TRANSNATIONAL COMMON GOODS
This page intentionally left blank
Transnational Common Goods Strategic Constellations, ...
10 downloads
584 Views
2MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
TRANSNATIONAL COMMON GOODS
This page intentionally left blank
Transnational Common Goods Strategic Constellations, Collective Action Problems, and Multi-level Provision
Katharina Holzinger
TRANSNATIONAL COMMON GOODS
Copyright © Katharina Holzinger, 2008. All rights reserved. First published in 2008 by PALGRAVE MACMILLAN® in the United States—a division of St. Martin’s Press LLC, 175 Fifth Avenue, New York, NY 10010. Where this book is distributed in the UK, Europe and the rest of the world, this is by Palgrave Macmillan, a division of Macmillan Publishers Limited, registered in England, company number 785998, of Houndmills, Basingstoke, Hampshire RG21 6XS. Palgrave Macmillan is the global academic imprint of the above companies and has companies and representatives throughout the world. Palgrave® and Macmillan® are registered trademarks in the United States, the United Kingdom, Europe and other countries. ISBN-13: 978–0–230–60585–5 ISBN-10: 0–230–60585–0 Library of Congress Cataloging-in-Publication Data is available from the Library of Congress. A catalogue record of the book is available from the British Library. Design by Newgen Imaging Systems (P) Ltd., Chennai, India. First edition: December 2008 10 9 8 7 6 5 4 3 2 1 Printed in the United States of America.
Contents
List of Tables
ix
Preface
xi
1 Introduction
1
Part I Strategic Constellations in Common Goods Provision 2 Common Goods Theory and Analytical Framework 2.1 Classical Public Goods Theory 2.2 Empirical Research on Common Goods 2.3 Common Goods and Collective Action Problems 2.4 Common Goods, the Prisoners’ Dilemma, and the State 2.5 Social Situations of Common Goods Provision
11 12 21 28
3 Case Studies 1: Attributes of the Goods 3.1 Cost-benefit Configuration: Credit Ratings as an Information Good 3.2 Demand-side Properties: Credit Ratings, Environmental Pollution, and Systemic Risk 3.3 Supply-side Properties: Global Warming, Biodiversity, and Locally Unwanted Land Uses
43
4 Case Studies 2: Attributes of the Groups 4.1 Heterogeneous Preferences: Capital Income Tax Coordination in the European Union 4.2 Heterogeneous Capabilities: Tax Coordination and International River Pollution
29 36
43 55 69 83 85 99
vi
5
CONT ENT S
Case Studies 3: Attributes of the Social and Institutional Context 5.1 Standards and Regulations in the Market for a Common Good: Credit Ratings 5.2 Trade Regimes and Regulatory Competition for Environmental Standards
6 Strategic Constellations and Collective Action Problems 6.1 Properties of Common Goods and Strategic Constellations 6.2 Collective Action Problems: A Typology
111 112 122 137 139 147
Part II Strategic Effects of Multi-level Provision of Common Goods 7 Transnational Common Goods and Multi-level Systems: Analytical Framework 7.1 Transnational Common Goods and National Borders 7.2 Theories of Multi-level Systems and Multi-level Provision of Common Goods 7.3 Multi-level Analysis: The Role of Homogeneity and Heterogeneity 8
Case Studies 4: Homogeneity and Heterogeneity 8.1 Homogeneity and Heterogeneity of the Groups: Lakes, Rivers, and Biodiversity 8.2 Heterogeneity of Homogeneous Constituencies: Global Warming
163 164 169 174 179 179 184
9 Case Studies 5: Distortion Effects through Representation 189 9.1 Distortion through Self-interested Representatives: Capital Income Tax Coordination 190 9.2 Distortion through Cooperation: European Monetary Union 194 10 Case Studies 6: Distortion Effects through Aggregation 10.1 Strategic Constellations in the L1, L2, and Single-level Games: Locally Unwanted Land Uses
209
210
CONT ENT S
10.2 Aggregation Effects in Cooperative Multi-level Solutions: Locally Unwanted Land Uses
vii
212
11 Conclusion
227
Notes
235
References
237
Index
253
This page intentionally left blank
Tables
2.1 2.2 2.3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.1 4.2 4.3 4.4 4.5 5.1 5.2 5.3 5.4 5.5 5.6
Traditional Taxonomy of Public Goods Extended Taxonomy of Common Goods Prisoners’ Dilemma in Common Goods Provision Provision of Ratings by Investors with Low Benefits Provision of Ratings by Investors with High Benefits Provision of Ratings by Investors: Private Benefits Higher than Private Costs Provision of Nonexcludable Ratings by Investors Exploitation of Common Pool Resources: Environmental Pollution Reduction of Systemic Risk in Global Financial Markets Summation Technology: Global Warming Weakest-link Technology: Biodiversity Best-shot Technology: Siting of Locally Unwanted Facilities Capital Income Tax Coordination with Revenue-oriented Governments Capital Income Tax Competition with Benefits-oriented Governments Capital Income Tax Competition with Heterogeneous Governments Capital Income Tax Competition of Small and Large Member States Environmental Pollution of International River Basins Provision of Ratings by Heterogeneous Borrowers Provision of Ratings with Limited Capital Supply Provision of Ratings in the Presence of Obligatory Rating Requirements Regulatory Competition with Homogeneous Countries Regulatory Competition with Heterogeneous Countries Type of Standards and Trade Regimes
15 19 30 50 51 52 56 61 68 74 78 82 95 97 98 101 106 115 117 120 125 126 129
x
5.7 5.8 5.9 5.10 6.1 6.2 6.3 6.4 6.5 7.1 8.1 8.2 9.1 9.2 9.3 9.4 9.5 9.6 10.1 10.2 10.3 10.4
TA B L E S
Regulatory Competition with Product Standards and Trade Barriers Regulatory Competition with Unilateral Harmonization Advantages Regulatory Competition with Product standards and Free Trade Results of the Analysis of Regulatory Competition Cost-benefit Configurations in Common Goods Provision Preference Orders for Four Symmetric Games Six “Hybrid” Games with Heterogeneous Players Pure Coordination and Matching Pennies Typology of Collective Action Problems Possible Combinations of Homogeneous and Heterogeneous Actors Global Warming and Type Y Actors in the Level 1 Game Global Warming and the Representatives in the Level 2 Game Capital Income Taxation Game between Taxpayer and Government Preferences of Level 0 Actors toward the Introduction of the Euro Percentage of Citizens Favorable to a Common Currency The Euro Game among Heterogeneous Governments Network Effect in the Euro Game The Euro Game among Homogeneous Governments The Siting Game at the Level of L0 Actors Costs and Benefits of the Facility The Siting Game in a Representative Democracy: Even Distribution of Voters The Siting Game in a Representative Democracy: Strongholds of Parties in the Jurisdictions
130 132 133 135 142 145 146 153 154 176 186 187 193 200 202 203 204 207 211 213 222 223
Preface
T
he foundations for this study were laid within the institutional and intellectual context of a wider research program on “Common Goods: Governance in Multiple Arenas,” which was being conducted by Adrienne Héritier at the Max Planck Project Group “Common Goods: Politics, Law, and Economics” in Bonn, Germany. The book developed from a number of case studies that used the perspective of public goods theory to analyze the problems arising from transnational environmental problems and the globalization of financial markets. As the approach seemed useful, the original cases have been transformed and complemented by additional case studies to fit into the overarching framework for the analysis of transnational common goods provided in this book. Four colleagues have inspired and influenced this work. Adrienne Héritier, now at the European University Institute, Florence, set me on the track of analyzing multi-level provision of common goods. She also encouraged me to take examples from two policy fields: transboundary environmental problems and global financial markets. Reinhard Zintl, University of Bamberg, called my attention to the fact that different properties of common goods may lead to different strategic constellations for their provision and thus to different collective action problems—which is one of the main underlying themes of this study. Fritz W. Scharpf, Max Planck Institute for the Study of Societies in Cologne, contributed another basic research question: Does the provision of transnational and global common goods within multi-level governance systems change the strategic constellations, and if so, in which direction? Finally, my study was in many ways inspired by Elinor Ostrom, Indiana University. Most importantly, I tried to respond to her call for more theoretical differentiation: As common goods problems are empirically so different, we cannot expect one theoretical model, the prisoners’ dilemma, to fit all situations. I am indebted to all four of them not only for conceptual
xii
P R E FA C E
discussions, but also for reading and commenting on draft chapters of the study. A number of other persons read earlier versions of individual chapters, commented on them, discussed conceptual questions with me, or helped me in some other way: Todd Sandler, Philipp Genschel, Michael Zürn, Joachim Behnke, Johannes Schmidt, Dieter Kerwer, Dirk Lehmkuhl, and Christoph Knill. I am grateful to them for the openness and friendliness with which they responded to my requests and for the considerable time some of them invested. As I am not a native English speaker, I needed some linguistic support. I gratefully acknowledge the assistance and advice provided by Mary Kelley-Bibra and Darrell Arnold.
Chapter 1
Introduction
T
he provision of transnational and global common goods becomes increasingly important as a consequence of economic globalization and of technological developments. For example, several big and worldwide financial crises since the 1990s showed that as a consequence of new financial instruments, such as derivatives or hedge funds, the decisions of individuals or single institutions may have substantial negative effects on the international economy. The protection against the so-called systemic risk has therefore become a global common good. Another example is mobility of capital that became a political goal for reasons of the creation of wealth, but lead to massive capital flight and tax evasion. As a consequence, international coordination of tax policies became necessary. So far, attempts to resolve these problems had only limited success. The same is true for some global environmental problems, in particular, climate protection. On the other hand, both in the financial sector and in the environmental field some international agreements have been concluded. Examples comprise the convention for the protection of the ozone layer, the biodiversity convention, or the Basle accords on the protection against systemic risk in banking. How can it be explained that international cooperation proves extremely difficult with some of these problems, while in other cases solutions are found relatively easily? The common good character of these problems and the lack of a central power in the international sphere are not sufficient to explain these differences. The classical analysis traces the problem of providing public goods to two properties of the goods, namely nonrivalry in consumption and nonexcludability from consumption. This theory lead to two propositions: First, the problem of common goods provision is identified with the strategic constellation of a prisoners’ dilemma. Second, common goods cannot
2
T R A N S N AT I O N A L C O M M O N G O O D S
be efficiently provided by the market and thus have to be provided by the state. These views of the problem of common goods are not wrong in general, however, they must be qualified in several respects. The idea of a common good is not necessarily equivalent to the strategic situation of a prisoners’ dilemma. Situations in which common goods are to be provided differ in many respects, and these differences influence the strategic situation. The characteristics of those situations include not only attributes of the goods themselves, but also attributes of the actors and collectives concerned, as well as attributes of the social and institutional context. All of these attributes influence the strategic constellation, which is therefore not necessarily a prisoners’ dilemma. The second proposition needs also some qualification. A state solution is neither always necessary, nor is it always possible or desirable. Most important in the context of this study is the argument that in cases of transnational common goods there is no single nation-state that can provide the good. The cooperation of several states is required. Transnational and global common goods have to be provided within multi-level governance systems, such as international regimes or the European Union (EU). Consequently, it is necessary to analyze under which conditions and how the multi-level structure affects the strategic constellations. Aims of the Research Building on these considerations, the following questions are addressed by this study: 1. Context characteristics and collective action problems What are the strategic constellations and incentive structures created by different characteristics of the goods, the actors and groups concerned, and their social and institutional context? Which types of collective action problem do these strategic constellations imply? 2. Multi-level provision of transnational common goods Do the strategic constellations change if transnational common goods must be provided by multi-level systems? Under which conditions will multi-level provision change the strategic constellation compared to single-level provision, and in which direction? Knowledge about the strategic constellation of actors and the types of collective action problems posed by the different kinds of common
INTRODUCTION
3
good problems allows to judge the differences between them and to understand the variance in successful provision of these goods at the international level. Several features of the problems create these differences and have to taken into account: the various characteristics of the goods to be provided, the differences in actor constellations, the variance in the institutional context, and, finally, the exact form of political multi-level governance structure. To answer the above questions, this study develops a concept for the analysis of common goods provision that is based on public good theory and matrix game analysis. Using a common methodology, the concept allows for the comparison of a wide range of empirical cases. In the first place, the study is an analytical approach that aims at further developing and differentiating the theory of common goods provision, and in particular of multi-level provision. However, in the course of the book the approach will be applied to a number of largescale real world problems. Theoretical Background and Main Argument The study is rooted in the economic analysis of public goods. However, it is primarily interested in real world political problems. It is not interested in finding optimal solutions and thus less normative than economics. I use the basic economic concept of public goods; however, I do not go into the very technical approach of mechanism design. Moreover, the study was inspired by Elinor Ostrom’s call for theoretical differentiation in the analysis of common goods problems. In several ways, this study is complementary to Ostrom’s work. First, it is primarily theoretical analysis, while Ostrom provides a great number of empirical cases. Second, Ostrom concentrates on one type of common goods, namely common pool resources, while I deal with a broad range of common goods. Third, with respect to the spatial scope, Ostrom selects local problems, while I choose transnational and global problems. Fourth, the consideration of multi-level provision is one of the central issues in my study, whereas it is only a minor aspect in hers. The main argument of the book proceeds as follows: The provision of transnational common goods is affected by the various attributes of the social situations in which the goods are to be provided, including the characteristics of the goods, the actors, and the institutional context. These attributes affect the strategic constellation and determine the type of collective action problem posed. This will be
4
T R A N S N AT I O N A L C O M M O N G O O D S
demonstrated through the analysis of a number of empirical cases with the help of matrix games. The characteristics of multi-level governance structures are of particular importance for the potential to resolve problems associated with the provision of transnational common goods. It will be shown under which conditions these characteristics affect the strategic constellations of common good provision and when multi-level provision of common goods makes a difference as compared to provision within a nation-state. Again, this will be demonstrated using empirical cases. Methodology Three methodological elements of this study may need some explanation. First, the status of the empirical cases analyzed needs clarification, and the selection of the policy areas requires justification. Second, matrix games are used almost exclusively throughout the book for the analysis of the strategic constellations in the empirical examples. The advantages and limitations of matrix game analysis deserve a few words. Finally, one remark is in order as to the methodology used for multi-level analysis. In this book a number of cases of common goods provision will be used for demonstration. Most of these cases involve large-scale political problems; they are concerned with transnational or global common goods and affect multiple levels of government. It is not the aim of these case studies to explain the empirical outcome or to fully reconstruct the story. Instead, certain characteristic attributes of the situation will be selected and varied according to analytical criteria. Thus, the individual case studies can best be called rational reconstructions of certain important and central aspects of a common goods provision situation. The cases serve to illustrate theoretical points, but—other than the usual cover stories of matrix games—they are not mere “toy examples.” Sometimes only one aspect is chosen for reconstruction; sometimes the models come closer to a complete reconstruction. In some cases, the models carry at least the potential for explanation of some elements of the story, as in the example of tax coordination in the EU. Some of the cases serve as illustrations for more than one attribute. For example, credit rating is used as an example of a nonrival but excludable good, and as an illustration of how certain legal rules affect the strategic constellation. Most cases are used both in the singlelevel and in the multi-level analysis parts.
INTRODUCTION
5
Therefore, the models lie in between very general theory and empirical description. They try to bridge the hiatus between theory and empirics by introducing some of the many empirically relevant factors into the theory. I incur the risk that this may neither satisfy the theorist nor the empiricist; however, I am convinced that we can learn more about common goods problems and their solution by bringing the theory a little closer to the empirical problems. The cases are selected from two policy fields: environmental policy and financial market regulation. Environmental policy is a classic case of transboundary common goods. It was only recently acknowledged that global financial markets also pose transnational common goods problems, after the worldwide liberalization of capital markets has shown that there are substantial risks associated with the free movement of capital. As a systematic comparison of policy areas is not intended, it is not necessary to select fields such that causal inferences are possible. The reason for taking two policy fields is simply to have a broader background against which to generalize. The analysis and conclusions drawn are valid in general, and not only for a certain policy area. This is underlined by the use of examples from two fields that are different in many respects—but not with respect to the common good character of the examples. Basic types of strategic constellation can be represented by 2 3 2 matrix games, although this is a great simplification. In common goods problems found in the real world there are usually more than two actors. There are also usually more than two strategies available. Moreover, actors will be confronted with some degree of uncertainty, and measuring costs and benefits will not be easy. In empirical common goods problems, players will often be able to communicate. Finally, in cases of repeated interaction and ongoing relationships among the actors, the single-shot game does not truly reflect the strategic situation: Many equilibria are possible if a game is played repeatedly. However, 2 3 2 games have the merit of demonstrating a given strategic structure very clearly; and much can still be learned by the analysis of stage games under complete information. They are very good models of the basic types of collective action problem, and they reveal a surprising amount of information as to how these problems can be solved. Their simplicity is at the same time their great advantage. As Morrow (1994: 312) states in his advice section for building models: “The single most important principle in modeling is simplify, simplify, simplify . . . there is no reason to rush to the most sophisticated techniques.” In fact, matrix games are the most parsimonious means of analyzing collective action problems.
6
T R A N S N AT I O N A L C O M M O N G O O D S
In general, my arguments are based on the assumptions of perfect rationality and of the homo oeconomicus “in his most ugly and perfect form” (Zintl 2001: 44). Although it is sensible to acknowledge that motivations such as altruism, norm orientation, and practical reason exist empirically, it is not a good basis for theoretical analysis to work with a mixture of motivations, as long as we do not know exactly under which conditions or to which degree we are confronted with which motivations (Zintl 1994, 1995). Thus, for theoretical analysis one has to pick one of the motivations. Picking norm orientation or altruism, however, would bereave the analysis of common goods and collective action problems of their point. There is no specific technique that is employed for multi-level analysis. Matrix games are used both for the single-level analysis and for the multi-level analysis. The second research question, whether and under which conditions multi-level provision changes the strategic constellation, implies the comparison of strategic constellations in single-level situations and in multi-level situations. Therefore, the strategic constellation that would result for a given common goods problem in a political single-level system is simply compared with the constellation that would result in a multi-level system within the same collective. Theoretically, it is sufficient to restrict the analysis to a two-level system. Plan of the Book Part I of the book deals with the strategic analysis of transnational common goods without taking into account that they have to be provided within multi-level systems. It assumes just “one level” of actors, states, or individuals. Chapter 2 starts with a brief overview of classical public good theory, introduces the terminological and conceptual points of departure, and argues that a wider approach to the analysis of common goods is desirable than has been pursued so far. In chapter 3, a number of properties of the common goods themselves, such as cost-benefit configurations, rivalry and nonrivalry, and various aggregation technologies, will be examined with respect to effects they have on the strategic constellation. Credit ratings, systemic risk in financial markets, global warming, biodiversity, and locally unwanted facilities are used as illustrations. In chapter 4, the focus is on the attributes of the collective that shall provide the common good. The attempts of capital income tax coordination among the EU member states and environmental pollution of international river basins will be used to illustrate the problems of heterogeneous collectives. Chapter 5
INTRODUCTION
7
analyzes the effects of different institutional contexts on the strategic constellation. Two cases are examined: the regulation of credit ratings and the effects of different trade regimes on the strategic constellation of states in environmental regulatory competition. Chapter 6 draws some general conclusions from the analysis thus far. A typology of five distinct basic collective actions problems is developed. Part II of the book deals with the particular aspect that transnational common goods have to be provided within multi-level systems of governance. Does the necessity to provide common goods within multi-level systems change the strategic structure of the situation and, if so, under what conditions? In chapter 7, the foundations for the analysis of common goods provision within multi-level systems are laid. Both a review of the literature and a discussion of the main problems of multi-level provision indicate that heterogeneity of actors at the different levels of government plays a decisive role for strategic differences between single- and multi-level systems. The role of homogeneity and heterogeneity of individual actors at the lower level and their representatives at the upper level is therefore analyzed in chapter 8. Pollution of international lakes and rivers, as well as biodiversity and global warming serve as examples. Chapter 9 examines those combinations of homogeneity and heterogeneity, which in fact lead to a change in the strategic constellation. The examples used to illustrate several kinds of distortion effects are capital income taxation, the introduction of the Euro, and the siting of a locally unwanted facility. Chapter 10 concludes.
This page intentionally left blank
Part I
Strategic Constellations in Common Goods Provision
This page intentionally left blank
Chapter 2
Common Goods Theory and Analytical Framework
T
his chapter gives an overview of theories of common goods and of the related empirical research. The theoretical literature originated and is still largely produced in economics; however, there has been considerable experimental and field research done by economists and social and political scientists. The basic economic concepts and the results of empirical approaches are presented in sections 2.1 and 2.2. Next, some terminological and conceptual distinctions are introduced. In section 2.3, I define and justify my use of the terms “common goods” and “collective action problems” in this book. The overview of public goods research shows that more detailed analysis is needed in two respects. First, many authors believe that the state should provide public goods. In section 2.4, this view will be discussed and qualified. Second, the overview of research makes obvious that the problems posed by common goods are manifold. Not only are there many different classes of goods that are not purely private, these goods have also to be provided within the framework of a social situation characterized by many different conditions. The strategic constellation of providing a particular common good is not only affected by characteristics like rivalry and excludability, but also by other characteristics of the social situation. Therefore, I argue in section 2.5 that a wider approach to the analysis of the problem of common goods is desirable that takes into account more of these circumstantial factors.
12
T R A N S N AT I O N A L C O M M O N G O O D S
2.1
Classical Public Goods Theory
The theory of public goods is rooted in welfare economics and public finance. Its basic concern is the observation that some goods show properties, which make them ill suited for purely private provision or individual use. As a consequence of these properties the market will produce inefficient results. The goods are provided or preserved in suboptimal quantity or quality, and thus a welfare optimum cannot be achieved. Although the origin of public goods theory is usually traced back to Samuelson’s (1954, 1955) seminal paper, parts of the theory had already been developed before. The existence of externalities, which are central to the concept of public goods, and their undesirable effects on collective welfare, had been discussed among economists since the early twentieth century. So-called Pigouvian taxes or subsidies, named after Pigou (1920), who had first proposed them, have been advocated to correct the inefficiencies induced by externalities. Before Samuelson tackled the problem of provision of public goods, Wicksell (1896) and Lindahl (1919) had already concerned themselves with the problem of “just taxation” for the provision of public goods provided by governments. It was the contribution of Samuelson (1954), which finally launched the discipline of public good theory. A huge body of literature has developed since, and models of the problem, normative solution proposals, as well as definitions and classifications have become increasingly differentiated. There are, however, few comprehensive presentations of the state of the art. The authoritative work is still The Theory of Externalities, Public Goods, and Club Goods by Cornes and Sandler (1996). Basic introductions into the theory can also be found in textbooks on welfare theory (e.g., Feldman 1997) or on public economics (e.g., Laffont 1988). The following brief overview of public goods theory is given on the basis of a number of important contributions. This includes discussion of the suboptimality problem of public goods, the tragedy of the commons, and definitions and classifications.
The Problem of Public Goods Paul Samuelson was the first to develop a model of public goods provision. He defines collective consumption goods by nonrivalry of consumption: Collective consumption goods are “goods . . . which all enjoy in common in the sense that each individual’s consumption of such a good leads to no subtraction from any other individual’s
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
13
consumption of that good” (1954: 387). According to this definition public goods are viewed as goods with extreme positive external effects: If one person’s activity has an effect on another person’s utility function, this is called an external effect. In case of a nonrival good, one person’s activity appears in the utility functions of everybody else within the scope of the public good. If one person provides the good, everybody can use it. In his model Samuelson derives the conditions for the optimal quantity of a “collective consumption good.” The optimal quantity of the public good is achieved, if the economy is in a Pareto-optimal state, that is, a state where one person can only be made better off by making another person worse off. The Samuelson optimality condition for public goods requires that the sum of the marginal utilities of the public good should equal the marginal cost of production of the public good. This is equivalent to saying that the aggregate net benefit of the good has to be maximized. The problem with this optimality condition is, that “no decentralized pricing system can serve to determine optimally these levels of collective consumption” (Samuelson 1954: 388). The so-called first fundamental theorem of welfare economics states that supply and demand are coordinated through the price mechanism in such a way that a general competitive equilibrium and, thus, a Pareto-optimal state of the economy are achieved (for a proof see Feldman 1997: 47–51). The theorem rests on the following conditions: perfect competition; neoclassical production functions, that is, no economies of scale exist; and convex indifference curves, that is, consumers have rational preferences. In the case of public goods, the theorem is no longer valid. The market is not able to secure provision efficiently, “other kinds of ‘voting’ or ‘signalling’ would have to be tried” (Samuelson 1954: 388). The reason for this is the fact that in the presence of public goods, rational and selfish individuals have an incentive not to reveal their true preferences for the public good. If a purely nonrival good has been provided, no one can be excluded from its use. Thus rational individuals have the opportunity and want to free ride on the good, while nobody will be ready to bear the costs for its provision. Consequently, the good will not be provided at all or it will at least not be provided in sufficient quantity. This is a suboptimal state, as many individuals would benefit from the good. Therefore, spontaneous market coordination cannot be relied upon in the case of public goods. Individual rationality leads to collectively suboptimal results.
14
T R A N S N AT I O N A L C O M M O N G O O D S
Collectively organized provision, however, faces the problem of rational incentives, as well. If someone tries to organize provision to overcome the suboptimal state and in doing so tries to collect contributions from potential users according to their personal benefit from the good, this is doomed to failure: Rational potential users will misrepresent their true preferences, to pay less, as they still are able to use the full amount of the good. Again, the contributions collected will not suffice to provide the optimal amount of the good. As Samuelson (1954: 388–389) puts it: “It is in the selfish interest of each person to give false [emphasis in original] signals, to pretend to have less interest in a given collective consumption activity than he really has.” Thus, Samuelson shows that the market mechanism does not produce optimal results. However, he also shows that other mechanisms, including state provision, face a problem of obtaining correct information about the preferences of the collective for the public good. Whether we rely on the market or on the state, in this theory the consequence will be underprovision of the public good. The Tragedy of the Commons In another classic article Garrett Hardin (1968) applies public good theory to the problem of the commons. As Samuelson, he directs attention to a situation, where the first fundamental theorem of welfare economics is not valid. Other than Samuelson, he thinks of “open access resources” or of “common property” rather than collective consumption goods. His examples are the commonly used village green, environmental pollution, and overpopulation. Thus, he thinks of resources provided by nature and used commonly by a certain group or by humankind. These resources have in common that their consumption is subject to rivalry, while nobody is or can be excluded from their consumption. Hardin illustrates his argument with the “tragedy of the commons.” In a village commons it can be expected that each cattle herder will try to keep as many cattle as possible on the pasture. As the land has a certain carrying capacity, this arrangement may not cause any problems for a long time. At a certain point, however, the benefits from using the pasture start decreasing as a consequence of overgrazing. Each rational cattle herder will look at his or her individual utility from adding one additional animal to his or her herd. The marginal benefits of an additional animal will fall fully to the individual herder, while all the herders will share the marginal cost of an additional animal in terms of decreasing benefits as a result of overgrazing. Therefore
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
15
it is rational for each and every cattle herder, to add another animal to his or her herd. “Therein is the tragedy. Each man is locked into a system that compels him to increase his herd without limit—in a world that is limited. . . . Freedom in a commons brings ruin to all” (Hardin 1968: 20). Therefore, in Hardin’s view “open access” or “common property” are arrangements that, only under conditions of nonrivalry, cause no harm. Under conditions of rivalry and scarcity they lead to the overuse of a resource. Definitions and Classifications The two seminal contributions discussed earlier used two different criteria for defining the problematic goods: nonrivalry of consumption and nonexclusion from consumption. In fact, both definitions extend to different phenomena, although both criteria may apply to the same good. The following classification of goods, which uses both criteria, was first introduced by Musgrave and Musgrave (1973) and can now be found in most textbooks on public economics or public finance. Whereas purely private goods are characterized by both rivalry in and excludability from consumption, purely public goods show properties of nonrivalry and nonexcludability. All other goods in-between these two extremes are usually called impure public goods (Cornes and Sandler 1996: 9). Two important subclasses of impure public goods are common pool resources (CPRs), including common property resources (commons; Hardin and Baden 1977) and open access resources, and club goods. Commons and open access resources have been dealt with above. Club goods are goods that are excludable, but partially nonrival, and thus similar to the marketable public goods in table 2.1. Examples of marketable public goods are computer programs, videos, books, or similar information goods. As both rivalry and excludability may exist in degrees, the criteria create two orthogonal scales rather than four distinct classes (Cornes and Sandler 1996: 9). I will come back to this in the following text. Table 2.1
Traditional Taxonomy of Public Goods
Rivalry of Consumption Nonrivalry of Consumption
Excludability
Nonexcludability
private goods marketable public goods
open access resources pure public goods
16
T R A N S N AT I O N A L C O M M O N G O O D S
The discussion of the defining properties of public goods in the economic literature is sometimes confusing (see also Blümel, Pethig, and von dem Hagen 1986: 244–245). First, there is no consensus over which criterion is the crucial characteristic of a public good. Some argue it is nonrivalry, while others say it is nonexcludability (Malkin and Wildavsky 1991: 358–359). Second, there is an “inflation” (Blümel, Pethig, and von dem Hagen 1986: 244) of terms used to substitute or differentiate the term “public good.” Different terms have been used to denote the same thing. For example, Samuelson’s “collective consumption good,” Musgrave’s “non-rivalry of consumption,” Buchanan’s “indivisibility of benefits,” and Blümel, Pethig, and von dem Hagen’s “joint consumability” all have essentially the same meaning. Moreover, the same terms have been used to denote different things. For example, nonexclusion sometimes has the meaning of nonexcludability for technical or economic reasons; but sometimes it means factual nonexclusion, that is, the absence of exclusive property rights (Blümel, Pethig, and von dem Hagen 1986: 248, 257). Third, a number of additional or different characteristics of public goods have been discussed, some of which are related to the two main criteria, some of which are independent. Examples are the concepts of rejectability, scarcity (Keohane and Ostrom 1995: 15), absence of exclusive property rights (Cheung 1970), or supply by the public sector (Buchanan 1968). Blümel, Pethig, and von dem Hagen (1986: 245–259) provide an overview of some of the terms, notions, and taxonomies, which—despite its clarity—leaves the reader confused because of the sheer amount of different approaches. Fourth, the phenomena that have been subsumed under the term “public good” have been increasingly extended. Not only have national defense, public security, or the legal system been labeled public goods, but also public regulation, deregulation, redistribution, democracy, justice, or even markets. “The ultimate ubiquity of public goods has been demonstrated by Bonus . . . who observed that the privateness of private goods is itself a public good” (Blümel, Pethig, and von dem Hagen 1986: 244–245). Despite the expansion and confusion at the levels of phenomena and terminology, there is still some common ground and a common belief that the term “public good” is useful. Most economists would agree that the notion of externalities is at the core of public good theory (Samuelson 1954; Bonus 1980; Blümel, Pethig, and von dem Hagen 1986; Cornes and Sandler 1996). Pure and impure public goods are special cases of externalities. The presence of externalities
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
17
can be traced back to the properties of (partial) nonrivalry or (partial) nonexcludability of some goods. Nonrivalry leads to positive external effects, while common use of rival goods leads to negative externalities. Externalities are basically interdependencies of production or utility functions of different persons, which remain uncompensated in a world where only the price mechanism is at work. This leaves us with a rather broad class of nonprivate goods, however, as there is hardly any activity that has no external effects. Differentiations and Extensions It has already been mentioned that the two main characteristics of public goods are present in degrees. With respect to exclusion goods are distributed over a scale, where at one extreme the cost of exclusion from the good is prohibitively high, and at the opposite end exclusion is costless. Nonexclusion as a legal regime or political decision, however, need not be correlated to high exclusion cost, but may have other reasons. Rivalry in consumption also occurs in varying degrees. Partially rival goods are usually called “congestible.” These are goods that are not perfectly rival; they can be shared, but not endlessly. Typical examples are roads or swimming pools. The more people enter a road or pool the more the utility decreases for each user. If these goods are excludable, as it is the case with swimming pools and roads, they are called club goods (Buchanan 1965). According to Cornes and Sandler, a “club is a voluntary group deriving mutual benefits from sharing one or more of the following: production costs, the members’ characteristics, or a good characterized by excludable benefits” (1996: 347). Club goods are different from pure public goods with respect to six features: first, they are subject to congestion or crowding; second, nonmembers of the club can be excluded; third, a club rests on voluntary membership, that is, the benefits of the good can be rejected; fourth, there must be a mechanism that monitors utilization of the club good, such that members can be charged tolls and nonmembers excluded; fifth, club goods can be provided to a population by more than one club; clubs then partition the population and may compete; sixth, while with pure public goods one is faced with only one allocation decision, the optimal quantity of the good, for club goods it is also necessary to decide on the optimal membership size (Cornes and Sandler 1996: 347–351). Adams and McCormick (1993: 111) and Blümel, Pethig, and von dem Hagen (1986) present a classification in which the consumption dimension has three classes: rival, congestible, and nonrival.
