamental Rights Jurisprudence—Part II’, Common Market Law Review 32/2 (1995), pp. 579–627; A.Stone Sweet, Governing with Judges. Constitutional Politics in Europe (Oxford: Oxford University Press 2000). 27. On the left-right dimension in EU politics, and the difficulties of isolating stable preferences along this dimension for the supranational institutions, see S.Hix, ‘The Study of the European Community: The Challenge to Comparative Politics’, West European Politics 17/1 (1994), pp. 1–30; Pollack, ‘The Engines of Integration?’; and contributions by K.McNamara, S.Hix, L.Hooghe, and M.Pollack in the forum on ‘Integrating Left and Right: Studying EU Politics’, ECSA Review 11/3 (1998). 28. Cf. Majone, Regulating Europe; K.Armstrong and S.Bulmer, The Governance of the Single European Market (Manchester: Manchester University Press 1998). 29. On the problem of democratic legitimacy as an agency cost, see Majone, ‘The Regulatory State’. 30. Stein, ‘Lawyers, Judges’, p. 1. But see also, e.g., J.Weiler, ‘A Quiet Revolution: The European Court of Justice and Its Interlocutors’, Comparative Political Studies 26/4 (1994), pp. 510–34; Alter, ‘Who are the Masters’. 31. For works which illuminate these tensions, see M.Höreth, ‘No Way Out for the Be ast Unsolved Legitimacy Problem of European Governance’, Journal of European Public Policy 6/2 (1999), pp. 249–68; F.Scharpf, Governing in Europe: Effective and Democratic? (Oxford: Oxford University Press 1999); Majone, ‘The Regulatory State’. 32. On the notion of dynamic relationships in the economic P-A literature, see J.Pratt and R. Zeckhauser, ‘Principals and Agents: An Overview’, in J.Pratt and R.Zeckhauser (eds.), Principals and Agents: The Structure of Business (Boston: Harvard Business School Press 1985); B.Holmström and J.Tirole, ‘The Theory of the Firm’, in R.Schmalensee and R. Willig (eds.) Handbook of Industrial Organization, Vol. 1 (Amsterdam: North Holland 1989); D.Sappington, ‘Incentives in Principal-Agent Relationships’, Journal of Economic Perspectives 5/2 (1991), pp. 45–66. 33. O.Williamson, ‘Transaction Cost Economics and Organizational Theory’, in O.Williamson (ed.), Organization Theory: From Chester Bernard to the Present and Beyond (New York: Oxford University Press 1995), p. 216. 34. F.Scharpf, ‘The Joint-Decision Trap: Lessons from German Federalism and European Integration’, Public Administration 66 (1988), pp. 239–78.

Delegation to supranational institutions

39

35. Stetter, ‘Regulating Migration’; J.Goetschy, ‘The European Employment Strategy’, ECSA Review 13/3 (2000), pp. 4–6; J.Mosher, ‘Open Method of Coordination: Functional and Political Origins’, ECSA Review 13/3 (2000), pp. 6–7. 36. For a P-A analysis of the Amsterdam Treaty, see Pollack, ‘Delegation in the Treaty of Amsterdam’. 37. The analysis of this process is developed in Tallberg, European Governance. 38. I am indebted to Alec Stone Sweet for the conceptualisation of this feed-back effect. 39. Cf. Alter, ‘Who are the Masters’; A.Slaughter, A.Stone Sweet and J.Weiler (eds.), The European Courts and National Courts: Doctrine and Jurisprudence (Oxford: Hart Publishing 1998). 40. Stone Sweet and Caporaso, ‘From Free Trade’, p. 127. 41. The analysis of this process is developed in Tallberg, European Governance; Tallberg, ‘Supranational Influence’. 42. For a recent attempt to analyse the independent influence of a range of international organisations, see B.Reinalda and B.Verbeek (eds.), Autonomous Policy Making by International Organizations (London: Routledge 1998). 43. Moravcsik, ‘Preferences and Power’, p. 509. 44. See the chapter by K.McNamara in this volume, as well as general sociological institutionalist arguments as developed in P.DiMaggio and W.Powell, ‘The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields’, American Sociological Review 48 (1983), pp. 147–60; W.Powell and P.DiMaggio, The New Institutionalism in Organizational Analysis (Chicago: University of Chicago Press 1991); W. Scott and J.Meyer et al., Institutional Environments and Organizations: Structural Complexity and Individualism (Thousand Oaks: Sage 1994).

Rational Fictions: Central Bank Independence and the Social Logic of Delegation

KATHLEEN R.McNAMARA The insulation of central banks from the direct influence of elected officials has been one of the pre-eminent examples of the practice of delegation to non-majoritarian institutions. More countries increased the independence of their central banks during the 1990s than in any other decade since World War II.1 This wave of institutional delegation showed little regard for region, sweeping across countries as diverse as Albania, Sweden, Kazakhstan, and New Zealand. Central bank independence has been promoted by international organisations such as the OECD and the IMF as a benchmark of good governance. It was also used by European Union (EU) leaders as an obligatory criteria for entry into Economic and Monetary Union (EMU). The EU’s new European Central Bank (ECB), established in 1999, takes central bank independence to the extreme, with only weak channels of political representation and oversight. Central bank independence has achieved an almost taken for granted quality in contemporary political life, with little questioning of its logic or effectiveness. Indeed, the theoretical rationale behind the delegation of political authority to independent central banks is straightforward and appears ironclad in its logic: the preference of politicians chasing votes in the next election will be to manipulate the economy in ways that make the populace happy in the short term, disregarding the potential for their monetary policies to produce economic trouble in the long run. Thus, it seems reasonable to assume that central bank independence is a necessary solution to a functional economic policy problem, and that it is this efficiency logic that has produced the dramatic move towards increased independence over a wide swath of nations. This chapter will challenge this conventional wisdom. On the theoretical level, it will argue that advocates of central bank independence rely on a series of contestable arguments about the relationship between democracy, policy making, and economic outcomes. First, although advocates of central bank independence argue for the need to insulate monetary policy from politics, severing ties to democratic representatives and relying on technocratic expertise does not apoliticise monetary policy. Rather, delegation to independent central banks produces partisanal policies, with significant distributional effects that raise important questions of democratic accountability. Second, the government-central bank relationship does not necessarily capture the most significant

Rational fictions

41

influences on economic outcomes, such as the role of societal groups in shaping macroeconomic conditions. In regard to the empirical evidence, the chapter puts forward findings on delegation to independent central banks that cast doubt on the severity and nature of the problem purportedly solved by delegation, that is, the pernicious inflationary effects of democracy on policy making, as well as raising questions about the linkages between delegation and superior economic outcomes. If the conventional wisdom is misleading, why then have we seen the spectacular spread of central bank independence? Remarkably, delegation in this realm has been applied as a ‘one size fits all’ solution across nations even when the economic problems it is designed to address are absent. My argument is therefore that governments choose to delegate not because of narrow functional benefits but rather because delegation has important legitimising and symbolic properties which render it attractive in times of uncertainty or economic distress. The spread of central bank independence should be seen as a fundamentally social and political phenomenon, rooted in the logic of organisational mimicry and global norms of neoliberal governance. Organisational models are diffused across borders through the perceptions and actions of people seeking to replicate others’ success and legitimise their own efforts at reform by borrowing rules from other settings, even if these rules are materially inappropriate to their local needs. This dynamic is rational and instrumental, as suggested by theories of delegation within the principal-agent framework, but only when placed within a very specific cultural and historical context that legitimises delegation—the culture of neoliberalism. Moving to an independent central bank appears to shelter monetary policy from the evils of democracy and partisanship while in fact solidifying a specific set of ideologies and partisan positions which favour certain groups in society over others. The conventional justifications for delegation obscure these dynamics in ways that make central bank independence more acceptable to all. The discussion proceeds as follows. First, it investigates the theoretical literature on the topic of central bank independence, spelling out the logical arguments which form the basis for the recommendation that monetary policy be delegated to a non-majoritarian institution. It offers some critiques of the theoretical arguments that central bank independence is the most effective form of economic governance and outlines the weak empirical evidence for the merits of delegation. It then provides an overview of some of the arguments and evidence on why countries have chosen central bank independence. It contrasts the conventional wisdom, which suggests that central bank independence is adopted due to pressing functional necessity, against a literature on the sociology of organisations that suggests that central bank independence is determined by a social process of cross-national institutional diffusion. The chapter concludes by examining the broader political implications of the argument. It suggests that while central bank independence may be a highly legitimate organisational form in terms of the contemporary ideology of neoliberalism, it can have troubling implications for democratic legitimacy and accountability. This is particularly true in the case of the European Central Bank, and the tensions that flow from this legitimation problem remain unresolved.

The politics of delegation

42

THE THEORY OF CENTRAL BANK INDEPENDENCE Central bank independence is one of the most prominent examples of the delegation of policy making to non-majoritarian institutions. The trend towards independence is demonstrated in Figure 1, which depicts the

FIGURE 1 CENTRAL BANK INDEPENDENCE OVER TIME

Sources: Pre-1990s data on legal central independence by decade comes from A.Cukierman, Central Bank Strategy, Credibility, and Independence: Theory and Evidence (Cambridge: MIT Press 1992). Data on legal CBI in the 1990s comes from S.Maxfield, Gatekeepers of Growth: The International Political Economy of Central Banking in Developing Countries, Table 4.1. Post-1994 CBI data comes from the European Monetary Institute, Convergence Reports 1998 and 1999, and from press reports of national central bank legislation in non-EU countries. A central bank is independent if it recieves a score of .35 or higher from Cukierman et al. for the period 1950–89. After 1989, central banks are assumed to remain independent once they reach the threshold identified by Maxfield. Post-1994 central banks are coded by author using the Cukierman standard. variation over time in the number of central banks that can be classified as legally independent. Yet, perhaps because monetary policy is such a seemingly technical and arcane research area, much of the broader political science discussion of principal-agent issues occurs separately from the discussion in economics about central bank independence, with some important exceptions.2 This first section thus attempts to situate the logic of central bank independence within the principal-agent literature and lays out, as simply as

Rational fictions

43

possible, the conventional wisdom regarding the logic of delegation in this issue area. After summarising the arguments for central bank independence, the chapter offers some conceptual critiques of this logic before examining the empirical evidence regarding the costs and benefits of central bank independence. The basic premise of principal-agent (P-A) theory is that in certain instances, one actor (the principal) may gain from delegating power to another actor (the agent) if there is an expectation, first, that the agent’s subsequent actions will be aligned with the principal’s preferences and, second, that moreover there is some advantage to moving policy capacity to the agent.3 Although developed in the context of American congressional politics, P-A analysis offers a ready-made framework that directs our attention to similarities across a wide range of activities of delegation. It has the potential to highlight nuances in the interplay between key actors, for example, the role of EU member states as principals and the role of central banks as agents. In the case of EU institutions, in particular, P-A offers a more dynamic and potentially more productive understanding of the mutual dependence of agents and principals than is possible using intergovernmental or functionalist approaches.4 The focus of this chapter, however, is the decision to delegate, examining the rationale for the act of moving control for day to day governance activities out of the hands of those who at first glance should be most desirous of keeping it. As outlined in the Thatcher and Stone Sweet introduction to this volume, the P-A analysis identifies common reasons why principals choose to delegate, one of which, the resolution of commitment problems, is widely given as the reason for central bank independence. Thatcher and Stone Sweet note that commitment problems have a distinct logic in the delegation game. The assumption underlying delegation in these areas, as we shall see, is that it will allow principals to overcome the obstacles lying in the way of more optimal policies. Drawing on P-A analysis, Thatcher and Stone Sweet note that commitment problems also tend to produce delegation with minimal ex post control over agents, as the institution in question needs to appear as delinked as possible from the principal if it is to appear credible. The spread of central bank independence certainly matches this general logic and its predictions about the form of delegation. However, P-A analysis cannot fill in the theoretical content for understanding specific sets of causal dynamics which cause commitment problems and enable their resolution through institutional design. For this, we need to turn to more specific, substantive theories. In the case of central bank independence, such theories are drawn mostly from work in political economy on the interaction between political influences and economic outcomes. These theories nestle inside the broader P-A framework, although not without certain important tensions, as we shall see. In the area of central banking, the justification for delegation is its ability to solve the problem of inflation purportedly caused by political involvement in monetary policy making. Central banks are generally charged with controlling the flow and supply of money, most importantly through the setting of interest rates, but also through such things as bank reserve requirements. There are at least three major macroeconomic indicators important to the health of the economy: GDP growth rates, unemployment levels, and inflation. It is a concern about inflation, however, that has dominated the debate over the merits of delegating policy making to independent banks. Theorists have hypothesised about two different ways for politics to impact negatively on the functioning of central banks and cause inflation: through electoral effects and through

The politics of delegation

44

partisan dynamics. The working of the democratic process, in this portrayal, may have pernicious effects on monetary stability, and far sighted politicians and their publics should seek the removal of central banking from the direct influence of elected officials as the appropriate cure. The two logics are analysed in turn. First, electoral politics are hypothesised to provide an overwhelming incentive for politicians to try to buy votes by stimulating the economy in advance of an election, most notably by lowering interest rates and easing the money supply in advance of election day.5 The result is a ‘political business cycle’, where the economy grows and contracts in tandem with the election schedule. This pattern is viewed as being problematic, because, while it may deliver votes, it can provide only short term benefits to the public while causing inflation in the long run. It may also be destabilising to the economy as a whole, producing cycles of boom and busts, or high growth and recession.6 A second problem with political control for some theorists of central banking lies in the effect of partisan politics, that is, the tendency for political parties to attempt to distinguish themselves in the eyes of the voters by advocating different economic policies. The partisan problem does not lie per se in partisanship itself, but from a specific type of partisanship, that is, the historical political choice of parties on the left, be they the Democratic Party in the United States, the Socialists in France, or the Social Democrats in Germany, to pursue more expansionary policies than the parties of the right. These leftist parties have traditionally targeted their appeal to workers, not investors, and therefore may give more priority to growth and unemployment than to price stability.7 Logically, therefore, these parties on the left may be more likely to pursue expansionary policies that, again, may be desirable in the short run but in the long run may produce unacceptably high levels of persistent inflation. Inoculating central banks from these partisan and electoral effects by placing monetary policy in a technocratic realm, separate from politics, is the policy prescription for the hypothesised shortcomings of democracy. These electoral and partisanal challenges, supported by deductive arguments from the rational expectations approach, mean the only way out of this conundrum is for the central bank to be able to commit credibly to keeping inflation low.8 Delegation to nonrepresentative institutions is seen as the key way to enhance commitment.9 By removing the bank from democratic pressures and establishing that it is free from political influence, the central bank may be able to convince actors in the economy that it has no incentive to manipulate the money supply for political gain. The most positive scenario, in this line of reasoning, is that once credibility is established the independent central bank can undertake surprise reflations of the economy, sparingly and at unpredictable times, ultimately producing more effective monetary expansions without inflationary side effects. Assuming for the moment that policy makers have a long run view of their selfinterest such that the solution of delegation will be attractive, the functional logic of central bank independence sets up the next important question: how should policy makers go about establishing a credible delegation of policy authority to the central bank? The achievement of central bank independence is generally viewed as dependent on at least three factors: a low degree of political involvement in personnel matters within the bank; the financial separation between the central bank and the government; and the policy independence of the central bank from political directives from the government.10

Rational fictions

45

Independence in personnel matters is usually assumed to be highest when there are long, non-renewable terms of office, arms-length appointment procedures, and very high barriers to dismissal of central bank authorities. Financial independence is important to the overall degree of independence as governments that rely on their central banks for credit or management of the government debt are by definition much more intertwined in central bank policies. Finally, the setting of monetary policy itself is subject to degrees of independence, and theorists have often separated out goal independence from instrument independence. If the central bank is free to set the final objectives of monetary policy, be it zero inflation or smoothing output, the government has delegated policy goals. The government may also give the central bank discretion over how it achieves those goals, that is, the instruments used such as inflation targeting or monetary targeting. These two are not always found together, for example, the European Central Bank is directed to pursue the goal of price stability, although it is not further defined numerically, while there is no effort to specify the instruments the ECB should use to further that goal. Obviously, independence is a continuum, and legal independence is not a perfect guide to independence in practice. Some banks, such as the pre-EMU Dutch central bank, may be behaviourally independent and treated as such by society and government, although their statutes do not ascribe them as much legal independence as other banks. Despite disagreement over the exact practical contours of independence, the arguments regarding the delegation of monetary policy are compelling. Monetary policy is an uncertain realm that can have great impact on the macroeconomy, for ill or good. Delegation seems to provide a one size fits all solution to the politicisation of monetary affairs. However, the theoretical rationales rest on shakier foundations that might be apparent at first glance. Below, two critiques of the essential premises of central bank independence are put forward. The first critique confronts the question of whether severing ties to democratic representatives and shifting to technocratic expertise within an insulated institution does truly apoliticise monetary policy making; the second point asks whether focusing solely on the government-central bank relationship captures the most significant influences on monetary policy outcomes. Both questions are answered in the negative, challenging the functional logic of delegation. THE POLITICS OF CENTRAL BANK INDEPENDENCE My first critical argument concerns the basic premise underlying the logic of central bank independence: delegation is warranted because of the need for economic expertise to provide more optimal, politically neutral policy solutions, policies that are not readily accomplished in the context of political intervention.11 Delegation to central banks is attractive in part because it seems to place priority on improving aggregate welfare, on making the economy work better for the majority of people by taking monetary policy away from the vicissitudes of electoral and partisanal politics. This logic is illusory, however. While severing the direct institutional ties to elected officials appears to create an apolitical environment for policy making, central banks continue to make policies which have important, identifiable distributional effects and thus remain resolutely political and therefore partisanal institutions.

The politics of delegation

46

As Joe Stiglitz has written, ‘the decisions made by central bankers are not just technical decisions: they involve trade-offs, judgments about whether the risks of inflation are worth the benefits of lower unemployment. These trade-offs involve values’.12 The values at the core of delegation to independent central banks are neoliberal in nature, as the purported effects of the electoral and partisanal influences on central banking all ultimately centre on the risk of inflation and its potential for detrimental long term effects on the economy.13 Thus, although principal-agent analysis focuses on delegation as a procedural solution, delegation in the area of central banking is a substantive choice as well. The privileging of price stability over growth and employment has important consequences. Those with money to save and invest may benefit from a low and stable rate of inflation, although if it falls so low as to choke off growth entirely, their investments will suffer as well. Those who rely on wages will tend to be helped more by a growing economy which maintains high levels of employment. Very high levels of inflation will dampen the investment and economic activity that workers rely on as well as eroding the capital of the wealthier groups. There is also an intergenerational component to the distributional effects of inflation. Older, retired workers will be more affected by inflation as they rely on investments and savings, whereas their children and grandchildren may be more affected by slowdowns in the economy. These distributional consequences of central bank independence have been subject to relatively little analysis in the literature or in the popular discourse, as the logic of central bank independence projects a procedural and political neutrality to the process of delegation that mutes questions about the values being traded off in pursuit of price stability.14 Given these distributional impacts, delegation raises important democratic accountability questions. In the general principal-agent literature, one of the potential benefits of delegation is to move policies closer to the desires of the median voter, desires that for some reason are difficult to achieve without such delegation. Thus, it can be argued that delegation is actually more democratic than allowing politicians to hold sway over policy for their own ends, subverting the common good. However, the argument in favour of central bank independence has evolved differently, positing in effect that the desires of the median voter should not guide policy. In fact, one influential article argues that the ideal central bank appointee will be more conservative and anti-inflationary than the median voter, as the latter might favour more growth without factoring the potentially negative longer run consequences of expansionary policies.15 Indeed, the structure of central bank independence does not make it likely that the median voter’s preferences are captured. The majority of individuals appointed to central bank boards, even when the ruling party is on the left, are from the private banking or investment communities, with a very small representation from industry and virtually no representation from other sectoral groups such as labour. This makes it less likely that a diverse spectrum of views will be represented. The limited role for political representatives in government positions in the formulation and execution of monetary policy may make it difficult to rein in central bankers that deviate too far from social norms regarding the management of the economy.16 However, the argument for central bank independence made by many economists and policy makers pays little attention to this danger of too much autonomy on the part of the agent, because of the underlying assumption that the principal (the national government) does not necessarily know its own interests. This is in contrast to much of the principal-

Rational fictions

47

agent literature, which has tended to stress the potential for difficulties on the agent side, arising from too much delegation or control shifting away from the principals.17 Two types of procedural problems are identified as potentially arising from delegation: agency ‘shirking’ and agency ‘slippage’. Shirking and slippage imply some sort of noncompliance by the agent such that the agent is no longer following the goals of the principal.18 The purposefully low level of direct political oversight in the area of central banking means that it is likely that independent central banks will have an intentionally high degree of agency slack. This is a positive development if you are confident in the merits of governance by highly conservative central bankers who exceed the preferences of elected officials for low inflation. However, such delegation may produce monetary authorities who pursue the goal of low inflation with too much zeal and thus have the potential to stave off needed growth and employment in the economy, without much leeway for the principals (that is, the governments) to correct this policy drift. Finally, a further and related conceptual critique of the principal-agent framework of central bank design is that it may overly emphasise the importance of a narrow set of dynamics between the government and the central bank at the expense of the broader societal dynamics that play a key role in economic policy outcomes. A strong argument can be made that central bank independence is a behavioural, not legal or organisational, phenomenon, and that it is more a function of societal relations, shared expectations and other such variables, rather than rules and institutional designs.19 Delegation does not occur in a political vacuum, and to what degree the effectiveness of the institutional form is actually endogenous to prior political relationships is a key question not adequately addressed in much of the debate. Persuasive logical arguments can be made in support of explicit and transparent linkages to electoral institutions, be it as official oversight or informal interactions, in order for the central bank to have the political support and policy co-ordination needed to achieve positive policy outcomes. In the case of Germany, for example, a complex set of societal understandings and relationships underpins the Bundesbank’s success in achieving growth and stability.20 This broader context of interlocutors in the policy process is outside the logic of delegation to independent central banks. DELEGATION AND OUTCOMES IN PRACTICE Conceptual critiques are all very well, but what about the empirical case for central bank independence? Can it be demonstrated to have been successful in ameliorating inflation or improving economic conditions across the various national settings? The rational choice institutionalist logic of delegation is based on the idea that policy makers choose these institutional designs in the expectation that they will address a compelling problem and produce better outcomes—a more optimal level of inflation in conjunction with employment and growth. Political influence, in the logic of central bank independence, can be dysfunctional for the economy, and the positive outcomes achieved from delegation therefore can be argued to outweigh concerns about the loss of democratic accountability. To evaluate the necessity and efficacy of delegation in the management of money, three questions of the empirical research on central banking are asked. Can it be

The politics of delegation

48

demonstrated (1) that there has indeed been a problem with democracy that central bank independence must solve, that is, a pattern of political business cycle behaviour or partisanal bias producing inflationary outcomes; (2) that inflation itself can be empirically demonstrated to be highly detrimental such that it presents a compelling rationale for central bank independence; or (3) that central bank independence does indeed produce more positive economic outcomes, outcomes that better match the long term interests of policy makers and citizens than those achieved with politically dependent central banks? Empirical evidence on these points would certainly provide support for the functional argument regarding central bank independence. On the first issue of whether electoral or partisanal influences on monetary policy are a critical factor producing high levels of inflation, the empirical evidence is mixed at best. A recent survey of the political business cycle literature and the effects of electoral politics on central banks assessed 25 years-worth of studies and found that the literature’s evidence is thin, and the ‘principal conclusion is that models based on manipulating the economy via monetary policy [for political gain] are unconvincing both theoretically and empirically’. The author goes on to argue that ‘explanations based on fiscal policy conform much better to the data and form a stronger basis for convincing theoretical model of electoral effects on economic outcomes’ than do monetary policy manipulations.21 Governments may try to influence the outcomes of elections by using the economic levers at hand, but those tools tend to be traditional pork barrel politics, that is, spending projects that may stimulate the economy, rather than manipulations of the money supply. Neither have analysts found convincing evidence of systematic partisanal differences in monetary policy since the 1970s. While work done on cross-national experiences in the early post-war era demonstrated evidence for partisanal effects, these effects have declined over the past few decades, in advance of the widespread delegation of central banks.22 The last few decades of experience have produced strikingly convergent monetary policies in the EU, for example, oriented towards low inflation policies regardless of what party is in power, whether it was the leftist Socialists in Spain or the Christian Democratic right in Germany. This convergence preceded the move to central bank independence, suggesting that other factors beyond the organisational delegation of monetary policy are at work in producing price stability, despite democratic influences. The second key question concerns the empirical evidence on the negative effects of inflation itself. The literature on delegation to central banks takes as given the fact that inflation is extremely detrimental and must be avoided at all costs. Nonetheless, Stiglitz is one prominent economist who has questioned this assumption, and he offers persuasive empirical evidence for his position.23 He points out that an older literature evaluates the costs of inflation in terms of such things as ‘menu costs’, ‘shoe leather costs’, tax distortions and increasing noise in the price system; however, the empirical estimates of the deadweight losses from these factors are relatively small. A newer literature has looked at how inflation might affect the level of output and growth. Studies by Bruno and Easterly have found that inflation rates need to be quite high, in excess of 40 per cent per year, to be very costly, and that below that level there is little evidence of a high inflation/low growth trap.24 In addition, studies have done cross-country comparisons of the effects of inflation on growth.25 Echoing the earlier findings, they find that while very

Rational fictions

49

high levels of inflation are detrimental to growth, lower levels do not seem to have the same impact. Stiglitz goes on to present his own evidence regarding two other common assumptions about why inflation should be kept extremely low: that inflation will accelerate uncontrollably if left unchecked, and that it is costly to reverse.26 In fact, Stiglitz finds no statistical evidence that inflation builds on itself but rather that when it has been rising it is likely to reverse its course. He also finds that adjustment occurs in the economy in more effective ways than is assumed by inflation hawks. He notes that we still have an imperfect understanding of the way the economy works, and that this should make us very cautious about prioritising inflation fighting above all things. Inflation is not clearly detrimental to the economy, and pre-emptive contractionary policies that attempt to dampen inflation while subsequently slowing growth and employment are policies ‘based on articles of faith, not on scientific evidence’.27 The third and final empirical question to be addressed is the question of whether central bank independence does in fact produce better economic outcomes overall. In contrast to other areas of policy delegation, which rarely offer systematic evidence about the effects of delegation, there is a large literature in economics which attempts to determine the relationship between central bank independence and macroeconomic outcomes.28 Despite careful efforts on the part of scholars, however, the evidence for superior economic outcomes from central bank independence remains inconclusive. The essential problem is that the strength of the correlation found between central bank autonomy and outcomes such as lower inflation is extremely sensitive to the (highly contested) criteria used to measure independence, the time period chosen, and especially the countries included in the sample.29 This latter point is important because if one looks at developing countries as opposed to advanced industrial ones, the positive impact of central bank independence on inflation simply disappears. This should hardly be surprising, because politics in developing countries is often guided by informal rules rather than laws and formal procedures and because such countries lack the range and depth of institutions necessary for full policy implementation and co-ordination. Even in advanced industrial countries, where the correlation between central bank independence and lower inflation seems strongest, the case is not clear-cut. Adam Posen claims that ‘differences in central bank autonomy and reputation are not the sources of inflation differences among the advanced industrial countries…all the published arguments used to trace a causal link between [central bank independence and low inflation] are not borne out by economic reality’.30 What really matters in the long term struggle against inflation, Posen finds, is a strong financial sector interested in price stability and willing and able to influence policy making to get it. Central bank independence and lower inflation, in this view, are linked not because one causes the other but rather because both are caused by the political effectiveness of a particular interest group coalition.31 Other scholars have come up with similar findings, noting that independent central banks will have little positive long-term impact on inflation unless there is a societal consensus on the need for price stability. Students of Germany, for example, have argued that the real source of the Bundesbank’s effectiveness is not its statutory independence but rather a widespread public acceptance of the need to make inflation fighting a primary goal of economic policy and the Bundesbank’s ability to act in concert with the

The politics of delegation

50

‘social partners’, that is, labour and business, in keeping down wage and price increases.32 In sum, the empirical work on delegation to independent central banks casts doubt on the severity and nature of the problem purportedly solved by delegation, that is, the pernicious effects of democracy on policy making, particularly with regard to inflation, as well as raising questions about the linkages between delegation and superior economic outcomes. This raises a basic question. If the premise of delegation is that policy makers choose these institutional forms, such as central bank independence, because they anticipate superior economic outcomes in their country with delegation, why might they continue to do so even if such evidence is not forthcoming? The section below suggests an alternative explanation for the diffusion of central bank independence. WHY DELEGATE? THE SOCIAL BASIS OF DIFFUSION OF ORGANISATIONAL FORMS If central bank independence rests on shakier functionalist foundations than is usually assumed, why has there been such a widespread move towards this organisational form? In this section, an alternative sociological perspective for the logic of delegation is outlined.33 In this approach, the choice of organisational form is linked to social processes that legitimate certain types of institutional choices as superior to others. Governments choose to delegate monetary policy to independent central banks because it is instrumentally rational given a particular cultural environment, one that rewards this organisational form. In this sociological perspective, it is the symbolic properties of central bank independence that carry substantial weight in explaining policy diffusion, rather than the expressed functional properties of delegation. A comprehensive set of sociological institutionalist arguments about the sources and working of organisational design have been put forth by Scott and Meyer and can be applied to the diffusion of central bank independence.34 This approach departs from the conventional wisdom of delegation in at least three ways: in the definition of what organisations are and where their rules come from; in what mechanisms drive the adoption of specific organisational forms; and in the nature of rationality in organisational design. First, the rule and practices of organisations arise, in this sociological account, from cultural and social processes, rather than being derived from local functional needs. Organisations, meaning visible structures and routines, are direct reflections and effects of wider cultural and symbolic patterns or templates. The definition of organisation is thus very ‘thick’, as organisations both embody and shape broader social and cultural dynamics. Therefore, central banks (as organisations) must be analysed in terms of the broader social environment within which they are situated—the ideas, norms, and culture of the moment—as it is this environment that profoundly shapes their form and practices. By situating central bank independence within a national or transnational culture of neoliberal economic policy making, which privileges price stability as an absolute good, it is easy to see why central bank independence, with its substantive bias towards low inflation outcomes, is a rational strategy. Further, in an increasingly globalised international financial market, central bank independence is one way of signalling to

Rational fictions

51

investors a government is truly ‘modern’, ready to carry out extensive reforms to provide a setting conducive to business.35 Note, however, that this signal is understood as such even if the statistical relationship between this organisational form and superior economic outcomes may not be borne out in practice. Delegation to central banks is thus a very rational adaptation to a specific cultural environment which rewards certain organisational forms over others, in part because of the real distributional effects that such delegation may provoke. In this process of organisational diffusion, Scott and Meyer point out that specific local functional needs may not be the central source of an organisational structure, but rather designs are borrowed from other environments and then applied locally despite important differences across settings.36 For example, it may not be the specific circumstances of the national economy that produces delegation to independent central banks, but rather the template of the central bank may be suggested by other national experiences that are perceived as successful. Thus, a country in the depths of a recession may increase central bank independence even though slow growth, not inflation, is the key policy challenge—policy makers use central bank independence to signal to investors that they are credibly following a reformist path. The transfer of the template occurs therefore even though there may be important discrepancies in the needs and contours of the national political economies that make the replication of that success unsure or unlikely.37 For example, a country such as Ecuador, without the legal and political institutions to truly emulate independent central banks such as the ECB, may rationally pursue this organisational design for its symbolic properties although central bank independence ends up being meaningless in practice. Second, in the sociological institutional perspective, the causal mechanisms driving the adoption of an organisational form, such as central bank independence, work through constitutive or phenomenological aspects and are socially constructed in conjunction with material circumstances. For example, markets themselves do not speak: the ‘signals’ that they send about standards of appropriateness occur through the evaluations and judgements of specific social actors operating within a dense cultural environment that shapes their interpretation of market signals. Credibility, the keystone of the central bank independence framework, is a perceptual variable, and thus constituted by social processes. The sociological perspective contextualises this process so as to illuminate the interplay between particular functional or material pressures, the subsequent interpretation of these pressures by actors, and the consequent creation of shared perceptions about those pressures and how to solve them. The view that central bank independence is a necessary component of good governance did not magically fall from the heavens, but rather was created. This process of social construction, for these theorists, permeates the creation and diffusion of organisational forms. Applied to central bank independence, the development of rules regarding the appropriate mode of central bank design can be seen to be an inherently social process occurring between human agents in interaction with each other and their material environments. Political elites face a host of challenges in governing their economies, and the social environment that they move within may provide central bank independence as a solution, even if not actually tailored to their specific circumstances. The sources of social interactions and linkages between actors in the central banking field are numerous, found in the international and regional monetary policy fora which link national central

The politics of delegation

52

bankers and economic policy leaders, such as the Group of 7, the International Monetary Fund, the Organisation for Economic Cooperation and Development, and the Bank for International Settlements. The development of epistemic communities and expert consensus on the part of these officials is well documented, and is likely to have contributed to the diffusion of central bank independence.38 The shared education of these officials within elite economics departments, centred in American universities like MIT or the University of Chicago, is also a foundational aspect of the process of socialisation and acculturation that produces conformity in outlook and beliefs about the workings of the economy.39 The creation of meaning can also occur in more diffuse ways, such as in the development of a conventional wisdom in tandem with reporting and editorialising by financial media such as the Financial Times, which has asserted that ‘The argument[s] for central bank independence appear overwhelming’, or the Economist, which has declared that ‘The intellectual case for independent central banks [has been] more or less won’.40 The third contention of this sociological perspective is that the processes and patterns that create and change organisations are both rationalised and rationalising. Modern sociocultural environments tend towards a process of rationalisation, that is, ‘the creation of cultural schemes defining means-ends relationships and standardising systems of control over activities and actors’.41 The spread of a standardised form of political organisation, such as the delegated governance model, is a direct result of these social processes. Actors must first perceive similarities across different national settings for the diffusion of a particular organisational form to make sense; abstraction into common categories, such as ‘the market’ or ‘the state’ is necessary for the use of templates to be rational. Cognitive or conceptual theorising about the appropriate functions of central banks, for example, and the development of a body of research that provides a recipe for success, regardless of setting, is in part the product of a modern, rationalised world culture. Part of that rationalised culture is rooted in the highly developed theorising of the academic discipline of economics. We take for granted the appropriateness and necessity of the sort of comparison and rationalisation implicit in the discussion of central bank independence; however, a case can be made that it is equally sensible to assume that differences across national institutional, political, and economic settings will make the transfer of templates without alteration ineffective, or possibly detrimental, to those local economies.42 Given these three sets of processes, sociologists argue that institutional isomorphism, or the similarity of organisational form across settings of social interaction, will be the expected outcome, as actors borrow those models collectively sanctioned as successful even though they may be decoupled from or incongruent with functional needs. Such organisational isomorphism has received a great deal of study among sociological institutionalists and economic sociologists.43 For these theorists, the replication and diffusion of organisational forms is provoked in part by the need to find legitimacy in terms of the prevailing norms, rather than adaptation to straightforward functional problems. The contrast with the prevailing wisdom of delegation to independent central banks as an instrumentally rational decision, the perspective outlined in the first section of this chapter, is striking.

Rational fictions

53

PROBING THE SOURCES OF CENTRAL BANK INDEPENDENCE This chapter has so far articulated two quite different sets of arguments about why delegation in the monetary realm might be a sensible choice. Although the arguments make very different claims about the process by which delegation occurs, they both predict similar outcomes of institutional isomorphism, that is, the rise in central bank independence over time demonstrated earlier in Figure 1. How can we therefore adjudicate empirically between these explanations? Below, a series of hypotheses is offered for what we should observe given the two logics, material and social, that drive the respective theories. Although a complete test of these two approaches is not possible here, the purpose is to provide some suggestive illustrations of the limits and potential of each explanation. First, the purely functional approach implies that countries increase the independence of the central banks because they believe delegation will solve their economic problems, particularly high inflation. Although the present discussion challenges the logic of this solution and the evidence that independence produces superior outcomes, it may be that governments are responding to a set of material circumstances for which central bank independence is viewed as offering a rational solution, even if the actual correctness of this analysis is flawed. Thus, empirically we should examine whether or not the material economic circumstances in advance of the decision to delegate do indeed match the conditions that independence is meant to improve on. Most prominently, inflation rates should be persistently high in advance of the decision to delegate. Unfortunately, the empirical work on central banks has focused overwhelmingly on the relationship between independence and economic outcomes instead of analysing the sources of the policy of delegation, with the exception of Maxfield’s seminal research.44 This is an important oversight. Investigating the timing of moves to central bank independence in light of national macroeconomic conditions might provide clues as to the fit between the functionalist story or the social institutions one. Maxfield’s study of the sources of central bank independence in emerging economies, which focuses on the desire of developing states to signal their credibility to international investors, notes that central bank independence and inflation do not seem to have a close temporal relationship, arguing that ‘inflation was very severe in developing countries in the 1970s and 1980s, yet central bank independence decreased in the 1970s and changed little in the 1980s’.45 A thorough econometric analysis is needed to assess fully the relationship between macroeconomic indicators and decision to increase central bank independence. A rough and ready preliminary look at the data on inflation, growth, and employment appears to demonstrate a mixed relationship to delegation (see Appendix 1). What is clear is that there may be strong regional trends in the data. The newly independent states of the former Soviet bloc have all moved to independent central banks in the context of sometimes extremely high inflation, but the macroeconomic indicators for these states are sketchy at best. Policy makers’ assessment of the economic necessity of central bank independence was short-circuited, as the newly created banks were given immediate legal independence after the break-up of the Soviet Union. A second regional trend is apparent

The politics of delegation

54

in the European Union member states, which moved to make their banks independent in conjunction with the Maastricht Treaty on European Union and its requirement for Economic and Monetary Union. The majority of EU states had been experiencing very low inflation throughout the 1980s and 1990s accompanied by slow growth and high and persistent unemployment. These economic conditions do not match those for which delegation of authority for monetary policy would appear ideal. If local material conditions do not appear to warrant the move to central bank independence, this would make more plausible the contention of the sociological approach. The adoption of a standardised form, institutional isomorphism, may be occurring through a process of social diffusion of organisational models. The conditions affecting diffusion in this account are quite different from the conventional wisdom on central bank independence.46 Such mimetic behaviour is a way to legitimise organisations, particularly under conditions of uncertainty, where means-ends relationships are unclear or there is no agreement on performance criteria.47 Imitation is a rational strategy under the above conditions. Indeed, central banking is an area of significant uncertainty, as questions are pervasive over the measurement of the money supply, the formulation of economic projections, and many of the causal linkages in the macroeconomy remain poorly understood. Given the degree of uncertainty in the world, particularly when it comes to the workings of the economy, how do social mechanisms of diffusion provide a rational source of organisational forms? Two broad categories of linked mechanisms for such diffusion are possible: coercive isomorphism, in which political power and legitimacy are the driving forces; and normative isomorphism, which arises from the processes of professionalisation and socialisation within networks.48 Coercive isomorphism ‘results from both formal and informal pressures exerted on organisations by other organisations upon which they are dependent and by cultural expectations in the society within which organisation functions’. These pressures can be ‘felt as force, as persuasion, or as invitations to join in collusion’.49 For example, if carrots and sticks (incentives and sanctions) are used on the part of international institutions, such as the IMF, this could be felt as coercion on the part of domestic actors, who may adopt central bank independence to meet the expectations of the IMF rather than because their own national conditions seem to warrant it. The broader package of domestic institutional reforms linked to the so-called ‘Washington Consensus’ may be reflective of this process.50 The European Union states can also be evaluated in terms of this sort of dynamic, as legal independence on the part of the national central banks is one of the criteria for entry into Economic and Monetary Union specified in the Maastricht Treaty on European Union. While the IMF and the EU are intergovernmental bodies, and thus one can argue that member nations in part determine these conditions, not all member governments have equal say, either formally on the part of the IMF which awards votes in terms of economic weight, or informally, in an EU where bargaining is never purely symmetrical. The second mechanism, normative isomorphism, can reinforce the first, more coercive, one. Persuasion may occur through the development of conceptual models, such as those promoting central bank independence outlined in the first part of this chapter, which gain authority and legitimacy by their advocacy on the part of prominent analysts. The models that proscribe delegation to independent central banks derive from the discipline of economics, which in the west is unified behind a single methodology and

Rational fictions

55

intellectual cannon. These models are diffused through professionalised networks of economists and economic policy makers and become institutionalised in the IMF, or spread through shared education and training, or through authoritative media sources and communities of financiers and international investors. The combination of coercive and normative pressure for conformity to organisational form in the area of central banking are formidable. Hewing to the precepts of central bank independence is thus a natural outcome for transition states from the former Eastern Bloc, as the need to appear credible and mimic the institutions of nations successful in stabilising their economies is extremely pressing, driving institutional isomorphism regardless of their particular local economic needs.51 Establishing legal independence for their new central banks at the start of their transitions was a way to legitimise the new regime and send a signal to allies, investors, and international institutions that the government would play by the rules of the global political economy. Developed states are not immune from these processes either. For example, the Bank of Japan was made independent in 1996, after a multiple year recession that far from being marked by inflation, has seen deflationary trends, with price increases hovering around one per cent or below. The move to delegate monetary policy was taken as a part of the ‘Big Bang’ package of economic reform, which proposed a variety of deregulatory actions in an effort to jump-start the Japanese economy and move its political economy a little closer to that of the Anglo-Saxon model. Delegation was a rational choice in the context of the culture of neoliberalism, but not to address the purported inflationary tendencies of democracy as per the economic literature on central bank independence. The European shift to central bank independence also challenges the functional argument and points to the role of institutional diffusion of organisational form. In the early 1990s, almost all of the EU countries with dependent central banks moved to independence in anticipation of joining EMU, for which such legal independence was a legal requirement. In 1992, Portugal and Italy delegated monetary policy to their central banks, followed in 1993 by Belgium, France, and Greece, and, in 1994, Spain. These states had largely achieved historically low rates of inflation by the 1990s, ranging from two to five per cent. The organisational form of their national banks was dictated by the rules of entry into EMU, and the ECB, which makes monetary policy for Euroland, is structured in the Maastricht Treaty as the world’s most independent central bank. A coercive dynamic of German negotiating strategy, which made the ECB’s independence a necessity, plus the surrounding legitimation of normative support for delegation across economic elites caused the independence criteria—not objective material needs on the part of the majority of the member states, whose monetary policies had been oriented towards price stability for some time.52 Today, the ECB reigns as the most politically independent central bank in the world. The Maastricht Treaty states that the ECB cannot seek or take instructions from any EU or national entity or any other body. Its independence is arguably more secure than that of any national bank because a modification of its statute would require an amendment to the Treaty, which can only occur with unanimous agreement among the member states, rather than legislative majority within a regular political system. Whereas most national central banks have a routinised system of consultation with other government and societal bodies, the ECB is only minimally linked to the political bodies of the EU. Policy makers from the Council of Ministers and the Commission are granted observation status in ECB meetings, and the

The politics of delegation

56

ECB must fulfil a number of reporting commitments to political bodies and the ECB’s president is required to appear before committees of the European Parliament. In sum, some of the most visible recent moves to central bank independence do not seem to fit the material, functional model but rather suggest the merits of testing an alternative approach focused on the symbolic value of central bank independence and the legitimacy of non-representative institutions as an organisational form. These cases suggest that instead of assuming the validity of the conventional wisdom about delegation in the monetary realm, we should undertake research agenda which can systematically test alternative propositions and clarify the social and political logics of central bank independence. CONCLUSION Central bank independence, one of the most prominent examples of delegation to nonmajoritarian institutions of the past few decades, rests on contestable theoretical arguments and inconclusive empirical evidence about the relationship between democracy, policy making, and economic outcomes. This chapter has challenged the logic of delegation in the monetary policy realm, and linked delegation to substantive policy outcomes which have important distributional effects not adequately addressed by proponents of central bank independence. To understand why governments have chosen this organisational form, an alternative theoretical framework drawn from institutional sociology is proposed. In this alternative view, delegation is a culturally rational strategy achieved through coercive and normative processes of institutional isomorphism. The symbolic properties of central bank independence, in signalling agreement with a broader series of economic management principals and in conveying credibility to external audiences about the economic and political character of a government, are critical to explaining the spread of this organisational form. The tensions in the conceptual assertions of the theory of central bank independence and scanty evidence on the beneficial economic effects of delegation in this policy area raise important questions about the legitimacy of this form of governance and the role of democratic accountability in delegated institutions.53 Central bank independence is viewed by political, business, and most academic elites as a highly desirable and legitimate policy, which accounts for its spread as an organisational form. However, delegation is not unproblematic. Central bank independence is designed to skew policies away from the preference of the majority of voters, and therefore does not necessarily have a broader legitimacy in terms of national publics and societal groups. The ECB is a particularly salient example of this conundrum of legitimacy, for it embodies the narrow approbation among elites that is necessary for social diffusion of such forms of delegation, but also demonstrates the disquieting aspects of delegation from democratic processes in the absence of a broad mandate for governance.54 The extreme independence of the ECB is meant to reassure financial and business elites that price stability will trump other economic goals and that Europe’s economic policy will be appropriately insulated from the demands of labour and other domestic interest groups. The decisions to free Europe’s central bank from political control and focus narrowly on fighting inflation, in other words, were not ‘technical’ or ‘apolitical’ as most advocates of

Rational fictions

57

independence argue. Instead, as with all difficult policy choices, they involved trade-offs among competing goods and will benefit some constituencies more than others. This is not surprising or wrong, indeed it is the meat and potatoes of politics, but it is often obscured in both the academic and policy making discussion of central bank independence. The attempt here has been to question some of the assumptions about the functionality of central bank independence, and suggest instead that it might be better understood as a social process that is highly political in its both its sources and effects. NOTES For very helpful comments on this chapter, I thank Nicholas Jabko, Keith Whittington, Mark Thatcher, Alec Stone Sweet, participants in the workshop at the European University Institute, and an anonymous reviewer. I also thank Sheri Berman for invaluable discussions on the topic of democracy and central banking. Elizabeth Bloodgood provided excellent research assistance. 1. S.Maxfield, Gatekeepers of Growth: The International Political Economy of Central Banking in Developing Countries (Princeton: Princeton University Press 1997), p. 3. 2. K.Dyson, K.Featherstone and G.Michalopoulos, ‘Strapped to the Mast: EC Central Bankers between Global Financial Markets and Regional Integration’, Journal of European Public Policy 2/3 (Sept. 1995), pp. 465–87; Robert Elgie and Erik Jones, ‘Agents, Principals and the Study of Institutions: Constructing a Principal-Centred Account of Delegation’, Working Documents in the Study of European Governance 5 (Nottingham: The Centre for the Study of European Governance, University of Nottingham Dec. 2000). 3. Reviews of the literature include T.Moe, ‘The New Economics of Organization’, American Journal of Political Science 28 (1984), pp. 739–77; K.Shepsle and B.Weingast, ‘Positive Theories of Congressional Institutions’, Legislative Studies Quarterly 19/149 (1994), pp. 145–79. 4. M.Pollack, ‘Delegation, Agency, and Agenda Setting in the European Community’, International Organization 51/1 (Winter 1997), pp. 99–134. 5. W.Nordhaus, ‘The Political Business Cycle’, Review of Economic Studies 42 (1975), pp. 169–90; A.Alesina, ‘Macroeconomics and Politics’, in S.Fischer (ed.), NBER Macroeconomics Annual 1988 (Cambridge: MIT Press 1988), pp. 13–69; A.Alesina, ‘Politics and Business Cycles in Industrial Democracies’, Economic Policy 4 (April 1989), pp. 57–98. 6. Alesina, ‘Macroeconomics and Politics’. 7. D.A.Hibbs, ‘Political Parties and Macroeconomic Policy’, The American Political Science Review 71 (Dec. 1977), pp. 1467–87; D.Cameron, ‘The Expansion of the Public Economy: A Comparative Analysis’, The American Political Science Review 72 (Dec. 1978), pp. 1243– 61. 8. The rational expectations approach argues that actors in the economy will catch on to the manipulation of the money supply and will start to figure in the inflationary effects of increased money supply into their wage demands, prices, and investment decisions. As they do so, these inflationary expectations will themselves create inflation. See R.J.Barro and D. Gordon, ‘Rules, Discretion, and Reputation in a Model of Monetary Policy’, Journal of Monetary Economics 12 (July 1983), pp. 10–121; S.Lohmann, ‘Optimal Commitment in Monetary Policy: Credibility versus Flexibility’, American Economic Review 82 (March 1992), pp. 273–86; A.Cukierman, Central Bank Strategy, Credibility and Independence: Theory and Evidence (Cambridge: MIT Press 1992).

The politics of delegation

58

9. A.Alesina and G.Tabellini, ‘Credibility and Politics’, European Economic Review 32 (1988), pp. 542–50. 10. J.De Haan, ‘The European Central Bank: Independence, Accountability and Strategy: A Review’, Public Choice 93 (1997), pp. 395–426, especially p. 398. 11. The validity of the assumptions of rational expectations that underpins the central bank independence logic constitutes a separate, additional critique. See C.Goodhart, ‘Game Theory for Central Bankers’, Journal of Economic Literature 32 (March 1994), pp. 101–14. An analysis of the ideological dimensions of the rational expectations approach is Ilene Grabel, ‘Ideology and Power in Monetary Reform: Explaining the Rise of Independent Central Banks and Currency Boards in Emerging Economies’, presented at a conference on ‘Power, Ideology, and Conflict: The Political Foundations of 21st Century Money’, 31 March–2 April 2000, Cornell University, Ithaca, NY. Grabel also stresses the distributional consequences of delegation. 12. J.E.Stiglitz, ‘Central Banking in a Democratic Society’, Economist 142/2 (1998), pp. 199– 226, quote p. 216. 13. This can be characterised as a conflict between a traditional Keynesian view of macroeconomic policy, which prescribes demand management through activist monetary policy, and a more neoliberal monetarist view, which prescribes targeting inflation control as the singular goal of a central bank. See P.Hall, ‘The Movement from Keynesianism to Monetarism’, in S.Steinmo, K.Thelen and F.Longstreth (eds.), Structuring Politics (New York: Cambridge University Press 1992); P.Hall, ‘Policy Paradigms, Social Learning, and the State’, Comparative Politics 25/3 (April 1993), pp. 275–96; K.Dyson, Elusive Union: The Process of Economic and Monetary Union in Europe (London: Longman 1994); and K. McNamara, The Currency of Ideas: Monetary Politics in the European Union (Ithaca: Cornell University Press 1998). 14. An exception is the societal analysis provided in J.B. Goodman, ‘The Politics of Central Bank Independence’, Comparative Politics 23/3 (1991), pp. 329–49. 15. K.Rogoff, ‘The Optimal Degree of Commitment to an Intermediate Monetary Target’, Quarterly Journal of Economics 100/4 (1985), pp. 1169–89. 16. See Goodman, ‘Central Bank Independence’, especially pp. 334–6, on historical variations in independence in European central banks. 17. E.g. R.D.Kiewert and M.D.McCubbins, The Logic of Delegation: Congressional Parties and the Appropriations Process (Chicago: University of Chicago Press 1991), p. 5. See Elgie and Jones, ‘Agents, Principals and the Study of Institutions’. 18. M.McCubbins and T.Page, ‘A Theory of Congressional Delegation’, in M.McCubbins and Terry Sullivan (eds.), Congress: Structure and Policy (New York: Cambridge University Press 1987). 19. A.Posen, ‘Why Central Bank Independence Does Not Cause Low Inflation: There Is No Institutional Fix for Politics’, in R.O’Brien (ed.), Finance and the International Economy 7 (Oxford: Oxford University Press 1993). 20. J.Goodman, Monetary Sovereignty: The Politics of Central Banking in Western Europe (Ithaca: Cornell University Press 1992); K.R.McNamara and E.Jones, ‘The Clash of Institutions: Germany in European Monetary Affairs’, German Politics and Society 14 (Fall 1996), pp. 5–31. 21. A.Drazen, ‘The Political Business Cycle after 25 Years’, in B.Bernake and K.Rogoff (eds.), NBER Macroeconomics Annual, 2000 (Cambridge, MA: NBER 2000), pp. 3–4. Drazen does not address, however, a more nuanced and promising analysis, which incorporates the role of domestic institutions and international capital mobility: W.R.Clark and M.Hallerberg, ‘Mobile Capital, Domestic Institutions, and Electorally Induced Monetary and Fiscal Policy’, American Political Science Review 94 (June 2000), pp. 323–46. 22. See K.McNamara, The Currency of Ideas, ch. 6; G.Garrett, Partisan Politics in a Global Economy (Cambridge: Cambridge University Press 1998).