18
T R A N S N AT I O N A L C O M M O N G O O D S
Whereas the idea of congestible goods differentiates the basic taxonomy of public goods internally, another concept was introduced that extends or complements the taxonomy: the concept of network goods or network externalities. Network goods are defined as goods, where the “utility that a given user derives from a good depends upon the number of other users who are in the same ‘network’ as is he or she” (Katz and Shapiro 1985: 424, 1994). Thus, an additional consumer of the good causes a positive network externality, as the other users benefit from that individual’s participation in the network without compensating him or her. Classical examples for goods causing technological network externalities are telecommunications systems (e.g., telephone or e-mail) or other goods based on physical networks (water, electricity, gas, or railways). It is obvious that the utility of using e-mail increases as the number of other users of e-mail increases, because the opportunities to communicate with other people via this medium rise. Another classic example refers to standards in the case of system goods that consist of several components such as computer hardware and software, or the videotaping format. Network goods may be excludable, as in the case of physical networks (a user of AOL can be charged), or nonexcludable, as is often the case for standards (a user of a common computer language cannot be charged). The existence of network externalities presupposes nonrival consumption; it is a special case of nonrivalry or in fact “more than nonrivalry.” The problem with network externalities is inefficient allocation and market failure, as with all externalities. At the core of the problem is the fact that the marginal individual benefit of participating in a network is less than the marginal social benefit. As Liebowitz and Margolis (1994: 141) put it: “It is the tragedy-of-the-commons problem turned on its head.” As a consequence, it may be difficult to reach the critical mass of network members necessary to finance the network. The problem arises both at the outset, when a network is to be built up, as well as with transition from one network (or standard) to another. The second problem is therefore too many and too small networks. This will happen, if the critical mass can be achieved, but the positive network externalities cannot be exhausted (Blankart and Knieps 1992). Koelliker (2001) presents a taxonomy of goods that includes network externalities. A more complete taxonomy that takes into account all differentiations and extensions discussed might look like the one in table 2.2. The two dimensions used are a combination of the criteria excludability from consumption and factual exclusion as a result of a legal of political regime, as well as the effect of additional users on the benefits
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
19
Table 2.2 Extended Taxonomy of Common Goods Excludability and Exclusion Effect of Additional Users on Consumption
Excludability (“Costless”) and Exclusion
Excludability and (“Political”) NonExclusion
Nonexcludability (“Prohibitive Costs”)
Rivalry
pure private good (bread, cigarettes)
CPRs, common property resources (nature parks)
Congestibility
club goods (public pools, theaters, roads)
Nonrivalry
marketable public goods (information goods, videotapes, books) positive externalities in network goods, standards for system goods (EMU, phone, e-mail, computer software) team goods (seesaw, games, sport)
nonmarketed impure public goods (free theater in public places) political public goods (basic education, standard wages) politically set standards for system goods (TV color technology)
CPRs, open access resources (fish in the ocean, ozone layer) nonmarketable impure public goods
Complementarity
Minimum number of users
pure public goods (solar energy, antenna TV) standards for “metaphorical” networks (languages, programming languages)
from consuming a good. The latter includes rival, congestible, nonrival, and complementary consumption. The last category “minimum number of users” denotes cases where the consumption of a good requires technologically a minimum number of users. Thus, it is not about the standard feature of a network good, that a critical number of users are necessary to cover the costs of the network. A simple example of such a “team good” would be a seesaw, which requires that two persons use it, as it otherwise cannot be consumed. Chess or many card games require a minimum or exact number of players before they can be enjoyed. The review and the table show that there are a great variety of goods that are not purely private, as there are externalities related to them. However, the distinctions and extensions discussed so far are by no means exhaustive. It is important for the analysis to look very carefully at the good in question. Seemingly very simple goods may exhibit characteristics that reveal that the taxonomy presented earlier
20
T R A N S N AT I O N A L C O M M O N G O O D S
is by far not sufficient to capture all relevant aspects. Some additional aspects important for analysis shall be mentioned here: First, the commodity, or the physical “thing” itself, is not to be confused with the good it provides. Different goods provided by the same “thing” may exhibit different characteristics. A river, for example, may serve as a waterway for transportation of persons or goods (congestible); it may serve as a reservoir for drinking water (rival); as an opportunity for bathing (congestible); as an absorber for polluting substances (rival); or it may provide us with fish (rival). For some purposes, it is easily excludable (transportation, drinking water), for others exclusion is more difficult (bathing, fishing, pollution). Second, there is the problem of joint production. “[B]read as a source of nutrition is rival in consumption because any bread that one person eats is unavailable to anyone else. What is not rival is seeing someone else eat bread [emphasis in original]” (Adams and McCormick 1992: 110). The latter might be a positive externality. Another obvious example of that type is smoking cigarettes where the externality is usually judged to be negative. The externality of the consumption of a private good may justify intervention, but it does not turn the good as such into a public one. However, the combination of private and public goods generated by one and the same activity may change the incentives for provision. Cornes and Sandler (1994) show that with such goods, voluntary contribution to the public good can be perfectly rational. Similarly, Olson (1965: 66–91) proposed to stimulate contribution to the public good by combining it with “selective incentives,” that is, goods that can be privately appropriated. Third, a very basic distinction of common goods is whether the goods are provided by nature, as fish in a body of water, or whether they have to be provided by humans. This is again no clear-cut distinction, as in most cases some investment is necessary to exploit naturally provided goods. Still, there is a difference. The problem of the commons is one of overuse of a natural resource, while the problem of public goods is one of underproduction of a humanmade good or service. If a pure public good (perfectly nonrival) is provided by nature, such as radiation from the sun, there exists no problem at all. Nonrivalry, in this case, can be seen as a gift from nature. Fourth, time and space are important factors in combination with rivalry or nonrivalry. Books are seemingly very simple goods. They can obviously be marketed as private goods, as they are excludable. However, there may be a problem of underprovision, as books are partially nonrival. The nonrivalry of reading books is not perfect: At
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
21
one point in time, usually only one person can read one and the same book (although there is the possibility to read it aloud to a number of listeners). Over time, however, a large number of people can read it. That there is rivalry at one point in time facilitates market provision; that there is nonrivalry over time enables us to have the institution of public libraries. In general, in the case of congestible goods, time and space play a role. Crowding on a bridge, for example, takes place in given periods of time and only when space is somehow limited. Clubs goods often have a spatial dimension. This is true, for example, for bridges, airplanes, roads, and theaters. Other nonrival goods, as well, have a given spatial scope, given by an acoustical, optical, or otherwise physical radius. Music or a play performed in a pedestrian zone is nonrival only within the range where it can be seen or heard. Environmental pollution of air or water has a certain reach given by physical conditions. 2.2
Empirical Research on Common Goods
Although most of the literature on common goods is theoretical, there is also a growing body of empirical research. The first experiments on public good provision were carried out in the 1970s by sociologists and psychologists, quickly followed by experimental economists and political scientists. The experimental research became a comprehensive body of work, as new experiments were developed in response to previous ones. This kind of research has often been interdisciplinary (e.g., Ostrom, Gardner, and Walker 1992, 1994). Field research, however, has mostly been done by political scientists. In this section, I will first present the main results from the experimental approaches and then turn to field research at the national and transnational level. Experimental Research on Common Goods Experimental research on public goods has rapidly developed, after its start in the 1970s. As hundreds of experiments have now been conducted, it is almost impossible to give a comprehensive account. In the following, a brief overview will be given of the rationale for public good experiments, what kinds of experiments have been conducted, and what their most important results have been. There will be no detailed description of individual experiments or discussion of more specific results. The reason why social scientists and economists first turned to experimental research on public goods can be found in the main
22
T R A N S N AT I O N A L C O M M O N G O O D S
theoretical proposition of the theory of public goods. It is difficult to test with field data that the actual provision of a particular public good is suboptimal, and that individuals free ride and do not reveal their true preferences for the good. What can be observed is only the actual amount of a public good provided; as nobody knows what the optimal amount would be, the difference between the two is unknown and cannot be taken as a measure for the extent of free riding. Therefore, for researchers interested in the question of to what extent free riding and underprovision of public goods is an empirically relevant factor “this problem presented a natural opportunity for laboratory experiments” (Roth 1995: 30). In the very first public good experiments conducted by psychologists or sociologists, subjects were simply asked for their willingness to pay for a certain public good (Bohm 1972; Dawes, McTavish, and Shaklee 1977; Marwell and Ames 1979). These experiments yielded the result that the subjects do in fact contribute voluntarily to a public good. These early results, which contradicted the economic theory, provoked experimental studies by economists. Kim and Walker (1984) and Isaac, McCue, and Plott (1985) conducted experiments where repetition was introduced. Their main result was that in the first round of the experiments subjects contributed at a considerable level. However, in later rounds contributions dropped severely. Related to public goods research are experiments designed to test the level of cooperation in a prisoners’ dilemma. The problem of public goods can be represented as a prisoners’ dilemma game (see section 2.4). The prisoners’ dilemma experiments lead to similar results as the public goods experiments. In one-period (“single-shot” games) experiments, there is a fair amount of cooperation (Rapoport and Chammah 1965). In experiments with repeated games, cooperation decreases in the later rounds. Subsequently, a huge number of experimental studies have been conducted. A great number of interesting factors that might influence contributions to public goods have been tested: • size of “stakes” (marginal returns, value of public good, Isaac, Walker, and Thomas 1984), • group size (Ledyard 1995: 151–155; Sandell and Stern 1998), • heterogeneity of benefits (Bagnoli and McKee 1991; Rapoport and Suleiman 1992), • decision rules (Banks, Plott, and Porter 1988), • monitoring and sanctioning (Ostrom, Gardner, and Walker 1992; Lubell and Scholz 2001),
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
23
• group identification and nonanonymity (Orbell, van de Kragt, and Dawes 1988; Frey and Bohnet 1996), • past experience from interactions (Gautschi 2000; Lubell and Scholz 2001), • fairness (Camerer 1997, Fehr, Kirchler, Weichbold, and Gächter 1998), and • reciprocity norms (Orbell and Dawes 1991; Cosmides and Tooby 1992; Ostrom 1998). For some factors the evidence is pretty clear, while in other cases the effects are unclear or still unknown. Moreover, there are many interaction effects among all these factors. In sum, there is a huge body of experimental literature now. However, as Ledyard concludes, “there is precious little comparability, and perhaps as a result a lot of uncertainty remains about behavior in public goods environments” (1995: 141). There is consensus, however, on three main results of the experimental research (Ledyard 1995: 121; Ostrom 1998: 5–7). In one-period experiments and in the initial rounds of repeated experiments, some 40 to 60 percent of experimental subjects provide contributions. Thus, cooperation is larger than zero but smaller than would be optimal. Individuals do not regularly play the equilibrium strategies predicted by economic theory. 1. Voluntary contributions decrease with repetition. Cooperation declines in the later rounds; however, there is still no universal defection or free riding. Thus, behavior is not consistent with the game theoretic logic of backward induction, which predicts that rational individuals will not cooperate in a finitely repeated prisoner’s dilemma game. 2. Face-to-face communication raises the level of contributions significantly. This finding has been replicated many times, both in one-period and in repeated experiments.
Field Research on Local and Regional Common Goods Field research has mostly concentrated on one specific subset of common goods, namely local CPRs and has been influenced very much by the work of Elinor Ostrom. The research group around Ostrom has collected a large body of field data on the management
24
T R A N S N AT I O N A L C O M M O N G O O D S
of CPRs. Although these studies stem from researchers in many different disciplines such as anthropology, human ecology, rural sociology, history, economics, and political science, they have been gathered and synthesized by Ostrom under a political science research perspective. The group has collected literally thousands of case studies (Ostrom 1990: XV). Ostrom has summarized the empirical and theoretical conclusions from this project in her book, Governing the Commons (1990). Ostrom’s point of departure was the extreme normative conclusions usually drawn from the economic model of public goods. On the one hand, Hardin (1968) and others (Demsetz 1967; Posner 1977) proposed as the best solution to the tragedy of the commons that common property should be privatized whenever possible. On the other hand, most theorists concluded that the market failure associated with public goods and externalities should be remedied by the state. They assumed that only an external power was able to prevent rational actors from free riding. The proposals ranged from state provision of the public good to central regulation of local CPRs (Ostrom 1990: 8–13). Ostrom’s point is that there are intermediate solutions between the market and the state and that the users of a CPR may well be able to resolve their problem without the intervention of an external authority (Ostrom 1990: 13–18). Ostrom analyzes “small-scale CPRs, where the CPR itself is located within one country and the number of individuals affected varies from 50 to 15,000 persons who are heavily dependent on the CPR for economic returns” (1990: 26). The analysis is further restricted to renewable resources, to situations where substantial scarcity exists, and to situations where the appropriators are in a symmetric position. The case studies include inshore fisheries, mountain meadows, groundwater basins, irrigation systems, and communal forests. Thus Ostrom’s analysis is not about pure public goods, club goods, or other nonrival goods. It is about rival resources commonly used by a specific group. It is not about transnational problems and not about large-scale problems such as the depletion of the ozone layer, the world’s oil resources, or the allocation of satellites in space. Since individuals earn their living from the resource, they have a much stronger motive to solve problems of exploitation than an average citizen’s motive to solve the ozone layer problem. Finally, problems arising from strong heterogeneity of interests are excluded. My emphasis on Ostrom’s selection criteria is not meant to diminish her results in any way; rather, I want to spell out what kinds of common goods problems have been extensively studied in field
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
25
research. Ostrom’s focus of the analysis implies that the probability of successful self-governance might be higher in this case than in the case of large scale problems, as she acknowledges herself (1990: 26). However, it is by no means clear that self-governance would not be possible for large-scale problems, other types of common goods, or problems including heterogeneity. There is simply a lack of field studies for those kinds of problems. Ostrom’s book can also be read as a call for the development of a more differentiated theory. She selected a subset of common goods, CPR dilemmas, for which her results are valid. This subset is not uniform; there are various subgroups of CPR dilemmas such as those classified in Gardner, Ostrom, and Walker (1990: 340–346). Her argument is that the general models of public goods provision and commons exploitation are correct, but that they do not fit each individual problem. Thus, theories should be further developed before any policy recommendations are made. “The observation that the world is more complex than is presented in these models is obvious, and not useful by itself. What is needed is further theoretical development that can help identify variables that must be included in any effort to explain and predict” (Ostrom 1990: 183). This call for differentiation is also valid for transnational common goods and was a motive for the underlying analysis. Field Research on Transnational and Global Common Goods In political science, similar theoretical questions to those posed by the theory of common goods have been treated in the subdiscipline of international relations. In international relations theory voluntary cooperation and “governance without government” (Rosenau and Czempiel 1992) become relevant, as there is no supra-state that could provide public goods at this level. Transnational and global public goods are thus a “natural counterpart” to CPRs of the type studied by Ostrom. They differ with respect to some properties. The problems at the transnational or global level are usually large-scale in terms of geographic scope, number of concerned actors, as well as salience and relevance to humankind as a whole. Interests and endowments are often heterogeneous at this level. A solution imposed by the state is impossible, voluntary cooperation and self-governance are the only ways out. This implies, however, that it is fruitful to ask the same theoretical questions: When is voluntary cooperation likely and under which conditions is governance without government successful?
26
T R A N S N AT I O N A L C O M M O N G O O D S
The theoretical parallels between local CPRs and international problems were first acknowledged by Keohane, McGinnis, and Ostrom (1993: 2–5) and Keohane and Ostrom (1995). Most problems in international relations are in fact CPRs or public goods at the international or global level. Prominent examples are global environmental problems or international peacekeeping. The leaders of national governments face problems of collective action similar to individuals exploiting CPRs. Thus, Ostrom’s research on self-governance found a parallel in international relations theories that “criticized the view that world government or hegemonic rule were necessary for international cooperation on economic or ecological issues” (Keohane, McGinnis, and Ostrom 1993: 3; see also Keohane 1984; Snidal 1985; Young 1989). The contributions in Keohane and Ostrom (1995) focus on two factors that may have an effect on the amount of cooperation at the local and global levels: the number of actors and the heterogeneity among actors. These variables were chosen because there were different hypotheses relating to them in the literature on CPRs, on the one hand, and on international relations, on the other (Keohane and Ostrom 1995: 4–10). While Ostrom found that mere changes in the number of actors did not have strong effects, in international relations it was generally assumed that increasing the number of actors increases the difficulty of cooperation. In addition, while the CPR literature had argued that heterogeneity inhibits cooperation, in international relations theory it was claimed that heterogeneity might facilitate cooperation (Martin 1995). The first to use explicitly the concepts of economic public goods theory in his analysis of current global challenges was Sandler (1997, 2004). He distinguishes different types of international public goods and examines the effects of several factors that may enhance or hinder international cooperation. Furthermore, he asks which problems are likely to be solved in a decentralized manner and which can only be resolved by supranational institutions. He develops principles for the design of such institutions. Some of his main positive and normative conclusions are: • The evolution of cooperation is less likely in a global than in a local setting. • Cooperation among nations is most likely if the problem affects the current generation, is well understood, and involves only a small number of neighboring countries. In the case of global problems that are characterized by uncertainty and that affect only future generations negatively, cooperation is least likely.
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
27
• As problems differ along a number of dimensions, the need for intervention differs across problems. • Only if independent action of nations cannot solve the problem of international collective action, should cooperative forms be introduced. • Increasing income inequality among nations will affect different types of public goods differently. While in some cases increased inequality may be conducive to cooperation, in others it will hamper a solution (Sandler 1997: 212–214). In recent years, a number of theoretical and empirical studies of global and transnational public goods have been published (Dasgupta, Mäler, and Vercelli 1997; Buck 1998; Barkin and Shambaugh 1999; Kaul, Grunberg, and Stern 1999; Kaul, Conceicao, and le Goulven 2003; Kaul and Conceicao 2006). These volumes provide applications and empirical studies of problems of international cooperation, which explicitly take on the perspective of public goods theory. Most of the contributions deal with natural resources such as river basins, oil fields, tropical deforestation, desertification, and especially population growth (Libecap 1995; Rogers 1997; Sandler 1997); or with environmental problems such as global warming, the ozone layer, marine pollution, acid rain, nuclear waste, and biodiversity (Barrett 1999; Mitchell 1995; Ehrlich and Ehrlich 1997; Sandler 1997). There are also studies on such classic public goods as international peace and security, the dispersion of revolutions, or international terrorism (Sandler 1997; Mendez 1999). Other contributions relate to global health problems (Zacher 1999; Chen, Evans, and Cash 1999), to knowledge and information goods like the prevention of Internet crime (Stiglitz 1999; Spar 1999; Sy 1999), to cultural heritage and diversity (Serageldin 1999), and, finally, to financial markets (Wyplosz 1999). The literature on transnational and global public goods displays a particular interest in two factors: the heterogeneity of actors (Keohane and Ostrom 1995) and considerations of equality, equity, and distributional justice (Sandler 1997, chapter 7; Rao, Kapstein, and Sen, all in Kaul, Grunberg, and Stern 1999). Both factors were already present in the experimental and the CPR research, but they have been much more emphasized in the literature on international public goods. It seems that these problems became more obvious at the level of international cooperation. Empirical common goods research has so far yielded two main results. First, the strong theoretical predictions of general free riding,
28
T R A N S N AT I O N A L C O M M O N G O O D S
or “zero contribution,” to the public good have not been supported by experimental evidence. There is a substantial amount of cooperation and voluntary contribution. Second, field research has shown that groups confronted with the problem of providing or preserving a common good have been able to do so without the intervention of a central government. Successful self-organization of local CPR and also transnational common goods is a widespread fact. 2.3
Common Goods and Collective Action Problems
In this book I am not concerned with the definition and classification of common goods as such, but rather with the different political problems they pose and with the solutions to these problems. However, starting with a definition of the term “common goods” seems necessary for two reasons. First, the term “common goods” is relatively new. Thus, the introduction of yet another term must be justified. Second, as we have seen, there is already a host of semantically or conceptually related terms such as “public goods,” “collective goods,” “commons,” “collective action problems,” “social dilemmas,” and so forth. Some authors use some of them as synonyms, while they seem to mean different things when used by other authors or in other disciplines. In this context it is important to clarify what a term extends to. Throughout this book, the term “common good” will be used as a synonym for all goods that are not purely private. A purely private good is understood as a good that causes no externalities, that is, its production or consumption by an agent has no effects on the utility of other agents. Thus, common goods are defined as goods characterized by the presence of externalities. This definition has the advantage of capturing a number of different notions of goods such as pure public goods, club goods, CPRs, congestibles, or network goods. It is justified to subsume these various notions—and the related phenomena—under the term “common goods,” as common forms of action are distinctive for their use or for their provision. Depending on the exact nature of the good, common use may be unavoidable (at least at reasonable cost), possible, or even profitable; and the optimal provision of these goods may require some form of common (collective) action. The notion of an externality relates to this “commonness” of a good, as it implies the interdependency of individual utility functions relating to the production or consumption of goods.
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
29
Thus, the term does not mean something new or fundamentally different from the term “public good,” as it is used in expressions like “public goods theory”; rather, it is used as a collective noun for a variety of phenomena that are usually treated under this heading, not all of which, however, are pure public goods (Héritier 2002a). The other terms denoting types of common goods will be used as in table 2.2. Particularly, “public good” refers to nonrival goods, and “CPR” to rival and nonexclusive goods. The term “collective action” refers to joint action by a number of individuals to achieve and distribute some gain through coordination or cooperation. All difficulties that arise in the pursuit of collective action goals will be called collective action problems. It should be noted that this is a very wide definition that includes, for example, problems of coordination or distribution. It is thus different from the usual notion that equates collective action problems with social dilemmas, that is, situations where rational individual actors are trapped to produce collective suboptimal results in terms of aggregated welfare. In particular, this definition does not equate the term collective action problem with the prisoners’ dilemma, which is surely too restrictive as Taylor (1987: 18) shows. However, it is also much wider than Taylor’s definition, which states “that a collective action problem exists where rational individual action can lead to a strictly Pareto-inferior outcome, that is, an outcome which is strictly less preferred by every individual than at least one other outcome” (Taylor 1987: 19). This definition is only concerned with the welfare or Pareto-optimality aspects of collective action, while I use the term also to characterize coordination problems, potential inequality, or instability of the outcome of collective action. However, this does not include problems for collective decision making, which are not the result of strategic behavior, for example, problems of fundamental uncertainty about the consequences of alternative actions. It will become clear in chapter 6 why I choose such a wide definition. 2.4
Common Goods, the Prisoners’ Dilemma, and the State
It has often been concluded from traditional economic analysis that public goods should be provided by the state. Provision by the state is traditionally legitimized by the failure of the market mechanism to provide common goods efficiently. Game theoretic analysis leads to a similar result. As Cornes and Sandler note, noncooperative game theory has increasingly been employed for the modeling of problems of
30
T R A N S N AT I O N A L C O M M O N G O O D S
collective action by political scientists, economists, sociologists, and philosophers: “In recent years, two-player binary-choice models have been used extensively in economics to represent externalities, public goods contributions, . . . arms races, and other phenomena” (Cornes and Sandler 1996: 305). The exploitation of commons or the provision of public goods is generally understood as the strategic interaction of individuals within a certain strategic constellation, the prisoners’ dilemma (Taylor 1987: 13–14; Gibbons 1992: 27–29; Cornes and Sandler 1996: 19). Contributions to public goods can be presented as simple 2 3 2 matrix games (table 2.3). Two players, A and B, have the choice between two strategies: They can either “contribute one unit to the public good” or “not contribute.” Each of the four possible outcomes is associated with a certain payoff for each player (the “row player,” A’s payoff is given first), whereby 4 represents the first and 1 the last preference of the players. Both players’ first preference is for the outcome where they themselves do not contribute, but their opponent does. Their second preference is that both players contribute and have the benefit of the full amount of the good. Their second-to-last preference is that neither contribute and thus the good is not provided. The players’ last preference is the situation in which they themselves contribute to the good but their opponent does not. Given this incentive structure rational individuals will end up with an alternative that is both collectively and individually undesirable: Nobody contributes to the public good. The strategy “do not contribute” is a dominant one for both players, meaning that they will choose this strategy irrespective of what the other players does. Consequently, the only stable equilibrium is the outcome where the common good is not provided at all. This outcome is a
Table 2.3
Prisoners’ Dilemma in Common Goods Provision
Game Matrix
Player B Contribute one unit
Do not contribute
Contribute one unit
3, 3
1, 4
Do not contribute
4, 1
2, 2
Player A
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
31
Nash equilibrium, that is, it represents the players’ best responses to each other. The best responses of a player to the other player’s strategy choices are underlined in the matrix. At a Nash equilibrium both players’ payoffs are underlined. Conventionally, the two strategies in a prisoners’ dilemma are labeled “cooperate” (C) and “defect” (D). In a common goods game, “cooperate” is equivalent to “contribute” and “defect” is equivalent to “do not contribute”. The terms “cooperate” and “defect” imply the idea of an attempt to achieve a “cooperative” outcome—an outcome where some gain is realized through the voluntary cooperation of both players. Such a cooperative outcome, however, will not be achieved in a prisoners’ dilemma, as there is a strong incentive to defect after any cooperative agreement reached. Even if both players can communicate and would agree to contribute in a negotiation, for both of them there is an incentive to take a free ride later. If they keep their promise, they risk ending up with the least preferred outcome; if they do not keep their promise, they have a chance to obtain the most preferred result. As each must reckon that the other undertakes the same reasoning, both will defect from the agreement. Therefore, in theory, a prisoners’ dilemma can only be resolved by a binding contract to be enforced by an exogenous power—for example, the state. As Taylor points out, Thomas Hobbes was the first theorist who used this argument as a justification for the existence of the state. The public good Hobbes was concerned with was domestic peace and security. In Hobbes conception, the preferences of citizens for peace are analogous to the preferences of players in a prisoners’ dilemma game (Taylor 1987: 17). This line of argumentation implies that the problem of common goods is equated with the problem of solving a collective action problem of the prisoners’ dilemma type, and that both problems can only be overcome by the intervention of the state. There is a double equation here: Common goods are equated with the prisoners’ dilemma, and both are equated with the necessity for state intervention. This view of the problem of common goods is not wrong, in general. However, it must be qualified in several respects. There is a twofold break in the double equation. First, not every problem of common goods provision or preservation necessarily leads to the strategic constellation of a prisoners’ dilemma. Second, the existence of a common goods problem or the presence of a dilemma constellation does not imply that a state solution is necessary, possible, or successful in any case.
32
T R A N S N AT I O N A L C O M M O N G O O D S
Common Goods and the Prisoners’ Dilemma For many authors the prisoners’ dilemma is the canonical representation of the problem of common goods (Hardin 1971; Lichbach 1996). However, it has long been acknowledged that the strategic constellation associated with the provision of common goods is not necessarily a prisoners’ dilemma. For example, Godwin and Shepard (1979) have shown that CPRs have many different characteristics and that the prisoners’ dilemma does not properly represent all of them. Hirshleifer (1983, 1985) has shown how the aggregation technology—that is, the way individual contributions add up to the socially available quantity of the good—affects the structure of the games and thus the equilibrium solutions (chapter 3 for details). Taylor (1987: 34–40) presents a similar argument, referring to thresholds for the provision of the public good. If there is a minimum number of contributions necessary to provide the good, or if one or a certain number of contributions is sufficient to provide the good for all, the game is usually not a prisoners’ dilemma, but a coordination game (Holzinger 2008; Sandler 2008). Public goods, CPRs, and other collective action problems have thus often been analyzed as various types of coordination games (Runge 1984; Sandler and Sargent 1995), for example, as volunteers’ dilemmas (Rapoport 1988; Diekmann 1992; Weesie and Franzen 1998). Taylor adds that “hybrids” of the above mentioned games are also possible (1987: 39), where the players have different preference orders. This points to the fact that, in general, the strategic structure of the game depends on the exact relation of the individual and collective costs of the contributions and of the benefits from the good. It is well possible that the incentives for the individuals are such that the good is provided in sufficient quantity, regardless of being a nonrival or nonexclusive good. It might happen that both players individually are rationally motivated to contribute (Cornes and Sandler 1996: 310), or that one of the players values the good so highly that he or she will provide it regardless of what the other player does (Taylor 1987: 39−40). In these cases there is no problem of providing a common good and no suboptimality. According to the classic definition, there is thus no collective action problem. In Taylor’s view there is also no collective action problem present in assurance games. These games have two Nash equilibria, whereby one of the equilibria is Pareto-superior and is preferred by both individuals. Definitely, there is no dilemma in such a situation. However, there are still problems of coordination to
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
33
solve. In games like the one Taylor mentions (1987: 39−40), where one individual provides the good for the other, there is no problem of efficient provision, however there may be a problem of distribution or fairness (Camerer 1997). Thus, these constellations do pose as well what is called a collective action problem in my terminology. I will come back to this in chapter 6. Necessity of State Intervention It is neither legitimate to conclude from public goods theory that common goods should always be provided or preserved by the state, nor is it legitimate to conclude that the market failure will be remedied by state intervention. There are several reasons why the state may not be able to provide better solutions to common goods problems than the market, or why the state is not needed to solve the problems. These reasons are summarized here. First, the state may not be able to provide or regulate common goods more efficiently than the market. As Samuelson (1954) has shown, the state faces a problem of obtaining correct information about the preferences of the public for the common good and can thus not be relied upon to produce an optimal amount of the good. This problem has been tackled by economists through the development of normative mechanisms for public goods provision. However, the solutions developed by “mechanism design” research (Blümel, Pethig, and von dem Hagen 1986) ignore the procedures actually used in political decision making on common goods by the affected groups, states, and multi-level systems. States are not governed by central bodies whose desire is to maximize the collective welfare, but by collaboration of a great number of actors pursuing individual and institutional interests within formal and informal political decisionmaking processes, who codetermine the political goals and the outcome. Therefore coordination mechanisms and decision-making procedures influence the solutions of collective goods problems, as well. Thus, economics and game theory diagnose market failure, but it remains open, if state intervention leads to better results. Second, in many instances of common goods private provision is possible because of the character of the good or the preferences of potential users. Some goods are perfectly marketable although they exhibit some externalities. This is especially true for information goods, such as books, newspapers, and computer software, or for network goods. In these cases no state provision is necessary although state intervention—for example, subsidizing—may sometimes seem
34
T R A N S N AT I O N A L C O M M O N G O O D S
desirable. Regardless of the character of the good, there are some situations where common goods will be provided or protected by the users themselves. This might be the case if in a “privileged group” (Olson 1965: 47) some of the users have much more to gain from the good than others, and value the good so highly that they are willing to provide it at their own cost. Thus, there are situations where the private provision of common goods is perfectly incentive-compatible and no external intervention by the state is needed. Third, game theoretic analysis has shown that voluntary cooperation is possible in prisoners’ dilemmas. In the case of infinitely repeated play, it is a rational strategy to cooperate in each round of the game. It is, however, also rational to defect in each round, or to play any other strategy. Cooperation is possible as a result of reciprocal threats to sanction a player who defects in a certain round, by defecting in later rounds. With different kinds of threats, different strategies and outcomes can be supported. In an infinitely repeated prisoners’ dilemma any combination of feasible and individually rational payoffs can be achieved. This is the meaning of the so-called folk theorems (Gibbons 1992: 88–102; Morrow 1994: 268–279). Taylor (1976, 1987, Chapters 3 and 4) has presented this argument in the context of common goods provision for two-player as well as N-player prisoners’ dilemmas. Cooperation can emerge no matter how many players there are (Taylor 1987: 104). There are a number of conditions for cooperation that have to be met, however. Especially there have to be “conditional cooperators,” that is, players who cooperate on the condition that all other players do so. The more players there are, the more unlikely it is that these conditions will be met. Thus, in infinitely repeated games the resolution of the problem of common goods is theoretically possible without the help of the state, as a mere consequence of the fact that there is a common future of the actors. However, this result is not reliable, as suboptimal outcomes are also possible. Fourth, the theory states that without the possibility of concluding binding and enforceable agreements actors will not be able to cope with the dilemma. Usually, this is interpreted to imply that there has to be an external power to secure the enforcement of an agreement. This power, however, does not necessarily have to be a nation-state or a subnational authority. The external power could also be a private organization that enforces the agreements. The so-called Mafia solution of the prisoners’ dilemma (Holler 1983) is an example that is plausible in the context of the original cover story of the game: The Mafia can ensure that the prisoners do not confess by credibly threatening a
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
35
prisoner who sings with death. This example is rather unpleasant, but it is just to the point: From the viewpoint of theory, it is sufficient that binding agreements are possible and contracts are enforceable. The power that ensures this need not be a nation-state. Fifth, the empirical research presented in section 2.2 has shown that common goods can be provided and prisoners’ dilemmas can be solved without the help of the state. The general result of experimental research is that a substantial amount of cooperation takes place (Ledyard 1995). Case studies show that commons can be governed without the intervention of an exogenous power (Ostrom 1990). In practice CPRs can be managed by the concerned collectives themselves. The work of Ostrom shows that in fact the power to enforce can be created “internally.” Sixth, the solution to common good problems by the state is not always possible. Political borders and the scope of common goods are not always congruent. In the case of common goods that do not inherently possess a spatial scope, like defense, this does not cause any problem. Their scope can be adjusted to the scope of the jurisdiction. In case of common goods, however, that do inherently possess a geographical scope, like many environmental goods, the provision of the common good requires the cooperation of several jurisdictions or several levels of jurisdictions. If the scope is transnational or global, the cooperation of several or all states is required. In cases of transnational or global common goods, a state solution is impossible because there is no state at the global level. Taylor points out that at the international level states are themselves in a state of anarchy: “Hobbes himself noted that ‘Sovereigns’, who alone can save people from the state of (domestic) ‘War’, are themselves in a ‘state of nature’, without a ‘common power to keep them all in awe’ ” (1987: 166). Moreover, nation-states can produce collective action problems at the international level by providing common goods at the national level. The provision of defense against external enemies may lead to an armaments race among nations, creating the need for the common good of peace. It has often been concluded that a supranational state is necessary for the solution to international collective action problems. As long as no such state exists, however, international negotiation of the affected states is necessary. In sum, the theory of public goods can only to a very limited extent be judged as a valid justification of the state. Classical public goods theory tells us that the state is no panacea; game theoretic reasoning and the results of empirical research tell us that cooperation may sometimes be possible without the help of the state; and, finally there
36
T R A N S N AT I O N A L C O M M O N G O O D S
is not always a state that could fulfill the task of solving a common goods problem. 2.5
Social Situations of Common Goods Provision
The empirical research presented in section 2.2 has shown that the problem of common goods provision is not as clear cut as the economic model states. There are many variables that affect the outcome of common goods provision because situations where common goods shall be provided differ with respect to many properties. Common goods are defined by the (partial) presence of (one of) the two demand-side characteristics of the goods: nonrivalry of consumption and nonexcludability from consumption. However, the two basic defining properties are not the only attributes to play a role in the provision of common goods. The need for common goods arises in a given social environment, and the goods have to be provided within a certain social setting. These social situations have many different properties. Such differences may stem from properties of the good itself, but also from properties of the affected actors, or from other circumstances, such as the distribution of property rights. In many cases these attributes influence the incentives for the actors and therefore the strategic constellation. This implies that these attributes also determine the type of collective action problem the actors are exposed to, as well as the possibilities for finding a solution to the problem. Some of these attributes, such as group size, have already been the subject of extensive theoretical or empirical analysis. The research question has often concerned how these properties influence the degree of actual cooperation in dilemmas. However, the strategic constellation will influence the degree of actual cooperation, but it is not equivalent to it. Moreover, real world situations of common goods provision vary in many dimensions. The resulting strategic constellations may therefore be very different, and in some cases cooperation may be part of a rational strategy. As real world situations of common goods provision are so different, the classic models are adequate only under very restrictive assumptions. Theorists must make assumptions to keep their models simple. However, as Ostrom (1990: 184) puts it, [M]any of these assumptions are equivalent to setting a parameter (e.g. the amount of information available to participants, or the extent of communication) equal to a constant (e.g. complete information, or no
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
37
communication). Because the resulting model appears to be relatively simple, with only a few “moving parts”, it may be considered by some to be general, rather than the special model that it is. Apparent simplicity and generality are not, however, equivalent. Setting a variable equal to a constant usually narrows, rather than broadens, the range of applicability of a model.
The analyzes presented in the following chapters also work with a set of restrictive assumptions. They take into account, however, a number of particulars, which are kept constant in the classical prisoners’ dilemma model. It is neither possible nor fruitful to develop a specific model for each real world contingency. It is possible and fruitful, however, to develop models that take into account certain recurrent attributes of real world situations. In this way one can determine whether these recurrent attributes have a systematic effect on the strategic constellation. As “an endless number of attributes could be posited” (Ostrom 2002: 27), those attributes of a given common goods problem should be selected, which are considered most important. A single general theory for all common goods, however, no longer seems appropriate (Ostrom 2002: 29). Aggarwal and Dupont (1999) have pursued an approach along these lines in their article on “Goods, Games, and Institutions.” They observe that “the links between the characteristics of goods, the nature of strategic interaction between actors, and the effectiveness or need for international institutions have not been systematically treated” (1999: 393). This is not fully correct, as in political science and in economics some work of this kind has been done in recent years, for example, by Sandler (1997, 1998, 2004), Barrett (1998, 1999), or Mäler and De Zeeuw (1998). Nevertheless, Aggarwal and Dupont’s statement is well taken, as it implies that much more of this type of work is needed (Sandler 1998: 223). There are many properties of common goods provision situations that have not yet been systematically analyzed as to their consequences for the costs and benefits of the actors, and thus for strategic constellations. Characteristics of Common Goods Provision In the following I give a short overview of the potential properties exhibited in situations in which common goods are provided. This is not intended to be a comprehensive list of attributes. Only a few examples are mentioned. Broadly, four categories of attributes influence strategic constellations: (1) properties of the good itself, (2) properties
38
T R A N S N AT I O N A L C O M M O N G O O D S
of the actors, (3) properties of the groups involved, and (4) properties of the institutional context (4). Attributes of the Goods Many attributes of the goods themselves have a clear impact on the strategic constellation. First of all, this is true for the “defining properties” of common goods: Nonrivalry and nonexcludability are the properties that make free riding possible and provide the incentive not to contribute. These attributes of goods are the factors that lead to the social dilemma, that is, a situation where individual rationality does not lead to a collective optimum. More specifically, nonrivalry is the cause of undersupply in the case of pure public goods, and nonexcludability the cause of overuse in the case of CPRs (Haveman 1973). There are other demand-side properties of common goods, which may influence the incentive structure; for example, the nonrejectability of a common good or bad. It can also be of strategic importance, whether the good extracted from a CPR is mobile, like fish, or stationary, like wood (Schlager, Blomquist, and Tang 1994). In chapter 3, I explore in more detail the strategic difference between CPRs and pure public goods. The importance of supply-side properties for the incentive structure in the provision of common goods has first been shown by Hirshleifer (1983, 1985); the term “technology of public supply aggregation” was coined by Cornes and Sandler (1996: 184–190) and Sandler 1997: 47). Whether the contributions of the individual actors to a common good are additive or not, and whether they can be substituted for each another, is of crucial importance for the strategic constellation. In general, production functions of common goods—that is, how the costs of contributions turn into the benefits of the good—as well as allocation functions, which define how the benefits are assigned to the individuals, affect the strategic structure (Marwell and Oliver 1993; Ostrom 2002: 32–37). The strategic differences caused by different aggregation technologies or production functions will also be examined in chapter 3. Public goods are also different with respect to their geographical or personal scope: There are goods, which can be easily adjusted to the borders of given territorial jurisdictions. Examples are national defense, national criminal law, national rules intended to restrict risk in financial markets, or communal systems of waste treatment. There are other problems, however, where the spatial scope of the problem is determined by nature and cannot easily be influenced by political decisions. Examples are most environmental problems: Air or water
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
39
pollution does not stop at national or communal borders; climate change is a global problem, because it affects the whole globe as a result of physical conditions (Holzinger 2000). This difference becomes important, if multiple levels of governance are involved in a common goods problem. I will come back to this attribute of common goods provision in chapter 7. Attributes of the Actors Many studies of common goods provision are devoted to the question of whether the assumption of individual rationality of actors in economic theory is empirically valid or not (Ostrom 1998). The empirical research has shown that real actors do not always behave as the rational choice model predicts. There is a substantial amount of conditional cooperation, a willingness to sanction, a concern for fairness, as well as trust and reciprocity (Ledyard 1995; Ostrom 1998). Whether actors are altruists, whether they behave according to social conventions or to norms of fairness and justice, whether they comply with contracts, or whether they act boundedly rational—under each condition the outcome will be different than the outcome achieved by rational actors. These motivational factors are not analyzed in this book. The variation of other attributes of common goods provision and of multilevel governance employs game-theoretic modeling. The idea of a game-theoretic representation of incentive structures presupposes that the assumption of individual rationality be made. It is fully acknowledged that altruism, social norms, and other motivational factors play an important role empirically. However, rational actors form the basis for the analysis here. The information actors have about the situation is another relevant factor. In common goods models, it is generally assumed that actors have complete information about their own and the other actors’ strategies and payoffs for each possible outcome. Often this condition will not be fulfilled for real actors. Actors sometimes may not exactly know which strategies are available to them; the may not know what the payoffs for the actors are; they may or may not know which strategies have been or will be chosen by their counterparts; finally, actors may even suffer from uncertainty about their own payoffs, as a common goods problem may be characterized by fundamental uncertainty about the outcomes of certain actions. In general, I will work with the assumption of imperfect but complete information in this book, that is, actors know the strategies and payoffs of other actors; however, they cannot observe all previous or simultaneous moves.