Rational fictions

59

23. Stiglitz, ‘Central Banking’. An earlier assessment of inflation along similar lines is B.Barry, ‘Does Democracy Cause Inflation?’ in L.N.Lindberg and C.Maier (eds.), The Political Economy of Inflation and Economic Stagnation (Washington, DC: Brookings Institution 1985), pp. 280–317. 24. M.Bruno and W.Easterly, ‘Inflation and Growth: In Search of a Stable Relationship’, Federal Reserve Bank of St. Louis Review 78 (May–June 1996), pp. 139–46. 25. R.J.Barro, Determinants of Economic Growth (Cambridge: MIT Press 1997); S.Fischer, ‘The Role of Macroeconomic Factors in Growth’, Journal of Monetary Economics 32/3 (1993), pp. 485–512. 26. Stiglitz, ‘Central Banking’, pp. 212–13. 27. Ibid., p. 215. 28. E.g. Cukierman, Central Bank Strategy, Credibility and Independence; T.Persson and G. Tabellini (eds.), Monetary and Fiscal Policy (Cambridge: MIT Press 1994); S.Eijffinger and J.De Haan, ‘The Political Economy of Central Bank Independence’, Special Papers in International Economics 19 (Princeton: International Finance Section, Princeton University May 1996). 29. J.Forder, ‘Central Bank Independence: Reassessing the Measurements’, Journal of Economic Issues 33 (March 1999), pp. 23–40. 30. Posen, ‘Why Central Bank Independence Does Not Cause Low Inflation’. 31. Goodman, ‘The Politics of Central Bank Independence’. 32. Goodman, Monetary Sovereignty; McNamara and Jones, ‘The Clash of Institutions’; P.Hall and R.Franzese, ‘Mixed Signals: Central Bank Independence, Coordinated WageBargaining, and European Monetary Union’, International Organization 52/3 (Summer 1998), pp. 505–36. 33. I offer a suggestive account here; a systematic empirical study of the motivations of national governments is necessary to ascertain whether the sociological explanation is correct. 34. R.W.Scott, J.Meyer and associates, Institutional Environments and Organizations: Structural Complexity and Individualism (Thousand Oaks, CA: Sage 1994); J.Meyer and B. Rowan, ‘Institutionalized Organizations: Formal Structures as Myth and Ceremony’, American Journal of Sociology 83/2 (1977), pp. 340–63; W.W.Powell and P.J.DiMaggio, The New Institutionalism in Organizational Analysis (Chicago: University of Chicago Press 1991); R.W.Scott and J.Meyer, ‘Developments in Institutional Theory’, in Meyer and Scott, Institutional Environments and Organization, pp. 1–8. 35. Maxfield, Gatekeepers of Growth. 36. Scott and Meyer, ‘Developments in Institutional Theory’, p. 2. 37. McNamara and Jones, ‘The Clash of Institutions’. 38. See for example, E.Kapstein, ‘Between Power and Purpose: Central Bankers and the Politics of Regulatory Convergence’, International Organization 46/1 (Winter 1992), pp. 265–88; G.J.Ikenberry, ‘A World Economy Restored: Expert Consensus and the Anglo-American Postwar Settlement’, International Organization 46/1 (Winter 1992), pp. 289–322; P.Hall, ‘The Movement from Keynesianism to Monetarism’; K.McNamara, ‘Where Do Rules Come From? The Creation of the European Central Bank’, in N.Fligstein, W.Sandholtz and A.Stone Sweet (eds.), The Institutionalization of Europe (Oxford: Oxford University Press 2001). 39. See W.J.Barber, ‘Chile con Chicago: A Review Essay’, Journal of Economic Literature 33/4 (Dec. 1995), pp. 1941–9; M.Fourcade-Gourinchas, ‘The National Trajectories of Economic Knowledge: Discipline and Profession in the United States, Great Britain and France’ (unpublished Ph.D. thesis, University of Harvard, 2000). 40. The Financial Times, 12 Nov. 1992, p. 20; The Economist, 28 Aug. 1993, p. 16. 41. Scott and Meyer, ‘Developments in Institutional Theory’, p. 3. 42. This debate is central to the academic discussion of globalisation. See S.Berger and R.Dore, National Diversity and Global Capitalism (Ithaca: Cornell University Press 1996).

The politics of delegation

60

43. P.Tolbert and L.Zucker, ‘Institutional Sources of Change in the Formal Structure of Organizations: The Diffusion of Civil Service Reforms, 1880–1935’, Administrative Science Quarterly 28/1 (1983), pp. 22–39; P.DiMaggio and W.Powell, ‘The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields’, American Sociological Review 48 (April 1983), pp. 147–60. 44. Maxfield, Gatekeepers of Growth. 45. Ibid., p. 70. 46. D.Strang and J.W.Meyer, ‘Institutional Conditions for Diffusion’, ch. 5 in Scott and Meyer, Institutional Environments and Organization, pp. 100–112. 47. J.G.March and J.P.Olson, Ambiguity and Choice in Organizations (Bergen: Universitetsforlaget 1976); DiMaggio and Powell, ‘The Iron Cage Revisited’. 48. This analysis draws on DiMaggio and Powell’s classic analysis of the sources of organisational diffusion, ‘The Iron Cage Revisited’. They also highlight a third process, mimetic isomorphism, where copying occurs because of uncertainty. Here, I assume uncertainty as a background causal condition, and focus on the two mechanisms by which uncertainty gets translated into outcomes, coercion and normative persuasion. 49. DiMaggio and Powell, ‘The Iron Cage Revisited’, p. 67. 50. D.Rodrik, ‘Does One Size Fit All?’ Brookings Trade Policy Forum 1999 (Washington DC: Brookings Institution 1999). 51. A similar dynamic may have driven the copying of Western-style constitutions in Eastern Europe. 52. McNamara, The Currency of Ideas, ch. 6. 53. S.Berman and K.R.McNamara, ‘Bank on Democracy: Why Central Banks Need Public Oversight’, Foreign Affairs 78 (March/April 1999), pp. 2–8. 54. N.Fligstein and K.R.McNamara, ‘The Promise of EMU and the Problem of Legitimacy’, Center for Society and Economy Policy Newsletter (University of Michigan Business School, Spring 2000, www.bus.umich.edu/cse); see also A.Verdun, ‘The Institutional Design of EMU: A Democratic Deficit?’, Journal of Public Policy 18/2 (1998), pp. 107–32; and R.Elgie, ‘Democratic Accountibility and Central Bank Independence: Historical and Contemporary, National and European Perspectives’, West European Politics 21/3 (July 1998), pp. 53–76.

Rational fictions

61

APPENDIX 1 LEGAL CENTRAL BANK INDEPENDENCE AND MACROECONOMIC TRENDS (5 YEARS PRECEDING) Country Date of Legal CBI Five Years Prior Inflation Growth Unemployment Latin American States Argentina

Chile

Colombia

Ecuador

Mexico

1992

24.9

8.7

7.0

1991

171.7

8.9

6.0

1990

2314.7

0.1

9.0

1989

3080.5

−6.2

7.0

1988

343.0

−1.9

6.0

1987

131.3

2.6

5.0

17.0

10.0

5.0

1988

14.7

7.4

6.0

1987

19.9

5.7

8.0

1986

19.5

5.7

9.0

1985

30.7

2.4

12.0

1984

19.9

7.8

27.0

3.5

9.0

1991

30.5

2.1

10.0

1990

29.1

4.3

10.0

1989

25.9

3.4

9.0

1988

28.1

4.1

1987

23.3

5.4

55.0

3.7

9.0

1991

48.8

4.4

9.0

1990

48.4

2.3

6.0

1989

75.7

0.2

8.0

1988

58.2

10.5

7.0

1987

29.5

−6.0

7.0

97.5

2.0

3.0

15.5

3.6

3.0

1989

1992

1992

1993 1992

Venezuela

Country

The politics of delegation

62

1991

22.7

3.6

3.0

1990

26.7

4.4

*

1989

20.0

3.3

*

1988

114.2

1.2

*

31.4

7.3

8.0

1991

34.2

10.4

10.0

1990

40.7

6.5

10.0

1989

84.5

−8.6

9.0

1988

29.4

5.8

7.0

1987

28.1

3.6

9.0

1992

Date of Legal CBI

Five Years Prior

Inflation Growth Unemployment

Transitional Economies Albania

Belarus

Bulgaria

Czech Republic

1991

35.5

−27.7

9.0

1990

*

−10.0

10.0

1989

*

9.8

7.0

1988

*

−1.4

7.0

1987

*

−0.8

6.0

1986

*

5.6

6.0

2200.0

−13.2

2.0

1993

1188.0

−7.0

1.0

1992

1074.5

−9.7

1.0

1991

94.1

*

0.0

1990

*

*

*

1989

*

*

*

333.5

−11.7

11.0

1990

21.6

−9.1

2.0

1989

6.4

−0.5

*

1988

2.5

2.5

*

1987

2.7

5.7

*

1986

2.7

5.6

*

20.8

0.6

4.0

1994

1991

1993

Rational fictions

Estonia

Hungary

Kazahkstan

Latvia

1992

11.0

−8.5

*

1991

59.0

−15.9

*

1990

10.8

−0.4

*

1989

1.4

4.5

*

1988

0.2

2.5

*

89.0

−8.2

8.0

1992

1069.0

−23.3

5.0

1991

210.6

−11.9

2.0

1990

*

*

1.0

1989

*

*

1.0

1988

*

*

*

36.4

−11.9

9.0

1990

28.9

−4.3

2.0

1989

17.0

−0.2

*

1988

15.5

−0.1

*

1987

8.6

4.1

*

1986

5.3

4.7

*

1662.3

−9.2

1.0

1992

2568.0

−14.0

1.0

1991

147.0

−13.0

0.0

1990

*

*

*

1989

*

*

*

1988

*

*

*

951.2

−32.9

2.0

1991

124.4

−8.3

*

1990

*

*

*

1989

*

*

*

1988

*

*

*

1987

*

*

*

1993

1991

1993

1992

Country

Date of Legal CBI

Lithuania

1994

63

Five Years Prior 1993

Inflation Growth Unemployment 72.1

−9.8

5.0

410.4

−16.2

4.0

The politics of delegation

Poland

Romania

Russia

Slovak Republic

Ukraine

64

1992

1020.3

−35.0

4.0

1991

224.7

−13.4

0.0

1990

*

*

*

1989

*

*

*

70.3

−7.6

*

1990

585.8

−11.6

*

1989

251.5

0.2

*

1988

60.2

4.1

*

1987

25.2

2.0

*

1986

17.8

4.2

*

161.1

−15.1

3.0

1990

4.7

−7.4

*

1989

0.9

−5.8

*

1988

2.6

−0.5

*

1987

1.1

0.8

*

1986

0.7

2.3

*

874.7

−10.4

6.0

1992

1353.0

−18.5

5.0

1991

92.7

−12.9

0.0

1990

*

*

*

1989

*

*

*

1988

*

*

*

23.0

−3.7

13.0

1992

11.0

−8.5

11.0

1991

59.0

−15.9

7.0

1990

10.8

−0.4

*

1989

1.4

4.5

*

1988

0.2

2.5

*

1445.3

−14.0

*

1991

91.2

−13.4

*

1990

*

*

*

1989

*

*

*

1991 and 1992

1991

1993

1993

1992

Rational fictions

65

1988

*

*

*

1987

*

*

*

3.8

−1.5

8.8

1992

3.6

1.6

7.3

1991

3.2

1.6

7.0

1990

3.5

3.0

7.0

1989

3.1

3.6

8.0

1988

1.2

4.7

10.0

2.8

2.7

11.6

1992

2.9

2.2

10.3

1991

2.8

1.2

9.3

1990

3.1

2.2

8.9

1989

3.2

4.1

9.4

1988

2.8

4.5

10.0

Advanced Economies Belgium

France

Country Greece

Italy

Portugal

1993

1993

Date of Legal CBI

Five Years Prior

1993

Inflation Growth Unemployment 14.5

−1.6

9.7

1992

14.8

0.7

8.7

1991

19.5

3.1

8.0

1990

20.4

2.7

7.0

1989

13.7

3.8

8.0

1988

13.5

4.5

8.0

5.6

1.3

11.1

1991

7.3

1.4

10.9

1990

7.5

2.2

11.0

1989

6.2

2.9

12.0

1988

6.7

4.1

12.0

1987

6.0

3.1

12.0

10.6

1.9

4.1

1991

11.4

2.3

4.0

1990

13.4

4.4

5.0

1989

12.6

5.1

5.0

1992

1992

The politics of delegation

Spain

New Zealand

Japan

Britain

Ireland

66

1988

9.5

7.5

6.0

1987

9.4

6.4

8.0

3.8

2.3

24.2

1993

4.3

−1.2

22.7

1992

6.9

0.7

18.4

1991

5.9

2.3

16.0

1990

6.7

3.7

16.0

1989

6.8

4.7

17.0

7.5

0.9

7.0

1988

6.4

−0.4

6.0

1987

15.7

0.4

4.0

1986

13.2

2.1

4.0

1985

15.4

0.8

*

1984

6.2

4.9

*

−1.4

5.0

3.3

1995

−0.6

1.5

3.1

1994

0.2

0.6

2.9

1993

0.6

0.3

2.5

1992

1.7

1.0

2.2

1991

1.9

4.4

2.1

2.9

2.2

4.7

1997

2.9

3.5

5.7

1996

3.3

2.6

7.4

1995

2.5

2.8

8.1

1994

1.5

4.4

9.4

1993

2.8

2.3

10.3

5.6

8.9

7.4

1997

3.5

10.7

9.8

1996

2.3

7.7

11.5

1995

2.7

9.5

12.1

1994

1.7

5.8

14.1

1993

5.2

2.6

15.5

1994

1989

1996

1998

1998

Rational fictions

Country Finland

Netherlands

Sweden

67

Date of Legal CBI Five Years Prior Inflation Growth Unemployment 1998

2.9

5.0

11.4

1997

2.1

6.3

12.6

1996

−0.2

4.0

14.6

1995

4.1

3.8

15.4

1994

2.0

4.0

16.6

1993

2.3

−1.1

16.4

1.9

3.7

4.1

1997

2.0

3.8

5.5

1996

1.1

3.0

6.6

1995

1.8

2.3

7.1

1994

2.3

3.2

7.6

1993

1.9

0.8

6.5

0.5

3.8

5.6

1998

1.3

3.0

6.5

1997

1.2

2.0

8.0

1996

1.4

1.1

8.1

1995

3.5

3.7

7.7

1994

2.4

4.1

8.0

22.8

0.2

21.0

1990

16.7

−1.4

20.0

1989

9.2

4.9

17.0

1988

5.9

−1.9

*

1987

5.9

−0.7

*

1986

14.0

−0.2

*

13.6

0.4

9.0

1991

19.8

1.2

10.0

1990

16.7

2.3

9.0

1989

21.4

2.2

7.0

1988

18.8

3.5

*

1987

19.7

8.7

*

1998

1999

Non-Latin American Developing States Algeria

Egypt

1991

1992

The politics of delegation

Turkey

Pakistan

Vietnam

1989

68

63.3

−0.3

9.0

1988

75.4

4.4

8.0

1987

38.8

7.2

*

1986

34.6

8.7

*

1985

45.0

4.5

11.0

1984

48.3

6.3

*

9.8

1.9

5.0

1992

9.4

7.8

6.0

1991

11.7

7.7

6.0

1990

9.7

5.6

3.0

1989

7.2

4.7

3.0

1988

3.3

4.8

3.0

25.0

5.0

*

1991

67.0

4.0

*

1990

67.0

5.1

*

1989

35.0

8.1

*

1988

394.0

5.1

*

1987

316.7

2.5

*

1993

1992

Notes: *=Data not available Data on legal CBI comes from Maxfield, Gatekeepers of Growth: The International Political Economy of Central Banking in Developing Countries (Princeton: Princeton University Press, 1997), Table 4.1. Post-1994 CBI data comes from the European Monetary Institute, Convergence Reports, 1998 and 1999. Data is given for the five years prior to legal independence (I–1, I–2, I– 3…) for three measures: inflation, growth and unemployment rate. Inflation is the annual percent change in the price deflator of the national GDP. Growth is the annual percent change in national GDP. Poland starts from 1991 (CBI date used). Czech Republic and Slovak Republic figures prior to 1993 are Former Czechoslovakia numbers. Macroeconomic data comes from World Bank, World Development Indicators, 2000 or from IMF, World Economic Outlook, October 1992, October 1993, and May 2000.

Constitutional Courts and Parliamentary Democracy

ALEC STONE SWEET Constitutional judges possess broad powers to govern, in conjunction with other state officials, by virtue of an explicit act of delegation. In the terminology of delegation theory, constitutional courts are agents. These courts, when considered as functional solutions to the mixed dilemmas of contracting and commitment, appear to conform, paradigmatically as it were, to the delegation theorist’s preferred logic of institutional design. This article explores this formulation critically, in light of the themes laid out in the introduction,1 and in light of the deep transformations of parliamentary governance now taking place in Europe as a result of the diffusion of constitutional review.2 The discussion proceeds in three parts. The first examines the organisational logic of conferring constitutional review authority on a specialised court, rather than on the judiciary as a whole. By constitutional review, is meant the authority to evaluate the constitutionality of public acts, including legislation, and to annul those acts as unlawful when found to be in conflict with the constitutional law. I then turn to constitutional politics, the sources and consequences of conferring review powers on a state organ. The second part focuses on formal elements of the delegation of powers to constitutional courts. Following Majone,3 I argue that the agency metaphor that animates standard Principal-Agent (P-A) research on the politics of delegation is less appropriate than a metaphor of trusteeship, given that relevant ‘political property rights’ have, for all practical purposes, been transferred to constitutional judges. The third part summarises the major outcomes of constitutional politics, providing an explanation of the most important similarities and differences among my four cases. This explanation does not rely on concepts provided by delegation theory, per se. THE CONSTITUTION AND SEPARATION OF POWERS One can always interpret the emergence of a new institution in purely functionalist terms. Doing so will displace the (potentially crucial) question of if and how pre-existing cognitive maps, ideologies, and institutional forms constrained the choices available to those who would delegate. If functional logics are always present in episodes of conscious institutional design, focus on them is unlikely to tell us much about the (analytically prior) process of how menus of options were constituted in the first place.

The politics of delegation

70

This latter process deserves our attention further, to the extent that it conditions expectations about the likely consequences of delegating review powers, and conditions how the legitimacy of new institutions will be understood and debated (see also McNamara, this volume). A functional logic for the establishment of constitutional courts is laid out in the second part. The organisational logic of these courts, however, must be considered in light of inherited structures, in that certain deeply embedded ideas and organisational forms—models of governance—radically narrowed the possibilities for review in Europe. These structures gave advantages to the Kelsenian court, as a particular institutional solution to a particular institutional problem: how does a polity guarantee the normative superiority of the constitution, and of human rights provisions, without empowering the judiciary? The alternative, the American model of judicial review, was hardly considered. The successful diffusion of the Kelsenian court across Europe depended heavily on the belief that it could be attached to the existing architecture of the state with minimal disruption to the established order of things. The Old Constitutionalism: Legislative Supremacy and Separation of Powers In Europe, where a deep political hostility toward judges has reigned over most of the last two centuries, the delegation of power to judiciaries has been viewed as a necessary evil. Since statutes do not interpret, apply, or enforce themselves, political rulers needed courts; and judicial authority was made more palatable through separation of powers doctrines. These doctrines sought to distinguish the political function (legislating and administering) from the judicial function (the resolution of legal disputes by a judge through application of the legislator’s law). Constitutions commonly expressed these doctrines in three, linked, ways: (1) they prohibited judicial review of legislation; (2) they denied courts jurisdiction over fundamental rights; and (3) they formally subjugated judicial to legislative authority. Put simply, these doctrines assert, and then seek to sustain, the supremacy of parliamentary statute within the legal order. Even today, with legislative sovereignty fatally weakened, these ideas retain a hold on the imaginations of legislators and other political elites. Prior to the appearance of the Kelsenian constitutional court, it was widely assumed that constitutional review was incompatible with parliamentary governance and the unitary state. The parliamentary model owes its popularity to its capacity to combine centralised political authority with responsible, representative democracy. The system privileges an ideology, that of majority rule, which is realised through legislative supremacy and its corollaries. American-style judicial review, by contrast, was thought to ‘fit’ only in polities where legislative sovereignty had been rejected—such as where ‘separation’ of powers meant ‘checks and balances’ among co-equal branches of government—or where a judicial ‘umpire’ of federalism was needed. Generally, forms of non-judicial, constitutional review, took root only in the German federations, Switzerland, and Austria.4

Constitutional courts and parliamentary democracy

71

The New Constitutionalism Successive waves of democratisation in this century (for example, Germany and Italy after World War II, Southern Europe in the 1970s, the whole of Central and Eastern Europe after 1989) have transformed the juridical basis of the European state. New constitutions typically proclaim a long list of human rights; they establish a mechanism of defending the normative supremacy of the constitution; and they stipulate (nonlegislative) procedures for how the constitutional law is to be amended.5 The framers of these new constitutions, almost always the leaders of political parties acting in constituent assemblies,6 rejected American-style judicial review.7 Suspicious of judicial power, political parties sought to maintain, as far as possible, the viability of traditional separation of powers. The ‘ordinary judges’—by which is meant all non-constitutional judges, including those who sit on the various specialised courts—remain precluded from reviewing the constitutionality of statute and other public acts. Instead, constituent assemblies conferred review authority on a constitutional court. Kelsen’s Court The modern European constitutional court is the invention of the Austrian legal theorist, Hans Kelsen. Kelsen developed the European model of constitutional review, first in his role as the drafter of the constitution of the Austrian Second Republic (1920–34), and then as a legal theorist. His followers and close collaborators were present at the founding of the Federal Republic of Germany, and they proposed a variant of the Austrian system as an alternative to American judicial review. Kelsen’s legacy was secured when constitutional reformers in Spain, Portugal, and most of Central and Eastern Europe rejected the American, but adopted the European, ‘model’ of constitutional review. We can break down this model into four constituent components. First, constitutional courts enjoy exclusive and final constitutional jurisdiction. Formally, constitutional judges possess a monopoly on the exercise of constitutional review, while the judiciary remains prohibited from engaging in review. Second, terms of jurisdiction restrict constitutional courts to the settling of constitutional disputes. Constitutional judges do not preside over judicial disputes or litigation, which remain the function of judges sitting on the ordinary courts. Instead, specifically designated authorities or individuals ask questions of constitutional courts, challenging the constitutionality of specific legal acts; constitutional judges are then required to answer these questions, and to give reasons for their answers. Their decisions are final. Third, constitutional courts have links with, but are formally detached from, the judiciary and legislature. They occupy their own ‘constitutional’ space, which is neither clearly ‘judicial’ nor ‘political’. Fourth, unlike the situation in the US,8 constitutional courts may review legislation before it has been enforced, as a means of eliminating unconstitutionality prior to harm being done. Thus, in the European model, the ordinary judges remain bound by the supremacy of statute in the legal order, while constitutional judges are charged with preserving the supremacy of the constitution.

The politics of delegation

72

The Constitution and Normative Hierarchies Kelsen’s constitutional theory remains a standard, scholarly reference point for contemporary debates about the legitimacy of review, and thus deserves brief discussion.9 Kelsen focused on the problem of determining legal validity, that is, on the issue of how legal rules and public acts are to be invested with normativity, their formal authority as binding law, enforceable through sanctioned state power. Simplifying, Kelsen viewed a system based on legislative sovereignty as logically incomplete, and indeed unstable, and sought to ground the legality of state action more formally in a supra-legislative body of rules, a Grundnorm. He argued that any given act could only be considered valid, or normative, if it is enabled by, and does not conflict with, a specific—and formally superior—legal rule. Furthermore, all legal rules, in order to confer validity on lower order rules, must be capable of being enforced by a judge or ‘jurisdiction’. The move to higher law constitutionalism and constitutional review established closed, self-referential, hierarchically ordered, systems of norms. The validity of every legal rule depends upon the validity of another, higher order, legal rule; and the hierarchy of norms itself depends, for its own systemic validity, on the constitution, as Grundnorm. A ministerial decree or a police action taken in pursuance of a statute must respect the terms of that statute or be invalid, as controlled by legally designated judicial authority; and the statute itself must conform to constitutional dictates or be invalid, as controlled by constitutional judges. The legality of any norm (which reduces to constitutional legality), and the legitimacy of the legal system, are virtually one and the same thing. That is, the constitutional law, as validated by the sovereign people, comprises both a source of law, in and of itself, and the ultimate source of legitimacy for all other legal norms. Kelsen’s model of the juridical state can easily be translated into the language of delegation theory. The distinguishing feature of P-A models is that they link, as in a chain, authoritative acts of delegation from one constitutionally recognised authority to another. In Europe and North America, these acts typically take a highly legalistic form. The sovereign people (first-order principals) ratify a constitution, which delegates power to governmental bodies, like legislatures and courts.10 The statute is the normative instrument through which governments and legislatures (agents of the electorate, but second-order principals vis-à-vis ordinary judges and administrators) delegate certain specific responsibilities and powers to the courts and the administration. Principals can therefore be identified by virtue of the constitutional authority they possess to delegate powers through a specific type of normative instrument. Agents are constituted, and their activities mandated or circumscribed, through these acts of law making. In this system, the ultimate source of authority (and of the legitimacy of all delegated powers), is the constitution, which is assumed to express the will of the sovereign people. And the normativity of the constitution is guaranteed through the delegation of constitutional review powers to constitutional judges. The Positive and Negative Legislator Kelsen understood that any jurisdiction that exercised the constitutional review of statute would inevitably participate in the legislative function, hence the insistence on

Constitutional courts and parliamentary democracy

73

concentrating review in a specialised court (rather than diffusing powers of review throughout the judiciary). He nonetheless distinguished what parliaments and constitutional courts do when they legislate. Parliaments, he argued, are ‘positive legislators’, since they made law freely, subject only to the constraints of the constitution (for example, rules of procedure). Constitutional judges, on the other hand, are ‘negative legislators’, whose legislative authority is restricted to the annulment of statute when it conflicts with the law of the constitution. Kelsen’s distinction between the positive and the negative legislator relies almost entirely on the absence, within the constitutional law, of enforceable human rights. Although this fact is ignored by his modern-day followers, Kelsen explicitly warned of the ‘dangers’ of bestowing constitutional status to human rights, which he equated with natural law, because a rights jurisprudence would inevitably lead to the obliteration of the distinction between the ‘negative’ and the ‘positive’ legislator. Through their quest to discover the content and scope of the natural law, constitutional judges would, in effect, become super-legislators.11 Although Kelsen’s constitutional theory is today orthodoxy, contemporary Kelsenians claim that constitutional courts function to protect constitutional rights, and that this function is basic to the legitimacy of review.12 Nonetheless, Kelsenians ritually rehearse the negative versus positive legislator distinction as a means of bolstering these courts’ legitimacy.13 CONTRACTING AND DELEGATION In the introduction to this volume, it was proposed that some of the variation in the scope of delegation to non-majoritarian institutions might be explained by the nature of the functional problem that principals seek to resolve. We hypothesised, in particular, that the greater the commitment problem, the more likely the principals will grant relatively more discretion to an agent, and the more likely ex post controls will be relatively weak. This formulation possesses an obvious attractiveness for explaining crucial aspects of the move to constitutional review in Europe. I have argued here that constitutional-relational contracting generated a massive functional demand for an institution whose purpose would be to monitor and enforce compliance with the constitution, and especially the respect for human rights. In the first part I characterised the new constitutionalism in terms of a new hierarchy of norms, or a new state theory, but constitutions may also be seen as contracts—a set of nested bargains—among those who negotiate them. Most post-World War II European constitutions were in fact produced by intense, often conflictual, negotiations among the main, national political parties. Each party, or group of parties, arrayed or clustered on the left, right, or in the centre of the ideological spectrum, brought to the bargaining table their own constitutional preferences, ideas about how the new polity was to be constructed. Each party or grouping fought to enshrine these ideas as constitutional provisions. Negotiations proceeded in light of the relative power of the participants in the constituent assembly convention, and the extent to which constitutional preferences either converged or diverged.

The politics of delegation

74

Constitutional Bargains Parties contract with each other in order to achieve joint purposes, that is, in order to provide benefits they could not expect to realise on their own. In the case of political elites negotiating the rules of a democratic regime, such benefits include the potential to rule, to direct the polity with political legitimacy. Of course, in establishing parliamentary democracy, each political party knows that its capacity to determine events will ebb and flow depending upon relative electoral fortunes. Each party hopes to see its policy preferences substantiated as law, through legislation. Yet the parties also know that the construction of stable rules for competition among them is a necessary first step to governing. Constitutional contracting—the choice of rules governing the interplay of policy makers and policy bodies—is often easier than choosing among those policy alternatives that are possible within the rules.14 Given the consensus in favour of parliamentarianism, the parties were able to come to agreement on the basic rules, not least, since each could see itself, over time, benefiting from them. In Germany, Italy, and Spain, negotiations produced four main outcomes. First, the contracting parties established parliamentary systems of government, using relatively familiar institutional templates. The other three outcomes ran counter to political centralisation. Constitutions provided for federalism (Germany) or strong regionalism (Italy and Spain), but only after long and contentious debate. The third outcome, the codification of an enforceable body of fundamental rights and liberties, proved to be even more difficult to achieve. Elster has argued that ‘norm-free bargaining’—where ‘the only thing at stake is selfinterest’—is most likely to result in a settlement, whereas ‘norm conflicts’ frequently lead to ‘bargaining impasse’ since the parties interact with one another from the standpoint of radically opposed social values.15 Arguments about rights are inherently arguments about social values. Indeed, the general problem of determining the catalogue and content of rights provisions intermittently paralysed constituent assemblies. Simplifying what were enormously complex politics, stalemates were typically broken by giving partial victories to everyone. Major parties enshrined their own preferred set of rights, albeit watered down by the horsetrading. Despite these disagreements, new European constitutions bestow a privileged status on rights provisions. In Germany, Italy, and Spain, constitutions announce rights before state institutions are established and governmental functions distributed. Partly in consequence, academic lawyers and some judges consider rights to possess a kind of ‘supraconstitutional’ status. As discussed below, this privileged status is reinforced by rules governing constitutional amendment, which tend to treat non-rights provisions as more flexible, and rights provisions as more rigid and immutable. Thus, although the constitutional law is viewed as positive law in these countries, parts of that law—rights— can be interpreted as expressing (or codifying) natural law. The scope of constitutional rights contained in the German, Italian, and Spanish constitutions, and in texts incorporated into the French constitution by the French constitutional court, is far more extensive than the American bill of rights. European charters typically include the traditional ‘liberal’, or negative’, rights and freedoms, for example of speech, assembly, religion, equality before the law, and due process. And

Constitutional courts and parliamentary democracy

75

they enshrine newer, more ‘collective’, or ‘positive’, rights, for example, to education, employment, trade union activity, health care, the development of ‘personality’, and leisure. Further, most new European constitutions list the duties of citizens (for example, to military service, to pay taxes, to educate their children), but also of the state (for example, to provide public heath care, free education, or unemployment insurance). The structure of these provisions constitute implicit delegations of enormous law making discretion to constitutional judges. Although a few rights are declared in absolutist terms—the most important being ‘equality before the law’ provisions, found in all four countries, and ‘human dignity’ in Germany—the great majority of rights are expressly limited. To illustrate, here are a handful of expressly ‘limited’ rights: • In Spain, art. 20.1 proclaims the right to free expression, which art. 20.4 then ‘delimits’ with reference to ‘other rights, including personal honor and privacy’. Art. 33.1 declares the right to private property, while art. 33.3 provides for the restriction of property rights for ‘public benefit’, as determined by statute; • In Italy, art. 21.1 announces that ‘the press shall not be subjected to any authority of censorship’, while art. 21.6 provides that ‘printed publications…contrary to morality are forbidden’, and art. 21.7 states that parliament possesses the responsibility to ‘prevent and repress all [such] violations’; • In Germany, art. 2.1 states that ‘everyone shall have the right to the free development of her personality in so far as she does not violate the rights of others or offend the constitutional order or moral code’. Art. 10.1 proclaims that the ‘privacy of posts and telecommunications shall be inviolable’, while art. 10.2 states that ‘this right may be restricted [by] statute’. • In France, art. 11 of the 1789 Declaration of the Rights of Man declares that ‘every citizen may…speak, write, and print freely, but is responsible for the abuse of this liberty in circumstances determined by statute’. The 1946 social and economic principles of the 1946 include the proclamation that ‘the right to strike is exercised according to the laws that regulate it’. It should come as no surprise that, today, the crucial task of the constitutional judges is to determine the proper scope and limit of any rights provision in relation to another provision announcing a right, or to one that defines the duties or powers of government. The fourth outcome was the adoption of the Kelsenian court. Constitutional Courts The normative logic of a constitutionalism that limits legislative sovereignty by recognising the rights of individuals and the prerogatives of subnational government all but necessitates the establishment of a means of enforcing these rules. Further, in occupied Germany and Italy, the Americans insisted that new constitutions include rights and a means of protecting them. American-style judicial review was nonetheless rejected in these states, and across Europe thereafter, for diverse reasons. Most important, a majority of political elites remained hostile to sharing policy making authority with judiciaries; left-wing parties, especially, fiercely opposed judicial review, seeing in it the spectre of the dreaded ‘government of judges’. As discussed, Kelsenian constitutional review provided a means of defending constitutional law as higher law, while retaining

The politics of delegation

76

the general prohibition on judicial review. The Spanish Constituent Assembly, which modelled that country’s system on the German, never seriously considered not establishing a Kelsenian court. In the debates, only the Spanish Communists opposed moving to constitutional review. France has been left out of the discussion so far because the new constitutionalism emerged there by a different, far more circuitous, route. Unlike Germany and Italy, the Americans did not occupy France. Other logics favouring constitutional review are weak or non-existent. The state is unitary not federal, with a tradition of weak regional and local government. The two post-war constitutions (1946, 1958) did not establish enforceable rights; indeed, statutory supremacy remained unquestioned dogma. The constitution of the Fifth Republic was not so much the product of inter-party bargaining, but rather the choice of one political force, the entourage of General de Gaulle, acting almost unchecked. And the Left voted against the new constitution in parliament, but the Right majority prevailed. The Gaullists replaced France’s traditional, British-style, parliamentary system with a ‘mixed presidential-parliamentary’ one, strengthening the executive. The constitution established a Constitutional Council, but its purpose was to guarantee the dominance of the executive (the government) over a weakened parliament. Beginning in 1971, however, the Council began to assert its independence. In that year, for the first time, it declared a government-sponsored law unconstitutional on the grounds that the law violated constitutional rights. This decision paved the way for the incorporation of a charter of rights into the 1958 constitution, a charter that the Council has taken upon itself to expand and enforce creatively.16 Thus, for the first time, and against the wishes of the Gaullists and all of the other political parties in 1958, France has both a bill of rights and an effective constitutional court. Constitutional Courts, Relational Contracting, and Commitment Constitutional contracting itself produces a demand for the establishment of something like a constitutional court. Generally, contracts can be said to be ‘incomplete’ to the extent that there exists meaningful uncertainty as to the precise nature of the commitments made. Due to the insurmountable difficulties associated with negotiating rules for all possible contingencies, and given that, as time passes, conditions will change and the interests of the parties to the agreement will evolve, all contracts are incomplete in some significant way.17 Most agreements of any complexity are generated by what organisational economists call ‘relational contracting’. The parties to an agreement seek to broadly ‘frame’ their relationship by agreeing on a set of basic ‘goals and objectives’, fixing outer limits on acceptable behaviour, and establishing procedures for ‘completing’ the contract over time.18 Modern European constitutions—complex instruments of governance designed to last indefinitely, if not forever—are paradigmatic examples of relational contracts. Much is left general, even ill-defined and vague, as in the case of rights. Generalities and vagueness may facilitate agreement at the ex ante, constitutional moment. But vagueness, by definition, is normative uncertainty, and normative uncertainty threatens to undermine rationales for contracting in the first place. The establishment of constitutional review can be understood as an institutional response to the incomplete contract, the linked problems

Constitutional courts and parliamentary democracy

77

of uncertainty and enforcement. Each party to it has an interest in seeing that other parties obey their obligations, and will be reprimanded when failing to do so. Review functions to clarify, over time, the meaning of the contract, and to monitor compliance. The constitution itself, through rules of jurisdiction, delegates authority to constitutional judges, determining how they are to exercise review.19 Jurisdiction Simplifying, there are three basic types of review jurisdiction: abstract review, concrete review, and the individual complaint procedure. Abstract review is ‘abstract’ because the review of legislation takes place in the absence of litigation (in the absence of what Americans call ‘a concrete case or controversy’). Concrete review is ‘concrete’ because the review of legislation, or other public act, constitutes a separate stage in an ongoing judicial process (that is, litigation in the ordinary courts). In individual complaints, a private individual alleges the violation of a constitutional right by a public act or governmental official, and seeks redress from the court for this violation through direct petition. These modes of (and other important variations in) constitutional review jurisdiction were earlier elaborated by Kelsen.20 Abstract review processes result in decisions on the constitutionality of legislation that has been adopted by parliament but has not yet entered into force (France), or that has been adopted and promulgated, but not yet applied (Germany, Italy, Spain). It is initiated by specifically designated, elected politicians. Executives and legislators (France, Germany, Spain), federated member states or regional governments (Germany, Italy, Spain), or an ombudsman (the Defender of the People in Spain) may, within prescribed time limits, attack legislation as unconstitutional before the constitutional court. These attacks are made in writing, in documents here called referrals. In practice, nearly all referrals are made by members of opposition parties, against legislation proposed by the majority, or governing, party or parties. Referrals usually suspend the legal force of the referred law (France, Germany, Italy) pending a ruling, but in Spain the law may be applied notwithstanding a referral. Concrete review is initiated by the judiciary (Germany, Italy, Spain) in the course of litigation in the courts. Judicial officials are authorised to refer constitutional questions— is a given law, legal rule, judicial decision, or administrative act constitutional?—to constitutional judges. The general rule is that a presiding judge must go to the constitutional court if two conditions are met: (1) that the constitutional question is material to litigation at bar, that is, who wins and who loses depends on the answer to the question; and (2) there is reasonable doubt in the judge’s mind about the constitutionality of the act or rule in question. Referrals suspend proceedings pending a review by the constitutional court. Once rendered, the constitutional court’s judgment is sent back to the referring judge, who then decides the case with the help of the ruling. Constitutional decisions are binding on the judiciary. Ordinary judges are not permitted, at least in theory, to determine the constitutionality of public acts on their own (but see the section below on Constitutional Adjudication). Individual complaints (a Verfassungsbeschwerde in German, an amparo in Spanish) bring private actors into the mix. Once judicial remedies have been exhausted, individuals (Germany, Spain), and an ombudsman (Spain) have the right to go to directly

The politics of delegation

78

to constitutional judges. In Spain, individuals may attack the act of any public official that they think has violated their constitutional rights, with one important exception. Technically, the amparo may not be used to attack a statute, although this occurs under certain conditions. In Germany, persons who believe that they have suffered as a result of a specific infringement of the constitutional law, by a public official, may file a complaint.21 A statute may be targeted, as long as the complainant’s rights have been abridged in some ‘personal’ and ‘direct’ way. In both countries, appeals of judicial decisions (on the grounds that due process has been denied) comprise, by far, the largest class of complaints. Composition Two modes of appointment exist, nomination and election. Where nomination procedures are used, the appointing authority simply names a judge or a slate of judges; no countervailing confirmation or veto procedures exist. Such is the case of France, where all constitutional judges are named by political authorities (the Presidents of the Republic, the Assembly, and the Senate). Italy and Spain have mixed nomination and election systems. Where election systems are used, a qualified majority (a 2/3 or 3/5 vote) within a parliamentary body is necessary for appointment. Because in Germany, Italy, and Spain no single party has ever possessed a super-majority on its own, the qualified majority requirement effectively forces the parties to negotiate with each other in order achieve consensus on a slate of candidates. This bargaining process occurs in intense, behindclosed-doors negotiations. In practice, these negotiations determine which party will fill vacancies on the court, with allocations usually roughly proportionate to relative parliamentary strength. Compared with the US, where federal judges serve life terms, members of European constitutional courts serve either nine or 12 year, non-renewable, terms. Agency and Trusteeship When those who possess or manage political authority effectively transfer, through an act of delegation, their political property rights to a new institution, the P-A framework loses much of its relevance (see the introduction to this volume). In Majone’s terms, ‘standard agency theory does not provide satisfactory models’, either for analysing relations between those who have delegated and the new institution, or for ‘understanding the governance structure in which such relations are embedded’.22 Instead, Majone proposes a model of trusteeship in which ‘relational contracting’ and ‘imperfect commitment’ lead political rulers to delegate broad ‘fiduciary’ powers to a particular kind of agent—a trustee—and then to guarantee that trustee’s independence. As I have argued here and with respect to the Treaty of Rome and the European Court of Justice:23 new European constitutions are relational contracts, par excellance; constitutional courts have been delegated broad powers to govern by those who, at the same time, have largely relinquished the powers usually held by principals; and mechanisms of control that do remain to control the court are relatively weak. Put differently, the move from the old to the new constitutionalism replaced a system of legislative sovereignty, in which governments and parliamentarians (that is, political parties) governed, with a system of

Constitutional courts and parliamentary democracy

79

trusteeship in which governments and parliaments are being governed by rules and processes largely outside of their control. To illustrate, consider this stylised account of the ordinary judge as agent. The judge’s principal is the parliament, and the normative instrument parliament controls—statute— fixes, substantively, the terms of the judge’s policy mandate. According to separation of powers dogma, the judge is a ‘slave to the codes’, whose purpose it is to enforce parliament’s will, as that will is expressed in legislation. In reality, we know that European judiciaries engage in such extensive interpretation of the various legal codes that these statutes often mean only what the courts say they mean. Nevertheless, the principal remains in charge. If ministers or parliamentarians notice that a judge has applied a statutory provision in a way that they did not intend and do not like, the law can be changed. Thus, to the extent that an agency problem can be identified, it can be corrected: the principals overturn judicial decisions by reworking the normative instrument that they themselves directly control, thus precluding the offending judicial interpretation. Other things being equal, the decision rule governing the P-A relationship—a majority vote of the parliament—favours control by the principals, and this rule is constitutionally frozen into place. (Further, judicial officials are typically employed and managed by ministries of justice which, we know empirically, gives political elites some leverage over judicial outcomes.) If for ordinary adjudication, the relationship between principals and agents appears straightforward, the situation becomes rather confused for constitutional adjudication— that is, if one insists on describing the situation in P-A, rather than fiduciary, terms. One basic service that constitutional judges render is to regulate the actions of the government and parliament themselves. Constitutional judges have the duty to control the exercise of legislative authority and all of those acts pursuant to the adoption of statute. Depending upon the relevant constitutional rules in place, the political parties may be able to overturn constitutional decisions, or restrict the constitutional court’s powers, but only if they can reconstitute themselves as a jurisdiction capable of amending the constitutional law. This last point deserves emphasis: legislators or ministers are never principals in their relationship to constitutional judges. Further, the decision rules governing constitutional revision processes are more restrictive than those governing the revision of legislation. These decision rules, as will be demonstrated, heavily favour the continuous dominance of constitutional judges over the interpretation of the constitutional law. In practice, some members of political parties who exercise executive or legislative powers participate in some of the functions usually associated with principals, since they appoint members of the courts, and they can sometimes initiate revision of the constitution in order to avoid the consequences of constitutional rulemaking on the part of the judges. Nonetheless, by establishing (1) the normative superiority of the constitution, (2) a review organ, and (3) specific procedures for constitutional revision, they have radically reduced their own influence over the development of the constitution. If they may sometimes act as principals, they are more often merely ‘players’ within the rule structures provided by the constitution. They compete with each other in order to be in the position to govern, and, once in power, they legislate, among other things. Constitutional adjudication is implicated in the exercise of legislative power. If in exercising review authority, the judges simply controlled the integrity of parliamentary procedures, and not the substance of legislation, the judges would be relatively minor

The politics of delegation

80

policy makers, akin to Kelsen’s ‘negative legislator’. But the judges possess jurisdiction over rights that are, by definition, substantive constraints on law making powers. The political parties thus transferred their own entirely unresolved problems—what is the nature and purpose of any given rights provision? and what is the normative relationship of that provision to the rest of the constitutional text?—to judges. This transfer constitutes a massive, virtually open-ended, delegation of policy making authority. Similarly, review jurisdiction organises the elaboration of higher law rules governing federalism and regional autonomy in Germany, Italy, and Spain. To the extent that it is costly to activate constitutional review of legislation, the importance of the review within policy processes, and the authority of judges over outcomes, would be mitigated. But initiating abstract review is virtually costless for oppositions; concrete review procedures entail delays for the litigants, but other costs are essentially borne by the state; and individual complaints can be scribbled by anyone on a crumpled piece of notebook paper. To the extent that those authorised to refer constitutional questions to constitutional judges find it difficult to translate their (policy) interests into a constitutional claim, constitutional adjudication would be less likely to be central to the work of legislators and the ordinary judges. But the huge, expansive array of rights provisions available for instrumental use on the part of the majority’s opponents in fact invites such use. Even so, if it were relatively easy for the governing majority to overturn the case law of the court, or to curb the judges’ powers, the court’s authority over the legislature and the courts might be fleeting. What kinds of control mechanisms are available to the political parties is therefore relevant to much policy autonomy judges will presumably wield. Mechanisms of Control We can sort control mechanisms into two broad categories: direct and indirect. Direct controls are formal (they are established by explicit rules) and negative (they annul or authoritatively revise the court’s decisions, or curb the court’s powers). For the purposes of this discussion, let us assume that the principal is that entity (or entities) which has the authority to revise the constitution in order to overturn a decision of the agent, the constitutional court. To exercise control by direct means, the principal must succeed in amending the constitution. The rules governing constitutional amendment are relatively permissive of agent autonomy, through restricting the principal’s control. Most of the German constitution can be revised by a two-thirds majority vote of the Bundestag and the Bundesrat; however, the Basic Law (art. 19.2, art. 79) explicitly precludes any restriction on ‘the essential content of a basic right’; content, of course, is to be determined by the Court. Many interpret these clauses as meaning that constitutional rulings interpreting rights provisions could not be overturned by the constituent power. In Spain, constitutional revision is normally accomplished by a three-fifths vote of the Cortes and the Senate, upon initiation by either chamber or the government. But when a proposed amendment concerns a rights provision, a two-thirds majority in each chamber is required; and if this majority is achieved, then parliament is dissolved, and the proposed amendment is submitted to the people for ratification by referendum. The Italian constitution can be amended by majority vote of the lower and upper houses, deliberating separately; but

Constitutional courts and parliamentary democracy

81

unless a two-thirds majority vote of both is secured, a ratification referendum is required if one-fifth of either house, or five regional councils, or 500,000 voters, so request. In Spain and Italy, core rights provisions are considered by many legal scholars to be immune to revision. As in Germany, such provisions are consider to be possessed of ‘supraconstitutional’ status (a natural law position). In fact, no rights provision has been changed in Spain or Italy, despite recurrent calls to do so. In France, where rights were not meant by the founders to be included in the constitution of the Fifth Republic, the revision process is the most permissive. Political parties can transform themselves into a constitutional assembly on their own: once both parliamentary chambers have adopted the text of a constitutional amendment, a constitutional congress is organised, wherein deputies and senators, sitting together, revise the constitution by a three-fifths vote. Further, the French court is the only one of the four constitutional courts to have determined that rights provisions are not privileged constitutional norms; every part of the constitution is apparently open to revision as long as prescribed procedures are followed. Indirect controls are informal and indirect. These mechanisms are effective in as much as the agent internalises the principal’s interests, or takes cues from the revealed preferences of the latter, and acts accordingly. It is generally assumed that the extent to which the agent does so is commensurate to the credibility of the threat that its principal will activate direct controls. Indirect controls operate according to the logic of deterrence: the more credible the threat of punishment, the more the agent will constrain itself by behaving as if the principal’s interests were its own. This control is registered as an anticipatory reaction operating as a constraint on the court’s behaviour. If this assumption holds, we have no good reason to think that constitutional courts (or for that matter the ECJ, when it interprets the treaties) are systematically constraining themselves for fear of being punished. CONSTITUTIONAL ADJUDICATION AND PARLIAMENTARY GOVERNANCE The introduction to this volume defined governance generically, as the mechanisms, or processes, through which the rule systems in any social system are adapted on an ongoing basis to the needs and purposes of those who live under them.24 In parliamentary systems, the government and legislature govern by building or refining the various codes and other legal regimes. Administrators govern when they apply rules to situations and to people. When a court resolves a legal dispute by interpreting (an exercise akin to legislating) and applying (an exercise akin to administering) statutory provisions, the judge, too, governs.25 The constitutional law comprises a body of rules that authoritatively conditions each of these activities. Thus, because constitutional judges possess broad powers to interpret authoritatively, and thereby (re)construct, the constitutional law (the normative context for all other public decision making), constitutional judges govern to the extent that they actually exercise these powers. There now exists a substantial social science literature on the impact of constitutional courts on West European politics.26 Rather than rehearse the empirical findings of this literature, the argument will focus summarily on how and to what extent constitutional

The politics of delegation

82

adjudication has transformed the nature of parliamentary governance. This transformation can be observed and to some degree measured by examining the impact of constitutional decision making on the work of legislators and judiciaries. Delegation theory can tell us something, but only relatively little, about the pace, content, or depth of these changes. Discretion and Autonomy The introduction to this volume raised a classic issue of delegation theory, namely how to conceptualise and account for the gaps that inevitably develop once new institutions actually begin to perform their assigned tasks. Such gaps, this volume asserts, ought to be conceived in terms of a theoretical ‘zone of discretion’, which is ‘constituted by (a) the sum of delegated powers (policy discretion) granted by the principal to the agent, minus (b) the sum of control instruments available for use by the principal to shape (constrain) or annul (reverse) outcomes that emerge as the result of the agent’s performance of tasks. A situation of trusteeship, wherein the agent exercises fiduciary responsibilities, the zone of discretion is, by definition, unusually large. In some places and in some domains, the discretionary powers enjoyed by constitutional courts are close to unlimited. Nothing in delegation theory can tell us what constitutional courts will actually do with their discretion, no functional theory of institutional design, per se, could. The best one can do is to generate predictions about outcomes in light of two variables: the distribution and intensity of the principals’ preferences for outcomes (assuming multiple principals), and the presumed effectiveness of control mechanisms (as built into the design of the delegation). We can, of course, assume that the framers of constitutions wished to see certain outcomes produced and others avoided. But, given the ‘zone of discretion’ enjoyed by constitutional judges, we should not expect the ex ante policy preferences of principals to be decisive. In fact, the outcomes to which we now turn were neither intended nor anticipated at the ex ante moment of institutional design. Constitutional Adjudication and the Legislator The impact of constitutional adjudication on the polity proceeds through constitutional decision making, which results in law making. By constitutional decision making is meant the process of determining the meaning of a given constitutional provision (or set of provisions) in order to resolve a dispute about the constitutionality of a public act, including statutes and the exercise of public authority (for example, administering and judging) pursuant to statute. In deciding, the constitutional court simultaneously resolves a legislative dispute and enacts the constitution. This enactment, or constitutional rule making, clarifies or revises the constitutional law, by authoritatively interpreting it. The effects, or impact, of constitutional rule making on the work of the legislator can be both direct and indirect. When a court declares a bill or statute unconstitutional, it vetoes the bill. The court has intervened directly in the legislative process, in Kelsen’s phrase, as a ‘negative legislator’. Such annulments are rare, but often spectacular, political events. Important legislation vetoed by constitutional courts include the liberalisation of abortion in Germany (1975, 1992) and Spain (1985), the nationalisation of industry and financial institutions in France (1981), the reform of German university governance (1973), the bid to introduce affirmative action in France (1982), and, in all

Constitutional courts and parliamentary democracy

83

countries, important revisions of the penal codes, and legislative moves to privatise industry and to establish anti-trust regimes for the press and television sectors. Yet constitutional judges have also come to exercise substantial, ‘indirect’ authority over legislating, to the extent that case law, in any given policy domain, meaningfully guides the decision making of legislators in that domain. There are two archetypal forms of indirect impact, both of which result when legislators anticipate the policy preferences of the constitutional court, as these preferences are revealed through constitutional decision making. The first, ‘autolimitation’, refers to the exercise of self-restraint on the part of the government and its parliamentary majority in anticipation of an annulment by the constitutional court.27 We observe autolimitation when the government and its parliamentary majority take decisions, during the legislative process, that (1) sacrifice initially held policy objectives in order to (2) reduce the probability that a bill will either be referred to the court, or be judged unconstitutional. Hundreds of bills have been altered, even gutted by such decisions including the French decentralisation (1982) and press (1984) laws, and the industrial co-determination bill in Germany (1976). The second form of anticipatory reaction, ‘corrective revision’, takes places after a bill has been annulled. Corrective revision refers to the re-elaboration of a censured text in conformity with the court’s decision in order to secure promulgation. These legislative processes are highly structured by case law, to the extent that the judges have already made their legislative choices explicit, and that oppositions work to monitor the majority’s compliance with the ruling. The logic of corrective revision is straightforward. Once a constitutional court has annulled a bill and then gone on to state in precise detail what a constitutional version of the bill would look like, the legislative majority is faced with a choice. In theory, it has four options. First, it can engage in a corrective revision process, securing constitutionality by deferring to the policy preferences of the constitutional court. This option is almost always the one selected. The majority knows that the surest way to secure promulgation of beleaguered legislation is to concede part of their law making authority to constitutional judges. Second, it can forego the legislation entirely. This is rarely a viable option for important pieces of legislation, since it is usually better for the majority to get some part, rather than nothing, of what it wanted. Third, it could seek to circumvent the court’s ruling, for example, by creatively reformulating the legislation. In such cases, law makers are, in effect, playing ‘chicken’ with the court, daring the judges to annul the bill a second time. Although they often interpret a court’s ruling as narrowly as possible, to allow for maximum legislative discretion over the precise terms of the correction, it seems there is no important case where legislators have revised a censured bill but blatantly ignored the court’s dictates.28 A final option exists: the majority can revise the constitution in order to make constitutional those acts that had been censored, repudiating the court in the process.29 The feasibility of this option depends on how difficult it is to revise the constitution.30 Corrective revision leads to the dominance of constitutional courts over policy outcomes. One finds relatively high levels of such dominance over, among others, penal competition and electoral law, and over legal regimes governing expropriation and privatisation, abortion, education, and industrial relations. The extent to which constitutional courts intervene in legislative processes and shape outcomes varies cross-nationally, as a function of three factors.31 The first is the existence

The politics of delegation

84

of abstract review jurisdiction. Where abstract review does not exist (for example, Italy), the court’s capacity to shape outcomes is reduced, other things being equal. The second is the number of veto points in the legislative process (the extent of centralised, executive control over the policy process as a whole). Where there are relatively more veto points, oppositions will make use opportunities to block or water-down the majority’s bills before turning to the court, and the parliament will adopt fewer ‘radical’ reforms. The case of France lies at one end of the spectrum (the court is the usually the only veto point available to the political minority), the cases of Germany and Italy lie at the other. A third factor, the relative development of the constitutional law through constitutional rule making, is endogenous to the politics described here, to the extent that constitutional case law ‘feeds back’ onto the legislature, reproducing the same behaviours and reinforcing the same logics that provoked constitutional review in the first place. Oppositions go to the court in order to win what they would otherwise lose in the ‘normal’ processes (they would be outvoted). Abstract review petitions enable constitutional courts to construct the constitutional law, to extend techniques of control over law making activities, and (the same thing) to make policy. As the constitutional law expands to more and more policy areas, and as it becomes ‘thicker’ in each domain (more dense, technical, and differentiated), so do the grounds for constitutional debate. The process tends to reproduce itself, and in so doing, the court’s authority over the legislator grows, and the legislator’s discretion reduces. Constitutional Adjudication and the Judiciary Although traditional separation of powers doctrines preclude judicial review of legislative acts, the new constitutionalism enables review by Kelsenian courts, and organises a set of interactions between the judiciary and constitutional judges. Concrete review processes require ordinary judges to participate in the scrutiny of legislation, weakening (at least potentially) the domination of the codes over judicial decision making. But judicial officials also have an interest in activating constitutional review to the extent that they wish (1) to participate in the construction of the constitutional law, and (2) to remove unconstitutional (and perhaps unwanted) laws. Both are new powers for them. The constitutional court also benefits from the system. Concrete review enlists potential litigants and the judiciary in a general, relatively decentralised effort to detect violations of the constitution; judicial officials provide the constitutional court with a case load; and the rules governing the process generally favour the constitutional court’s control over outcomes. Given this mix of formal rules and corporate interests, we can only expect that concrete review will undermine the dogma of the juridical supremacy of legislative acts even further, subverting traditional separation of powers schemes. In fact, the development of constitutional review has transformed the role and function of the law courts, a development known as ‘the constitutionalisation of the legal order’. By constitutionalisation I mean the process through which: • constitutional norms come to constitute a source of law, capable of being invoked by litigators and applied by ordinary judges to resolve legal disputes, including in the domain of private law;

Constitutional courts and parliamentary democracy

85

• the constitutional court, through its jurisdiction over concrete review referrals and individual complaints, comes to behave as a kind of super-court of appeal for the judiciary, involving itself in the latter’s tasks of fact finding and rule application; and • the techniques of constitutional decision-making become an important mode of argumentation and decision-making in the ordinary courts. Constitutionalisation is partly the logical, normative consequence of the direct effect of rights provisions, and in part the product of complex dialogues between constitutional judges and the judiciary. There is space here only to mention the most important outcomes of this process, and to discuss variation among cases.32 Constitutionalisation has subverted three very powerful, deeply entrenched dogmas about the character and functioning of Continental legal systems. First, the traditional notion that the various legal codes in each country constituted more or less autonomous realms, governed by different sources of law and different principles of adjudication, is gradually being replaced by a new view, Kelsenian in inspiration, that the constitutional law unifies these domains into a more or less coherent legal order. Second, the quasiofficial myth of judges as slaves of the codes, prohibited from creatively interpreting and rewriting the laws, has been shattered. The very existence of constitutional review subverted the ‘sacred’ nature of statute within the legal order; and the practice of constitutional review, by processing concrete review referrals, for example, socialised judges into a new role, that of protecting the legal order from those acts contaminated by unconstitutionality. Third, orthodox accounts of the division of powers between constitutional judges and the ordinary judges have been rendered obsolete (see below). The fact that these dogmas persist has much to do with the inability of judges, legal scholars, and political elites, faced with constitutionalisation, to reconstitute the legitimacy of the juridical order differently than they have in the past. These commonalities mentioned, there is also significant cross-national variation in the pace and scope of constitutionalisation. Generally, the greater the level of interaction between the constitutional court and any given court system (civil, administrative, labour, tax, and so on), the more the distinction between constitutional jurisdiction and ordinary jurisdiction collapses. That is, as constitutionalisation deepens, ordinary judges necessarily behave as constitutional judges—they engage in principled constitutional reasoning and resolve disputes by applying constitutional norms. And, as constitutionalisation deepens, constitutional judges become more deeply involved in what is, theoretically, in the purview of the judiciary: they interpret the facts in a given dispute, and they review the relationship between these facts and the legality of infraconstitutional norms. Ordinary judges begin to treat the codes, more overtly and explicitly than they had previously, less as a set of sacred commands issuing from the sovereign, and more as a system of rules that must be co-ordinated with other systems of rules in light of changing conditions. Cross-national differences in the scope, pace, and intensity of constitutionalisation appear to be closely tied to the existence, or non-existence, of particular modes of review. In Germany and Spain, where abstract review, concrete review, and individual complaint procedures coexist, extensive constitutionalisation has proceeded rapidly. Indeed, well over 90 per cent of all individual complaints allege that a judicial decision has failed to take into account the constitutional law or the constitutional court’s decision making. For the court to decide on the merits of such claims, it must delve deeply into the workings of

The politics of delegation

86

the judiciary, and it has the power to impose its own preferred outcome on any recalcitrant judge. In Italy, the absence of an individual complaint mechanism has reduced the capacity of the constitutional court to control judicial outcomes (as the absence of abstract review mechanisms has reduced the court’s capacity to control legislative outcomes). In consequence, constitutionalisation has been more gradual and piecemeal, and the high civil court, Cassazione, has retained substantial interpretive autonomy. In France, where promulgated statute retain their formal, sovereign character, and no formal links between ordinary and constitutional jurisdictions exist, a primitive form of constitutionalisation can nevertheless be observed. Last, there is increasingly compelling evidence that the ordinary judges are moulding, by creatively interpreting, the various codes, without first referring questions to the constitutional court.33 Judges read constitutional principles into the codes, through statutory interpretation (principled, constitutional construction of statutes), and sometimes they have gone beyond, or even contradicted, more restrictive interpretations of rights issued by the constitutional court. By foregoing constitutional referrals, ordinary judges enhance their own autonomy. In this way, both legislative intent and the alleged monopoly of the constitutional court on constitutional interpretation are subverted. CONCLUSION In summary, both discrete organisational and more general functional logics underpin the diffusion of the Kelsenian constitutional court across Europe. These logics are easily synthesised. Historical legacies, notably in the form of existing models of governance and rigid separation of powers doctrines, reduced the options available to those who sought to resolve their own social dilemmas through delegating to a non-majoritarian institution. Little in delegation theory, however, can explain the major outcomes produced by constitutional politics. Indeed, the ongoing exercise of review by constitutional judges has served to (re)construct the constitutional law in ways that subvert the schema initially assumed to legitimise the role and function of review itself. Kelsen was right. The delegation of rights jurisdiction to a constitutional organ, effectively insulated from political controls, has engendered the gradual but inevitable collapse of separation of powers doctrines. Today, ministers and parliamentarians govern with constitutional judges. And the judiciary, partly in interactions with constitutional judges, has radically expanded its capacity to control policy outcomes.