40
T R A N S N AT I O N A L C O M M O N G O O D S
Attributes of the Group Attributes of the group include group size, the anonymity or nonanonymity of actors, the cohesion in groups, homogeneity or heterogeneity within the group, and whether the group members interact repeatedly, that is, have an ongoing relationship. The most important strand of research on attributes of the group, spawned by Olson (1965), deals with the effects of group size. While game theory predicts that the number of players does not in principle change the strategic constellation, empirical research has yielded different results (Isaac, Walker, and Williams 1994; Güth and Kliemt 1995). Changes in group size often imply changes in other variables, for example, in the marginal rates of return. Thus, the pure effect of the group size is difficult to isolate. Whether group size affects cooperation negatively or positively depends on how other factors, especially cost and allocation functions, or chances of monitoring, are influenced by a change in the number of actors. As there is already a large body of research on this, group size will not be dealt with here. The homogeneity or heterogeneity of actors is a further important aspect. Heterogeneity may stem from different benefits of the good, different costs of contributing, different strategies open to actors, and other forms of heterogeneous conditions that affect the preferences or the capabilities of the actors. It has been claimed that it is easier for heterogeneous actors to find a solution to the dilemma (Martin 1995). While this may be valid in certain situations, it is by no means a general truth, as has been shown, for example, by Hausken and Plümper (1999). Only few experiments have worked with heterogeneous actors, but they found that heterogeneity has a negative effect on cooperation (Bagnoli and McKee 1991; Rapoport and Suleiman 1993). Chapter 4 examines how different kinds of heterogeneity change the strategic structure of a common goods problem. Chapter 8 analyzes the effects of heterogeneity in conjunction with multi-level provision. Attributes of the Social and Institutional Context The social and institutional setting includes rights, rules, and conventions that apply in the respective situations, as well as opportunities among the actors to communicate, monitor, and sanction. As reported in section 2.2, experimental research into common goods has analyzed the effects of communication and sanctions (Ostrom, Gardner, and Walker 1992; Ledyard 1995). So far, not much research has been done on the effects that rules and institutions have on the
C O M M O N G O O D S T H E O R Y A N D A N A LY T I C A L F R A M E W O R K
41
incentive structure for common goods. The work of Ostrom, Gardner, and Walker (1994) is a notable exception. Rules, laws, and institutions may change the costs and benefits of the common good, for all or for some of the actors. A certain assignment of property rights provides an important example for a given institutional environment. Ostrom, Gardner, and Walker (1994: chapter 4) give a few examples for the assignment of different types of property or exploitation rights in fishing and show how changes in four different rule configurations change the game. Rules for decision making are another example for institutions: The effects of the unanimity rule on contributions were tested experimentally by Banks, Plott, and Porter (1988). As common goods provision does not happen in a social and institutional vacuum, there are always social conventions or legal rules that must be taken into account in an analysis of the situation. An infinite number of different institutional arrangements may affect the strategic constellation. Only a few examples will be examined here. In chapter 5, the effects of state regulation on the provision of credit ratings and the impacts of different trade regimes on regulatory competition for environmental standards are analyzed as examples for the consequences of particular institutional settings on the strategic constellation. If common goods are to be provided in multi-level government systems, this is also an institutional variable that affects the strategic structure. Multi-level provision is examined in chapters 7 through 9.
This page intentionally left blank
Chapter 3
Case Studies 1: Attributes of the Goods
In this chapter some properties of common goods will be examined
with respect to effects they have ceteris paribus on the strategic constellation. Only some basic and important characteristics will be taken into account. First, I would like to draw attention to a fundamental feature: the relation of the individual costs of contributions to the individual benefits derived from the good. Depending on the cost-benefit configuration, a good that exhibits the typical properties of a common good might very well pose no collective action problem at all. This will be demonstrated using credit ratings as an example. Second, the classic demand-side properties will be examined. Do different combinations of the rivalry and the excludability dimension result in different strategic constellations? Here, an exclusive and a nonexclusive nonrival good will be compared, as well as a rival and a nonrival exclusive good. The examples used are credit ratings, environmental pollution, and systemic risk in global capital markets. Third, the supply-side properties of common goods will be examined. I vary three extreme cases of aggregation technologies of the contributions to a good: summation technology, weakest-link technology, and best-shot technology. Global warming, biodiversity, and the siting of locally unwanted facilities are used as illustrations. 3.1 Cost-benefit Configuration: Credit Ratings as an Information Good It is the relation of the individual costs of contribution to the individual benefits derived from the good that determines the incentives for the individuals in a common goods provision situation. Even when a
44
T R A N S N AT I O N A L C O M M O N G O O D S
good shows some of the characteristics of a common good, its provision does not necessarily pose a collective action problem. Sometimes, the provision of a good is not even collectively desirable, given the individual cost-benefit relation. It is then not only individually, but also collectively rational not to contribute to the good. To be a common good, something must not only exhibit attributes of “commonness,” but it must also be valued as a good by the individuals and the collective. The cost-benefit configuration is not really a property of the good; however, it makes a “thing” desirable or undesirable, and thus a good or a bad. An analysis of individual contribution costs and benefits from a good is necessary before a certain strategic constellation like the prisoners’ dilemma can be diagnosed. As Cornes and Sandler (1996: 310) put it, “the configurations of benefits and costs are behind the payoff configuration assumed by a given game situation.” Each analysis of the strategic constellation in a common goods provision situation thus has to start by making reasonable assumptions about the individual costs and benefits in a concrete situation, and by making assumptions about how these add up to the payoffs achieved in the interaction of the players. While the latter depends on the kind of the good, the former depends on the valuations of the players. This method will be employed throughout the book. The example of credit ratings is used to demonstrate how varying assumptions about individual costs and benefits affect the structure of the game. I begin by providing some background information about credit ratings. Next, I analyze credit ratings as an information good and consider how costs and benefits affect the provision of the good. This analysis is continued in section 3.2, when I show how excludability influences the game, and in section 5.1, when I examine the effects of the use of ratings in public regulation. Credit Ratings Credit ratings are evaluations of the creditworthiness of borrowers. They are standardized assessments of the probability that a debtor will default on his or her payments, be it interest or capital repayment. Many different actors are usually involved in judging creditworthiness. Private investors, banks, and financial analysts must make such judgments. The most prominent form of such evaluations, which will be in the center of the following analysis, are credit ratings provided by specialized rating agencies. Rating agencies assess the bonds of several types of issuers: states, municipalities, banks,
AT T R I B U T E S O F T H E G O O D S
45
insurance companies, and industrial firms. Usually a certain emission of a bond is rated; however, the rating can also refer to the issuer as an organization. Ratings have long been an important instrument for the evaluation of borrowers by potential investors in their country of origin, the Unites States. After the liberalization of capital markets during the 1980s, the importance of ratings in international capital markets has increased. This is due to two factors. First, as a result of US regulation, access to important segments of investors, such as US pension funds, is limited for borrowers who have not been rated by an officially acknowledged rating agency. Second, a process of “disintermediation” has taken place since the 1980s. Traditionally, banks acted as financial intermediaries between borrowers and investors. Banks borrowed money from their savings customers, and then lent the money at their own risk to the borrowers. In recent times this role of banks has diminished, as mutual funds now get money from the depositors and pump it directly into the capital market. Corporate firms and other borrowers increasingly issue bonds instead of classic bank credits to finance their investments (Lütz 1997, 1999; Sinclair 2002; Strulik 2002). Private and institutional investors, such as mutual funds, pension funds, or insurance companies, diversify their portfolios, and thus need information on the creditworthiness of many different borrowers. Whereas banks have been able to assess the credit ratings of individual borrowers with whom they have had long-term credit relationships, institutional investors take more and more recourse to the standardized evaluations of rating agencies (Lütz 1999; Sinclair 2002; Strulik 2002). What are the functions of credit ratings for the capital market, or, more specifically, for the investors? Investors have a need for information that allows them to evaluate the creditworthiness of borrowers. For holders of bonds, however, it rarely pays to obtain and process all the information needed to judge realistically the credit standing of the issuer, not only at the time of emission, but over the whole term of the bond. The cost of gaining information and monitoring the issuer would certainly exceed its expected utility. Rating agencies provide this information in standardized and easily accessible form, as they categorize bonds or issuers according to their risk to default on payments. They use combinations of letters and/or numbers to distinguish classes of risk, ranking from high-quality “investment-grade” to low-quality “speculative-grade” emissions (Cantor and Packer 1994: 3; Basel Committee on Banking Supervision 2000: 23−24). This way, they transform the subjective uncertainty of investors into
46
T R A N S N AT I O N A L C O M M O N G O O D S
calculable risk (Strulik 2002: 321–325). The service of rating agencies significantly decreases the cost of information for the investors. In doing so, it also decreases the costs of debt for the borrowers, as investors who are not sufficiently informed about the risk of default would demand higher compensation for their risk, and thus higher interest rates would be the consequence. Low-interest rates and lowrisk borrowers would have to exit the market, which, in turn, would lead to higher interest rates, and so forth. Capital would become expensive and scarce (v. Randow 1996: 547). There are numerous similar information goods, where products or services are tested and rated by experts according to a standardized system. Examples are the ratings of hotels and restaurants, cars, personal computers, or of specific properties of products, like their effects on the environment. The results of the tests are usually published periodically and the publications are sold to the users of the information. In most cases, the ratings are provided by firms that belong to the publishing sector. Usually, it is not the firm offering the rated product, who pays for the rating. In the credit rating industry, however, the rated body—whether state, municipality, bank, or industrial firm—not only pays a fee for the rating, but it also actively seeks the rating. About 75 to 80 percent of the major rating agencies’ income is obtained from fees charged to issuers (v. Randow 1996: 553; Sinclair 2002: 285). Why is this the case? There are a number of possible explanations, which will be discussed in the course of the analysis of credit rating. However, not all rating agencies charge fees to the rated entities. There are still a number of them that charge fees only to the subscribers of their journals and newsletters, or to the users of their databases (Basel Committee of Banking Supervision 2000: 25). Moreover, the situation has not always been like that. Until the 1970s the major credit rating agencies obtained their income by selling their publications, as well. Why has the demand for rating switched from the investors to the borrowers? A short account of the history of credit rating serves to give some hints. Robert Dun and John Bradstreet published the first mercantile rating guides in the mid-nineteenth century (Cantor and Packer 1994: 1). Another precursor to ratings were statistical periodicals issued at the end of the nineteenth century, with the aim to provide financial markets with information about important businesses, and especially about unreliable borrowers like the railroad sector. Poor’s published the first “Manual of the Railroads of the United States” in 1868, while Moody’s started its “Manual of Industrial Statistics” in
AT T R I B U T E S O F T H E G O O D S
47
1900, which proved to be a “gold mine” (Sinclair 2002). After the financial crisis of 1907, these publications began to actually provide assessments of creditworthiness. The world economic crisis of 1929 was a further stimulus to the development of the rating business, and new rating firms entered the market. In 1931, for the first time a regulation was introduced in the United States that referred to ratings. Bank holdings of bonds had to be rated BBB (triple B) or better (investment-grade), if they were to be carried at book value; speculative-grade bonds were to be written down to market value. In 1936, banks were prohibited altogether from holding bonds that were not rated BBB or better (Cantor and Packer 1994: 5). Thus, ratings became a standard requirement for any emission of bonds in this phase (Kerwer 2002). After the spectacular default of Penn Central in 1970, investors became wary and hesitant, and thus a liquidity crisis threatened industrial companies. From the mid-1970s onward, a number of new and specific regulations came into force in the United States that made use of ratings. In this situation, borrowers began to actively seek ratings, and it became standard practice that new emissions have at least one rating. This started the transition to charging issuers instead of investors. With an increasing demand by borrowers for ratings, the agencies felt they were able to charge fees to them. Standard and Poor’s started charging municipalities in 1968; Moody’s and Fitch charged corporate bonds from 1970 onward (Cantor and Packer 1994: 4). During the 1980s, the use of credit ratings spread all over the world as a result of the liberalization and globalization of financial markets. American institutional investors, like pension funds, are central lenders not only for American, but also for international borrowers. American investors, however, are subject to US regulations not just with national but also with international investments. Thus, foreign borrowers have to be rated by a US rating agency, if they want to have access to this segment of the capital market. To fulfill the US requirements, ratings have to be issued from “nationally recognized statistical organizations.” Currently there are five American agencies recognized by the US banking supervision body, the “Securities and Exchange Commission” (Basel Committee on Banking Supervision 2000: 46). Finally, in the period after 1990, the global diffusion of ratings led to an increased use of ratings in international and national regulations. In 1996, an amendment of the Basle Accord on minimum capital requirements for banks of 1988 was adopted, which refers to
48
T R A N S N AT I O N A L C O M M O N G O O D S
ratings in specifying market risk and the necessary capital requirements. These rules were taken over into the European Union (EU) Capital Adequacy Directive (Council Directives 1993/6 and 1998/31) and have been strengthened in the Basle II Accord. For members of the EU, the rules are obligatory from January 1, 2007 (Council Directives 2006/48 and 2006/49); in the United States, implementation has been postponed to 2009. Credit Rating as a Common Information Good In this section, the analysis is restricted to the situation that obtained in the early phase of credit ratings and is still present in some countries and instances. This situation has three characteristics. First, the investors pay for the ratings, in the form of a subscription fee for a journal or for the access to individual ratings. Second, ratings are generally unsolicited. Until 1970, this was true for all ratings and it is still true today, whenever the investors pay. All agencies, which charge fees only to subscribers, carry out unsolicited ratings (Basel Committee of Banking Supervision 2000: 25). Third, access to ratings is limited to subscribers, which implies that exclusion is possible—at least to a certain extent. This was obviously possible in the early phases of credit rating, and it still is possible today. With only one exception, the ratings of agencies that charge fees to investors are not public (Basel Committee of Banking Supervision 2000: 25–27). In the following, I examine why and under what conditions rational investors contribute to the production of the information service provided by a rating agency. It was mentioned earlier that the individual cost of obtaining sufficient information about the creditworthiness of a borrower will often exceed the benefit this information can provide. However, if an information service provider like a rating agency collects and processes the information, costs for the users of the information will significantly decrease. There are three reasons why an information service provider can overcome the problem of high individual information costs. First, specialized agencies can use the information they have once acquired—for example, information on a state—to rate other entities—for instance, a bank or firm within this state. Thus, there are economies of scale in the production of the necessary knowledge. In addition, there are advantages to specialization: The agencies develop specific expertise. Second, each additional rating increases the amount of comparable material and, thus, the utility of ratings for its users (v. Randow 1996: 549). Third, and most importantly in the
AT T R I B U T E S O F T H E G O O D S
49
context of this analysis, the information and knowledge created by the agencies is a common good. Once provided, all potential investors can use the information about a borrower’s relative creditworthiness. The nonrivalry of information opens up the possibility that the users share the costs of information production. Information goods are often marketable public goods (table 2.2). This presupposes that potential users of the information can be excluded from consumption. It will be discussed in section 3.2 that in case of ratings this is true only to a certain extent. For the first step of the analysis, I assume that credit ratings are an excludable information good. The ratings are provided by an external agency, they are nonrival, but excludable, and the investors can share the costs of their provision through subscribing the rating. The costs of the rating are assigned to the investors by the agency, whereby the total costs are divided among the subscribers in equal shares. Thus, the individual costs decrease with the number of subscribers. Cost sharing is an important feature of the game. It represents here all three of the advantages of information provision by a specialized agency mentioned earlier. Three different situations will be looked at: (1) the individual benefit b is lower than the individual share of the costs c of the rating (b , c); (2) the individual benefit exceeds the total costs (b . C); and (3) the individual benefit exceeds the individual share of the costs, but is lower then the total cost C (c , b , C). There are two investors with identical preferences and thus identical payoff functions. They have two strategies: they can either contribute to the provision of a rating (subscribe, S), or not contribute (not subscribe, ~S). Table 3.1 gives the payoffs for the two investors, I and J, for each strategy combination. A general formulation of the payoffs is given along with an ordinal formulation. In the game matrix the general payoff is given in the first line, and the ordinal formulation in the second. Best answers and Nash equilibria are underlined. Investor I receives benefit b, whenever she subscribes, and receives nothing, whenever she does not subscribe. The cost share of c (c 5 ½C) is assigned to her, if investor J also subscribes; if J does not subscribe, she has to pay the full amount of C. Under the assumptions of case 1, her first preference is not to subscribe, her second preference is that both subscribe, and her last preference is, that she subscribes while J does not. The same considerations apply to investor J, as this is a fully symmetric game. The game in table 3.1 is a harmony game. There is a unique equilibrium in dominant strategies. At the same time, this equilibrium represents the unique Pareto-optimal outcome. From the perspective of both game theory and welfare economics this is an ideal incentive
50 Table 3.1
T R A N S N AT I O N A L C O M M O N G O O D S
Provision of Ratings by Investors with Low Benefits
Assumption, case 1
b , c, c 5 ½C
Strategy Benefit from Combination Rating Investor I
Cost of Subscription
Payoff
Ordinal
I: S
J: S
b
c
b2c
2
I: S
J: ~S
b
C
1
I: ~S I: ~S
J: S J: ~S
0 0
0 0
b2C 0 0
3 3
All factors are identical for investor J.
Game Matrix
Investor J
Subscribe
Subscribe
Not subscribe
b 2 c, b 2 c 2, 2
b 2 C, 0 1, 3
0, b 2 C 3, 1
0, 0 3, 3
Investor I Not subscribe
structure. The collectively optimal outcome is achieved. Both investors have a dominant strategy not to subscribe. The reason for this is the cost-benefit relation. The benefit is less than the individual cost of subscription for each player; hence, it does not make sense to contribute to the provision of the information good. There is no dilemma here. Still, this is a common good, as it is nonrival. Each player benefits from the others’ contribution, however the cost-sharing effect does not suffice to achieve a positive net benefit. It is perfectly sensible that the rating will not be provided in this game, as it is neither individually nor collectively desirable. The good is not “a public bad” as such, as it has some positive value, but its net benefit is negative. Under such conditions, however, ratings would not have been provided, either individually, or by specialized agencies. Thus, this cannot be the correct model of the US situation at the beginning of the century, when rating agencies developed. Table 3.2 shows what happens if the cost-benefit configuration is changed according to the assumption of case 2. If the individual benefit is very high, so that it exceeds not only the individual costs share,
51
AT T R I B U T E S O F T H E G O O D S
Table 3.2
Provision of Ratings by Investors with High Benefits
Assumption, case 2
b . C, c 5 ½C
Strategy Benefit from Combination Rating Investor I
Cost of Subscription
Payoff
Ordinal
I: S
J: S
b
c
b2c
3
I: S
J: ~S
b
C
I: ~S I: ~S
J: S J: ~S
0 0
0 0
b2C 0 0
1 1
2
All factors are identical for investor J.
Game Matrix
Investor J
Subscribe
Subscribe
Not subscribe
b 2 c, b 2 c 3, 3
b 2 C, 0 2, 1
0, b 2 C 1, 2
0, 0 1, 1
Investor I Not subscribe
but also the total costs, then the order of preference changes. The investors prefer a situation, where both subscribe, as they can share the costs. However, both are willing to bear the costs of the rating individually, as they still have a net benefit. Their last preference is not to subscribe. This game is again a harmony game, although its outcome is different. Both players have a dominant strategy to subscribe. In such a situation, the information on creditworthiness would probably be provided without the existence of a rating agency. The investors might do it on their own, neglecting the property of nonrivalry. However, they might also sense that, once obtained, the information can be sold to others and, thus, costs could be reduced. As already discussed, such a cost-benefit configuration is not very realistic for evaluating the creditworthiness of borrowers and emission of securities. In table 3.3 the cost-benefit configuration of case 3 is considered. The investors’ private benefits are now higher than their private share of the costs, but lower than the total cost of the rating. This is the most realistic picture of the situation at the beginning of the rating
52 Table 3.3 Costs
T R A N S N AT I O N A L C O M M O N G O O D S
Provision of Ratings by Investors: Private Benefits Higher than Private
Assumption, case 3
c , b , C, c 5 ½C
Strategy Benefit from Combination Rating Investor I
Cost of Subscription
Payoff
Ordinal
I: S
J: S
b
c
b2c
3
I: S
J: ~S
b
C
1
I: ~S I: ~S
J: S J: ~S
0 0
0 0
b2C 0 0
2 2
All factors are identical for investor J.
Game Matrix
Investor J
Subscribe
Subscribe
Not subscribe
b 2 c, b 2 c 3, 3
b 2 C, 0 1, 2
0, b 2 C 2, 1
0, 0 2, 2
Investor I Not subscribe
business. Carrying out a full rating individually did not pay. However, if the costs could be shared among many users, provision was collectively beneficial. The preference order changes again. The investors’ first preference is that both subscribe to the rating. Next, they would like the situation where they do not subscribe, irrespective of what the other player does. The worst case would be that they themselves subscribe, but the other player does not, as then the total costs exceed their benefits and they end up with a loss. The game in table 3.3 is an assurance game. There are two Nash equilibria in pure strategies, one at a low, and one at a high level of collective benefits.1 The equilibrium, where both players subscribe to the rating, is Pareto-optimal; the second equilibrium is suboptimal. None of the actors has a dominant strategy. The investors’ best strategy depends on the strategy of the other player. Subscription only makes sense if the other player subscribes as well. The problem here is coordinating the strategies such that both choose the same strategy, or better, that both choose to contribute to the rating. The early rating agencies were the entrepreneurs who served not only as information
AT T R I B U T E S O F T H E G O O D S
53
service providers, but also as the coordinators of potential users of a common good. The agencies helped the investors to organize cost sharing. Marketable public goods, like excludable credit ratings, are not so different from private goods. With private goods, if individual costs are lower than individual benefits, than collective costs are also lower than collective benefits, and vice versa. Thus, individually rational decisions lead to collectively rational outcomes; harmony games are played, whether the goods are provided or not. If individual benefits are above total costs, or if individual costs are above individual benefits, there is similar harmony of individual and collective interests in the case of marketable public goods. However, if individual benefits are above individual costs, but below total costs, the harmony is somewhat disturbed. There is a chance that the good will not be provided, although it would be collectively desirable and Pareto-optimal. In the case of credit ratings, this is caused by the nonrivalry property. However, there is a good chance to find a solution without state intervention, as it is not too difficult to optimally coordinate action in an assurance game. As nonrivalry opens up the possibility of cost sharing among a collective of users, collectively organized provision is efficient and possible. Private, profit-maximizing firms can do this, as the case of rating agencies shows. Cost sharing could also be self-organized by a group of users of the information. This would not be a club, however, as information is not congestible and there is no optimal size of the group (in fact the largest group is the best one). In practice, there would not be so much difference between the two solutions: A collective would have to organize and finance a “production unit,” and it would have to find a modus of sharing costs. The problem is less easy to solve, if exclusion from the benefits of the good is impossible. In section 3.2, I examine what happens to credit ratings if they are assumed to be pure public goods rather than marketable ones. It is the purpose of the present section to show that ratings will not be provided if costs exceed the benefits, or that they will be provided if individual benefits exceed the costs. It is obvious that goods are not desirable if the costs are higher than the benefits, either at the individual or the collective level. It is not very interesting that under such conditions goods would not be provided—common goods theory is about cases where goods are collectively desirable, but will not be provided. The point here is to demonstrate that the relation of costs and benefits significantly influences the strategic structure of a common goods game. Under certain cost-benefit configurations, voluntary
54
T R A N S N AT I O N A L C O M M O N G O O D S
provision is possible even if the good has the properties of a common good. Under other configurations, voluntary provision will not happen because the common good is not collectively desirable. In this case it is not a good, strictly speaking. Whether an entity is valued by an individual or by a collective as a good, whether it is judged to be a bad, or whether its net benefit is negative are consequences of the costbenefit configuration, irrespective of the properties that determine that entity’s commonness. Therefore, it is impossible to conclude from the presence of one of these properties that there is a collective action problem. The theoretical analysis of a given problem of common goods provision needs to start with (empirically adequate) assumptions about costs and benefits for the individual players that determine the cost-benefit configuration. Which cost-benefit configurations are possible depends on the exact circumstances. The conditions used in the credit rating example are: • symmetry: the players have identical strategies and payoff functions; • linearity: individual costs and benefits are aggregated in a linear manner to collective costs and benefits; • fixed costs for an indivisible unit of common good; and • costs can be shared among the players. Under these conditions five different configurations can be distinguished: (1)
c.b & c.B
⇒ C.B
(2)
c.b & c,B
⇒ C.B
(3)
c.b & C.b
⇒ C,B
(4)
c,b & C,b
⇒ C,B
(5)
c5b
⇒ C5B
negative net benefit, individual costs exceed collective benefits negative net benefit, individual costs less than collective benefits positive net benefit, total costs exceed individual benefits positive net benefit, total costs less than individual benefits individual and collective net benefit is 0
For the credit rating example, configurations 3 (case 3) and 4 (case 2) were analyzed. No distinction was made between configurations 1 and 2 (case 1), as the collective benefit was not considered here. The fact that it is necessary as a first step to analyze the cost-benefit configuration of a situation does not imply that this is sufficient for
AT T R I B U T E S O F T H E G O O D S
55
predicting the strategic constellation. Game structures are also influenced by other factors. Most important are the dimensions of rivalry and excludability, which determine how payoffs are assigned to strategy combinations. For example, it was the assumption of excludability from credit ratings that determined that investor I has no benefit at all from the rating if he or she does not subscribe (third and fourth lines in tables 3.1, 3.2, and 3.3). If this assumption is changed, the games might also change under all three configurations. While I will analyze nonexcludable ratings in the next section, I will no longer systematically vary cost-benefit configurations. Instead I will work with the most realistic case (case 3), where cost sharing enables investors to provide credit ratings. 3.2 Demand-side Properties: Credit Ratings, Environmental Pollution, and Systemic Risk This section is devoted to the strategic influence of the most important demand-side characteristics of common goods, namely nonexcludability and nonrivalry. As private goods are not the topic here, of the traditional four-cell classification of common goods, three cells are relevant: pure public goods, marketable public goods, and common pool resources (CPRs). Two conditions will be varied: the change from an excludable to a nonexcludable, nonrival good; and the change from a nonrival to a rival, nonexcludable good. The first comparison will be exemplified by the continuation of the analysis of credit ratings. The second comparison will be illustrated by environmental pollution and by systemic risk in financial markets. Excludable and Nonexcludable Credit Ratings Ratings were treated as marketable public goods, so far. This is true, however, only to a certain extent. Even if fees can be charged to subscribers of single ratings or journals, once disclosed, it is not possible to completely exclude nonpayers from the information. A reader of the journal or of a single rating report can convey the information to others orally; more than one person can read the publication; the publication can be copied; and so on. The more standardized the information is, the easier it can be passed on to nonpayers. A combination of alphabet letters can be very easily passed on. If one wants to read the differentiated reasoning behind the rating, one probably needs a copy of the report. Technology plays a role, as well. So long as there were no copying machines available to everybody, and
56
T R A N S N AT I O N A L C O M M O N G O O D S
investors were interested in detailed information about a firm, exclusion was not complete, but it worked sufficiently. Subsequently, investors diversified their portfolios and became less interested in detailed information while, at the same time, it also became easier to copy the information. As a result, exclusion became more difficult. Nowadays, ratings are placed on public Web sites or mentioned in the advertisements of a company. There is factual nonexclusion. This no longer poses a problem to the agencies, however, as the borrowers now cover a large part of the cost of ratings. Ratings, at least in their standardized form of a combination of alphabet letters, are effectively a pure public good for most investors, although many institutional investors still subscribe to the publications of rating agencies—albeit at relatively low prices. Table 3.4 gives the payoffs for an investor in case of a nonexcludable credit rating. The cost-benefit configuration is the same as in table 3.3. There is only one difference: If investor I does not subscribe or solicit the rating herself, but investor J does (third line), then investor I has the full benefit without any cost. Thus, nonexcludability offers the chance Table 3.4 Provision of Nonexcludable Ratings by Investors Assumption, case 2
c , b , C, c 5 ½C
Strategy Benefit from Combination Rating Investor I
Cost of Subscription
Payoff
Ordinal
I: S
J: S
b
c
b2c
3
I: S
J: ~S
b
C
1
I: ~S I: ~S
J: S J: ~S
b 0
0 0
b2C b 0
4 2
All factors are identical for investor J.
Game Matrix
Investor J
Subscribe
Subscribe
Not subscribe
b 2 c, b 2 c 3, 3
b 2 C, b 1, 4
b, b 2 C 4, 1
0, 0 2, 2
Investor I Not subscribe
AT T R I B U T E S O F T H E G O O D S
57
of free riding on the other’s contribution. The preference order of the investors is as follows: their first preference is that only the other subscribes; their second and third preferences are that both subscribe, respectively do not subscribe; their least preferred situation is that they themselves pay, while the opponent free rides. This game is now in fact a prisoners’ dilemma. Both investors have a dominant strategy not to subscribe to the publication, although both would be better off, if they did subscribe. Thus, if ratings are nonexcludable, they will not be provided at all. This obviously contradicts reality, as there are indeed public ratings, and the publications of rating agencies can still be sold. This model is inadequate for two reasons. First, the data, which is easily publicly available, does not satisfy the needs of, for example, institutional investors and financial analysts. They want to read indepth reports and to make use of the full coverage of the rating journals. Today, the subscription fees are so low (usually in the order of tens of dollars), compared to the direct costs and opportunity costs of free riding on the material, that it does not really pay to free ride. Second, the subscription fees are so low because there is now another mechanism to finance credit ratings: the borrowers pay for it. Thus, credit rating is now almost a pure public good, provided by the borrowers (through the agencies) for the investors. Cantor and Packer (1994: 4) explain the transition to charging issuers (partially) by nonexcludability: Agencies initially provided public ratings of an issuer free of charge, and financed their operation solely through the sale of publications and related materials. However, the publications, which were easily copied once published [my emphasis], did not yield sufficient returns to justify intensive coverage. As the demand on rating agencies for faster and more comprehensive service increased, the agencies began to charge issuers for ratings.
With nonexcludability, the explanation is not fully convincing, as obviously the ratings had paid very well for the agencies before. The availability of copying machines may have played a role, as already mentioned. The Internet came much later. Cantor and Packer mention as another factor the rising demand for ratings, without telling us, who exerted the demand: investors or borrowers. Finally, they offer as a second explanation the scarcity of capital supply after the default of Penn Central and a shift of demand to borrowers (1994: 4). Obviously, nonexcludability cannot fully account for the transition from investor-financed to borrower-financed credit rating. That credit rating has effectively become a pure public good, is more the
58
T R A N S N AT I O N A L C O M M O N G O O D S
consequence of charging the issuers, than it is the cause. I pursue the question of what has induced borrowers to provide credit ratings at their own cost, in section 5.1. The introduction of nonexcludability has changed the strategic constellation from an assurance game to a prisoners’ dilemma in this case. This does not imply that each time that nonexcludability is present, we are confronted with a prisoners’ dilemma. If we had started from another game, for example, from another cost-benefit configuration in the credit rating game, the introduction of nonexcludability would have yielded a different result. It is true that a dilemma is more likely if both nonrivalry and nonexcludability are present, but it is by no means necessarily so. There are still other factors, which may so strongly affect these basic features such that the dilemma can be overcome. This will be demonstrated in section 3.3. Environmental Pollution In the following the difference between rival and nonrival excludable goods is explored. CPRs are strategically different from the nonrival, pure public goods, as rivalry implies that goods may become scarce over time, while a truly nonrival good does not imply any scarcity. The effect of rivalry will be illustrated using the example of environmental pollution. Environmental goods are in many instances CPRs. Individuals and firms cannot normally be excluded from the use of the air, oceans, lakes, rivers, the landscape, or the atmosphere for a number of different purposes. The resources are used as reservoirs for discharges from human production or consumption activities. This is equivalent to environmental pollution. The quality of the environmental good deteriorates as a result of pollution and thus limits other kinds of uses of these goods. In addition, the resources may be depleted as a result of overuse of its stock or flows. The environmental goods are characterized by rivalry relating to most uses. Their stocks and their fruits are rival in consumption. However, the goods exhibit rivalry not only with respect to their quantity, but also with respect to their quality. Environmental pollution typically deteriorates the quality of the resource. In a CPR problem, the strategic structure is determined by the relationship of collective and individual marginal benefits. As a result of rivalry in consumption and scarcity of the open-access resource, the total benefit that can be derived from the good increases at the beginning (at low “amounts” of use), but begins to decrease after some point. Each unit of use causes negative externalities to the other
AT T R I B U T E S O F T H E G O O D S
59
users, which diminish the collective benefit. Therefore, collective marginal benefit, that is, the benefit of an additional unit of use, decreases. It is positive, as long as the total “amount” used is low and the good’s carrying capacity is sufficient. At a certain point, however, the collective benefit function turns, and collective marginal benefits become negative. In the end, the collective benefits become zero, which is equivalent to a collapse of the environmental good. As long as the collective benefits derived from the resource increase, the individual benefits increase as well. The individual benefits, however, still increase after the marginal collective benefits have become negative. As all users share the negative externality caused by each additional unit of use, the marginal individual benefit of an additional unit is still positive, while the collective marginal benefit becomes negative. Thus individuals continue to use the resource. This is the “tragedy of the commons” as described by Hardin (section 2.1). Only after the marginal individual benefits become negative, do individual users start cutting back their use of the good. The relationship of collective and individual benefits splits the process of exploiting the resource into three segments or phases. In phase I, both individual and collective benefits increase with additional units of use. Before collective benefits become negative, the scarcity of the good may not even be realized, and is in fact irrelevant for the decisions of the individual actors. In phase II, collective marginal benefits become negative, while individual marginal benefits are still positive. Individuals may begin to realize scarcity but still have an incentive not to react optimally to it. In phase III, both collective and individual marginal benefits are negative, and thus, it no longer pays for the individual to increase use. The overuse now drives the users’ surplus to zero, and thus they will gradually stop exploitation (Haveman 1973: 287). In each of the three phases, the users face a different strategic constellation. For the model, the example of pollution of a lake shall be used. The lake and its users could be situated everywhere. I shall assume that it is an international body of water, like the Great Lakes in Canada and the United States (Verweij 2000), or Lake Constance, which borders Germany, Switzerland, and Austria (Blatter 2000). The selection of this example serves an additional end: The analysis of pollution of international lakes will be contrasted later with the analysis of international river basins (sections 4.2 and 8.1). The lake in this example is an open-access resource. The players are two states that both use the water of the lake for different purposes, for example, for the extraction of drinking water and for the absorption of industrial and
60
T R A N S N AT I O N A L C O M M O N G O O D S
household discharges. The discharges are distributed within the water and deteriorate the quality of the water for other purposes. Thus, if one player uses the absorption capacity of the water, this produces negative externalities that are shared by the other player. The externalities are bidirectional. The players have two strategies: “to pollute” (P) or “not to pollute” (~P) the water. More exactly, “(not) to pollute” means (not) to add an additional unit of pollution. “Polluting” is a side effect of some other activity of the users that yields benefits for them. “Not to pollute” is equivalent to not undertaking this activity and thus sacrificing those benefits at the level of individual players, such as firms or consumers. As the players in the model are states, “to pollute,” in this case, means in fact not to restrict the pollution by individual users through regulation, while “not to pollute” means to regulate. To keep the model simple, only the direct benefits of the use of the resource and the negative externalities will be considered at the individual and collective level. Costs of undertaking the polluting activity will be ignored.2 Costs would be equivalent to the investments necessary to extract from the resource, in the sense of cost of harvesting a resource. For the three phases of pollution of the environmental good (or exploitation of the resource), different conditions are valid relating to the size of marginal individual and collective benefits, and the negative externalities. In phase I, marginal benefits, MB and mb, are both positive and the individual marginal benefit, mb, is higher than the negative external effect, e, caused by an additional unit of the polluting activity. In phase II, the collective marginal benefit, MB, is negative, the individual marginal benefit, mb, is positive, and mb is smaller than e but greater than ½e, which is the share of the additional externality each player has to bear for each additional unit of pollution. In phase III, the collective marginal benefit, MB, is again negative, while the individual marginal benefit mb is still positive. However mb is now smaller than ½e, so that benefit minus externalities becomes negative. The conditions and the preference orders are given in table 3.5. The general formulation is the same for all three phases. The varying assumptions about benefit and externalities change only the preference orders given in the last three columns of the table. The three game matrices show that in each phase the two states are in a different strategic constellation. In phase I, as long as the negative externalities have no severe effects on collective and individual benefits, the states play a harmony game: Both choose to add another unit of pollution. In phase II, again, both states have a dominant strategy
Table 3.5
Exploitation of Common Pool Resources: Environmental Pollution
Assumptions, phase I Assumptions, phase II Assumptions, phase III
MB.0; mb.e; e,0 MB.0; ½e,mb,e; e,0 MB.0; mb.½e; e,0
Strategy Individual Externalities Combination Marginal Benefit
State A
Payoff
Collective Marginal Benefit
Ordinal, phase I
II II
nP
nP
0
0
0
0
2
3
4
nP
P
0
½e
2 ½e
mb 2 e
1
1
2
P
nP
mb
½e
mb 2 ½e
mb 2 e
4
4
3
P
P
mb
e
mb 2 e
2mb 2 2e
3
2
1
All factors are identical for State B.