Constitutional courts and parliamentary democracy

87

NOTES 1. M.Thatcher and A.Stone Sweet, ‘Theory and Practice of Delegation to Non-Majoritarian Institutions’, this volume. 2. This chapter is based partly on themes developed in A.Stone Sweet, Governing with Judges: Constitutional Politics in Western Europe (Oxford: Oxford University Press 2000). 3. G.Majone, ‘Two Logics of Delegation: Agency and Fiduciary Relations in EU Governance’, European Union Politics 2/1 (2001), pp. 103–22. See also T.Moe, ‘Political Institutions: The Neglected Side of the Story’, Journal of Law, Economics, and Organisation 6 (1990), pp. 213–53. 4. K.von Beyme, ‘The Genesis of Constitutional Review in Parliamentary Systems’, in C. Lanfried (ed.), Constitutional Review and Legislation: An International Comparison (BadenBaden: Nomos 1989); A.Stone, The Birth of Judicial Politics in France (Oxford: Oxford University Press 1992), chs. 2, 9. 5. In the old constitutionalism (for example, in the French Third Republic), constitutional rules could be revised or otherwise recast through statutory instruments. 6. The exception is France, see Stone, The Birth of Judicial Politics in France, ch. 3. 7. Only the example of Greece might comfort proponents of American judicial review, although there are a few curious mixed systems (for example, that of Estonia). Some older systems, for example, Sweden, possess systems of constitutional judicial review. 8. But see A.Stone, ‘Qu’y a-t-il de concret dans le contrôle abstrait aux États-unis?’, Revu française de droit constitutionnel 34 (1998), pp. 227–5 9. H.Kelsen, ‘La Garantie Juridictionnellede la Constitution’, Revue du Droit Public 44 (1928), pp. 197–257. 10. Germany (art. 20.2): ‘All governmental authority emanates from the People. It shall be exercised by the People by means of elections and voting, and by specific legislative, executive, and judicial organs’. France (art. 3): ‘Sovereignty resides in the Nation. No body, no individual, may exercise authority which has not been expressly granted’. Italy (art. 1): ‘Sovereignty belongs to the People, who exercise it in the manner and within the limits laid down by the Constitution’. Spain (art. 1.2): ‘National sovereignty belongs to the Spanish People, from whom all powers of state emanate’. 11. Kelsen, ‘La Garantie Juridictionnelle de la Constitution’, pp. 221–41. 12. Although this point will not be explored further here, the orthodox position of modern Kelsenians, and many judges, is that rights possess a kind of ‘supraconstitutional’ status (their contents can not be altered by constitutional revision). This is akin to a natural law, not a Kelsenian or positivist, position. 13. See Stone Sweet, Governing with Judges, ch. 5. 14. G.Brennan and J.Buchanan, The Reason of Rules: Constitutional Political Economy (Cambridge: Cambridge University Press 1985). 15. J.Elster, The Cement of Society: A Study of Social Order (Cambridge: Cambridge University Press 1989), pp. 215, 244–47. 16. This charter is composed of rights texts mentioned or contained in the preamble to the 1946 constitution, which itself is mentioned in the preamble to the 1958 constitution. 17. A ‘complete’ contract ‘would specify precisely what each party is to do in every possible circumstance and arrange the distribution of realised costs and benefits in each contingency so that each party individually finds it optimal to abide by the contract’s terms’, P.Milgrom and J.R.Roberts, Economics, Organization and Management (Englewood Cliffs, NJ: Prentice-Hall International 1992), p. 127. 18. Ibid., pp. 127–33.

The politics of delegation

88

19. Although most important rules of jurisdiction are constitutional in nature, some jurisdictional details are regulated by statutes called organic laws. 20. Kelsen, ‘La Garantie Juridictionnellede la Constitution’. 21. In Germany, local officials can also file complaints if they believe that the constitutional prerogatives of local government have been abridged by the public actions of another level of government. 22. Majone, ‘Two Logics of Delegation’, p. 104. 23. Stone Sweet, Governing with Judges’, and A.Stone Sweet and J.Caporaso, ‘From Free Trade to Supranational Polity: The European Court and Integration’, in W.Sandholtz and A.Stone Sweet (eds.), European Integration and Supranational Governance (Oxford: Oxford University Press 1998), pp. 92–133. 24. A.Stone Sweet, ‘Judicialisation and the Construction of Governance’, Comparative Political Studies 31 (1999), pp. 147–84. 25. See the contribution of M.Shapiro to this volume. 26. For example, S.Kenney, W.Reisinger and J.Reitz (eds.), New Approaches to Law and Politics in Europe (New York: MacMillan 1999); D.P.Kommers, The Constitutional Jurisprudence of the Federal Republic of Germany (Durham, NC: Duke University Press 1997); C.Landfried, Bundesverfassungsgericht und Gesetzgeber (Baden-Baden: Nomos 1984); C.Landfried (ed.), Constitutional Review and Legislation: An International Comparison; M.Shapiro and A.Stone (eds.), special issue of Comparative Political Studies 26 (1994); ‘The New Constitutional Politics of Europe’; Stone, The Birth of Judicial Politics in France; Stone Sweet, Governing with Judges; M.L.Volcansek (ed.), special issue on Judicial Politics in Western Europe, West European Politics, 15 (1992); M.L.Volcansek, Constitutional Politics in Italy (London: Macmillan 2000). 27. The term was first coined by L.Favoreu, ‘Décentralisation et Constitution’, Revue du Droit Public 98 (1982), pp. 1259–95. In Germany, the phenomenon has been studied extensively by Landfried, Bundesverfassungsgericht und Gesetzgeber. Game theoretic models of autolimitation are currently being developed, see Georg Vanberg, ‘Abstract Judicial Review, Legislative Bargaining, and Policy Compromise’, Journal of Theoretical Politics 10 (1998), pp. 299–326; and A Stone Sweet, ‘Rules, Dispute Resolution, and Strategic Behavior: Reply to Vanberg’, Journal of Theoretical Politics 10 (1998), pp. 327–38. 28. The German government was urged to do so by its parliamentary supporters after the 1975 abortion ruling, and the French government considered doing so in reaction to the nationalisations decisions (1981–82). 29. The legislature still must re-adopt the law subsequent to the entry into force of the constitutional revision. 30. I know of only one important instance: in 1993, the French constitution was revised to enable the adoption of a law, previously annulled in important respects, on immigration and asylum. 31. For a fuller discussion see Stone Sweet, Governing with Judges, ch. 2.

Constitutional courts and parliamentary democracy

89

32. For a summary and comparative perspective, see Stone Sweet, Governing with Judges, pp. 114–25. For Spain, see P.Bon, ‘La Constitutionnalisation du Droit Espagnol’, Revue Française de Droit Constitutionnel 5 (1991), pp. 35–54. For France, see L.Favoreu, ‘Le Droit Constitutionnel, Droit de la Constitution et Constitution du Droit’, Revue Française de Droit Constitutionnel 1 (1990), pp. 71–89; For Italy, see A.Pizzorusso, ‘The Italian Constitution: Implementation and Reform’, Jahrbuch des Öffentlichen Rechts der Gegenwart 34 (1985), pp. 105–21. For Germany, see P.E.Quint, ‘Free Speech and Private Law in German Constitutional Theory’, Maryland Law Review 48 (1989), pp. 247–347. 33. There is no systematic literature on the phenomenon, but see G.Bognetti, ‘Direct Application and Indirect Impact of the Constitution in the Italian Legal System’, in Italian National Reports to the XIth International Congress of Comparative Law, Carcas (Milano, Giuffré Editore 1982); R.Errera, ‘Recent Decisions of the French Conseil d’Etat’, Public Law 2 (1998), pp. 152–4; S.D.-R.Guanter, ‘Employee Privacy in Spanish Labor Relations’, Comparative Labor Law Journal 17 (1995), pp. 122–38; and Quint, ‘Free Speech and Private Law in German Constitutional Theory’.

Institutional Choice and Bureaucratic Autonomy in Germany

MARIAN DÖHLER A growing body of literature refers to a transformation of state structures in OECD nations that consists of an expansion of non-majoritarian institutions, particularly independent administrative agencies. Among others, this thesis is advocated by Majone, who argues that the transition from the positive to the regulatory state generates a functional need to adopt a US-like agency model.1 Evidence can be detected at the EU level, but also at the national level. The argument is backed by current research about sectoral regulation whereby regulatory agencies are used as safeguards or initiators of competition in formerly state-supplied areas such as water, energy or telecommunication.2 No less supportive for the ‘agencification’ thesis is the ‘New Public Management’ discourse in which the separation between policies and operations is strongly recommended as a tool to enhance government efficiency.3 A closer look at this phenomenon reveals three dimensions. First, agencies, defined as administrative units separated from federal or central government departments, appear to gain more political relevance, due mainly to their regulatory functions. Second, there is an increase in numbers of newly created agencies, measurable in a number of OECD nations, and most vigorously so in the UK where the policy/operations split is almost completely implemented by hiving off so-called executive agencies from ministerial departments.4 Finally, either as a result of organisational decentralisation or their managerial functions, agencies are expected to become more independent from top-down political interference, thereby changing the command-control model of bureaucracy. Taken together, these dimensions lend political importance to the process of agencification. The analytical framework to deal with this phenomenon has mainly been institutional convergence. But if agencification is explored within the German context, some striking differences occur. Within the system of government institutions agencies only play a subordinate role. In the official language they are usually referred to as ‘non-ministerial federal administration’.5 Furthermore, the idea of constructing administrative agencies as non-majoritarian institutions, protected from political interference into single decisions, has not gained ground. Rather, bureaucratic autonomy is still seen as contrary to the principle of democratic legitimacy that requires elected politicians to retain control. A final peculiarity is the resistance to using regulatory agencies as a standard response to the re-regulation of deregulated industries. Mostly policy makers try to find a solution

Institutional choice and bureaucratic autonomy in Germany

91

within the existing framework, often via associational self-regulation. These observations are admittedly based on general indicators and the presumed deviation from the OECD mainstream may appear less pronounced if, for example, the cross-national equivalence of the dependent variable (that is, independent agencies) is defined differently. What sets the German case apart, however, is the combination of three factors that reduce the political role of administrative agencies. The first task of this discussion, then, is to explain their peculiar position in the German politico-administrative system and their lack of political independence. A second more theoretical aim is derived from the principal-agent (P-A) perspective that dominates research about delegation. Even if the P-A approach could be used as a framework, open to different hypotheses, most applications exhibit a strong magnetism for rationalist behavioural assumptions. P-A analyses are usually built around the problem of political control and typically include the following assumptions.6 Administrative agencies mainly emerge out of a ‘clash of interests’—all actors in this game, that is, legislators, presidents, interest groups and bureaucrats, pursue their own, often contradictory preferences, thus justifying the assumption that the form of delegated functions is based on highly rational and instrumental considerations; because these contradictions are built into the structure of an agency, a constant ‘fear of shirking’ leads politicians to produce regulations, rules, reporting requirements, and other measures for bureaucratic control; and this in turn causes an enduring ‘conflict over control’ fuelled by the willingness of all actors to gain a competitive advantage whenever possible. These assumptions are not necessarily wrong or irrelevant, but they privilege an analytical perspective that is biased to exclude motives other than rational-instrumental. By taking up this problem, the second aim is to provide evidence for the argument that the instrumental rationality of institutional design and control decisions should not be overestimated. This theoretical predisposition may appear somewhat outdated since modifications of some hard-core P-A assumptions have appeared over the past decade,7 mainly through the integration of institutional factors, accepting non-utilitarian motives, and empirical arguments about a political process lacking a functional order, beset with goal ambiguities and post-hoc rationalisations. Despite such analytical extensions, P-A analysts still have problems escaping from the rational choice world of ‘sophisticated economics and simple politics’.8 Two of the persistent objections against the functionalism implicit in rationalist theories, recently summarised by Pierson,9 also apply to mainstream P-A theories. First, actors are not necessarily instrumental in all their decisions, not only because they adhere to the logic of appropriateness, but also because the question of proper design is not raised. Second, policy makers often do not anticipate long or even medium term consequences of their choices. This is quite reasonable when considering the difficulties of assessing causal relations between policy goals and particular institutional arrangements.10 The explanatory concept applied here, therefore, emphasises the bounded rationality model and its derivates.11 More specifically, it is assumed that institutional choices are often based on no or low preference decisions. Hence, even the term ‘choice’ may be misleading because of its preference-based image. This is not to say that interests or goaloriented preferences are suspended or irrelevant; they do play a role in almost every political decision. However, this does not necessarily imply that policy makers are

The politics of delegation

92

concerned with the problem of administrative design. The complexity of the political problems needing to be addressed is only one reason to concentrate on substantial aspects instead. Other reasons are the availability of reliable and undisputed administrative models or a lack of party competition, both of which would prevent actors from exploiting organisational questions. Whether policy makers will act this way depends from the institutional setting within which they act, and this inevitably raises the question of how the institutional forces work. Among those scholars who have tried to elaborate the ‘institutions matter’ argument, Moe and Caldwell deserve special attention as their analysis is a rare example to link different modes of executive control with the structure of government in a systematic way. The initial assumption is that the institutional form of government, such as the separation of powers in the US or the parliamentary system in the UK, ‘programs a whole array of system features’12 by generating distinct sets of incentives. The position and performance of agencies is affected by the degree to which control is unified, the varying responsiveness of legislators to interest groups, the accruing need to formalise durable deals, and the degree of agency politicisation. Together this forms a ‘genetic code’ for bureaucratic structures and the way they are controlled.13 The following analysis will consider these analytical perspectives. OBSERVATIONS ON AN ‘AGENCY LAGGARD’ German observers like to stress the multiplicity of the units of the federal administration as reflected in variations in their size, function, legal status, and political or economic relevance. Currently, there are no less than 639 different non-ministerial public authorities (Behörden) at the federal level, ranging from rather obscure, hardly noticed units to huge and well-known agencies such as the Federal Environment Office or the Federal Cartel Office (FCO).14 According to a fine-grained legal framework, four basic types of federal authorities can be distinguished: Federal agencies (Bundesoberbehörden), self-governing bodies (öffentlich-rechtliche Körperschaften and Anstalten), public enterprises (Bundesunternehmen), and charged administrations (Bundesverrichtungsverwaltung). The first two types belong to the sphere of public law. The other two are private law organisations. Among these types, there is a clear hierarchy, allowing core government functions (hoheitliche Aufgaben) only to be delegated to federal agencies.15 This type of agency enforces major administrative functions at the federal level such as banking, insurance or drug control, and cartel law. Although the legal rules are rather vague, a recurrent pattern has emerged to delegate functions with decreasing political relevance to the other types, that is, touchy political functions are likely to go hand in hand with closer ministerial control. Typically social security funds, public banks or government research institutes are organised as selfgoverning bodies, whereas public enterprises and charged administrations mainly have entrepreneurial functions that are thought to perform better outside the restrictions of the public sector. If the following analysis is confined to the roughly 40 federal agencies, the reason is that only this type of authority performs functions, including regulation, that have the chance to enter the sphere of politics, thereby rendering the question of design and control more interesting.

Institutional choice and bureaucratic autonomy in Germany

93

What all federal authorities have in common is their subordination, in varying degrees, to some sort of ministerial oversight. Due to their private law status, public enterprises and administrations charged with missions are only subject to a control at arm’s length, mainly by ministerial bureaucrats acting as shareholders. In the case of self-governing bodies, oversight is restricted to legal aspects (Rechtsaufsicht). Federal agencies are additionally exposed to technical oversight (Fachaufsicht),16 a remarkably indeterminate piece of public law that allows ministerial departments to issue instructions on virtually every substantial aspect of agency activities, including the reversal of single decisions. There are a few exceptions to this rule, most notably the FCO and the Regulatory Agency for Telecommunication and Postal Services (RegTP), both of which are partly excluded from ministerial instructions. Cases like these have been dealt with in legal thinking as ministerialfreie Räume (area free from ministerial oversight).17 This legal doctrine, based on the idea of a continuous chain of democratic legitimacy and control, provides that an exemption from ministerial oversight should be limited to very special and well-founded cases. Although it is stressed that parliament is equally empowered to grant a fairly wide range of discretion to agencies, this is couched in a rule-exception scheme that clearly limits the appearance of autonomous authorities. This is well supported by general construction principles of public administration in Germany. At the federal level, the so-called Ressortprinzip (departmental principle) is essential for executive organisation. Article 65 of the constitution stipulates that, within the policy guidelines set by the chancellor and cabinet decisions, ‘each federal minister shall conduct the affairs of his department independently and in his own responsibility’. In practice, this rule creates a domain for each department that is often eagerly defended and not only applies to substantial policies but, almost to a greater extent, to the internal affairs of federal agencies. Ministers appoint agency heads, decide about their budgets and command a far-reaching power (Organisationsgewalt) to reorganise, downsize, merge, or abolish agencies within their competence (Zuständigkeit).18 If this requires rewriting the enabling law, the majority factions in parliament usually follow suit. Without any explicit reference, the departmental principal has given birth to the equally heavyweight normative principle of ministerial responsibility. As in most parliamentary democracies, this doctrine requires individual ministers to appear before parliament and, if asked, to report, answer questions, and to take the blame for omissions or failures within their competence.19 But, on top of this common obligation, ministerial responsibility has had quite unique repercussions on the role assigned to federal agencies. Even if important political functions are delegated to them, which is not unusual, they retain a subordinate status and are not given the chance to become ‘non-majoritarian’ in terms of escaping ministerial control. The federalist structure of the German polity is another important factor for defining the role of agencies. In fact, a substantial number of federal laws are not implemental by the federal goverment but instead are enacted by the Länder. If devolving public funtionsto decentralised administrative units in an aim of creating agencies, then the federal allocation of functions has certainly reduced these kind of pressures. However, allocating functions across different levels of government (central/local) should not be confused with the devolution of functions at one level, as is the case if the Länder administration is regarded as the functional equivalent to British executive agencies.20 In principle there are enough policies to be delegated to agencies at the federal level. What

The politics of delegation

94

seems more important about the argument of federalism is the legislative veto right that the Länder can use to block the expansion of federal administrative bodies with regional or local field offices. However, this veto can be, and is, bypassed if a new agency is organised solely at the federal level. This has severely reduced the ability of the Länder to restrict the creation of new federal agencies, especially during the post-war reconstruction era. The Länder continue to be concerned about the growth of federal competencies and functions, and in energy or rail regulation they oppose new regulatory agencies at the federal level. But if federal policy makers agree to create, for example, a new agency for consumer protection, it is difficult for the Länder to avoid such a decision. Thus, on the one hand, federalism has reduced the need to think about decentralising federal functions but, on the other hand, is not a severe restriction for creating new agencies at the federal level either. Aside from federalism, the frequent integration of interest groups into policy making is regarded as the trademark of German politics. Often this takes the shape of associational self-regulation replacing government interventions and therefore rendering new regulatory agencies superfluous. This argument is supported by examples such as health care, energy, or consumer protection. But two caveats need to be considered as well. The first is that self-regulation is embedded in a highly legalised political setting21 that throughout makes self-regulation and government regulation more complementary than mutually exclusive. Typically, agreements are checked for their conformity with government policy goals and their enforcement is supervised by public authorities or even translated into binding law. Agencies, for example in banking or insurance regulation, strongly rely on associational agreements instead of replacing them. Second, functional equivalence may also work in the reverse, as is the case with associational agreements in energy policy whose function is to set rules for a more competitive electricity and gas market. Not only the contents, but also the implementation of these agreements are supervised by the federal Department of Economics and the introduction of a regulatory agency in the energy sector has been used as a threat to force associations to conform with federal government’s policy goals.22 Since the introduction of the RegTP, such a threat is no longer improbable and receives more support as deregulation shows perceptible price reductions. This short overview leads to two conclusions. First, the legal-constitutional framework does not prescribe a single best way for administrative organisation. There are neither strong restrictions on the range of delegation nor constitutional doctrines that strictly prohibit autonomous agencies. There is considerable leeway for using agencies, including independent ones, as a tool for government policies. Second, although federalism and associational self-regulation reduce the functional need for regulatory institutions, they do not represent serious restrictions either. EX-ANTE DESIGN BETWEEN INSTITUTIONAL CHOICE AND INSTITUTIONAL DYNAMICS In rationalist accounts institutions result from deliberate choices made to achieve substantial policy goals, to reduce transaction costs, or to overcome collective decision making problems. It is therefore hard to imagine that setting up a new agency is not

Institutional choice and bureaucratic autonomy in Germany

95

accompanied by all those ‘deck-stacking’ elements23 usually considered as the bread and butter of politics in P-A analyses. However, an alternative way to conceptualise the process of agency creation is to assume that institutional dynamics, or ‘creeping institutional change’,24 prevail over conscious and goal-oriented institutional choices. Following this idea, result-based institutional reform could be expected to decrease if existing government institutions have the capacity to relieve policy makers from engaging in time-consuming and costly considerations about alternative structures, control problems and the consequences of agency design. Presumably, this will be the case if policy makers have no incentives to put these problems at the very centre of the legislative struggle, mainly because institutional or other factors provide reliable standard solutions to agency design and control. The argument, illustrated in the following cases, is that the German political system contains several attributes that allow institutional dynamics to prevail over institutional choice. The 1950s were not only the decade of economic reconstruction, but also the heyday of agency creation. No less than 19 Bundesoberbehörden were set up during this period. Among them was the FCO, founded in 1957 after a long and bitterly fought legislative battle known as the ‘seven years war’.25 After initial hostility from German industry calmed down, the FCO became almost celebrated, not only due to its tough anti-trust enforcement, but also for its independent status. How can this be reconciled with the picture presented above? Part of the explanation is to be found in the vague rules establishing the agency’s independence.26 The law only mentions that general instructions must be published in the federal register (§ 49 GWB). One controversial interpretation, among others supported by FCO officials, is that this prohibits special instructions (Einzelweisungen). But this could be, and is, also understood the other way around, that is, there is no obligation to publish special instructions. In the overseeing Department of Economics the view is still held that the FCO’s independence is only a customary right (Gewohnheitsrecht) without any formal guarantee.27 But even these concessions seem misplaced in the midst of an otherwise hierarchydominated administrative organisation. Due to pressures on German government by the American occupational forces to implement anti-trust legislation, the FCO’s status was often regarded as conforming to the US tradition of independent agencies. However, a closer look reveals that American influence evaporated over the years. The exemption of the divisions from ministerial instructions was justified in parliament with reference to the Federal Patent Office and the Federal Insurance Oversight Agency,28 both of which already embodied ‘court-like’29 decision making bodies. But these historical legacies are only part of the explanation. By looking at the broader context it seems more likely that the FCO as well as the Bundesbank, equipped with an even more pronounced degree of independence, emerged in a situation in which granting autonomy to federal agencies was still an open question. At this point, the constitutional doctrine of ministerial responsibility had not yet solidified in practice, nor had other delegation-hostile concepts become politically effective. With the gradual emergence of this doctrine and its derivates, both Bundesbank and Bundeskartellamt were perceived as exceptions.30 This not only prevented a tradition of mimetic isomorphism among other federal agencies, the notion of singularity even made it redundant to justify why independence was not granted to other agencies. As German policy makers experienced their hands being tied whenever they were exposed to uncomfortable decisions by non-majoritarian institutions, including

The politics of delegation

96

the Constitutional Court, a silent consensus emerged not to proliferate this type of Nebenregierung (supplementary government).31 The irony of both the FCO and, to a greater degree, the Bundesbank is that they became widely recognised for being independent from political intrusions but at the same time contributed to an institutional dynamic in which subordinating agencies to ministerial oversight was perceived as a democracy enhancing constitutional standard. This process did not surface as a series of conscious decisions. The ‘rule’ to subordinate federal agencies has remained more an implicit preference than a clear-cut regulation. Students of public administration have consistently observed low interest by politicians for administrative design and reform.32 As is often the case if political preferences are diffuse or non-existent,33 the ministerial bureaucracy becomes the main actor, and this frequently happens within the domain of the Rechtsformenwahl (choice of legal status). However, there is no indication that ministerial choices have gone beyond the application of routine legal models. In those few cases in which this process is mentioned at all, the focus is on privatisation, that is, the transition from public to private legal status.34 Questions related to agency design or controls are largely neglected. Certainly, the enactment of enabling laws is by no means free from political disputes. But the focus is clearly on the range of functions delegated to an agency rather than on the range of discretion or administrative autonomy. A more recent case to exhibit exactly this pattern is financial market regulation.35 Since the 1980s, German banks, insurance companies, and large industrial firms felt the increasing need to move the cartelised and self-regulation based German model of capital market regulation closer towards US-like standards. Driven by anxieties over loss of attractiveness for foreign investors, especially in the stock and securities market, the financial community successfully lobbied for the introduction of the Bundesaufsichtsamt für den Wertpapierhandel (BAW) in 1994. Despite the US Securities and Exchange Commission, an independent commission by definition, figuring prominently in the debate preceding the enabling law, no thoughts were given to the question of political control. The BAW was almost automatically constructed as a federal agency subject to ministerial oversight. If there is a credible commitment to protecting a particular policy it is situated on the abstract level of having sectoral government regulation, leaving open the backdoor for the federal government to intervene in agency policies. The efficiency of the BAW to regulate stock exchange transactions, such as preventing insider trading, certainly depends more on the law enforcement capacity than on the degree of independence. But the straightforward attitude to adopt a particular legal form reflects a persistent pattern of relying on a tried and tested agency model. This underlines that the German way is not to have no delegation at all, but to restrict bureaucratic autonomy almost automatically by subordinating agencies to ministerial oversight. Drug regulation is one of the few occasions on which controls on agency decisions have been thoroughly discussed during the legislative process.36 The law, enacted in 1976, contained an unusually large number of measures to empower countervailing interests. A system of commissions, staffed with external experts including representatives from the pharmaceutical industry, plays an important role in the drug admission procedure. A reporting requirement to parliament was included for the first time, and post-market drug surveillance was arranged as a co-operative, somewhat ponderous procedure, again embracing a broad range of external experts. Certainly, this

Institutional choice and bureaucratic autonomy in Germany

97

piece of legislation is no example for an ‘auto-pilot’ pattern of agency design. It shows that German policy makers may well consider frontloaded measures to temper agency activism, thus conforming to rational-instrumentalism. On the other hand, drug regulation also reflects interest accommodation through expert committees or co-operative arrangements as a typical pattern of the German administrative rule system.37 The most puzzling case in recent years was the introduction of the RegTP. Starting its operation in 1998, the agency was immediately received as a major innovation in at least three aspects. First, the agency adopted a highly visible, almost political role that triggered an unusual amount of attention and conflict. Second, it is regarded as the only ‘real’ regulatory agency, representing the arrival of the regulatory state in Germany.38 Third, it is the first federal agency ever since the FCO back in 1957 to receive a limited exemption from ministerial oversight. These peculiarities reflect the whole field of telecom regulation being embedded in a deregulation-globalisation discourse and such large-scale political processes certainly have the power to overthrow national standard operating procedures. The creation of the RegTP has not swept away the model of a hierarchical administrative structure, but has demonstrated that pressures, primarily from the EU, to conform to a new regulatory regime exist and should be taken into account when considering the future of the German model. During the legislative deliberation preceding the 1996 telecommunication act, the problem of agency monitoring and independence was discussed at some length.39 The unusual willingness of policy makers to address this subject and to enact a limited exemption from ministerial instructions shows a clear deviation from institutional dynamics. This is underlined by the original plan to set up the RegTP as the highest federal authority (oberste Bundesbehörde), enjoying an independent status equal to the Bundesbank. But lawyers soon reminded policy makers about the logic of appropriateness,40 thus redirecting attention towards imitating the FCO model rather than inventing a completely new kind of agency.41 This indicates that the autopilot of German institutional dynamics was not completely disengaged. The internal structure of the RegTP owes much to the FCO. Six collegial decision chambers were set up for some important agency functions such as licensing and price regulation. Analogous to the FCO divisions, these chambers are widely regarded as being free from ministerial instructions. But this is only inferred from their court-like image and nowhere guaranteed in the law. The Telecommunications Act (TKG) only stipulates that general instructions by the responsible Department of Economics have to be published in the federal register (§ 66 TKG). This has prompted a sophisticated and controversial legal debate, the result of which is that ministerial instructions should be issued very carefully. Thus, the RegTP’s much proclaimed independence is shrouded in vagueness, as in the case of the FCO. It remains vulnerable to instructions as long the federal government remains a major shareholder of Deutsche Telekom AG, and the labour market problems of this and other former monopolists, caused by privatisation, are relevant for policy makers. In practice, the overseeing department has issued several general and special instructions.42 Again, if agency design was meant to offer credible commitment, it went only halfway. The conclusion to be derived from this section is not that conscious choices about agency design are non-existent, but that they are embedded in an institutional dynamic allowing policy makers a low preference stance with respect to monitoring and control

The politics of delegation

98

measures. The underlying rationale could be described as ‘design without grand decisions’.43 As single policy-related cases will always show more or less pronounced deviations from this standard model, their importance should not be overestimated, as is the case with the RegTP. If we take a P-A point of view, yet an other question occurs. German policy makers apparently reduce the transaction costs of agency creation by treating the problem of monitoring and control as already solved through ministerial oversight. The question, then, is whether there are any incentives to work on changing this? Answering this question requires an analysis of the incentives generated by the broader institutional setting of government. THE INSTITUTIONAL ORDER OF GOVERNMENT Following Moe and Caldwell, the identification of institutional genetics should start with the relationship between executive and legislative powers. In this respect, the two models of presidential and cabinet/parliamentary government represent the most basic institutional configurations.44 In a presidential system of government, such as the US, the separation of power charges two equally legitimised institutions with controlling the executive. This ‘parallel’ system not only hampers a unified control but also triggers continuing conflicts between both constitutional powers.45 Furthermore, legislators in Congress have strong incentives to react to group pressures as they depend on their support for re-election. As a result, agencies are exposed to a broad and sophisticated repertoire of political measures, all directed at maximising the gains or minimising the losses of politicians and clienteles. Making an agency independent is equally used either as a shield against competing congressional and presidential influence or to expose agencies to countervailing interests. To sum up, the American political system places federal agencies and commissions under a multiple-principal regime that generates strong incentives to make use of an extensive repertoire of supervisory measures. No such running strategic battles exist in the German parliamentary system. As opposed to the US case, German government is based on a ‘sequential’ order of institutions, that is, parliament, cabinet, and the executive are arranged consecutively, thus allowing a unified and usually hierarchical administrative control. With the federal government being elected and supported by the majority parties in parliament, the likelihood of conflicting interests between majority factions and cabinet is extremely low. Furthermore, the responsiveness of legislators to demands of interest groups is boiled down through strong discipline of parliamentary factions. Since political parties command the decisive resources for re-election of individual MPs, they are less inclined to follow group pressures and more to subordinate to a fairly coherent party line. Administrative agencies are thereby situated in a clearly defined environment. Overseeing competencies rests with the responsible minister, thus reducing the incentives for legislators to be interested in such an ‘executive-centred’ administration. In contrast to the US system where executive and ‘legislative-centred’46 agencies exist at the same time, their German counterparts are only confronted with a single principal. Consequently, neither parliamentary sponsorship of agencies nor micromanagement aspirations have appeared. Certainly, parliamentary controls of executive action are regarded as an important constitutional function, just as in most other democracies. But

Institutional choice and bureaucratic autonomy in Germany

99

by all measures, legislative controls over the executive in Germany are not designed to exercise a perceptible influence over the federal bureaucracy.47 This brings ministerial responsibility back in. Implicit to the concept is a transfer of agency monitoring from parliament to individual cabinet ministers. Parliamentary controls are seen as sufficient even if they remain ‘sporadic’ and ‘coincidental’.48 Their function is more to offer a threat than actual enforcement. Part of the German version of ministerial responsibility is its direct linkage to administrative organisation. Underlying is the idea that individual political responsibility can be invoked only as long and as far as ministers possess influence over their portfolios. The best guarantee to achieve this goal is the unlimited ministerial right to issue instructions. This, in turn, requires a hierarchical model of organisation and processes.49 It is this very connection with democratic legitimacy that makes hierarchy an almost indispensable element of administrative structure. During day-to-day work it is not expressed in a rigid command-and-control fashion. The hierarchy principle, routinely referred to in law and administrative science textbooks, is even increasingly hidden under the language of a modernisation discourse that stresses co-operation and decentralised responsibility. But in case of conflicts or diverging opinions, hierarchy serves as a ‘rule of last resort’,50 covering other forms of interaction. It is definitely among the elements constituting the genetic code of the German administrative landscape. Ministerial responsibility, however, is not a sufficient explanation for low political conflict on agency operations. Since parliament is divided between majority and opposition parties, the latter might be expected to exercise a more vigorous control. Of course, the typical game is that majority parties practice solidarity with cabinet, whereas opposition parties critically assess government policies whenever possible. However, the focus of partisan disputes is more on individual ministers’ ability to head a department and does not concern agency policies. It makes more sense for opposing parties to aim at cabinet and not to waste energy on agencies which, at any rate, are appropriately thought of as a ministerial appendage rather than as political actors in their own right.51 This also makes interest group pressures for changes in agency policies more effective if they are directed at the incumbent minister who is equipped with oversight authority, and not at the parliamentary opposition that can, at best, try to mobilise the public. The most frequent technique used by groups or industries to influence an agency’s operation is through administrative courts.52 By doing so these actors contribute to the transformation of unsolved political conflicts into legal problems. Of course, judicial activism may backfire and force legislators to react. But, as is illustrated by the long-lasting atomic energy conflict, the legalisation strategy also keeps conflicts pending without forcing painful political decisions.53 Aside from incentives generated by the institutional order of government, an important reason for policy makers not to be concerned with agency design and control is the performance of bureaucracy. If it is based on ‘neutral competence’,54 that is, agencies could be expected to serve the party in power as well as to enforce law in a neutral, public interest-oriented way, the likelihood increases that opposing parties agree to keep agencies out of politics because they want to benefit equally from an efficient bureaucracy if they assume power. This comes close to the situation in Germany, but is somewhat more complex due to the contradictions concerning the role of bureaucracy.

The politics of delegation

100

Even though the traditional separation between politics and administration is no longer explicitly advocated, there is still a tacit preference for a de-politicised administration, especially if connected with the implementation process. This is based on the following factors. First, a neutral civil service is fixed in several laws and it is taken for granted that no additional safeguards are needed on top of existing regulations to ensure that administrators will respect the political will of parliament. A second point is that party politics are expected to end if a law has passed parliament,55 largely because administration is seen as law-enforcement and not as a continuation of policy making outside parliament. Third, there is a great lack of feedback from the implementation process to parliament,56 allowing legislators to be satisfied with passing the law and not being forced into laborious evaluation activities. Even if a de-politicised administration and a neutral process of law implementation are fictitious, there are no pressures for policy makers to move away from these comfortable ideas. Unsurprisingly, this system is not free from incoherence. Despite the fact that ministerial responsibility is recognised without reservations, public perceptions, including those of opposition parties, of course, are at odds with its implementation. If instructions are made public, which does not happen very often, negative responses prevail as ministers are suspected of serving partisan or clientelist purposes. Ministerial instructions are more generally thought of as a tool for correcting bureaucratic non-compliance rather than providing policy direction beyond what is already fixed in law.57 This has forced incumbent ministers as well as ministerial officials to take an almost hypocritical stance. Whereas they have no problem in emphasising their hierarchical position vis-à-vis agencies in general, they dislike admitting that single agency decisions have been overturned or reversed by ministerial instructions. Facing the political costs of public punishment,58 ministers and their bureaucrats are inclined to treat all agency matters, including instructions, as an internal administrative affair.59 The underlying rationale is that the legitimacy of a hierarchical mode of governance requires that agency activities are administrative in nature, thus avoiding any suspicion that ministerial instructions may collide with, or even reverse, parliamentary decisions that have defined agency tasks in the first place. EX-POST CONTROLS AND THE MALLEABLE MEANING OF BUREAUCRATIC AUTONOMY Usually, P-A analyses start with the assumption that delegation almost by definition is creating a control problem.60 As the previous sections have shown, to a considerable extent German politicians are released from conflicts over control and fear of shirking. Consequently, the problem of ex-post controls may appear less relevant. But reconsidering the argument that the advantage of independent agencies is the use of pure expertise for addressing complex problems and to generate credible policy commitments and continuity,61 it might at least be asked how German policy makers cope with a system of agencies that apparently is not well equipped to harvest the fruit of independence. To begin with, German politicians would probably disagree with the account that agencies are not independent. There are even typical statements, supposing that a certain

Institutional choice and bureaucratic autonomy in Germany

101

agency is professionally independent (fachlich unabhängig) and the minister decides only questions of political relevance. But the arbitrary element in this distinction, which obviously serves the purpose of dispensing with the need for further outside scrutiny, is that it is within the supervising department’s competence to decide whether a case is still professional in nature or already political. The resulting leeway for oversight is enormous. It even allows ministerial departments to impose their political rationality on scientific judgements. In such case, agency expertise is not challenged outright as being wrong; rather, ‘scientists are expected to consider economic and political factors in the course of producing a reasoned, scientific judgment’.62 Thus, in the German perception, expertise does not hinge on autonomy but can be achieved whilst taking account of political obligations. The somewhat indeterminate meaning of agency independence is not limited to Germany. Already a modest comparative outlook reveals that bureaucratic autonomy is not as sacred as it may appear at first sight. A recent British survey has shown, for example, that general managers constantly struggle with ‘political interference on operational management decisions’.63 Swedish central agencies, which are typically described as being independent from their political principals, also display a different picture. Larsson describes the Swedish system ‘as one in which ministers have as much influence over the bureaucracy as in other countries, but without the corresponding responsibility’.64 Finally, it is well known that the independence of US agencies and commissions is not based on isolation from political influence, but is rather attained through a precarious balance between Congress, interest groups, and the president, all struggling for influence over agency policies. Political independence may be achieved if no single actor has more influence than his competitors,65 but this is not necessarily the case. On the one hand, these observations call for a more thorough determination of the concept of agency autonomy. The important point, however, is that independence is obviously framed in different mindsets in such a way that British, Swedish, and American politicians accept political independence as an instrument to achieve policy goals. In contrast, the idea that administrators should make policies is rejected in German normative thinking. One of the practical consequences is that overseeing departments try to suppress obvious policy ambitions from below. This is often done by denying that agency decisions are political and insisting that they are only technical or legal in character. Majone has argued that there is a functional need to have more non-majoritarian institutions, because of pressures to conform to international regulatory regimes and the domestic problem of overcoming ‘short-termism’.66 The abstract solution to both problems is to delegate policies to institutions that are detached from electoral cycles and are trustworthy because they have nothing to gain if they give in to partisan or clientelist demands. As to the credibility problem, Majone probably overestimates the ensuing need for non-majoritarian institutions. The cases of the BAW and RegTP illustrated that the German style of handling pressures is to pretend that the institutionalisation of a public function is demonstrating sufficient honest effort to enforce a regulatory policy. Given their complexity, the credibility of this strategy is only damaged if ministerial micromanagement reverses the overall goal to a significant extent. Credibility is probably a less ambitious aim of government policies and is accomplished on a lower level as well.

The politics of delegation

102

Second, non-majoritarian institutions are not the only way to achieve policy continuity. In Germany, continuity already results from a system of dispersed government powers, complemented by a culture of negotiated policies. This allows the relevant economic and political interests to be considered in the political as well as in the administrative decision making process.67 The persistence of this ‘policy of the middle way’68 is especially visible after partisan changes of government, which have not yet been accompanied by radical policy changes. With continuity being a general characteristic of German policy making, it is less urgent to protect agencies against the U-turn aspirations of subsequent governments. In a way this is reflected in the patterns of ministerial oversight that show no signs of fixing policies for the long term. Administrative guidelines and other measures to cope with agency policies are dealt with in an almost secret fashion and therefore are not subject to public scrutiny or open political deliberations. Furthermore, ministerial oversight is not statutorily defined in terms of intensity, regularity, or instruments. Not least due to the departmental principle, it is up to each department to decide how to deal with an agency. Frequently used monitoring techniques are reporting requirements, prechecking press releases or reserved agreements in relation to more important decisions. Agency officials in turn know which questions could bother the overseeing department and protect their position by providing first-hand information to the department. It must be added, however, that the day-to-day work of most agencies is not greatly subject to interference by ministerial instructions. Agencies may be left alone for long periods, either because there is no need for closer inspection, no interest, or no personal capacity. On balance, the oversight behaviour of German ministerial bureaucrats more closely resembles the ‘fire alarm’ than the ‘police patrol’69 approach. But instructions remain an option and agency officials anticipate this. A by-product is that strong agency activism has remained a rare exception. Delegation is often referred to as an act of granting broad powers to administrative agents. A peculiarity of the German understanding is to think of delegation from the opposite direction. According to the German constitution, ‘the content, purpose and scope’ (Art. 80 para. 1 GG) of the range of delegated power must essentially be written into a parliamentary law. Pre-defining the ‘essentials’ of administrative action requires very detailed laws that tend to reduce administrative discretion. And, in fact, there is some evidence that German administration is equipped with less discretion than is found in France, Britain or the US.70 At the same time, however, increasing complexity of legislative goals has caused parliamentary laws to include infinite law terms such as ‘public safety’, ‘unreasonableness’, or ‘urgency’.71 In cartel law even the central concept of ‘competition’ is left open to definition.72 More recently, the relevance of ‘informal or ‘cooperative administrative action’ was stressed as an emerging procedural standard that opens greater leeway to agencies. The impression of German administrators as lagging behind the discretionary powers in other countries may therefore be overdrawn. But this is not the problem. Regardless of deviations in practice, the constitutional doctrine of delegating only plainly circumscribed functions to administrative agents leads, if even implicitly, to a view of administrative agencies as non-political actors and restricted to applying the law. This also explains why German policy makers do not perceive administrative discretion as a severe problem. As compared to US legislators, who deliberately employ a wide range of constraint categories, including spending limits,

Institutional choice and bureaucratic autonomy in Germany

103

legislative veto, rule making requirements or exemptions as instruments of political control,73 most of these measures, in theory at the disposal of German legislators as well, are usually not considered in the law making process, because it is assumed that the average law contains sufficiently detailed guidance for administrators. A consequence of this somewhat circular argument is that legislators are not forced to behave like typical principals who are obsessed with the problem of misuse of administrative discretion. PRESSURES FOR CHANGE The recent debate on delegation and non-majoritarian institutions has largely been caused by the increasing importance of such institutions for the traditional model of representative democracy. As this process continues, it would be unrealistic to expect Germany to be left completely unaffected. Pressures for change indeed emerge from different sources. The process of European integration is a major source of pressure. Majone puts forward the most explicit arguments in this respect.74 They are based on the assumption that EU policies contribute to the transformation of the positive state, which was characterised by nationalised industries, redistributive policies, and taxing and spending as main instruments, into a regulatory state, which concentrates on rule making and generating competition. This new mode of governance is said to replace the old, rather centralised, command-control-based model of government administration by a more decentralised variety of commissions and agencies. Even if the overall story seems persuasive, at least from a German perspective, the assumed pressure for national adoption appears more diffuse and less powerful for institutional reform. Grande and Eberlein also put forward the argument of an emerging regulatory state, but are more cautious as to the ensuing administrative ‘architecture’75 that may take different shapes depending on the regulated sector. The often remarked strong sectoralisation of the German political economy, which is reflected in the views of actors who like to emphasise the peculiarities of their policy field, supports the conclusion that inter-sectoral spill-overs of the agency model (as seen in the case of the UK) are inhibited by divergent modes of governance.76 The adaptive pressure of EU deregulation policies are therefore not only channelled and potentially absorbed through the genetics of national politicoadministrative systems, but also meet policy sectors with variable adaptive rationalities. Quite a similar conclusion is drawn by Börzel and Risse, who argue that there must a perceptible ‘misfit’77 between European and national policies and institutions until adjustment pressures can succeed. This is not to say, however, that European policies still lack an impact. A less visible but nonetheless important change resulting from deregulation is related to the mandate of agencies. Since the early 1990s the traditional agency function of state oversight (Staatsaufsicht) with its hierarchical connotations and anti-competitive bias is slowly being replaced by a less industry-sponsoring but more consumer-oriented mandate, increasingly aimed at facilitating sectoral competition. So far Germany has resisted a visible agencification as a response to EU policies. But it is likely that EU liberalisation strengthens the public awareness and support for reregulation and thereby makes the use of agencies more acceptable in policy domains such as energy or food safety.

The politics of delegation

104

A second source of change is the recent NPM discourse in which agencification of federal or central government executive structures is explicitly advocated as a tool to increase public sector performance.78 Underlying this discourse is an entrepreneurial model of agency behaviour that advocates the splitting of operational functions from policy functions, the latter remaining in conventional ministerial departments. In the German version of NPM, the focus is on decentralising responsibility and contract management,79 and its main success has so far been at the level of agenda setting and prompting activities for internal reorganisation, modernising civil service law and personal management, use of information technologies, improving team work and cooperation in federal authorities, and so on. This makes the federal government look like a radical moderniser, however, reform activities are still circumventing basic institutional features, especially the instrument of agency independence. An assessment of the impact of NPM on the traditional hierarchical model has to consider that the mixture of organisational decentralisation, performance controls, contracts, and operational independence makes it is hard to measure non-compliance with the NPM prescription. It would be no problem either to follow the NPM agenda by decentralising or granting more operational autonomy to a number of agencies with unimportant service functions, such as civil and military procurement, foreign language translation or maritime affairs, while retaining hierarchical control over the political relevant ones. Aside from such an adoption à la carte, the question remains whether and to what extent a core area of federal administration possesses enough institutional inertia to resist an ongoing modernisation process. The pressure for change will presumably not be exercised through direct demands for institutional reform. Considering the opaque conditions under which federal agencies operate, NPM discourse could instead shed light on outdated modes of hierarchical control and activate a re-examination of the benefits of bureaucratic autonomy. Currently, however, neither incumbent politicians nor ministerial bureaucrats seem to be willing to accept changes.80 Despite such reservations against the institution-transforming power of EU-driven deregulation and the modernisation discourse, it has to be acknowledged that the pressures for change, if only incremental, are nagging at the traditional administrative model. An outline of likely changes for the German case follows in the concluding section. CONCLUSIONS This discussion has used the notion of institutional genetics to explain the subordinated status of federal administrative agencies in Germany. Four elements shape the genetic code: (1) the conflict reducing executive centrism of agencies; (2) the unchallenged status of ministerial responsibility that makes (3) hierarchy an important construction principle for administrative organisation; and (4) the idea of a neutral, de-politicised civil service that allows policy makers to trust in uniform implementation standards. The extent to which this code determines ex-ante agency design and ex-post controls depends on whether institutional dynamics or institutional choice dominates. In the German case, an institutional dynamic, prevailing over long periods, has routinely ‘auto-piloted’ those parts of political decisions that involved the design and control of agencies. Typical P-A considerations such as fear of ‘shirking’ or conflict over control are thereby mostly

Institutional choice and bureaucratic autonomy in Germany

105

absorbed and relax the necessity for policy makers to consider alternative options. The result is a low preference posture in relation to agency design and control. These questions are clearly more regarded in an instrumental and choice-based fashion in political systems such as the US or a polity in the making such as the EU, in which the separation of constitutional powers creates multiple P-A relationships. As opposed to the sequential structure of the German political system, such variability opens the window to pursue policy goals via agency design and control. As we know from genetics, evolution is difficult to suppress in the long run—even stable institutional settings may be forced into evolutionary adjustment. With regard to the three characteristics of the German agency landscape, it is possible to make a ranking of expected persistence. The least protected attribute is the resistance to create new agencies with regulatory powers. Pressures either from the EU or domestic sources are likely to push forward this instrument more firmly in the immediate future. Somewhat more durability can be expected of the political role of agencies. Due to their executivecentred position, there are few incentives for political actors to giving them a greater political role, although the functions delegated to them might be important. Additionally, on the normative dimension, agency activism is still contrary to the preferred division of labour between politics and administration. Finally, bureaucratic autonomy is least likely to succeed over the traditional model of hierarchical governance for similar reasons. Neither pressures to make credible commitments nor efforts to secure policy continuity can be translated into a need for more non-majoritarian institutions. These characteristics of the German case are not evident at first sight since observers usually turn their attention to more obviously expressed interests of political actors. But previous findings suggest that non-instrumental, unreflected or low preference behaviours are equally relevant as goal-oriented and rationalist forms of action and therefore deserve a similar degree of analytic attention. NOTES The author would like to thank the editors of this volume, Dominik Böllhoff, Jochen Franzke, Werner Jann and Kai Wegrich for their helpful comments on a first version of this chapter. 1. Giandomenico Majone, ‘The Rise of the Regulatory State in Europe’, West European Politics 17 (1994), pp. 77–101; idem (ed.), Regulating Europe (London: Routledge 1996); idem, ‘From the Positive to the Regulatory State: Causes and Consequences of Changes in the Mode of Governance’, Journal of Public Policy 17 (1997), pp. 139–67. 2. Mark Thatcher, ‘Institutions, Regulation and Change: New Regulatory Agencies in British Privatised Utilities’, West European Politics 21 (1998), pp. 120–47; Robert Baldwin and Martin Cave, Understanding Regulation: Theory, Strategy, and Practice (Oxford: Oxford University Press 1999); Dominik Böllhoff, ‘The New Regulatory Regime—The Institutional Design of Telecommunications Regulation at the National Level’, Paper for the Conference on Common Goods and Governance Across Multiple Arenas, Max-Planck Project Group on the Law of Common Goods. Bonn, 30 June–1 July 2000. 3. OECD, The OECD Report on Regulatory Reform (Paris: OECD 1997); Frieder Naschold, Werner Jann and Christoph Reichard, Innovation, Effektivität, Nachhaltigkeit. Internationale Erfahrungen zentralstaatlicher Verwaltungsreform (Berlin: edition sigma 1999); Christopher Pollitt and Geert Bouckert, Public Management Reform. A Comparative Analysis (Oxford:

The politics of delegation

106

Oxford University Press 2000). More recently, the OECD has been actively promoting independent agencies through its PUMA initiative: ‘Governance of State Agencies and Authorities’, Focus 18 (2000), p. 6. 4. In the British case, see Francesca Gains, ‘Implementing Privatization Policies in “Next Steps” Agencies’, Public Administration 77 (1999), pp. 713–30; some comparative data are presented in Colin Talbot et al., ‘The Idea of Agency. Researching the Agencification of the (Public Service) World’, Paper for the American Political Studies Association Conference, Washington DC, August 2000. 5. This is the title of an official compilation of federal administrative units. Cf. Bundesminister des Inneren, Die nichtministerielle Bundesverwaltung (loose-leaf edition, 2000). 6. This sketchy summary is based on: Mathew D.McCubbins, Roger G.Noll and Barry R. Weingast, ‘Administrative Procedures as Instruments of Political Control’, Journal of Law, Economics and Organization 3 (1987), pp. 243–77; Terry M.Moe, ‘The Politics of Structural Choice: Toward a Theory of Public Bureaucracy’, in Oliver E.Williamson (ed.), Organization Theory from Chester Barnard to the Present and Beyond (New York: Oxford UP 1990); Murray J.Horn, The Political Economy of Public Administration. Institutional Choice in the Public Sector (Cambridge, MA: Cambridge University Press 1995); David Epstein and Sharyn O’Halloran, Delegating Powers: A Transaction Cost Politics Approach to Policy Making under Separate Powers (Cambridge: Cambridge University Press 1999). 7. Cf. Moe, ‘The Politics of Structural Choice’; several contributions in Kristen Renwick Monroe (ed.), The Economic Approach to Politics (New York: HarperCollins 1991); David B.Spense, ‘Administrative Law and Agency Policy-Making: Rethinking the Positive Theory of Political Control’, Yale Journal on Regulation 14 (1997), pp. 407–50. 8. James E.Alt and Kenneth A.Shepsle, ‘Rules, Restrictions, Constraints: Structure and Process in the New Institutional Economics’, Journal of Institutional and Theoretical Economics 154 (1998), p. 740. 9. Paul Pierson, ‘The Limits of Design: Explaining Institutional Origin and Change’, Governance 13 (2000), p. 477. 10. Johan P.Olsen, ‘Modernization Programs in Perspective: Institutional Analysis of Organizational Change’, Governance 4 (1991), p. 127. 11. James G.March, A Primer on Decision Making. How Decisions Happen (New York: Free Press 1994). 12. Terry M.Moe and Michael Caldwell, ‘The Institutional Foundations of Democratic Government: A Comparison of Presidential and Parliamentary Systems’, Journal of Institutional and Theoretical Economics 150 (1994), p. 172. 13. Ibid., p. 192. 14. Bundesministerium des Inneren, Moderner Staat—Moderne Verwaltung. Zwischenbilanz— Chancen und Veränderungen (Berlin 2000), p. 16. Still the most comprehensive overview is Roman Loeser, Die Bundesverwaltung in der Bundesrepublik Deutschland—Bestand, Rechtsformen und System der Aufbauorganisation. Vol. I (Speyer: Speyerer Forschungsberichte 50, 1987). 15. Klaus Stern, Das Staatsrecht der Bundesrepublik Deutschland. Vol. II (München: Beck Verlag 1980), p. 817. 16. Roman Loeser, ‘Neubestimmung der Steuerungsingrenzen der Ministerialinstanz gegenüber Bundesoberbehörden?’, Verwaltung und Fortbildung 23 (1995), pp. 275–91. Some empirical findings are summarised in Marian Döhler, ‘Das Modell der unabhängigen Regulierungsbehörde im Kontext des deutschen Regierungs- und Verwaltungssystems’, Die Verwaltung 34 (2001), p. 74. 17. Carl-Peter Fichtmüller, ‘Zulässigkeit ministerialfreien Raumes in der Bundesverwaltung’, Archiv des öffentlichen Rechts 91 (1966), pp. 297–355.

Institutional choice and bureaucratic autonomy in Germany

107

18. Cf. Janbernd Oebbecke, Weisungs- und unterrichtungsfreie Räume in der Verwaltung (Köln: Kohlhammer 1986), p. 36; Dodo Traumann, Die Organisationsgewalt im Bereich der bundeseigenen Verwaltung (Baden-Baden: Nomos 1998). 19. Klaus Kröger, Die Ministerverantwortlichkeit in der Verfassungsordnung der Bundesrepublik Deutschland (Frankfurt/M.: Athenäum 1972). 20. Helmut Klages and Elke Löffler, ‘New Public Management in Germany: The Implementation Process of the New Steering Model’, International Review of Administrative Sciences 64 (1998), p. 46. 21. Cf. Michael Kloepfer and Thomas Elsner, ‘Selbstregulierung im Umwelt- und Technikrecht’, Deutsches Verwaltungsblatt (1996), pp. 964–75; Ian Bartle and Markus M.Müller, The Administration of Regulation: Self Regulation State Agencies (University of Exeter Discussion Paper No. 3, 2000), p. 14. 22. Süddeutsche Zeitung, 1 Feb. 2001, p. 25. On the background cf. Lars Kumkar, Zur institutionellen Ausgestaltung der Strommarktregulierung: brauchen wir eine eigenständige Regulierungsbehörde für den Stromtransport? (Kiel: Institut für Weltwirtschaft, Kieler Diskussionsbeiträge 371, 2000). 23. Cf. McCubbins et al., ‘Administrative Procedures as Instruments of Political Control’, p. 261. 24. Roland Czada and Uwe Schimank, ‘Institutionendynamiken und politische Institutionengestaltung: Die zwei Gesichter sozialer Ordnungsbildung’, in Raymund Werle and Uwe Schimank (eds.), Gesellschaftliche Komplexität und kollektive Handlungsfähigkeit (Frankfurt/M.; New York: Campus 2000), p. 30. 25. Knut W.Nörr, Die Republik der Wirtschaft: Recht, Wirtschaft und Staat in der Geschichte Westdeutschlands. Vol I: Von der Besatzungszeit bis zur Großen Koalition (Tübingen: Mohr Siebeck 1999), p. 157. 26. Jürgen Rodrega, Zum Problem aufsichtsfreier Verwaltung durch das Bundeskartellamt: untersucht am Beispiel des Fusionskontrollverfahrens (Frankfurt/M.: Lang 1992), pp. 33–40; Wernhard Möschel, ‘Die Unabhängigkeit des Bundeskartellamtes’, Ordo 48 (1997), pp. 241–51. 27. Handelsblatt, 23 July 1996, p. 40. 28. Knut W.Nörr, Die Leiden des Privatrechts: Kartelle in Deutschland von der Holzstoffkartellentscheidung zum Gesetz gegen Wettbewerbsbeschränkungen (Tübingen: Mohr Siebeck 1994), p. 216. 29. On the relevance of this notion cf. Fritz Rittner, ‘Das Ermessen der Kartellbehörde’, in Horst Bartholomeyczik et al. (eds.), Festschrift für Heinz Kaufmann (Köln: Heymanns 1972), p. 315. 30. Cf. Arnold Köttgen, ‘Der Einflufß des Bundes auf die deutsche Verwaltung und die Organisation der bundeseigenen Verwaltung’, Jahrbuch für öffentliches Recht 11 (1962), p. 258; Oebbecke, Weisungs- und unterrichtungsfreie Räume, p. 45. 31. This term, still pouring out the flavour of an unconstitutional measure, was used by former chancellor Konrad Adenauer to underline his rejection of an independent Bundesbank. Cf. Volker Hentschel, Ludwig Erhard: ein Politikerleben (Berlin: Ullstein Verlag 1998), p. 364. 32. Werner Jann, ‘Zur Entwicklung der öffentlichen Verwaltung’, in Thomas Ellwein and Everhard Holtmann (eds.), 50 Jahre Bundesrepublik Deutschland, PVS special issue 30 (Opladen: Westdeutscher Verlag 1999), pp. 526–7; Eckard Schröter, ‘A Solid Rock in Rough Seas? Institutional Change and Continuity in the German Bureaucracy’, in B.Guy Peters and Jon Pierre (eds.), Politicians, Bureaucrats and Administrative Reform (London: Routledge forthcoming 2001). 33. Friedrich Bischoff and Michael Bischoff, ‘Parlament und Ministerialverwaltung’, in HansPeter Schneider and Wolfgang Zeh (eds.), Parlamentsrecht und Parlamentspraxis (de Gruyter: Berlin/New York 1989), p. 1468.

The politics of delegation

108

34. Roman Loeser, System des Verwaltungsrechts. Vol. II: Verwaltungsorganisationsrecht (Baden-Baden: Nomos 1994), p. 90; Arthur Benz and Klaus H.Goetz, ‘The German Public Sector: National Priorities and the International Reform Agenda’, in idem (eds.), A New German Public Sector? Reform, Adaptation and Stability (Aldershot: Dartmouth 1996), p. 21. 35. Susanne Lütz, ‘Die Rückkehr des Nationalstaates? Kapitalmarktregulierung im Zeichen der Internationalisierung von Finanzmärkten’, Politische Vierteljahresschrift 38 (1997), pp. 475–97. 36. Axel Murswieck, Die staatliche Kontrolle der Arzneimittelsicherheit in der Bundesrepublik und den USA (Opladen: Westdeutscher Verlag 1983), p. 204. 37. Helge Sodan, Kollegiale Funktionsträger als Verfassungsproblem (Frankfurt/M.: Metzner 1987), p. 419; Fritz Niklisch, Dieter Schottelius and Hellmut Wagner (eds.), Die Rolle des wissenschaftlich-technischen Sachverstandes bei der Genehmigung chemischer und kerntechnischer Anlagen (Heidelberg: C.F.Müller 1982). 38. Edgar Grande and Burkhard Eberlein, ‘Der Aufstieg des Regulierungsstaates im Infrastrukturbereich’, in Roland Czada and Helmut Wollmann (eds.), ‘Von der Bonner zur Berliner Republik. 10 Jahre deutsche Einheit’; special issue of Leviathan 19 (Opladen: Westdeutscher Verlag 1999), p. 641. 39. Susanne K.Schmidt, ‘Privatizing the Federal Postal and Telecommunications Services’, in Benz and Goetz, A New German Public Sector?, p. 62; Klaus Oertel, Die Unabhängigkeit der Regulierungsbehörde nach §§ 66 ff. TKG (Berlin: Duncker & Humblot 2000), p. 200. 40. Ernst-Joachim Mestmäcker, ‘Regulierende Mißbrauchsaufsicht’, in Bernhard Großfeld et al. (eds.), Festschrift für Wolfgang Fikentscher zum 70. Geburtstag (Tübingen: Mohr Siebeck 1998), p. 562. 41. Böllhoff, The New Regulatory Regime, p. 28. 42. Cf. Tagesspiegel, 9 July 1998, p. 22; ‘Über die Regulierung des Marktes neu nachdenken’, Blickpunkt Bundestag 7 (2000), p. 45; Monopolkommission, Dreizehntes Hauptgutachten 1998/1999. BT-Drs. 14/4002, p. 61. 43. Johan P.Olsen, ‘Institutional Design in Democratic Contexts’, Journal of Political Philosophy 5 (1991), p. 218. 44. Moe and Caldwell, ‘The Institutional Foundations’, p. 173. 45. For an overview, see Robert S.Gilmour and Alex A.Halley (eds.), Who Makes Public Policy? The Struggle for Control Between Congress and the Executive (Chatham, NJ: Chatham House 1994). 46. These terms are used by Joel D.Aberbach and Bert A.Rockman, ‘Bureaucracy: Control, Responsivenenss, Performance’, in Adbo Baaklini and Helen Desfosses (eds.), Designs for Democratic Stability: Studies in Viable Constitutionalism (Armonk, NY: M.E.Sharpe 1997), p. 83. 47. Hartmut Klatt, ‘Verwaltungskontrolle durch das Parlament’, in Hans-Georg Wehling (ed.), Verwaltung und Politik in der Bundesrepublik (Stuttgart: Kohlhammer 1986); Klaus von Beyme, Der Gesetzgeber. Der Bundestag als Entscheidungszentrum (Opladen: Westdeutscher Verlag 1997), p. 300; Joachim Krause, ‘Der Bedeutungswandel parlamentarischer Kontrolle: Deutscher Bundestag und US-Kongreß im Vergleich’, Zeitschrift für Parlamentsfragen 30 (1999), p. 543. 48. Horst Dreier, Hierarchische Verwaltung im demokratischen Staat (Tübingen: Mohr Siebeck 1991), p. 133. 49. Dreier, Hierarchische Verwaltung, p. 134. 50. Helmut Wiesenthal, ‘Auf dem Wege zu einer allgemeinen Theorie der Koordinationsmedien’, BISS public 9 (1999), p.109. 51. The former head of the Federal Environment Agency reports that if agency experts are invited to appear before a parliamentary committee ‘we inform the Government beforehand of the contents of our presentation and try to settle any differences that may exist’. Heinrich

Institutional choice and bureaucratic autonomy in Germany

109

von Lersner, ‘The German Federal Environment Agency’, in Alexander Kreher (ed.), The New European Agencies, EUI Working Paper RSC No. 96/49 (Florence: European University Institute 1996), p. 93. 52. Fritz W.Scharpf, Die politischen Kosten des Rechtsstaats. Eine vergleichende Studie der deutschen und amerikanischen Verwaltungskontrollen (Tübingen: Mohr Siebeck 1970), p. 38. 53. Roland Czada, Administrative Interessenvermittlung—am Beispiel der kerntechnischen Sicherheitsregulierung in den Vereinigten Staaten und der Bundesrepublik Deutschland (Habil. thesis, University of Constance 1992), p. 325. 54. Moe and Caldwell, ‘The Institutional Foundations’, p. 167. 55. Thomas Ellwein, Das Dilemma der Verwaltung. Verwaltungsstruktur und Verwaltungsreformen in Deutschland (Mannheim: BI-Taschenbuch Verlag 1994), p. 115; Peter Katzenstein, Policy and Politics in West Germany: The Growth of a Semi-Sovereign State (Philadelphia: Temple University Press 1987), pp. 273 and 382–4. 56. Cf. Helmuth Schulze-Fielitz, Theorie und Praxis parlamentarischer Gesetzgebung (Berlin: Duncker & Humblot 1988), p. 511; Wolfgang Zeh,‘Gesetzesfolgenabschätzung— Politikgestaltung durch Gesetze?’ in Werner Jann et al. (eds.), Politik und Verwaltung auf dem Weg in die transindustrielle Gesellschaft (Baden-Baden: Nomos 1998), p. 371. 57. Paul Kirchhof, ‘Mittel staatlichen Handelns’, in Josef Isensee and Paul Kirchhof (eds.), Handbuch des Staatsrechts, Vol. III (Heidelberg: Müller 1988), para. 217. 58. McCubbins et al., Administrative Procedures, p. 252. 59. Meinrad Schröder, ‘Grundfragen der Aufsicht in der öffentlichen Verwaltung’, Juristische Schulung 26 (1986), p. 372. 60. McCubbins et al., Administrative Procedures, p. 246; Horn, The Political Economy, p. 40. 61. Majone, ‘From the Positive to the Regulatory State’, p. 152. 62. Rob Coppock, ‘Interactions between Scientists and Public Officials: A Comparison of the Use of Science in Regulatory Programs in the United States and West Germany’, Policy Sciences 18 (1985), p. 387. For a case study, see Heike Bruns, ‘Das Umweltbundesamt (UBA) im Spannungsfeld zwischen Wissenschaft und Politik’, in Michael Kotulla (ed.), Umweltrecht und Umweltpolitik (Heidelberg: Physica Verlag 1998). 63. Colin Talbot, quoted in Pollitt and Bouckert, Public Management Reform, p. 135. For a specific case, see Clare Hall, Colin Scott and Christopher Hood, Telecommunications Regulation: Culture, Chaos and Interdependence Inside the Regulatory Process (London: Routledge 2000). 64. Torbjörn Larsson, ‘Cabinet Ministers and Parliamentary Government in Sweden’, in Michael Laver and Kenneth A.Shepsle (eds.), Cabinet Ministers and Parliamentary Government (Cambridge: Cambridge University Press 1994), p. 79; Stig Montin, ‘Central State Government Reforms in Sweden’, in Christoph Reichard and Werner Jann (eds.), Central State Governmental Reform in a Comparative Perspective (Washington DC et al.: Brookings Institution forthcoming), p. 5. 65. Cf. Martin Shapiro, Independent Agencies: US and EU. Jean Monnet Chair Paper No. 34 (Florence: European University Institute 1996), p. 9. 66. Giandomenico Majone, Temporal Consistency and Policy Credibility: Why Democracies Need Non-Majoritarian Institutions (EUI Working Paper RSC No. 96/57 1996), p. 1. 67. Gerhard Lehmbruch, ‘Die Große Koalition und die Institutionalisierung der Verhandlungsdemokratie’, in Max Kaase and Günter Schmid (eds.), Eine lernende Demokratie: 50 Jahre Bundesrepublik, WZB-Jahrbuch 1999 (Berlin: Edition Sigma), pp. 41–61. 68. Manfred G.Schmidt, ‘West Germany: The Policy of the Middle Way’, Journal of Public Policy 7 (1987), pp. 135–77. 69. McCubbins et al., Administrative Procedures, p. 250; Wolfgang Welz, Ressortverantwortung im Leistungsstaat (Baden-Baden: Nomos 1988), p. 197.

The politics of delegation

110

70. Cf. Scharpf, Die politischen Kosten des Rechtsstaats, p. 20; Martin Bullinger (ed.), Verwaltungsermessen im modernen Staat (Baden-Baden: Nomos 1986); Katzenstein, Policy and Politics in West Germany, p. 383. 71. Georg Nolte, ‘General Principles of German and European Administrative Law—A Comparison in Historical Perspective’, Modern Law Review 57 (1994), p. 193; Dreier, Hierarchische Verwaltung, pp. 18591. 72. André R.Fiebig, ‘The German Federal Cartel Office and the Application of Competition Law in Reunified Germany’, Pennsylvania Journal of International Business Law 14 (1993), p. 387. 73. Epstein and O’Halloran, Delegating Powers, p. 112. 74. Majone, ‘From the Positive to the Regulatory State’, p. 149. 75. Grande and Eberlein, ‘Der Aufstieg des Regulierungsstaates’, p. 641. 76. Bartle and Müller, The Administration of Regulation, p. 19. 77. Tanja A. Börzel and Thomas Risse, When Europe Hits Home: Europeanization and Domestic Change. Paper presented at the Annual Convention of the APSA, Washington, DC, 31 Aug.–3 Sept. 2000, p. 6. 78. Pollitt and Bouckert, Public Management Reform, p. 165. 79. Naschold et al., Innovation, Effektivität, Nachhaltigkeit, p. 16; Ole M.Resen, ‘From Bismarck to Benchmark? German Federal Administration at the Cross-roads’ (Ph.D. thesis, European University Institute 2000). 80. Resen, ‘From Bismarck to Benchmark?’, p. 272; Döhler, ‘Das Modell der unabhängigen Regulierungsbehörde’, p. 86.

Delegation to Independent Regulatory Agencies: Pressures, Functions and Contextual Mediation

MARK THATCHER The 1980s and 1990s saw sweeping reforms of regulation in Western Europe. Most attention has been focused on privatisation, the ending of state monopolies and the expansion of EC regulation.1 However, a key element was the creation of independent or semi-independent regulatory agencies at the national level. They have proliferated, spreading both across new domains and in countries which previously had few such agencies. Once established, they have often become central actors in decision making. Moreover, their role raises issues of democratic legitimacy and accountability. The creation, design and consequences of independent regulatory agencies (‘IRAs’) represent a classic example of delegation to non-majoritarian institutions. They are created by legislation; hence elected officials are their principals. They are organisationally separate from governments and headed by unelected officials. They are given powers over regulation, but are also subject to controls by elected politicians and judges. Cross-national comparison of delegation to regulatory agencies allows important theoretical issues to be addressed. Much of the principal-agent literature is based on experiences outside Western Europe, particularly of delegation by Congress to regulatory agencies.2 Western Europe enables the role of non-US national features to be analysed; moreover, comparison across countries provides variation in these features. The richness of the European experiences also allows isomorphism, policy transfer and learning across countries and across domains to be studied.3 The study uses principal-agent frameworks that are based on the functional logic of delegation as its starting point. It finds that, in response to powerful pressures, elected officials delegated to IRAs to fulfil functions such as overcoming information asymmetries, blame shifting, commitment, and dealing with complex, technical issues. However, it also argues that many features of delegation to IRAs cannot be explained by a simple analysis of pressures on elected policy makers and the advantageous functions that agencies performed for them. Contextual factors such as state traditions, policy learning and political leadership strongly mediate institutional responses to pressures. They mould whether delegation to agencies or another institutional arrangement is chosen, the timing of change and the institutional features of delegation. Moreover, an

The politics of delegation

112

analysis of functions and formal institutional structures cannot fully capture the consequences of delegation to regulatory agencies which have often been unanticipated. The discussion begins with an overview of the creation and key institutional features of IRAs in four countries (Britain, France, Germany, and Italy) across selected domains. It then discusses principal-agent models of delegation to IRAs and their ability to explain the patterns found in Western Europe. Thereafter it turns to contextual factors to offer a fuller explanation, notably of variations across countries and patterns of delegation to agencies that run counter to those expected by functionalist explanations. Finally, it looks at the consequences of delegation to IRAs. INDEPENDENT REGULATORY AGENCIES IN WESTERN EUROPE: AN OVERVIEW The 1980s and 1990s saw the creation of many new IRAs in Western Europe.4 Two broad groups can be distinguished. The first are agencies regulating the operation of markets. Examples include utility regulators, general competition authorities and financial bodies. They were typically given powers to prevent ‘unfair competition’, enforce licences and sometimes control prices. They also received powers over ‘social’ or distributional aspects of markets, for instance to ensure universal service or to protect specified groups. The second group is responsible for promoting ‘public interest’ goals other than competition. It is more heterogenous and includes agencies for the environment, safety (at work or of food, for instance), and racial and gender equality. They were given powers that include setting standards, issuing licences, prohibiting unauthorised supply and providing information. Delegation also involved strengthening existing agencies, especially in the 1990s. General competition authorities were given additional powers to enforce legal requirements; safety agencies received further responsibilities and enforcement instruments; the scope and legal standing of financial regulators such as stock exchange commissions and financial services authorities were reinforced.5 Defining an ‘independent regulatory agency’ is exceedingly difficult, given the diversity of agencies. Each country has its own legal doctrines and definitions: Italy has ‘independent administrative authorities’;6 France also has AAIs (Autorités administratives indepéndentes);7 in Germany, there are several types of agency—public and private law agencies, federal and Land agencies, most within or subordinate to ministries, with even ‘independent agencies’ such as the Federal Cartel Office and the telecommunications regulator, the RegTP, being (legally) subject to supervision and instructions from ministries;8 Britain does not have a legal doctrine of independent agencies. Following the definition of delegation and non-majoritarian in the Introduction, minimum requirements for inclusion as an IRA refer to the formal institutional position and comprise the following: the agency has its own powers and responsibilities under public law; it is organisationally separated from ministries; it is neither directly elected nor managed by elected officials. The autonomy of IRAs in practice from their principals and from regulatees is considered separately under the consequences of delegation. Table 1 sets out the spread of market and public interest regulatory agencies in key policy

Delegation to independent regulatory agencies

113

domains and the date of their creation (dates in brackets refer to when the first IRA in the domain was created). Whilst detailed comments on specific examples are offered below, certain general features are noteworthy. In all four countries, numerous independent regulatory agencies were created in the 1980s and 1990s. They were established earlier and more extensively in Britain than in other countries. Cross-domain comparison shows that market regulation agencies have spread more widely than public interest regulators. Certain agencies have been excluded from the table because they do not meet the requirements for delegation set out above. In France and Germany, bodies have been established in areas such as sexual equality, racial discrimination and the environment, but they are not independent regulatory agencies to which powers are delegated. In the case of France, they are largely consultative. Thus, for example, an agency for the environment was established in 1990 (the Agence de l’environnement et de la maîtrise de l’énergie) but its functions are to encourage, provide information and offer guidance; the Observatoire de la parité has similar functions for gender equality; in the railways, the CSSPF (Conseil Supérieur du service public ferroviaire) lacks independence from the government. In Germany, agencies such as those for food safety or the environment are mostly units within federal and/or Länder ministries. All regulatory agencies face continuing controls by elected officials—nominations, annual budget allocations, and requirements to report to legislatures. However, delegation to independent regulatory agencies has taken many institutional forms. The formal organisational position of agencies differs. Sometimes agencies are explicitly made independent in enacting legislation (for instance, in Italy); at other times, their institutional position is a unit subject to the supervision of a ministry and/or to its instructions (for instance, the Federal Cartel Office and the

TABLE 1 INDEPENDENT REGULATORY AGENCIES IN BRITAIN, FRANCE, GERMANY AND ITALY IN SELECTED DOMAINS Domain

Britain

General Competition competition Commission 1998 (1948) and Office of Fair Trading 1973

France

Germany

Italy

Conseil de la Concurrence 1986 (1977)

Bundes Kartellamt [Federal Cartel Office] 1957

Autorità Garante della Concor renza e del Mercato 1990

Regulieru ngsbehörde für Telekommunikation und Post (RegTP) 1996

Autorità per le Garanzie nelle Comunicazioni (AGCOM) 1997

Telecom Oftel (Office of Autorité de munications Telecommunications) Régulation 1984 des Télécom munications 1996 Energy

Ofgem (Office of Gas and Electricity Markets) 2000

Commission de Régulation de l’Energie

Autorità per l’energia elettrica ed il gas

The politics of delegation

(1989)

114

2003 (2000)

1995

Water

Ofwat (Office of Water Services) 1989

Railways

Office of Rail Regulator and Strategic Rail Authority 1999 (1993)

Postal services

Postal Services Commission 1999

Regulierungsbehörde für Telekommunikation und Post (RegTP) 1996

Media

Independent Television Commission 1990 (1954)

Conseil Landesmedienanstalt Supérieur de for each Land l’Audiovisuel 1989 (1982)

AGCOM (see telecommunications) 1996

Stock Exchange/ shares

Financial Services Authority 1997 (1986)

Commission des operations de bourse 1996 (1967); Conseil des marchés financiers 1996

Consob (La Commissione per le società e la borsa) 1974

Bafin (Bundesanstalt für Finanzdiensleistungen) 2002

Environment Environment Agency 1996 Food safety

Food Standards Agency 1999

Privacy

Data Protection Registrar 2000 (1984)

Commission nationale de l’informatique et des libertés 1978

Garante per la tutela dei dati personali 1996

Notes: 1 Dates refer to the creation of the IRA; dates in brackets refer to the date on which an IRA was first created in the domain. 2 Germany: (i) Media—each Land has a Landesmedienanstalt, which is a legal entity under public law, has a degree of financial and regulatory independence but within the supervision of the Land which ensures that it exercises its powers within the framework of legal and regulatory provisions. (ii) Railways—in 1994, an Eisenbahn-Bundesamt was created, responsible for technical and legal aspects of planning and operation and for safety, but subordinate to the Transport Ministry; economic regulation lay with the Cartel Office. (iii) There are several regulatory bodies for finance, but they lack the independence to qualify as NMIs; the most important is the Bundesaufsicht für des Wertpapierhandel created in 1995, but it is within the jurisdiction of the finance ministry; for reforms, see S. Lutz, ‘The Revival of the Nation-

Delegation to independent regulatory agencies

115

state? Stock exchange regulation in an era of globalized financial markets’, Journal of European Public Policy 5/1 (1998), pp. 153–68. 3 France: there are agencies for the environment (the Agence de l’environnement et de la maîtrise de l’énergie), railways (the Conseil Supérieur du service public ferroviaire) and gender equality (the Observatoire de la parité), but these lie within ministries or lack their own significant powers.

telecommunications regulator in Germany) or even a non-ministerial government department (the utility regulators in Britain). Thus ‘independent’ agencies include many bodies which are ‘semi-independent’ from legislatures and governments; for ease of reference, they are referred to henceforth as independent. There are also cross-national variations in the powers delegated: for example, in the utilities, British regulators have powers over licence modification, unlike their continental counterparts. The controls imposed vary: thus, for instance, legislatures in France, Germany and Italy usually have powers over nomination of regulators, unlike the British parliament. Table 2 summarises some of the typical features of IRAs in each country. PRINCIPAL-AGENT MODELS OF DELEGATION TO INDEPENDENT REGULATORY AGENCIES For principal-agent models, IRAs represent a prime example of delegation (transaction cost models are grouped within principal-agent approaches, since, although they use a slightly different vocabulary, they share the same essential elements).9 The creation and design of IRAs are a matter of institutional choice. Principals, in this case elected officials, calculate the costs and benefits of delegation.10 They delegate because they find that agencies can perform functions useful for them to deal with pressures and

TABLE 2 TYPICAL INSTITUTIONAL FEATURES OF IRAs Feature

Britain

Number of regulators

Nomination Power

Germany

Italy

Until 2000, single Multi-person person for utilities, commissions for others; moving towards multiperson

Multi-person (Presidential structure but decentralised internal decision making)

Multi-person

Ministers

Separate places for nominees of President, Prime Minister and heads of parliamentary houses

Federal government, but sometimes on advice of legislature (for RegTP)1

Separate places for nominees of President and two houses of parliament

Autorité

Within scope of

Independent

Organisational Non-ministerial

France

The politics of delegation

status

government department

administrative indépendente

116

supervisory authority responsibilities of a ministry (usually Economics ministry)

Notes: 1. The President of the Cartel Office is appointed by the Federal Government as a ‘technical’ civil servant (thus, excluding the possibility of ‘political retirement’ as provided for other senior civil servants), whereas the senior members of the RegTP have fixed four year terms (based on a private law contract and not a civil servant arrangement); however, decisions by both the RegTP and the Federal Cartel Office are made by court-like chambers that include solely of civil servants.

problems. If they delegate, they also choose the formal institutional form (notably the powers delegated and controls imposed) that minimises ‘agency losses’ arising from ‘shirking’ (divergence of the preferences of IRA and its principals) or ‘slippage’ (institutional design causing IRA decisions to differ from those desired by principals). In regulation, the central pressures and problems faced by elected officials include: making promises about future behaviour that are believed by other actors; taking decisions that are increasingly technical and complex; introducing policies that are unpopular; responding to the demands of powerful international/supranational organisations. Delegation to IRAs can perform useful functions for elected officials to assist them in responding to pressures and problems.11 The most important function of agencies is to enhance credible commitment. This may be vis-à-vis investors, especially in industries that require large capital expenditure: private investors must have sufficient reassurance that they will recover their expenditures with sufficient profit—otherwise, they will not invest, causing major difficulties for countries.12 It can also be vis-à-vis voters and organised interests—to enable governments to make long-term promises in return for immediate benefits (notably votes).13 Other functions include shifting blame, responding to information asymmetries and dealing with international organisations.14 Thus principal-agent models point to the functions performed by agencies and the pressures on elected officials to understand the spread of IRAs. If pressures increase, so does delegation (assuming prior equilibrium and no change in the costs of delegation). The institutional design of delegation is seen in terms of minimising costs, especially agency losses, for the principals. THE FUNCTIONAL REASONS FOR DELEGATION TO INDEPENDENT REGULATORY AGENCIES IN EUROPE Governments and legislatures in Western Europe faced increased pressures and problems in regulation during the 1980s and 1990s. These included: the unpopularity of policies designated as desirable or necessary; increased technical knowledge requirements; lack of confidence by actors vital for successful policies in policy domains; increased supranational regulation. IRAs performed functions that were useful for elected officials to respond to these pressures. Elected officials faced policy choices that they believed were necessary but would be unpopular. IRAs allowed them to shift blame for decisions. Thus, for instance, promoting

Delegation to independent regulatory agencies

117

racial and gender equality became seen as important by British political leaders and groups, but was greatly disliked by large sections of the electorate. Creating specialised agencies such as the Commission for Racial Equality or Equal Opportunities Commission in Britain allowed governments to appear to pursue ethical policies, but delegate the difficult tasks of enforcement.15 Agencies regulating competition provided a buffer between controversial decisions and the government. Thus, for instance, in the utilities, recently privatised incumbents often increased prices, especially for domestic users who had been cross-subsidised by large, commercial ones; regulators responsible for regulating such firms could take part of the blame for tariff increases. A similar rationale applies to general competition regulation, where takeovers and mergers involved losers as well as winners, the former often including employees who lost their jobs after successful bids. Regulation became much more technical in the 1980s and 1990s. Liberalisation of markets increased the technical requirements of regulation. In the utilities, regulators had to grapple with the costs and terms of interconnection among networks, standard setting, and preventing abuses of dominant position. New issues that emerged onto the regulatory agenda were frequently complex and involved very high levels of scientific expertise. Recent examples include food safety and environmental protection; the BSE issue exposed the difficulties for civil servants and ministers in even understanding the dangers of policy decisions, let alone forming an assessment of the risks posed. Juridification meant that public decisions had to be based on evidence and properly reasoned in order to withstand judicial challenges that became more common. Governments faced powerful interests—companies and non-governmental organisations (for example, environmental groups)—with high levels of information and the resources and desire to challenge them, including in the courts. Increased information requirements made it more difficult for elected politicians to produce clear benefits for voters from their regulatory decisions. Even issues that were more directly relevant to voters became increasingly linked to arcane matters that were incomprehensible to non-specialists. Thus price controls became enmeshed in questions of costs of capital, comparable rates of return, long-run incremental cost and tariff baskets—matters that were far from exciting for ministers. Complexity made it more difficult for ministers and legislators to participate effectively in decision making. Delegation to agencies of ‘unsexy’ technical policy making and implementation functions became more attractive to elected politicians (and their generalist civil servants). IRAs offered governments and legislatures a means for making credible commitments, a particular problem in domains and countries with long histories of unpredictable intervention by elected politicians. The utilities have been subject to constant activity by governments for short-term and non-economic objectives. Almost all privatisations have been accompanied by the creation of IRAs, allowing government to provide more credible commitments to investors and hence boost their revenues from the sales.16 Similarly in France, the creation of the Conseil des marches financiers in 1996 was driven by the desire to increase the international standing and attractiveness for overseas investors of the Paris Stock Exchange.17 IRAs have also been used for credible commitments vis-à-vis citizens and consumers. Scandals involving food and environmental safety and financial services have been followed in many countries by the creation of independent agencies, as trust in governments or self-regulation fell;18 the

The politics of delegation

118

most blatant case was the creation of the Food Standards Agency in Britain after the BSE scandal. In France, the first agency recognised as an autorité administrative indépendente in 1978, the CNIL (the Commission nationale de l’informatique et des libertés) was a response to a controversial law to merge several government databases containing personal details about citizens, whilst the creation of an audiovisual agency in 1982 was an attempt to reassure public opinion and journalists that the new government of the Left would allow journalistic freedom.19 Greater regulation by international organisations has provided impetus for delegation to IRAs. EC regulation has to be transposed and implemented by member states, and has accounted for a high proportion of national regulation.20 Although national governments may have great power at the EC level in deciding regulation, EC decisions can be controversial, difficult to implement and impose heavy costs. Thus, for example, EC environmental and food standards are often not met and require high expenditure in ‘laggard countries’. When EC requirements run counter to traditional standards and ways of operating, they have frequently been unpopular—for instance, food requirements. EC legislation allowing liberalisation and introducing re-regulatory rules in sectors such as the public utilities offers rights for frequently disgruntled new entrants to attack incumbents and public authorities. It has reduced the capacity of governments to pursue traditional industrial policies whilst also exposing them to greater challenge. Faced with EU demands, delegation to IRAs has had many advantages for national governments, notably offloading controversy, shedding difficult implementation decisions and ensuring an interlocutor for Brussels. The establishment of utility agencies in France and Italy was closely linked to EC legislation. When EC liberalisation made traditional statist industrial policies more difficult and opened the prospect of actions against state-owned former monopolists such as Electricité de France, France Télécom, ENI and Telecom Italia, the French and Italian governments rapidly responded by creating agencies in telecommunications and energy.21 Environmental regulation in Britain provides another interesting instance. Decades of under-investment, notably by water and sewerage authorities when in public hands, meant that the quality of water and beaches often fell below EC standards. The creation of the Environment Agency in Britain (and its predecessor the National Rivers Authority) and the water regulator Ofwat allowed the government to shift responsibility onto independent agencies for dealing with the legacy of past policy choices, including unpopular decisions to raise prices in order to pay for quality improvements. Pressures on policy makers and the functional advantages of delegation offer insights into the choice between creating IRAs and establishing bodies that are called agencies but do not have delegated powers and lack independence. The latter have been used in domains such as social and environmental policy in countries such as France and Germany. The four pressures on elected officials identified above were particularly strong in the regulation of markets, especially after privatisation, liberalisation and EU legislation. They were weaker in other fields, such as race relations or equal opportunities and differed in being concerned with public unpopularity more than technical expertise, credible commitment, and dealing with international obligations. Here legislatures may have wished to shift blame and hence delegation may have been more symbolic,22 or concerned with providing information and raising awareness. Thus they have not created IRAs, contenting themselves with ‘observatories’ and bodies within ministries (for

Delegation to independent regulatory agencies

119

instance, for the environment, racial and gender equality in France) or have delegated fewer powers than for market regulators (in the case of the British Equal Opportunities and Racial Equality Commissions). THE LIMITS OF A SIMPLE FUNCTIONALIST EXPLANATION The general fiunctional advantages of delegation to independent regulatory agencies for elected officials to respond to pressures offer an excellent starting for analysis of delegation. Nevertheless, they are insufficient to explain the patterns seen in Western Europe in the 1980s and 1990s. Many of the pressures for change and advantages of delegation were long-standing, but had not led to substantial delegation to independent regulatory agencies before the 1980s and the 1990s. There were scandals, lack of credible commitment and unpopular policies in domains such as utilities policy or financial regulation in several countries. Moreover, the timing of agency creation has differed. IRAs have generally been created earlier in Britain than in the other three nations. A rare exception is offered by general competition regulation, where Germany has had a powerful specialised agency since 1957, much earlier than other countries, but the puzzle remains that this example was not replicated in other domains.23 Finally, several of the functional pressures on elected officials could have been met through institutional reforms other than delegation to IRAs—for instance, by strengthening government departments to deal with technical complexity, greater use of constitutional entrenchment to increase the credibility of commitments or increased powers for the EU to shift blame and implementation problems. Powerful functional pressures led to delegation in some countries but not in others in the same domains. Britain has created IRAs in more domains than other countries. Despite its history, problems of racial conflict and powerful feminist movements, Germany has not created IRAs for racial or gender equality, whereas Britain has done so. Even in the utilities, variations remain in 2001: Britain has IRAs in each sector; France and Italy lack an IRA for postal services, railways and water, and France also one for gas; Germany only has an IRA for telecommunications and postal services,24 with other utility sectors being covered by the Federal Cartel Office.25 Even in the face of similar powerful pressures, countries have made different institutional choices. Thus, in electricity, all four countries have faced EU regulation, technical complexity and the need for credible commitment given the long-term nature of the industry; nevertheless, Germany has not created a sectoral regulator.26 Comparison of the institutional features of delegation also reveals the limitations of focusing solely on the functions performed by agencies. IRAs enjoy differing organisational positions, ranging from being non-governmental departments (the utility regulators in Britain) to being legally independent (for example, the Italian Competition Authority). Within the same issue domain, there are cross-national differences, despite the presence of similar reasons for delegation such as the need for credible commitment and international regulation. In Britain and Germany, general competition authorities and sectoral utility regulators enjoy greater powers than in France and Italy. General competition authorities in the first two have stronger tools to block mergers and takeovers

The politics of delegation

120

and enforce competition rules, while utility regulators in Britain have powers not only to enforce but also to seek to modify licences. At times, cross-national findings run counter to certain functional pressures. Thus, although France and Italy faced more significant problems of credible commitment than Britain and Germany given their traditions of political intervention and instability, they have delegated fewer powers to their sectoral regulators and have created many more controls, notably over nominations and reporting, in the hands of elected officials.27 The number of regulators within IRAs and the selection procedures for choosing them have taken many forms—there is no one ‘European model’. In Britain, single-person regulators have been used in the utilities during the 1980s and 1990s (although this is now being altered28) whereas the general competition authority, the Commission for Racial Equality and the Equal Opportunities Commission are headed by boards with a chairperson. Regulators are nominated by ministers, without the need even for parliamentary approval. In contrast, on the continent, agencies regulating competition are headed by multi-person boards. Complex arrangements have been made to ensure that the legislature and executive share nominations, and that different political formations are able to control at least some nominations. Thus, for example, in telecommunications, the Italian communications regulator AGCOM has nine members, with its head being nominated by the president on proposal of prime minister; four are elected by the Chamber of Deputies and four by the Senate; in Germany, the telecommunications regulators are nominated by the federal government on the proposal of an Advisory Council, composed of members of the two legislative chambers, the Bundestag and the Bundesrat. The length of terms of agency members in Britain is also typically shorter than in continental countries, especially for market regulators. Cross-domain comparison equally suggests that functional advantages can explain only some of the patterns of delegation to IRAs. In certain cases, similar institutional forms have been used within countries across domains with very different characteristics. The clearest example is the utilities in Britain, where regulators very similar to Oftel in telecommunications (Ofwat for water, Offer for electricity and Ofgas for gas) were created during the 1980s, although pressures for delegation (such as the need for technical complexity or the extent of supranational regulation) have varied greatly. CONTEXTUAL FACTORS MEDIATING PRESSURES FOR DELEGATION Cross-national and cross-domain variations in delegation to IRAs suggest that there is no automatic link between functional advantages of delegation and the creation of IRAs. Instead, the choice of delegation as a response must be explained rather than assumed. It must be analysed in its context as well as a response to pressures and functional advantages for governments and legislators. Contextual factors aid a better explanation of the patterns of delegation found across countries and domains: differences in the spread of IRAs and the timing of their creation; the various institutional forms of IRAs chosen. Four factors are examined here: policy learning and institutional isomorphism; state traditions and structures in regulation; political leadership; the broader institutional context of West European states.

Delegation to independent regulatory agencies

121

Learning and Institutional Isomorphism Policy learning and institutional mimetism were important in the spread of independent regulatory agencies.29 Within countries, ‘snowball’ effects meant that once an apparently successful model of an agency existed, it was copied in other domains. Thus, for instance, in Britain, the Commission for Racial Equality followed the lack of opposition to the creation of the Equal Opportunities Commission.30 Similarly, the example of Oftel (the Office of Telecommunications) was rapidly copied in other utility sectors in Britain.31 In France, the officially recognised first autorité administrative indépendente (the CNIL) was only created in 1978, but thereafter similar bodies were gradually established in other regulatory domains such as broadcasting, telecommunications and electricity.32 In Italy, mimetism followed the establishment of the general competition authority in 1990. Copying took place despite considerable contrasts in the features of the sectors and hence functional reasons for delegation to IRAs. It gave rise to strongly national patterns of delegation, especially in agency design. It also helps to explain cross-national variations in timing of IRAs: agencies spread after a powerful example had been created. Cross-national learning and institutional isomorphism were also important in certain domains, especially highly internationalised ones such as telecommunications. The creation of Oftel was influenced by the example of the Federal Communications Commission.33 In turn, Oftel served as an example for the spate of national telecommunications regulators that were established even in countries with little tradition of such bodies (for instance, Italy). Within Europe, the EC Commission brought together national regulators and encouraged cross-national fertilisation of ideas, aiding the spread of the new ‘model’ of independent regulators. Thus, for example, in telecommunications and energy, groups of senior regulators were formed (for example, the High Level Telecommunications Regulators group or the Florence and Madrid groups for electricity and gas). Dynamics within domains appear to operate: once ‘successful’ examples appear in several European countries, they are copied in other countries. Moreover, both learning and ‘coercive isomorphism’ apply, as bodies such as the EC Commission put pressure on countries that have become the ‘odd men out’ (for example, Germany in the field of electricity or Italy in general competition regulation before 1990).34 State Traditions and Structures State traditions and existing state structures affected delegation to agencies. Establishing IRAs was easiest in Britain, where regulatory commissions dated from the nineteenth century—for example, in the railways.35 When difficult problems arise for governments, they find a tradition of creating IRAs. In contrast, Italy had little tradition of independent agencies and the process of delegation only began in earnest after the successful creation of the Competition Authority in 1990. In France there existed a long-standing suspicion of ‘independent’ agencies that would either not be properly independent or which fragmented the unity of the state; however, in the late 1970s and 1980s judicial doctrines emerged that were more favourable.36 Germany provides an interesting case of opposing forces: on the one hand, successful federal agencies existed (notably the Federal Cartel Office); on the other hand, there were strong traditions of self-regulation and of subnational autonomy. The result has been a ‘German model’ of reliance on selfregulation and the Federal Cartel Office.37

The politics of delegation

122

State traditions and structures have marked the institutional design of delegation to IRAs, and hence variations by country and domain. There is a long history in most continental European countries of direct parliamentary and sub-national involvement in utility services, and legal and political doctrines that they are ‘public services’. These have justified legislative powers over agency nominations. In countries where the legislature has played an important role in a domain and has greater independence from the executive, it has maintained important controls over agencies. Thus in France, Germany, and especially Italy, many nominations are in the hands of the legislature. In contrast, the British parliament has almost no powers over the choice of regulator, which is made by ministers. Political Leadership Political leadership was needed to establish IRAs. Countries with strong leaderships could respond more quickly to pressures for delegation than those without such leadership. Reform was easier in Britain, with its single-party majoritarian system, than in Italy, which was dominated by unstable weak governments during most of the 1980s and 1990s. The contrast is best seen in information and communications. The Thatcher governments rapidly created Oftel in 1984 and reformed television broadcasting (notably by establishing the Independent Television Commission). It took several years for Italy to create its telecommunications and media regulator (AGCOM—the Autorità per le Garanzie nelle Communicazioni), due to difficulties over media ownership, making it one of the last countries in Western Europe to set up an independent telecommunications regulator. In other domains, British governments have been able to delegate to IRAs rapidly when confronted with scandals: they created and/or strengthened financial services and food safety agencies in the 1990s following problems such as pensions misselling, insider trading or BSE. Political leadership also influences the choice of agency form. In a country with majoritarian political systems and few constitutional veto points such as Britain, it is possible to create single-person agencies nominated by the executive (notably in the utilities in the 1980s). This appears an exception. In most countries, multi-partyism and multiple veto points (for instance, between heads of governments and presidents, or between the two legislative chambers) have encouraged an agency form that ensures representation for several parties and players. State Reforms Delegation to independent agencies must also be seen in the broader context of reforms of state structures, especially privatisation, liberalisation and new public management. Thus, in Italy, independent agencies spread in the 1990s after the end of the first republic, as the public sector was reformed.38 The creation of agencies to regulate competition in the utilities has usually accompanied structural reforms such as the sale of suppliers, ending state monopolies or breaking up vertically integrated public firms. In Britain, the creation of the utility regulators was undertaken in the same legislation that privatised the state firms. The creation of an energy regulator in Italy accompanied the partial sale of ENEL and liberalisation of the market, whilst the liberalisation of electricity in France

Delegation to independent regulatory agencies

123

was also met with a new independent agency.39 Moreover, the greater the spread of new public management doctrines, the more agencies were created—such doctrines being more popular and arriving earlier in Britain than in France and Germany.40 THE CONSEQUENCES OF DELEGATION TO INDEPENDENT REGULATORY AGENCIES Principal-agent frameworks focus on the extent to which elected officials can minimise agency losses through the formal institutional design of delegation, notably by choosing which powers are delegated to the agency and which controls are imposed on it.41 Whilst a useful starting point, analysis shows that delegation to IRAs has had broad consequences, well beyond the issues of ‘slippage’ and ‘shirking’ traditionally at the heart of principal-agent models, that many effects were unanticipated at the time of delegation and that several cannot be captured solely by studying formal institutional arrangements, but require inclusion of contextual factors. Relationships between IRAs and Elected Officials after Delegation Initially, new regulatory agencies appeared to conform to national patterns of policy making. In Britain, regulators have tended to be drawn from ‘the great and the good’ that is, mostly non-party partisan experts and officials.42 In contrast, the allocation of posts in continental countries has been more clearly party political, with nominees frequently being either party politicians or clearly identified with a party.43 Ex post formal controls have rarely been used, especially in agencies regulating competition; for instance, dismissals of regulators have been extremely infrequent and government powers to reverse decisions little exercised. In contrast, informal methods have been much more important. In many countries, government officials and regulators have been in continuous dialogue. Thus, focusing on formal controls appears to be inappropriate; less formal networks of influence and linkages are the usual means whereby governments and legislatures have sought to control their agents. Indeed, the degree of public conflict between them has generally been very limited in the initial stages of agencies’ lives. Governments maintained control and well-established relationships continued between regulatees and governments—for instance, between large suppliers such as France Télécom or Telecom Italia and their governments. Yet, over time, the relationships between agencies and their principals have altered, often in ways that cannot be ‘read off’ formal arrangements. Agencies have tended to gain power and importance. They have become key actors in decisions, acquiring expertise, reputations and political weight. They have developed objectives and conceptual frameworks that have structured their policy domains. Thus, in Britain, the utility regulators have expanded their roles thanks to their expertise, information, constant contact with interested parties and conceptual frameworks, as well as the use of their formal powers.44 In Germany, the Federal Cartel Office has gained great independence in practice from ministers, who have almost never used their powers to overturn its decisions; indeed, the Office has moved to use quasi-judicial procedures, with investigations and court-like ‘chambers’ taking decisions. Even in countries with

The politics of delegation

124

little tradition of independent agencies, such as France and Italy, the new regulators have gradually established their own separate profile and ideas. One prominent example is the Italian Competition Authority, which has become a powerful and respected body, able to act effectively in politically contentious fields.45 The agencies have become more independent of elected officials, able to challenge their views, sometimes even in public. Thus the Italian Competition Authority and the French telecommunications regulator ART disagreed publicly with government plans for licensing UMTS services in 1999– 2001 and were partly responsible for changes in government choices.46 The survival, strength, and independence of IRAs vis-à-vis governments were frequently unanticipated at the time of their creation. Thus the British telecommunications regulator Oftel was designed as a temporary measure to ‘hold the fort’ until competition arrived; 17 years later it is still going strong. Similarly, expectations that utility and competition agencies in France and Italy would be highly politicised and under the control of elected officials because nomination procedures gave powers to legislatures have not been realised. Policy Making IRAs have influenced policy making, sometimes in directions that were surprising given the formal institutional framework. They have altered the processes of decision making. Although delegation legislation offered few provisions, IRAs have introduced new procedures—for example, public hearings and consultation documents. New actors have been allowed to enter the ‘regulatory space’47 —competitors, foreign firms, and user groups. Much more information has been published. Closed cosy relationships between ministers, officials, and privileged firms have been weakened, particularly those involving ‘national champion’ firms. Relationships have become more legalistic, with a host of new formal rules replacing informal norms;48 this has been the case even in domains with strong inheritances of ‘self regulation’ such as financial markets, including the City of London. The creation of agencies has frequently also been followed by greater antagonism and conflict between traditional ‘insiders’ and the state, on several occasions leading to court cases. Thus, in Germany, the telecommunications regulator, RegTP, has locked horns on many occasions with Deutsche Telekom, with the latter challenging decisions before the courts. Agencies have also have influenced substantive policy. Regulators of competition have often advanced policies of liberalisation and effective competition, sometimes unexpectedly. In Britain, this was striking—in the original statutes, advancing competition was only a secondary duty in legislation creating agencies such as Oftel; yet they made it their primary aim.49 In France and Italy, agencies such as the general competition and telecommunications agencies have developed powerful norms based on fair competition, contrary to state traditions. Further Institutional Reforms The behaviour of IRAs has affected further institutional reforms, for governments have altered arrangements in the light of experience. Such processes have seen much learning, both from previous national experience and from overseas.50 In general, IRAs have been given further powers, balanced by more refined monitoring controls over their actions for

Delegation to independent regulatory agencies

125

legislatures and regulatees. Under the 1998 Competition Act the Monopolies and Mergers Commission in Britain received powers to punish anti-competitive behaviour based on Articles 80 and 81 of the Treaty of Rome. Following concerns about lack of clarity in objectives and transparency, together with lack of independent representation of consumers,51 the energy regulators were strengthened under the Utilities Act 2000, which widened their powers to prevent and punish anti-competitive behaviour, but also tightened procedural requirements and provided consumers with their own Council. Accountability and Legitimacy The effects on accountability and legitimacy of delegation to independent agencies are complex. Independent agencies with their own policy making powers pose problems for traditional West European models of accountability centred on legislatures that can be held to account by the electorate.52 They are unelected bodies that by definition enjoy discretion. Many were set up with few procedural guidelines and with broad objectives, enhancing their discretion. However, the new institutional arrangements need to be compared with their predecessors, when powers lay in the hands of governments and legislatures. In practice, the accountability and legitimacy of policy makers before the creation of independent regulatory agencies was often low.53 Regulation of domains such as the environment or the utilities was opaque at best, secretive at worst. Legislative control was largely a fiction in systems with strong executives enjoying disciplined parliamentary majorities and much greater information and expertise than legislatures. Even within the executive, effective control by ministers was difficult and rare—they relied on civil servants who were often generalists and had limited knowledge and understanding of technical matters. Thus delegation to IRAs took place in a situation in which legislative models of accountability were closer to fiction than reality. In practice, several forms of accountability and legitimacy have been developed for agencies. One has been the use of controls available to elected officials (both executives and legislatures) such as appointments. More important has been ‘answerability’. Agencies have explained their actions and been obliged to face questioning for them. They have been answerable to legislatures—for example, by legal duties to produce annual reports, by their officials appearing before legislative committees and by being subject to administrative scrutiny bodies such as the National Audit Office in Britain or the Cour des Comptes in France. Answerability has extended to regulatees and consumers. Agencies have published consultation documents before taking decisions, set up public meetings and hearings in which they have explained their reasoning, and listened to criticisms. They have provided reasons and explanations for their decisions, together with supporting data, thus making it easier to raise objections than previously when ministerial choices were announced with little further information. Occasionally, forms of appeal or challenge have been created; thus, for example, the licence modification procedure for utilities in Britain whereby if a regulatee and regulator cannot agree, the latter can make a reference to the Competition Commission, offers a form of appeal. Accountability via judicial control has increased: agency decisions have been challenged via administrative review, a process that has been much easier thanks to the much greater amount of information and reasoning provided.