Game Matrix, Phase I
State A
Not pollute
State B Pollute
Not pollute
Pollute
0, 0 2, 2
½e, mb 2 ½e 1, 4
mb 2 ½e, 2 ½e 4, 1
mb 2 e, mb 2 e 3, 3
Game Matrix, Phase II
State A
Not pollute
State B Pollute
Not pollute
Pollute
0, 0 3, 3
½e, mb 2 ½e 1, 4
mb 2 ½e, 2 ½e 4, 1
mb 2 e, mb 2 e 2, 2
Game Matrix, Phase III
State A
Not pollute
Not pollute
Pollute
0, 0 4, 4
½e, mb 2 ½e 2, 3
mb 2 ½e, 2 ½e 3, 2
mb 2 e, mb 2 e 1, 1
State B Pollute
62
T R A N S N AT I O N A L C O M M O N G O O D S
to pollute, but, as the collective benefits are now decreasing because of the growing externalities, the game has changed to a prisoners’ dilemma. The states would be better off, if they would chose not to pollute, but they are entrapped in the tragedy of the commons. In phase III, the externalities of their actions now “hit back” with negative repercussions on the individual benefits; as a consequence not to pollute becomes a dominant strategy. This last game is again a harmony game, but harmony is now achieved by reduction of the polluting activity. Thus, rivalry and the effects of the negative externalities caused by overuse of the resource as an absorption facility finally drive the individual actions into the collectively optimal direction. At that point, however, it may be too late for recovery of the resource. This model only works, however, if the situation is perfectly symmetrical: All users of the resource undertake a polluting activity and all suffer from the polluting activities of the other users. In the case of heterogeneous actors, the result is different as will be shown in section 4.2 (see also Haveman 1973). Environmental pollution serves to show the strategic constellations present over three phases of exploitation of a CPR. How is this different from a pure public good, that is, a good that is both nonexcludable and nonrival? Unlike the case of credit rating, where the effect of taking away the property of excludability from a nonrival good could be shown by using that same good, the effect of taking away the property of rivalry from a nonexcludable good cannot be shown with the same good. This is due to the fact that rivalry in consumption is an inherent property of goods, while excludability is often a social or political construction. Therefore, another example will be employed to show the basic strategic properties of pure public goods: the reduction of systemic risk in capital markets. Reduction of Systemic Risk in Global Financial Markets As has become obvious earlier, credit ratings do not really pose a big problem for collective action, because they are marketable information services, and investors and borrowers have an incentive to provide such ratings. Nevertheless, there is a lot of regulation of credit risk in the financial markets, which, for example, takes recourse to ratings or to minimum capital requirements for institutional lenders. The justification for the existence of these regulations is the concept of systemic risk (Kaufman 1995; Burghof 1998: 50−55; Hellwig 1997).
AT T R I B U T E S O F T H E G O O D S
63
According to a study of the “Group of Thirty” (1997: 3), “[s]ystemic risk may be defined as the risk of a sudden, unanticipated event that would damage the financial system to such an extent that economic activity in the wider economy would suffer. To qualify as systemic [emphasis in original], shocks must reverberate through and threaten the financial system, not just some small part of it.” At the heart of the concept are so-called contagion effects, which are various forms of externalities. Contagion effects may occur as a result of the actual direct exposure of some other financial actor or sector to the damaged sector, or indirectly, because of suspected exposures. Thus, both factual interdependencies and the expectations of actors play a role in contagion. Contagion involves negative external effects, as the private costs of the initial failure of a bank or other financial institution are substantially lower than the social costs of a contagious crisis. Individually rational risk management by financial actors may lead to a level of systemic risk that is higher than that would be optimal at the collective level (De Bandt and Hartmann 2002). This market failure is the justification for state intervention into risk management in the financial sector (Burghof 1998: 50−52; Spencer 2000: 202). As mentioned, market failure may have two reasons: first, the expectations of market participants and, second, the actual interdependencies. The idea behind the first element of systemic risk is that of a “generalized” bank-run. In a bank-run, depositors liquidate their claims on the basis of information about the bank, for example, that the bank has a significant number of bad loans. If this information spreads widely among its clients, a panic may ensue, forcing the bank into failure. In a contagious bank failure situation, news of the bank failure spreads and causes clients of other banks to also withdraw their deposits, because they assume the conditions that were responsible for the failure of the first bank may also apply to their bank (“homogeneity assumption,” Burghof 1998: 82). It might also be the case that depositors do not react to the actual information on a bank’s bad loans, but just to the signal that other clients withdraw their deposits (“informational cascades,” Burghof 1998: 82). Such expectations of market participants become self-fulfilling. In this way, the failure of the first bank has contagious effects on the whole banking sector. In a systemic crisis the run may not be confined to the banking sector, but may spread into the bond markets, where participants will also try to liquidate their claims on financial institutions suspected to be weaker, shifting their portfolios to other institutions perceived stronger (“flight to quality,” De Bandt and Hartmann 2002).
64
T R A N S N AT I O N A L C O M M O N G O O D S
The behavior of clients, however, need not be irrational, as De Bandt and Hartmann (2002) show. If depositors receive the information that their bank has in fact made a number of bad loans, then their reaction is a rational adaptation of expectations. The outcome of “bank failure” would be efficient in this case, compared to the situation where a bank continues to operate and accumulate losses. If depositors react to information that is uncertain, then the reaction, to withdraw, can be right or wrong, and the outcome efficient or inefficient, depending on whether the information was correct or not. Finally, if market participants react to the signal that others withdraw their funds, then financial institutions are forced into liquidation without good reasons. The outcome, in the form of a self-fulfilling panic or “pure” contagion, is inefficient. This implies that the collectively optimal level of the probability of failure is not zero (De Bandt and Hartmann 2002). There needs to be a certain risk of failure to induce financial institutions to act responsibly. While pure panics and informational cascades should be avoided, rational reactions that force institutions deeply involved in debts into failure are in order. The second channel of contagion is the actual interdependency of financial institutions. Different institutions, regions, or sectors in the financial market have overlapping claims on each other. When on institution or sector suffers a failure, the other regions suffer a loss, because their claims in the institution or region are lost, or decline in value. If the spillover effects are sufficiently strong, the crisis spreads to other institutions, sectors, or regions (Burghof 1998: 91−93; Allen and Gale 2000). Apart from overlapping claims, additional transmission mechanisms are derivative financial instruments, as there is often shortage of liquidity in these markets, as well as settlement risks, which result from time lapses between the booking and the actual settling of transactions. In general, contagion can occur within the banking system, the bond markets, and in payment and settlement systems (De Bandt and Hartmann 2002). Over the last decades, numerous domestic crises in banking and financial markets have occurred. Well-known examples of domestic crises are the shutdown of the Bank of Credit and Commerce International (Kapstein 1994: 155−159) and the collapse of the Barings Bank in 1995 in the United Kingdom, or the Russian banking crisis of 1998. Most countries in Asia, Eastern Europe, and Latin America have experienced financial collapses (Horowitz and Heo 2001), and “more than a dozen have suffered systemic shocks that cost more than 10 percent of GDP to resolve” (Group of Thirty 1997: 4). Among the OECD countries, Japan, the United States,
AT T R I B U T E S O F T H E G O O D S
65
France, Spain, Norway, and Sweden have sustained financial crises that were resolved by their respective governments. Government intervention prevented these crises from spreading into the international financial system; however, a number of international shocks have also occurred since the 1970s. The first one was the failure of the German “Bankhaus Herstatt” in 1974, whose shutdown led to spillovers into international markets. The Latin American debt crisis of the 1980s is another prominent example (Kapstein 1994: 81−85). The problem threatened the solvency of many—especially many US banks. More recent examples include the Mexican liquidity crisis of 1995, which affected many countries that were merely suspected of having similar problems, and the East Asian crisis of 1998 (Group of Thirty 1997; Large 1998). Another case is the real estate credit risk crisis in the United States in 2007/2008 that quickly spread over the world. The risk that domestic shocks cascade through the international financial markets has substantially increased during the last three decades. This is a result of a number of developments in the global financial system. Volatility of exchange and interest rates, and thus prospects of gain, have increased since the 1970s. Some states, such as the Organization of the Petroleum Exporting Countries (OPEC), developed huge financial surpluses, which had to be invested in foreign countries. This was eased by deregulation: Most countries have liberalized their financial systems and have abolished capital controls, which, in previous times, aimed at walling off the domestic markets from global influences. Technological changes in communications systems have substantially decreased the costs of cross-border financial transactions. Financial products became more and more diverse, complex, and risky. The accounting systems, on the other hand, did not keep pace with increased volatility and complexity of products. Former clients of financial intermediaries started dealing on their own account, and thus a process of disintermediation took place. The growing importance of major institutional investors, like insurance companies or pension funds, made the financial markets more competitive and contributed to the globalization of the sector. All these trends have increased enormously the amount of foreign exchange, cross-border claims, and securities transactions (Group of Thirty 1997; Rotberg 1992: 11–14). There are, however, some positive developments, as well. Measurement of credit and market risk has been improved; portfolios are increasingly diversified; the use of collaterals has been enhanced; equity capital has been increased in major institutions; disclosure of information about risks, which cannot be
66
T R A N S N AT I O N A L C O M M O N G O O D S
read from the balance sheets, has been improved (Group of Thirty 1997). These developments have also contributed to the emergence of integrated global financial firms. International financial transactions are increasingly concentrated in a relatively small number of internationally active institutions. These 50 to 100 “core institutions” are commercial and investment banks; other big players, like international insurance companies, cause less risk to the system, as their failure implies fewer spillovers onto other actors than does the failure of the banks. The core institutions are usually well capitalized, headquartered in well-supervised countries, and heavily interdependent (Group of Thirty 1997: 39−42). Direct and indirect risk exposures within this group are complex and subject to rapid change. Thus, within these institutions, contagion through actual mutual exposure is more important than contagion through pure information cascades. This group of globally active financial institutions presents the largest threat to global systemic risk and is thus an important target of financial regulation. Thus far, domestic crises could often be prevented from spreading internationally. Usually this occurred through direct government intervention at the national level. Governments managed successfully to resolve crises on an ad hoc basis (Gavin and Hausmann 2000). Government rescue of failed financial institutions to avoid spillovers, however, has two major disadvantages. First, it causes high costs to taxpayers. Precautionary measures that prevent failure of financial institutions are less expensive for governments. If there are less initial failures, there is less systemic risk as a result of both contagions through interdependency and through informational cascades. Second, government bailout of failed institutions may create moral hazard: If financial institutions know that they are “too big to fail” or “too sophisticated to fail,” this does not provide them with very strong motivation to behave responsibly (Large 1998: 15; De Bandt and Hartmann 2002: 260). This risk of moral hazard leads one school of thought, the “free banking” school, to argue that the best reaction is a reduction of intervention by banking supervisors. The market will sort out the problems, insane firms will fail, and people will learn how to better manage risks (Burghof 1998: 97; Large 1998: 14; Spencer 2000: 211–218). Most economists and national governments, however, argue that systemic risk is a kind of market failure and therefore should be regulated at the national and international levels (Group of Thirty 1997). There are several kinds of regulatory measures that are already in use
AT T R I B U T E S O F T H E G O O D S
67
or that could be introduced (Aghion, Bolton, and Dewatripont 2000). A solution used in many countries is deposit insurance; this does not avoid moral hazard, however. The most important type of regulation is minimum capital requirements for financial institutions. Capitaladequacy ratios have not only been used at the national, but also at the international level. OECD prudential regulation, the EU Capital Adequacy Directive (Council Directives 1993/6 and 1998/31), as well as the Basle Accords of 1987 and 2004 (Kapstein 1989, 1992; Genschel and Plümper 1996; Plümper 1996: 201–213; Hofmann 2001; Strulik 2002) all use minimum capital requirements for the reduction of credit risk. Some authors even argue in favor of capital controls, and this implies de-liberalization of international capital markets (Khor 2000; Yongding 2000). If it is agreed that contagion causes negative externalities, the reduction of systemic risk by prudential behavior of financial institutions or by regulation through national authorities or supranational agreements can be interpreted as a pure public good. A low level of systemic risk would be a public good, while a high level of systemic risk would be a public bad. Nobody can be excluded from a low level of systemic risk in national financial markets or global financial markets, because all market participants enjoy the benefits of higher security. A low level of systemic risk is also nonrival, as its “consumption” by one investor does not in any way change the level of risk for other investors. It is a transnational, or even global, public good, because, after the liberalization of capital markets, free movement of capital between most states is possible. Thus, foreign investors cannot be excluded from a low level of systemic risk in national markets. As core financial institutions are transnationally interdependent, and as depositors in a certain bank or investors in a certain security come from many different countries, the risk itself is transnationally contagious. Contributions to the public good can be made at two different levels. First, the financial institutions themselves can contribute to a lower level of global systemic risk by applying strict rules for the management of liquidity, credit, and other kinds of financial risks. Second, national authorities can contribute to the public good by regulating financial institutions, for example, by minimum capital requirements for certain types of risk exposures. As with environmental pollution, there are two kinds of actors at different levels: the concerned market actors themselves, or the states and their authorities as regulators. Because in the model of environmental pollution, states have been chosen as the players, in this model, core financial institutions are chosen.
68
T R A N S N AT I O N A L C O M M O N G O O D S
The financial institutions have two strategies: to apply “strict rules” (S) for risk management or to apply “lax rules” (L). Strict rules might imply that, in general, a higher share of credits is balanced by equity, or that the share varies according to the level of risk associated with different kinds of credits. Strict rules are costly in terms of additional equity costs. This cost difference to lax rules is the cost, c, of contribution to the public good. The benefits, b, of a contribution are given by the individual benefits of a reduction of systemic risk through a single contribution. It is assumed that the costs of the individual contribution are higher than its benefits, as equity costs are fairly high and the contribution of one institution will only slightly reduce systemic risk. Both players enjoy the effect of a contribution, but the cost is borne solely by the contributor. Thus, each financial institution will prefer a situation where the other institution reduces systemic risk and they themselves enjoy the benefits without bearing the costs. The players’ second preference is that both apply strict rules; the third preference that both apply lax rules. The situation that financial institutions least prefer is where they themselves make the contribution while the other firm does not. Payoffs and preference order are given in table 3.6. Table 3.6 Reduction of Systemic Risk in Global Financial Markets Assumption
2b . c . b . 0 Strategy Benefit from Combination Strict Rules
Institution I
Cost of Strict Rules
Payoff
Ordinal
I: S
J: S
2b
c
2b 2 c
3
I: S
J: L
b
c
1
I: L I: L
J: S J: L
b 0
0 0
b2c b 0
4 2
All factors are identical for institution J.
Game Matrix
Institution J
Strict rules
Strict rules
Lax rules
2b 2 c, 2b 2 c 3, 3
b 2 c, b 1, 4
b, b 2 c 4, 1
0, 0 2, 2
Institution I Lax rules
AT T R I B U T E S O F T H E G O O D S
69
This structure is a prisoners’ dilemma, as one would expect in the case of a pure public goods game. Both institutions would prefer a situation where both apply strict rules, to a situation where both apply lax rules. However, both have a dominant strategy not to reduce the level of systemic risk, as their costs of unilateral reduction of risk exceed their benefits of doing so. This analysis is only valid, however, if there are no other factors that strongly influence the strategic constellation—especially, the game changes with its aggregation technology. In this example, a summation technology has been assumed: Individual contributions simply add up to the collective benefit. In the next section, I will show what happens, if this assumption does not hold. 3.3 Supply-side Properties: Global Warming, Biodiversity, and Locally Unwanted Land Uses Apart from the cost-benefit configurations and the classical demandside properties of the good, there are also supply-side properties, which affect the strategic constellation. The term supply-side refers to attributes of the production of the goods rather than to attributes of its consumption. Most important is the production function or the “aggregation technology” of common goods, that is, the way the contributions are aggregated. In the following, three extreme forms of aggregation technologies will be introduced, before an example for each of them is given. Global warming serves as an example for summation technology, biodiversity is used as an illustration for weakestlink technology, and the siting of locally unwanted land uses (LULUs) exemplifies best-shot technology. Aggregation Technologies Traditionally it has been assumed in public good models that the total amount, X, of a public good available to the collective is the sum of the individual contributions, xi. Hirshleifer (1983, 1985) points out that this “summation technology” (X 5 ∑i xi) is not the only possibility of an aggregation technology. He treats two cases of other production technologies where the good can only be provided as a fixed total amount whose level is determined by a single contribution. For “weakest-link technology” goods the total quantity is determined by the smallest contribution [X 5 mini (xi)], for “best-shot technology” by the largest contribution [X 5 maxi (xi)]. The two aggregation functions are extreme cases. Other functions in between are also
70
T R A N S N AT I O N A L C O M M O N G O O D S
possible (Cornes and Sandler 1996: 186−187; Sandler 1998, Arce and Sandler 2001). Hirshleifer provides two intuitive examples for weakest-link and best-shot technologies. His example of a weakest-link common good is about protection against flood on a circular island. Each citizen owns a wedge-shaped slice of the island and each builds a dike along the coastal line of the slice. As he assumes in a state of anarchy, the height of the dike is decided solely by the individuals. Protection against flood is here only as good as the lowest dike—the sea will penetrate at the slice with the lowest dike and flood the whole island. The dike can be seen as a chain: Each link is necessary for achieving the common good and the weakest link determines which level (quality or quantity) of the good can be achieved. The contributions are not additive and they cannot—physically—be substituted for each other. A piece of dike higher than the average cannot compensate for a piece that is lower. However, a player who would like to build a very low dike could be financially compensated by the others, to induce him or her to also build a minimum height dike (Vicary and Sandler 2002 for this argument). Hirshleifer’s cover story for a best-shot aggregation function is as follows: A city is protected against nuclear attack by several antimissile batteries. All of them are supposed to fire at a single incoming nucleararmed missile, which will devastate the whole city if not destroyed by the antimissile devices. In this situation, the best defensive shot is sufficient to provide the good for the city. Again, the contributions are not additive, a single shot is enough. In Hirshleifer’s example the contributions cannot be substituted for each other, as “the best shot” is required. This is not true for all aggregation technologies equivalent to “best shot.” One can easily imagine situations where each individual is capable of providing the single necessary contribution. In a similar example, the city has to be protected against a dragon attack. If there are several equally experienced dragon-slayers, any of them can go and kill the dragon. Then, the question is: Who will do it? In sociology and social psychology, best-shot situations are known as volunteer’s dilemmas or as missing hero dilemmas (Schelling 1978; Diekmann 1992; Weesie and Franzen 1998). Different aggregation technologies result in different strategic constellations. In terms of matrix games, summation technology leads to a prisoner’s dilemma, weakest-link technology to an assurance game, and best-shot technology to a chicken game (Sandler 1997: 46–59; Sandler 1998; Holzinger 2001). Aggregation technology and the possibility of substitution are independent dimensions.
AT T R I B U T E S O F T H E G O O D S
71
Aggregation technologies may vary when the individual physical contributions are perfect substitutes, or, as economists say, are anonymous. Nevertheless, they may also be combined with physical contributions that are not interchangeable. However, the empirical difference between those situations where substitution is possible and those where it is not, does not always affect the strategic situation. Predominantly, the strategic situation is determined by the aggregation technologies. The possibility of substitution may still make a strategic difference. This will depend on the specific circumstances, for example, the number of actors. Capturing strategic constellations by two-by-two matrices implies reduction and simplification in comparison to the modeling technique Hirshleifer has used. In particular, the strategy set is continuous in Hirshleifer’s model as the individuals can choose their contribution quantities as they like. In a two strategy matrix, players can only choose to contribute or not to contribute, or between high and low levels of contribution, respectively. In a two-player game, two contributions is the maximum, we thus talk only about zero, one, or two contributions to the common good. In symmetric games, it is also generally assumed that the contributions are equal (the players contribute “one unit”), while the contributions in Hirshleifer’s model differ quantitatively in the mathematical formulation (largest contribution) and qualitatively in the examples used (best shot). It is the qualitative difference, which prevents physical contributions from being substituted for each other. In an environment of two players, two strategies, and of equal contributions the equivalent to a weakest-link technology is the requirement that both players must contribute to provide the common good. In an environment with more than two players, the equivalent is that all or at least n players contribute. The equivalent to a best-shot technology is that the contribution of one player is sufficient to provide the good. If there are more than two players, the equivalent is that n players’ contributions are sufficient for provision. In the case of a summation technology, the contributions are restricted to a maximum of two in a 2 3 2 game, and to a maximum of n in the n 3 2 game. If contributions are allowed to be nonsubstitutive, this does not lead to a difference for summation technology goods—as long as the good can be provided in “degrees.” For best-shot technology goods, there is a difference between both players being capable of providing the good interchangeably and only one of the two players being capable of provision. There is no strategic interaction in the latter case: The player capable of provision just decides if he or she will or will not provide on
72
T R A N S N AT I O N A L C O M M O N G O O D S
the basis of his or her personal cost-benefit analysis. In the case of a two player game with weakest-link technology, there is no difference whether physical contributions are substitutive or not. The weakestlink character of the good requires both contributions anyway. However, if there are more than two, for example, m players, and at least n contributions are necessary, there is a difference. In the case of substitutive contributions m!/(m2n)!n! coalitions of n players are able to provide the good, while in the case of perfectly nonsubstitutive contributions only one such coalition is possible. These are different strategic situations. To illustrate some of these contingencies, I will go through a number of politically relevant applications from three fields: global warming, biodiversity, and siting conflicts. Global Warming In the case of global warming, the common good is a certain composition of the atmosphere. This composition keeps the climate and, as a consequence, the biosphere on earth within the parameters that we have adapted to, and to which we have accommodated our lives, culture, economic activities, and so on. In principle, all species on earth are concerned, although only humans can contribute to the preservation of the atmosphere. The composition of atmospheric gases and its effect on biosphere and humans is a CPR. Nobody can be excluded from enjoying its positive effects. However, it is used as a reservoir for dumping emissions, and with respect to this, there is rivalry. The common good is destroyed by the emission of six different gases, which change the composition of the atmosphere to produce the greenhouse effect. The greenhouse effect leads to global warming, and this in turn is expected to have serious consequences, for example, the flooding of low lands, an increase in heavy storms, and negative impacts on world food supply (Sandler and Sargent 1995; Sandler 1997: 99−102; Tietenberg 1997, part II). The effects are unevenly distributed over the world; some states will suffer first, and severely, while in other states only marginal effects will be felt. The same is true for the contributions to the destruction of the atmosphere: Some states emit much more of the greenhouse gases than others. In addition, global warming must be viewed as a unidirectional, intergenerational externality (Tietenberg 1997, part IV). The contributions to the preservation or restoration of a composition of the atmosphere such that there are no negative climate effects, are reductions in climate gas emissions. The reductions do not only produce benefits for the global climate, but also costs
AT T R I B U T E S O F T H E G O O D S
73
where the emissions are cut. The emissions result mainly from human production or consumption activities. A decrease in emissions also means a decrease in benefits from these activities. The contributions to the greenhouse effect are additive: The higher the level of greenhouse gas emissions, the greater the damage caused. The more the emissions decrease, the more of the common good will be preserved (see also Sandler 1998: 225). The necessity for emission reduction is generally accepted. The contributions are also substitutive: It does not matter where and by whom the greenhouse gases are emitted or reduced, respectively. The level of common good achieved is determined solely by the total amount of emissions, or by the sum of the contributions in terms of emission reduction. Not only are quantities of each greenhouse gas substitutive, but the six gases can also be substituted for each other. The effects of each gas in the atmosphere are different, of course, but their effects on the climate can be compared. They are usually expressed as GWP: their “greenhouse warming potential” relative to CO2. The above matrix game models the strategic situation of two states A and B, which have the options to contribute one unit of emission reduction (R) or not to contribute (~R). As the contributions are additive and substitutive the aggregation technology follows the summation rule. The actors are homogeneous, the game is symmetric. One unit of emission reduction creates a benefit of b and a cost of c, with c . b, to each player. The costs and benefits of each strategy combination for the states are given in table 3.7. This game is a prisoners’ dilemma. Both states have a dominant strategy not to contribute. This is a result of the assumptions of additivity and of the individual benefit of one unit being smaller than the individual costs. If the individual benefit is greater than the costs, the game changes into a harmony game, where both states have a dominant strategy to contribute. If c were greater than 2b, the game would also change to a harmony game. However, as in the prisoners’ dilemma both states would have a dominant strategy not to contribute. Unlike in the prisoners’ dilemma, the (~R, ~R) outcome would be socially optimal, as the collective net benefits are negative. Therefore, the interpretation of protection against global warming as a prisoner’s dilemma is only valid if the cost-benefit relation is as specified in the game earlier—and if we do not take into account heterogeneity, or intergenerational effects. The strategic situation between the states could also be modeled differently for other reasons. For example, Sandler and Sargent (1995; Sandler 2008) focus on the aspect of international treaty formation where a minimum number of signatories
74
T R A N S N AT I O N A L C O M M O N G O O D S
Table 3.7 Summation Technology: Global Warming Assumption
2b . c . b . 0 Strategy Benefits from Combination Emission Reduction
State A
Costs of Emission Reduction
Payoff
Ordinal
A: R
B: R
2b
c
2b 2 c
3
A: R
B: ~R
b
c
1
A: ~R A: ~R
B: R B: ~R
b 0
0 0
b2c b 0
4 2
All factors are identical for state B.
Game Matrix
State B
Reduction
Reduction
No reduction
2b 2 c, 2b 2 c 3, 3
b 2 c, b 1, 4
b, b 2 c 4, 1
0, 0 2, 2
State A No reduction
are required. Under these conditions, the strategic situation is similar to that of a weakest-link technology, as at least n contributions are required (Holzinger 2008). This model of global warming looks only at two aspects, namely the aggregation technology and the cost-benefit configuration. There are some more important characteristics of the strategic situation in case of global warming that would have to be taken into account to arrive at a realistic model of situation. As Barrett (1999: 198) notes, the 2 3 2 prisoners’ dilemma is not an adequate representation of the problem, as the climate game is at the level of states played by approximately 200 players, and as these have not only the two strategies of contributing and not contributing, but a continuous strategy set of cutting from 0 to 100 percent of their emissions. However, global warming has intergenerational effects. There are intertemporal externalities, that is, the negative externalities of current actions will be felt much later. This changes the cost-benefit relation for the current actors, as the benefits from cutting emissions are strongly discounted,
AT T R I B U T E S O F T H E G O O D S
75
while the costs of cutting emissions are calculated at the full present value. Furthermore, the states are heterogeneous both with respect to their contribution to global warming and to their preference for an effective climate protection policy (Barrett 1999; Benedick 1999). This heterogeneity implies distributional conflict, which can explain a great deal of the difficulties to reach and implement an agreement that the states actually face. These aspects may have greater effects on the strategic constellation than the aggregation technology (Barrett 1999). I will come back to the heterogeneity problem posed by global warming in chapter 8. Biodiversity Biological diversity is a very complex good. It is a good for several reasons. First, it facilitates ecosystem functions, which are of crucial importance for the continued habitability of the earth, for example, carbon exchange, regulation of surface temperature and local climate, or protection of soils. Second, the biosphere is, and will continue to be, the source of many products, such as food, fibres, and chemicals, and will serve as an input for biotechnology. Third, biodiversity is the basis for the development of new crop and livestock varieties and the improvement of existing ones. Fourth, it provides aesthetic, scientific, and cultural values and serves recreational purposes (OECD 1996: 19−25; Ehrlich and Ehrlich 1997). Although some of these goods are private in character, others, like the ecosystem functions mentioned earlier, are common goods. Therefore biodiversity as such may be treated as a common good. Biodiversity is in danger, as human activities destroy ecosystems, exterminate species, and threaten genetic variability. Protection of biodiversity means that we restrict our productive, consumptive, and recreational activities. This causes costs. Biodiversity is difficult to define and to measure, however. The definition of biodiversity in article 2 of the United Nations Convention on Biological Diversity is not about the protection of single entities. It emphasizes diversity or variability: Biological diversity is “the variability among living organisms from all sources including, inter alia, terrestrial, marine and other aquatic ecosystems and the ecological complexes of which they are part; this includes diversity within species, between species and of ecosystems.” Thus, the goal of biodiversity is pursued at three levels—genes, species, and ecosystems. The following illustrations and examples are mainly taken from species diversity. At each level, the diversity goal consists of two components: “One is a measure of how many different
76
T R A N S N AT I O N A L C O M M O N G O O D S
living things there are, the other is a measure of how different they are” (OECD 1996: 19). If biodiversity were to be measured only by counting the number of species, it could be achieved in degrees. The preservation of existing species or the creation of new ones would count as a contribution to biodiversity. The more species would be preserved, the more of the biodiversity goal we would have achieved. The contributions to biodiversity would be both additive and substitutive and, thus, the strategic situation would be a prisoners’ dilemma as in the global warming case—provided the individual costs of preserving one species were higher than its individual benefit but lower than the collective benefit. Intuitively, however, the species are not substitutes for each other, at least not in general. The breeding of a new fungi species may be viewed as a sufficient substitute for a similar fungi species that is extinct. However, the extinction of a bird would not be sufficiently compensated for by the cultivation of a new bacteria species. Also, to stay not only within the same order but even within the same family, the extermination of the elephant will not be judged as offset by the preservation of the manatee (the sea cow), the next and only remaining relative of the elephant. Here the second element of species biodiversity invariably plays a part, that is, how different the species are. The examples given above are of the so-called taxonomic diversity: Ten species in the same genus are valued as less diverse than ten species, each in a different genus. There is, however, no generally accepted way of balancing species richness (the mere number of species) against taxonomic diversity. Another important factor for substitutability of species is their ubiquity or their endemism, respectively. Biodiversity has a spatial dimension: It is measured with respect to a certain area, an ecosystem, a continent, and the planet. A species endemic to a certain area is clearly more worthy of preservation in this area than an ubiquitous species. Nevertheless, it may even be desirable for an ubiquitous species to be preserved within a certain area, as its extinction could cause the destruction of an ecosystem. This possibility hints at a third factor of different importance of species, and thus, of different degrees of substitutability. The extinction of a dominant species within an ecosystem may lead to the breakdown or significant change of the whole system—it may often be impossible to replace it by another species, even if closely related, without grave repercussions (Woodward 1994; OECD 1996: 20−25). As a consequence of all this, one could argue that biodiversity has to be viewed as a chain where the preservation of each and every
AT T R I B U T E S O F T H E G O O D S
77
species is required to achieve the goal. Each species counts, as each has its unique function within the global ecosystem. The contributions to the global common good of biodiversity are not substitutive. However, at the global scale this interpretation of the goal is not convincing either, as a certain degree of substitution of species, as well as a certain reduction of species richness seems possible without the complete loss of the good. The idea of a chain is more plausible for small ecosystems or biotopes. For modeling purposes one could think of a biotope where two dominant (or two crucial) species live. They interact in way such that the system breaks down if one of the two species vanishes. There are two actors who are each capable of protecting one of the species. Both species have to be preserved, as the loss of each one leads to the extinction of the complementary one. This is not an unrealistic scenario, although in most cases the interaction of more than two species is necessary to preserve an ecosystem, or sometimes only one dominant species may cause the breakdown of the biotope. An example of a two species case was Lake Nakuru in Kenya. The lake was inhabited by a very dense and dominant population of a specific blue alga, which served as food for a certain species of dwarf flamingo, which also had a large population and was a dominant one within this ecosystem. As a result of an unknown event, the alga suddenly vanished within one year. The dwarf flamingos followed. In the event, a green alga species population grew. It serves as food for plankton, which in turn now feeds the flamingo ruber (Remmert 1992: 309−311). The assumptions for the model are the same as stated earlier: There are two actors who can each preserve (P) or not preserve (~P) one species; payoffs are symmetric. The contributions are nonadditive and nonsubstitutive, the aggregation technology is of the weakest-link type. The individual benefit of achieving the goal (preservation of both species and the biotope) is b for each player. The costs of each contribution are c, with b . c. The cost and benefits of each strategy combination for the actors are given in table 3.8. This game is an assurance game. There are two Nash equilibria in pure strategies (this game has a third equilibrium in mixed strategies, which is Pareto-inferior), one at a low, and one at a high level of preservation (in the model: preservation and nonpreservation). Only the preservation equilibrium is Pareto-optimal. None of the actors has a dominant strategy; their best strategy depends on the strategy of the other. Preservation of one’s own species only makes sense if the other player preserves as well. The real problem here is of coordinating the
78
T R A N S N AT I O N A L C O M M O N G O O D S
Table 3.8 Weakest-link Technology: Biodiversity Assumption
b.c.0 Strategy Combination
Actor A
Benefits from Species Preservation
Costs of Species Payoff Preservation
A: P
B: P
b
c
A: P A: ~P A: ~P
B: ~P B: P B: ~P
0 0 0
c 0 0
Ordinal
b2c c 0 0
3 1 2 2
All factors are identical for actor B.