The politics of delegation

126

CONCLUSION Principal-agent frameworks offer an excellent starting point to analyse delegation to IRAs. They point to the fiunctional reasons that lead elected officials to choose delegate, such as desires to shift blame, the need for credible commitments and to respond to greater technical and international demands on national policy makers. In Western Europe, delegation to IRAs aided elected officials to face new or increased pressures, such those as arising from privatisation, liberalisation, technological and economic developments, clashes between public opinion and desirable objectives, and the effects of supranational organisations. Yet a purely functionalist and formal institutional account of delegation is inadequate. Its limits can be seen through comparison across countries and policy domains and over time. Despite facing similar pressures, countries have made diverse choices over delegation to independent regulatory agencies. Britain has had the greatest delegation in the 1980s and 1990s, establishing independent agencies in both the regulation of competition and in social and environmental fields. Fewer agencies have been created in other countries. In terms of new agencies and delegation in the 1980s and 1990s, Germany has delegated least. On several occasions, the findings run counter to those expected. France and Italy have had less delegation than Britain, despite facing greater problems of credible commitment. There have been many differences in the timing of delegation, its extent and its institutional form across nations. Conversely, cross-domain comparison within countries shows that, despite functional pressures for delegation, different choices have been made over whether to establish independent agencies. Thus, in certain domains, independent agencies have become the norm—notably telecommunications and general competition regulation. Other domains have many fewer agencies—for instance, gender and racial equality, the environment and water. Finally, the same institutional form has sometimes been copied across domains within a country despite significant differences in the pressures for delegation. Functional pressures do not determine a single institutional response. Contextual factors strongly influence political choices by mediating pressures for delegation, institutional responses to those pressures and the effects of delegation. In some countries and/or domains, policy learning, state traditions, and structures encourage policy makers to respond to pressures and problems by delegating to IRAs. Britain is an example where contextual factors have aided delegation: policy problems are focused on the central government executive; thanks to state traditions and the influence of learning from the United States, delegation to independent regulatory agencies is often proposed as an institutional response. However, in other countries and domains, contextual factors can retard or prevent delegation to IRAs. Alternative models of institutional structures can rival IRAs. Weak political leadership, numerous veto points and constitutional obstacles can make delegation to IRAs difficult and slow. The German federal system and state traditions hindered delegation to IRAs, despite pressures such as supranational regulation or privatisation and liberalisation; instead, regulatory bodies within ministries have been used. In France and Italy, delegation to IRAs only spread after initial successful agencies had been established; it remains largely limited to the regulation of competition, with few public interest IRAs.

Delegation to independent regulatory agencies

127

The importance of contextual factors is also visible in decisions over the institutional form of IRAs. In the face of similar problems, national policy makers have different choices over the extent of delegation, the powers delegated, and the controls imposed on independent regulatory agencies. Even in a domain such as telecommunications, characterised by powerful transnational forces, supranational regulation and much crossnational policy learning, significant differences remain in the institutional form of IRAs. Conversely, the same forms are copied across domains within nations, despite differing pressures. The consequences of delegation to IRAs are far from fully captured by analysis of the purposes of delegation and formal institutional arrangements. Informal linkages that affect the distribution of power and policy making develop. IRAs can be actively engaged in developing their role and power. Outcomes can be unexpected at the time of the original delegation. Context influences the wider effects of delegation to agencies. Broader analysis is needed beyond pressures, functional advantages, and the formal institutional frameworks of delegation. Contextual factors allow variations in responses to pressures and functional advantages of delegation to IRAs to be understood, together with the ways in which such agencies operate and develop. NOTES The author expresses his thanks for comments on earlier drafts to the reviewer, to Martin Lodge and to participants at the ECPR conference at Grenoble. He thanks Maria Kampp, Francesco Salerno and Danilo von Sperling for research assistance. He also acknowledges support from ESRC grant no. L216 25 2007 within the ESRC Future Governance programme, and the Centre for the Analysis of Risk and Regulation. 1. See J.S.Vickers and V.Wright (eds.), The Politics of Privatization (London: Frank Cass 1989); H.Feigenbaum, J.Henig and C.Hamnett, Shrinking the State (Cambridge: Cambridge University Press 1999); V.Wright (ed.), Privatisation in Western Europe (London: Pinter 1994); G.Majone, ‘Paradoxes of Privatisation and Deregulation’, Journal of European Public Policy 1/1 (1994), pp. 53–69. 2. See, for instance, M.McCubbins, ‘The Legislative Design of Regulatory Structure’, American Journal of Political Science, 29/4 (1985), pp. 721–48; R.Calvert, M.McCubbins and B.Weingast, ‘A Theory of Political Control and Agency Discretion’, American Journal of Political Science 33/3 (1989), pp. 588–611; and, more generally, D.Epstein and S. O’Halloran, Delegating Powers (Cambridge: Cambridge University Press 1999); Britain is sometimes included—see B.Levy and P.Spiller, Regulation, Institutions and Commitment (Cambridge: Cambridge University Press 1996); the major exception that offers a European perspective is G.Majone, Regulating Europe (London: Routledge 1996). 3. Cf. C.Radaelli, ‘Policy Transfer in the European Union’, Governance 13/1 (2000), pp. 25–43. 4. For cross-national studies in key domains, see Majone, Regulating Europe; G.Giraudi and M.Righettini, Le autorità amministrative independenti (Rome: Laterza 2001); B.Doern and S.Wilks (eds.), Comparative Competition Policy: National Institutions in a Global Market (Oxford: Oxford University Press 1996); D.Coen and M.Thatcher (eds.), Utilities Reform in Europe, special issue of Current Politics and Economics of Europe 9/4 (2000). 5. Examples include the 1998 Competition Act in Britain, which for decades had been seen as a country with weak general competition authorities relative to Germany (cf. Wilks and Doern, Comparative Competition Policy), the reform of the COB (Commission d’Operations de la Bourse) in France, the move from informal non-statutory regulation of the City to much

The politics of delegation

128

more formalised rules underpinned by statute, notably by the 1986 Financial Services Act, the establishment of the Financial Services Authority in 2001, and statutory regulation of Lloyds of London; in Italy, Consob, the Stock Exchange regulator, was greatly strengthened in 1985. 6. Cf. S.Cassese and C.Franchini (eds.), I garanti delle regole (Bologna: Il Mulino 1996); R. Perez, ‘Autorità Indipendenti e Tutela dei Diritti’, Revista trimestrale di diritto pubblico 1 (1996), pp. 115–47; F.Gobbo and G.Zanetti (eds.), Istitutiozioni e mercato: il ruolo delle Autorità nell’economia italiana, special issue of L’Industria, 21/4 (2000); E.Martinelli, ‘Autorità independenti e politica’, Amministrare 30/12 (2000), pp. 127–46; Giraudi and Righettini, Le autorità amministrative independenti. 7. For discussions, see Conseil d’État, Rapport Public 2001 (Paris: La documentation Française 2001); C.Teitgen-Colly, ‘Les instance de regulation et la constitution’, Revue de Droit Public 36/1 (1990), pp. 153–262; C.-A.Colliard and G.Timsit (eds.), Les autorités administratives indépendentes (Paris: Presses Universitaires de France 1988); M.-J.Guédon, Les autorités administratives indépendentes (Paris: Librairie Générale de Droit et de Jurisprudence 1991). 8. Cf. contributions by Marian Döhler and Stephen Wilks/Ian Bartle, this volume. 9. For major general studies in addition to those discussed in Thatcher and Stone Sweet, this volume, that gives an important place to delegation to independent regulatory agencies, see M.Horn, The Political Economy of Public Administration (Cambridge: Cambridge University Press 1995); Levy and Spiller, Regulation, Institutions and Commitment; M. Fiorina, Congress: The Keystone of the Washington Establishment (New Haven CT: Yale University Press 1977); and ‘Legislative Choice or Regulatory Forms: Legal Process or Administrative Process?’, Public Choice 39 (1982), pp. 33–71; Majone, Regulating Europe. 10. The most systematic model of costs and benefits is Murray Horn’s transaction cost approach. He examines the choice between delegation to an independent agency, to courts and to an executive agency. He also considers decisions over key features of institutional design, notably the extent of delegation, the entrenchment of the agency (its protection from action by future legislatures to alter its organisational structure), its independence from the incumbent legislature and the extent of procedural rights given to private interests specially to participate in the agency’s decision making. Legislators make their choices by weighing up four sets of transaction costs: decision making costs, passing legislature for governments and participation in decision making for private interests; agency costs, as agencies fail to comply with the intentions of legislators; commitment costs, the risk that future legislators may intervene; uncertainty costs, concerning the future costs and benefit of legislation. 11. Cf. G.Majone, ‘The Regulatory State and its Legitimacy Problems’, West European Politics 22/1 (1999), pp. 1–24. 12. Levy and Spiller, Regulation, Institutions and Commitment. 13. Horn, The Political Economy of Public Administration. 14. Cf. Fironia Congress, for blame shifting; cf. Thatcher and Stone Sweet, ‘Theory and Practice of Delegation’, this volume. 15. Even so, their powers were limited, notably by being constrained to assist individuals bringing cases rather than doing so themselves—for details, see G.Appleby and E.Ellis, ‘Formal Investigations: The Commission for Racial Equality and the Equal Opportunities Commission as Law Enforcement Agencies’, Public Law (Summer 1984), pp. 236–76; V. Sacks, ‘The Equal Opportunities Commission Ten years On’, Modern Law Review 49 (1986), pp. 560–92. 16. Cf. Levy and Spiller, Regulation, Institutions and Commitment. 17. See Conseil d’État, Rapport Public 2001, p. 268. 18. For discussions of financial services reforms involving a mix of self-regulation and statutory regulation, see J.Black, Rules and Regulators (Oxford: Clarendon Press 1997), ch. 2; A. Page, ‘Financial Services: The Self-Regulatory Alternative?’ in R.Baldwin and C.

Delegation to independent regulatory agencies

129

McCrudden (eds.), Regulation and Public Law (London: Weidenfeld and Nicolson, 1987); and M.Moran, The Politics of the Financial Services Revolution: The UK, US and Japan (London: Macmillan 1990). 19. Conseil d’État, Rapport Public 2000; S.Hessel, ‘La Haute Autorité de la communication audio-visuelle’ in Colliard and Timsit, Les autorités administratives indepéndentes. 20. Cf. Majone, Regulating Europe. 21. Giraudi and Righettini, Le autorità amministrative independenti, pp. 136–9; Conseil d’État, Rapport Public 2001, pp. 271–4. 22. Cf. B.Mitnik, The Political Economy of Regulation (New York: Columbia University Press 1980). 23. See contributions by Stephen Wilks and Marian Döhler in this volume. 24. Its agency for railways in not independent see M.Lodge, ‘On Different Tracks: Institutions and Railway Regulation in Britain and Germany’ (Ph.D. thesis, London 2000). 25. Cf. D.Coen and A.Héritier, ‘Business Perspectives on German and British Regulation: Telecoms, Energy and Rail’, Business Strategy Review 11/4 (2000), pp. 29–37; A.Héritier and S.K.Schmidt, ‘After Liberalization: Public-Interest Services in the Utilities’, in F. Scharpf and V.Schmidt (eds), Welfare and Work in the Open Economy (Oxford: Oxford University Press 2000). 26. B.Eberlein, ‘Institutional Change and Continuity in German Infrastructure Management: The Case of Electricity Reform’, German Politics 9/3 (2000), pp. 81–104. 27. For a discussion of the many powers of elected politicians over CONSOB, the stock exchange regulator in Italy, see Martinelli, ‘Autorità independenti e politica’; and, generally, Perez, ‘Autorità Indipendenti e Tutela dei Diritti’; for a legal discussion of the limits to the independence of Autorités Administratives Indépendentes in France, see Guédon, Les autorités administratives indépendentes, pp. 62–86. 28. Cf. the 2000 Utilities Act setting up a Board for energy, and the proposed Communications Commission. 29. Cf. W.W.DiMaggio and P.J.Powell ‘The Iron Cage Revisited’, in P.J.Powell and W.W. DiMaggio (eds.), The New Institutionalism in Organizational Analysis (Chicago: University of Chicago Press 1991). 30. C.McCrudden, ‘The Commission for Racial Equality’, in R.Baldwin and C.McCrudden (eds.), Regulation and Public Law (London: Weidenfeld and Nicolson 1987). 31. Ofwat was created for water, Offer and Ofgas for electricity and gas (today merged as Ofgem), and the Office of Rail Regulator for the railways. 32. For discussions, see Conseil d’État, Rapport Public 2001. 33. Cf. M.Thatcher, The Politics of Telecommunications (Oxford: Oxford University Press 1999). 34. Cf. Eberlein, ‘Institutional Change and Continuity in German Infrastructure Management’; G.Giraudi, ‘Europeizzazione delle politiche pubbliche e mutamento instituzionale: la politica antitrust italiana’, Revista Italiana di Scienza Politica 30/2 (2000), pp. 257–94. 35. See C.D.Foster, Privatisation, Public Ownership and the Regulation of Natural Monopoly (Oxford: Blackwell 1992); J.Foreman-Peck and R.Millward, Public and Private Ownership of British Industry 1820–1990 (Oxford: Clarendon Press 1994). 36. Teitgen-Colly, ‘Les instances de regulation et la constitution’. 37. Cf. I.Bartle and M.Müller, Causes and Consequences of Regulatory Transformation: A British-German Comparison (Exeter: Discussion paper 3, 2000). 38. See Cassese and Franchini, I garanti delle regole; Perez, ‘Autorità Indipendenti e Tutela de Diritti’. 39. Conseil d’État, Rapport Public 2000; P.Ranci, ‘La regolazione elettrica in Italia’ Industria 21/4 (2000), pp. 749–64. 40. C.Pollitt and G.Bouckaert, Public Sector Management Reform: A Comparative Analysis (Oxford: Oxford University Press 2000); H.-U.Derlien, ‘Actor Constellation, Opportunity

The politics of delegation

130

Structure and Concept Feasibility in German and British Public Sector Reforms’, in H. Wollmann and E.Schröter (eds.), Comparing Public Sector Reform in Britain and Germany (Aldershot: Ashgate 2000). 41. For discussions of institutional design of controls, see especially M.D.McCubbins and T. Schwartz ‘Congressional Oversight Overlooked: Police Patrols versus Fire Alarms’, American Journal of Political Science 28/1 (1984), pp. 165–79; R.D.Kiewiet and M.D. McCubbins, The Logic of Delegation (Chicago: University of Chicago Press 1991); M.D. McCubbins, ‘The Legislative Design of Regulatory Structure’, American Journal of Political Science 29/4 (1985), pp. 721–48; on limiting agency discretion, see D.Epstein and S. O’Halloran, ‘Administrative Procedures, Information and Agency Discretion’, American Journal of Political Science 38/3 (1994), pp. 697–722. 42. Thus, for instance, the heads of the utility regulators were experts such as Professor Steven Littlechild (Offer) or Professor Sir Bryan Carsberg (Oftel). For figures, see M.Thatcher, ‘Regulation after Delegation: Independent Regulatory Agencies in Europe’, Journal of European Public Policy 9/6 (2002), pp. 954–72. 43. For example, the head of the ART, Jean-Michel Hubert, a Chirac loyalist, was chosen over the very experienced and respected Bruno Laserre, who had headed the regulatory unit within the PTT ministry for several years. 44. T.Prosser, Law and the Regulators (Oxford: Clarendon Press 1997); M.Thatcher, ‘Regulation, Institutions and Change: Independent Regulatory Agencies in the British Privatised Utilities’, West European Politics 21/1 (1998), pp. 120–47; C.Hall, C.Scott and C.Hood, Telecommunications Regulation (London: Routledge 2000). 45. For instance, it was important in decisions over privatisation, liberalisation and licensing in telecommunications and electricity. 46. The choice of an auction element in UMTS licensing, against earlier suggestions that a pure beauty contest would be used—the authority insisted that a more open competitive system be used. 47. L.Hancher and M.Moran (eds.), Capitalism, Culture and Regulation (Oxford: Oxford University Press 1989). 48. Cf. G.Majone ‘The Rise of the Regulatory State in Europe’, West European Politics 17 (1994), pp. 77–101; and ‘From the Positive to the Regulatory State: Causes and Consequences of Changes in the Mode of Governance’, Journal of Public Policy, 17/2 (1997), pp. 139–68. 49. Thatcher, ‘Regulation, Institutions and Change’; idem, The Politics of Telecommunications; B.Carsberg, ‘Injecting Competition into Telecommunications’, in C.Veljanovski (ed.), Privatisation and Competition (London: IEA 1989). 50. For France and Italy, see Conseil d’État, Rapport Public 2001; Giraudi and Righettini, Le autorità amministrative independenti. 51. European Policy Forum/Hansard Commission, The Regulation of Privatised Utilities (London: Hansard Society 1996). 52. Cf. Majone, ‘The Regulatory State and its Legitimacy Problems’. 53. For Britain, see Forman-Peck and Milward, Public and Private Ownership of British Industry; Foster, Privatisation, Natural Monopoly and the Regulation of Natural Monopoly.

The Unanticipated Consequences of Creating Independent Competition Agencies

STEPHEN WILKS with IAN BARTLE This contribution is informed by four features of competition policy agencies in Europe. First, such agencies have become ubiquitous over the past 20 years; second, all such agencies have been granted more or less independence as non-majoritarian institutions; third, delegation to agencies originally displayed a strong element of symbolism; fourth, in practice agencies have defined new roles with real impact on the industrial economies of the countries within which they operate. The extent to which governments have created independent agencies to administer competition law and policy is quite remarkable, both in its consistency and in the depth of delegation. There are available a series of theoretical explanations as to why such levels of independence should have been established, what is less well theorised is the consequences of such delegation. Since competition agencies have by now built up a considerable history of operation it is possible to evaluate the ways in which they have exploited their independence. This chapter argues that the original decisions to delegate, and the design of the agencies, were motivated by a need to reassure and to appear to act. The agencies had a strong symbolic element and had a ‘constitutional’ significance. They were not expected to be very active in developing policy or in its implementation. As time has passed, however, the agencies have become more activist and have contributed to policy through a demonstration effect, showing what can be done; as a source of technical expertise; and as an available agency of implementation to be enhanced or adapted by subsequent governments. The consequence of creating agencies has been to populate the policy area with actors (agents) who have their own priorities, interpretations, and influence. It is therefore possible to analyse the working out of a classic principal-agent relationship in which the agents have become entrenched and influential, although they have also responded to the changing preferences of changing principals. How much of policy development is due to agency independence, and how much to the evolving utilities of principals, is a matter returned to in the conclusion. This introductory section reviews the spread of competition agencies across Europe. Study of the operation of the competition agencies parallels in the micro-economic field the delegation of power to the central banks analysed by McNamara, and is of at least equal importance. The chapter examines the main regimes in Europe—the European

The politics of delegation

132

Union itself, Germany and the UK. The second section outlines the pattern of delegation to competition agencies and picks up various dimensions of the concept of independence. Section three examines the ‘decision to delegate’; section four reviews the creation of agencies and the institutional design of delegation; and section five looks at the consequences of delegation. Section six comes back to the pattern of the delegation of powers and the unanticipated consequences. Competition policy came to the fore in Europe during the 1980s in parallel with the decline in Keynesianism and in nationally based industrial policies.1 As the conventional industrial policy instruments of nationalisation, planning, selective intervention, and subsidy were abandoned by Western governments, so a new portfolio of privatisation, deregulation, incentivisation, and market-friendly re-regulation took their place. The new policies rested upon and exploited the market and, as they became more central to the philosophy of late twentieth century government, so the market itself became a dominant concern of policy makers. In this environment competition policy, as a generic regime of horizontal regulation, designed to protect the workings of the market, became steadily more central to economic policy. The governmental principals became more positively disposed to the substantive implementation of competition laws and the agencies experienced a more supportive environment. For the EU itself, competition policy had an added allure. Competition policy is aimed at extending the operation of the market within and between economies. It is antagonistic to sectoral and national market segmentation and is hence a vehicle to facilitate the globalisation of markets and market integration. The priority they attached to market integration inspired the authors of the Rome Treaty to build competition policy into the design of the Commission, and prompted the single market evangelists to facilitate the expansion of the competition competence in the run-up to 1993. The exploitation by the European Commission of its hitherto latent power to shape the European economy through the competition rules has produced something close to a revolution in the regulation of European industry. The growth in influence of DGIV (the Directorate General for Competition) has overshadowed the BKA (Bundeskartellamt). It has catalysed a transformation in the UK competition regime and has spawned the creation of competition agencies in every member country. Thus in 1986 the French created a new independent authority, the ‘Conseil de la Concurrence’. In 1990 Italy created a new five person independent Competition Authority with Giuliano Amato as President. In 1993 the Dutch enacted a new competition law with an independent agency modelled on DGIV, and in the same year Sweden replaced its former competition ombudsman with a new competition office again modelled on DGIV.2 This orgy of policy borrowing extends to the aspirant member states who are required to create competition agencies. The Europe Agreements negotiated with the Central and Eastern European countries require them to establish competition regimes and agencies and we saw, during the 1990s, the sprouting of new independent competition authorities from Warsaw to Bratislava.3 Competition policy is a fact of European life and the operation of competition regimes affects daily lives to a remarkable degree—from the football we watch to the price of cars and choice of holidays. Evaluation of competition policy is central to modern ideas of economic freedom and economic power as well as to assessment of the economic prosperity that the exploitation of market efficiencies should generate.

The unanticipated consequences of creating independent competition agencies

133

DELEGATION AS A FEATURE OF COMPETITION POLICY There has been a tendency for legislatures and executives to opt for extreme delegation in the operation of competition policy. In the terminology of our approach such agencies enjoy a substantial ‘zone of discretion’. In the case of American antitrust the zone of discretion is large and is exercised by the Federal Trade Commission created in 1914. The FTC provides something of a benchmark against which to evaluate European independent agencies. Its independence is ensured by the ‘Commission’ formula with five commissioners appointed for seven year staggered terms. It is required to be bipartisan (with commissioners from both parties) and is responsible to Congress. Like all regulatory commissions it combines the roles of legislation, policy making, and administration. It is one of the components of the regulatory ‘fourth arm of government’ that has so exercised American regulatory theorists and fuelled the principal-agent debate. The limits on the independence of the FTC include the standard budgetary needs, the political influence of presidential appointments and, above all, the control of the courts. There is also the parallel presence of the DOJ Antitrust Division. In effect the USA has two anti-trust agencies, the DOJ and the FTC. Although they administer different legislation their jurisdictions overlap extensively and are co-ordinated by networks and agreements. They present an intriguing experiment in institutional design. The expectation would be that the DOJ would be far more responsive to presidential priorities. The FTC, by contrast, was designed by the Progressives to be neutral, expert, consistent, and to act as a ‘depoliticised’ guardian of a competitive market order.4 The European interest in delegation to such non-majoritarian institutions is relatively recent. It has been stimulated by the path-breaking work of Majone and by the ‘democratic deficit’ in Europe which, Majone suggests, requires a more sophisticated assessment of the legitimacy of regulatory agencies.5 The competition agencies are independent to varying extents but are not separate independent agencies in the formal sense of the American regulatory commissions or the new breed of European agencies, the so-called ‘ten agencies’.6 As regards their non-majoritarian status, insulation from control by elected politicians is mixed. In the UK, the Competition Commission has clear political independence, the main political influence coming from the appointment of the chairman and the members by the Secretary of State for Trade and Industry. The OFT is a ‘non-ministerial government department’ and is also formally independent, although here again the minister appoints the Director General and the DTI has influence over staffing and the budget, and is influential in decision making, especially over merger control. In Germany, the BKA is a federal agency under the supervision of the Ministry of Economics. Some observers would think it conceptually problematic to regard it as a ‘non-majoritarian’ agency. However, given the de facto autonomy of the BKA, it seems acceptable to analyse it an a non-majoritarian organisation and this is certainly the approach adopted by Majone.7 For Europe, the Commission is conventionally regarded as non-majoritarian, although it is worth recalling that Commissioners are national politicians and are susceptible to pressures exerted by their home countries. Thus, here too, majoritarian influences are not insignificant and DGIV can been ranked, along with the OFT, as enjoying only medium independence from majoritarian influences. In the following section the rationale for delegating independence to agencies is examined in detail but it is necessary to examine first what is meant by ‘independence’.

The politics of delegation

134

The primary axis of regulatory independence is freedom from control by government and hence by political parties. When agencies are said to be ‘independent’, it is political independence that is usually meant. Political independence is intended to reduce improper lobbying and the pursuit of party political advantage. It is a guarantee of impartiality, fairness, and consistency which is typically seen as essential if the public is to have confidence in an agency. By implication the emphasis is put on ‘expert’ criteria based on objective principles, be they technical, legal, or economic. For economic regulation, political independence may also be seen as a proxy for independence from business. Since business can mobilise massive financial and political pressures, politicians are especially susceptible to business lobbying. The real argument for the independence of competition agencies is to insulate them from the unholy alliance of politicians and business. It has also been recognised, however, that formal independence is often a facade. The sociology of regulation stresses the risk of ‘capture’ where regulators become too close to, and dependent on, the regulated.8 In competition policy, because of the episodic and exceptional nature of regulation, this is less likely, but there are other informal sources of influence. Horn observes that ‘the single most important dimension along which these [regulatory] institutions vary is their degree of independence from the political process’,9 and he puts unusual stress on the way the political process affects the labour market and the employment calculations of key decision makers within the agencies. In addition to the rather obvious distinction between formal and informal independence, the competition arena raises the question of symbolic or constitutional independence. On a day-to-day basis it is often convenient for politicians to be able to point to a competition agency as a symbolic safeguard against monopoly or restrictive practices. Since abuse of market power is endemic in every market economy and is indeed a major component in the competitive strategies of most large companies, the elimination of anti-competitive practices (rather like the elimination of racial intolerance) is an impossible goal. Yet, as with racial equality, it is essential for governments to confirm support for free competition. This role has been projected onto the European level. As we noted in 1996: ‘This role of DGIV …gives it and the competition rules almost a “constitutional” status…. Just as the prospective European central bank will guaranteed sound money across Europe, so the Competition Authorities will guarantee a sound market.’10 From this perspective the competition agency should have an independence that goes beyond formal jurisdiction and the pursuit of given policies. It requires an independence of mind and embodies a philosophical or ideological commitment to the benign operation of the market economy. To prejudice this independence could thus be seen as ideologically unsound. From this perspective the granting of independence to competition agencies comprises a statement about the ideological mission of the state. EXPLAINING THE DECISION TO DELEGATE There are competing theoretical perspectives on the reasons why governments establish regulatory agencies and why they delegate to them particular combinations of powers. Some of these theories are as old as the study of modern administration. Thus the creation

The unanticipated consequences of creating independent competition agencies

135

of independent regulatory agencies by the Progressives in the United States was informed by the theories of Woodrow Wilson and his contemporaries about the separation of politics and administration, about the importance of technical expertise, and about the need to create neutral agencies free from corruption. Succeeding generations of administrative theorists, on both sides of the Atlantic, studied the dynamics of control and accountability, efficiency and equity, using tools ranging from Weberian neutrality to cybernetics and administrative culture.11 In the German tradition delegation was defined and circumscribed by the commitment to the Rechtsstaat, the adherence to an all encompassing system of laws sustained by the administrative courts. In the British tradition delegation was controlled in a framework of trust resting on a culture of commitment to the ‘public interest’ shared by the generalist administrative class. Hood et al. point to the persistence of the British model of control which they characterise as ‘a relatively non-litigious judicial environment, a pattern of high compliance with light-rein control, a lack of elaboration of formal legal rules’.12 These contrasting administrative traditions continue to influence agency design.13 Although it shares an interest in ongoing control, the challenging ‘positive politics’ literature, based on rational choice theory, has focused to a greater extent on the initial decision to establish an agency. The focus on the ‘moment of creation’—the ‘enacting legislation’—allows rational choice theorists to ask rigorous analytical questions. The questions are acutely political in that they seek to ascertain the calculations of political gains that motivated the progenitors. This focus on the decision conveys a potential to reveal the ‘genetic code’ of the agency and the law. A code which may pre-determine later developments. Perhaps traditional analysts, and even historians, might accept the metaphor of the genetic code14 (see also Döhler, this volume). What is more provocative, and analytically challenging, is the idea that legislators embark on a conscious decision both to pre-determine the later behaviour of the agency, and to limit the ability of future generations of legislators to change that behaviour. In a stimulating analysis Macey suggests that agencies can be ‘hard-wired’ to favour certain interests.15 The rational choice literature suggests that this can be undertaken in a calculating fashion that trades political, electoral support from influential groups for access to agencies or predetermined agency behaviour. The more effectively an agency can be ‘hard-wired’, the greater the price legislators can extract from interested groups. The creation of an agency can thus be seen and analysed as an exercise in institutional design that not only resolves current problems, but which pre-determines future patterns of action in ways which favour the creators and their constituencies. Analysis of this creative process has been taken forward in the field of institutional economics by writers such as Horn,16 who has advanced a rigorous, transaction cost-based approach to institutional design. At the core of this approach is the concept of ‘credible commitment’. As noted in the introductory chapter, commitment problems arise when an enacting coalition is unable to guarantee the durability of future benefits. There is always the risk that future legislators will supercede legislation and that future administrative action will reverse current arrangements. The enacting coalition can reduce that risk by delegating authority to an agent that is relatively independent from legislative or bureaucratic influence.17 But in order to maximise beneficial effects the agencies must be designed in such a way as to minimise divergencies from the original distribution of benefits. In a curious way, therefore, Horn is analysing the deliberate creation of the ‘path dependence’

The politics of delegation

136

that in a different theoretical setting would be associated with the assumptions of historical institutionalism. For Horn, of course, path dependency is consciously constructed on the basis of a rational calculation of advantage. Nonetheless, there is an interesting confluence of ideas here which at least agree that originating structures do dictate future choices. There is thus a quite different emphasis employed by historical or sociological institutionalists. From this perspective rationality is bounded and choices are limited. When agencies are created, or legislation enacted, the participants often have very limited room for manoeuvre. They are constrained to pursue standard, orthodox, and ‘appropriate’ choices in a process of path dependency.18 Institutional theory offers a persuasive analysis of constraints which can be summarised in a trilogy of categories: the cognitive (legislators only conceive of a limited set of possibilities); the regulative (only a limited range of organisational forms are available, means triumph over ends); and the normative (only certain choices are acceptable in a given ethical, cultural, or ideological setting).19 For rational choice theorists and institutionalists alike, the enacting legislation is an important source of continuing influence, and the actions of the legislators, conscious or unconscious, rational or perverse, self-seeking or altruistic, are of fundamental importance. One important question here is ‘who are the legislators’? In the American political system, where the institutional economics literature has blossomed, it makes sense to analyse congressmen and to evaluate their electoral motivation. In the UK system of fused executive/legislative actors, mobilised by political parties, the targets of analysis are more complex and considerably more opaque. Thus the Weingast and Moran study of congressional control over the FTC20 could not have been replicated in the UK. Congressional control over budgets, new legislation, and appointees would in the UK be under the control of various ministers and their civil servants. Major legislative enactments in Britain and in Germany rest heavily on the bureaucracy which will define options, negotiate choices, draft legislation, and help to pilot it through the legislative chambers. The analytical equivalent of the congressional legislators lies somewhere between the Cabinet Committee on new legislation and the ‘Bill Team’ of ministers, civil servants, and advisers who work to pass the legislation. The positive politics approach thus provides welcome stimulation in its emphasis on agency design, which highlights the politics of the process. These emphases can be borrowed without taking on board the analytical certainties of the rational choice approach, which, in any case, is far more difficult to apply in the European setting. One common factor, however, is the use of principal-agent (P-A) theory which can be employed pragmatically. In his eclectic account of regulation and delegation, Mitnik employs the P-A approach to good effect, suggesting that ‘the science of politics is in the broadest sense the study of agency relations’.21 His account of the multiple reasons which lead to delegation to agents does not include ‘credible commitment’ (a concept which emerged later—although he does stress a ‘continuity’ problem), but he does pick up one aspect that is very germane to competition policy. He points out that legislators may not want the task performed at all. They may have made a collective decision to authorise the task because of pressures from constituents and wish to appear to favour the action. To resolve this dilemma, legislators

The unanticipated consequences of creating independent competition agencies

137

can deliberately delegate the activity to an agency that will not act or, if it does act, will do a minimal or poor job.22 Here he is picking up the agency dimension of ‘symbolic politics’. This builds on the well-established understanding that political acts can be purposive and goal orientated, or they can be ritualistic and reassuring. They can, in other words, be instrumental or expressive. In his review of symbolism in politics, Edelman cites antitrust as a prime example of symbolic politics.23 In a typical use of paradox he asserts that it is not unusual in antitrust decisions to give the rhetoric to one side and the decision to the other, and he pleads in support Thurman Arnold’s view that ‘the actual result of the antitrust laws was to promote the growth of great industrial organization by deflecting the attack on them into purely moral and ceremonial channels…every scheme for direct control broke to pieces on the great protective rock of the antitrust law’.24 We come back to the employment of symbolic politics in the creation of the European agencies. In reconciling the P-A and sociological approaches, Stone Sweet and Thatcher (Introduction) suggest that, when making decisions about delegation to non-majoritarian agencies, elected politicians behave rationally but select from a menu of choices that has emerged from wider social processes. In some cases, as in Germany and, to a lesser extent, in the UK, the enacting legislation came into play only after a wide and exhaustive survey of a substantial menu of choices. In later cases, for Southern and for Eastern Europe, choice was conditioned by emergent standard practice exemplified in foreign models. The menu became shorter, very much more specific, and defined by EU norms. This suggests a possible life cycle of agency innovation whereby in the early days of developing a policy area the menu of choice is wide, and becomes steadily narrower as politically acceptable or functionally efficient models become enshrined in EU practice (and perhaps in EU legislation). Stone Sweet and Thatcher suggest four ‘functional logics of delegation’. Here, five possibilities are reviewed. First is the completely conventional need to harness expertise. Operation of competition policy and law is relatively technical. Cases are complex and have to be analysed by industrial specialists, by economists, and by lawyers. This is a classic argument which stresses information asymmetries and the efficiency of rule making. Second is a perceived need to insulate the administration of competition law from party political interests. The reasons for this include credibility, legitimacy, neutrality, and consistency. The credibility argument is based on the need to make clear to the regulated community and the public that the agency is not subject to partisan pressure from politicians. In P-A terms it constitutes ‘credible commitment’. The same argument applies to central banks who have to be seen to be acting according to neutral economic data and not in pursuit of political advantage. Only by ‘de-coupling’ administration of policy from the political process can it take on long term credibility as a fixed programme of action. The ‘legitimacy’ argument is different and relies on the ability to appoint members to an agency who have ‘standing’ in society and who have earned respect for their independent judgement. This is important in British public life. The ‘neutrality’ argument is less to do with political neutrality and more to do with minimising clashes of interest. Again, especially in the UK, an agency under the control of majoritarian ministers would run the risk of regulation being linked to other policy areas. The extent

The politics of delegation

138

of UK intervention in the economy was such that civil servants preoccupied with regional policy, exports, or employment would have found it impossible to disassociate those priorities from their operation of competition policy. Finally, ‘consistency’ is perhaps another way of stressing the commitment problem but it is especially important in the British two party system where alternation of strong party government is regarded as desirable and has historically carried almost an assumption that controversial policies will be reversed. A third motive for delegation is in order to respond to demands from business. The business view of competition law is Janus-faced. Business is often a strong proponent of regulation. It provides certainty, it helps to control irresponsible competition and it opens up a realistic possibility for business to control the agency concerned (the possibility of ‘capture’). This motive can be variously analysed. It is consistent with theories of the structural power of business, with theories of business lobbying, or with an ideological basis of business-friendly political parties seeking to pursue ideological goals. More or less cynical economic theories of private interest regulation go back in the United States to at least 1940,25 well before Stigler popularised the view that regulation operated primarily in the interests of the regulated industries.26 As we have seen, this motive can be seen to generate a ‘symbolic’ response. It seems that governments have wanted to signal a commitment to the market and to free competition without being willing to attack the monopolies, oligopolies, and restrictive practices which often had pragmatic uses and were strongly defended by vested interests (and, indeed, by respectable economic theory). The perfect solution was thus the ineffective agency. Eisner questions whether in the USA ‘some of the early initiatives were anything more than symbolic efforts designed to assuage the demands of mobilised constituents’.27 Precisely the same question could be asked of the first 20 years of UK policy or the first 20 years of European policy. Certainly, in all regimes, organised business consistently worked to weaken the implementation of competition policy—the other side of the Janus face. A fourth reason to delegate to agencies is the opposite of the symbolic rationale, a concern that competition policy would be administered effectively and would become hugely unpopular. Sometimes known as ‘blame avoidance’, it is an ancient human reaction, labelled, appropriately enough, the ‘Pontius Pilate principle’ by Hood,28 it is present in all competition regimes. When powerful companies and industrial interests come clamouring to politicians or bureaucrats it is immensely helpful to assert agency independence just as they would assert judicial independence. Fifth is a constitutional argument, which stands outside the scope of institutional economics analysis. Independent competition authorities have been regarded in some countries as of quasi-constitutional importance. They guarantee a market system in countries that have had a record of the state over-riding the market. This applies to Germany, Japan, and Korea, as well as the former Communist countries of Central and Eastern Europe. To some extent this is a further argument about symbolic politics, but it is an argument of a different nature. The constitutional argument is not about ‘policy symbolism’ but about grand principles of political philosophy. Just as independent courts symbolise the rule of law, so a competition agency symbolises commitment to the free market. It is an important ideological statement which lies at the heart of the European project and reaches a peak of importance in Germany, as we see below. In the United States, antitrust tradition also symbolises the role of the competition agencies in

The unanticipated consequences of creating independent competition agencies

139

sustaining economic democracy as going hand in hand with political democracy. The following section turns to specific national cases. CREATION OF AGENCIES: THE CASES Analysis of the delegation of authority to non-majoritarian competition agencies starts with the enacting legislation. The analysis is complicated by the fact that the originating legislation is now relatively old. Reconstructing the reasoning of the legislators, discovering their incentive structures, and the associated transaction costs, requires historical research. This is, however, an exercise worth undertaking. Although each policy area has changed significantly during the growth of the competence, and each area has been amended by subsequent legislation, nevertheless, each agency remains in place. UK Enactment The first British agency, the Competition Commission, can be traced directly back to the Monopolies and Restrictive Practices Commission Act 1948 and, up to 1999, was known as the Monopolies and Mergers Commission. The Commission has operated as an investigative and administrative appeals body and employs part-time independent commissioners. The parallel policy making and prosecuting body is the Office of Fair Trading (OFT) headed by a Director General who is a public official with a considerable degree of independence. The OFT is the central agency in a British model of competition policy which is quite distinctive. The model has been ‘administrative’ with no recourse to the courts in relation to oligopolies, monopolies, and mergers. The system rested on persuasion and negotiation in pursuit of a definition of ‘the public interest’. The British administrative machinery has always been highly pragmatic, unwilling to condemn monopoly, and suspicious of legal purism and the prohibition approach. But while the British system did not have the prominence or the legal coherence of the American or German systems, it did share with them the qualified support of industry. British industry was in favour of the OFT, supportive in principle of the Competition Commission, and very hesitant about the reform of British policy towards a European-style prohibition approach in the 1998 Competition Act. The origins of the Competition Commission encapsulate a perennial problem with applying theories of rational calculation to enacting legislation. It is the problem of disentangling the role of politicians and that of civil servants in a fused executive. Historical research29 indicates that the design of the Act and the agency originated in a paper written in 1943 by Hugh Gaitskell, then a temporary civil servant, later Leader of the Labour Party and for many ‘the best Prime Minister Labour never had’. Gaitskell advocated a statutory investigatory commission, but with powers of decision remaining with ministers. The Commission was to have minimal levels of discretion and its influence would rest on advice, analysis, and publicity. Gaitskell’s emphasis on ministerial decision was, in the British context, also a code for reinforcing civil service discretion. The Act was eventually passed by a Labour government in 1948 with Harold Wilson the architect as President of the Board of Trade. Wilson has been seen as unduly influenced by business in the design of the Commission.30 Certainly he and his officials

The politics of delegation

140

were concerned to maintain good working relations with industry, and the Commission was structurally biased in favour of industry by the investigatory principle (no presumption that monopoly was ‘bad’), by having a workload controlled by civil servants, and by the part-time principle (Commission members were independent but marginalised). There is therefore room to suggest that the enacting legislation created a ‘business-friendly’ organisation; and similarly that Wilson created a ‘symbolic’ body which diverted pressure for more forceful legislation. The second, and more significant, British agency, the Office of Fair Trading, was created by the 1973 Fair Trading Act passed by the Heath government. Some considerable debate and negotiation went into the design of the OFT. While the Competition Commission could be seen as a variant on the well established theme of the administrative tribunal, the OFT was in contrast unusual and distinctive. Rational choice analysts would regard the main features of the design as calculated to ‘hard wire’ the agency to seek certain goals and to maintain the benefit structure intended by the enacting legislation. It is useful to draw from Horn’s lucid categorisation of the dimensions of institutional choice. The major choice variables he identifies are: (i) extent of delegation to the administrative level; (ii) governance structure of the administrative agent; (iii) rules imposed on the agent which dictate third party access; (iv) the nature of subsequent legislative monitoring; (v) provision of capital (sales or grants) and labour (employment rules). This is a useful definition of choices but it is necessary to adapt the categories to concentrate on delegation, expertise (rather than governance), access, and the labour market. The question of legislative oversight is far less significant in Europe than in the United States. Thus the questions become: (i) how much independence was the agency given? (ii) what access biases were built in? (iii) what labour market biases were built in? (iv) what technical/expertise biases were built in? In answer to the first question, the agency was given very considerable independence. Substantial powers of consumer protection and prosecution of competition policy were vested in the Director General of Fair Trading. The DG was appointed for a five-year term, renewable by the Secretary of State for Trade and Industry, and was mandated ‘to keep under review the carrying on of commercial activities in the UK’31 in order to implement consumer protection and competition measures and to advise the minister. The Director General idea was novel, momentous, and devised by civil servants. This was a named individual, not a politician or a civil servant, intended to be an independent public figure and with authority to recruit staff to his own office. This is the model of the British single person regulator which was used as the blueprint for the utility regulatory offices and the regulators appointed from 1984 onwards (see Thatcher, this volume).

The unanticipated consequences of creating independent competition agencies

141

In the British system of government, which is fixated on the concept of ‘ministerial responsibility’, full independence for any officials other than judges is difficult to establish. The independence of the DG was distinctive but not complete. The DGFT is appointed by the minister (not the crown) and some of the most significant powers were to be exercised in consultation with the minister. Interestingly, with something of a parallel with the reserve powers of the German Economics Minister, the Act also gave powers of political over-ride to prevent monopoly investigations (these powers have never been used). The civil service was very aware of the sort of powers they were conceding. Independence was regarded as important partly for procedural reasons, in order to insulate the machinery from accusations of mixed motives (the neutrality argument mentioned above) but independence was to be qualified. The DTI civil servants who drafted the Bill discussed their design with their opposite numbers at the Civil Service Department. A senior official from the CSD underlined the fact that the ‘appearance’ of independence was the crucial factor: Should the Director General be Completely Independent? The arguments point two ways. You want to give him the appearance of considerable independence. In part this is to avoid any suspicion that the Department’s ‘sponsorship role’ puts it in the pocket of the industries sponsored. On the other hand you want to retain ministerial control over the making of References… Given this degree of control my own view—and I think you agreed when we discussed it—is that the appearance of independence can be adequately ensured by setting up a separate Office and there is no need to strengthen it unduly.32 Even accepting that the independence of the OFT from the minister and from civil servants within the ministry was limited, how did the legislators influence how the OFT would use its independence? The second question addresses the access biases built into the legislation. The FTA provided for a ‘Consumer Protection Advisory Committee’ and hence seemed to anticipate a privileged input (which never materialised). It also required the office to work with the MMC as well as with the minister in arrangements that became known as the ‘tripartite system’. This represented a division of powers unusual within British government with the executive function split between the DGFT and the ministry, limited rule making powers residing in the office, and the appeals, or quasi-judicial function allotted to the MMC. A third question focuses on labour market biases. The Act was open-ended but the arrangements encouraged civil service staffing. With the majority of the senior staff coming from the DTI, a predisposition to conform to civil service standards and priorities was built into the office. A fourth aspect can be seen in the technical biases of the office. Although in practice it was clear that lawyers and economists would be required, there is no stipulation as to the size, tenure, or expertise of the staff. This left it open for staff to be generalist civil servants, professionals, or people seconded from a business career. The complex evolution of the agencies and the 1973 Act is analysed at length in a recent book.33 Since this was an industrial regulatory enactment, the biggest interest

The politics of delegation

142

constituency was organised business, hence the CBI was deeply involved in the design and legislative process. The CBI was keen on the DGFT/OFT formula. It wanted insulation from politics, it was exceedingly anxious to exclude the courts from the operation of policy and it wanted to retain appeals to the MMC. Ministers moved some way, but certainly not the whole way, towards the CBI position, but it would be hard to point to direct electoral calculations or payoffs. It would be especially difficult since the animating force in facilitating, designing, and passing the legislation was the civil service. The creation of the OFT was certainly helpful for civil servants in the DTI. al lowed th em to t controversial and difficult activities, which created conflict with their industrial constituents, out of the ministry and into the office. From their point of view, the creation of the OFT was beneficial for the civil servants involved and on a general level it was beneficial for relations between Conservative ministers and organised industry. The level of industrial support was consolidated by the appointment as the first DGFT of a leading industrialist, John Methven, who had himself been a prominent industrial lobbyist who had pressed for some at least of the provisions which were incorporated into the Fair Trading Act. German Enactment After nearly a decade of impassioned debate, West Germany passed the GWB (Law Against Restraints on Competition) in 1957 as a post-occupation measure strongly supported by Ludwig Erhard, then Economics Minister. The Act was widely seen as a victory for industrial interests. Although based on a prohibition principle, the law had many sectoral exemptions, it excluded merger control (which was to follow in 1973), and allowed many cartels. The Act created an independent Federal Cartel Office, the BKA (Bundeskartellamt). The first draft of the GWB (Gesetz gegen Wettbewerbsbeschränkungen) was presented to Erhard in 1949. This ‘Josten draft’ proposed a major law based on ordoliberal principles, embodying a concept of administrative control, and based on ‘the creation of an independent monopoly office that would strictly enforce a set of norms designed to combat harmful uses of economic power and protect “performance competition”’.34 Ludwig Erhard was CDU/CSU Economics Minister from 1949 to 1963 and the GWB was a central element in his vision of the social market economy. The creation of the GWB went hand in hand with the creation of the social market economy and was a long drawn out affair. The concept of Ordnungspolitik developed during these years rested on something close to an economic constitution which would constrain government as well as private economic actors. The ordo-liberal debate that underpinned the GWB emphasised the role of Ordnungspolitik as an abstract and universal commitment to market principles guaranteed by the rule of law. Under this philosophy government had no role of detailed intervention in the economy, which should instead be regulated by market principles of free competition policed by an independent agency. That agency was the BKA, which in a peculiar way was seen almost as ‘outside’ or ‘above’ government. In Gerber’s interpretation of this philosophy,

The unanticipated consequences of creating independent competition agencies

143

competition law is not merely a pragmatic tool for achieving specific policy ends, but an important part of a society’s “constitution”. It is a matter of long term legal and economic structures… In Germany, legal principles and methods became the central dynamic of the story rather than policy and administrative discretion.35 Gerber therefore suggests that independence is almost metaphysical. The job of the BKA is to safeguard the integrity of the market system itself and not merely to secure particular administrative efficiencies. In P-A terms, therefore, the German independent competition agency can be seen as a quasi-constitutional solution to the problem of credible commitment. The commitment problem faces legislators with a need to bind future governments and to embed political authority in institutional arrangements. This is precisely the exercise upon which Erhard and the ordo-liberals were engaged. They sought to establish a lasting economic constitution that would limit the ability of future governments to intervene in the industrial economy. A recent source observes that ‘it is important to note that the government itself is also constrained by these rules’.36 What is less clear is whether Erhard was motivated by calculation of electoral advantage. In the eight years it took to pass the GWB, the government made a number of substantial concessions, mainly to business interests. Business in Germany was unpersuaded by ‘American-style’ antitrust and preferred traditional cartel arrangements. In Erhard’s finally successful enactment, the BKA was created as an independent federal agency under the supervision of the Economics Ministry. The minister appoints the president and can give general or individual directions to the office although the general directions have to be published in the Federal Gazette (in practice, directives are almost unheard of). This pattern was repeated in the second amendment to the law, which introduced merger control. The minister was permitted to over-ride decisions on merger cases and has done so in a small number of very controversial cases. Ambiguity continues as to the precise constitutional ability of the ministry to control the office but in practice the BKA has developed extensive independence. This is symbolised by physical separation. While the government was in Bonn the BKA was located in Berlin. With the creation of the new capital in Berlin the BKA has shifted to Bonn, with the curious sideeffect that appeals now go to the Dusseldorf Court of Appeal rather than to the Berlin Court. There are thus some clear structural similarities between the BKA and the OFT. It is interesting to see whether there were also clear biases in organisational design and, as with the OFT, the discussion focuses on access, the labour market and expertise. No special privileged access to the BKA was created by the enacting legislation. But since it was located under the supervision of the Economics Ministry it was placed in a specific legal context with organic links to the civil service. In order to emphasise its quasiconstitutional status, appeal against BKA decisions is to the civil and not to the administrative courts.37 Legal challenge is therefore facilitated, which opens up avenues of access for industry. Certain interests were privileged in another sense by being excluded from the coverage of the law. Long and bitter debates resulted in concessions which excluded such important industries as transport, agriculture, insurance, and energy from the exercise of the cartel provisions. Further, small and medium sized enterprises

The politics of delegation

144

enjoyed substantial concessions and the BKA has been sympathetic to the needs of the strong German Mittelstand.38 The labour market influence tied the BKA into a set of German civil service norms. The first staff were simply transferred from the Economics Ministry (just as OFT staff were transferred from the DTI). The first president, Eberhard Guenther, served until 1976. He was trained as a lawyer and had an evangelical belief in the social market, but he and his successors all came from the competition sections of the Economics Ministry.39 For presidents, therefore, the BKA comes as the apogee of a civil service career. For other staff, the legalistic German civil service provides the incentive structures. The technical bias of the enacting legislation was to emphasise a legalistic approach. This came in a number of mutually reinforcing effects. The whole thrust of the ordoliberal argument was to elevate the status of law and to regard it as superordinate. The BKA’s legal remit would protect it against political intervention. This was reinforced by the German administrative tradition which emphasises legal process and a legal training. The whole was consolidated by the necessity of preparing cases and decisions in such a way that they would survive scrutiny by the civil courts. European Enactment A further European model of delegation and independence is offered by the European Commission. Competition policy is a unique European competence because Articles 85 and 86 of the Rome Treaty (now Articles 81 and 82 of the Treaty of Amsterdam) require the Commission to deal directly with industry. This obligation was given effect by Council Regulation 17 of 1962, which must rank as one of the most important ever enacted. Regulation 17 gives the Commission direct powers to investigate, decide, exempt, and punish breaches of the competition rules. Moreover, Regulation 17 gives the Commission a superior role over the national competition authorities by giving it a monopoly over the operation of the crucial Article 81 (3) which grants exemptions from the prohibition on agreements which restrict competition. It is almost certain that the politicians in the Council at the time had little conception of the potential for independent action latent in Regulation 17. Instead it lay like a slowly ticking bomb for over 20 years until it exploded in the faces of the national governments during the 1980s as the Commission developed competition policy as an influential component of European economic regulation. In this case, the route of delegation is ‘upwards’ to the Commission as a supra-national body. In turn, the Commission entrusts competition to Directorate-General IV, currently headed by Mario Monti (who succeeded Karel Van Miert in 1999). DG IV has always operated with a considerable degree of administrative and formal independence from both the Council and the European Parliament.40 Nonetheless, major decisions have to be taken or approved by the Commission collectively. This can introduce perverse political dynamics and has aroused strong criticism, especially from the purist Germans who have argued for the creation of a ‘truly independent’ European Cartel Office.41 The European regime is a prohibition system based on administrative action with marked similarities to the German model. Indeed, Gerber is clear that there is a direct parental link between the ordo-liberal prescriptions being developed in Germany in the late 1940s and early 1950s

The unanticipated consequences of creating independent competition agencies

145

and the competition articles incorporated into the ECSC Treaty (Treaty of Paris, 1951) and the Rome Treaty.42 THE CONSEQUENCES OF DELEGATION The thrust of the argument in section four is that the initial decision to establish agencies and to delegate to them had a large symbolic component. The agencies were established with ambitious goals but with modest resources, with conformist staffing and with hidden limitations on their independence. The agencies have had decades in which to develop and the impact of the enacting legislation has inevitably been reduced by time, by further amending legislation, and by the evolving relationship between the agencies as agents and the ministries and government as principals. In each case the agencies had a mission built in to their goals. Crudely, the German mission was legally to defend the market economy, the European mission was market integration, and the British mission was to protect the public interest. The enacting legislation favours some groups and interests and there was an element of ‘hard-wiring’ which favoured certain interests, such as the German energy industries or the British Department of Trade and Industry. All these European Acts also had an element of ‘softwiring’. Rather contrary to an institutional economics explanation the choice of administrative systems (rather than the American judicialised system) meant that the competition agencies could be used as agents of intervention. This aspect of competition policy has always aroused great controversy, but the reality is that governments were provided with one more potential lever with which to influence industry. This can be seen in UK ministerial involvement in merger control, or in the involvement of the whole European Commission in difficult merger or state aid cases. More salient than the continued privileging of certain interests or the pattern of political intervention has been the ability of the agencies to exploit their independence to pursue aspects of their missions. Agencies have had the advantage of continuity of leadership, as can be seen in all regimes, with the prime examples being Lord (Gordon) Borrie, Director General of Fair Trading for 16 years (1976–92); Eberhard Guenther and Wolfgant Karrte, Presidents of the BKA for 18 years (1957–76) and 16 years (1976–92) respectively; and Manfred Caspari and Claus-Dieter Ehlermann as Director Generals of DGIV (1980–90 and 1990–95 respectively). It is an interesting unwritten rule that every Director General of DGIV to date has been German. A full analysis of the developing role of each agency is beyond the scope of a single chapter and the analysis must therefore draw conclusions from pre-existing material. The BKA was able to exploit its potential for independence in the early years thanks to the benign oversight of Erhard as Economics Minister and then as Chancellor. He resisted intervention by the Economics Ministry and established precedents which allowed the office to develop a reputation for fierce independence and a fearless pursuit of competition as a principle. Independence extended to internal procedures and the individual decision making units operate with striking autonomy in respect of the sectors of the economy for which they are responsible. The relative success of the BKA resulted in a strengthening of the law through a series of six amendments, the most important of which was the second amendment in 1973 which established a system of merger control.

The politics of delegation

146

Subsequent amendments and recent legislation, including the important Energy Act which came into effect in 1998, have removed the sectoral exclusions. In recent years, however, the BKA has become steadily more marginalised. The reasons for this include the opening of the German market to European competition, the evolution of economic doctrine towards ‘workable competition’, the growth of DGIV as an alternative and superior body, and the by now taken-for-granted idea of the free market. But the decisive factor has been the overly literal way in which the BKA has pursued its legal mission. In Sturm’s analysis he points to the persistent challenge of the Court of Appeal, which has overturned a number of important decisions. The pattern continues. In 1999–2000, the Berlin court overturned a decision on Lufthansa and the Dusseldorf court overturned a ferry decision (Scandlina).43 Sturm sees the office as too preoccupied with a strict application of the law, on a case-by-case basis, without considering developments in policy or economic theory; hence, ‘what remained was caseorientated intervention, and incremental competition policy without an overall vision and, what is equally important, without broad public support’.44 It would seem, therefore, that the BKA used its independence to good effect in its earlier days and did provide a quasi-constitutional defence of the German social market economy. It did so under a long-serving and determined president, but since the mid1970s it has not exploited its delegated powers to redefine its mission. Instead it has become unduly self-referential and inward-looking. It has become just one more regulatory agency whose effectiveness and hence reputation has been limited by a lack of support from industrial policy makers and case-by-case opposition from industry working through the courts. It seems something of an irony that an agency explicitly established as a bastion of legal certitude should have been undermined by its inability to win cases in the courts. From that point of view, the original decision to put appeals against the BKA into the civil courts was significant and unexpectedly weakened the administration of the law. In the UK, limited delegation in 1948 led to fresh legislation in 1956 and 1973 creating new agencies (a specialised court in 1956, the OFT in 1973). A strong input from organised business was influential but not dominant in the creation of each. The OFT did not exploit its delegated powers to become an innovative consumer protection agency and it found itself limited by continued oversight from politicians and officials in the DTI. Even when public policy and economic sentiment swung towards liberalisation after 1979, the OFT only gradually developed its legitimacy and activism. Nevertheless, a steady if unspectacular record of success through the 1980s brought a far more effective and respected agency by the 1990s.45 An important element in its increased impact was the development of expertise. The British agencies were established on a rationale of protecting the public interest rather than deploying expert analysis. Over the years since 1973, however, they have steadily developed a stronger economic input and used economic analysis, and economic doctrine, to build their credibility. It is significant that the new Chairman of the Competition Commission (Derek Morris) and the new DGFT (John Vickers) are both (Oxford) economists. Curiously, therefore, while the BKA declined in salience the OFT increased. Also, whilst both agencies combine legal and economic expertise the economists have become dominant in the UK agencies. The OFT’s moment arrived with the 1998 Competition Act which has increased its independence, given it powers comparable to those of DGIV, and

The unanticipated consequences of creating independent competition agencies

147

enhanced its legal standing. Both the OFT and the Competition Commission have therefore seen the gradual extension of independence based on a measured exploitation of their delegated powers. Within the British system, which stresses trust and responsible behaviour, the reasonable and co-operative actions of the competition agencies has been met with a steady increase in their delegated authority and a major expansion in their power and status as a result of the Competition Act. The shift from the early symbolic role has been marked and during that shift the relationship with companies and organised business has been both important and ambiguous. Business, individually as well as collectively, has found it difficult to formulate its preferences. The instinct is to argue for a weak competition policy and for agencies that are independent but with good access for industry. But there is also a strand within industry, of a more Thatcherite bent, that has argued for the virtues of a fully competitive economy and has welcomed the freedom of opportunity for entrepreneurs to break into new markets. In the face of divided business the OFT and successive governments have strengthened competition policy to an extent that would not have been anticipated by Harold Wilson in 1948 or Ted Heath in 1973. The consequences of delegation of competition policy to the European Commission represents perhaps the most spectacular case of unanticipated consequences. The intentions of the enacting legislators from the founding six members are opaque and, of course, those countries like the UK, which joined later, had to accept the status quo. It is very doubtful, however, that the enacting legislators has any clear idea of what impact the competition rules would have. Their incorporation into the treaties, rather like the German legislation, had a strong symbolic content. The treaty powers were signalling that Europe was committed to a free market and a competitive, integrated economy, this despite the nationalisation, intervention and planning that characterised the contemporary French economy in particular. For 25 years the competition rules were honoured in the breach. Then came a decade of radical exploitation of the historically dormant delegated powers during the 1980s. The Commission took action against restrictive practices, against cartels and to a lesser extent against monopolies (abuse of a dominant position). Significant fines began to be imposed and an apparatus of control was erected which no substantial European business could afford to ignore. The machinery was brought to a peak of effectiveness by the merger control regulation, eventually passed in 1989, so that by the early 1990s ‘DGIV had transformed itself from a sleepy, ineffectual backwater of Community administration into a formidable machinery for economic integration’.46 The new power to control large European mergers was especially important, but DGIV also moved into utility regulation, control of public monopolies and control of state aid. In these latter areas it was acting against governments rather than against companies and could be seen as fully undertaking the role of credible commitment. The national governments had created an agency that was binding them to observe their free market commitments. The reasons for the exploitation of the powers of DGIV during the 1980s are fairly clear. They lie in political entrepreneurship exercised by Commissioners (especially Peter Sutherland and Leon Brittan), in a series of strongly supportive and teleological judgments by the ECJ, and in a self-confident assertiveness by a new generation of determined young lawyers in the Directorate General. Favourable environmental conditions included the influential neo-liberal commitment to market solutions and the move towards completion of the single market as both a political and an economic goal.

The politics of delegation

148

DGIV expanded its jurisdiction with a fierce independence which rested on legal foundations. The partnership of the competition lawyers and the ECJ proved irresistible, but it generated a legalistic preoccupation that sometimes defied economic logic and which worried both national governments and academic observers. To this extent the original enactment, based on a set of legal principles and enforced by the European Court, had momentous consequences for the future exploitation of delegated powers. The contrast with the German regime was that the ECJ shared the integrative mission of DG IV, whereas the Berlin Court of Appeal acted as a brake on the BKA. A final consequence of the European agency arrangements was to permit an element of political intervention. DGIV is part of the Commission and in the intense lobbying over merger, cartel, and state aid cases the Competition Commissioner has occasionally been over-ruled. This has provoked huge controversy and has created a new problem of credible commitment. The logical response, re-emphasised recently by Majone,47 is to press for more effective delegation to a fully independent agency in the form of a European Cartel Office. This would truly be the counterpart of the European Central Bank. CONCLUSIONS This chapter has sought to explain a consistent pattern in the administration of competition policy which is expressed in the delegation of powers to independent competition agencies. There are both conventional and unconventional explanations. The conventional explanations include a need to create expertise, to allow flexible administration to cater for an uncertain future, to achieve blame avoidance, and to create ‘credible commitment’. The unconventional explanations lie in the symbolic component of agency design and also in a ‘constitutional’ component. The symbolic component would be shared by agencies charged with equally impossible tasks, such as equal opportunities and environmental protection. Government motives in establishing agencies which are substantially symbolic may appear cynical, but it is a familiar political device. The constitutional component is genuinely unusual. Historically for the American and the German regimes, and nowadays for the European and UK regimes, the competition agencies are important guarantors of equality of access to the competitive market. Further, they proclaim that the state itself is committed to a market system and, to this extent, the agencies are constitutionally significant and relatively immune from abolition. The chapter has also charted the consequences of delegation. It notes that delegation had its limits and that the agencies studied are considerably less independent than is commonly supposed and far less independent that the American independent commissions. Their independence is relative within a larger administrative apparatus but in each case a remarkable degree of continuity, stretching over decades, has helped to expand the level of independence. The consequences of delegation are both internal and external. Among the internal consequences, the agencies have built up expertise and become significant reservoirs of legal and economic specialisation. In the German and European settings, excessive legalism has been noted. In the UK setting, economic expertise has moved the agencies towards a narrower and less critical judgement of market behaviour. Externally the agencies have established legitimacy by virtue of their

The unanticipated consequences of creating independent competition agencies

149

expertise and decades of responsible administration. By now the legitimacy has become attached to the agencies, their functions, and operational processes and is not dependent on the government in office or majoritarian processes. The survival of agencies is striking, especially in the UK where an entirely new legal structure has been grafted on to the existing agency arrangements. The external consequences are in each case a considerable contribution to the consolidation of a liberal political economy in Europe and the integration of the European single market. In P-A terms the agents have exploited their discretion but have done so in an environment in which the principals have also evolved their preferences towards more active, substantively implemented policy. In the cases of Germany and the UK it could be argued that the principal’s preferences were dominant, not so in Europe. The assertive development of European policy was led by DG IV as an agency that had to a large extent escaped from detailed control by its principals, the member states. In an intriguing process of feedback the success of DG IV in turn precipitated strengthening of the national regimes. This cycle of policy reinforcement is still under way.48 From this perspective, the unanticipated consequences of delegation to the competition agencies in Germany, the UK, and the EU are threefold. First, the competition agencies have shifted from a passive, symbolic defence of a competitive market economy to an aggressive promotion of market freedoms. Second, the agencies have substantially escaped from the influence of industrial interests. They are relatively impervious to industrial lobbying and do not give priority to views from leading industrialists and industrial organisations. The tripartism and corporatism of the period of creation has been superceded and ‘escape’ rather than ‘capture’ has occurred. Third, the agencies have stepped from the arms of business into the arms of lawyers and economists. The agencies are employing more purist economic criteria and their actions are driven by economic efficiency rather than by the public interest. A consequence of delegation has been to narrow the criteria employed by agencies so that the old, broad, balancing public interest criteria have been replaced by a far narrower and dogmatic focus on market efficiency. The agencies have turned symbolism into reality, replaced hypocrisy with activism, and in reinforcing the market are in danger of accentuating rather than compensating for some of the undesirable outcomes of the operation of market forces. NOTES The authors would like to thank the volume editors for their comments, similarly the participants in the two preparatory workshops. Additional thanks are due to Markus Muller and Oliver James. The research was generously supported by an award from the Anglo-German Foundation for the Study of Industrial Society. 1. See B.Doern and S.Wilks, ‘Conclusions’, in idem (eds.), Comparative Competition Policy: National Institutions in a Global Market (Oxford: Clarendon Press 1996), p. 344. 2. See D.Gerber, Law and Competition Policy in Twentieth Century Europe (Oxford: Clarendon Press 1998), pp. 406, 409, 412; M.Drahos, Convergence of Competition Laws and Policies in the European Community (Kluwer, 2001), p. 374. 3. S.Wilks, EU Competition Policy and the CEECs: The Institutions of a New Market Economy (Thorkil Kristensen Institute, Paper 23/98; South Jutland University Centre 1998).

The politics of delegation

150

4. The literature on American antitrust is vast, but see M.Eisner, Antitrust and the Triumph of Economics: Institutions, Expertise and Policy Change (Chapel Hill: University of North Carolina Press 1992); B.Weingast and M.Moran, ‘Bureaucratic Discretion or Congressional Control? Regulatory Policymaking by the Federal Trade Commission’, Journal of Political Economy 91/5 (1983), pp. 765–800. 5. See G.Majone, Independence vs. Accountability? Non-Majoritarian Institutions and Democratic Government in Europe (Florence: EUI Working Papers, European University Institute 1994); G.Majone et al., Regulating Europe (London: Routledge 1996), pp. 278–81. 6. See E.Chiti, ‘The Emergence of a Community Administration: The Case of the European Agencies’, Common Market Law Review 37 (2000), pp. 307–43. 7. Majone et al., Regulating Europe, pp. 271, 285. 8. See T.Makkai and J.Braithwaite, ‘In and Out of the Revolving Door: Making Sense of Regulatory Capture’, Journal of Public Policy 1 (1995), pp. 61–78. 9. M.J.Horn, The Political Economy of Public Administration: Institutional Choice in the Public Sector (Cambridge: Cambridge University Press 1995), p. 43. 10. S.Wilks with L.McGowan, ‘Competition Policy in the European Union: Creating a Federal Agency?’ in Doern and Wilks (eds.), Comparative Competition Policy, p. 226. 11. For example, P.Self, Administrative Theories and Politics (London: Allen and Unwin 1972), ch. 4. 12. C.Hood et al., Regulation Inside Government: Waste-Watchers, Quality Police and SleazeBusters (Oxford: Oxford University Press 1999), p. 72. 13. See C.Pollitt and G.Bouckaert, Public Management Reform: A Comparative Analysis (Oxford: Oxford University Press 2000), p. 53. 14. S.Wilks, In the Public Interest: Competition Policy and the Monopolies and Mergers Commission (Manchester: Manchester University Press 1999), p. 117. 15. J.Macey, ‘Organizational Design and Political Control of Administrative Agencies’, Journal of Law, Economics and Organization 8/1 (1992), pp. 93–125. 16. Horn, The Political Economy of Public Administration. 17. Ibid., pp. 18–19. 18. See B.G.Peters, Institutional Theory in Political Science; The ‘New Institutionalism’ (London: Pinter 1999) for an introduction. 19. See W.R.Scott, Institutions and Organizations (London: Sage 1995), p. 33. 20. Weingast and Moran, ‘Bureaucratic Discretion or Congressional Control?’. 21. B.M.Mitnick, The Political Economy of Regulation: Creating, Designing, and Removing Regulatory Forms (New York: Columbia University Press 1980), p. 326. 22. Ibid., p. 335. 23. M.Edelman, The Symbolic Uses of Politics (Urbana: University of Illinois Press, 1964), p. 26. 24. T.Arnold, The Folklore of Capitalism (New Haven: Yale University Press 1937), pp. 212, 215. 25. See Mitnick, The Political Economy of Regulation, p. 111. 26. G.Stigler, ‘The Theory of Economic Regulation’, Bell Journal of Economics and Management Science 2/1 (1971), pp. 3–21. 27. M.Eisner, Regulatory Politics in Transition (Baltimore MD: Johns Hopkins University Press 1993), p. 45. 28. C.Hood, ‘Keeping the Centre Small: Explanations of Agency Type’, Political Studies 26 (1978), p. 41. 29. This section draws on Wilks, In the Public Interest. 30. Ibid., p. 83 and H.Mercer, Constructing a Competitive Order: The Hidden History of British Antitrust Policies (Cambridge: Cambridge University Press 1995). 31. Fair Trading Act 1973, section 2.

The unanticipated consequences of creating independent competition agencies

151

32. Public Records Office, PRO, FV 60/57, IC3.460 Pt 1, ‘Monopolies and Restrictive Practices Bill: Instructions to Parliamentary Counsel’, exchange of correspondence with Civil Service Department (10 Sept. 1971) cited in Wilks, In the Public Interest, p. 186. 33. See Wilks, In the Public Interest. 34. Gerber, Law and Competition Policy, p. 273. 35. Ibid., p. 331, see also pp. 246–7. 36. Drahos, Convergence of Competition Laws and Policies, p. 241. 37. R.Sturm, ‘The German Cartel Office in a Hostile Environment’, in Doern and Wilks, Comparative Competition Policy, p. 201. 38. P.Baake and O.Perschau, ‘The Law and Policy of Competition in Germany’, in Majone et al., Regulating Europe, pp. 138–9. 39. Drahos, Convergence of Competition Laws and Policies, p. 249. 40. See D.Goyder, EC Competition Law (Oxford: Clarendon Press, 3rd edn. 1998); and M.Cini and L.McGowan, Competition Policy in the European Union (London: Macmillan Press 1998), pp. 39–41. 41. See S.Wilks and L.McGowan, ‘Disarming the Commission: The Debate over a European Cartel Office’, Journal of Common Market Studies 33/2 (1995), pp. 159–73. 42. Gerber, Law and Competition Policy, p. 264. 43. OECD, Germany, Country Report 1999–2000, OECD website, pp. 22, 26. 44. Sturm, ‘The German Cartel Office’, p. 218. 45. See S.Wilks, The Office of Fair Trading in Administrative Context (London: CRI Discussion Paper 1994). 46. Wilks with McGowan, ‘Competition Policy in the European Union’, p. 225. 47. G.Majone, ‘The Credibility Crisis of Community Regulation’, Journal of Common Market Studies 38/2 (2000), pp. 273–302 at 292. 48. See C.D.Ehlermann, ‘The Modernization of EC Antitrust Policy: A Legal and Cultural Revolution’, Common Market Law Review 37 (2000), pp. 537–90.

Judicial Delegation Doctrines: The US, Britain, and France

MARTIN SHAPIRO Law discovered delegation long before political science did. In another contribution to this volume, Stone Sweet treats the constitutional allocation of powers as a delegation from the people as principal to the segments of government as agents. Typically law has used the term delegation for a second principal-agent relationship, the delegation of some of its legislative power by the constitutionally empowered legislature to the constitutionally defined executive branch or, as several other pieces in this symposium point out, to so-called independent agencies. In British law the products of this secondary delegation are actually called ‘delegated legislation’. Thus we may conceive of a sort of nested set of delegations with the people delegating law making power to a legislature by constitution and the legislature delegating some of that law making power to administrative agencies by statute. Indeed, a little used and obscure piece of American constitutional law, the ‘non-delegation’ doctrine, treats the matter precisely in this way. More prominent, however, in legal thinking is concern for the second delegation—that is, delegation by statute, by an enactment of the legislature, of some of its law making authority to administrative agencies independent and otherwise. The delegation by legislatures to administrators of law making power by statute tends to set in motion one of those dynamics to which Stone Sweet traces the institutional authority of courts. The delegator or principal seeks to assure that the delegatee or agent will exercise the delegated law making power for the purposes and within procedural and substantive constraints established in the delegating statute. The delegator frequently resorts to judicial review as one means of policing the statutory conditions it has placed on its delegations and/or such judicial review is constitutionally required. Even where review is required, the intensity and intrusiveness of review is often determined either by the general provisions of an administrative code, general administrative review statute or case law, or by the particular statute making the particular delegation. This setting of courts to watch agencies on behalf of legislatures is notorious for creating its own principal-agent problems, ones that are a particular subset of those discussed by Stone Sweet. Whenever we establish courts as third party resolvers of two party conflicts, inevitably we pass some rule, law, norm, policy making authority to courts. It is impossible for them to resolve conflicts without rules. It is impossible for any outsider to provide rules complete enough and specified enough to resolve all the conflicts that judges face. Judicial elaboration and supplementation is necessary if the

Judicial delegation doctrines

153

courts are to do their conflict resolution job. So if either constitutions or statutes choose to police delegations by defining situations in which someone may sue a government agency or defend himself from agency suit by alleging agency violation of the conditions of delegation, then, unavoidably, the judges are assigned some law making authority. This general problem is aggravated by certain special features of the delegated legislative situation. The first arises if constitution or statute requires the reviewing court to determine whether the agency’s rule or delegated law meets a general reasonableness standard or achieves the substantive purposes, standards, or constraints set out in the statute that delegates supplementary law making authority to the agency. It turns out that the only way judges can judge whether an agency has made a reasonable rule, or a rule that satisfies the substantive goals of the statute, is to go through the same policy analysis process the agency did and see whether they reach the same policy outcome the agency did.1 If the judges come out somewhere near the agency, then they find the agency’s action reasonable and in accord with the statutory commands. Otherwise they do not. Judges themselves, or statutes or constitutions, may somewhat constrain this judicial second-guessing of agency policy analysis. It may be required that reviewing courts give some level of deference to agency fact-finding or technical expertise. It may be explicitly or implicitly required that judges approve any agency rule that some reasonable persons, knowing what the agency could reasonably have known, might have made rather than the rule that the judges themselves would have made. Whatever the self- or other imposed constraints on judicial policy analysis, however, ultimately the only way reviewing judges can determine whether an agency rule conforms to the legislation delegating rule making authority to it is to go through the same thought processes the agency did and see if they end up in at least roughly the same ball park as the agency did. There is a second problem, or perhaps the same one by another name. One issue in nearly every judicial review of administrative rule making is whether the rule is ‘lawful’, that is whether it is in accord with the statutes enacted by the legislature. This question in turn typically depends on interpreting the statutory language. If a statute empowers an agency to make rules about ‘vehicles’ in wilderness areas, and an agency rule bans horses, whether the rule is lawful depends upon whether a horse is a vehicle within the meaning of the statute. Quite apart from all the post-modernist flummery about the indeterminacy of all text, reviewing courts often are confronted with two, plausible, rival interpretations of a statute. One is the agency’s interpretation. The other is that of the party challenging the lawfulness of the agency’s rule. Here again there may be self- or other imposed constraints on judicial choice. It may be required that courts give great deference to agency statutory interpretation. Ultimately, however, courts simply cannot give up the claim that they are the final authority on what the law says. Without such a claim courts simply lose all institutional authority. Courts inevitably must decide whether the agency delegatee has properly met the terms of the delegation. If it is true that any ‘interpreter’ of law inevitably is to some degree a law maker, the temperature tends to get particularly high when judges interpret regulatory statutes delegating supplementary law making authority to agencies. The causes of this high temperature are offered and illustrated in the Bartle/Wilks, Thatcher and Thatcher/Stone Sweet contributions to this volume. Legislatures delegate to agencies because of unresolved policy preference conflicts within the legislature, need for high levels of technical expertise to assure policy success and/or desire to shift political responsibility.

The politics of delegation

154

An additional motive, particularly where regulation of private economic conduct is involved, is the belief that agency rule making may provide more rapid, flexible responses to constantly changing market conditions than would the law making process in the legislature. For all of these reasons, statutes that delegate law making authority are likely to be written in particularly broad or vague language. It is axiomatic that the broader the textual language, the greater the discretion of the interpreter to supply independent meaning. Courts reviewing the lawfulness of delegated legislation often do, and fairly often must, actually make the law of the delegating statute. Thus when the legislature as principal chooses to police its delegation of law making authority to its administrative agent through the institution of judicial review, unavoidably it has also chosen to delegate law making authority to courts. And so it must confront the problem of policing them as well. Why do legislatures so often make this choice of policing by litigation? Litigation shifts the costs of policing away from the legislature onto the litigating parties. It provides feedback loops to both principal and agent, feedback loops that particularly highlight troubles. Most importantly, litigation markets provide a combined police patrol and fire alarm system. Those persons with substantial monetary or ideological interests in particular delegated legislation will continuously monitor the agent. Thus the legislature gets the regularity of police patrols together with decentralised mechanisms for alerting principals to unwelcome behaviour by agents. Lawsuits are fire alarms. By specifying in their statutes broader or narrower categories of persons who may seek judicial review of delegated legislation, legislatures may choose to have more or less patrolling and alarm sounding. Delegating legislatures are not always in a position to freely choose or tailor this policing mechanism. Where judicial review exists, the judicial reviewer will itself be one of the determiners of the size, shape, and velocity of the litigational market created. It does so not only by announcing its own standing rules but also by creating incentives or disincentives to enter the market through announcing legal doctrines and rendering streams of decisions either more or less favourable to potential litigators. There are a number of (partial) alternatives to judicial review for policing delegated legislation. One is executive supervision such as British ‘treasury control’ or Office of Management and Budget oversight of agency rule making in the US. Another is direct legislative control through ‘legislative veto’ provisions. Agency rule making may also be controlled by legislatures through their general powers of appropriating and legislating. Both by post hoc reward and punishment, and by anticipated reaction, agency rule making (like all other agency activity) can be subjected to control by the legislature itself as various studies noted in other contributions to this volume indicate. Depending on constitutional arrangements, judicial review may simply be excluded by a provision in delegating statutes or the intrusiveness of review can be fine-tuned by dozens of legal devices incorporated in constitutions, statutes, case law, and/or administrative practice. Judicial review standards or tests may be written to either encourage or discourage active review. Judges are more likely to overturn an agency made rule under a ‘reasonableness’ standard than if they are told they may only strike down ‘arbitrary and capricious’ rules. (Of course, the fixed words of any review standard, such as ‘reasonable’ or ‘arbitrary and capricious’ are themselves subject to interpretative manoeuvre by the judges.) Agencies may be subjected to more or less elaborate

Judicial delegation doctrines

155

procedural and evidentiary rules that generate more or less chance of agency procedural error that can serve as a successful ground for judicial review. Delegating statutes may be worded so precisely that there is little room for judges to come up with rival interpretations to those of the agencies or so broadly that no rule the agency might adopt could possibly violate them. The remedies that courts can offer to parties injured by unlawful rules may be broader or narrower, influencing both the power and incentives of courts to intervene and the incentives of prospective parties to go to court. Finally the institutional qualities of courts can be manipulated and exploited in various ways to encourage or discourage activist, policy oriented judicial intervention. Judicial review may be maximally separated from administration, vested in an independent judiciary. At the other extreme, judicial review may be exercised by tribunals that are placed within administrative organs and only minimally insulated from the administrators being reviewed. In France, for instance, the administrative courts are separate from the regular courts and from the agencies, but the highest administrative court is a section of the Council of State, which is an organ of executive control over administration. Both agencies making delegated legislation and courts reviewing them are legally subordinated to the statutory commands of the legislature. Their capacity to disobey their principal derives from their need to ‘interpret’ those statutory commands. Thus, the more easily and rapidly the legislature can rewrite its statutory commands to ‘correct’ ‘mistaken interpretations’, the more it can constrain its administrative and judicial agents. The ultimate weapon of legislatures against policy making by reviewing courts is responding rapidly by corrective amendments to the statutes the judiciary has interpreted in ways that the legislature does not like. Thus two party, strong party discipline, parliamentary systems are likely to experience highly deferential judiciaries. But deferential to whom? If we take the UK as our example, the result of a rapid parliamentary response time is not that the judges police the agencies in the interest of the legislature but rather that the agencies police the courts in the interests of themselves. If a court should find that the agency has acted unlawfully, in the sense that its rule is in conflict with the parliamentary statute delegating rule making authority to the agency, the result is highly unlikely to be that the rule will be changed to conform to the statute. Far more likely is that the statute will be changed to conform to the rule. The agency, typically one of the ministries, will get the cabinet to order the parliament to ‘correct’ the judicial ‘misinterpretation’ of the delegating statute by an amendment to that statute confirming the agency’s rule. The contrast with two party, weak party discipline, presidential systems is obvious. Given Congress’ extreme difficulty in passing statutes, US courts’ statutory interpretations are likely to survive legislative correction for a long time or for ever. Congress will not discipline a court unless the court’s interpretation generates sufficient opposition from a sufficient number of factions in each house to generate in turn majority coalitions in both houses to which the president is willing to acquiesce. Judicial findings that agency rules are unlawful are far more likely to generate changes in the rules than in the congressional statutes. We should also expect, then, that multi-party, strong party discipline, parliamentary systems will experience high levels of judicial policy making. Such systems might also be expected to experience long delays before ‘correcting’ legislation could respond to judicial policy initiatives. A number of factors may, however, militate against judicial

The politics of delegation

156

activism in multi-party, strong party discipline, parliamentary systems. The coalition building problems of such systems may be handled in part by giving each party in the governing coalition complete control over one or more ministries with only a facade of collective cabinet responsibility. Under such conditions each ministry will be in something like the British position vis-à-vis courts and parliament. If a court should hold a ministry rule unlawful the ministry may be able to invoke the coalition agreement to get the cabinet to order the parliamentary governing coalition to pass whatever correcting legislation the ministry needs to reaffirm the lawfulness of its rule. Alternatively, judges in a polity suffering from the legislative inertia of multi-party coalition parliamentary systems may acquiesce in the sub-silentio transfer of legislative functions to the administration. They may abandon serving as a policing agent for a principal too weak to actually be a principal, but they do so without desiring to become a principal themselves. To the extent that institutional arrangements lead to a reviewing judiciary that identifies with administration, this judicial quiescence will be reinforced. There remains one further major aspect to this story of legislative principals, administrative agents and courts. This aspect returns us to the constitutional law with which we began. In the US and nearly all European states there is now some form of judicial review of delegated legislation to ensure that such legislation is in accord not only with national statutes but with supra-statutory norms. Such norms may include national constitutions, European Union legal norms, the European Convention on Human Rights, international law and/or ‘entrenched’ or ‘basic’ statutory human rights provisions. Stone Sweet’s contribution to this symposium sketches the rise of ‘higher law’ judicial review in Europe. In all these situations above and beyond enforcement on the agency of the terms of the delegating legislation, the principal becomes increasingly notional and the problem of policing the policing agent increasingly intractable. We have seen that the ultimate hold of legislative principals over their judicial agents is amendment of the statutes that the judges interpret. The harder the amendment process, the greater the judicial principalagent problem. Where entrenched statutory provisions, constitutions, or transnational agreements provide the text for judicial interpretation, national legislatures have a hard time amending. Indeed, the principal shifts from the relatively concrete national legislature to principals that are either extremely internally divided, like the Council of the European Union, or nearly fictitious, like the ‘people of the United States’ who ‘ordained’ ‘this constitution’. The less real, active, and immediately reactive the principal, the greater potential for insubordination enjoyed by the judicial agent. Legislation (rules) were really made at bureau, not at departmental level. In France and the UK, too, those creating administrative law jurisprudence saw a mix of majoritarian and technocratic elements in both cabinet ministries and other agencies not headed by political appointees. The principal focus will be on courts as non-majoritarian agents. The main argument is that when legislature as principal delegates law making to either a majoritarian or nonmajoritarian government agency and chooses judicial review as a policing mechanism, the legislature, whether it wants to or not, consciously or unconsciously, also delegates law making power to a non-majoritarian government agency, the courts. The comparative task is to track similarities and differences in the quantity and quality of this secondary delegation in the US, the UK, and France.

Judicial delegation doctrines

157

In doing so, a number of the general concerns of this project particularly ought to be borne in mind. First, the delegation involved here is not one to an agent either normatively or empirically totally subordinated to the principal. Judiciaries are supposed to be and often actually are relatively neutral and independent. Administrative law is highly reflexive in character, that is, it operates under an internal, relatively autonomous, lawyer driven dynamic of its own as well as being responsive to outside influences. Second, in many situations, legislative losers will more or less deliberately seek to expand the discretion of judicial reviewers. They do so because they hope to win at the stage of review of delegated legislation what they lost at the stage of passage of the delegating legislation. So review often markedly increases uncertainty about prospective winners and losers even above the routine uncertainties of litigation. Thirdly, delegations to courts are usually delegations to agents with a long, independently established record of legitimacy. This legitimacy may facilitate delegation but also tends to increase the risk of agent defection. Counterbalancing the special long-term legitimacy resources of courts as courts, however, is another dynamic already referred to. Judicial review of delegated legislation always raises the question of the relative legitimacy of two agents, the initial delegatee and the judicial reviewer. The more obvious it becomes that the judicial reviewer is second-guessing the administrative agency, and the more obvious it is that the guesses concern matters about which the agency knows far more than the judicial reviewer, the more the favourable legitimacy balance of the court is depleted. Against this background, we may examine the constitutional and statutory judicial review of delegated legislation in three states: the US, the UK, and France, in each instance concentrating on the chronic problems encountered by a principal in setting one agent to quell the potential insubordination of another. This is only a preliminary survey of some terrain in the hope that some hypotheses implicitly offered can be rendered more explicit and some new hypotheses stumbled upon. THE UNITED STATES A description of the US law of delegation probably should begin with the so-called nondelegation doctrine.2 Literally, that doctrine holds that one to whom authority has been delegated may not re-delegate it to another. Through the constitution the American people have delegated their law making authority to Congress. Congress, therefore, may not re-delegate any of its law making authority to anyone else. If the non-delegation doctrine were really taken seriously much of American government would be unconstitutional. Certainly, the thousands of pages of rules and regulations issued by executive branch agencies and independent regulatory commissions would be. In fact, the Supreme Court does not prohibit delegation of congressional law making authority. It holds that the doctrine only requires that, when Congress delegates, it announces standards or purposes in its delegating statute that will guide the law making of the delegatee. Often, however, the court has accepted as sufficient to meet the demands of the non-delegation doctrine statutory language that provides no more precise standard than that the delegated legislation be ‘in the public interest’ or be issued to serve ‘public convenience and necessity’ or be aimed at achieving ‘clean air’.

The politics of delegation

158

Standing alone, the non-delegation doctrine is a fake. The real significance of the doctrine lies not in its independent effects but in its effects when combined with another doctrine.3 The Supreme Court has long held that where two interpretations of a congressional statute are possible and one interpretation would create the risk that the court would have to find the statute unconstitutional, the court will choose the other interpretation so as to ‘save the constitutionality’ of the statute. Parading as a judicially self-restrained deference to Congress designed to avoid striking down congressional statutes whenever possible, the doctrine is actually an activist judicial strategem. By merely raising the risk that the statute read a certain way would be unconstitutional, the court gives itself a licence to read it some other way. Indeed, it gives itself a licence to read the statute in a way that is clearly different than the way Congress intended, because, if it were read the way Congress intended, it might be unconstitutional. Complete judicial deference to agency interpretations of the statutes delegating law making power to them would allow agency ‘bootstrapping’. That is, the agency could interpret the legislation as giving it far broader law making authority than Congress had actually intended. Yet agencies have the expertise and continuous implementation experience that make it difficult for a court to say that its own independent interpretation of the delegating statute ought to be preferred to that of the agency that received the congressional delegation. The non-delegation plus saving constitutionality doctrines give federal courts an extra dollop of legitimacy when striking down agency interpretations that the justices see as bootstrapping. The court may say of a particular agency interpretation of a statute that, however plausible it may be, it raises constitutional problems so that the court has been driven to choose an alternative interpretation. The bulk of US jurisprudence, however, does not concern the constitutionality of delegating legislation but the lawfulness, as opposed to the constitutionality, of the delegated legislation. The usual issue is whether an administratively created regulation is in accord with or violates not the constitution but some congressional statute. Each US statute delegating law making authority to administrators usually contains its own specific procedural and substantive rules or standards for the delegated legislation, that is for how the administrative regulation shall be made and what should be in it. In 1946 Congress enacted the Administrative Procedure Act (APA) which provided a general set of procedural rules for the making of regulations. That act does not supplant but only supplements the provisions of delegating statutes. Since the passage of the APA new legislation delegating the power to make regulations typically will make reference to the APA and announce any special deviations from it or supplements to it that are to govern the particular delegation.4 We shall see that in England and France most administrative law doctrines do not distinguish between delegated legislation and other kinds of administrative decisions. Because of the APA, that distinction is always important in the US. The APA distinguishes ‘informal rule making’ from other kinds of administrative decision making and sets up special procedural rules and judicial review standards for it. The APA struck a compromise between pro- and anti-bureaucratic forces. Where administrators were making decisions about the legal entitlements and obligations of particular individuals, that is where they were engaging in administrative adjudication, they were to act pretty much like courts and be subject to roughly the same kind of appellate judicial review that a trial court is. It should conduct a kind of a trial using procedural rules similar to those

Judicial delegation doctrines

159

used in a court, before a kind of a judge, who should compile a record like a trial court record and render a kind of verdict. The losing party should be allowed to appeal ‘on the record’ to a regular court which would review the initial outcome the way an appeals court reviews a trial court outcome. On the other hand, where an administrative entity was enacting delegated legislation, that is rules or regulations of general application, because it was making law, it should act roughly like a legislature. It need not hold a formal trial-like proceeding, compile a record, limit its considerations to ‘evidence’ formally presented by ‘parties’, or write a formal justification of the rule it arrived at any more than a legislature would have to in making a law.5 The APA did, however, pay some attention to the principal-agent problem. The APA did formally require of delegated law makers some things Congress did not formally require of itself. ‘Informal rule making’ under the APA required that the rule maker give advance notice that it intended to make a rule, invite public comment, publish a concise and general statement of the basis and purpose of the rule, and publish the rule. Judicial review of agency rules was to be available. But reviewing courts were to strike down a rule only if it was unlawful, arbitrary, and capricious, or an abuse of discretion. Although the informal rule making provisions imposed some requirements on the agencies, they were clearly meant to avoid the formal, adversarial, evidentiary, procedural, record keeping, and opinion writing requirements that applied to courts. In the 1950s the orthodoxy of the administrative law of judicial review of informal rule making was judicial deference to the ‘expertise’ of the administrative agencies. Courts deferred to administrative interpretations of the statutes delegating ‘informal’ rule making authority to them. In the absence of any requirement of a formal record of the facts underlying a rule making or of adversarial offerings of facts, courts presumed that agencies had the facts to support the rules they had made. The APA review standard of ‘lawfulness’, that is, that the rule made should be in accord with the statute that had delegated the power to make the rule to the agency, was easily met if the courts deferred to the agency on issues of law and presumed in favour of the agency on issues of fact. The APA arbitrary and capricious standard was read by the courts as a lunacy test. An agency rule was to be struck down only if no reasonable person could have made it. In the 1950s most of the administrators and most of the judges were New Dealers. They were hardly going to call one another crazy. Even without shared party allegiance, it is institutionally difficult for one organ of government to say that another belongs in the loony bin. Between about 1965 and 1980, without amendment of the relevant passages of the APA, all of this changed. Part of the change results from deliberate congressional action. Conscious that agency rule making was itself a complex, extended political process, in many new delegating statutes Congress set new procedural rules for that process designed to favour some interests and policies over others in the rule making struggle.6 Much of the new statutory language of procedure, however, was borrowed directly from procedural innovations introduced independently by the judiciary. The judiciary, through its case law, radically amended the APA. Congress did not, but the language of many new congressional statutes echoed, and thus endorsed, those judicial amendments. The federal judiciary’s transformation of the administrative law of rule making is tracked in detail elsewhere.7 It is enough here to briefly state the results. The APA had simply required notice, an opportunity for public comment and a concise and general

The politics of delegation

160

statement of basis and purpose. It had not required the making of a record. That omission was clearly deliberate. Indeed, ‘informal’ rule making is called informal to distinguish it from another kind of APA-defined rule making called ‘formal’ rule making, which must be ‘on a record’. The APA set the very permissive ‘arbitrary and capricious’ judicial review standard. Again Congress’ intention to leave the agencies relatively free in informal rule making is evidenced by the stricter judicial review standard the APA sets for several other kinds of administrative actions. By the end of the judicial transformation of the APA, agencies were giving not one notice but typically three successively more detailed ones. The rounds of comments now often included oral hearings and public meetings and grew to thousands of pages of printed testimony. The rule making agency was required to respond persuasively to each and every comment made. Rather than a concise and general statement of basis and purpose the courts required the agency to produce an exhaustive ‘rule-making record’ demonstrating that the agency had responded to every point raised by commentators and, indeed, relevant points that they had failed to raise. Rather than showing that they were not lunatics, the agency had to demonstrate to the reviewing court that it had made the best rule possible. Judges continued to say that they deferred to administrative expertise, would not substitute their own policy preferences for those of the agencies and, where the statute was ambiguous, followed agency interpretations of their statutes so long as they fell short of bootstrapping. In fact, however, the Courts of Appeal often did impose their own policy preferences and their own statutory interpretations. They announced that they were ‘partners’ with the agencies in the rule making process. Given that they were the ultimate deciders of whether a rule actually became law, judges were actually the senior and very far from silent partners. It does not appear that principal-agent thinking was much in the minds of the judges. The judicial initiatives do not seem to have been designed to produce maximum congressional control over rule making. Instead they appear to have been inspired largely by pluralist democratic thinking. The new procedural rules were designed to achieve greater transparency and participation in the rule making process for the benefit of interested groups and individuals. The point was not to increase congressional surveillance but rather non-governmental participation in rule making. In that sense congressional control was actually weakened. Many interests could hope to gain by rule making participation what they had lost in the earlier process of congressional statute making. Congress could manipulate procedural rules to favour certain interests or outcomes at the rule making stage, but that very indirect intervention demonstrated how much the courts had strengthened non-congressional players in the total law making process of statute making plus rule making. The transformation also dramatically raised the ‘set an agent to watch an agent’ paradox that we have seen is raised by all judicial review. Congress provided for judicial review in every one of the new delegating statutes. It acquiesced in and often positively supported more active judicial review. It often even provided especially liberal standing provisions in the new statutes, allowing more people to sue to force the agencies to live up to their statutory duties. When the courts declared themselves partners in the rule making business, however, they were declaring themselves partners with the agencies not Congress. More judicial review meant more judicial participation in rule making. The new regulatory statutes typically were complex texts that offered mixtures of aspirational

Judicial delegation doctrines

161

general language, some incomplete specific commands, intentional and unintentional loopholes, and vague language papering over unresolved congressional policy differences. Judicial enforcement of such statutes on the rule making agencies necessarily involved much judicial interpretation of the statutes. Given the best will in the world by judges to avoid imposing their own policy preferences, interpretation of such statutes necessarily and unavoidably involves judicial policy choices because the statutes themselves frequently had not clearly, consistently, and completely made such choices. The higher the procedural standards and the more demanding the judicial review standard, the more judges are likely to find that the agency choice did not fall within the reasonable set of choices implicit in the statutory language. The more judges are exposed to a full-scale rule making record of a complete dialogue between the agency and all interested parties, the more likely they are to conclude that it was unreasonable for the agency not to adopt whatever the judges consider the best possible rule. What is reasonable about choosing a rule that cleans less air at higher cost rather than more air at lower cost? Where the reviewing court had serious doubts about the wisdom of the agency rule, it could strike it down on pure substantive grounds as ‘arbitrary and capricious’ now transformed. Or it could use the enormous statutory interpretation opportunities opened by the new statutes to strike down the rule as ‘unlawful’, that is, not in accord with the commands of the delegating statute properly interpreted. Or the court could simply search the huge record of the extended, transparent, participatory process that the courts themselves, and often Congress, had demanded and find some procedural flaw which necessitated invalidating the rule. There always is such a flaw if you require elaborate enough procedures and look hard enough for mistakes. The Supreme Court several times warned the Courts of Appeal to stop using statutory interpretation and procedural nit picking as disguises for what was in reality policy based judicial intervention against the agencies. In the very process of doing so, however, the Supreme Court actually adopted the transformations of the APA that the Courts of Appeal had achieved and so left the gates open while chiding the reviewing courts for going through them.8 Clearly, American courts have been making a lot of such choices. From the 1960s onwards they were invalidating many agency rules, only some of which openly violated clearly stated congressional intentions. Indeed, recently there has been a considerable backlash in legal circles calling into question the degree of judicial activism.9 The transformation has resulted in an enormously costly and very slow rule making process. Judicial policy choices may be even more suspect than those of technocrats in areas where the rule making record has become, at the command of the judges themselves, too long and complex for the judges to understand. Yet hardly anyone in the US proposes to give up the greater transparency and participation that the transformation imposed on rule making. Instead, what is sought is that somehow participation become more deliberative and less adversarial and that judges restrain their tendencies to impose their policy preferences on the rule makers. The transformation of US law of delegation in the last half century seems to have been guided far less by concerns for delegation itself, or for principal-agent problems, than by a desire to render a particular stage of the law making process more democratically participatory and transparent and more productive of substantively good results. To the extent that Congress as principal is advantaged by better researched and better explained

The politics of delegation

162

work product from its agents exercising delegated law making authority, the transformation has benefited the principal. On the other hand, the transformation has greatly increased opportunities for non-congressional influence on the congressional agents engaged in rule making. It has also greatly aggravated the standard paradox of judicial review. Judges set up as policing agents to watch administrative agents on behalf of congressional principals themselves create principal-agent problems for Congress. To a substantial degree they may become independent influences and even controllers of the agents’ behaviour. The less capable of clear, unified commands the congressional principal is, the greater the principal-agent problems raised by review become. All of this, of course, presumes that Congress wants to be the principal. In those instances in which Congress prefers to actually transfer policy making authority elsewhere, judicial review simply assures that the elsewhere is more diffuse. Of course, Congress need not make a black and white choice about its principal status. Typically it retains some ultimate policy control to itself but willingly turns over much policy discretion to others. When Congress chooses to initiate or acquiesce in active judicial review, it is turning over more policy discretion to courts. Congress has far less budgetary control over courts than over administrative agencies. Unlike agencies, courts do not usually have agendas that seek new federal legislation so that Congress cannot control courts as it does agencies by granting or withholding new legislation. The principal control of Congress over reviewing courts is its power to amend its own legislation because courts purport to be acting to enforce that legislation. Congress’ well-known institutional barriers to rapid legislation leave American administrative review courts under less control than are such courts in countries where political leadership can quickly achieve new statutes. ENGLAND10 In one version, the story of judicial review of delegated legislation in England is a classic principal-agent story. The conventional view in the twentieth century has been that the unwritten British constitution provides for absolute parliamentary sovereignty. It follows that any general norms enacted by bodies other than parliament can have the force of law only when they are enacted under and are in accord with parliamentary laws delegating parliamentary law making authority to non-parliamentary bodies. Judicial review of such delegated legislation is then exclusively ultra vires review; that is, solely review of whether the secondary norms are within the limits of the delegated powers granted to someone else by parliament.11 Just as English courts may not overturn an act of parliament because parliament is sovereign, they may not overturn an act of a nonparliamentary body so long as that body is acting by delegated parliamentary authority and within the limits set by parliament on its delegation. The only function of judicial review is to police the agent on behalf of the delegating principal. Even within this vision, no distinction is made between democratic and nondemocratic and only limited distinctions even between public and private delegatees. The same ultra vires review is applied to delegated law making by cabinet ministers and ministries who are directly politically responsible to the elected parliament and to a wide array of agencies, tribunals, local governments, planning authorities, government

Judicial delegation doctrines

163

corporations, and so on, which exercise delegated law making authority but are not directly responsible to parliament. While some complex distinctions exist between public and private law, courts exercise roughly the same ultra vires review when parliament delegates public law making power to private bodies such as professional associations or corporations exercising monopoly powers by virtue of state charter as it does when parliament delegates to the cabinet departments. In only one way are ‘democratic’ executives favoured over non-democratic ones. Ministers may re-delegate law making authority delegated to them to their ministry subordinates while other executives cannot usually re-delegate.12 Under the conventional view—perhaps especially under the conventional view—the real sticking point for delegation in England lies, as we have seen it does generally, in the intentions of the principal. If the parliamentary principal delegates without stating or knowing what its policy purposes are in making the delegation, or if it more or less explicitly states that it wishes to delegate to an agent without any limits on the agent’s discretion, how is ultra vires review to work? In fact not only do many parliamentary statutes disguise the failure to reach a policy outcome under vague statutory language, but also many such statutes explicitly state a parliamentary intention to leave its agent unlimited and uncontrolled except by largely fictional cabinet responsibility to parliament and the more real threat of the next election. Parliament does so by ‘subjectivity’ clauses which provide that the minister may act not when specific statutorily defined conditions exist but whenever ‘in the minister’s view’ action is warranted. ‘Privity’ clauses exclude judicial review. On these fronts, the courts have tried desperately to protect the principal against itself and to protect their own powers as policing agents for the principal. They have virtually ignored privity clauses and have tried to narrow subjectivity clauses. And, more generally, like reviewing courts everywhere, English courts have pretended as best they can that statutes that either state no purpose or multiple unprioritised purposes or contradictory purposes do contain purposes sufficiently articulated to be judicially imposed on agents. As of 1950, however, and in startling parallelism to the US at that date, English judges were not actually intervening at all in delegated law making matters. Although they did resist privity clauses, they bowed to subjectivity clauses and bowed even more deeply to the interpretations of principals’ intentions offered by the agents enacting delegated legislation. An agent allowed unchallenged interpretation of what vires his principal has allowed him is unlikely even to be found to be acting ultra vires. As of 1950, ultra vires might serve to police the principal-agent relation between central and local government or government and private authority. Very occasionally, a local government or special purpose authority might be found to have exceeded the jurisdiction granted it by parliament, but never the ministries. In the wake of World War II, and in the enthusiasms of socialism, confidence in executive discretion was at its height. In national politics in general the long movement to reverse the principal-agent relationship of parliament and cabinet was accelerating, with little parliamentary desire or capacity to subordinate the executive. In judicial politics the judges had little taste for policing a noble executive agent, particularly when the principal was making a virtue of subordinating itself to its supposed agent. Only one small cloud troubled this climate of judicial deference to executive law making. The cloud had to do with reasonableness. It is extremely distasteful to English

The politics of delegation

164

judges to proclaim that even the sovereign may make unreasonable laws. Conventionally English judges have spoken of three ‘heads’ of ultra vires review: lawfulness, procedural regularity, and reasonableness. Lawfulness, of course, was the core principal-agent concern. Did the delegated law making conform to the terms set down in the delegating parliamentary statute? Reasonableness would appear to be independent of the principalagent relationship. Whether or not the agent had acted within the scope of authority delegated to him, was his action reasonable? Parliament might delegate to a cat ministry the authority to make regulations for the control of cat diseases. When the cat ministry enacted a regulation requiring all cats to be inoculated against a disease which had not actually appeared anywhere in 300 years, certainly it was acting within its vires but was it acting reasonably? Common law courts had never made good a claim that they could strike down parliamentary statutes as unreasonable. Yet they made the claim, even in 1950, that they could strike down delegated legislation as unreasonable. If parliament is free to make unreasonable laws itself, why is it not free to delegate the power to make unreasonable laws to others? Common law courts took care of this problem by saying that they surely could not impute to parliament an intention that its agents make unreasonable laws. Thus an unreasonable delegated law must be ultra vires. Yet if judges were to judge the reasonableness of the action of agents, then the crucial problem of judicial review noted earlier arose. By setting up courts as a policing agent to watch its delegated law making agent, the principal would be creating a second and even more difficult principal-agent problem. As of 1950 the English courts dealt with this problem by clinging to reasonableness review but driving it down to an absolute minimum. Reasonableness review always presents a spectrum between two extremes. At one is extremely intrusive judicial review. Surely it is unreasonable to make any rule less than the best rule that could be made. It is unreasonable to say ‘I know that something else would do the job better and cheaper, but nonetheless I choose to do it less well and more expensively’. Thus a court engaged in reasonableness review requires the administrator to set out all the alternatives and show why the one chosen is best. We have seen this end of the spectrum in the US. At the other extreme is a test that reads ‘Something is unreasonable if no reasonable person could have chosen to do it’. This is the US arbitrary and capricious test as written and as initially employed by US courts. In England it is the ‘Wednesbury’ unreasonableness test, named after a famous case.13 English courts clung to reasonableness, but reduced it to a lunacy test.14 Administrative acts were unreasonable only if no reasonable person could have done them. Some judge might occasionally say that a local authority ought to be committed, but he could hardly say that about a cabinet member, and in fact never did. All of the above is the conventional story, clear in 1950, and still propounded by eminent English judges and scholars. But since 1950 another story has been developed. It is easier to follow in academic writing and the off-the-bench writing of certain judges than it is in the actual behaviour of English courts. Also it is a story about administrative law in general, not about the administrative law of judicial review of delegated legislation in particular. Unlike the situation in the US, where the APA tends to create a specialised administrative law of ‘informal’ rule making, English administrative law doctrines, such as ultra vires and reasonableness, tend to be general and apply to all administrative decisions, including both the making of regulations and particular decisions about the treatment of specific individuals.