Game Matrix
Actor B
Preserve
Preserve
Do not preserve
b 2 c, b 2 c 3, 3
c, 0 1, 2
0, 2 c 2, 1
0, 0 2, 2
Actor A Do not preserve
strategies such that both choose the same strategy, or better, that both choose preservation. If the individual benefit from the biotope is less valued than the individual cost of preservation (b , c), the game is a harmony game again. Both players have a dominant strategy not to preserve their species. As the biotope—under this cost-benefit assumption—has a collectively negative net benefit, their behavior is socially optimal. If we observe that biotope or species preservation does not take place in reality this may simply be due to the fact that the actors value the biotope less compared to the preservation costs. However, it may also be due to heterogeneity. The biotope example, however, is still not completely plausible, there is a problem with respect to the actors. One could imagine a scenario where, for example, one of the actors is a farmer and the other a tourist. Both are capable of preserving one species and both suffer from the breakdown of the biotope. In general, the actors capable of preservation are political actors who can decide on the necessary regulations. If two species in a small biotope are to be preserved
AT T R I B U T E S O F T H E G O O D S
79
there will only be one territorial jurisdiction, in the earlier example the government of Kenya. One can, however, think of examples from species protection where two territorial jurisdictions are involved. Brown bears, for example, can be found in Europe and North America. Both Europe and North America can preserve their brown bear populations. If the goal were just to preserve the species the bear populations could be considered to be substitutes, although the aggregation function is not additive. It would be sufficient then, if the bears were protected in only one of the two areas. In the two-player case, this aggregation technology is equivalent to the best-shot technology. There are two varieties of bears, however: the European brown bear in Europe and the grizzly bear in Canada, Alaska, and the United States. Let us assume that biodiversity relating to bears is considered to be achieved only if both varieties are preserved. The bear populations are then no longer substitutes. The contributions of both continents are required to achieve the goal, and we are again in a weaker-link technology game. In Europe there are brown bear populations in Southern Europe, Eastern Europe, and Scandinavia. The goal to preserve the European brown bear is nonadditive and the populations are substitutive. Let us assume that political regulation prescribes that the brown bear be preserved in at least five populations in five different states. Then we have a scenario where nonadditive, but substitutive contributions lead to a weakest-link technology game (at least n), not to a best-shot technology (one is sufficient). As there are bear populations in more than five states, several different coalitions of five states could be formed to contribute to preservation. It should be noted here, that the aggregation technology is only one aspect of the problem of biodiversity. As in case of global warming, the modeling of biodiversity as an assurance game does not capture the whole problem. If the biodiversity problem were in fact an assurance-like coordination game, the negotiations of the biodiversity convention should have been easy. This was by no means the case: Negotiations were characterized by hard-nosed, positional bargaining; the agreement was rather imposed on the parties by officials of the United Nations Environmental Program under deadline pressure before the Earth Summit in 1992; finally, the convention was not ratified by the United States and renegotiated later (Moremen 1995; Raustiala 1997). These difficulties can be attributed to heterogeneity of preferences of the states. There is a host of distributional conflict associated with biodiversity protection, not only between the states at the international level, but also within nation-states between different
80
T R A N S N AT I O N A L C O M M O N G O O D S
private actors, such as the agricultural and other economic sectors, or environmentalists and the general public (Görg and Brand 2000). Siting of Locally Unwanted Facilities Siting conflicts are a very plausible example of a best-shot or a single-contribution aggregation technology. Siting conflicts arise from so-called LULUs, “locally unwanted land uses” (Mazmanian and Stanley-Jones 1995). The main characteristic of the goods associated with siting decisions is that the scopes of their benefits and their costs differ. The projects have a spatial dimension. While only the neighbors of a specific project suffer from negative external effects of the projects, the benefits generally accrue to a larger group of persons. LULUs are often public projects, but private projects may have the same consequences. Examples are the building of roads, motorways, or airports; the construction of waste management facilities; nuclear power plants and disposal facilities; prisons or psychiatric clinics; or the zoning of protected areas, which restricts economic activities. As the examples show, in many cases the goods provided by the project are common goods (often club goods), in other cases, like industrial facilities, the main benefits of the good are purely private. The common good attribute shared by all siting projects arises from their negative external effects on the neighbors. As a side effect the goods provided produce a nonrejectable public bad. Typical externalities are health or environmental risks caused by the emission of chemical substances, annoyances by noise, feelings of insecurity because of unusual people, or monetary losses because of a decrease in real estate values or restriction of economic activities. The siting conflict modeled here is assumed to be about a public project providing a good, which has the characteristic of a club good. One may think of a waste management facility or a nature reserve. The benefits of the project accrue to all citizens of the state or county, that is, to all club members. The contributions to the provision of the good consist of taxes for all citizens and/or fees for the users. The neighbors of the site, however, have additional costs to bear as they suffer from the negative externalities. Thus, the neighbors, for example, the inhabitants of communities, which can offer a site for the project, pay an additional, site-specific contribution. In some cases there are also site-specific benefits, for example, additional jobs for the community. They work as selective incentives. I will neglect this aspect here and only consider the site-specific costs.
AT T R I B U T E S O F T H E G O O D S
81
If a jurisdiction, the county for example, wants to provide such a locally unwanted land use good, a “hero community” is needed that offers the site. The physical contributions are nonadditive: A single site suffices to provide the good for all county citizens. Depending on the type of project, the contributions may or may not be substitutive. In the case of a nuclear waste facility, it is possible that only one specific site is suitable. The nature reserve, as well, has to be a specific area. In the case of waste management facilities or prisons, probably several communities in the county could offer a suitable site. If in fact only one site is suitable, no strategic game between potential site contributors will be played. The site will be offered voluntarily, if the benefits for the community are higher than the costs; otherwise the respective community can only be forced by state authority for the sake of the common good and/or it can be compensated for suffering the negative externalities. I will assume that both players are able to offer a site, and that the sites are substitutes for each other. The specific assumptions for the model are as follows: there are two communities with potential sites, which have symmetric payoffs. The benefit from the project is b for each community. The costs of the externalities are e, with b . e. The physical contributions are nonadditive, but substitutive. The two strategies are to offer the site (O) or not to offer the site (~O). Table 3.9 gives the costs and benefits of each strategy combination for the communities. The resulting game is a game of chicken. It has two Nash equilibria in pure strategies (plus a Pareto-inferior equilibrium in mixed strategies), where the payoffs to the players are different for each outcome. There is thus not only a coordination problem but also a problem of reaching agreement. As both players strive for a different equilibrium, there is a risk that they end up without a site for the project, which is the individually and collectively least desirable solution. Lack of coordination of the players at the solution whereby both offer the site is less likely, but possible. The pure coordination problem could be solved by communication. Simple arrangement (coordination by communication) is not sufficient however, as both players aim at the solution whereby the other one offers the site. A bargaining process is necessary to find an agreement as to which of the two Pareto-optimal equilibrium solutions will be chosen. The example presupposes that the project is both individually and collectively beneficial. In case individual costs are higher than individual benefits, the game turns into a harmony game: There is a unique equilibrium in dominant strategies where no site is offered. This is a Pareto-optimal solution, while the outcome of both players offering
82 Table 3.9
T R A N S N AT I O N A L C O M M O N G O O D S
Best-shot Technology: Siting of Locally Unwanted Facilities
Assumption
b.e.0 Strategy Benefits from Combination the Project
Community A
Negative Externalities
Payoff
Ordinal
A: O
B: O
b
e
b2e
2
A: O
B: ~O
b
e
2
A: ~O A: ~O
B: O B: ~O
b 0
0 0
b2e b 0
3 1
All factors are identical for Community B.
Game Matrix
Community B
Offer site
Offer site
Do not offer site
b 2 e, b 2 e 2, 2
b 2 e, b 2, 3
b, b 2 e 3, 2
0, 0 1, 1
Community A Do not offer site
the site would be inferior. The outcomes where only one player contributes are also Pareto-optimal, but they are no equilibria. The three different aggregation technologies presented in this section produce different matrix game structures: prisoners’ dilemma, assurance game, chicken game, and harmony games. As harmony games pose no dilemma between individual and collective rationality, there is no collective action problem to solve. The other three strategic constellations pose different collective action problems. The three game structures are of a universal nature, however. They do not necessarily arise only as a result of a certain aggregation technology; they may also be the result of completely different attributes of situations of common goods provision. Furthermore, situations where a specific aggregation technology applies will have other attributes, which may change the game structures again. The next chapter introduces additional features.
Chapter 4
Case Studies 2: Attributes of the Groups
While in the previous chapter the attributes of the common goods
themselves were examined, in this chapter the focus is on another aspect of the situation: the attributes of the collective. Attributes of the group are, for example, the number of actors concerned, the anonymity or nonanonymity of the members of the collective, their homogeneity or heterogeneity, and the duration or repeatedness of their interactions. There is already much theoretical and empirical research on the number of actors and repeated interaction (sections 2.2 and 2.5). The questions of anonymity (Frey and Bohnet 1996) and heterogeneity (Keohane and Ostrom 1995) have found much less attention in the literature. In this chapter, I concentrate solely on the effects of the heterogeneity of actors on the strategic structure of common goods provision games. The analysis of heterogeneity is especially interesting for political science, as it reveals that there are problems in common goods provision other than inefficiency, namely, problems of power and distribution. Moreover, it is interesting in the context of this book, because heterogeneity is a crucial factor when a common good has to be provided in a multi-level institutional context. Keohane and Ostrom (1995: 7; Snidal 1995: 62−64; Mitchell 1995) distinguish three dimensions of heterogeneity: actors’ capabilities, their preferences, and their information and beliefs. They define “capabilities broadly to refer to actors’ assets, whatever form these assets may take, . . . preferences refer to evaluations of the individual benefits and costs of policies (in view of actors’ expectations of likely resulting outcomes) and of the outcomes themselves.” I will use this distinction as well, but will restrict my analysis to heterogeneous
84
T R A N S N AT I O N A L C O M M O N G O O D S
capabilities and preferences. Information can be interpreted as part of actors’ capabilities, anyway (Libecap 1995: 187). Heterogeneous capabilities may include differences of the actors relating to property rights or other rights that are relevant for common goods production, to physical endowments, or to the size of possible investments into common good production. Heterogeneous preferences may stem from different valuations of the common good, from different benefits derived from the good (for other reasons than pure valuation), and from different costs of contribution. The contributions in the volume of Keohane and Ostrom (1995) are concerned with the question of how different types of heterogeneity affect the possibility of cooperation. Olson (1965) argued that “privileged” actors, that is, those with a preponderant interest in a certain public good, would unilaterally provide it. In international relations, the theory of hegemony (Keohane 1984) implies a similar argument: The existence of a single dominant actor in the international system is to the benefit of all nations, if this actor is willing to provide international public goods. From this argument, it has often been concluded in international relations that extreme heterogeneity enhances the chances for cooperation. Martin (1995) provides another argument that supports this proposition: Heterogeneity opens up the possibility of package deals and compensation. Mitchell (1995) presents the case of oil pollution of the oceans, where heterogeneity at different levels of actors has lead to cooperation. In the common pool resource (CPR) literature, on the other hand, it is generally held that homogeneity promotes cooperation (Ostrom 1990: 89), while heterogeneity impedes it. Libecap (1995) concludes that heterogeneity of different kinds makes it more difficult for the actors to achieve agreements in bargaining. Thus, predicting the effects of heterogeneity on cooperation is obviously a complex task that needs a more differentiated approach. As Snidal puts it: “the impact of heterogeneity is heterogeneous” (1995: 63; Hausken and Plümper 1999). In this chapter, I am not concerned with the effect of heterogeneity on cooperation, but solely with its effect on the strategic constellation of actors. This, in turn, influences the possibility of cooperation, but it is not equivalent to it. In this chapter two examples will be analyzed. The attempts of capital income tax coordination among the European Union (EU) member states serves as an illustration of heterogeneous preferences of the players. In this case, heterogeneity of preferences is not the result of urgency, but rather of different relative valuation of two goals relevant for tax coordination in the EU, namely, tax revenue and
AT T R I B U T E S O F T H E G R O U P S
85
political benefits from sufficient capital supply (section 4.1). However, the heterogeneity of member states is also, in part, a consequence of different endowments—in this case, the relative size of capital markets. The problems of environmental pollution of international river basins will be used as a second illustration of the effects of different endowments. In this case, negative externalities are unidirectional. This can be interpreted as a result of a natural advantage of the upstream country, or of property rights that give an advantage to the upstream country (section 4.2). The case studies in chapter 5 are also based on heterogeneous actors. 4.1
Heterogeneous Preferences: Capital Income Tax Coordination in the European Union
The coordination of capital income tax in the EU has not been very successful for a long time. It took the EU 35 years to achieve a cooperative agreement on coordinated measures of savings taxation. Several earlier attempts of the European Commission and EU member states to harmonize national policies on capital income taxation have failed. Finally, a preliminary agreement was reached in June 2000, which was—after some changes—confirmed by the European Council in June 2003. Although this agreement is surely an achievement in preventing capital movements induced by tax differentials, it still mirrors the conflict between two groups of member states: those who want to prevent tax evasion and capital flight from their countries, and those who have a strong interest in attracting foreign capital. Capital income taxation by several states belonging to a common financial market can be analyzed as a problem of managing a common pool resource. More precisely, harmonization of taxes throughout the market is a common good that exhibits weakest-link characteristics. The good cannot be provided unless all states coordinate their tax policies. Thus, the strategic constellation can be expected to be a coordination game, which should be easy to solve. As this does not correspond to the historical facts, however, there must be additional factors that affect the strategic constellation. It will first be shown that the game turns into a prisoners’ dilemma, if the governments value other benefits associated with a large domestic capital stock higher than they do the revenue from capital income tax. Furthermore, EU member states are heterogeneous regarding their governments’ relative valuations of tax revenue and other economic and political benefits from capital, as well as regarding the size of domestic capital stock. If this is taken into account, harmonization
86
T R A N S N AT I O N A L C O M M O N G O O D S
of capital income tax is, for all states, not the first preference. If a state is small and if its government is strongly interested in its financial sector, but much less so in tax revenue, the government might prefer tax competition to tax coordination. This could explain why it is so difficult to achieve a negotiated agreement in a EU of heterogeneous members. I start with a short overview of the history of capital income tax harmonization in the EU. Next, I briefly refer to theories of tax competition and analyze tax harmonization as a problem of providing a common good. Third, the basic model of tax coordination among homogeneous states is presented. Finally, two forms of heterogeneity are introduced: heterogeneity of utility function of governments, and heterogeneity of states with respect to the size of their capital market. Capital Income Tax Coordination in the EU As early as 1960 the first expert committee was set up by the European Commission to analyze the effects of different capital income taxation policies of the member states on the functioning of the common market. However, only after a French initiative aimed at concerted measures regarding direct taxes, did the Commission publish a first program on the harmonization of direct taxes in 1967. Taxation of private capital income was among the problems addressed in this program (Genschel 2002).1 Capital income (interests and dividends) is basically to be taxed by the country of residence at the marginal rate of personal income tax, irrespective of the country of origin of the capital income. All residents of a country shall be subject to the same taxation whether the income is earned domestically or abroad. However, on the basis of OECD rules, states may also levy so-called withholding taxes on all capital income produced by domestic sources. In fact, most EU member states did so in the 1960s. Withholding tax rates differed, but were in general much lower than personal income tax rates. Whereas residents could deduct domestic withholding taxes from personal income taxes, this was not possible for withholding taxes paid abroad for income from foreign investments. The Commission identified three problems of different taxation policies among the member states. First, double taxation of capital income from foreign investment led to barriers to the free movement of capital. Second, double taxation implied an incentive to invest more capital in countries with no withholding taxes or low withholding tax rates. Thus, different tax rates led to distortion of capital allocation.
AT T R I B U T E S O F T H E G R O U P S
87
Third, double taxation only takes place, if capital income is declared to the tax authorities. As capital investment in foreign countries is difficult to monitor, for private tax payers, tax evasion is easier with foreign than with domestic investments. The lower the foreign withholding tax rate, the greater is the benefit from tax evasion. Therefore, in the case of tax evasion, there is also an incentive to invest more capital in low-tax countries. This again leads to distortion of capital allocation. The Commission discussed two solutions to these problems. The first solution combined the abolition of withholding taxes with the development of a transboundary system of information exchange. In such a system, banks in the countries of investment would inform the fiscal authorities in the countries of residence of capital owners about their earnings from interest or dividends. An information exchange system would be an optimal solution, since it both avoids distortion of capital allocation and prevents tax evasion. Still, the Commission rejected this idea for two reasons. First, this practice would collide with the principle of bank secrecy, which even has constitutional status in some member states. Second, the Commission expected massive capital flight from the common market to the outside world. The Commission’s preferred solution was therefore complete harmonization of withholding taxes combined with full deductibility in the country of residence. This way double taxation as well as distortion of competition for capital would be avoided. However, the incentive for tax evasion remains. The higher the harmonized withholding tax rate (the smaller the difference to personal income tax), the lower is the incentive for tax evasion. On the other hand, the higher the common withholding tax rate, the greater is the incentive to invest capital outside the European market. The member states reacted differently to the proposal (Genschel 2002: 138−145). Conflict centered mainly on the level of the withholding tax rate. France, Belgium, and Italy preferred a higher level than the proposed 10 percent, as they wanted to avoid tax evasion and a decrease of tax revenues owing to liberalization of the capital markets. The Netherlands and Luxembourg rejected the rate as too high, because they were interested in efficient capital markets, and the free movement of capital within and outside the member states. The Council of Finance Ministers did not come to an agreement. Thus, the first attempt at harmonizing capital income taxes ended in failure. Subsequently, the problem lost its relevance. Currency crises and the breakdown of the Bretton Woods system caused the member states to deliberalize international capital transfer (Kapstein 1994).
88
T R A N S N AT I O N A L C O M M O N G O O D S
The aim of a common European capital market had been given up for considerable time. The next initiative to harmonize capital income taxation was launched by the Commission in 1989 within the framework of the internal market program. At that time, double taxation was no longer a problem within the EU, as practically all member states had introduced some kind of deduction of foreign withholding taxes from domestic personal income taxes (Genschel 2002: 142). Furthermore, during the 1980s, liberalization of capital had again become the dominant philosophy in Europe as well as worldwide. The aim of the elimination of all capital controls was no longer limited to the European internal market but applied to capital transfers with the outside world, as well. The Commission focused therefore on the problem of distortion of capital allocation as a result of tax evasion. The proposal was not very ambitious from the beginning (Genschel 2002: 142–143). Again, the idea of an information exchange system was rejected because of the principle of bank secrecy and the assumption of high administrative costs. At the heart of the proposal was a minimum withholding tax on interests of 15 percent, where some exceptions were possible. This proposal had two shortcomings. First, as the tax rate was low, compared with income tax rates, there was still a considerable incentive for tax evasion. Second, the minimum standard still allowed for different tax rates among member states. Consequently, distortion of capital allocation was only slightly reduced, but not avoided. The reaction of the member states to this proposal was again divided (Genschel 2002: 147−148). France and Italy requested a higher tax rate and that dividends also be included; Belgium and Portugal criticized the exception of Eurobonds; Denmark and the Netherlands preferred an information exchange system. On the other side, the United Kingdom and Luxembourg were completely against a common withholding tax, as they feared massive capital flight from the internal market. This time, the position of Germany turned out to be pivotal. Germany had introduced a national withholding tax on interests in January 1989, and then surprisingly decided to abolish it in April 1989. There were several reasons for this. First, there was an immediate response of the capital markets: Massive capital flight took place and as a consequence, interest rates increased, bonds turnover decreased, and the exchange rate for German mark decreased (Genschel and Plümper 1997: 632). Second, as a result of a relatively flourishing economy there were no budgetary problems in Germany at that time. Finally, voters had reacted negatively to the tax and the
AT T R I B U T E S O F T H E G R O U P S
89
new minister of finance, Theo Waigel, used this opportunity to start in office with a popular measure. As a consequence of this change of German policy, the proposal for a harmonized withholding tax rate at the European level failed again. During the 1990s conditions changed again. Germany reintroduced its withholding tax in 1992. This led to massive capital export to Luxembourg and considerably lower tax revenues than expected (Genschel and Plümper 1997: 633). Luxembourg’s neighbors, especially Belgium, France, and Germany, felt disadvantaged by its refusal to introduce any kind of taxation of foreign capital income (Genschel and Plümper 1997: 632). As a result of the economic recession, most EU member states faced serious budgetary problems, which were aggravated by the fact that the governments were required to meet the strong criteria for the European Monetary Union. In addition, international liberalization of capital markets had made the problem of capital flight more severe. In this situation, a memorandum of the European Commission warned that tax evasion, tax revenue losses, and distortion of competition in the European capital markets required a common policy against capital flight. The Council of Ministers agreed and charged the Commission to develop another proposal for a directive on interest taxation. The European Commission’s proposal of 1998 was modest (Genschel 2002). It relied on the harmonization of withholding taxes on transboundary capital income only, while taxes on domestic interests from capital could still be different. Second, the proposal was based on the so-called coexistence model: Member states should be allowed to opt between a harmonized withholding tax and a European information exchange system. In 1993, Belgium and Germany had already suggested the coexistence model. This initiative had failed because of stiff resistance from the United Kingdom and Luxembourg. Third, the proposal included an obligation for the EU to negotiate guarantees with third countries that capital incomes of EU citizens would be taxed at the same level as it would be within the EU. During the negotiations that followed within the Council, these provisions were further watered down. The member states’ positions differed on the tax rate and the division of the withholding tax revenue between countries. Luxembourg, Austria, and the United Kingdom argued that the cooperation of all relevant third countries should be a precondition for the European solution. Great Britain demanded that Eurobonds should be exempted from the tax, as otherwise London’s City would be severely disadvantaged. Negotiations again failed at the 1999 summit of Helsinki.
90
T R A N S N AT I O N A L C O M M O N G O O D S
In June 2000, the agreement of Feira was made possible by a spectacular change of position by the British government (Genschel 2002: 143). Britain introduced a national system of information exchange for banks and fiscal authorities, and now argued at the European level that only such a system could secure sufficient protection against tax evasion and distortion of competition. With the exception of Luxembourg and Austria, all member states now agreed on a European system of information exchange. Austria and Luxembourg declared that they would not sacrifice their bank secrecy policy, because the relevant third countries would also not do so. The compromise found at the Feira summit mirrors this divergence of positions. Only transboundary interest incomes will be taken into account. The ultimate solution will be based on a system of information exchange, while the coexistence model applies for a transitional period of approximately 10 years. In the meantime, negotiations not only with Switzerland, Liechtenstein, Monaco, Andorra, and San Marino, but also with the United States, are to guarantee that these countries introduce similar measures (Genschel 2002: 143). In the event the Commission revised its proposal (European Commission 2001). Now, the final solution was to be based on a system of information exchange, while the coexistence model should apply only for a transitional period of approximately 10 years. In the meantime, negotiations of the Commission with Switzerland, Liechtenstein, Monaco, Andorra, San Marino, and with the United States, were to guarantee that these countries introduce equivalent measures (European Council/Ecofin 2000). From 2011 onward, all countries should be obliged to use an automatic reporting system. For the transitional period, three countries were permitted to apply a minimum withholding tax of 15 percent during the first 3 years and 20 percent during the remaining 4 years on nonresidents interest income. These countries were Luxembourg, Austria, as well as Belgium that now also opted for the withholding tax. At the end of 2001 the Commission started to negotiate with the third states. It reported to the Council on the results of these negotiations in November 2002 (European Commission 2002; Holzinger 2005 for details). On this basis, the EU Council negotiated the proposed directive again in December 2002 and January 2003. The agreement of January 21, 2003 includes the coexistence of the two systems. Twelve member states introduced an automatic system of information exchange in January 2005. Belgium, Luxembourg, and Austria apply a withholding tax. The rate will be 15 percent from 2005, 20 percent from 2008, and 35 percent from 2011 onward.
AT T R I B U T E S O F T H E G R O U P S
91
The three states will change to an automatic reporting system only after Switzerland and the other tax havens agree on systems of information exchange on request. The revised proposal of the directive was finally adopted on June 3, 2003. This agreement mirrors the conflicts and divergence of governments’ interests within and beyond the EU. There are still two groups of countries: those which prefer automatic information exchange, and those which are prepared to accept only minimum withholding taxes. The agreement allows for the permanent coexistence of these two systems, as the general introduction of information exchange within the EU depends on Switzerland’s and the other third states’ willingness to accept OECD rules on information exchange on request, as well as on negotiations with further non-EU countries, which Switzerland wants to have included in the cooperative arrangement. Tax Harmonization as a Common Good The growing relevance of international tax competition has two sources. As Deheija and Genschel (1999: 403) put it: “As the level of taxation reaches 30, 40, or even 50 percent in welfare states, the premium on tax avoidance and tax evasion rises. At the same time, the costs of doing so go down.” The level of economic integration and the liberalization of markets, especially of trade and capital markets, achieved in recent decades, have made it much easier for taxpayers to avoid domestic taxes by shifting the tax base to a foreign country. In the case of capital income taxation, the mobile tax base implies a double danger and a double temptation for national governments. A high-tax government may not only loose some of its tax revenue by capital flight, but also the economic and political benefits associated with a large domestic capital stock and capital market: Capital supply and investment will decline, interest rates increase, the financial sector suffers, and growth and employment are at risk. With low tax rates, however, governments can attract capital from high-tax countries and thus improve both tax revenue and economic figures and political benefits. As a consequence of this strategic situation, tax competition may lead to a downward spiral or “race to the bottom” of tax rates and fiscal revenues (Frey 1990; Sinn 1990, 1997). Suspicions are that this could mean that the EU would turn into a “single large tax haven” (Giovannini and Hines 1991: 172). In economic theories of tax competition, two contrasting views can be found (Keen 1993). In the literature on public finance, tax competition is considered to have negative effects on public welfare.
92
T R A N S N AT I O N A L C O M M O N G O O D S
Tax competition leads to underprovision of common goods, and tax harmonization in the EU would therefore increase overall welfare (Bucovetsky and Wilson 1991). The public choice literature, on the other side, emphasizes the fact that governments are not benevolent actors and that their interests do not coincide with that of taxpayers, which gives rise to “political distortions” (Frey and Eichenberger 1996) and waste of tax money. Tax coordination is considered to be a cartel of governments at the cost of the taxpayer, while tax competition is a welfare increasing check on governments. On this view, constitutional rules that forbid tax coordination are desirable (Brennan and Buchanan 1980). Finally, there are economic models that use an objective function for governments that includes two variables: the welfare of the citizens and a “waste” variable that accounts for imperfections in the political process (Edwards and Keen 1996; Myles 2000). Not surprisingly, in these models, the effects of tax coordination on welfare are ambiguous. In this chapter no such normative questions are posed. I am not concerned with welfare or with the taxpayer’s benefit, although the game between the fiscal authorities and the taxpayer is interesting in itself. I will come back to this in chapter 9. Here I am interested in the rational reconstruction of cooperation among governments who compete for a mobile tax base. Is it possible to explain the European difficulties to cooperate by modeling the strategic constellation? Problems of tax coordination have been characterized as a prisoner’s dilemma (Hallerberg 1996), as coordination game (Radaelli 1998), or as constant-sum game (Deheija and Genschel 1999). Modeling strategic constellations as matrix games is a tricky task, however, as has been demonstrated in the earlier chapters. Matrix games are very sensitive to changes in assumptions. The type of game depends on the exact circumstances that determine the preferences of the governments and thus their payoffs for each possible strategy combination. Tax competition games may, for example, differ with respect to the taxes concerned. It makes a difference whether transfer-pricing rules for multinational companies (Radaelli 1998) or capital income taxation (Deheija and Genschel 1999) is analyzed. The problem of collecting taxes on capital income by several governments in a common market where capital is perfectly mobile can be seen as a common pool resource problem (see also Koelliker 2001). Consumption is rival, as the tax base can only be taxed up to 100 percent, and exclusion is difficult, as the capital is mobile. The resource is the tax base, that is, the invested capital. The states exploit the resource, tax revenue is their payoff, and the tax rate is equivalent
AT T R I B U T E S O F T H E G R O U P S
93
to the rate of extraction. If there were only one user of the resource (the complete capital stock) the problem would reduce to an optimization problem comparable to those in cases of renewable resources (Perman, Ma, and McGilvray 1998; Young 1992). If there are more users of a common resource, a strategic dimension is added and thus a collective action problem may arise. Capital income taxation is distinct from the problem of common use of the village green where all inhabitants can graze their sheep, in that the users are confined to their territories while the resource is mobile. It can best be compared to fishing problems, where the fishermen are assigned to a certain territory, whereas the fish is mobile. If there is too much harvesting pressure in one territory, the resource moves to another. The fish, so to speak, is attracted by low taxation and leaves the country in the case of high taxation. At least this is true if there is perfect and costless capital mobility and if taxpayers are perfectly rational. Taxpayers can move their capital to another territory where the tax rate is lower or where there is no tax at all. As a consequence of this, the production function in the tax case is given by the Laffer curve (Deheija and Genschel 1999: 408). A raise of the tax rate directly increases tax revenue (tax rate effect), but indirectly reduces tax revenue, as the tax base moves out (tax base effect). The possibility of out-migration reveals another property of the capital income tax case: As long as there is only one single country with lower or no taxes, the capital will move there. Coordinated capital income taxation is a weakest-link common good. The common good is the ability to effectively collect tax from capital owners in the common capital market. The governments will only achieve this goal to the extent that they coordinate their taxes and tax rates. If there is a weakest link, that is, a country with no or low tax rates, capital will move into this country. Thus, in the case of capital income taxation, the strategic constellation can be expected to be a coordination game, or, more specifically, an assurance game. Tax Competition: Coordination Game or Dilemma? In this section a model of tax coordination will be introduced, where two governments compete for a mobile tax base. They have the choice to levy a capital income withholding tax (T) or not to do so (~T). Countries A and B are identical regarding size of capital stock and to the preferences of governments. Capital is perfectly mobile between
94
T R A N S N AT I O N A L C O M M O N G O O D S
the two countries, capital owners are perfectly rational, and there are no transaction costs. This implies that there is complete capital flight from country A to country B, if A levies the tax, while B does not. The payoff function for the government includes two elements. The first element is the revenue, r, from the tax, which is determined by the tax rate and the tax base. The second element is political benefits, b, which are associated with the domestic capital market: A flourishing banking and finance sector, sufficient capital supply, low interest rates, investment, innovation, growth, and employment are (partly) a consequence of a large and sound capital market. Since the state of the economy plays a great part in voters’ evaluations of a government’s performance, these benefits are not only economic but also political. For the model, it is assumed that both elements of the payoff function are nonnegative (r $ 0, b $ 0). For the tax revenue, this is evident. The political benefits can be negative in reality; however, the assumption made here is just a simplification that does not affect the result. There is no problem in setting a lower limit to the payoff function, as it is the relative size of payoffs that determines the strategic constellation. A political loss is equivalent to no political benefit in the model. A second assumption concerns the relative weight of r and b in the payoff function. Initially, it is assumed that r is greater than b (r . b). This assumption has to be justified. The revenue from the withholding tax for each outcome can, in principle, be measured and calculated in advance, even if empirically this might prove difficult. The political benefits of capital stock for the governments are not easy to estimate, however. Even if it is possible to estimate the tax base effect of a certain tax rate, its effects on interests, investment, growth, and employment are difficult to measure, as there are many other factors that influence these figures. Predictions are even more problematic. Finally, the political benefits for governments are not necessarily identical with the economic benefits, as it is no simple task to determine how the economic effects turn into public and voter support. Still, it can be assumed that there is some positive correlation of a well-functioning capital market, its positive economic effects, and political benefits. From this discussion it should have become evident, that there is also a subjective factor in the values of the political benefits. The same applies to the tax revenues. The payoffs depend on the governments’ relative valuations of r and b. To be precise we should talk about the values of r and b for the government and denote u(r) and u(b). Governments will not, in general, value the tax revenue more than
95
AT T R I B U T E S O F T H E G R O U P S
the political benefits or vice versa. This will depend on the actual circumstances. For example, if the overall budgetary situation is positive, a government may be much more interested in the political benefits of a large capital market, whereas in a situation of high budget deficit, the revenue may become more important. A government that is susceptible to the interests of the financial sector might prefer to keep the domestic capital stock as large as possible and sacrifice tax revenue. This could be a consequence of the hard lobbying by the financial sector or of the fact that it is a relatively big sector within this country’s economy. Thus, sometimes a government may be revenueoriented (u(r) . u(b)), sometimes political benefits-oriented (u(b) . u(r)). To keep the notation simple, this can be approximated by assuming r . b or b . r. If we could measure r and b, and if the governments’ utilities were a linear transformation of r and b, it would be equivalent anyway. The first model assumes revenue-oriented governments. Table 4.1 gives the payoffs for the two identical countries A and B for each strategy combination. The benefit from tax revenue can only be Table 4.1
Capital Income Tax Coordination with Revenue-oriented Governments
Assumption
r.b$0 Strategy Combination
Government A
Tax Revenue
Political Benefits
A: T
B: T
r
b
A: T A: ~T A: ~T
B: ~T B: T B: ~T
0 0 0
0 2b b
Payoff
Ordinal
r1b 0 2b b
4
All factors are identical for government B.
Game Matrix
Government B
Tax
Tax
No tax
r 1 b, r 1 b 4, 4
0, 2b 1, 3
2b, 0 3, 1
b, b 2, 2
Government A No tax
1 3 2
96
T R A N S N AT I O N A L C O M M O N G O O D S
realized when both governments levy the tax. In case a government itself levies no tax, there is no tax revenue; in case the other government applies no tax the capital moves out. There are political benefits, b, from domestic capital, when both governments have a tax, or when both do not. A government that levies no tax gains the political benefit of both domestic and foreign capital (2b), as foreign capital migrates in. A government that introduces the tax has no political benefits at all, because the domestic capital moves out. The first preference of the governments is thus that both have a tax, the second that only the other government has a tax, the third that both have no tax, and the least preferred situation is that they themselves levy the tax, while the other government does not. The game in table 4.1 belongs to the class of coordination games. More precisely, it is an assurance game, which has both a Paretooptimal and a suboptimal Nash equilibrium. This confirms the expectation for weakest-link goods. Only if there is no weakest link, the common good (tax revenue for both states) will be provided. If there is a weakest link, one single tax haven in this case, the good will not be achieved. There is no tax revenue, although one government enjoys double political benefits. In the case of two tax havens, the good will not be provided, but there are equally distributed political benefits: There is no capital flight, but no tax revenue, either. It should be easy to find a solution to this collective action problem. Communication of the two governments should be sufficient. They should agree on the strategies to levy the withholding tax, as this is the optimal equilibrium. Thus, tax coordination should be easy to achieve. This does not at all correspond to the above observations of attempts to harmonize tax policies in the EU. This is due to the fact that the assumptions of the model are not yet sufficiently realistic. In the following model, benefits-oriented governments are assumed. This is captured by the assumption that benefits are greater than revenue (b . r). The payoffs in table 4.2 are constructed as in table 4.1, however, their value and thus the preference order changes. The game is now a prisoners’ dilemma. The Nash equilibrium implies tax competition and no tax at all, or a “race to the bottom” of tax rates. Even if the players agree on a contract to play Pareto-optimal outcome strategies (T, T), they both have an incentive to defect afterward. The suboptimality of this strategic constellation should be obvious to the EU member states. For a relatively small number of states, which are in a permanent relationship in their capacity as EU members, it should thus not be too difficult to find a negotiated agreement. The EU institutions have both the authority to decide on a cooperative
97
AT T R I B U T E S O F T H E G R O U P S
Table 4.2
Capital Income Tax Competition with Benefits-oriented Governments
Assumption
b.r$0 Strategy Combination
Government A
Tax Revenue
Political Benefits
Payoff
Ordinal
r1b 0 2b b
3
A: T
B: T
r
b
A: T A: ~T A: ~T
B: ~T B: T B: ~T
0 0 0
0 2b b
1 4 2
All factors are identical for government B.