Judicial delegation doctrines

165

Here is a very brief version of the story. Since 1950 English judges have become much less deferential to administrators and have increasingly insisted that administrators act justly and fairly. They have been more willing to go behind administrative professions of public interest to what has actually been going on. They have come to weigh individual rights more heavily as against administrative claims of pragmatic pursuit of public policy. The result has been a new flowering of administrative judicial review with the judge actively looking over the administrator’s shoulder. Judges have become less deferential to administrative interpretations of delegating statutes. Where individual rights are at stake, they have moved to a less judicially restrained test of reasonableness than ‘Wednesbury’,15 demanding that administrators provide some reasoned defence of their sacrifice of individual to public interest beyond a mere claim that what they have done is not insane. This story is told in a pure parliamentary sovereignty and a modified parliamentary sovereignty version. In the pure parliamentary sovereignty version,16 courts have simply taken more seriously their role in policing the principal-agent relationship between delegating parliament and delegated administrators. Parliament itself has engaged in a progressive series of actions designed to strengthen judicial review. These include a new judicial review statute, increasing statutory specificity in the granting of individual rights judicially enforceable against administrators, accession to the European Convention on Human Rights, accession to the European Communities, and the new statutory Charter of Rights. Courts are doing more for their principal because it has commanded them to do so. Finally, parliament itself, through a series of reforms, particularly of its committee system, has evidenced a greater desire to reassert its principal status vis-à-vis the government. More active judicial review is in harmony with this parliamentary development. The modified parliamentary sovereignty story runs as follows.17 Conventional twentieth century ultra vires doctrines were designed to finesse a longstanding tension between parliamentary sovereignty and the independence of common law and common law courts. Both reasonableness and natural justice or procedural regularity as ‘heads’ of judicial review are common law standards of judicial review of administration. Law making has always been a shared power of parliament and the common law courts. At least where parliament has not spoken specifically to the contrary, common law courts apply their own common law standards to administrative actions including delegated legislation. Moreover, by parliamentary action the British constitution has now come to include a host of individual rights grounded in the European Convention on Human Rights, the treaties of the European Union and the UK Charter of Rights but independently interpreted by English judges who are part of a European community of independent judges which itself transcends national sovereignties. Therefore, and again a certain US parallel suggests itself, judges and administrators are partners in assuring the reasonableness and respect for rights of delegated legislation. Parliament and common law courts are partners in demanding that administrators act lawfully, that is, in accordance with parliamentary statutes, reasonably, justly, and in accordance with legally guaranteed individual rights. One ‘head’ of judicial review, ‘error of law’, avoids the parliamentary sovereignty versus independence of the common law courts issue entirely. For when reversing an administrative action for ‘error of law’ the reviewing court purports to be enforcing the

The politics of delegation

166

will of the parliamentary principal on the administrative agent. The clear error the court claims to discover is a conflict between the administrative action and a parliamentary statute.18 As in the United States, the question of how much deference the judges ought to pay to administrative interpretation arises. Judges favouring judicial restraint may speak of ‘clear error of law on the face of the record’. And, of course, as noted earlier, a court that invokes ‘clear error’ against administrators risks a new intervention by parliament backing the administrators against the judges’ interpretation. Those are the stories. What has actually happened? Here we very quickly reach the limits of current social science.19 The number of filings for judicial review and cases heard has risen dramatically. If immigration cases are excluded, the number of wins against government administrators has not greatly increased. The courts have been quite activist in immigration matters. The number of wins against private entities wielding public authority is up, but it could hardly help but be, as privatisation has multiplied such entities. The UK scholars and judges who push for more active review keep talking about a half dozen, maybe now a dozen, leading cases of more activist review, but they have to keep talking about those few cases. Most of them are decisions against local governments, not the national ministries. There are many signs that the national government is taking law more seriously, and that the national courts are too. It is possible that there are hardly any judicial reversals of national government decisions because the national government is now so careful about lawfulness. There has finally been one leading case in which a piece of delegated legislation of a national ministry has been declared unlawful, and unlawful on unreasonableness grounds. That there is only one such case, and that it is about immigration and one of the most fundamental of individual rights, indicates how uncertain the picture is. The case is R. v. Secretary of State for Social Security, ex parte Joint Council for the Welfare of Immigrants.20 A previous ministry regulation had specified that aliens who had entered unlawfully and then made asylum claims would not be issued work permits. The previous practice had been to grant them the same subsistence benefits as other unemployed persons. A new ministry regulation declared them ineligible for such benefits until their asylum claims had been adjudicated. The court declared this regulation unlawful because unreasonable. The thrust of the decision was that it was unreasonable to require the alien to pursue his asylum claim at the cost of starving to death before it could be finally adjudicated. A fundamental right to life is involved. The ministry in effect was nullifying asylum rights guaranteed by a treaty to which parliament had assented. The court is so embarrassed by its policy intervention that one of the judges goes out of his way to say that the court does not question the minister’s policy of discouraging asylum seeking and that the minister is perfectly free to cut asylum seekers’ subsistence benefits to anything short of actual starvation. This case will be cited in future as a telling episode in the post-1950 judicial review story, but it is not quite clear what tale it tells. The bottom line is that English legal experience does not offer a great deal of support for any of the current political science modes of analysis of delegation. The formal legal doctrine makes little distinction between delegated legislation situations and those involving purely administrative implementation decisions. While the conventional twentieth century ultra vires view does emphasise courts as policers of principal-agent relations, the newer view, and that of the nineteenth century is that common law courts

Judicial delegation doctrines

167

are to be as concerned with the reasonableness, fairness, and rights of delegated legislation as with its obedience to the will of the delegator. If English courts really are much more active than they used to be, and we are not sure they are, that activity appears to be less a matter of judicial loyalty to parliament than of an increasingly influential view among judges that they are themselves independent actors insuring that others act lawfully, justly, fairly, and in accord with rights without much regard for whether those others are principals or agents or neither. Finally, while it is quite clear that English judges are more willing to act against lesser authorities than against government ministers and ministries, it is not at all clear that this reluctance is based on concerns about democracy rather than pragmatic concerns about relative political power. FRANCE The French Constitution provides for two primary law makers. The legislature makes lois. The executive, that is the prime minister and president, enact décrets. Both parliamentary statutes and these executive décrets may delegate supplementary law making powers to cabinet ministries, independent agencies, or private entities. The general term for such secondary legislation is réglement, meaning a general norm as opposed to a decision about a particular matter or person. However, many pieces of delegated legislation issued by ministries are entitled décret, and so are many pieces of delegated legislation issued by subnational government bodies. So some décrets are primary legislation, exercises of the law making authority directly invested in the executive by the constitution. Other décrets are delegated legislation whose bases are provisions either of parliamentary statutes or executive decrees delegating law making authority.21 In theory there is no constitutional judicial review of delegated legislation. The Council of State reviews delegated legislation for lawfulness, that is whether a piece of delegated legislation is in accord with the commands of the delegating statute and other laws. Among those other laws are international treaties and conventions to which France is a signatory and ‘general principles of law’. Given that France is a signatory of the European Convention on Human Rights and that French courts have announced many individual rights under the heading ‘general principles of law’, the Council of State exercises much rights review over secondary legislation although technically that review is lawfulness not constitutional review. The Council of State also exercises reasonableness or proportionality review over delegated legislation. Delegated legislation must achieve the goals of delegating legislation fairly, and without undue burdens on other goals or interests. Thus the Council of State in theory has review powers over delegated legislation as great or greater than American and arguably greater than British courts. The Council’s rights review can be as intrusive as American constitutional review and, at least until the recent passage of UK rights legislation, more intrusive than English review. Even now the Council has far greater remedial powers for rights violations by delegated legislation than do the English courts. For the English rights legislation authorises reviewing courts to declare a rights violation but not to actually invalidate the offending delegated legislation. The Council may do both.

The politics of delegation

168

The Council of State’s reasonableness review goes far beyond ‘Wednesbury’ unreasonableness and more openly and directly addresses policy issues than does US judicial review. The judicial section of the Council of State is both the highest French administrative court and a part of an institution of the executive whose duty it is to assure not simply lawful administration but good administration. Precisely because the Council of State is a part of administration rather than an independent judiciary, it is freer than a court would be to take up not only questions of lawfulness but substantive policy issues. The members of the judicial section of the Council, however, are career members of the Council as a whole, typically have served before and may serve again in other sections of the Council and often have been and may later again be seconded from the Council to one of the ministries or other organs of administration. Their experience qualifies them for activist review of administration but also intimately identifies them with those they are reviewing. In the light of its vast reviewing powers and its intimate connection with those it is reviewing, the key question becomes just how much does the Council intervene in delegated law making? Even more than in the British case, the question is very difficult for any outsider to answer. Although the Council fairly routinely strikes down the decrees of the smaller government units, it almost never strikes down a decree of the central government. Of course, here the usual problem of anticipated reaction arises. In the light of the very great capacity of the Council to review, and its demonstrated skills at doing so, those drafting secondary legislation would be foolish not to work hard to anticipate Council objections. There is every reason to believe that the normally intimate style of Parisian governance with much telephoning and lunching is operative here. Given the number of Council of State people seconded to the ministries, connections are particularly intimate. Not only is someone from the Council very likely to have been involved in the drafting of any decree, but the Council has a formal advisory role in the drafting process. The Council will have told the delegated legislation writers what is and is not lawful and reasonable at the drafting stage long before litigation challenging a decree comes to it as an administrative court. There is little reason why the Council as a court should disapprove a decree which as a paramount administrative organ it has had a significant role in drafting. Thus the Council of State may be a powerful and effective police agent over those wielding delegated law making authority, but much of the policing over those in central government may come at pre-litigation stages and leave no easily visible record. Where the prime ministership and the presidency are held by the same political party and that party enjoys a majority in the Assembly, it seems probable that the Council acts on behalf of the maker of primary law to assure that its delegates make the secondary law it wants. In earlier periods of multi-party coalition politics, and today when there is cohabitation, it is more difficult to see what principal the Council would be serving. Indeed, given the Council’s own internal character as a high esprit de corps career, professional civil service, and its intimate linkages with the rest of the French career civil service, it is quite possible that the Council is far less than the faithful agent of either the legislature or the political executive. ‘General principles of law’, elaborate but ultimately open ended concepts of ‘public order’ and ‘public service’, the doctrine of proportionality, that is that the implementing means employed to achieve a legislated goal not be excessive, plus expanding rights review, give the Council enormous independent discretion.

Judicial delegation doctrines

169

Certainly the Council’s actual case decisions formally recite, as do all French legal decisions, the statutory provisions on which they purport to rest. The language of statutory intent is frequently employed in the more discursive portions of the Council’s decisions. Yet it is also clear that the Council has its own long-term perceptions of good legal and administrative practice quite independent of the particular commands of particular pieces of primary legislation. The actual balance between Council concerns for enforcing statutory mandates and enforcing its own independent views of good law and good administration when policing those making regulations may shift case-by-case and cannot be specified with any precision. Of all the administrative review courts in the world the French Council of State exhibits the highest level of integration, continuity, and collective identity among its judges. The Council, through its case law, has virtually created the administrative law of France. It has long been an island of stability in a legal and political system subject to periodic crises. It is an acknowledged elite of the elites in a political system dominated by elites. A part of the French executive, it is unified where the executive power itself is divided. It long served and developed its institutional identity ‘under’ a French parliament noted for its divisions and incapacities. It is not clear that the Council’s peculiar role in the French legal system is adequately captured by principal-agent similes. It is clear that whenever the French legislature or executive chooses to delegate law making authority to entities subject to Council review, the delegator opens for itself not one principal-agent problem but two. COMPARING THE US, ENGLAND, AND FRANCE Now a brief comparative summary is in order. In all three polities, judicial review is commonly attached to and difficult to avoid in law making delegations. In all three, reviewing courts seem as concerned, probably more concerned, with imposing their own standards of justice, fairness, individual rights, good procedure, and reasonableness as they do with imposing the will of the principal. In all three, it is hardly possible to arrive at any precise measure of the degree to which judicial reviewers modify the balance of policy priorities specified by the delegating statute in the direction of their own priorities or the delegatees’ priorities. This problem is rendered even more insoluable by the frequent failure of delegators to themselves clearly state their policy priorities in primary legislation. It is difficult to assess national differences in levels of judicial activism, but it is also difficult to account for those differences. It is fairly clear that American courts have been far more active in striking down delegated legislation than have British or French. No doubt in part such differences arise from differences in training, recruitment, and career patterns of judges in the three countries. The American-English difference, however, is almost certainly partially a function of British cabinet dominance over parliament. In Britain, one of the agents has come to dominate the principal and by doing so also to dominate the other agent. An English reviewing court knows that should it find a piece of delegated law making unlawful under the primary law, the government will probably order parliament to change the primary law to fit the secondary one. Reviewing courts in the US enjoy far better odds of not being legislatively reversed. The French-American difference must in some part result from Council of State influence on the drafting of

The politics of delegation

170

delegated legislation before it is litigated. While the makers of regulations in the US also to some degree seek to anticipate the reaction of reviewing courts, Council of State early intervention goes far beyond this kind of anticipated reaction. In all three countries the potential restraint exercised over reviewing courts by potential or actual reversal by the primary law maker or delegator is being reduced by reviewing courts’ increasing reliance on law ‘higher’ than that of the delegating statute, be it constitution, common law, general principles of law, reasonableness, justice, fairness, or international agreements. Particularly the emphasis on human rights as somehow superior to mere statutory commands which pervades contemporary legal and political discourse increasingly moves courts in all three countries from agent towards principal. Similarly, the increasingly complex task of regulation of private enterprise in a nonsocialist, high tech, globalised economy tends towards broader and more ambiguous delegations by primary law makers to technocratic agents. Courts may respond to this tendency by opposing it, exploiting it to increase their own participation in regulatory policy making or deferring to technocratic expertise. While the Supreme Court has made some non-delegation noises, basically the courts of all three countries have acquiesced in increasing delegation. The courts of all three experience tensions between activism and deference. The Americans have been the most active, but with some current signs of retreat. The English have been least active, but show a rising level of activism, although not as great a rise as those favouring activism claim. Because French activism does not leave much of a litigation footprint, it is hard to measure it either over time or in comparison to that in other states. Given the Council of State’s integration into the whole administrative, technocratic apparatus it seems unlikely that it has served as a major opponent of technocratic government. Unlike US and UK courts, it is a government insider and so has not pressed as much for greater transparency and participation as the US and to a lesser extent English courts have. On the other hand, its ‘reasonableness’ review has been far more active than the English and more overt but probably less really intrusive than the US. Similarly, its ‘general principles’ review has been far more activist than the still disputed claims of British courts to move beyond a parliamentary sovereignty view of ultra vires. Certainly the primary basis of review in all three countries has been lawfulness, that is whether the delegated legislation is in accord with the delegating legislation and with other statutory law. In this sense, the primary role of courts is as one agent policing another on behalf of a principal, the primary law maker. In all three, however, judicial ‘interpretation’ of the statutes is unavoidable in the performance of this primary role. Such interpretation must to some degree become a vehicle for judicial law making outside and beyond the principals’ intention. Of course the vaguer and more ambiguous or internally contradictory the principal’s statute, the more independent of the principal’s intent the judicial interpretation becomes. That independence is constrained by the capacity of the principal to ‘correct’ judicial ‘errors’ of interpretation. And that capacity in turn depends on the principal’s facility in passing amending statutes. Thus we would expect the most judicial activism in the US because of Congress’ multi-roadblock law making process and the least in the UK with its draconian Cabinet control of parliamentary law making. Relative activism is difficult to measure, however, because of the phenomenon of anticipated reaction. UK and French administrative review courts

Judicial delegation doctrines

171

may have a greater influence on delegated legislation than the record of litigation would indicate because delegatees are very careful to anticipate their preferences. The US may appear more activist than it has been because it took time for the regulation writing administrative agencies to adjust to the rapid transformation of administrative law that occurred in the US and tailor their regulations and their regulation writing procedures to it. Beyond anticipated reaction, direct participation by Council of State personnel in the regulation writing process may make the visible level of litigation a quite untrustworthy measure of its activism. However uncertain the measurement of judicial activism in all three countries, the very role of policing agent for a delegating principal in some degree and inevitably turns the courts into principals imposing their own purposes on the administrative delegatees. Also, inevitably, there is a tension or rivalry between judicial statutory interpretation and administrative statutory interpretation. If courts entirely defer to administrative interpretation they abandon their primary role as policing agent. If they do not defer, to some degree they put themselves in the position of the ignorant telling the knowledgeable what to do. In all three countries administrative law jurisprudence provides us with some clues that a simple principal-agent model is not fully satisfactory for other reasons in addition to those just stated. That jurisprudence also warns us that issues of democracy may not be the only or the primary ones of interest to the judges. In all three, roughly the same norms are applied to public and private delegates of law making power. In all three, roughly the same norms are applied to administrative entities directly subordinated to political or majoritarian executives and those which are not. In France, where some decrees are themselves primary legislation flowing directly from the constitutional authority of the executive and other decrees are delegated legislation flowing from parliament’s primary legislation, the Council of State reviews both under roughly the same norms except of course that independent executive decrees need not be derived from parliamentary statutes. This French situation highlights a final important commonality. In all three countries administrative review courts do far more than test delegated legislation for lawfulness in the sense of compliance with the laws enacted by the legislature. All three do some form and degree of ‘reasonableness’ review designed to strike down substantively very bad regulations—how bad varying from country to country and time to time. All three do higher law review testing regulations not only against the specifics of the delegating statutes but against constitutional provisions, and/or general principles of law and/or treaty law and/or ultra vires notions that run beyond any direct statutory basis. Most importantly, all three, on some legal basis or another, are doing individual or human rights review. Such review always necessarily involves some balancing of the interests pursued by the delegated legislation against other interests that have somehow been given priority and called rights. All such balancing in turn pushes courts towards examining whether the legislation being reviewed achieves its purposes at the least possible cost to rights. For even where the state has a pressing enough interest to invade ‘rights’, surely it ought to invade them as little as possible to protect its interest. Thus, every rights court ultimately must to some degree substitute its legislative judgement for that of the law maker being reviewed, whether a primary or secondary law maker. For ultimately rights review must come down to judges asking whether one interest or another ought to prevail legally and whether even justified invasions of rights have been

The politics of delegation

172

accomplished by the best possible law—that is, the one legal alternative among many that invades ‘rights’ the least. Those, of course, are precisely the questions that congresses and parliaments should and do ask. In any event, the administrative review courts of all three countries seem as anxious, or more anxious, to ask whether delegated legislation meets various judicial criteria of goodness, fairness, justice, reasonableness, and rights protection as they are to ask whether delegated legislation is in accord with the will of the primary law making delegator or principal. NOTES 1. Martin Shapiro, The Supreme Court and Administrative Agencies (New York: Free Press 1968). 2. See Symposium, ‘Democracy and Delegation’, Cardozo Law Review 20 (1999), pp. 731– 1018; Whitman v. American Trucking Associations, 121 S. Ct. 903 (2001). 3. See Whitman v. American Trucking Association, 121 S. Ct. 903 (2001) and the cases cited there. 4. See ‘The Administrative Procedure Act: A Fortieth Anniversary Symposium’, Virginia Law Review 72 (1986), pp. 215–492. 5. Martin Shapiro, ‘APA: Past, Present, Future’, Virginia Law Review 72 (1986), pp. 447–92. 6. See particularly, Roderick Kiewiet and Matthew McCubbins, The Logic of Delegation (Chicago: University of Chicago Press 1991). 7. See Richard Stewart, ‘The Reformation of American Administrative Law’, Harvard Law Review 88 (1975), pp. 1667–813; Martin Shapiro, Who Guards the Guardians (Athens, GA: University of Georgia Press 1988). 8. Under some conditions Congress, of course, may correct the courts’ ‘misinterpretations’. See Barry Weingast, A Positive Theory of Statutory Interpretation (Palo Alto: Hoover Institution 1991); John Ferejohn and Barry Weingast, Limitation of Statutes: Strategic Statutory Interpretation (New York: Columbia Law School 1991). 9. See Sidney A.Shapiro, ‘Administrative Law After the Counter Reformation: Restoring Faith in Pragmatic Government’, University of Kansas Law Review 48 (2000), pp. 689–750. 10. While at various points I have used UK or British as appropriate, in this discussion of administrative law per se I have used England and English because Scottish courts operate under quite a different basic law system than the common law and Northern Irish courts, although common law courts, operate relatively independently. Where the High Court or the House of Lords reviews delegated legislation applicable to the whole of the UK, they are, of course, serving both as English and UK courts. 11. See Christopher Forsyth, ‘Of Fig Leaves and Fairy Tales: The Ultra Vires Doctrine’, Cambridge Law Journal 55 (1996), pp. 122–48. 12. H.W.R.Wade and C.F.Forsyth, Administrative Law (Oxford: Oxford University Press, 8th edn. 2000), pp. 865–6.

Judicial delegation doctrines

173

13. Wednesbury Cpn. v. Ministry of Housing and Local Government (No. 2) (1966) 2 Q.B. 275. 14. See Lord Lowry in Brind v. Secretary of State for the Home Department (1991) 1 AC 696; Lord Diplock in Council of Civil Service Unions v. Minister for the Civil Service (1985) AC 374; Lord Irvine, ‘Judges and Decision-Makers: The Theory and Practice of Wednesbury Review’, Public Law (1996), pp. 59–78. 15. See Bingham LJ in R. v. Ministry of Defence, ex parte Smith (1996) 1 All ER 256; Grainne de Burca, ‘Proportionality and Wednesbury Unreasonableness’, European Public Law 3 (1997), pp. 561–86. 16. Most authoritatively stated in successive editions of H.W.R.Wade, Administrative Law, the latest of which is H.W.R.Wade and C.Forsyth, Administrative Law. (Oxford: Clarendon Press, 7th edn. 1994). 17. See Paul Craig, ‘Ultra Vires and the Foundations of Judicial Review’, Cambridge Law Journal 57 (1998), pp. 63–90; Sir John Laws, ‘Illegality: The Problem of Jurisdiction’, in M. Supperstone and J.Goudie (eds.), Judicial Review (Oxford: Oxford University Press, 2nd edn. 1998); Sir Stephen Sedley, ‘Human Rights: A Twenty-First Century Agenda’, Public Law 37 (1995), pp. 386–407. 18. See R. v. Monopolies and Merger Commission ex parte South Yorkshire Transport Ltd. (1993) 1 WLR 23. 19. See Susan Sterett, Creating Constitutionalism? The Politics of Legal Expertise and Administrative Law in England and Wales (Ann Arbor: University of Michigan Press 1997). 20. (1997) 1 WLR 275. 21. In general for this entire section see Rene Chapus, Droit Administratif General Tome 1 (Paris: Montchrestien, 13th edn. 1999); Yves Meny (ed.), Il Consiglio di stato in Fracia e in Italia (Rome: Il Mulino 1994).

Learning from the Americanists (Again): Theory and Method in the Study of Delegation

MARK A.POLLACK It is no secret that European(ist) scholars have largely followed—at a greater or lesser distance—their American(ist) colleagues in the formulation of theories about delegation of powers to non-majoritarian institutions. This leading role for Americanists reflects in part the longer historical tradition of governance through independent regulatory agencies and their accompanying mechanisms of administrative law and congressional oversight, and in part the greater influence in America of rational choice scholars, who were the first to theorise explicitly about the control relationship between legislative principals and bureaucratic agents. In recent years, the Europeanist literature has begun to catch up with the Americanists in theoretical terms, in part because the European Union and its member states have resorted increasingly to delegation to non-majoritarian agencies and central banks, and in part because European scholars have imported the relevant literatures on regulation and principal-agent interactions. This chapter argues that Europeanists can once again learn from American debates, not only about theory, but also about the methods that we use to test our hypotheses about delegation to, and autonomous action by, non-majoritarian institutions. The argument is not that Europeanists should mechanically adopt the theories and methods advocated by Americanists (for as we shall see many of their methods run into problems when applied outside the US context), but rather that Americanists appear to have thought more systematically about the difficulties of testing hypotheses about delegation and agency, and that these methodological debates have in turn pushed Americanists in new directions that are being followed—at an ever-decreasing distance—by Europeanists. The chapter is divided into three sections. The first section discusses the challenge of testing hypotheses about the conditions under which executive or judicial agents might enjoy some degree of autonomy from their legislative principals; it suggests that Americanists have made considerable progress in overcoming these challenges, taking full advantage of the extraordinary availability of quantitative as well as qualitative data in the American context. At the same time, however, it notes that many Americanists have recently despaired of ever being able definitively to test hypotheses about agency. This despair, it is argued, is premature, but it suggests that students of delegation and agency outside the American context should learn from the experience of Americanists,

Theory and method in the study of delegation

175

and draw from that literature those methods that are in fact transferable to the European and other contexts. The second section examines the development in American politics of a second and more recent wave of principal-agent analysis which, having given up on the notion of measuring agency directly, focuses instead on devising and testing hypotheses about the conditions under which legislative principals might delegate authority and discretion to bureaucratic agents. The third and final section examines some preliminary applications of the principal-agent approach to the European Union and to the comparative study of European parliamentary democracies, and proposes a research agenda for the comparative study of national-level delegation in the parliamentary systems of Western Europe. AGENCY VS. POLITICAL CONTROL: THE PROBLEM OF MEASUREMENT In their introduction to this volume, Alec Stone Sweet and Mark Thatcher have introduced the basic elements of principal-agent analysis, and the most common hypotheses regarding the conditions under which executive and judicial agents should be able to act autonomously and influence political outcomes. As Stone Sweet and Thatcher point out, the American politics literature on agency arose largely in the context of a debate over the autonomy and accountability of American regulatory agencies. According to this early debate, such agencies were depicted either as obedient servants which implement without distortion the mandates given to them by the US Congress (the ‘Congressional dominance’ school) or as runaway bureaucracies implementing their own preferences or those of their industrial clients at the expense of a Congress whose oversight is lax and ineffective (the ‘Congressional abdication’ or ‘runaway bureaucracy’ school). Later, the American literature developed beyond this binary debate, in favour of a more subtle hypothesis according to which the discretion of an agent—say, a regulatory bureaucracy like the Federal Trade Commission or the Environmental Protection Agency—depends on the relative preferences of the agent and its principals in the Congress and the presidency, and on the specific oversight mechanisms (administrative and oversight procedures) that serve to constrain agents in the pursuit of their preferences. American regulatory agencies, for example, may be constrained by more or less restrictive administrative procedures laid down in the act of delegation (including inter alia the level of detail prescribed by legislation, the instruments available to the agents, requirements to hold hearings and report regularly to Congress, the threshold and guidelines for judicial review of agency acts, and the possibility of removing and appointing heads of agencies), and by more or less active oversight by congressional committees and subcommittees. Ceteris paribus, the American literature tells us, we should expect agencies to be most responsive to their legislative principals when control mechanisms are extensive, and most autonomous when administrative procedures are few and oversight is weak. However, if the hypotheses generated by principal-agent analyses are relatively straightforward, the methods for testing these hypotheses are far less self-evident. In

The politics of delegation

176

essence, the core problem of testing hypotheses about agent autonomy is that agents such as US regulatory agencies, the European Commission, or courts (see Alec Stone Sweet’s contribution to this volume) may rationally anticipate the reactions of their principals, as well as the probability of sanctions, and adjust their behaviour to avoid the costly imposition of sanctions. If this is so, then agency behaviour that at first glance seems autonomous may in fact be subtly influenced by the preferences of the principals, even in the absence of any overt sanctions. Indeed, as Weingast and Moran point out, the more effective the control mechanisms employed by the principals, the less overt sanctioning we should see, since agents rationally anticipate the preferences of the principals and the high probability of sanctions for shirking, and adjust their behaviour accordingly. Studies of agency autonomy that rely on the frequency of sanctioning are therefore likely to run into the methodological problem of observational equivalence, namely that the absence of sanctions is consistent with both the obedient servant and the runaway bureaucracy scenarios, each of which predicts, albeit for different reasons, the rarity of sanctions.1 Despite these problems, Americanists have devised a number of methods to test hypotheses about the existence and (to a lesser extent) the determinants of presidential and congressional control over US regulatory agencies. In some cases, these scholars rely on process-tracing case studies to study the interaction between congressional oversight committees and the regulatory agencies they oversee. In a 1985 study of the Federal Communications Commission (FCC), for example, Ferejohn and Shipan test a ‘signalling’ model of principal-agent interactions, in which congressional committees signal to regulatory agencies their intent to overrule regulatory decisions through new legislation, in order to influence apparently ‘independent’ regulatory decisions. Specifically, Ferejohn and Shipan examine the FCC’s 1982 decision to charge telephone consumers a monthly access charge as part of the ongoing deregulation and reform of long-distance service in the United States. Although such a move was widely welcomed by economists and by some members of the industry, congressional representatives feared that the FCC decision would have political consequences for Congress, and both the House and Senate oversight committees held hearings regarding the decision and introduced legislation which—if adopted on the floor of both houses—would have overturned the decision. Using extensive qualitative evidence from the hearings, the authors demonstrate that Congress acted with the deliberate intention of putting pressure on the FCC, which in turn revised its decision four separate times, apparently in response to congressional pressure, before adopting a scaled-back access charge in 1984, with implementation delayed until 1985. At the same time, however, Ferejohn and Shipan also note that the FCC did not abandon its plan entirely, but rather compromised in a series of steps until the threat of congressional legislation had passed. The result is a picture of congressional-agency relations in which the former can influence the latter—albeit imperfectly, and with varying amounts of agency slack—through a clearly signalled intention, and a credible threat, to legislate in the case of agency shirking.2 Nevertheless, as Calvert, Moran and Weingast point out in their study of the Federal Trade Commission (FTC), case studies may in fact prove misleading in studies of agency behaviour, because of the problem of observational equivalence.3 In any given case, they argue, the absence of sanctions may reflect either autonomous agency behaviour (the ‘runaway bureaucracy’ view) or the rational anticipation of congressional preferences by agencies (the ‘congressional dominance’ view). Indeed, even a case where the Congress

Theory and method in the study of delegation

177

does sanction an agency—as it famously did with the FTC in 1979 and 1980—such sanctions may reflect either a legislative correction of an agency engaging in shirking, or a change in preferences among the principals themselves. For this reason, the authors adopt a large-n, comparative-statics research design to test their hypothesis that the behaviour of regulatory agencies is responsive to the preferences of their political principals in congressional oversight committees. Put simply, their research design consists in measuring the preferences of congressional principals (the independent variable) and the behaviour of regulatory agents (the dependent variable), and establishing a correlation between these two variables. For the preferences of the principals, they consult individual voting scores of liberalism collected by Americans for Democratic Action (ADA) for the chairmen and members of the relevant House and Senate oversight committees, producing indices of committee support for agency activism over a period of 15 years straddling the events of 1979, and demonstrating the sharp change in committee preferences as a result of membership turnover between 1976 and 1979.4 Similarly, to measure agency behaviour, they examine the distribution of various kinds of cases (consumer credit, textiles, and mergers) brought by the FTC in any given year, on the assumption that a change in the distribution of cases represents a change in the priorities of the agency. Over the entire period, they demonstrate, the preferences of congressional committees and chairmen correlate positively with changes in the distribution of the FTC’s case-load, suggesting that the latter are politically responsive to the changing preferences of the former, even during the pre-1979 period which witnessed no apparent conflict between the agency and its congressional principals. In a later article, however, Terry Moe takes Calvert, Moran and Weingast to task for focusing exclusively on congressional preferences as an explanation of FTC behaviour, and ignoring competing explanations such as the internal reorganisation of the Commission during the 1960s and 1970s, or additional pressures brought to bear by the president and other political actors. The problem, Moe suggests, is not simply theoretical but also empirical, insofar as the Calvert-Moran-Weingast design fails to control for any non-congressional influences on FTC behaviour, and especially on the mix of cases, which he demonstrates varied considerably over time for reasons that may have had little or nothing to do with the preferences of Congress. The lesson of this study, according to Moe, is that any test of principal-agent hypotheses should not focus exclusively on the correlation between principals’ preferences and agency behaviour, but should be attentive to and control for competing explanations, particularly those internal to the agency itself.5 In this context, a 1988 article by Wood adopts a research design similar to Weingast and Moran, while simultaneously addressing some of Moe’s criticisms, including consideration of the history and internal politics of regulatory bureaucracies as well as the roles of the president and Congress as multiple principals. In empirical terms, Wood examines the behaviour of the Environmental Protection Agency (EPA) during the early years of Ronald Reagan’s presidency, when the traditional activism of the EPA, which had been supported by the president during the Carter years, came under attack as a result of electoral turnover in one of the key principals. Specifically, Wood argues, the EPA’s vigorous enforcement activities came under assault from two fronts. First, immediately upon taking office, the Reagan administration appointed a new EPA administrator, Anne Gorsuch, and other personnel committed to reducing the activism of the agency and

The politics of delegation

178

providing regulatory relief to industry. Second, at Reagan’s request, the Congress drastically reduced the agency’s budget for enforcement activities from fiscal year 1982 onwards. According to standard principal-agent models, Wood argues, we should expect to see EPA enforcement activity decline during this period, as the agency’s principals employed both personnel decisions (from January 1981) and budgetary cuts (from October 1981) in order to restrict the agency’s discretion, and hence its traditional activism. In addition, he argues, we should expect to see EPA activity revive after the March 1983 nomination of the moderate William Ruckleshaus to replace Gorsuch. In order to test these hypotheses, Wood adopts research design that, rather than measuring EPA activity before and after Reagan’s inauguration, ‘zooms in’ on agency behaviour, measuring the EPA’s monitoring and enforcement actions month by month, using information obtained under the Freedom of Information Act. By adopting such a finegrained measure of the dependent variable, Wood is able to examine the timing of changes in agency behaviour, and hence isolate more effectively the effects of personnel decisions (in January 1981 and March 1983, respectively) and budgetary decisions (from fiscal year 1982 onwards). Interestingly, Wood’s findings speak to both method as well as theory. If we aggregate EPA monitoring and enforcement actions annually, he demonstrates, the pattern seems to demonstrate precisely the effects predicted by a model of perfect presidential control, namely a sharp drop-off in enforcement actions in 1981, with an upward trend resuming after Ruckleshaus’ appointment in 1983. However, when broken down into monthly indicators, the same data reveal additional details, including a post-inaugural spike in enforcement actions during the two months straddling Reagan’s inauguration, followed by a gradual decline during the first six months of the administration, and then a sharp decline from October 1981, suggesting that the budgetary constraints introduced during the 1982 fiscal year 1982 had a greater impact on EPA activity than the personnel policies of the new administration alone. Perhaps most interestingly, Wood also finds that, after the initial shock of the new administration and the subsequent cuts in its budget, EPA enforcement activity actually increased gradually during 1982, even before the resignation of Anne Gorsuch; and that enforcement activity after 1983 actually surpassed the levels reached during the Carter administration. From this evidence, Wood concludes that the new Reagan administration was unable to overcome the EPA’s inherent preference for strong environmental regulations, and that ultimately the EPA bureaucracy was willing and able to resume vigorous enforcement actions despite the existence of clear pressure for restraint from both the president and Congress. He ends on a methodological note, warning that fine-grained, ‘high-resolution’ data is required to capture the precise timing of changes in agency behaviour, and to pinpoint the effects of specific shocks on that behaviour.6 In sum, American students of regulatory bureaucracy have studied the existence and the determinants of agent autonomy through both case study analysis and through large-n studies of the correlation between principals’ preferences and agency behaviour. Both approaches have strengths as well as weaknesses. Case studies like Ferejohn and Shipan’s study of the FCC have the advantage of tracing the respective preferences and interactions between principals and agents, focusing on the methods used by principals to exert pressure on agents (for example, the threat of legislative overruling) and the timing and nature of concessions by the latter. However, as Weingast and Moran point out,

Theory and method in the study of delegation

179

single case studies raise significant issues of case selection, and they may mask the ‘invisible’ effects of rational anticipation by agents. By contrast, large-n studies like those by Calvert, Moran and Weingast and by Wood have the advantage of demonstrating the existence and strength of correlations between principals’ preferences and agency behaviour over a large number of cases and over time. As we have seen, however, such quantitative studies are generally more successful in demonstrating that congressional principals exert some degree of control over the bureaucracy (for example, Calvert, Moran and Weingast), or conversely that the bureaucracy possesses some degree of autonomy (for example, Wood), than they have been at assessing the degree of control or autonomy in each case, much less the determinants of control and autonomy (although Wood’s design fares better along this criterion than do Calvert, Moran and Weingast). Furthermore, such studies rely on quantitative indices of actor preferences and agency outputs that are readily available in the United States, with its long-standing practice of keeping voting records, ADA scores, party breakdowns in the various branches of government, and records of individual agency decisions. Looking beyond the United States, the availability of such data is likely to vary considerably, raising additional questions of measurement as well as research design—a point which will be returned to below. FROM AGENCY TO DELEGATION Faced with these methodological obstacles, an increasing number of scholars have opted to concentrate their empirical analysis not on the behaviour of the agent, but rather on the delegation stage, where principals make their strategic decisions about the powers delegated to agents, and the control mechanisms established to limit their autonomy. Huber and Shipan, for example, argue that existing principal-agent analyses of regulatory bureaucracies have forced scholars to reconsider their views about the existence or absence of congressional control over bureaucrats. Such studies, however, are not without their drawbacks. One problem is that these studies tend to focus on the presence or absence of political control itself: either legislators control bureaucratic behavior or they do not. From the perspective of transaction cost theory, either claim is plausible and neither is likely verifiable… To validate an argument about the expected level of political control, one would have to establish that the actual level of control is consistent with the predicted level, given the transaction costs and the political institutions that are present. None of the theories we have discussed approaches a level of specificity that could serve as relevant for such tests. Even if they did, it would be difficult to imagine a data set that could be used to test such predictions. Not surprisingly, most existing evidence for the theories has involved case studies, which, while illuminating, can hardly be viewed as definitive.7 For this reason, Huber and Shipan argue, scholars are increasingly turning away from direct studies of agency behaviour, focusing instead the delegation stage:

The politics of delegation

180

A significant feature of the more recent research [they write] is its focus on variation in instruments of control as dependent variables. This represents a turn away from questions about the degree of control, which dominated much of the early literature. This change is sensible. If the question is, “Do legislatures control bureaucracies?” the answer is always somewhere in the gray area. Consequently, although it may be possible to establish empirically whether decision making is affected by present or past legislative preferences, it is much more difficult to make black or white claims about the presence or absence of control, and thus about the validity of particular theoretical arguments. Not surprisingly, then, empirical tests that focus on the instruments of control have proven easier to undertake.8 Indeed, an increasing number of American scholars have recently turned to the literature on the transaction costs of policy making to generate specific predictions about the conditions under which legislative principals will allocate discretion—defined by Epstein and O’Halloran as the net of initial delegation minus the administrative and oversight mechanisms established to limit shirking—to their agents.9 Simplifying tremendously, such studies typically point to three aspects of the political environment that influence the transaction costs of policy making, and therefore the desirability of delegating powers and allocating discretion to agents. (1) Imperfect information, and the demand for policy-relevant expertise. The existence of imperfect information—or, more specifically, legislators’ need for policy-relevant information—is perhaps the most important transaction cost cited by rational choice scholars as justification for the delegation of powers from legislators such as the US Congress to regulatory agencies like the EPA or the FTC. The argument here is straightforward: the empirical world is inherently uncertain, and legislators face constant demands for policy-relevant information about the state of the world (for example, the effect of a given policy on the natural environment or the safety requirements for civil aviation). According to Krehbiel, legislators create congressional committees precisely to provide their members with an incentive to acquire such policy-relevant information; yet, as Epstein and O’Halloran explain, the congressional committee system itself imposes transaction costs which may make it desirable for Congress instead to delegate power to a regulatory agency, which is assumed to provide policy-relevant expertise and thereby improve the quality of regulation.10 This core assumption—that regulatory agencies are created and delegate powers largely in order to provide expert, policy-relevant information to legislators—is central to the analyses of many recent studies, including Bawn, Epstein and O’Halloran, and Huber.11 In the view of these and other authors, legislative principals should delegate greater discretion to regulatory agencies as a function of the uncertainty in any given issue-area, although the latter is has proven quite difficult to measure in quantitative terms. (2) Credible commitments, and the demand for an independent, credible regulator. Despite the importance of imperfect information in principal-agent models, transactioncost theorists have offered a second, analytically distinct rationale for delegation, stressing the need for legislators to establish the credibility of policies. More specifically, students of regulatory bureaucracies and especially independent central banks have

Theory and method in the study of delegation

181

suggested that legislators, having adopted a given policy, might be tempted to relax that policy to appease concentrated interests or to increase their popular support on the eve of an election. For example, congressional legislators may find it technically or politically efficient to adopt a rigorous antitrust policy against business cartels and concentrations, or to close inefficient and expensive military bases, but in each case individual legislators would be tempted to be lenient faced with protests from concentrated interests (such as large firms or military bases) in their own constituencies.12 Similarly, legislators and/or governments may prefer to adopt tight, anti-inflationary monetary policies, but may find themselves unable to commit to this policy because of the subsequent temptation to reduce interest rates and stimulate the economy on the eve of an election. In this view (subsequently extended to the European Union), an independent regulator or central bank that was insulated from political pressures would enjoy greater credibility vis-à-vis political constituencies and markets, and legislators would benefit by delegating authority to such non-majoritarian organisations.13 If this view is indeed correct, we should expect legislative principals to delegate powers, not (only) in issue areas marked by relative uncertainty, but (also) in issue areas where legislators find themselves unable to commit credibly to a particular line of policy. (3) Conflicting preferences among principals, and between principals and agents. A third factor often cited in principals’ decision to delegate authority and discretion to an agent is the degree of conflicting preferences either (a) among multiple principals or (b) between principals and agents. In the first case, the existence of conflicting preferences among multiple principals increases the appeal of detailed administrative provisions, insofar as such conflicts complicate subsequent efforts to punish shirking by an agent.14 In the second and more intuitively obvious case, the existence of conflicting preferences between principals and agents increases the likelihood of significant agency costs, and therefore decreases the desirability of delegation for the principals. In their study of congressional delegation to regulatory bureaucracies, for example, Epstein and O’Halloran begin with the assumption that the president nominates the heads of independent regulatory agencies, and that divided government therefore serves as a proxy for conflicting preferences among congressional legislators on the one hand and regulatory agents on the other; and they do indeed find that Congress delegates less authority and less discretion to the executive branch during periods of divided government.15 Similarly, Huber suggests that ‘granting substantial autonomy to civil servants will have the greatest payoff to ministers when conflict of interest between ministers and civil servants is low. As conflict increases, the benefits of granting autonomy decrease’. Specifically, Huber hypothesises that the likelihood of such conflicts increases in the presence of politicised bureaucracies, coalition governments, and portfolio volatility, and that legislators and governments should therefore delegate less authority and discretion to civil servants under such conditions.16 Although systematic analyses of institutional choices for delegation are still relatively rare, recent empirical studies of congressional delegation by Epstein and O’Halloran, Bawn and others suggest that congressional legislators do indeed systematically delegate powers and allocate greater or lesser discretion to regulatory agencies, and that these institutional choices do indeed vary with at least some elements of the political environment.17 The promise of such studies, and their ability to measure delegation, discretion, and their relationship with various features of the institutional environment,

The politics of delegation

182

can best be illustrated through a brief discussion of Epstein and O’Halloran’s ambitious book, which constitutes the state of the art in the study of American congressional delegation. Epstein and O’Halloran begin by creating a formal model which in turn generates a set of nine hypotheses about the degree of discretion likely to be delegated to agents under varying conditions, the most important and generalisable of which are the presence of conflicting preferences (as measured by divided government) and the degree of uncertainty in any given issue area (for which they adopt two distinct measurements). In order to test these hypotheses, Epstein and O’Halloran create an updated version of Mayhew’s set of significant post-war legislative enactments, which serve as the sample for their study of delegation.18 For each of the 257 laws in their dataset, the authors adopt simple measures of delegation (that is, the percentage of major provisions which delegate powers to an executive agent) and discretion (an ordinal scale ranging from 1 to 14 and comprising the presence or absence of each of 14 distinct control mechanisms), and they provide a full chapter of descriptive statistics demonstrating that congressional decisions about delegation and discretion vary systematically across issue areas and over time. In later chapters, they analyse the correlation between their measurement of the dependent variable—discretion—and their key independent variables—including divided government and uncertainty—finding that these factors do indeed appear to influence the congressional decision to delegate power and discretion to executive agencies.19 Although we may collectively debate specific aspects of Epstein and O’Halloran’s study (including, for example, the measure of discretion, which weighs all 14 control mechanisms equally), it nevertheless breaks new ground in measuring systematically the congressional decision to delegate powers to executive agencies, and in testing hypotheses about the conditions under which greater or less discretion will be granted to such agencies. The implications of such studies for the comparative analysis of delegation are explored in the final section of this chapter. THEORY, METHOD, AND THE STUDY OF DELEGATION IN COMPARATIVE EUROPEAN POLITICS What can comparative Europeanists learn from the Americanists—and, conversely, what might comparative Europeanists be able to offer to Americanists in return—regarding the study of delegation? The question is not as simple as it may appear at first blush, and it seems clear that both the theory and the research designs of American studies can be applied beyond the US context only with considerable adaptation. Precisely because they were developed by Americanists with the aim of understanding the specific institutional features of the American political system—that is, congressional committees and regulatory agencies—the theories analysed above have been developed at a relatively low level of abstraction, which makes it difficult to transfer their insights to the very different, and primarily parliamentary, political systems of Western Europe.20 Similarly, many of the empirical methods devised by Americanists to test hypotheses about delegation in the United States presuppose the availability of data about the preferences and behaviour of the key actors, such as roll-call votes, ADA ratings of individual Congressmen, committee voting records, Congressional Quarterly legislative summaries, the partisan

Theory and method in the study of delegation

183

balance of the Congress and the presidency, and the regulatory output of agencies. While this or comparable data is available in some European countries, research designs created in an American context are likely to require significant adaptation in European cases. To the extent that rational-choice students of delegation have turned their attention to comparative politics, their initial efforts have been primarily theory-driven, seeking (1) hypotheses that ‘travel’ beyond the US case and (2) research designs and measures for variables such as actor preferences, agency behaviour, political conflict, political uncertainty, and agency discretion outside the US congressional setting. In Delegating Powers, for example, Epstein and O’Halloran include an afterward on comparative institutions. In it, they point out that their model of delegation is most obviously applicable to other presidential systems; even here, however, they concede that all presidential systems are not alike in the distribution of powers between the legislature and the executive, and that presidents in many other systems enjoy legislative powers such as a line-item veto or the exclusive right to introduce legislation in certain policy areas. Thus, the applicability of their model of congressional delegation to other systems would depend largely on the respective strengths of the executive and legislature in proposing and adopting legislation, and on other factors such as the nature of legislative committees that play a central role in their model. Moving beyond presidential systems to the parliamentary systems that constitute the majority of West European countries, Epstein and O’Halloran admit that the theory is less obviously applicable to the latter, but that testable hypotheses may nevertheless be generated regarding legislative-executive relations in parliamentary democracies. For example, they suggest that, insofar as conflicts emerge between prime ministers and parliaments, the latter should enjoy a greater incentive to create their own policy making prerogatives by delegating powers to parliamentary committees or independent agencies. Most broadly, they suggest, the delegation of powers in a given political system should be understood and explained ‘relative to the next best feasible alternative’, which is likely to vary across different political systems.21 Precisely because Americanist theories of delegation have been devised with a separation-of-powers system in mind, the earliest and most sustained research programme on delegation and agency outside the United States has been undertaken by students of the European Union. Although not a national political system, the Union has a number of characteristics—a strong, constitutionally defined vertical and horizontal division of powers, and extensive delegation to an executive branch, courts, an independent central bank and various regulatory agencies—which make it analytically similar to the US political system, and hence a promising testing ground for Americanderived theories of delegation and agency. In the EU case, member state governments are typically modelled as the collective principals, delegating powers to the Commission, Court, European Central Bank and other regulatory agencies either through the EU’s constitutive treaties or through secondary legislation. Applied to the European Union, principal-agent analysis therefore leads to the hypothesis that agency autonomy is likely to vary across issue areas and over time, as a function of the preferences of the member governments, the distribution of information between principals and agents, and the decision rules governing the application of sanctions or the adoption of new legislation.22 Not surprisingly, however, students of the Commission and other EU agents have encountered now familiar methodological obstacles to measuring Commission agency

The politics of delegation

184

and member state control, insofar as the Commission rationally anticipates the reactions of its principals and adjusts its behaviour in advance to avoid the costly imposition of sanctions. Here again, we are in the realm of anticipated consequences and observational equivalence. Thus, although there is no shortage of empirical studies asserting an independent causal role for the European Commission,23 many of these studies are guilty of selecting on the dependent variable for most likely cases of Commission influence; few make any attempt to identify the conditions for Commission influence; and even fewer attempt to deal systematically with the consequences of the ‘law of anticipated reactions’. In recent years, Susanne Schmidt and this author have undertaken preliminary efforts to test principal-agent hypotheses in the EU through process tracing and the use of comparative case studies (selecting cases across the independent variables hypothesised to determine Commission discretion). Such studies suggest that the Commission does indeed ‘matter’—in the sense of using its limited and variable discretion to pursue an integrationist agenda across a number of issue areas—but their cases do not constitute a representative sample of Commission activity, and additional studies are needed.24 More recently, a number of scholars have made the Americanist-inspired leap from the study of agent behaviour to the institutional design of delegation, with a special emphasis on the arcane issue of comitology, the committees of member state representatives established to supervise the Commission in its implementation of EU law. Although often depicted by legal scholars as the site of technocratic deliberation, in which the aim is collective problem solving rather than control over the Commission bureaucracy,25 comitology committees actually come in seven different variants with distinct voting rules, which have been shown in formal models to place varying degrees of constraint upon the Commission in its activities.26 In recent empirical studies, moreover, Rhys Dogan and Fabio Franchino demonstrate that the EU’s Council of Ministers adopts systematically distinct committee structures across issue areas, suggesting that comitology is indeed employed consciously as a control mechanism by member states. Both authors find (unsurprisingly, in light of P-A analysis) that the Commission has displayed a consistent preference for less restrictive procedures, and that this preference is shared by the European Parliament. However, they also find (again unsurprisingly) that the Council of Ministers regularly adopts more stringent procedures than those proposed by the Commission, and that the strictest procedures are clustered in certain issue areas such as social policy, phytosanitary standards, and environmental protection. In addition, Franchino’s analysis suggests that the Council adopts the most stringent regulations in areas characterised by high levels of uncertainty and conflict, although these findings should be regarded as tentative given the difficulty of finding reliable quantitative proxies for concepts such as uncertainty (which Franchino measures using a word count of the legislation in question) and conflict (measured by the number of amendments adopted by the Council).27 Moving beyond the EU, John Huber and other scholars have recently begun to formulate explicitly comparative hypotheses about delegation and discretion in various political settings, including parliamentary as well as separation-of-powers systems. In a recent article, for example, Huber, Shipan and Pfahler undertake a comparative analysis of delegation decisions across the 50 US states. Drawing from the transactions-cost literature, the authors hypothesise that various characteristics of the political environment should influence the decision by legislative principals to delegate greater or lesser

Theory and method in the study of delegation

185

discretion to state government agencies. Specifically, they hypothesise that a number of independent variables—including not only divided government, but also as a divided legislature, the professionalisation of the legislature, and the availability of a post hoc legislative veto (as an alternative means of control)—should correlate with the delegation of ex ante discretion in the area of health-care provision. The dependent variable of discretion, in this case, is measured by the number of new words adopted by a given state during a given year; although a less detailed and fine-grained measure than that adopted by Epstein and O’Halloran, the authors argue that the number of new words captures an important aspect of discretion, insofar as more detailed legislation constrains bureaucrats more than a shorter legislative command simply to implement a programme. In the event, the authors find support for their claims that divided government and legislatures do indeed influence delegation decisions, as do the professionalisation of the legislature (in cases of divided government) and the availability of a legislative veto as an ex post alternative to ex ante restrictions.28 In addition to their comparative study of US states, Huber and his colleagues have recently hypothesised that specific features of parliamentary systems of government— most notably portfolio volatility among ministers and political parties—might influence ministers’ decisions to delegate greater or lesser discretion to the civil service. Taking issue with the classic argument that portfolio volatility necessarily decreases ministerial control over civil servants, Huber and Lupia’s formal model of a delegation suggests that the effects of portfolio volatility on ministerial control may in fact vary as a function of a number of other variables, including the counter-intuitive finding that agency losses are most likely precisely where a bureaucrat’s preferences are closely aligned with those of the incumbent minister (but presumably not with those of the minister’s likely successor).29 More generally, Huber suggests that ministers in countries with high portfolio volatility need not accept passively the loss of ministerial control, but may act deliberately to establish institutions to limit the adverse effects of volatility by decreasing civil servants’ discretion. In empirical terms, he examines efforts at cost-containment in the health-care sector across 18 advanced industrial democracies, showing a correlation between short-term portfolio instability and efforts to control the rising cost of health care, but no correlation at all between long-term portfolio volatility and cost-control— suggesting that ministers in such countries do indeed design institutions to compensate for any agency losses resulting from endemic volatility.30 In a later article using data from the same countries, Huber finds that countries with higher portfolio volatility are more likely to adopt ‘global budgets’, which limit bureaucratic discretion in health-care expenditures, and allow ministers to achieve results similar to those found in countries with relatively stable cabinets. Although he concedes that such findings are tentative, Huber concludes with a call for more rigorous theoretical modelling of delegation decisions and continued efforts to test hypotheses about ‘how political and institutional features affect systematic variation in delegation processes across political systems’.31 Finally, moving beyond the relatively narrow theme of delegation to civil servants and other non-majoritarian institutions, Torbjörn Bergman, Wolfgang Müller and Kaare Strøm’s recent special issue of the European Journal of Political Research, broadens the focus to examine ‘parliamentary democracy and the chain of delegation’. Acts of delegation, and the resulting principal-agent problems that arise from them, are not limited to regulatory agencies and courts, the editors argue. Indeed, the very concept of

The politics of delegation

186

parliamentary democracy is predicated on the successful management of a chain of delegation with at least four discrete steps: 1. From voters to elected representatives. 2. From legislators to the executive branch, specifically to the head of government (the prime minister). 3. From the head of government (prime minister) to the heads of different executive departments. 4. From the heads of different executive departments to civil servants.32 From this perspective, ‘agency losses can occur in any link in the parliamentary chain of delegation, and only if the whole chain works can the democratic promise be fulfilled’.33 Accordingly, individual essays in the book isolate and examine the agency issues present within each of these links in the chain of delegation, and the ways in which the principals at each stage attempt to keep their agents accountable through various institutional mechanisms.34 By and large, the essays in Bergman, Müller and Strøm’s special issue are theoretical rather than empirical, and their methodological lessons for us here are therefore necessarily limited. Nevertheless, these studies represent the beginning of a promising comparative research programme, which decreases the parochialism of the American literature on delegation while at the same time generating hypotheses about delegation that can be tested rigorously and comparatively both within and across national boundaries. Indeed, following the essays presented in this volume, we may add three additional stages of delegation, which, although not part of the classic model of parliamentary democracy, have become increasingly widespread among West European democracies in the past two decades and can profitably be studied through the lens of principal-agent analysis: 5. From legislators or governments to private actors who deliver public services on behalf of the government; such delegation, in which government services are ‘contracted out’ to private firms or voluntary organisation, is a long-standing practice in most countries, but has arguably grown in importance in recent years.35 6. From legislators to non-majoritarian institutions deliberately selected as alternatives to ordinary government regulation in areas such as monetary policy, utilities, telecommunications or financial regulation.36 7. From legislators and governments to EU institutions.37 With these seven sets of questions, we have now travelled a substantial distance from the narrowly focused early literature on delegation to regulatory agencies in the United States to a much broader and explicitly comparative research agenda. The claim here is not simply that parliamentary democracies delegate powers to non-majoritarian institutions in the American mould—although the chapters in this volume clearly suggest that they do, and with increasing frequency—but more generally that the theoretically informed study of delegation has much to tell us about the nature of parliamentary democracy in contemporary Europe. In conclusion, the argument of this brief review has not been to suggest that Europeanists should slavishly follow the theory and methods of Americanists; indeed, it has been argued above that American theories of delegation need to be specified at a

Theory and method in the study of delegation

187

higher level of abstraction in order to travel beyond the specificities of the American separation-of-powers system, while methods must be adapted to the availability of data in EU and national European contexts. Nevertheless, students of comparative European politics can learn from the methods adopted by students of American politics, and indeed the data problems faced by these scholars are arguably less severe than for students of the more opaque EU institutions. In conclusion, therefore, two lessons for comparativists from the American literature can be offered—both of which offer significant grounds for optimism. First, regarding the study of agency behaviour, Americanists are correct to suggest that hypotheses about agency are notoriously difficult to test, and are unlikely to produce black-and-white findings in any event. However, the proper response to these problems is not to abandon the study of agency simply because single case studies fail to produce definitive results. Rather, the real methodological lesson to draw from past studies is that principal-agent analysts should employ carefully chosen, comparative case studies featuring variation across the hypothesised independent variables; and that these cases should be disaggregated in ways that allow us to both multiply the ‘observable implications of theory’ and trace the hypothesised causal mechanisms at work.38 In addition, national-level studies of agency in parliamentary systems could easily replicate research designs such as Wood’s, examining the ways in which partisan change in parliamentary systems influences—or fails to influence—the behaviour of ‘independent’ regulatory agencies in Europe. Second, and finally, Europeanists may also benefit from the recent turn among Americanists to the study of delegation decisions as institutional choices by legislative principals, which may differ over time and across issue areas as a function of other variables like uncertainty, the need for credible commitment, and conflict among principals and agents. Here again, one might argue that students of comparative, nationallevel European politics actually have an advantage in the study of delegation decisions, since the countries of Western Europe have in recent years undertaken a wave of delegation to regulatory agencies and courts, which can be studied comparatively both within and across nations using methods similar to those of Epstein and O’Halloran. At the very least, such comparative studies of delegation would allow us to map the dependent variable—that is, the delegation of powers and discretion to non-majoritarian institutions—across a large-n of both issue areas and countries. More ambitiously, recent studies by Huber and his colleagues have already demonstrated the possibility of measuring and testing hypotheses about delegation decisions across national borders. McNamara’s cross-national study of delegation to central banks in this issue, while putting forward a sociological explanation inherently hostile to Huber’s functionalist approach, also suggests the promise of comparative analysis of delegation. These studies, moreover, have only scratched the surface of the recent wave of delegation in competition policy, utilities, and telecommunications regulation, each of which offers a promising and empirically important arena for cross-national measurement and explanation of legislative decisions to delegate—or not to delegate—discretionary powers to regulatory agencies. In sum, the studies reviewed in this chapter suggest that Europeanists have begun to close not only the theoretical but also the methodological gap separating us from the Americanists—and that comparative European politics may offer a testing ground for the study of delegation as least as fertile as that of the United States.