Game Matrix
Government B
Tax
Tax
No tax
r 1 b, r 1 b 3, 3
0, 2b 1, 4
2b, 0 4, 1
b, b 2, 2
Government A No tax
outcome, as well as the power to secure compliance. Even if there are often no real sanctions for deviation, there are verdicts of the European Court of Justice. Furthermore, reactions and retaliation by other member states must be expected. Thus member states should agree on the coordinated solution, since this would guarantee all states a higher payoff than the tax competition equilibrium. Why was there not even a negotiated agreement for such a long time? One possible explanation is that not all member states consider tax coordination to be individually better than tax competition. The background for this is the heterogeneity of preferences among EU member states. Tax Competition of Heterogeneous Member States Not all European states are in the same situation regarding political benefits and tax revenue. Governments may have heterogeneous preferences for many reasons. Their financial sectors and capital markets are different in size and development; their economies may do well or may suffer from a recession; budget deficits may be more or less severe;
98
T R A N S N AT I O N A L C O M M O N G O O D S
capital income may account for more or less of overall personal income, and thus capital income tax would have a correspondingly larger or smaller share of the states revenue; taxpayers may be more or less prone to tax evasion. This variance in circumstantial factors will lead to different relative valuations of tax revenue and political benefits. Moreover, governments differ in their ideology: While conservative governments usually favor economic liberalization and free movement of capital, social-democratic governments emphasize state activity and the public provision of goods, and thus prefer to prevent tax evasion and to secure national tax revenues. Heterogeneity shall be introduced into the model as a next step. For the model it is sufficient to distinguish two types of governments: those who are revenue-oriented (r . b) and those who are benefits-oriented (b . r). As long as we are only concerned with the effect of heterogeneity on the strategic constellation, it is not necessary to know why the governments have these preferences. The payoffs are as in tables 4.1 and 4.2. Let government A be revenue-oriented and government B be benefits-oriented. The game matrix is given in table 4.3. This game is an asymmetric dilemma or unilateral defection game. There is a unique and suboptimal Nash equilibrium as in the prisoners’ dilemma. The game is different from the prisoners’ dilemma, because only the benefits-oriented government, B, has a dominant strategy not to tax; the revenue-oriented government prefers to levy the tax, if B does as well, but prefers not to tax, if B does not. This game combines a defection problem with a distributional problem. If the governments agree to tax coordination (T, T), only government B has an incentive to defect afterward. However, there is a distributional Table 4.3 Capital Income Tax Competition with Heterogeneous Governments Assumptions
Government A: r . b Government B: b . r
Game Matrix
Government B
Tax
Tax
No tax
r 1 b, r 1 b 4, 3
0, 2b 1, 4
2b, 0 3, 1
b, b 2, 2
Government A No tax
AT T R I B U T E S O F T H E G R O U P S
99
problem, as the Pareto-optimal tax coordination outcome represents the first preference for government A, but only the second preference for government B. The latter would prefer to have no tax in any case, even if government A applies a tax as well. This characteristic may account for the difficulty to agree on the coordinated solution and may thus explain the repeated failure of tax coordination in the EU. In the symmetric prisoners’ dilemma, both players have an incentive to defect if an agreement is reached, but both would prefer a common tax to the Nash equilibrium of no tax. In the asymmetric dilemma, it would be easier to secure compliance after an agreement than in the prisoner’s dilemma (as only one government has an incentive to defect), but it is more difficult to find an agreement in the first place. This model is based on the subjective preferences of governments, on their heterogeneous valuations of r and b, which implies a methodological problem. On the one hand, governments’ preferences are what counts in an explanation of strategic interaction of governments in the EU. On the other hand, the subjectivity of preferences makes their use as explanatory factors problematic. Preferences cannot be observed directly, although the reasons governments give as justifications for their positions and their behavior in negotiations can serve as a hint. There is no guarantee, however, that the reasons given are the “true” reasons. Therefore, a model is developed in the next section that relies on a single objective factor, namely, the size of a country’s capital stock. Countries are assumed to be heterogeneous relating to this factor. Heterogeneity, in this case, is not heterogeneity of preferences but of capabilities. Both countries have the same preference for tax revenue, but their endowments, the size of their capital stock, are different. 4.2
Heterogeneous Capabilities: Tax Coordination and International River Pollution
In this section, two illustrations for the effects of heterogeneous capabilities on the strategic constellation are given. First, the different size of countries’ capital markets serves as an example of different “natural” endowments of states. It provides another rationale for the European failures to agree on tax coordination so far. The contribution of both models of tax competition to the explanation of governments’ positions during EU negotiations will be examined at the end of this case study. A second example of different “natural” endowments is the pollution of international river basins, where upstream and downstream riparian countries are in different positions. This
100
T R A N S N AT I O N A L C O M M O N G O O D S
illustration can also be interpreted as an example of different assignment of property rights. Tax Competition among Small and Large EU Member States An approach to the European failures to agree on tax coordination similar to the one outlined above is its explanation by the fact that member states differ in size (Kanbur and Keen 1993; Deheija and Genschel 1999). According to these theories, tax revenue is the only element in the governments’ payoff function. Countries are heterogeneous relating to the size of their capital stock, or, in other words, the size of the tax base. Deheija and Genschel (1999) have put forward the argument convincingly, that it pays more for small-tax-base countries to prey on their neighbors’ capital by applying lower taxes, than it does for large-tax-base countries. The reason for this is the fact that, in an open economy, tax revenue is not linear in the tax rate. Capital income tax revenue follows the Laffer curve: It increases with the tax rate for some time but, with very high tax rates, the revenue strongly decreases, because the tax base moves to another country with lower tax rates. It pays for a country to prey on the other countries’ capital as long as the tax-base effect from the attracted foreign capital (increase of revenue) is greater than the tax-rate effect (decrease of revenue) from domestic capital. The smaller the domestic tax base and the more capital can be attracted—that is, the larger the foreign capital base—and the higher the tax differential, the more it pays to follow a low-tax strategy. Therefore, small countries are in a much better position in tax competition than large ones. Large countries, however, can try to limit the small countries prey by keeping the tax differential small, which may lead to a downward spiral of tax rates. The strategic model of Deheija and Genschel (1999: 410) shows, however, that the equilibrium in this “race to the bottom” is above zero. This is valid, as long as only tax revenues are considered as elements of the governments’ payoff functions. The capital stock size argument shall now be translated into the matrix game language used in this book. Country A has a large capital stock or tax base (cA), while country B has only a small one (cB), thus cA . cB . Both governments have two strategies: They can apply a high tax rate (tH), or a low tax rate (tL), with 0 , tL , tH , 1. Capital is again perfectly mobile and there are no transaction costs. This implies complete capital flight from the high-tax country to the low-tax country, if there is no tax coordination. We further assume that for the small
101
AT T R I B U T E S O F T H E G R O U P S
Table 4.4
Capital Income Tax Competition of Small and Large Member States Strategy Payoff Payoff Combination Government Government A B
Governments A and B
Ordinal Ordinal Government Government A B
A: t H B: t H
c At H
c Bt H
4
A: t H B: t L
0
(cB 1 c A)t L 0
1
4
3
1
c Bt L
2
2
A: t L B: t H (c A 1 cB)t L A: t L B: t L c At L
3
Game Matrix
Government B High tax
Low tax
High tax
c A t H , c Bt H 4, 3
0, (cB 1 c A)t L 1, 4
Low tax
(c A 1 cB)t L , 0 3, 1
c A t L , c Bt L 2, 2
Government A
country, B, the positive foreign tax-base effect (cAtL) of low taxes is greater than the negative domestic tax-rate effect (cBtH 2 cBtL); and that for the large country, A, the negative domestic tax-rate effect (cAt H 2 cA t L) of low taxes is greater than the positive foreign tax-base effect (cBtL). Therefore, government B prefers low taxes, while government A prefers high taxes in case of tax competition. Given these conditions, government B may benefit from undercutting A’s tax rate. Table 4.4 gives the payoffs for each strategy combination, as well as the preference order for both countries. If both governments choose high tax rates, they earn the high-rate revenue from their domestic tax base. If one government chooses the high tax rate and the other the low one, the high-tax government looses its tax base completely and has tax revenue of zero; in this case, the low-tax government gains the low-rate revenue of its domestic and the foreign capital stock. If both governments decide in favor of the low rate, both get the low-rate revenue from their domestic tax base. The preference orderings of the outcomes given in the last column are different for the small and the large country owing to the assumptions about the size of capital stock.
102
T R A N S N AT I O N A L C O M M O N G O O D S
The game has the same strategic structure as the game in table 4.3. It is an asymmetric dilemma, combining the defection and the distributional problem. For government B the low tax rate is a dominant strategy. Its first preference is the outcome where country A chooses the high tax while B applies the low tax. If there were a negotiated agreement on the outcome, where both countries levy the high tax, government B could be expected to defect afterward. Thus the low tax rate is the equilibrium. In fact, government B may not loose much, if the tax competition equilibrium is played instead of the tax coordination Pareto-optimal outcome. The smaller the tax differential, the closer government B will be to indifference between the coordinated (tH, tH) and the competition (tL , tL) outcome. Therefore, it will not be easy to convince government B that coordination of taxes is to the benefit of both governments. Although the combined tax revenues would be much higher under tax coordination, the gain is distributed unevenly. The small country does not have much to win by coordination. If this result is valid for the tax revenue component of the governments’ utility function alone, it will be reinforced, if political benefits from capital stock increases are taken into account. The effects of tax competition will shrink, however, if the assumptions of perfect capital mobility and perfect rationality of capital owners are given up. It is reasonable to assume that capital owners consider transaction costs, and the risks of transboundary capital investment are considered higher than those of domestic investment. Moreover, a certain percentage of the taxpayers may be honest and may not seek to evade taxes. This does not change the general results of the model, but it does diminish the effects. Environmental Pollution of International River Basins In section 3.2, the environmental pollution of an international lake was analyzed as a commons dilemma. In that example, the riparian countries used the lake as a wastewater receptor as well as for drinking water extraction. The individual net benefits of doing so were greater than the costs caused by the externalities, at least in the first and second phase of exploitation. As the externalities were shared among all countries, only in phase three did the benefits become smaller than the externality costs. This model presupposed homogeneity of the respective states in their capacities to cause and receive externalities.
AT T R I B U T E S O F T H E G R O U P S
103
In this section, the environmental pollution of an international river is examined. As with the international lake, the common good concerned is a CPR. This model differs from the model of lake pollution only with respect to one property: The riparian countries are now heterogeneous in their capacities to cause and receive externalities. While, in the case of a lake, the externalities are bidirectional; in the case of a river, they are unidirectional. On a river, there are upstream countries and downstream countries that are in a different position. The upstream countries are not affected by the actions of the downstream countries, while the downstream users are affected by the decisions of the upstream users. Thus, an upstream country does not play a strategic game; it will decide just on the basis of its own benefits and costs irrespective of the downstream player’s actions, which do not affect its payoffs. An upstream country has a “natural” advantage over a downstream country, because the upstream country cannot be reached by the externalities of actions that happen downstream. Rivers are the classic example for unidirectional externalities. They were used as an illustration in Coase’s (1960) seminal contribution to the internalization of externalities. Rivers, however, are not the only example. International groundwater streams are another instance. Long-range air pollution is also unidirectional, if there is a prevailing direction of wind. Even with lakes there may be unidirectional externalities, if there is a current in one direction, for instance, as there is in the Great Lakes in North America. Unidirectionality is a cause of heterogeneity. It is not the only cause, however. In the case of global warming, the externalities of emissions into the atmosphere are bi-, or more precisely, multidirectional. Each country emitting substances with greenhouse warming potential also suffers from the emissions of other countries. That the countries are basically homogeneous in their capacities to harm each other does not imply that the countries are fully homogeneous. They may still differ with respect to their behavior or their preferences. Some countries contribute much more to global warming than others; some countries will suffer much more from the effects of global warming than others; costs and benefits of a reduction of harmful emissions may be different as a result of several factors, among them the subjective valuations of the actors. Thus, there are many sources of heterogeneity, and unidirectional externalities is but one of them. It is a very clear instance of heterogeneous capabilities, however. Problems of international river basins have increased over the last century owing to growing populations and growing negative externalities of ever more kinds of water use (Al Baz, Hartje, and
104
T R A N S N AT I O N A L C O M M O N G O O D S
Scheumann 2002). These problems were first felt in the earlier industrialized regions of Europe and North America, but have now spread to the developing world. According to UN data, almost 250 river basins are shared by 2 or more countries. These river basins account for approximately 50 percent of the land area of the earth (Durth 1996: 5; Rogers 1997: 35). Although in most cases only two nations are concerned, there are many large international rivers, which have multiple riparians such as the Amazon (with 9), the Congo (7), the Niger (9), the Nile (11), the Rhine (7), or the Zambezi (9). With 15 riparian countries, the Danube is the most “multinational” river of them all. Whereas conflicts over international river basins in the industrialized countries are mostly caused by pollution, shortage of water is typically the cause of conflict in the Middle East and Asia (Rogers 1997: 39−40). The number of bi and multilateral treaties on the use of shared rivers corresponds to the number of conflicts. Between 1858 and 1992, altogether 127 treaties were concluded in Europe and North America (Durth 1996: 154). Despite the obvious need for a set of rules, there was no binding international legal framework to govern the use of international rivers at the international level, for a long time. There are, however, a number of conventional rules that have been worked out by the UN’s International Law Commission (the ILC draft of 1994), the International Law Association (the “Helsinki Rules” of 1966), and the World Bank (the “Operational Directive 7.50” of 1990) (Rogers 1997). In 1997, the UN General Assembly adopted the “Convention on the Protection and Use of Transboundary Watercourses and International Lakes,” in short the International Water Convention (IWC). However, by 2006, only 36 countries were parties to the treaty. What are the causes of externalities among riparian states? Basically, there are three kinds of functions of rivers. First, there are instream functions such as navigation and transportation, hydropower, flood storage, fishing, and bathing. Second, there are extraction functions such as irrigation, the extraction of drinking water, or cooling water for industrial purposes. Third, rivers are receptors for cooling water, and industrial and domestic wastewater. Most of these uses cause positive or negative externalities to downstream users. Some of the positive externalities are that hydropower may help to regulate the river, and that flood storage provides downstream flood protection. The negative externalities are more pervasive, however. Irrigation, municipal and industrial diversions, and recreation storage, all remove water from the system. Filling wetlands and urban development may increase flooding.
AT T R I B U T E S O F T H E G R O U P S
105
Wastewater treatment, agriculture, forestry, animal husbandry, and urban development add pollution and sediments to the riverbed, its surface, and groundwater (Durth 1996: 25; Rogers 1997: 41). For simplicity, this model will be restricted to negative externalities and pollution, although in principle it applies to other externalities, as well. There are two countries sharing a river, one upstream and one downstream. Both have two strategies available: to pollute the river (P) as a side effect of some benefit-yielding activity or not to pollute it (~P). The activity yields benefit, b, and produces a negative externality, e, with 0 , e , b, as in the model of lake pollution earlier. The upstream country, U, enjoys the benefit whenever it chooses to pollute, irrespective of what the downstream country, D, does; it does not enjoy the benefit if it chooses not to pollute. In any case, it does not suffer from externalities. The downstream country, D, also enjoys the benefits, if it pollutes, and has no benefits, if it does not pollute. D suffers from the externalities caused by U, however, in the case that U decides to pollute. The preference order of U contains only two elements (2 and 1), the first preference being to pollute and the last preference being not to pollute. This is a consequence of the fact that U’s decision is nonstrategic. The preference order for D has the usual four elements. Its first preference is to pollute, while U does not; the second and third preferences are that both pollute or, respectively, do not pollute; D’s last preference is that it does not pollute while U does. The cardinal and ordinal payoffs are given in table 4.5. The game in table 4.5 is a weak harmony or rambo game. It has a unique Nash equilibrium in dominant strategies at the outcome where both countries choose to pollute. The equilibrium is Paretooptimal, thus, the game does not pose a problem of suboptimality or inefficiency. I call it weak harmony, as it is different from a harmony game with respect to only one feature. Unlike a pure harmony game, it has a second Pareto-optimal outcome. Under the perspectives of welfare and coordination, this game does not pose any problem (Taylor 1987: 40). There are only reservations from a distributional perspective. While one country (U) is able to push its most preferred outcome through, the other (D) obtains only its second preference. This is why this type of incentive structure has been called a rambo game (Zürn 1992). One of the players has an advantage that helps him or her to achieve a (relatively) better outcome.2 The game in table 4.5 rests on the assumption that the externality of one unit of pollution is less than the individual benefit. If we assume the externality to be more severe, such that it is greater than b, this does not change anything for country U, as the externality does not
106 Table 4.5
T R A N S N AT I O N A L C O M M O N G O O D S
Environmental Pollution of International River Basins
Assumption
0,e,b Strategy Combination
Country U
Country D
U: P U: P U: ~P U: ~P
D: P D: ~P D: P D: ~P
Benefit
Costs
Payoff
Ordinal
b b 0 0
0 0 0 0
b b 0 0
2 2 1 1
b2e b
D: P
U: P
b
e
D: P D: ~P
U: ~P U: P
b 0
0 e
D: ~P
U: ~P
0
0
3 4 1
2e 0
2
Game Matrix
Country D
Pollute
Pollute
Do not pollute
b, b 2 e 2, 3
b, 2 e 2, 1
0, b 1, 4
0, 0 1, 2
Country U Do not pollute
appear in its payoff function. It does change the preference order of country D, however. Under this condition, D would prefer that both countries do not pollute to the situation where both pollute. This does not change the structure of the game. It is still a rambo game with the only difference being that D now obtains only its third preference in equilibrium. The distributional inequality in equilibrium becomes more severe. Irrespective of the size of the externality, both the upstream and the downstream countries have a dominant strategy to pollute. Although this result may be undesirable from an ecological point of view or from the viewpoint of a wider collective that suffers from the negative externalities of a polluted river, for the users of the river, this is Pareto-optimal; they are not in a dilemma. Nevertheless, the downstream country has an incentive not to accept the noncooperative equilibrium solution. It will try to negotiate with the upstream country, as its first preference is a solution
AT T R I B U T E S O F T H E G R O U P S
107
where U does not pollute, while D does. D can ask U for compensation of the externality caused by U, or D can offer compensation to U, if U gives up its polluting activity. Who will compensate whom in a cooperative solution depends on the distribution of property rights as the Coase theorem (1960) tells us. In a world where the right to use the river of the upstream country includes the right to pollute, the upstream country has no reasons to compensate the downstream country for the externalities. Therefore, Godwin and Shepard, who present a similar model of water quality in rivers, conclude: “The downstream user will continue to pollute and will bribe the upstream user to adopt a ‘treat’ strategy” (1979: 272). However, in a world, where the downstream user also has some right, say, the right to use clean water, or where the upstream user’s right to use the river is somehow limited, the upstream user may have to compensate the downstream user or refrain from the polluting activity. Both allocations of rights might be judged unfair by the disadvantaged actor (Benz, Scharpf, and Zintl 1992: 75–84). Whether the downstream country is able to compensate the upstream user depends on the relative size of payoffs. In the earlier example, D’s gain from a change to the (~P, P) outcome is not sufficient to compensate U. U would demand a minimum amount of b, while D would gain an amount of e. As e has been assumed to be less than b, D cannot compensate U within the game. However, in the case where b is less than e, country D would be able to do so (as long as b is the same for both players, which has been assumed hitherto). Compensation seems in fact to play a great role in real world solutions of upstream-downstream conflicts. Very often the treaties include some kind of package deals (Durth 1996: 158−160), equivalent to compensation in kind. This presupposes, however, that there is some kind of physical effect that works in the reverse direction. There are four major legal doctrines about sharing water in international river basins (Rogers 1997). The first is that a state has “absolute sovereignty” over the water within its territory. This implies that riparian countries to a body of water do not have any right to limit other countries’ uses of the water. This doctrine turns the “natural advantage” into a right. Downstream users are in a bad position under this rule: They have to compensate the upstream countries, if they want them to refrain from pollution. A second doctrine is that “the river belongs to its riparians.” This gives equal rights to upstream and downstream countries. A third legal theory focuses on the “optimum development of the river basin,” thus it takes the position of the collective welfare, instead of allocating rights to individual users. The
108
T R A N S N AT I O N A L C O M M O N G O O D S
fourth doctrine is based on “reasonable share or equitable use.” It respects the rights of the sovereigns of a territory, but it limits the use of the water to ensure reasonable or equitable shares for the other riparian countries. The international conventional rules mentioned earlier as well as the IWC are based on the fourth approach and work out rules, which aim at clarifying what “reasonable” and “equitable” might mean in a concrete conflict. However, as the conventional rules are not binding and the wording in the IWC is very ambiguous (Schroeder-Wildberg 2002), any country may take any stance, and conflicts are mostly resolved through negotiated agreements. In fairly integrated regions like the EU, where there are supranational institutions, laws and courts may play a role in assigning property rights. Durth (1996) gives an instructive example. Chlorides cause a major pollution problem in the Rhine. They have a negative influence on the quality of water for drinking and irrigation purposes. One of the major emittants of chlorides was the French “Mines de Potasse d’Alsace.” The Netherlands, as the downstream country, suffered from too high concentrations of chlorides and started to protest against these emissions in 1955. In 1976, the chloride agreement of the Rhine shoreline states was concluded. This could not be implemented, however, as it was not ratified by France for several years. In 1987, the agreement came finally into force after the riparian states had agreed to share the estimated costs of treatment, 600 million French Francs. In 1988, the Netherlands suddenly refused to further finance its part, as provided for by the chloride agreement. What had happened? In 1974, three horticultural firms in the Netherlands had instituted legal proceedings against the Mines de Potasse, as they suffered high losses from a too high concentration of chlorides. The Private Law Court of Rotterdam did not feel that it had jurisdiction over this international case, and so it passed it on to the European Court of Justice. The European Court ruled that both a court in France and a court in the Netherlands did indeed have jurisdiction over the case and passed it back to the court in Rotterdam. The Rotterdam court finally ruled that the Mines de Potasse had to pay compensation to the horticultural firms for damage caused by chloride emission. In 1988, the Court of Den Haag confirmed this ruling, which was the background for the Netherlands’ refusal to further compensate France for the negative externalities caused by its Mines de Potasse. The European Court of Justice, as well as the courts in the Netherlands had, in this way, limited the property rights of France, or more precisely, of the French Mines de Potasse. This redefinition of property
AT T R I B U T E S O F T H E G R O U P S
109
rights enabled the Netherlands to refuse to participate in the compensation of the French mining companies. A new chloride agreement was signed in 1989. The international action of private firms and the ruling of a supranational court made this possible. The fact that package deals, compensation, and property rights play a role empirically shows the appropriateness of the model developed earlier. The problem of common use of an international river basin is not a problem of welfare but of distributional justice and fairness. This is properly represented by the rambo game. Apart from the distributional problem, the main result of the model is that both upstream and downstream users of a river will pollute as long as no cooperative solution can be found. Empirically, this can be assumed to be true. The sheer number of treaties and agreements on international river basins shows that there is an incentive for the downstream countries to try and find a cooperative agreement.
This page intentionally left blank
Chapter 5
Case Studies 3: Attributes of the Social and Institutional Context
I
n this chapter, the effects of different social and institutional contexts on the strategic constellation of common goods provision are analyzed. The number of properties of the social context and of institutional rules valid in a certain situation is potentially infinite. Which social characteristics and which institutions apply, depends on the particular good and its context. A very instructive example in this respect is the analysis of the international seabed authority by Bräuninger and König (2000). As not every characteristic of the institutional environment heavily influences the problem at hand, for a model of the situation, such characteristics are chosen whose presence or absence effectively changes the game. Two illustrations are given here. First, the analysis of credit ratings is continued in section 5.1. In chapter 3, the question remained open why this information service, which primarily serves the investors, is now usually commissioned and paid for by the borrowers. The second illustration is about regulatory competition on environmental standards. The theoretical prediction has long been that, in the case of transboundary environmental goods, regulatory competition of states may lead to a “race to the bottom” of the environmental standards. Empirically, however, a “race to the bottom” has rarely been observed, while there are in fact examples of “races to the top” of the regulation. In section 5.2, the effects of different trade regimes on the strategic constellation of states in environmental regulatory competition are examined.
112
T R A N S N AT I O N A L C O M M O N G O O D S
5.1 Standards and Regulations in the Market for a Common Good: Credit Ratings The analysis of credit ratings in sections 3.1 and 3.2 did not answer the question why credit ratings are now provided by the issuers of securities instead of by the investors as before. What incentives do borrowers have to finance their own ratings? If they benefit from ratings, why didn’t they provide them from the beginning? In the following, I no longer concentrate on the perspective of the investors, but rather on that of the borrowers. Three different market and institutional environments are compared. First, I consider a world where no credit ratings exist. This refers to the situation in the United States before 1900, as well as to a situation that is still present in countries in which the capital markets are not developed or are based mostly on domestic lenders and borrowers. There are also countries with welldeveloped capital markets, which do not rely much on ratings, such as Germany.1 Second, a situation is modeled, where ratings are provided and taken into account by the investors voluntarily. This might be a picture of the phase between 1900 and 1930 in the United States or of the international capital markets before 1980 (Sinclair 2002). In addition, I assume that capital is scarce and that investors prefer borrowers who possess a rating. Finally, a situation will be considered, where credit ratings have become instruments in state regulation of capital markets. Institutional investors are obliged to take ratings into account and may even be prohibited from investing money in securities that do not have a rating. This correlates to the situation in the United States after 1930, as well as to international capital markets after their liberalization during the 1980s (Kerwer 2000). Capital Markets without Credit Ratings and with Sufficient Capital Supply For the following model, it is assumed that it is too expensive for the investors themselves to collect the necessary information about the creditworthiness of borrowers. Moreover, there is no collective effort to provide rating and no rating agencies exist. Would there be an incentive for borrowers to introduce ratings within a world where no ratings have existed as yet? Before the model is presented, the consequences of the introduction of systematic and comparable ratings into the capital market for investors and borrowers will be discussed. Basically, comprehensive and comparable credit ratings have two functions. First, they reduce uncertainty for the investors. In a world
SOCIAL AND INSTITUTIONAL CONTEXT
113
without ratings, investors face high uncertainty about the creditworthiness of potential borrowers. The information about the credit standing of a certain issuer of a security is to a high degree his or her private information. As issuers may have an incentive to lie about their creditworthiness, investors cannot rely on the information provided by them. The existence of reliable and neutral ratings of many different borrowers that are based on a set of comparable criteria and that evaluate borrowers in a standardized way enables investors to judge the risk that a borrower may default on his or her obligations. Through ratings, subjective uncertainty is transformed into risk: The investors can now calculate much better the relative risk of a given security. Thus, in a world with ratings, reasonable investors (not gamblers perhaps) will prefer bonds that possess a rating to those without one. Consequently, it will be easier for borrowers to gain access to capital, if they have a rating. Second, systematic and comparable ratings coordinate different types of investors and borrowers. In a well-functioning capital market, higher risks of default will be compensated by higher interests rates and vice versa, so that the expected utilities of high-risk and lowrisk securities converge. If a borrower gets a bad rating, this does not necessarily imply that he or she does not find lenders: Risk-loving investors may accept a bad rating in exchange for the chance to make high profits. Risk-averse investors, on the other hand, will prefer lowrisk and low-interest securities. Thus, ratings serve to coordinate corresponding types of borrowers and investors. Access to capital is not directly linked to the quality of a rating: As long as there are enough risk-loving and profit-oriented capital owners in the market, borrowers with lower grade ratings will still have access to capital. In times of increased risk sensitivity of capital markets, however, borrowers with low-grade ratings will face more difficulties in obtaining access to the market. A world without credit ratings is the worst scenario for investors, because they face high uncertainty and because risks are not adequately compensated for by higher profits. Owing to insufficient information, interest rates increase and, in the long run, high-interest and high-risk borrowers displace low-interest and low-risk borrowers. Therefore, investing money becomes ever more risky. A capital market without ratings is also not the best state of affairs for all issuers. In general, ratings make the world more difficult for borrowers, as they become more transparent to investors and their creditworthiness becomes comparable to that of others. However, in the presence of ratings, low-risk borrowers are not driven out of the market, and they
114
T R A N S N AT I O N A L C O M M O N G O O D S
have to pay lower interest rates. Thus, low-risk borrowers will prefer a situation with ratings to a situation without them. The lack of transparency in a capital market without credit rating is the best state of the world for the high-risk borrowers, but it is much less favorable to reliable and sound borrowers, who have to bear some of the cost (in the form of high interest rates) caused by the high-risk borrowers. Low-risk borrowers have not much to gain, however, if only they themselves are rated, as it is their relative creditworthiness that counts. For them it is important that everybody has a rating, so that they can be recognized as low-risk borrowers. Thus, the sound borrowers have an incentive to introduce ratings for all borrowers, while the less sound would rather have no ratings. The model of provision of ratings by borrowers thus starts with the assumption of heterogeneous players. There are two types of issuers of bonds. Borrower L has a low risk of default and, accordingly, can expect a positive rating (a high grade). Borrower H has a high risk of default, and they can expect a rather negative rating (a low grade). Both borrowers can expect to find investors, with or without rating, as there is sufficient capital supply and both types of investors, risk-loving and risk-averse, are present. Borrowers without rating pay the interest rate that is currently valid in the market of the respective segment of bonds. Negatively rated borrowers have to pay higher interests, while positively rated borrowers pay lower rates. This split of interest rates is only possible, however, if both borrowers possess a rating. The borrowers have two strategies: to commission a rating (R), or not to commission a rating (~R). The differences in interest rates result in benefits b, which are negative for the borrower H and positive for the borrower L. The benefit is zero, if none or only one of the borrowers has a rating, as in this case, there is no informational basis given for a differentiated treatment of the borrowers and thus the standard market rate will be paid. Soliciting a rating, costs a fixed amount c, which is smaller than the benefit b that results from the difference in interest rates. The payoffs are given in table 5.1. Both players are represented by the general formulation of the payoff, where b may take a positive or negative value; the difference between L and H becomes only visible in the ordinal payoffs. This game is a harmony game. It has two Pareto-optimal outcomes, one where L commissions a rating and H does not, and one where both do not commission a rating. The second outcome is the Nash equilibrium. L would prefer a situation where both have a rating. As H has a dominant strategy not to commission a rating, however, L’s best answer is also not to commission a rating. Therefore, in a world
115
SOCIAL AND INSTITUTIONAL CONTEXT
Table 5.1 Provision of Ratings by Heterogeneous Borrowers Assumptions
b . c; L: b . 0; H: b , 0 Strategy Combination
Borrower L and H
Benefit from Rating
Cost of Rating
Payoff
Ordinal L
Ordinal H
R
R
b
c
b2c
3
1
R
~R
0
c
1
2
~R ~R
R ~R
0 0
0 0
02c 0 0
2 2
3 3
Game Matrix
Borrower L
Rating
Rating
No rating
b 2 c, b 2 c 1, 3
0 2 c, 0 2, 2
0, 0 2 c 3, 1
0, 0 3, 2
Borrower H No rating
without ratings and with sufficient capital supply for both types of borrowers, the collective of borrowers has no incentives to introduce credit ratings. For the collective of borrowers, this is a harmony game and does not pose any welfare problem. For the collective of participants in the capital market, including investors, this is clearly a suboptimal state of affairs. Historically, it is obvious that investors— with the help of rating agencies—were able to change this situation and to provide credit ratings themselves. How did this change the incentives for borrowers? Capital Markets with Credit Ratings and Limited Capital Supply For the next model, it is assumed that credit ratings are present in the capital market and that they are taken into account by investors in their decisions. Technically, rating agencies provide the ratings. The costs are paid for by the investors’ subscriptions. However, ratings can also be commissioned and paid for by borrowers. The background for this is that the quality of ratings, which are solely provided on the basis of
116
T R A N S N AT I O N A L C O M M O N G O O D S
publicly available data, is limited. Ratings, which are based on confidential data from the borrowers, are more informative and reliable. To gain access to this data, the rating agencies depend on the cooperation of the issuers. What are the incentives for borrowers to cooperate in such a process, or even to commission and pay for a rating? It has been argued earlier that a capital market without ratings is the worst state of affairs for investors. Investors will in general prefer borrowers who possess a rating and can thus be assessed more easily. Therefore, in a world where many borrowers have a rating and capital supply is limited, borrowers with a rating have an advantage over borrowers without a rating. It is easier for them to access capital. As long as there are risk-prone and risk-averse investors in the market, in principle, both high-risk and low-risk borrowers will find investors. However, securities without a rating face difficulties, as their risks are not calculable for the investors. This model corresponds to the situation in the United States in 1970, when credit ratings were present and widely spread, and capital supply was short of capital demand as a result of increased risk-sensitivity among investors. In this phase rating agencies began to charge fees to the borrowers they rated (section 3.1). In the previous model, costs and benefits in terms of differences in interest rates of ratings have been considered. The first goal of borrowers, however, is to gain capital for financing their projects. Their payoff function has another component apart from the level of interests, namely capital itself. This component can be assumed to weigh more than interest, since a borrower would rather pay higher interests than not get capital at all. At least this is true for moderate interest rates; if capital costs are very high, some borrowers must sacrifice their projects. For the model in table 5.2 it is assumed that the benefit of having access to capital, a, is much greater than the benefit, b, from the interest rate difference. Capital supply is limited. In the cases where only one of the two borrowers has a rating, this one will get the capital, while the other will not. In the cases where the players either both possess a rating or both do not, they have an equal chance of gaining capital. Thus the expected utility of getting the capital is ½a in the case of two players. This is assumed to still count more than b. For simplicity, I will ignore here the fact that there may be a difference between the chances of the high-risk and the low-risk borrowers if the capital market is very risk-sensitive. All other assumptions are as they were in the earlier model. This game is a game of pure conflict.2 Each of the four outcomes is Pareto-optimal. The problem that arises here is one of pure distributional
117
SOCIAL AND INSTITUTIONAL CONTEXT
Table 5.2
Provision of Ratings with Limited Capital Supply
Assumptions
½a . b; ½a . b 1 c; b . c; L: b . 0; H: b , 0
Strategy Combination
Borrower L or H
Benefit from Capital
Benefit Cost of from Rating Interest
Payoff
Ordinal Ordinal L H
R
R
½a
b
c
½a 1 b 2 c
3
R
~R
a
0
c
4
4
~R ~R
R ~R
0 ½a
0 0
0 0
a2c 0 ½a
2
1 2
1 3
Game Matrix
Borrower L
Rating
Rating
No rating
½a 1 b 2 c, ½a 1 b 2 c 2, 3
a 2 c, 0 4, 1
0, a 2 c 1, 4
½a, ½a 3, 2
Borrower H No rating
nature. Borrower L prefers the outcome where both have a rating, over the outcome where neither has a rating. For borrower H it is the other way round. The other preferences are equal. The Nash equilibrium is at the outcome where both have a rating; it thus gives the distributional advantage to the low-risk borrower L. Both have a dominant strategy to solicit a rating. This result rests on a further assumption that has not yet been justified: ½a . b 1 c. If this assumption is given up, the payoff in the first line might become negative for borrower H, changing his or her preference order. This leads to a rambo game, again a merely distributional problem. Borrower L commissions a rating at the Nash equilibrium, while H does not. This shows again, how important the assumptions about costs and benefits are. Despite the change in this assumption, however, the result remains in the same class of games, namely, mere distributional conflicts. If heterogeneity were removed from the game, assuming the same level of risk or the same rating grade for all borrowers, thus focusing only on the aspect of facilitated access to the capital market for those borrowers with a rating, the game would become a prisoners’ dilemma.