The politics of delegation

188

NOTES The author would like to thank Alec Stone Sweet, Mark Thatcher, the participants in this volume, and an anonymous reviewer for comments on an earlier draft of this chapter. Responsibility for any errors or omissions is, of course, my own. 1. Barry R.Weingast and Mark Moran, ‘Bureaucratic Discretion or Congressional Control? Regulatory Policy-making by the Federal Trade Commission’, Journal of Political Economy 91/5 (1983), pp. 765–800. 2. John A.Ferejohn and Charles R.Shipan, ‘Congressional Influence on Administrative Agencies: A Case Study of Telecommunications Policy’, in Larry Dodd and Bruce Oppenheimer (eds.), Congress Reconsidered (Washington, DC: Congressional Quarterly Press, 4th edn. 1989), pp. 393–410. 3. Randall Calvert, Mark Moran and Barry Weingast, ‘Congressional Influence Over Policymaking: The Case of the FTC’, in Mathew McCubbins and Terry Sullivan (eds.), Theories of Congress: The New Institutionalism (Cambridge: Cambridge University Press 1987). For an earlier and highly influential version of the same argument, see Weingast and Moran, ‘Bureaucratic Discretion or Congressional Control?’. 4. ADA scores, tallied for decades by Americans for Democratic Action, are often used, alongside other indicators such as party affiliation, as a measure of congressional preferences along the left-right scale. For recent ADA scores, see the ADA website at: http://www.adaction.org/. 5. Terry M.Moe, ‘An Assessment of the Positive Theory of “Congressional Dominance”’, Legislative Studies Quarterly 12/4 (Nov. 1987), pp. 475–520. 6. B.Dan Wood, ‘Principals, Bureaucrats, and Responsiveness in Clean Air Enforcement’, American Political Science Review 82/1 (1988), pp. 213–34. 7. John D.Huber and Charles R.Shipan, ‘The Costs of Control: Legislators, Agencies, and Transaction Costs’, Legislative Studies Quarterly 25/1 (2000), pp. 25–52; quote from p. 35. 8. Ibid., p. 41. 9. David Epstein and Sharyn O’Halloran, Delegating Powers: A Transaction Cost Politics Approach to Policy Making under Separate Powers (New York: Cambridge University Press 1999). 10. Keith Krehbiel, Information and Legislative Organization (Ann Arbor: University of Michigan Press 1991); Epstein and O’Halloran, Delegating Powers. 11. Kathleen Bawn, ‘Political Control versus Expertise: Congressional Choice About Administrative Procedures’, American Political Science Review 89/1 (1995), pp. 62–73; Kathleen Bawn, ‘Choosing Strategies to Control the Bureaucracy: Statutory Constraints, Oversight, and the Committee System’, Journal of Law, Economics and Organization 13/1 (1997), pp. 101–26; David Epstein and Sharyn O’Halloran, ‘Administrative Procedures, Information, and Agency Discretion’, American Journal of Political Science 38/3 (1994), pp. 697–722; Epstein and O’Halloran, Delegating Powers; John. D.Huber, ‘How Does Cabinet Instability Affect Political Performance: Credible Commitment, Information, and Health Care Cost Containment in Parliamentary Politics’, American Political Science Review 92/3 (1998), pp. 577–92. 12. See, e.g., Konstantine Gatsios and Paul Seabright, ‘Regulation in the European Community’, Oxford Review of Economic Policy 5/1 (1989), pp. 37–60; Giandomenico Majone, Regulating Europe (New York: Routledge 1996); Giandomenico Majone, ‘The Credibility Crisis of Community Regulation’, Journal of Common Market Studies 38/2 (June 2000), pp. 273–302; Epstein and O’Halloran, Delegating Powers, pp. 1–4. 13. On credible commitments in the EU, see Majone, Regulating Europe; Majone, ‘Credibility Crisis’; and Andrew Moravcsik, The Choice for Europe: Social Purpose and State Power from Messina to Maastricht (Ithaca: Cornell University Press 1998).

Theory and method in the study of delegation

189

14. See for example Mathew McCubbins and Talbot Page, ‘A Theory of Congressional Delegation’, in Mathew McCubbins and Terry Sullivan (eds.), Congress: Structure and Policy (New York: Cambridge University Press 1987), pp. 409–25. 15. Epstein and O’Halloran, Delegating Powers. 16. John D.Huber, ‘Delegation to Civil Servants in Parliamentary Democracies’, European Journal of Political Research 37/3 (2000), pp. 397–413. 17. See, e.g., Epstein and O’Halloran, ‘Administrative Procedures’; Epstein and O’Halloran, Delegating Powers; Bawn, ‘Political Control versus Expertise’; Bawn, ‘Choosing Strategies to Control the Bureaucracy’. 18. See David Mayhew, Divided We Govern: Party Control, Lawmaking, and Investigations, 1946–1990 (New Haven, CT: Yale University Press 1991). 19. Epstein and O’Halloran, Delegating Powers. 20. I am grateful to Alec Stone Sweet for discussions on this point; see also Epstein and O’Halloran, Delegating Powers, pp. 240–44. 21. Epstein and O’Halloran, Delegating Powers, pp.240–44. 22. Mark A. ‘Delegation, Agency and Agenda Setting in the European Community’, International Organization 51/1 (Winter 1997), pp. 99–134. For recent summaries of the principal-agent approach, see Simon Hix The Political System Union (New York: St. Martin’s Press 1999), ch.2; and Jonas Tallberg’s excellent review in this volume. 23. See, e.g., the essays in Neill Nugent (ed.), At the Heart Union: Studies of the European Commission (Basingstoke: Macmillan Press 1997). 24. See Susanne K.Schmidt, ‘Sterile Debates and Dubious Generalizations: European Integration Theory Tested by Telecommunications and Electricity’, Journal of Public Policy 16/3 (1997), pp. 233–71; Mark A.Pollack, ‘The Engines of Integration? Supranational Autonomy and Influence in the European Union’, in Wayne Sandholtz and Alec Stone Sweet (eds.) European Integration and Supranational Governance (New York: Oxford University Press 1998), pp. 217–49. The study of the European Court of Justice is beyond the scope of this chapter, but see Alec Stone Sweet’s contribution to this volume, as well as the following examples of methodological ‘best practice’ in qualitative and quantitative analysis of ECJ case law: Alec Stone Sweet and T.L.Brunell, ‘Constructing a Supranational Constitution: Dispute Resolution and Governance in the European Community’, American Political Science Review 92/1 (1998), pp. 63–81; Geoffrey R.Garrett, R.Daniel Kelemen and Heiner Schulz, ‘The European Court of Justice, National Governments, and Legal Integration in the European Union’, International Organization 52/1 (1998), pp. 149–76; Bernadette Kilroy, ‘Member State Control or Judicial Independence: The Integrative Role of the Court of Justice’, paper delivered at the 1995 Annual Meeting of the American Political Science Association, Chicago, 31 Aug.–3 Sept. 1995.

The politics of delegation

190

25. See, e.g., Christian Joerges and Jürgen Neyer, ‘Transforming Strategic Interaction into Deliberative Problem-Solving: European Comitology in the Foodstuffs Sector’, Journal of European Public Policy 4/4 (1997), pp. 609–25. 26. See, e.g., Bernard Steunenberg, Christian Koboldt and Dieter Schmidtchen, ‘Policymaking, Comitology, and the Balance of Power in the European Union’, International Review of Law and Economic 16/3 (1996), pp. 329–44; Bernard Steunenberg, Christian Koboldt and Dieter Schmidtchen, ‘Beyond Comitology: A Comparative Analysis of Implementation Procedures with Parliamentary Involvement’, Aussenwirtschaft 52/1 (1997), pp. 87–112; and Pollack, ‘Delegation, Agency and Agenda Setting’. 27. Rhys Dogan, ‘Comitology: Little Procedures with Big Implications’, West European Politics 20/3 (1997), pp. 31–60; Fabio Franchino, ‘Control of the Commission’s Executive Functions: Uncertainty, Conflict and Decision Rules’, European Union Politics 1/1 (2000), pp. 63–92. 28. John D.Huber, Charles R. Shipan and Madalaine Pfahler, ‘Legislatures and Statutory Control of Bureaucracy’, American Journal of Political Science 45/2 (April 2001), pp. 330– 45. 29. John D.Huber and Arthur Lupia, ‘Cabinet Instability and Delegation in Parliamentary Democracies’, American Journal of Political Science 45/1 (Jan. 2001), pp. 18–33. 30. Huber, ‘How Does Cabinet Instability Affect Political Performance’. 31. Huber, ‘Delegation to Civil Servants’, quote from p. 410. 32. Kaare Strøm, ‘Delegation and Accountability in Parliamentary Democracies’, European Journal of Political Research 37/3 (May 2000), pp. 261–89; quote from p. 267. 33. Torbjörn Bergman, Wolfgang C.Müller and Kaare Strøm (eds.), ‘Parliamentary Democracy and the Chain of Delegation’, special issue of the European Journal of Political Research 37/3 (May 2000), pp. 255–60; quote from p. 257. 34. Bergman et al., ‘Parliamentary Democracy and the Chain of Delegation’, pp. 255–429. 35. See R.A.W.Rhodes, Understanding Governance: Policy Networks, Governance, Reflexivity and Accountability (Philadelphia: Open University Press 1997). 36. See the chapters by Thatcher and by Stone Sweet in this volume. 37. See Pollack, ‘Delegation, Agency and Agenda Setting’; Torbjörn Bergman, ‘The European Union as the Next Step of Delegation and Accountability’, European Journal of Political Research 37/3 (May 2000), pp. 415–29; and Jonas Tallberg, this volume. 38. For a good discussion, see Gary King, Robert O.Keohane and Sidney Verba, Designing Social Inquiry: Scientific Inquiry in Qualitative Research (Princeton: Princeton University Press 1994).

Abstracts

Theory and Practice of Delegation to Non-Majoritarian Institutions, by Mark Thatcher and Alec Stone Sweet The chapter summarises the analytical frameworks, questions, and empirical findings of the volume. It defines the key concepts used. It then sets out the principal-agent framework that explains delegation to NMIs through functional logics for principals. It sets out alternative explanations based on sociological and historical institutionalism. Thereafter, it relates the empirical findings of the volume to these wider debates about delegation. It argues that although functional demands for delegation can almost always be identified, purely functional accounts of delegation to NMIs are inadequate. Explaining the decision to delegate and the institutional forms of that delegation involves including and specifying interests, policy learning/ institutional isomorphism and institutional inheritances. Delegation has also had major consequences on the distribution of power, policy making processes and substantive policy choices, both through its direct effects, and via feedback effects. Finally, delegation has raised questions about the legitimacy and accountability of NMIs. Delegation to Supranational Institutions: Why, How, and With What Consequences? by Jonas Tallberg Why, how, and with what consequences do national governments delegate political authority to supranational institutions? Contrary to the static conceptions of delegation that dominate the existing literature, this article adopts a dynamic approach, where the stages of the delegation process are integrated into a coherent rational institutionalist framework. With demonstrations from the case of the European Union, the article argues that: (1) the expected consequences of delegation motivate governments to confer certain functions to supranational institutions; (2) the nature of these functions influences the design of mechanisms for controlling the institutions; (3) institutional design shapes the consequences of delegation by facilitating or obstructing attempts by the institutions to implement private agendas; and (4) the consequences of previous rounds of delegation affect future delegation, institutional design, and interaction, through positive and negative feed-back loops. Rational Fictions: Central Bank Independence and the Social Logic of Delegation, by Kathleen R.McNamara The conventional wisdom in support of central bank independence rests on a series of contestable arguments about the relationship between democracy, policy making, and

Abstracts

192

economic outcomes. Empirical work casts doubt on the severity of the inflationary problems purportedly solved by delegation to independent central banks, as well as raising questions about the linkages between delegation and superior economic outcomes. So why delegate? Theories of institutional isomorphism, or the copying of organisational models, provide an alternative sociological explanation of the diffusion of central bank independence. Drawing on this approach, it is argued that governments choose central bank independence because delegation has important legitimising and symbolic properties that are attractive to political leaders in times of economic uncertainty. Delegation to independent central banks is rational, efficient, and acceptable in a democratic society because of the cultural processes which define it as such, not because of the fiunctional requirements of economic management. Constitutional Courts and Parliamentary Democracy, by Alec Stone Sweet The chapter assesses the creation and subsequent evolution of systems of constitutional justice in West Europe, in light of delegation theory. The author argues that constitutional judges are better conceptualised as trustees, exercising fiduciary responsibilities, than as agents, who operate in the shadow of principals. The zone of discretion that organises the activities of constitutional courts is unusually large, in some contexts close to unlimited. The author then surveys why, and to what extent, constitutional adjudication has transformed the nature of parliamentary governance, focusing on the cases of France, Germany, Italy, and Spain. Notwithstanding important variation, certain trends are both pan-European and irreversible: traditional separation of powers doctrines are steadily eroding; legislators and administrators are being placed under the authority of an expansive, continuously evolving constitutional law; and the judiciary’s participation in law making processes is becoming more overt and assertive. Institutional Choice and Bureaucratic Autonomy in Germany, by Marian Döhler Although the delegation of government functions to non-majoritarian bodies such as independent agencies has accelerated throughout the OECD, Germany has followed a different path so far. In particular, administrative agencies have only rarely been granted autonomy from direct political control. The main argument is that a genetic code, inherent to the system of government institutions, has routinely auto-piloted choices that involved the design and control of agencies away from principal-agent-like considerations. The elements of this genetic code are outlined and the likelihood of their future persistence is assessed. Even if the pressure to create new agencies at arms length from government has increased, there are still strong incentives to keep them in a subordinate position. Delegation to Independent Regulatory Agencies: Pressures, Functions and Contextual Mediation, by Mark Thatcher Independent Regulatory Agencies (IRAs) have spread across many domains in Western Europe. The chapter examines selected examples in Britain, France, Germany, and Italy. A functionalist analysis of the pressures on elected officials and the functions that IRAs perform provide a valuable starting point for analysis. Nevertheless, it confronts crossnational and cross-domain variations in the timing of the creation of IRAs, their spread and their institutional forms. In order to offer a fuller account, contextual factors that

Abstracts

193

mediate pressures must be considered. These factors include learning and institutional isomorphism; state traditions and structures; political leadership; state reforms. Finally, IRAs have had far-reaching consequences that have often been unanticipated at the time of their creation. The Unanticipated Consequences of Creating Independent Competition Agencies, by Stephen Wilks with Ian Bartle The tendency for governments to create independent competition agencies is analysed in the cases of Directorate General IV of the Commission, the German Cartel Office, the Office of Fair Trading, and the Competition Commission. Analysis of the historical process of agency design, and the redefinition of agency missions, indicates a progression from a symbolic and constitutional rationale to a more material impact on contemporary market economies. Drawing loosely on principal-agent theory, changing agency roles are ascribed partly to the activism of independent agents, partly to the changing priorities of majoritarian principals. The unanticipated consequences of delegation include an escape from business capture but a shift to legalism or economic purism. Judicial Delegation Doctrines: The US, Britain, and France, by Martin Shapiro Legislatures frequently enact primary legislation that delegates secondary law making powers to administrative agencies. Judicial review designed to ensure that this secondary legislation is in accord with the primary legislation necessarily involves judicial interpretation of the primary legislation and hence also some degree of judicial law making. Both the relative degree of judicial law making and its causes may vary from country to country. Judicial review of secondary legislation in three countries is examined. Judicial activism is great in the United States and probably related both to congressional inefficiency in passing amending statutes and judicial recruitment and career patterns. The level of judicial activism in the UK has been low but may be increasing. It is severely constrained by the capacity of parliament rapidly to ‘correct’ judicial interpretations but encouraged by judicial career patterns. The formal decisions of the French Council of State show little judicial intervention against administrators’ secondary legislation but such intervention may occur extensively at the stage of agency regulation drafting rather than through formal review processes. Learning from the Americanists (Again): Theory and Method in the Study of Delegation, by Mark A.Pollack European(ist) scholars have largely followed their American(ist) colleagues in the formulation of theories about delegation of powers to non-majoritarian institutions, most notably through the application of principal-agent models of relations between legislative principals and their executive and judicial agents. This chapter suggests that Europeanists can once again learn from recent developments in both theory and method in the study of delegation in American politics. The first section discusses the methodological challenges of testing hypotheses about the conditions under which agents might enjoy some degree of autonomy from their legislative principals, and draws lessons from the recent Americanist literature. The section examines the development in American politics of a second wave of principal-agent analysis which aims to formulate and test hypotheses about the conditions under which legislative principals might delegate authority and

Abstracts

194

discretion to bureaucratic agents. The third and final section examines some preliminary applications of the principal-agent approach to the European Union and to the comparative study of European parliamentary democracies, and proposes a research agenda for the comparative study of national-level delegation in the parliamentary systems of Western Europe.

Notes on Contributors

Ian Bartle is Postdoctoral Research Fellow, Department of Politics, University of Exeter. Marian Döhler is Lecturer in the Faculty of Economics and Social Sciences, University of Potsdam. Kathleen R.McNamara is Assistant Professor of Politics and International Affairs, Woodrow Wilson School and Politics Department, Princeton University. Mark Pollack is Assistant Professor, University of Wisconsin-Madison and Visiting Professor (2001–2002), European University Institute. Alec Stone Sweet is Official Fellow, Nuffield College, Oxford and Chair of Comparative Government. Martin Shapiro is James W. and Isabel Coffroth Professor of Law, University of California, Berkeley. Jonas Tallberg is Research Fellow, Department of Political Science, Lund University. Mark Thatcher is Senior Lecturer in Public Administration and Public Policy, Department of Government, London School of Economics. Stephen Wilks is Professor of Politics and Deputy Vice Chancellor, University of Exeter. He is also a member of the Competition Commission, but writes here in a purely private capacity.

Index

abstract review 86–7, 90, 95 ADA 203, 206, 211, 217n.4 administration, law-enforcement and 113 administrative agencies, ‘clash of interests’ and 112 administrative courts, influence on agency operations 113 administrative law, reflective in character 179 advisory committees, European Committee and 30 AGCOM (autorità per le Garanzie nelle Communicazioni) 135, 138 Agence de l’environnement et de la maîtrise de l’énergie (France) 127 agencification 101–2, 118 agency activism, division of labour between politics and administration 120 agency behaviour 118, 202, 204–5, 216 agency costs 25–6, 32–3, 209 agency design and control, bureaucracy and 113 agency discretion and control 5–7 agency innovation, life cycle of 156 agency losses 4, 16, 130, 139, 214–15 agency shirking 28, 55, 139 agency slippage 55, 130, 139 agency vs political control, problem of measurement 201–6 agents 4, 12, 28–9, 32–3, 81, 91–2 Albania 47 Amato, Giuliano (Italian Competition Authority) 149 American bill of rights 83 American courts 184, 195–6 American Courts of Appeal, policy preferences and 183 American delegation theory, supranational politics in EU 1 American judicial review 15, 78–9, 85 American politics literature, American regulatory agencies and 201–2 Americanist theories of delegation, separation of powers system and 211 Americans for Democratic Action see ADA Amsterdam Treaty (1997) 39, 164 antitrust, symbolic politics 155 APA 181–3, 189 Arnold, Thurman 155 atomic energy conflict, legislation strategy and 113 Austria 79 Austrian constitutional court, democratisation 12

Index

197

Austrian Second Republic (1920–34) 79 ‘autolimitation’, definition 94 Bank of England, independent (1997) 13 Bank for International Settlements 61 Bank of Japan, independent (1996) 65 Bartle, Ian 9–12, 16–18, 175 Bawn, Kathleen 209 Belgium 66 Bergman, Torbjörn 214–15 BKA 108, 149, 151, 162–3 apogee of civil service career 164 appeals against in civil courts 167 Berlin Court of Appeal and 169 German Mittelstand 163 independence because of Erhard 166 blame shifting 27, 131, 157, 169 ‘bootstrapping’ 180, 183 Borrie, Lord Gordon (DGFT 1976–92) 165 Börzel, Tanja A. 118 bounded rationality 37, 103, 154 Britain see UK British ‘treasury control’ 176 Brittan, Leon 168 Bruno, M 57 Bundesaufsichtsamt für den Wertpapierhandel see BAW Bundesbank 10, 17, 56, 108 Bundeskartellamt see BKA Bundesrat 91, 135 Bundestag 91, 135 Caldwell, Michael 103, 111 Calvert, Randall 203–4, 206 Calvert-Moran-Weingast design, FTC behaviour and 204 Caporaso, James 40 Carter presidency, activism of EPA during 204 Caspari, Manfred (DGIV 1980–90) 165 CBI, DGFT/IFT formula and 161 central bank independence 1–2, 9–12, 208, 217 APPENDIX 1 71–6 behavioural not legal or organisational phenomenon 55 contestable theoretical arguments 66–7 credibility is perceptual variable 61 danger of too much autonomy 55 delegation and outcomes in practice 56–9 discipline of economics and 65 economic outcomes from inconclusive 58–9 evidence for outcomes inconclusive 58 functional necessity and 59 highly political social process 67–8 international organisations and 47

Index

198

Maastricht Treaty and 63–4 macroeconomic indicators and 63 monetarist policies and 17 partisanal policies 48 the politics of 53–6 probing the sources of 62–6 signalling that government is truly ‘modern’ 60 social basis of diffusion 59–62 theory of 49–53 central banks 1–2, 9, 11 acting on neutral economic data 156 analysed by McNamara 149, 217 area of uncertainty 64 delegation attractive because of improving aggregate welfare 53 flow and supply of money through interest rates 51 relationship between independence and economic outcomes 63 Central and Eastern Europe (after 1989) 79, 150, 157 charged administrations (Bundesverrichtungsverwaltung) 104 ‘choice’, preference-based image 103 City of London, regulation and 141 civil service 113, 119, 160–61, 164 CNIL 132, 136 coercive isomorphism 12–13, 64–5 comitology committees, European Commission and 212–13 Commission nationale de l’informatique et des libertés see CNIL Commission for Racial Equality (UK) 131, 135–6 competence (Zuständigkeit) 105 ‘competition’, cartel law and 116–17 competition authorities 9, 152, 157 competition law, business view Janus-faced 156–7 competition policy, delegation 150–52, 157, 164, 169 concrete review 86–7, 90, 95–6 Confederation of British Industry see CBI Congress active judicial review and 186 FCC and political consequences for 203 imperfect information and 207 law delegated to 180 oversight lax and ineffective 201 P-A and 125, 185 prefers to transfer policy making authority 185 procedural rules to favour certain interests 183–4 sanctioning an agency 203 Congressional Quarterly legislative summaries 211 ‘consistency’, commitment problem and 156 constitution and separation of powers 77–9 constitution and normative hierarchies 80–81 contracting and delegation 82 constitutional bargains 82–4 Kelsen’s court 79–80 new constitutionalism 79 old constitutionalism 78–9

Index

199

positive and negative legislator 81–2 constitutional adjudication 92 discretion and autonomy 92–3 exercise of legislative power 90 and the judiciary 95–8 and the legislator 93, 93–5 constitutional contracting 83, 86 Constitutional courts 77, 86, 92 components of 79–80 dilemma of political parties 9 discretionary powers 93 important legislation veoted by 93–4 constitutional judges European serve fixed term 88 ‘indirect’ authority over legislating 94 ministers and parliamentarians govern with 98 negative legislators 81 powers to govern 77 regulate actions of government and parliament 89 scope and limit of any rights provision 84 zone of discretion and 93 constitutional review 77, 97 constitutionalisation 96–8 Continental legal systems, constitutionalisation and 96 contracting and delegation 82 agency and trusteeship 88–90 composition 88 constitutional bargains 82–4 constitutional courts 85–6 constitutional courts, relational contracting and commitment 86 jurisdiction 86–8 mechanisms of control 90–92 ‘correcting’ legislation, judicial policy initiatives and 178 ‘corrective revision’ 94–5 Council of the European Union 42, 178 Council of Ministers 6, 66, 213 Cour des Comptes in France 142 courts 174–6 credible commitment concept 153, 156, 162,169, 208–9 CSSPF (Conseil Supérieur du service public ferroviaire) 127 customary right (gewohnheitsrecht) 107 delegated legislation in England, P-A story 186 delegation benefits of 26 benefits outweigh costs of 4 civil servants and NMIs 214 costs and benefits 25 definition 3, 25, 127 democratic accountability questions 54 efficacy of in management of money 56

Index

200

flawed because of large agency costs 37 from agency to 206–10 functional demands for 9, 143 IRAs and 127 justification in central banking is problem of inflation 51 links between stages of process 24 logic of in monetary policy realm 67 outcomes in practice and 56–9 power of private interests 17 procedural problems 55 process rather than one-off event 18 rational choice in context of neoliberalism 65 rational choice students and comparative politics 211 scholars focusing on 207 substantive choice in central banking 54 supranational institutions and 23 winners and losers from to NMIs 19 delegation theory 77, 81, 92, 98 Department of Trade and Industry (UK) see DTI Deutsche Telekom AG 110, 141 developing countries 58, 63 DGFT 160–61 DGIV 149, 151–2, 164, 168–70 direct controls, formal and negative 90–91 Director General of Fair Trading see DGFT Directorate General for Competition see DGIV directorates-general, European Commission 26–7 discretion 4 agents and credible commitments 33 American agents and 201 delegation and in political settings 213 effects of supranational in EU 35–6 functional problems and 31 independent agencies as unelected bodies 141 literature on transaction costs of policy and 207 number of new words and 213–14 private interests and 34 Dogan, Rhys 213 Döhler, Marian 10–11, 13–14, 17–18, 153 DTI 151, 160–61, 163 Dutch competition law, modelled on DGIV 149–50 dynamics of delegation 36–7, 38–41 Easterly, W. 57 Eastern Bloc, need to appear credible 65 Eberlein, Burkhard 117 EC legislation 13, 133 ECJ 1, 6, 9, 17, 25, 88 compliance of members and 26 constitutionalisation of Maastricht and dynamic effects 40 de facto constitutional court 35

Index

201

delegation to 23 enhances decision making efficiency 27, 31–2 EU and 12 expanded its authority through rulings 16 judicial and political competencies 26 no role in justice and home affairs 38 oversight instruments 29–30 partnership of competitive lawyers and 168–9 transformation of European legal system (1960s) 34 treaty making process 18 upgrading of enforcement powers (1990s) 39 Economic and Monetary Union see EMU Economist, The 61 ECSC Treaty (Treaty of Paris 1951) 165 Ecuador 60 Edelman, M. 155 Ehlermann, Claus-Dieter (DGIV 1990–95) 165 Eisner, M. 157 Electricité de France 133 electricity and gas, Florence and Madrid groups 137 Elster, J. 83 EMU 47, 64, 66 ENEL, partial sale of (Italy) 139 England administrative law jurisprudence 181 judicial delegation 186–92 parliament and ‘subjectivity’ clauses 187 privity clauses exclude judicial review 187 ultra vires review and 186 ‘Wednesbury’ unreasonableness test 189 English courts 186, 188, 191–2, 195–7 English judges 187–90, 192 ENI (Italy) 133 environment, few agencies for 143 Epstein, David 207, 209–11, 213, 217 Erhard, Ludwig 161–3, 166 EU 6, 208 competition policy 149 convergent monetary policies and low inflation 57 degree of control exercised by 24 delegated functions to supranational institutions 41 delegation of supranational powers in 26–8 delegation to 23 delegation to NMIs and central banks 200 demands of and delegation to IRAs 133 European Court of Justice and 12 feed-back loops in 38–9 independence of central banks for EMU 65 independent banks in conjunction with Maastricht Treaty 63–4 legal norms 178 low inflation in (1980s and 1990s) 64 NMIs and 101

Index

202

regulation in electricity 135 research programme on delegation 211 separation of constitutional powers 119 test of P-A hypotheses in 212 treaty making process 18 unanimous approval requirement 38–9 EU law, agencies and 13 Europe competitive policy agencies 148–50 constitutional review in 82 delegating to NMIs 9–11 decision 11–12 institutional forms 14–15 institutional legacies 13 policy learning and institutional isomorphism 12–13 free market and competitive integrated economy 168 functional reasons for delegation to IRAs 131–4 judicial review 178 Kelsenian court and 78 NMIs as ‘fourth branch of government’ 16 party political regulators 139 politics of delegation analyses beyond functional logics 20 symbolic efforts at competitive policy 157 European Agreements, Central and Eastern Europe 150 European Cartel Office 164, 169 European catalogue of rights 35 European Central Bank (ECB) 47, 49, 53, 66–7, 152, 169 European Commission 1–2, 6, 9, 17, 25 Council of Ministers or European Council 14 delegation to 23 empirical studies of independent causal role 212 European economic regulation 164 ‘guardian of the treaties’ 26–7, 31 infringement proceedings and centralised enforcement 40 internal market initiative (1980s) 34 less restrictive measures and 213 Maastricht Treaty and 38 MNI 151 national bureaucracies and 19 national regulators and 137 observation status in ECB meetings 66 oversight instruments 29–30 provides policy expertise 32 re-legislation by qualified majority 30 treaty making process and 18 trustee when it enforces competition rules 7 unanticipated consequences of delegation of competition policy 168, 170 upgrading of enforcement powers (1990s) 39 European constitutional court 35 European Convention on Human Rights 13, 178, 189–90, 192 European Court of Justice see ECJ European Free Trade Association (EFTA) 42

Index

203

European integration 17, 33–4, 117 European Journal of Political Research 214 European and Monetary Union see EMU European Parliament (EP), European Commission and 6, 213 European politics, theory, method and delegation in 210–17 Europeanist scholars, Americanist scholars, delegation of powers to NMIs 200 Europeanists, institutional choices by legislative principals and 216 Fair Trading Act (1973) see FTA FCC 13, 202–3, 206 FCO 104, 110, 134, 137 BKA and 162 independence and 10, 16–18, 107–8, 127, 140 federal agencies (Bundesoberbehörden) 104 Federal Communications, Oftel and 137 Federal Communications Commission see FCC Federal Gazette 163 feed-back loop, number of forms 40–41 Ferejohn, John A. 202–3, 206 Financial Times 61 food safety and environmental protection, scientific expertise 132 Food Standards Agency in UK, BSE scandal and 132, 138 former Soviet bloc, independent central banks 63 France 15, 66, 134 abstract review and 87 administrative courts separate from regular courts 177 administrative law 181 appointment constitutional judges 88 audiovisual agency (1982) 133 Autorités administratives indepéndentes (AAIs) 126 CNIL 132, 136 Conseil de la Concurrence (1986) 149 Conseil des marches financiers 132 constitution revision is most permissive 91 constitutional rights 83–4 creating administrative law jurisprudence 179 Declaration of the Rights of Man (1789) 84 delegation of law and 192–5 discretion 116 Fifth Republic and entourage of General de Gaulle 85 form of constitutionalisation 98 IRAs 126, 128–30, 134, 137–8, 143–4 judicial delegation 192–5 liberalisation of electricity 139 nationalisation and affirmative action 93 referrals in 97 social and environmental policy 133 Socialists in 51 telecommuications regulator (ART) and UMTS services 140–41 utility agencies 133 utility regulators 135, 140

Index

204

France Télécom 133, 139 Franchino, Fabio 213 French constitution, two primary law makers 192 French Council of State 192–7 French executive, enacts décrets 192 French legal system, P-A similes and 194–5 French legislator, lois and 192 functional analysis 8,10 Gaitskill, Hugh 158–9 GATT/WTO dispute-settlement mechanism 42 Gaullists, ‘mixed presidential parliamentary’ system 85 GDP growth rates, health of economy and 51 gender, few IRAs 143 Gerber, D. 162, 165 German Constitutional Court 10 German Federal Cartel Office see FCO German Länder 13, 105–6, 127 Germany 15, 79 639 different NMPs (Behörden) 104 abstract review 87, 97 agencification in 101–2, 118 appointment of constitutional judges 88 associational agreements in 106 ‘auto-piloted’ design and control agencies 119 Bundesbank’s success and 55–6, 58–9 bureaucracy and 154 competition authorities 157, 162, 165, 170 concrete review 87, 97 constitution revision by Bundestag and Bundesrat 91 continuity through dispersed government powers 116 delegated relatively little 10 democratisation 79 Democrats in 51 discretion in 116–17 drug regulation 109 economic reconstruction (1950s) 107 Federal Department of Economics 106–7, 110 Federal Environment Office 104 government based on ‘sequential’ order of institutions 111–12 individual complaints (verfassungsbeschwerde) 87, 97 IRAs 126, 128–30, 134–5, 138, 143 legal obstacles to NMIs 13 legislation after survey of choices 155 liberalisation of abortion (1975, 1992) 93 ministerial responsibility and 112 NMIs including the Constitutional Court 108 non-judicial constitutional review 79 NPM and 118 observations on an ‘agency laggard’ 103–6 parliamentary government in 83

Index

205

reduction of transaction costs and 111 referrals 87 regulation of electricity 137 review jurisdiction 90 rights in 84 scope of constitutional rights 83 sectoralisation of political economy 118 several types of agency 126–7 social and environmental policy 133 symbolic content of legislation 168 telecommunications 129 Telecommunications Act (TKG) 110 understanding of delegation 116 utility regulators 135 globalisation of markets 149 Gorsuch, Anne (EPA administrator) 204 governance 2, 23, 92 governments, political risk of delegation and 42 Grande, Edgar 117 Greece 66 Group of 7 (G7) 61 Guenther, Eberhard 163, 165 GWB 161–3 Heath government, Fair Trading Act (1973) 159 Heath, Ted 168 High Level Telecommunications RegulatorsGroup 137 historical institutionalists 8,11 Hood, C. 153, 157 Horn, M.J. 152, 154, 159 Huber, John D. 207, 209, 213–14, 217 human rights 35, 79, 84, 195, 198 IMF 41, 47, 61, 64–5 immigration 18, 191 imperfect information and 207–8 independent competition agencies 9, 148–52 consequences of delegation 165–9 internal and external 170 definition of independence 151 European enactment 164–5 explaining decision to delegate 152–8 four ‘functional logics of delegation’ 156 German enactment 161–4 limits of delegation to 169 UK enactment 158–61 unanticipated consequences of delegation 170–71 independent regulatory agencies see IRAs Independent Television Commission 138 indirect controls, informal and indirect 91–2 individual complaints 87, 96–7

Index

206

inflation 51, 54, 56–7 insider trading, BAW and 109 institutional choice, categories of 159 institutional design 4–5, 153, 212 institutional design of delegation 28 functional basis of EU control mechanism 29–33 principal’s control problem 28–9 institutional genetics, federal administrative agencies in Germany 111, 119 institutional isomorphism 62, 64, 65, 67 institutional order of government 111–14 intergoventalism, European integration and 33 intergovernmental bodies, IMF and EU 64 intergovernmental conference, ECJ’s treaty interpretations and 40 International Atomic Energy Agency (IAEA) 41 International Criminal Court 42 international governance structures, problems at national level and 23 International Monetary Fund see IMF international organisations, greater regulation by and IRAs 133 IRAs 9, 125 ‘answerability’ and 142 consequences of delegation 139 accountability and legitimacy 141–2 further institutional reforms 141 policy making 140–41 relationship between IRAs and elected officials after delegation 139–40 contextual factors mediating pressures for delegation 136 learning and institutional isomorphism 136–7 political leadership 138 state reforms 138–9 state traditions and structures 137–8 credible commitments for governments 132 defining 126–7 influence on policy making 140–41 limits of simple functionalist explanation 134–6 number and selection of regulators 135 useful for elected officials 130–31, 143 isomorphism 42, 108, 125 Italy 66 abstract review 87, 97 AGCOM 135, 138–9 agencies in 13, 15 Cassazione (high civil court) 97–8 Competition Authority (1990) 135, 137, 140, 149 concrete review 87 constitutional revision by majority vote 91 constitutional rights 83–4 ‘independent administrative authorities 126 IRAs 126, 128–30, 134, 136, 138, 143–4 nomination of constitutional judges 88 parliamentary government in 83 referrals 87 review jurisdiction 90 telecommunications regulator 137, 141

Index

207

utility agencies 133 utility regulators 135 Japan 157 judges 78, 89, 96–7, 174–6, 197 judicial activism 178, 185 judicial authority, separation of powers doctrines 78 judicial delegation 173–9 comparing US, England and France 195–8 judicial review 173–9, 195 England and 186–8, 190, 195 United States and 181–6, 195 judicial transformation of APA 182–5 justice and home affairs 38–9 Karrte, Wolfgant (president of BKA 1976–92) 165 Kazakhstan 47 Kelsen, Hans 79–82, 90, 93, 98 Kelsenian court 15, 78, 84, 95, 98 Korea 157 Large-n studies 206, 217 law, delegation and 173 Law Against Restraints on Competition (1957) see GWB law maker, interpreter of law 175 lawfulness 182, 188, 196 lawsuits, fire alarms 176 leftist parties, appeal to workers 52 legal aspects (Rechtsaufsicht), oversight and 104 legislators and judiciaries, impact of constitutional decision making 92 legislatures, delegate to agencies because of unresolved policy conflicts 175 liberal democratic theory, public authority and 2 liberalisation 133, 143–4 liberalisation of markets, technical requirements of regulation 131 litigation, policing away from legislature to litigating parties 175 long-term portfolio volatility, cost-control and 214 Lupia, Arthur 214 Maastricht Treaty (1993) 31, 38–9, 64, 66 Macey, J. 153 McNamara, Kathleen R. 9, 11–12, 78, 149, 217 Majone, Giandomenico 1, 7, 14, 77, 88, 101, 117 ‘democratic deficit’ in Europe 151 European Cartel Office and 169 management committees, European Commission and 30 market integration 149 Maxfield, S. 63 Mayhew, David 210 Methven, John (first DGFT) 161 Meyer, J. 59–60

Index

208

Miert, Karel Van 164 ministerialfreie Räume (areas free from ministerial oversight) 104 MIT 61 Mitnik, B.M. 155 MMC 141, 158, 161 modified parliamentary sovereignty 189–90 Moe, Terry M. 103, 111,204 monetary policy 56–7 monitoring, agent behaviour and 28 Monopolies and Mergers Commission see MMC Monti, Mario 164 Moran, M. 154, 202–4, 206 Moravcsik, Andrew 41 Morris, Derek (Chairman of Competition Commission) 167 Müller, Wolfgang 214–15 multi-party coalition parliaments, legislative functions 177–8 national courts, symbiotic relationship with ECJ 35 NATO 42 Nebenregierung (supplementary government) 108 negative externalities 9 negative legislator 81–2, 90, 93 neofunctionalism, European integration and 33 new constitutionalism, replaced legislative sovereignty 79, 88, 95 new constitutions 13, 79 new political management (NPM) 118–19 new procedural rules, greater transparency and 183 New Zealand 47 NMIs central bank independence and 49 competitive policy agencies 148 concepts and theory 2–3 agency discretion and control 5–9 alternatives to P-A logics 7–9 functional logics of delegation to 3–5 P-A approach 3 decision to delegate to 11 institutional legacies 13 Interests 11–12 policy learning and institutional isomorphism 12–13 definition 127 delegating to in Europe 9–11 delegation of powers and discretion to 217 expansion in OECD nations 101 institutional forms of delegation 14–15 international governance and 23 IRAs example of delegation to 125 more focused mandates than elected politicians 19 national-level factors and 13 ‘output legitimacy’ is justification for 18 part of ‘good governance’ 12

Index

209

policy continuity and 116 political dynamics and consequences 15–16 direct effects 16–17 feedback effects 17–18 legitimacy and accountability 18–19 reconfigured EU 19 separation of powers and delegation to 98 substantive policy decisions 17 Western Europe and 1–2, 9–10 non-governmental organisations (NGOs) 132 non-majoritarian institutions see NMIs norm-free bargaining, result in a settlement 83 normative isomorphism, coercive isomorphism 65 observational equivalence 202–3, 212 Observatoire de la parité, gender equality and 127 OECD 47, 61, 101–2 Offer (electricity regulator) 136 Office of Fair Trading (UK) see OFT Ofgas (gas regulator) 136 OFT 151, 158–61, 163, 167 Oftel (telecommunications) 136–7, 140–41 Ofwat (water regulator) in UK 133, 136 O’Halloran, Sharyn 207, 209–11, 213, 217 old constitutionalism 78–9 ombudsman (Defender of the People in Spain) 87 ‘open method of coordination’, employment, social affairs and education 38–9 ordnungspolitik 162 ordo-liberal debate in Germany 162, 164 Organisation for Economic Cooperation and Development see OECD Organisation for European Economic Co-operation (OEEC) 42 organisations, definition 59–60 P-A analysis 3, 24, 41 acts of delegation are problematic 28 agents and 16 agents exploiting their discretion 32, 170 alternatives to logics of 7–9 common reasons why principals choose to delegate 50 cross-national comparison 10 ‘deck-stacking’ elements 107 EU and 201, 212 European integration and 34 explanation of delegation and 42 focus on institutional choice through design 8 point of contention via agency costs 33 political control 102 regulatory bureaucracies and 206 relatively straightforward 202 trusteeship and 77 use in regulation and delegation 155

Index

210

P-A framework, delegation to IRAs 3, 125, 142 P-A hypotheses, principals’ preferences and agency behaviour 204 P-A literature 50, 54–5, 125 P-A models 25 delegation from one authority to another 81 delegation to IRAs 129–31 EPA and 204–5 functions performed by agencies 131, 197 imperfect information in 208 slippage and shirking 139 P-A problems, acts of delegation and 173, 178, 185, 214 P-A relationships 37, 89, 119, 148, 187 Paris Stock Exchange 132 parliamentary democracy 2, 78–9, 178, 215–16 parliamentary model, delegation and discretion in 213 parliaments, positive legislators 81 partisan politics, problem with political control 51 path dependence 154 pension mis-selling in UK 138 ‘people of the United States’ 178 Pfahler, Madalaine 213 Pierson, Paul 103 political business cycle 51, 56 politics, public governance 2 Pollack, Mark 1, 3, 11, 16, 26 Portugal 66, 79 Posen, Adam 58 positive politics literature, rational choice theory and 153 post-modernist, indeterminacy of all text 175 post-World War II European constitutions 82 power (Organisationsgewalt) to reorganise 105 pre-EMU Dutch central bank 53 Principal-Agent see P-A principals 3, 4, 6, 81 principals and agents, boundedly rational 37 principals of IRAs, elected officials 129 privatisation 108, 132–3, 143 Progressives in United States, theories of Woodrow Wilson 152 pure parliamentary sovereignty 189 qualified majority 30, 88 R. v. Secretary of State for Social Security, ex-parte Joint Council for the Welfare of Immigrants 191 racial equality 143, 152 rational choice approaches to politics 7, 159 rational choice institutionalists, P-A and cost analysis 24 rationalisation, modern sociocultural environments and 61–2 rationality, organisational forms 59 Reagan, President Ronald, behaviour of EPA 204 reasonableness 188–9, 191–2, 196–8

Index

211

Rechtsformenwahl (choice of legal status) 108 Rechtsstaat, German tradition delegation defined by 153 referrals 87 RegTP 104, 109–11, 127, 141 regulators’ powers, circumscribed 14 Regulatory Agency for Telecommunications and Postal Services see RegTP regulatory committees, European Commission and qualified majority 30 regulatory state, rule making and generating competition 117 Ressortprinzip (departmental principle) 105 reviewing courts, rival interpretations of statute 175 Risse, Thomas 118 Rome Treaty 14, 88, 141, 164–5 Ross, Stephen 25 Ruckleshaus, William (replaced Gorsuch at EPA) 205 Scharpf, Fritz 38 Schmidt, Susanne 212 Scott, R.W. 59–60 Self-governing bodies (öffentlich-rechtliche Körperschaften and Anstalten) 104 separation of powers 77–9, 98, 103, 111, 119, 211, 213 Shapiro, Martin 12–13, 17 Shipan, Charles R. 202–3, 206–7, 213 shirking 28, 32–4, 36, 38–9, 102, 119 short-term portfolio instability, rising cost of healthcare 214 sociological institutional arguments, organisational and 59 sociological institutionalism 8, 11, 42–3 sociological perspective 60–61 Southern and Eastern Europe, choice of legislation, foreign models 155 Southern Europe 79 Spain 66, 79 abstract review 87, 97 appointment of constitutional judges 88 concrete review 87, 97 constitutional revision by three-fifths of Cortes and Senate 91 constitutional rights in 83–4 individual complaint (amparo) 87, 97 Kelsenian court and 85 liberalisation of abortion (1985) 93 parliamentary government in 83 referrals 87 review jurisdiction 90 special instructions (Einzelweisungen) 107 statues, broad or vague language 175 Stigler, G. 157 Stiglitz, J.E. 53, 57 Stone Sweet, Alec 7, 9, 11–12, 14, 16–17 delegation and 40, 50, 155–6, 173–5, 178 P-A analysis 201–2 Strom, Kaare 214–15 Sturm, R. 166 supranational institutions

Index

212

active choice between alternative governance 25 blame shifting and 27–8 consequences of delegation 32 anticipated benefits and unwanted shirking 32–3 supranational discretion and effects of integration 33–6 control mechanisms and 30 delegation to 23 discretion enjoyed by 24 dynamics of delegation 36–41 European constitution and 35 institutional design of delegation 28 functional basis of EU control mechanisms 29–32 principal’s control problem 28–9 rationale of delegation 25 functional basis 25–6 supranational shirking 36, 38–9 Sutherland, Peter 168 Sweden 47, 150 Switzerland 79 symbolic politics 155, 157, 169 Tallberg, Jonas 9, 11–12, 14, 16–17 technical oversight (Fachaufsicht), federal agencies and 104 technocratic expertise, monetary policy and 48, 53 Telecom Italia 133, 139 testing hypotheses, executive or judicial agents and legislative principals 200 Thatcher government, Oftel (1984) 138 Thatcher, Mark 9–12, 17–18, 50, 155–6, 160, 175, 201 theoretical literature, monetary policy be delegated to NMIs 48 trade-offs 53–4, 67 ‘transaction cost’ analysis 24 ‘transaction cost’ economics, P-A framework 3 transaction costs of policy making 207–9 treaty revisions, intergovermental conference and 34 trustee, particular kind of agent 7, 88 UK administrative law jurisprudence 179 agencies police courts 177 BSE 132, 138 bureaucracy and 154 Competition Act (1998) 167 Competition Commission 151, 158–9, 167 constitution includes individual rights 190 delegated legislation 173 delegation and commitment to public interest 153, 165 discretion and 116 Environmental Agency 133 Equal Opportunities Commission 131, 135–6 executive agencies 105 executive powers centralised 15

Index

213

FTA 159, 161 fused executive/legislative actors 154 independent regulator in telecommunications 12 insider trading 138 inter-sectoral spill-overs of agency model 118 IRAs 126–30, 134, 137–8, 143 legislation after survey of choices 155 ministerial responsibility and 160 MMC 141, 158, 161 National Audit Office 142 NMIs achieved quickly 13, 17, 101 parliamentary system 103 position vis-à-vis courts and parliament 178 survival of agencies 170 symbolic efforts at competitive policy 157 two party system and strong party government 156 utility regulators 135–6, 138–40, 160 UK Charter of Rights 190 ultra vires, courts as policers of P-A relations 191 unemployment 17, 51, 54, 64 United Nations (UN) 41 United States antitrust and 150, 157 capital market regulation 109 constitutional judges serve life terms 88 courts’ statutory interpretation survive correction 177 delegation to IRAs 143 Democratic Party 51 discretion and 116 EPA 201, 204–5, 208 executive and ‘legislative centred’ agencies 112 Federal Insurance Oversight Agency 108 Federal Patent Office 108 Freedom of Information Act 205 FTC 150, 154, 201, 203–4, 208 institutional economics literature 154 judges as New Dealers in (1950s) 182 judicial delegation 180–6 judicial review 178 law of delegation 180–6 linkages by British policy makers 13 Office of Management and Budget 176 separation of powers 103, 111, 119, 211 Supreme Court 180, 184–5, 196 theories of private interest regulation 157 voting records 206 wide range of constraint categories 117 United States. Securities and Exchange Commission 109 University of Chicago 61 unreasonable delegated law, ultra vires 188 US Administrative Procedure Act see APA US anti-trust agencies (DOJ and FTC) 150

Index

214

US Environmental Protection Agency see EPA US Federal Trade Commission (1914) see FTC US states, analysis of delegation decisions 213 Utilities Act (2000) 141 Vickers, John (DGFT) 167 wages, helped by growing economy 54 ‘Washington Consensus’ 64 water, few IRAs for 143 Weberian neutrality 153, 160 Weingast, B. 154, 202–4, 206 West European politics, social science literature and 92 West Germany, GWB and 161 Western Europe delegation in parliamentary systems 201, 210–11 delegation to NMIs and difficulties 18 delegation to regulatory agencies and courts 217 independent agencies pose problems in 141 IRAs in 126–9 limitations exposed by analysis of delegation 20 patterns of (1980s and 1990s) 134 pressures in regulation (1980s and 1990s) 131 reforms of regulation (1980s and 1990s) 125–6 widespread delegation in democracies 215 Wilks, Stephen 9–12, 16–18, 175 Williamson, Oliver 37 Wilson, Harold 159, 168 Wood, B. Dan 204–6, 216 World Trade Organisation (WTO) 41 zones of discretion 5–7, 14, 16, 18 competition policy and 150 gaps and 92–3