118
T R A N S N AT I O N A L C O M M O N G O O D S
As in the pure conflict game, both borrowers have a rating at the Nash equilibrium. The incentives are such that they solicit a rating, although both would prefer not to solicit one. What is a dilemma for homogeneous borrowers, however, is a distributional problem for heterogeneous issuers and a common good for the investors. This analysis is valid, if capital is scarce and if investors generally prefer bonds with a rating to those without—which are realistic assumptions. Compared to the earlier model, a double change in the social environment has been assumed: First, capital supply has become scarce, and second, the investors are now prepared to consider ratings both as means for differentiating interests rates between high- and low-risk borrowers and for their decision whether to give capital to a certain borrower at all. These changes have turned a harmony game, where borrowers’ incentives lead them not to commission ratings, into a distributional problem among borrowers, where both highand low-risk borrowers will solicit ratings, although the low-risk borrowers gain relatively more from the presence of credit ratings. Capital Markets with Obligatory Credit Ratings In the last model, I analyze how the presence of state regulation influences the decisions of borrowers. State regulations that refer to ratings are designed to decrease credit risk. They target primarily institutional investors, but also issuers and even stock exchanges. Whereas the regulation of issuers is restricted to various requirements about disclosure of information according to the grades borrowers have received in ratings, there are many different forms of regulations for investors. In the United States, all kinds of institutional investors are obliged to consider credit ratings in their investment decisions, among them commercial banks, investment banks, savings & loan companies, insurance companies, pension funds, and mutual funds. Three basic types of regulation can be distinguished: investment restrictions, accounting prescriptions, and minimum capital requirements (Kerwer 2000). In most cases, the regulations prescribe that stricter rules apply if a certain minimum grade has not been achieved by a borrower or bond in the ratings of at least one or two acknowledged agencies. Financial regulation that relies on the results of credit ratings of specialized rating agencies is most widespread in the United States. To give a few examples: The very first US regulation in 1931 required banks to mark-to-market-value bonds with a rating lower than BBB; the second one, in 1936, prohibited the purchase of “speculative”
SOCIAL AND INSTITUTIONAL CONTEXT
119
securities (lower than BBB) by commercial banks. This was extended to savings & loans in 1989. In 1951 and 1975, two regulations imposed higher minimum capital requirements on lower rated insurance company and brokerage bonds; that is, the investors have to set aside a higher share of capital as a security for higher risks (Basel Committee on Banking Supervision 2000: 54). In other countries ratings are used far less in financial regulation. Most European countries apply only the rules provided by the socalled market risk amendment of the Basle Accord of 1988. European Union (EU) members are required to do so by the EU Capital Adequacy Directive (Council Directives 1993/6 and 1998/31). This amendment provides guidelines for minimum capital requirements banks have to set aside a “risk cushions” (Szego 1997). Germany does not even apply these rules, as is allowed for in the directive by a waiver. Australia, Hong Kong, New Zealand, and many Latin American countries apply some regulations, which refer to ratings (Basel Committee 2000: 42−43). The background for state and international regulation of credit risks is the protection of private investors as well the protection of the capital market against systemic risk. Private investors, in general, are not able to evaluate the risk of a given security, owing to lack of information, expertise, and experience. The regulation of institutional investors’ behavior toward risk prevents them either from selling highrisk bonds to their customers (in the case of banks) or from investing money from individual investors in highly speculative securities (in the case of funds). Furthermore, individual customers are protected against damage if their bank or fund collapses as a consequence of having pursued a too risky credit policy. To the extent that credit risk regulation aims at protecting private investors, it is a merit good. The state wants to protect individuals from the consequences of their own incompetence. It is not a common good, however, as there are no externalities. However, state regulation also helps to avoid failures of banks and other institutional investors. Regulation of institutional investors’ behavior toward credit risk decreases the systemic risk inherent in capital markets, and in this way it is a common good. The regulatory requirements, which do not allow institutional investors to invest in speculative grade bonds or prescribe different accounting rules and minimum capital reserves, imply that the investors will only invest in securities that posses a rating. Irrespective of what the regulation exactly forbids or prescribes, its main effect is that investors will not buy bonds without ratings. Thus, the analysis is restricted to the most fundamental effect implied in each of the
120 Table 5.3
T R A N S N AT I O N A L C O M M O N G O O D S
Provision of Ratings in the Presence of Obligatory Rating Requirements
Assumptions
½a . b; ½a . b 1 c; b . c; L: b . 0; H: b , 0
Strategy Combination
Borrower L or H
Benefit from Capital
Benefit Cost of from Rating Interest
Payoff
Ordinal Ordinal L H
R
R
½a
b
c
½a 1 b 2 c
2
R
~R
a
0
c
3
3
~R ~R
R ~R
0 0
0 0
0 0
a2c 0 0
2
1 1
1 1
Game Matrix
Borrower L
Rating
Rating
No rating
½a 1 b 2 c, ½a 1 b 2 c 2, 2
a 2 c, 0 3, 1
0, a 2 c 1, 3
0, 0 1, 1
Borrower H No rating
rules: Securities without ratings are actually excluded from the capital market for institutional investors. For the model, it is assumed that borrowers without a rating are effectively excluded from the market, whether they are high-risk or low-risk borrowers. All other assumptions are as they were in the earlier model. Table 5.3 gives the payoffs. As borrowers who do not commission a rating are practically excluded, the choice of the other player is irrelevant in these strategy combinations. There is no benefit from capital and the payoff is zero. Given the cost-benefit assumptions, the least preferred option for both players is thus not to commission a rating. If both players have a rating, access to capital is possible but limited by competition. If only one of the borrowers has a rating, this one will obtain the capital for sure. Thus, the preferred option for both is that they have a rating while the other does not; their second preference is that both have a rating. This structure is a harmony game. The Pareto-optimal Nash equilibrium is at the outcome where both issuers commission a rating. Thus, compared to the situation, where capital supply is limited, state regulation has not changed the result that ratings become general
SOCIAL AND INSTITUTIONAL CONTEXT
121
practice. However, it has turned the game from a distributional conflict into a harmony game. The effect of factual exclusion from the market is very strong. It turns the preferences of issuers upside down: Both have a dominant strategy to solicit ratings. The exclusion effect is even stronger than the heterogeneity of actors: The preference order for the low-risk and the high-risk borrower is the same. Again, this result hinges on the cost-benefit assumptions: If ½a . b 1 c is given up, the payoff of the high-risk borrower may become negative for the strategy combination (R, R). The result is a rambo game. In this case, heterogeneity is stronger than the exclusion effect and the high-risk borrower prefers to have no rating and stay out of the institutional investors’ market. This is not a plausible case, however, as having a rating and gaining access to capital seems in every way preferable to not having to pay for a rating and not having access to capital. This analysis is valid wherever there are regulations of the type prevailing in the US capital markets. In other segments of the market— for example, in Germany—borrowers still do not have much of an incentive to commission ratings. Non-US borrowers, however, who want to gain access to the flourishing US capital markets, need to have ratings. Furthermore, US issuers may seek capital in non-US domestic or international markets. The presence of regulation that requires rating in the largest and most attractive capital market has in general restricted access to capital. In such a situation, it pays for issuers of securities to improve their creditworthiness in the long run, so that they can commission a rating with an acknowledged agency and, accordingly, that they can expect to get an investment grade rating. State regulation of credit risk that obliges investors to invest only in rated securities can explain why there are in fact incentives for borrowers to commission credit ratings and to pay it. It can also explain why the demand for rating shifted from investors to borrowers after ratings had become a regular feature of the capital market. There was willingness to pay for a rating on the side of the borrowers and the rating agencies could make use of this. We have seen, however, that a similar—although less strong—effect is there if there is shortage of capital and investors prefer rated bonds to nonrated bonds. Whenever the access to capital for nonrated borrowers is limited, whether by state regulation or by the decisions of the investors, there is an incentive for the collective of borrowers to provide credit ratings on their own. In fact, the transition from investors being charged for the ratings to borrowers paying a fee in the 1970s happened at a time when there was reticence of investors to buy bonds and when a number of regulations
122
T R A N S N AT I O N A L C O M M O N G O O D S
were present. Nevertheless, it is surprising that the transition did not occur earlier, as there were already some regulations present in the 1930s. This might be a consequence of the fact that the early regulations did not cover all segments of the market. It might also be, however, that rating agencies simply did not sense that they could charge issuers for their ratings, or that there was fear that this could threaten the neutrality of their ratings in the eyes of investors. Obviously, it was the lack of demand for commercial bonds at the beginning of the 1970s that finally brought about the change in the structure of payment for ratings (v. Randow 1996: 553). 5.2 Trade Regimes and Regulatory Competition for Environmental Standards In this section, the example of regulatory competition in the environmental field will be used to show how different institutional rules affect the strategic situation and thus the provision of common goods. Before the effect of institutions, in this case, different trade regimes, can be examined, the basic problem of regulatory competition must be modeled. There are some important characteristics of the problem that have to be taken into account. As a consequence, three different conditions will be varied. The first dimension is whether the actors have homogeneous or heterogeneous preferences, the second is the type of environmental standard used, and the third is the prevailing trade regime. Before I go on to discuss the model, I shall consider why regulatory competition can be viewed as a common goods problem. Regulatory Competition as a Common Goods Problem The subject of environmental regulatory competition among states is environmental protection standards. The aim of environmental standards is to preserve a certain environmental common good—for example, clean air, clean water, a species, or a habitat—with respect to a certain pollutant or with respect to other forms of deterioration. What these goods have in common, to be considered relevant for regulatory competition, is their transboundary nature. “Purely domestic” regulations, for example, local noise standards set to preserve the good of a quiet environment, lie outside the scope of the following analysis. Transboundary common goods affect several states or, more generally, several jurisdictions, which may regulate the use of the environmental good on their own. However, because of externalities between the
SOCIAL AND INSTITUTIONAL CONTEXT
123
jurisdictions that result from the transboundary nature of the goods, the individual regulations will not be efficient. The spatial scope of the goods is determined by biological and physical conditions, and it cannot easily be adapted to political borders. Examples of such environmental goods are clean air, transboundary bodies of water, and the climate, to name a few. In many cases they are commons, which means that consumption is rival but not excludable. These goods are increasingly being destroyed as a consequence of productive and consumptive activities, for example, through emissions from products or production processes. Contributions to the preservation or restoration of these goods consist, first of all, in reducing emissions, or, less specifically, in restricting the destructive activities. Since we are talking about common goods, in general the reduction will not be achieved by voluntary or market behavior, but by the regulatory activity of the affected jurisdiction. State regulation—for example, an emission standard— can thus be viewed as providing the common good within the jurisdiction. At the super-jurisdictional level, the regulations of the individual jurisdictions must be interpreted as the jurisdictional actors’ contributions to the common good. In the following, the analysis will focus on the regulatory level. For the following analysis, it is not necessary to specify the good, the cause of its deterioration, and the type of regulation. However, for the purpose of illustration we may think of clean air polluted by the emission of nitrogen oxides. Nitrogen oxides cause long-range air pollution; the primary sources are large combustion plants and vehicles. They are regulated by limit values set to the emission of nitrogen oxides from large combustions plants, as well as by limit values for car emissions, among them nitrogen oxides. In times of a globalized economy not only are environmental goods transboundary, so are economic processes. Commodities that are subject to environmental regulation are traded internationally. Firms and economies are subject to international competition. In states with stricter environmental regulations, productions costs for firms are higher and the firms consequently suffer a competitive disadvantage in comparison to firms in states with laxer standards. At least, this is usually assumed in the literature on environmental regulatory competition. The classical research question about regulatory competition is whether and under which conditions this situation leads to a regulatory “race to the bottom” or to a “race to the top”. In the case of transboundary environmental problems, there is evidence of both
124
T R A N S N AT I O N A L C O M M O N G O O D S
outcomes (Vogel 1995, 1997; Zürn 1997; Kern 2000; Holzinger 2003; Holzinger and Knill 2008; Holzinger, Knill, and Sommerer 2008). However, a convincing explanation and clarification of the conditions that lead to one or the other result is still missing. Some important factors have been dealt with in the literature, such as the distinction between product and production standards (Scharpf 1996, 1997a), but most factors have not been systematically analyzed. Since this chapter only concentrates on three variables important in the wider context of this book, it will not be able to fill this gap in the research. However, it shows that the type of trade regime may contribute to the explanation. Homogeneous and Heterogeneous Actors The first factor to be varied is the homogeneity and heterogeneity of actors. The players are two states, which regulate the transboundary environmental good by emission standards. They have two strategies: namely employing high (H) or low (L) standards. This is analogous to making “a large contribution” or “a small contribution” to the common good. Homogeneous states have the same preferences regarding the two strategies; heterogeneous states have different preferences. The first case represents a symmetric game, as both players not only possess identical strategies but also identical payoffs. The second case represents an asymmetric game. I start with the symmetric case. The high standards are assumed to be equivalent to two units of contribution to the environmental good; the low standards are equivalent to one unit. The only further assumptions that must be made concern the relation of the costs and benefits of the standards. Two cases can be distinguished: (1) the benefits b per unit of contribution are lower than the costs c per unit, and (2) the costs per unit are lower than the benefits. Table 5.4 gives the model of case 1. The game in table 5.4 is a prisoners’ dilemma. There is a unique equilibrium in dominant strategies at the Pareto-inferior outcome. This is exactly what is usually expected in the provision of public goods. Both states choose the low-level regulation; the consequence is a “race to the bottom”. This is set in conditions of nonrivalry and nonexcludability in which the benefit from the individual contribution is lower than its cost. What happens in case 2 where the cost-benefit relation is reversed? The preference order for country A changes as follows: 4, 3, 2, 1 (in the order of lines in table 5.4). Not surprisingly, the result is a harmony
125
SOCIAL AND INSTITUTIONAL CONTEXT
Table 5.4 Regulatory Competition with Homogeneous Countries Assumption
c.b Strategy Combination
Country A
Benefit from Cost of Rating Subscription
Payoff
Ordinal
H
H
4b
2c
4b 2 2c
3
H
L
3b
2c
3b 2 2c
1
L
H
3b
c
3b 2 c
4
L
L
2b
c
2b 2 c
2
All factors are identical for country B.
Game Matrix
Country B High standards
Low standards
High standards
4b 2 2c, 4b 2 2c 3, 3
3b 2 2c, 3b 2 c 1, 4
Low standards
3b 2 c, 3b 2 2c 4, 1
2b 2 c, 2b 2 c 2, 2
Country A
game. The collectively optimal outcome is achieved, and therefore no exogenous action is necessary. This outcome is equivalent to the notion of a “race to the top”: Both states apply the high standards.3 The reason for this is the cost-benefit relation. The cost is less than the benefit of each unit of contribution for each player. Hence, it makes sense to contribute an additional unit. There is no dilemma here. Still, one would talk about a common good, as it is nonrival and nonexclusive. Each player also benefits from the others’ contributions. We would not talk about a collective action problem, however, as the good would be provided as a consequence of the evaluation of its benefits and costs by the actors. More realistic in the context of regulatory competition is an asymmetric situation. The two states have heterogeneous preferences in respect to the environmental good. In practice, this will be caused by two factors. First, there are states, which place a greater value on the environmental good. Their benefit from the preservation or restoration of the good is higher than that of other states. Second, there are states, which have lower contribution costs than others. It is cheaper for them to achieve a certain environmental standard. For the
126
T R A N S N AT I O N A L C O M M O N G O O D S
examples, these two factors need not to be distinguished. It is enough to distinguish “green” and “nongreen” countries. The green countries often have highly developed economies, and their citizens have a preference for a high environmental quality. These countries are able and willing to afford the economic costs of an ambitious environmental policy. The nongreen countries are usually economically less developed and the citizens do not consider the clean environment to be a priority. The economic costs of a strict environmental policy are high. These countries will therefore prefer less stringent regulations. The model is the same as above. The benefit of the environmental good per unit is now higher for the green country when compared to the nongreen country. The costs per unit contribution are equal. They imply c . b for the nongreen country and b . c for the green one. Thus, the variation in the payoffs is only a result of the different valuation of the environmental good. This way, the asymmetry of payoffs is sufficiently represented. The assumption of different cost structures does not change the incentive structure as long as the nongreen country values the costs of the contribution more than its benefits and the reverse is true for the green country. Table 5.5 gives the model. Table 5.5 Regulatory Competition with Heterogeneous Countries Assumptions
bP , c ,2bP for the nongreen country, N bR , c for the green country, G Strategy Combination
Countries G and N
Benefits Costs
Payoff P Payoff R Ordinal Ordinal N G
H
H
4b
2c
4bP 2 2c 4bR 2 2c
3
4
H
L
3b
2c
3bP 2 2c 3bR 2 2c
1
2
L
H
3b
c
3bP 2 c
3bR 2 c
4
3
L
L
2b
c
2bP 2 c
2bR 2 c
2
1
Game Matrix
Country G High standards
Low standards
High standards
4bP 2 2c, 4bR 2 2c 3, 4
3bP 2 2c, 3bR 2 c 1, 2
Low standards
3bP 2 c, 3bR 2 2c 4, 3
2bP 2 c, 2bR 2 c 2, 1
Country N
SOCIAL AND INSTITUTIONAL CONTEXT
127
The result is no surprise. The equilibrium is attained when the nongreen country chooses low standards and the green country, high standards. It is a Pareto-optimal equilibrium in dominant strategies. There is a second Pareto-optimal outcome, where both states use the high standards. There is no dilemma between individual and collective rationality and there is no problem in selecting the equilibrium. The game is a rambo game, which poses only a distributional problem. The nongreen country gets its first preference, namely low standards within its own jurisdiction and high standards within the green country’s jurisdiction. Consequently, different standards apply over the whole area. There are two regulatory areas, and in the case of product standards the market becomes segmented. There is neither a “race to the top” nor a “race to the bottom”. The Influence of Market Segmentation So far it has been implicitly assumed that market segmentation or nonsegmentation by different standards does not affect the national economies. However, the industries concerned are affected by market segmentation in two respects: The first aspect is whether and in which state differentiated standards lead to competitive advantages or disadvantages for the industries. Different standards, both product or production standards, may cause different costs for the industries concerned. If firms from both states trade their products within the whole market, different standards lead to competitive advantages for the firms in the nongreen country as well as to competitive disadvantages for the firms in the green country. The competitive disadvantage regarding product standards vanishes for the green country if barriers to trade are permitted for environmental reasons. In case of production standards, there is no way to escape the disadvantage for the green states’ industry. The second aspect is whether a common standard for the whole market implies harmonization advantages (or: segmentation costs). Advantages through harmonization can be expected if products are subject to different national environmental standards, if licensing procedures are different for these products, and if they have to be produced with different variations in each state. This is the case if the prevailing trade regime allows the states to wall off foreign products that do not fulfill the domestic environmental standards. In such a situation, the harmonization of standards leads to economic gains for the industries concerned as the average costs decrease. The example of limit values for car exhaust emissions fits into this pattern: They are product standards, the respective product is globally traded, and
128
T R A N S N AT I O N A L C O M M O N G O O D S
producing different varieties, as well as licensing procedures cause high costs. Thus, market segmentation is expensive for car producers, which is the reason why car emission limits have been harmonized voluntarily ever since the 1950s within the institutional frame of the United Nations Economic Commission for Europe (ECE). The rules developed there have been taken over into EU regulation, and only during the 1980s the EU began to introduce its own and more ambitious regulation (Holzinger 1994: 190–194). US standards for car emissions played a role as benchmark, because the US market for cars is an important market also for European car produces: The harmonization of US and EU standards promised efficiency gains. There are no harmonization advantages, when there is a rule of mutual recognition of the products and when the protection from foreign products is not permitted for environmental reasons. Products that are licensed according to different standards can be sold throughout the entire market. In such a situation the green country does not achieve its environmental goal. The regime leads to a positive externality for the nongreen country and a negative externality for the green one. However, this is only true relating to product standards. Production standards in general do not lead to any harmonization advantages or segmentation costs. The regulation of nitrogen oxides from large combustions plants is an example of production standards. Thus, the incentive structures for the preservation of an international common environmental good vary with two additional properties: 1. Are the instruments used to achieve the common good product standards, or are they production standards? 2. Does the prevailing trade regime permit trade barriers for the sake of the environment, or is it possible to enforce the mutual recognition of products? The difference of product and production standards relating to regulatory competition has already been analyzed by Scharpf (1996, 1997a) while the effect of trade regimes has so not been systematically examined. Altogether, four cases can be distinguished (table 5.6). How do the incentive structures in these four cases differ? In case 1, there is a harmonization advantage for both states. However, there are no competitive advantages or disadvantages because the product prices—as far as they are determined by the environmental standard— are the same for both states. In cases 2 and 3, there are no harmonization advantages, as markets are not segmented. The nongreen country has a competitive advantage, and the green one has a disadvantage. It
SOCIAL AND INSTITUTIONAL CONTEXT
129
Table 5.6 Type of Standards and Trade Regimes
Trade Barriers
Free Trade
Product Standards
Production Standards
(1) harmonization advantage no competitive dis/advantage
(4) no harmonization advantage small competitive dis/advantage, in/decrease of turnover
(2) no harmonization advantage competitive dis/advantage
(3) no harmonization advantage competitive dis/advantage
does not make a difference here whether product or production standards are concerned. In case 4, there are no market segmentation costs, as only production facilities are affected. The nongreen country has a competitive advantage, as it has lower production costs. However, this advantage is restricted to its own territory if the green country is able to erect trade barriers against products from the nongreen country where production is subject to low-level regulation. As a consequence, within the green country’s territory there is a loss of turnover for the nongreen country’s industry. The reverse is true for the green state. Its competitive disadvantage is restricted to the nongreen country’s territory, and its domestic turnover increases. The following assumptions will be made for all cases: The industries in both states sell their products throughout the entire market and have an equal share of the market. Furthermore, the harmonization advantage is the same for low and high standards, which is a plausible assumption. The models rest on these and further symmetry assumptions, as they assume that all other factors, such as market size or production costs are equal for both states. Case 1: Product Standards and Trade Barriers The harmonization advantage is the same for both countries. I shall assume that it weights less for the countries than their benefits of an additional unit of the common good, but more than their net benefits per unit. This is equivalent to saying that the economic advantages of a nonsegmented market are valued higher by the national regulators than the difference in benefits (minus the costs of regulation) from the common good between high and low standards. If the condition holds true for the green country with its positive net benefit per unit, it is trivially true for the nongreen one, whose net benefit per unit is
Table 5.7
Regulatory Competition with Product Standards and Trade Barriers
Assumptions
bP , c , 2bP for N; bR . c Strategy Combination
Countries N and G
for G; h . bi − c; h , b
Benefits
Costs
Harmonization
Payoff N
Payoff G
Ordinal N
Ordinal G
H
H
4b
2c
h
4bP 2 2c1h
4bR 2 2c1h
4
4
H
L
3b
2c
0
3bP 2 2c
3bR 2 2c
1
1
L
H
3b
c
0
3bP 2 c
3bR 2 c
3
3
L
L
2b
c
h
2bP 2 c1h
2bR 2 c1h
2
2
Game Matrix
Country G High standards
Low standards
High standards
4bP 2 2c1h, 4bR 2 2c1h 4, 4
3bP 2 2c, 3bR 2 c 1, 3
Low standards
3bP 2 c, 3bR 2 c 3, 1
2bP 2 c1h, 2bR 2 c1h 2, 2
Country N
SOCIAL AND INSTITUTIONAL CONTEXT
131
negative. If I would assume the harmonization advantage to be smaller than the net differential benefit between levels of standards it would have no substantial effects on the strategic constellation. The harmonization advantage becomes relevant where both states choose the same strategy, that is, high or low standards, respectively (table 5.7). This modeling leads to an assurance game, provided the harmonization advantage is sufficiently large (h . bi – c). It has two equilibria, one Pareto-optimal, the other suboptimal. The game is symmetric although the players are heterogeneous. The harmonization advantage drives their interest into the same direction. The optimal equilibrium represents the solution with harmonized high standards, the suboptimal one, and the outcome with harmonized low standards. Segmentation is no equilibrium. The problem is one of coordination: if the states communicate, they will easily agree to introduce high standards, as both states realize their highest payoff then. This situation leads to a “race to the top” regarding environmental standards. Without communication, however, there is a certain chance that the states will choose different strategies and “miss each other,” although the optimal equilibrium is focal. Only if h is very small (h < bi – c) will the game become a rambo game. The game changes if only one of the countries has a harmonization advantage. If, for example, the nongreen country’s industry does not export its products, it has no harmonization advantage. The nongreen state has a dominant strategy at low standards. The resulting equilibrium is suboptimal (the upper sets of lines in table 5.8). Both states choose low standards, which means that a race to the bottom will take place. If the nongreen state has a unilateral harmonization advantage, a unique equilibrium exists where both states choose high standards (lower sets of lines in table 5.8). Case 2: Product Standards and Free Trade In a free-trade regime, the mutual recognition of products manufactured according to different environmental standards can be pushed through. There are no market segmentation costs for the national industries. However, there is a competitive advantage for the country applying the low standards if the other country chooses the high ones, and vice versa. This is true as long as different environmental standards lead to different costs and product prices. There is no effect on competition if both states choose high or low standards. For simplicity, a symmetric competition effect (e) is assumed, where the competitive advantage and the disadvantage are of the same size. As in case 1 with the harmonization effect, it shall be assumed that the competition effect is smaller than the benefit b
132
T R A N S N AT I O N A L C O M M O N G O O D S
Table 5.8 Regulatory Competition with Unilateral Harmonization Advantages Assumptions
– country G has a unilateral harmonization advantage (upper sets of lines); – country N has a unilateral harmonization advantage (lower sets of lines)
Game Matrix
Country G High standards
Low standards
3, 4 4bP 2 2c, 4bR 2 2c 1 h
1, 3 3bP 2 2c, 3bR 2 c
4bP 2 2c 1 h, 4bR 2 2c 4, 4
3bP 2 2c, 3bR 2 c 1, 3
4, 1 3bP 2 c, 3bR 2 2c
4, 1 3bP 2 c, 3bR 2 2c
3bP 2 c, 3bR 2 2c 3, 2
3bP 2 c, 3bR 2 2c 3, 2
High standards
Country N
Low standards
from an additional unit of the common good, but higher than the net benefit of the difference between the high and low standards. In this case the states are in a classical prisoners’ dilemma constellation. It has a Pareto-inferior equilibrium in dominant strategies, where both countries choose the low standards. If all the earlier conditions are met, a free-trade regime will result in a race to the bottom. If the competition effect is small (e < bi – c), the game changes once again to the original rambo game (table 5.9). Case 3: Production Standards and Free Trade Whenever environmental regulations affect the production processes and not the products themselves, there is no market segmentation, and consequently no harmonization advantage. The production is stationary in each state, but the products are traded throughout the entire territory. The products themselves do not have different properties, but they are produced with different manufacturing processes. There is a dilemma for the country with high standards: On the one hand, in a free-trade regime, trade barriers on the basis of environmental considerations are not permitted on product properties themselves. On the other hand, in most cases the production processes cannot be monitored, and even if it is known that the products
133
SOCIAL AND INSTITUTIONAL CONTEXT
Regulatory Competition with Product standards and Free Trade
Table 5.9 Assumptions
bP , c , 2bP
for N; bR . c for G; e . bi 2 c; e , b
Strategy Benefits Costs Competition Combination Countries N and G
H
H
4b
2c
0
H
L
3b
2c
2e
L
H
3b
c
L
L
2b
c
Payoff N 4bP 22c
Payoff G 4bR22c
Ordinal Ordinal N G 3
3
3bP 22c2e 3bR22c2e
1
1
e
3bP 2c1e
3bR2c1e
4
4
0
2bP 2c
2bR2c
2
2
Game Matrix
Country G High standards
Low standards
High standards
4bP 2 2c, 4bR 2 2c 3, 3
3bP 2 2c 2 e, 3bR 2 c 1 e 1, 4
Low standards
3bP 2 c 1 e, 3bR 2 2c 2 e 4, 1
2bP 2 c, 2bR 2 c 2, 2
Country N
are produced according to lax production standards, the only possible reaction is to reject the products from this country. In a freetrade regime, however, the complete ban on the products is considered discriminatory, thus it is not permitted. Consequently, the green country has a competitive disadvantage, while the nongreen country (applying low standards) has a competitive advantage. It is possible that the consumers in the green country have such a high preference for the environment that they buy the more expensive domestic products because such products are more environmentally friendly. Then the industry in the nongreen country would experience a decrease in turnover. As it is not very realistic, this possibility will be ignored. The strategic constellation in case 3 is thus the same as in case 2: namely a prisoners’ dilemma. Equilibrium results where both states choose low standards. Case 4: Production Standards and Trade Barriers for Environmental Reasons In general, it is practically impossible to erect trade barriers against production standards even if they were permitted, because it is difficult to monitor production processes. Nevertheless, there are examples of trade embargoes for environmental reasons (Vogel 1997).
134
T R A N S N AT I O N A L C O M M O N G O O D S
If the green country does not allow the nongreen country’s products to be admitted into its market, the competitive advantage of the nongreen country is restricted to its own territory. In the same way, the competitive disadvantage of the green country arises only in the nongreen country’s territory. However, the nongreen country’s industry suffers a loss in turnover, as they cannot sell their product in the green country. Consequently, the green country’s industry experiences an increase in domestic turnover. Then, the turnover effects offset the competition effects, as long as the symmetry assumptions about market size and market shares are retained. Therefore, the game structure associated with this is once again the rambo game. There is one equilibrium in dominant strategies, where the green country chooses the high standards and the nongreen country the low ones. Results of the Analysis of Regulatory Competition The analysis has shown that regulatory competition in the environmental field may lead to several different outcomes, depending on the exact conditions. In the symmetric case, the outcome is a result of the relation between costs and benefits of the individual contributions to the common good. If the costs are lower than the benefits, we end up with a harmony game. If the benefits are lower than the costs, the actors are in a prisoners’ dilemma, and a race to the bottom of the environmental standards has to be expected. If the actors have heterogeneous preferences, the strategic constellation is a rambo game. Each state applies its most preferred standards. This leads to market segmentation. The analysis becomes more complex if assumptions about the trade regimes and the type of environmental standard are introduced on the basis of heterogeneous preferences. In a free-trade regime where mutual recognition of products is the rule, regulatory competition leads to a “race to the bottom,” as all states choose the low standards, irrespective of the type of standards. The incentive structure is a prisoners’ dilemma. In a trade regime where products can be excluded if they are manufactured according to low production standards, the market will become segmented. In the rambo game one state will introduce high standards, the other low ones. Finally, if it is also possible to apply domestic product standards to foreign products, it is in the interest of both states to coordinate their action and to apply the high standards. The related game is an assurance game. A “race to the top” regarding environmental standards can be achieved in this constellation (table 5.10).
SOCIAL AND INSTITUTIONAL CONTEXT
Table 5.10
Trade Barriers Free Trade
135
Results of the Analysis of Regulatory Competition Product Standards
Production Standards
(1) assurance (2) prisoners’ dilemma
(4) rambo (3) prisoners’ dilemma
These outcomes are a result of varying effects of harmonization of standards and competition among the countries’ industries in a common market. Whenever the harmonization and the competition effects are very small compared to the benefits and costs of providing the common good, a rambo game results, where each country employs its most preferred standards. In this case neither a “race to the bottom” nor a “race to the top” can be expected. From the perspective of welfare economics, the purely marketbased solution is worst. Free trade leads to inefficiency. With respect to efficiency, the solution, which permits protection from negative externalities leads to preferable outcomes. In case of product standards the Pareto-optimal solution is one of two possible equilibria; there is a coordination problem. In case of productions standards the equilibrium is one of two Pareto-optimal solutions. However, this result will be accompanied by distributional consequences, which could be considered unfair. The nongreen states decision to accept low standards occurs at the cost of the green state, which is forced to bear the negative externalities. That there is no clear and unique prediction of the effects of regulatory competition in the environmental field corresponds to empirical observations. A general “race to the bottom” has not occurred, and there are some clear cases of “races to the top,” especially in the field of car emissions regulation (Holzinger 1994; Vogel 1997). The analysis carried out here offers some suggestions for explanations. First, if the concerned industry has a lot to gain from harmonized product standards in global markets, a “race to the top” is a likely and rational result. Second, in the case of production standards, one would not expect a clear movement of regulations into either of the two directions, as long as there is no rigid free-trade regime that is relevant for the products of the regulated facilities. Third, in practice no strict freetrade regimes exist regarding many environmental matters. Both under EU and under World Trade Organization rules, trade barriers can often be justified on the basis of health and environmental reasons (Epiney 2000). Finally, even if a strict free-trade regime applies, the
136
T R A N S N AT I O N A L C O M M O N G O O D S
costs of environmental standards are often too low to be a relevant competition factor (Vogel 1997). There are a number of other factors characterizing the market and institutional environment, which have been held constant in this analysis. Some of them, especially relative market size of the countries, will also be important for the strategic constellation and may thus contribute to the explanation of the fact that there is obviously no universal trend in regulatory competition for environmental standards (Holzinger 2003).
Chapter 6
Strategic Constellations and Collective Action Problems
In the previous chapters a number of attributes of the social situations
of common goods provision were examined that affect the strategic constellation and the collective action problems involved. Among these properties were: the type of common good, the cost-benefit configurations, the aggregation technology of the goods, the homo- or heterogeneity of the actors, and some properties of the institutional context. The applications involved different kinds of individual and collective actors: individuals, firms, or political jurisdictions, such as states and communities. The strategic constellations in the illustrations included a number of different games. These games entail different kinds of collective action problems. Some of them are dilemmas—that is, the actors have an incentive to defect from the collectively optimal solution (prisoners’ and asymmetric dilemmas); others present coordination problems, where the actors face the risk of not being able to coordinate their strategies for a desirable outcome (assurance, chicken); still others involve inequality and distributional problems (rambo, chicken, pure conflict); finally, some of the games represent several problematic aspects or several kinds of collective action problems (chicken). What conclusions can be drawn from this analysis of so many and different examples? Which general lessons are implied? The first point is obvious: The simplistic conclusion—that the provision of common goods necessarily poses a collective dilemma—is not valid. There are so many factors influencing the social situations that it is impossible to draw general conclusions about the strategic constellations and the collective action problems posed by common goods provision. The models of the empirical cases presented earlier are extreme simplifications of
138
T R A N S N AT I O N A L C O M M O N G O O D S
the respective situations; yet, they produce a variety of strategic constellations. This implies that each common goods problem has to be analyzed individually and in some depth—if rash generalizations are to be avoided. The “problem adequacy” of the models developed for the empirical cases in chapters 3 through 5 has varied. In some cases, only a single aspect of the situation has been examined. This is true, for example, for the cases of global warming and biodiversity, where only the aspect of aggregation technology was of interest. For appropriate models of global warming or biodiversity protection, other aspects would have to be included in the analysis. Other cases have been analyzed in more depth. The model of tax coordination in the European Union (EU) is a rather good approximation of the problem. The regulatory competition model shows the influence of trade regimes, but still retains as constants a number of other factors that are important for empirical prediction. What can be judged as an appropriate model for a situation depends, of course, on the research question and on the aim of the analysis. The fact that differentiated analysis is needed to avoid overgeneralization does not imply that we are left with the only alternative way of analyzing individual cases. Some general conclusions can still be drawn. Identical combinations of the characteristics of a situation will yield an identical strategic structure. It is therefore possible to conclude, given the presence of certain attributes in a common goods problem, that certain kinds of collective action problems will be present and will have to be solved. In chapters 3 through 5, five properties were selected and several variations among each of these dimensions were examined. This indicates that the number of ways in which properties can be combined is huge. It is of no use to try to find out in a casuistic way what the corresponding game structure would be for each combination. However, the variation in single factors has systematic effects, provided that other factors are kept constant. The systematic effects of the five properties selected in chapters 3 through 5 will be summarized in section 6.1. Despite so many different combinations of attributes examined in the earlier chapters, the number of different strategic constellations proved to be limited. Some 30 different combinations have been looked at, but only 7 different games have resulted so far. These seven games represent three kinds of collective action problems: problems of efficiency (welfare, optimality), problems of coordination, and problems of distribution. This raises the question of how many different strategic constellations exist, and how many and which
COLLEC T IV E AC T ION PROBLEMS
139
combinations of collective action problems they reveal. In section 6.2, matrix games will therefore be classified according to some formal attributes. This formal classification produces a typology of collective actions problems. One more general remark is appropriate: In the earlier illustrations, the models are valid only for the collective of the (two) players, and not for the general public. In the case of credit ratings, these were the investors or the borrowers; in the cases of lake and river pollution, they were the riparian states, and in the case of tax coordination, they were the EU member state governments. What constitutes a harmony game (or a prisoners’ dilemma) for the collective of players need not be desirable (or undesirable) for the general public or for any other group of actors. In the classical example of the prisoners’ dilemma, it is the prisoners who face the dilemma of confessing. For society as a whole, however, it is desirable that the prisoners confess. This is the reason why chief witness rules have been introduced as an institution. There are similar examples in the cases presented earlier. If the borrowers play a pure conflict game that leads them to provide ratings, they offer a common good to the investors. If the riparians of a river play a rambo game in which the dominant strategy of all riparians is to pollute, this causes negative externalities for the wider public. If capital income taxation is a common good for European governments, which they can achieve by harmonization, a “cartel” of these governments is by no means in the interest of the European capital income taxpayer. Thus, the models should be read carefully with a view to which collective is to play the game. In modeling a situation of common goods provision one must ask: Which collective is something a common good for? Which collectives are concerned by the situation? The collectives, which are then in fact modeled, depend on the research question. If the aim is to include all relevant actors, this might mean that a game between borrowers and investors or between the riparians and the other people who suffer from the negative externalities has to be modeled. 6.1
Properties of Common Goods and Strategic Constellations
As mentioned, the analysis of each and every combination of properties is both impossible and pointless. However, single factors can be systematically varied, while other factors are kept constant. This is possible for four of the five properties examined. Social and
140
T R A N S N AT I O N A L C O M M O N G O O D S
institutional factors are so manifold and different that no systematic variation is possible on this very broad dimension. This would still be possible if a certain institutional rule were selected as a factor, which could take several “values.” The other four factors are sufficiently specific to allow the systematic variation of the respective dimension. Not all values in these dimensions have been taken into account in the illustrations earlier, nor will they be taken into account in the conclusions in the following text. Only three variations of demand-side properties have been looked at, although the related properties often come in degrees and not as distinct classes. Only three forms of aggregation technology have been examined, although there are many more possibilities. Similarly, costbenefit configurations are manifold and depend very much on the actual problem to be modeled. Only those producing critical differences in strategic constellations, given a certain combination of basic properties, will be looked at below. In the following, the results of systematic variation of the four factors will be summarized. While one property is varied, the others are given a constant value. These values are selected to be as “neutral” as possible, meaning that they are set to a standard assumption, to a value more “natural” than some other variation, or to an extreme case. The values selected are consistent with the usual assumptions in common goods analysis. They are: 1. The cost-benefit configuration is set to the standard condition that leads to a prisoners’ dilemma, given a pure public good with summation technology and homogeneous actors, that is: 2b . c . b. 2. The demand-side properties are nonrivalry and nonexcludability— that is, a pure public good is assumed. 3. The production function of the good follows the summation technology. 4. The actors are fully homogeneous. 5. No specific social or institutional factors apply to the situation of common goods provision. Cost-benefit Configurations Which strategic constellation yields the variation of different costbenefit configurations, given that conditions (2) through (5) apply? In section 3.1 the cost-benefit configuration was varied in reference to the example of credit ratings. This example has two properties that
COLLEC T IV E AC T ION PROBLEMS
141
differentiate it from these conditions. First, credit ratings are not a pure, but a marketable public good, with excludable benefits. Second, the provision of credit ratings was (implicitly) assumed to follow a bestshot aggregation technology: Each of the investors is able to provide the good on his or her own. Thus, the resulting games in section 3.1 are partly different from the ones based on the earlier conditions for the same cost-benefit configurations. Four cost-benefit configurations are distinctive, in the sense that they lead to different game structures and outcomes. These are: c . 2b . b 2b . c . b c5b b.c Table 6.1 gives the payoff functions based on conditions (2) through (5), and the preference orders of the outcomes for the four cost-benefit configurations. The strategies are to provide (P) or not to provide (nP) a unit of the good. The ordinal payoffs for all four costbenefit configurations are given in columns (1) to (4). Cost-benefit configuration (1), where the individual contribution costs are higher than the individual benefits, even if both provide a unit, yields a harmony game. The dominant strategy for both actors is not to provide the good, for it is not beneficial for them. Cost-benefit configuration (2), in which the individual contribution costs are higher than individual benefits, but lower than the total benefits if both players are providers, leads to a prisoners’ dilemma. The good is not provided, although it would be beneficial. In configuration (3) individual costs and benefits are the same. Both actors are indifferent which strategies to choose. This leads to the so-called degenerate coordination game, a game with four Nash equilibria. Either of these equilibria could be the real outcome. Thus, provision, nonprovision, and unilateral provision of the good are possible. In configuration (4) the individual benefits are higher than the individual costs of contribution. This leads to a harmony game where the dominant strategy for both actors is to provide the good. Thus, it is only cost-benefit configuration (2) that produces a collective dilemma. If individual benefits are higher than individual costs, the good is harmoniously provided. If individual costs are higher than the total benefits achievable given that all contribute, the good is harmoniously not provided. The following variations rest on the assumption of configuration (2).
142
T R A N S N AT I O N A L C O M M O N G O O D S
Table 6.1 Cost-benefit Configurations in Common Goods Provision Assumption (1) Assumption (2)
Assumption (3) Assumption (4)
c . 2b . b 2b . c . b
Strategy Benefits Costs Payoff Combination
Ordinal, configuration
(1) Actor A
c5b b.c
(2)
(3)
(4)
A: P
B: P
2b
c
2b2c
2
3
2
4
A: P
B: nP
b
c
1
1
1
2
A: nP
B: P
b
0
b2c b
4
4
2
3
A: nP
B: nP
0
0
0
3
2
1
1
All factors are identical for country B.
Game Matrices (1)
Actor A
Actor B P
nP
P
2, 2
1, 4
nP
4, 1
3, 3
(2)
Actor A
Actor B P
nP
P
3, 3
1, 4
nP
4, 1
2, 2
Harmony
(3)
Actor A
Prisoners’ dilemma
(4)
Actor B P
nP
P
2, 2
1, 2
nP
2, 1
1, 1
Actor A
Actor B P
nP
P
4, 4
2, 3
nP
3, 2
1, 1
Degenerate Coordination
Harmony
Demand-side Properties: Nonrivalry and Nonexcludability Three combinations of demand-side properties have been varied in section 3.2: nonrival and nonexcludable goods (nonexcludable credit ratings, systemic risk), nonrival but excludable goods (excludable credit ratings), and rival but nonexcludable goods (environmental
COLLEC T IV E AC T ION PROBLEMS
143
pollution). While the comparison between pure public goods and common pool resources (CPRs) in the cases of systemic risk and environmental pollution rests on the earlier conditions, the comparison between excludable and nonexcludable credit ratings is a little different, as credit ratings follow the best-shot aggregation technology. If a marketable public good is modeled on the basis of a summation technology, this does yield the same result—namely an assurance game. Thus, the results obtained in section 3.2 can simply be summarized here. Given the combination of conditions summarized earlier, the provision of a pure public good puts the players into a prisoners’ dilemma. The provision of a marketable public good leads to an assurance game. This is a plausible result, as it is usually possible to solve coordination games that do not involve conflict (like assurance or pure coordination games) without the intervention of an exogenous power, provided there is some mechanism that can coordinate the strategies. This is why these goods are marketable. Finally, the exploitation of a common pool resource is a prisoners’ dilemma in the second phase of exploitation, where individual marginal benefits are still positive, but collective marginal benefits are negative. The demand-side properties are responsible for two effects. The property of nonrivalry allows for cost sharing and the common use of a good. This is basically a positive property, although the players may wait for others to provide the good or for some signal that tells them others will also pay their share. As long as nonrivalry is not combined with nonexcludability, this does not cause a very large problem. Nonrival goods provided by nature do not imply any problem. The examples of marketable public goods and club goods show that these goods can be provided in a noncooperative environment. The property of nonexcludability allows for free riding. This is the most problematic property in the production of common goods. Whether it is combined with rivalry or with nonrivalry, it leads to collective dilemmas, which can only be solved by cooperation, that is, by concluding binding and enforceable agreements. Supply-side Properties: The Aggregation Technology Three extreme forms of aggregation technology have been varied in section 3.3: summation, weakest-link, and best-shot technologies. In all three examples, conditions (1), (2), (4), and (5) were fulfilled. Thus, the results of section 3.3 need only be summarized here: The
144
T R A N S N AT I O N A L C O M M O N G O O D S
summation technology yields a prisoners’ dilemma; the weakest-link technology an assurance game; and the best-shot technology a chicken game. More, generally, if contributions are fully additive, the result is a dilemma game. If there is some upper or lower threshold for contributions, a coordination game will arise. Heterogeneous Actors within the Group There are many forms and many causes of heterogeneity. In chapters 4 and 5, actors with heterogeneous preferences (tax coordination, regulatory competition, credit ratings) and heterogeneous capabilities (tax competition, pollution of international rivers) were examined. There was no systematic variation relating to the form or the cause of heterogeneity. Rather, the situations with homogeneous and heterogeneous actors were compared. The following result was reached: The introduction of heterogeneity tends to transform symmetric games into asymmetric ones. In the cases of river pollution and regulatory competition, rambo games resulted. These are basically asymmetric harmony games. In case of tax coordination, heterogeneity has led to asymmetric dilemmas, irrespective of the cause of the heterogeneity (preferences or capability). However, there were also examples in which games with heterogeneous actors proved to be perfectly symmetric. In the case of credit ratings, heterogeneous borrowers ended up in a pure conflict game and a harmony game—both symmetric. In the regulatory competition case, heterogeneous states played a symmetric assurance game. The type of game that results if players are heterogeneous depends on the game played, if the actors were homogeneous, and on exactly how the players are heterogeneous. More precisely, it depends on which kinds of different preference orders are combined in a game involving heterogeneous players. As each kind of preference order of the four outcomes can be combined with every other kind, there are as many possibilities as there are strategically distinct asymmetric games in ordinal terms. I will not go on to present all the games that imply ordinal heterogeneity. Instead, I will examine what happens to some symmetric games if heterogeneity is introduced. These games are the prisoners’ dilemma, the symmetric harmony game, the assurance game, and the chicken game. These obviously play a role in common goods provision, as their recurring appearance in the applications has shown. Table 6.2 gives the preference orders of the homogeneous players for the four games.
145
COLLEC T IV E AC T ION PROBLEMS
Table 6.2
Preference Orders for Four Symmetric Games Strategy Combination
Actors A and B
Harmony
Prisoners’ Assurance Chicken Dilemma
A: P A: P
B: P B: nP
4 3
3 1
4 1
3 2
A: nP
B: P
2
4
3
4
A: nP
B: nP
1
2
2
1
There are six possible combinations of preference orders (“hybrids” of games, Taylor 1987: 39), which can be formed by these four games and which are strategically distinct.1 The matrices in table 6.3 show which games the six hybrids represent. The following combinations of the preference orders in the four symmetric games produce the six hybrids: 1. 2. 3. 4. 5. 6.
Harmony and Assurance Harmony and Prisoners’ dilemma Harmony and Chicken Assurance and Prisoners’ dilemma Assurance and Chicken Prisoners’ dilemma and Chicken
In three cases the hybrid games are rambo games, and the combination of an assurance game and prisoners’ dilemma (4) produces an asymmetric dilemma. This supports the claim that asymmetric games result if the players are heterogeneous. However, the combination of harmony and assurance (1) produces a symmetric harmony game, although it looks a little different from the standard harmony game. The reason for this is that neither harmony games nor assurance games involve conflict between the players’ preferences (see table 6.3). The combination of assurance and chicken (5) produces a game that has no Nash equilibrium: Players have an incentive not to coordinate their strategies, which is the reason why these games are called discoordination games here. This game has no stable outcome, thus, it poses a problem of instability. The Pareto-optimal outcome, in this case (P, P), will not be achieved in a noncooperative environment. I will come back to this in the next section. Basically, the heterogeneity of actors can lead to any of these kinds of game, depending on which kinds of preference orders are combined. The inequalities, which are involved in the actors’ heterogeneity and which are present in their initial positions in a game, are usually mirrored in the
146
T R A N S N AT I O N A L C O M M O N G O O D S
Table 6.3 Six “Hybrid” Games with Heterogeneous Players (1)
Player A
Player B P
nP
P
4, 4
3, 3
nP
2, 1
1, 2
(2)
Player A
Player B P
nP
P
4, 3
3, 4
nP
2, 1
1, 2
Harmony
(3)
Player A
Player B P
nP
P
4, 3
3, 4
nP
2, 2
1, 1
Rambo
(4)
Player A
Player B P
nP
P
4, 3
1, 4
nP
3, 1
2, 2
Rambo
(5)
Player A
Player B P
nP
P
4, 3
1, 4
nP
3, 2
2, 1
Discoordination Game
Asymmetric Dilemma
(6)
Player A
Player B P
nP
P
3, 3
1, 4
nP
4, 2
2, 1
Rambo
structure and the outcomes of the games: Most, but not all, games with heterogeneous actors are asymmetric. Moreover, in most but not all cases, there is inequality in the Nash equilibrium of the game. This is not true for games (1) and (4) in table 6.3. Asymmetric games can lead to equality in equilibrium, as in game (4); symmetric games can lead to inequality in equilibrium, as in the pure conflict game in chapter 5, where the low-risk borrowers had a distributive advantage. Finally, in most cases asymmetric games reveal inequality in the Pareto-optimal outcomes. This shows that there is a distributive problem involved. However, to add another complication, distributive problems are also implied in some perfectly symmetric games with homogeneous actors: The chicken and battle of the sexes games have two equilibria, and there
COLLEC T IV E AC T ION PROBLEMS
147
is distributional conflict among the players regarding which equilibrium should be chosen. Inequality will be treated more systematically in the next section. To sum up, the presence of a single property, such as the nonrivalry of a good or the heterogeneity of the actors, does not allow us to draw conclusions about the nature of the strategic constellation associated with this situation. It must be checked, whether other properties of the situation also affect the constellation and could thus transform the game. If a certain property appears in different combinations of other characteristics of the situation, it may or may not lead to different games, depending on how strong the influence of the respective property is. On the other hand, combinations of very different properties may lead to the same game structures. As has been noted, approximately 30 different combinations of properties in 5 dimensions have yielded only 7 different types of games in the applications. Two more have been added in this section: degenerate coordination, and discoordination games. In the following section, I will turn to the question of how many different basic strategic constellations exist and which kinds of collective action problems they represent. 6.2 Collective Action Problems: A Typology It is obvious that a large number of different games may result if situations of common goods provision are modeled as proposed here. One may start with a simple situation, such as the provision of an additive public good by homogeneous actors and find that the associated game is a prisoners’ dilemma. Each new property modeled will change the game. More precisely, it will change payoffs; but this need not necessarily result in a change in the strategic structure. However, the large number of individual games does not imply that there are just as large a number of strategically different games. Furthermore, the still large number of strategically different games belongs to a small number of game types, which each pose the same type of collective action problem. Taxonomies of Matrix Games The analysis of matrix games dates back to the very beginning of game theory. The founding fathers of game theory, von Neumann and Morgenstern (1943), started the discipline with the analysis of matrix games, especially zero-sum games. Nash’s definition of equilibrium, of mixed strategy equilibria, and of his proof that every matrix game
148
T R A N S N AT I O N A L C O M M O N G O O D S
has at least one mixed strategy equilibrium were major steps in the development of the theory of matrix games (Nash 1951). During the 1950s, 1960s, and 1970s, matrix games were widely discussed and used for experimental purposes in the social sciences (e.g., Rapoport, Guyer, and Gordon 1976; Rapoport 1988; Ledyard 1995). At the theoretical level not much has been added since then. In the modern theory of games, as well as in game theoretic applications in economics, matrix games are no longer important. In political science, however, matrix games are still used as an analytical tool. This is fully justified, since they are good models of conflict, coordination, and cooperation among rational individuals or quasi-rational populations. In their theoretical history, matrix games have been classified according to different attributes. In the beginning, the distinction between zero-sum and nonzero-sum games became important. Later the number of Nash equilibria (one, two, or more) served as a criterion. Also, basic types of collective action problem played a role from the very beginning: Games were classified as prisoners’ dilemmas, chicken, or battle of the sexes games (Taylor 1987; Kreps 1990: 37), as volunteers’ dilemmas (Rapoport 1988), or as coordination games. However, attempts to produce systematic taxonomies have been rare. Rapoport and Guyer (1966) and Rapoport, Guyer, and Gordon (1976) develop a taxonomy, which uses several solution concepts (dominant strategies, natural outcomes, Nash equilibria) and some factors that can make equilibria instable, such as threat-vulnerability. In this way they are able to rank matrix games according to the stability of their solutions. Although this dimension is important for theory and for the empirical testing of games, it is not of much relevance for applications in political science. Much more interesting for political science is a typology that allows different types of collective action problem to be distinguished. These are in fact the kind of distinctions that can be found in political science books (e.g., Schelling 1960: 88; Taylor 1987: 34–60; Scharpf 1997b: 69–79). Prisoners’ dilemmas, chicken, and assurance games are analyzed with respect to the social problems of coordination, cooperation, or inequality (Ullmann-Margalit 1977), and they are used to model empirical problems of the respective types. In political science a rather comprehensive attempt to classify 2 3 2 matrix games has been undertaken by Zürn (1992). Zürn distinguishes four types of games that represent what he calls “problematic situations”: coordination games without distributional conflict, coordination games with distributional conflict, dilemma games, and
COLLEC T IV E AC T ION PROBLEMS
149
rambo games. He uses three criteria for defining these types: the number of Nash equilibria, the Pareto-optimality of Nash equilibria, and the notion of a qualified Pareto-optimum, which serves to distinguish the rambo games (1992: 161–165). The classification presented here ends up with similar results, although on the basis of somewhat different criteria. In contrast to Zürn’s taxonomy, it is a complete classification of all ordinal 2 3 2 games. However, the theoretically relevant types it yields are similar. Only two aspects of Zürn’s taxonomy are contested: I argue that zero-sum games also represent problematic situations and that games without a Nash equilibrium are a class in their own right, with specific problems. What my own approach and the approach of Zürn have in common is a very wide understanding of “collective action problem” (section 2.3) or of a “problematic situation” (Zürn 1992: 154). I will use “collective action problem” here for all situations in which strategic action produces a collective situation that is considered problematic and requires a political or societal response to avoid suboptimality, coordination failure, distributional conflict, or instability. A Classification of Strategic Form Games The classification of matrix games developed here allows each game to be identified as a certain type of collective action problem. The aim of the classification is to find and distinguish all important types of collective actions problems relevant for political science insofar as they can be represented by simple strategic games. For this purpose it is necessary to classify all strategically distinct matrix games according to certain properties, which reveal or represent collective action problems. The number of different matrix games is infinite, if cardinal payoffs are used. This is true, even if the games are restricted to two players, two strategies and one-shot situations. Therefore it is impossible to tell how many individual games belong to one class of games. However, how can one be sure to include all relevant classes of games if there are an infinite number of games? First of all, the ordinal formulation of payoffs is sufficient to capture differences in the strategic structure. Games with cardinal payoffs can be reformulated into ordinal ones without a loss of information about the strategic structure. It is thus sufficient to classify all games with ordinal payoffs that are strategically different. In a 2 3 2 game with strict ordinal preferences there are four outcomes, which are ranked from 1 to 4 by the players. Each player can place
150
T R A N S N AT I O N A L C O M M O N G O O D S
the numbers into the four cells of the matrix in 4! 5 24 ways. Since there are two players, there are a total of 24 3 24 5 576 possible combinations of strict ordinal payoffs. Not all of them represent different strategic structures, however. The game matrices are strategically equivalent whenever only the players, the rows, the columns, or both rows and columns are interchanged. Taking this into account, only 78 strategically distinct games remain. If indifference about the outcomes or strategies is permitted, however, the number of strategically different games rises to 732 (Rapoport, Guyer, and Gordon 1976: 14–17). For the purpose pursued here it seems sufficient to work with the 78 strictly ordinal games and to treat them as representatives of 2 3 2 games in general. However, some interesting “borderline” games result if the players are allowed to be indifferent about the outcomes. Therefore, I include some prominent games that are based on weak preferences. I have analyzed all these games with respect to a number of game theoretic properties: symmetry or asymmetry, the number of Nash equilibria, the number of dominant strategies, the number of Paretooptimal outcomes, Pareto-optimality at equilibria, conflict over outcomes, inequality of payoffs in Pareto-optimal outcomes and equilibria, and the Kaldor criterion.2 The classification was developed on the basis of the results of this tentative analysis. Only four of these formal properties have proven to be crucial for a distinction between basic collective action problems. These four properties of games and their equilibria are used firstly to classify matrix games and secondly to develop a typology of collective action problems. The properties can be easily identified for each game. The classification is empirically complete insofar as all 2 3 2 games can be subsumed under the disjunctive classes defined by the following four factors: Number of Nash Equilibria in Pure Strategies (none, one, two, or more) I have chosen to use the number of Nash equilibria in pure strategies as a criterion because mixed strategy equilibria seem empirically to be highly implausible solutions as long as one-shot games are considered. Pure strategies are what real actors will have in mind in the first place, and it is the number of pure strategy equilibria that determines whether problems of coordination or discoordination arise. If I had chosen to count the number of pure and mixed strategies, the distinct classes created would have remained the same anyway. The
COLLEC T IV E AC T ION PROBLEMS
151
difference would be that the categories of this factor would have to be relabeled as one mixed, one pure, three or more pure, or mixed strategy equilibria. Pareto-optimality of the Equilibria (optimal, suboptimal) The Pareto-optimality and suboptimality determines whether an outcome of the game entails conflict between individual and collective rationality. In cases of multiple pure strategy equilibria, it is distinguished whether all equilibria are Pareto-optimal or whether there are optimal and suboptimal equilibria. This criterion cannot be applied to games that do not have equilibrium in pure strategies. Conflict over the Valuation of the Outcomes (no, irrelevant, partial, pure) The conflict factor refers to the players’ valuation of the four possible outcomes. No conflict means that the players rank all outcomes equally. Pure conflict means that the players’ ranking of the outcomes is completely oppositional. In ordinal terms, the latter games are constant-sum games, in which the constant is 5. Zero-sum games would be strategically equivalent in a cardinal game formulation. The group of games that can be classified as pure conflict games, however, is not restricted to constant- or zero-sum games, as the cardinal formulation allows for nonlinearity.3 Conflict over outcomes is irrelevant if both players rank the same outcome first, and different valuation occurs consequently only with respect to the second, third, or fourth preference. For all other games, conflict is partial. Usually games in the last group are called “mixed motive” games. Equal Payoffs in Pareto-optimal Outcomes or Equilibria (yes/yes, yes/no, no) The previous three factors constitute common ground in matrix game analysis. They have long served as criteria for the classification of games, and the strategic relevance of these attributes is obvious. The factor of inequality introduces a distributional dimension into the typology, which has thus far only been taken up in the Zürn classification. The factor is interesting from a political science perspective, since the distributional inequality of the outcome of a game may produce incentives to negotiate or to defect, even if the solution is rational (i.e., if it is an equilibrium). This has convincingly been shown by economic experimental research on fairness, beginning with “ultimatum bargaining” experiments (Güth, Schmittberger, and Schwarze
152
T R A N S N AT I O N A L C O M M O N G O O D S
1982; Fehr, Kirchler, Weichbold, and Gächter 1998; Fehr and Schmidt 1999; Güth, Huck, and Müller 2001). Equality of payoffs in the equilibrium is measured at the ordinal level. Ordinal comparisons of equality are problematic, however. The rejection of outcomes by players who judge them to be unfair is usually based on the inequality of cardinal payoffs. Outcomes, however, which have the same rank in the preference order of both players, may well be associated with very different cardinal payoffs, and vice versa. If, in a battle of the sexes game, for example, the female gets her preferred equilibrium, it may well be that the male’s overall cardinal utility is higher, as he values both spending an evening together and the female’s preferred entertainment more than she herself does. On the other hand, it is not unreasonable to assume that players also judge an outcome as unfair where player 1 gets her first preference while player 2 gets only her third preference. This is implied in the usual interpretation of battle of the sexes games or other games with ordinal inequality in the equilibria. Implicitly, we seem either to assume that there is the same cardinal utility associated with the same rank in the preference order, or we also evaluate distributional equality on the basis of ordinal comparison. Nida-Rümelin (1991), does so, for example, when he compares the outcomes of matrix games with ordinal payoffs in respect to justice. I will follow this tradition and work with the concept of ordinal equality in classifying the 78 strategically different ordinal games. It should be clear, however, that the equality criterion makes more sense when applied to cardinal payoffs. For the classification, I have used the equality criterion in the following way: there is equality if both players rank an outcome the same; there is inequality if they rank an outcome differently. Since I am concerned with what makes a game problematic, I am interested in the equality or inequality of outcomes that would be collectively desirable or that form an equilibrium. If at least one player has a strong preference for equal (fair) outcomes, a Pareto-optimal outcome may not be achieved, or an equilibrium may even be rejected by one of the players because it is characterized by inequality. Games are not problematic with respect to distributional justice if there is a Pareto-optimal equilibrium, which is associated with equal payoffs. They may become problematic, however, if the Pareto-optimal equilibrium is characterized by inequality, or if they have no equal Paretooptimal outcomes at all. Thus, two questions have to be asked: First, is there a Pareto-optimal outcome with equal ordinal payoffs (yes or no)? Second, if yes, is this outcome a Nash equilibrium (yes or no)? This creates three groups of games: Those with Pareto-optimal and
153
COLLEC T IV E AC T ION PROBLEMS
Table 6.4
Pure Coordination and Matching Pennies
Player A
Player B
1
2
1
2, 2
1, 1
2
1, 1
2, 2
Pure Coordination
Player A
Player B
Head
Tail
Head
1, 0
0, 1
Tail
0, 1
1, 0
Matching Pennies
equal Nash equilibria (yes/yes); those with Pareto-optimal equilibria, which are characterized by inequality (yes/no); and those without any equal Pareto-optimal outcomes (no). The classification presented in table 6.5 is the product of these four factors. The complete product would include 72 cells. However, not all categories of each factor are applicable to all categories of all the other factors. For example, the two equilibria-related criteria cannot be applied to games without equilibrium in pure strategies. Other subclasses are empirically empty; for example, in no conflict or pure conflict games equality does not vary. Thus, several rows and columns can be omitted and only 30 cells remain. In this way, the classification becomes more informative and remains at least as fine. In table 6.5, factors (1) and (2) are given in the vertical dimension and factors (3) and (4) in the horizontal dimension. The table shows how the 78 strict ordinal games are distributed over the 30 cells of the classification. In addition, some prominent examples of each important class of games are given. The games in brackets are the ones that are based on indifference to outcomes or strategies. For example, “degenerate coordination,” is a game where both players are indifferent to their strategies. Two other well-known games that involve indifference are shown in table 6.4: the pure coordination game and “matching pennies.” Typology of Collective Action Problems The classification of matrix games makes it possible to distinguish between some basic types of collective action problem. Each type is defined by a certain combination of the four matrix game properties analyzed. Although the table has 30 cells, there are only 7 types interesting for collective actions problems. First, this is due to the fact that some of the cells are still logically or empirically empty. Second, not
Table 6.5 Typology of Collective Action Problems Conflict
No 3
Irrelevant 18
Equality
Yes/Yes 3
Yes/Yes 18
2 “pure harmony”
13
Nash equilibria in pure strategies: Number
Paretooptimality
Unique
Pareto-optimal
57
53
HARMONY
Partial (mixed motive games) 54 Yes/Yes 7
7 “weak harmony”
Yes/No 7
No 40
No 2
3
26 “rambo”
2 “constant-sum”
DISTRIBUTION PROBLEMS 1 “prisoner’s dilemma”
Suboptimal 4
Pure 3
3
DISTRIBUTION PROBLEMS
DEFECTION PROBLEMS Multiple 12
Pareto–optimal 6
(“pure coordination”)
1 “chicken”
COORDINATION PROBLEMS Both optimal and suboptimal 6 None
(not applicable)
1
5
PURE CONFLICT
“battle of the sexes”
DISAGREEMENT PROBLEMS
(“zero-sum”)
5 “assurance” (“degenerate coordination”) 6
1
INSTABILITY PROBLEMS
(“matching pennies”)
2
9
COLLEC T IV E AC T ION PROBLEMS
155
all factors produce crucial distinctions in each of the subclasses constituted by the other factors. Thus, several cells can be subsumed under one type of collective action problem. The types are shown in table 6.5 by bold frames. They have been labeled in accord with the dominant problem of a certain class of games. Sometimes classes of games combine several elementary collective action problems—for example, a coordination plus a distribution problem—and thus form a new characteristic type. In the following, definitions and descriptions of the seven types are given. The number of strategically different games of the 78 ordinal games belonging to each type and each subclass discussed is given in brackets. Harmony The first type to be distinguished is the group of harmony games. They pose no collective action problem at all. The have a unique Nash equilibrium and thus a stable solution, which is Pareto-optimal and maximizes joint gains. There is no conflict concerning the valuation of outcomes (or conflict is irrelevant), and there is no inequality in the players’ payoffs in the equilibrium. From a collective action perspective, the distinction between “pure” harmony games (2), where all outcomes are ranked equally by both players, and “impure” harmony games (13), where outcomes off the equilibrium path are ranked differently, is irrelevant. “Weak” harmony games (7) involve partial conflict and have two Pareto-optimal outcomes. However, these games are harmony games: that is, there is no inequality in a Pareto-optimal equilibrium. Mere Distribution Problems On the right-hand side of the harmony games, there is a group of games that have a unique and Pareto-optimal Nash equilibrium. They differ from harmony games only in that they possess a second Pareto-optimal outcome (which is not an equilibrium), and they do not maximize joint gains. There is partial conflict over the valuation of outcomes. In the first group, there are three games, in which there is inequality in the Pareto-optimal Nash equilibrium. In the second and larger group (26), each of the two Pareto-optimal outcomes (equilibrium or not) is characterized by inequality. Both groups pose a distributional problem. It is merely distributional since, from the perspective of coordination, stability, and welfare effects, there are no problems with these games. They have been labeled “rambo” games by Zürn (1992: 209). So far, they have not received much attention. There are two reasons why this presents a significant
156
T R A N S N AT I O N A L C O M M O N G O O D S
deficit in the political science literature: First, from an outside normative perspective, the inequality may be judged to be unfair. Second, from a positive perspective, if the strict game theoretic assumptions are given up, there may be an “irrational” incentive to defect, driven by the players’ considerations of fairness, or an incentive to renegotiate if communication is possible, as has been shown by experimental economics. Rambo games are probably very common and should not be neglected. Pure Conflict Further to the right in table 6.5 there are pure conflict games. In pure strategies they may have no equilibrium or just one equilibrium, and their equilibria (as all other outcomes) are by definition Paretooptimal. Constant-sum games and zero-sum games are well-known examples. They pose distributional problems because their equilibria inevitably entail inequality between the players. In the cases of “no equilibrium” instability is a problem. Thus, these games intersect partially with distributional problems (2), partially with instability problems (1). Defection Problems The class of games below the rambo games entail a problem of defection (4). They have a unique Nash equilibrium. However, this is a suboptimal equilibrium. There is partial conflict, and equilibria may result in either equality or inequality of players’ payoffs. The collective action problem associated with this group of games is a conflict between individual and collective rationality. The Pareto-optimal outcome (which would be better than the equilibrium not only collectively but also individually) is not achieved by the players, as there is an individual incentive to deviate should the other player conform. The players are tempted to defect. The most well-known example is the prisoners’ dilemma. Both the Pareto-optimal outcome and the Nash equilibrium are equality outcomes in the prisoners’ dilemma. However, there are also asymmetric dilemmas (3) in which the Pareto-optimal outcome is characterized by inequality. The distributional problems that go along with these games make the situation worse: Even if the players could conclude a contract and enforce it, it would be difficult for them to agree to the Pareto-optimal and Kaldor-efficient outcome because that outcome is associated with inequality. Moreover, the Nash equilibria in these games may be equality outcomes. If the actors strongly value equality, they have one more reason to choose the suboptimal equilibrium.
COLLEC T IV E AC T ION PROBLEMS
157
Coordination Problems Below the harmony games there is a class of games that pose a coordination problem. The reason for this is the fact that these games have multiple equilibria in pure strategies. Some of them are no conflict games (1), and some are irrelevant conflict games (5). Both may have two Pareto-optimal equilibria, as in the case of the “pure coordination” game. The “impure” coordination games have both an optimal and a suboptimal equilibrium, such as the assurance game or the degenerate coordination game. There is no inequality in their equilibria, and they maximize joint gains in the optimal equilibria. The main problem for the players is to coordinate their strategies such that at least one of the equilibria, and hopefully the optimal equilibrium, will result. Disagreement Problems The next class, to the right of the coordination games and below the defection games, combines several problems. Games belonging to this class have two equilibria in pure strategies, and both strategies are Pareto-optimal. However, both equilibria entail inequality. Thus players disagree about which equilibrium is to be selected. In its ordinal formulation the chicken game does not only have two Pareto-optimal Nash equilibria—both of which are inequality outcomes—it also has a third Pareto-optimal outcome, which is not an equilibrium, but which shows equality. The other five games in this class, including the battle of the sexes, possess only two Pareto-optimal outcomes. There is no problem with collective welfare in these games, but there is both a coordination and a distribution problem. Given this combination, the players have difficulties of finding an agreement. Both players prefer a different equilibrium, and there is thus a good chance that they will miss the equilibria. This is even true if both players act altruistically and pursue the other player’s preferred equilibrium. If the players can communicate and thus eliminate the coordination problem, they will still have difficulties agreeing on one of the equilibria. Instability Problems The last group is formed by discoordination games. These games have no pure strategy Nash equilibria. Therefore the criterion of Pareto-optimality cannot be applied. The main problem these games pose is instability. In these games, rational players generally want to prevent their strategies from intersecting—they discoordinate strategies. There is always a unilateral incentive to deviate from
158
T R A N S N AT I O N A L C O M M O N G O O D S
Pareto-optimal or joint gain-maximizing outcomes. Most of these games are partial conflict games (8); however, they may appear in pure conflict situations as well (1). A classical example that allows for indifference is “matching pennies” (table 6.4). As mentioned earlier, matrix games have many more strategically relevant attributes: for example, symmetry or asymmetry; the Kaldor criterion (allowing evaluation of whether joint gains are maximized or optimized in equilibrium); the number of Pareto-optimal outcomes; the number of dominant strategies; or the exact extent of conflict. In checking the sample of 78 games for whether they make any systematic and important difference and whether they create new and interesting subclasses of games, these attributes have been included in the analysis. They make no such systematic difference. Still, the additional criteria display some regularity. Two examples shall be given. First, joint gains are maximized only in the Paretooptimal equilibria of no conflict games and irrelevant conflict games. For all other games, joint gains are never maximized, although there are sometimes equilibria that optimize joint gains. This is true, for example, for the equilibria of pure conflict games. Second, symmetry is only very weakly correlated with classes of games. All no conflict games (3) and all pure conflict games (3) are symmetric; a great number of irrelevant conflict games are symmetric (11 out of 18), while only a small number of partial conflict games are symmetric (17 out of 54). It is especially noteworthy that there is no correlation between asymmetric games and inequality in Nash equilibria. The distribution of the 78 strictly ordinal preference games throughout the 7 classes shows that the number of strategically different games in the various problematic classes is significantly higher than the number of games in the harmonic classes. There are altogether 22 pure, impure, and weak harmony games. There are 3 games that pose only moderate distributional problems, and 26 more severe rambo games, 4 defection games, 6 coordination games, 6 disagreement games, 8 discoordination games, and 3 games of pure conflict. Thus, 56 of 78 strategically different games cause some sort of collective action problem. However, only four different dilemmas have been found. This result must be qualified, however. First, it is not clear what the distribution would look like if all 732 strategically different games were classified, including the indifference games. Second, and more importantly, the empirical relevance need not at all be correlated with the theoretical distribution of strategically different game models throughout the types of collective action problem. As of yet, not
COLLEC T IV E AC T ION PROBLEMS
159
much can be said about the empirical frequency of the collective action problems. It is true that representatives of most classes of collective action problems seem to have been frequently analyzed and applied—with the exception of harmony, rambo, and instability games. The fact, however, that the prisoners’ dilemma has received a great deal of attention in social science literature, while rambo games have not, does not make it possible to draw conclusions about the empirical relevance of the two games. This fact is probably a mere result of researcher’s perceptions and interests, and of the greater logical or normative attractiveness of the prisoners’ dilemma. Moreover, systematic information on the empirical frequency of the different games is difficult to acquire.
This page intentionally left blank
Part II
Strategic Effects of Multi-level Provision of Common Goods
This page intentionally left blank
Chapter 7
Transnational Common Goods and Multi-level Systems: Analytical Framework
The empirical illustrations used in this book are global common
goods in some cases, transnational common goods in most cases, and transboundary common goods in all cases. The analysis carried out in the previous chapters has not devoted any special attention to the fact that the respective goods are transboundary, transnational, or even global. The transboundary nature of goods (crossing jurisdictional borders) implies that they will be provided either by a governance structure at a central level, if such a central level exists, or by a system of multi-level governance. The analysis thus far treated the provision of these goods as single-level problems, however. Only one level of decision making has been looked at. The multi-level structure of the provision of transnational common goods is the subject of this part of the book. In chapter 7, the foundations for the analysis are laid. As this book deals primarily with attributes affecting the strategic constellations, the following questions will be addressed: Is the multi-level provision of a common good different from single-level provision of the same good? Are strategic constellations affected by the fact that goods have to be provided in multi-level governance structures, and if so, in which ways are they affected? The analysis presented here will be restricted to investigating the conditions, which may change the basic strategic constellation and the solution potential as a consequence of multi-level provision of the good. One particular fundamental and important aspect will be highlighted; namely, how the homogeneity or heterogeneity of actors at the various levels affects common goods provision.
164
T R A N S N AT I O N A L C O M M O N G O O D S
In section 7.1, the notions of transnational and global common goods is defined and the problem of the incongruence of national borders and transnational common goods is discussed. Section 7.2 provides a brief account of theories of multi-level systems and introduces the definition a multi-level system, as it will be used in the remainder of the book. In section 7.3, the possible combinations of homogeneity and heterogeneity of actors across the levels of governance are examined. 7.1
Transnational Common Goods and National Borders
This section starts defining transnational and global common goods. Next, the problem of the incongruence of national borders and transnational common goods is discussed, and options for dealing with the problem of incongruence and transnational spillovers are presented. Incongruence can have different causes. Most importantly, it can appear as a consequence of conditions given by nature or as a consequence of political decisions. Finally, the basic strategies available to overcome the problem of incongruence will be introduced: centralization, cooperation, and mutual adjustment. Only the strategy of cooperation creates multi-level systems. Definition of Global and Transnational Common Goods Kaul, Grunberg, and Stern define global public goods in the following way: “A pure global public good is marked by universality—that is, it benefits all countries, people and generations” (1999: 11). This definition is very demanding. It does not focus on the character of “something” as a good, as opposed to a bad, nor on the publicness of a good. The definition expands to include what makes a public good a global public good. It does require that all countries value “something” as a good, not that only a group of countries or a certain region do so. Moreover, it requires that all social groups within each country benefit from the respective good. Finally, it requires that the good meets the needs not only of present, but also of future generations. This concept, in fact, defines a universal public good. The definitions I use here are less demanding. Since preferences for a common good may be heterogeneous, it is not necessary that everybody values “something” as a good to call it a common good. What is assessed as a common good by some may be assessed as a
M U LT I - L E V E L S Y S T E M S
165
common bad by others. It is important, however, that the basic characteristics of a common good are present, that is, there are positive or negative externalities. A common good is called global, if the spatial scope of its externalities is global. The collective concerned by the good is scattered over the whole globe—even if not everybody in the world is convinced that the good is really a good. A common good is called transnational, if its externalities reach more than one nation-state. A transnational common good may concern only two states, a larger region, or even the whole globe. Thus, global common goods are a subset of transnational common goods. In short: A common good is transnational or global, if activities in one state cause positive or negative external effects, which are felt in other states or even over the whole globe. Most of the examples used in this book relate to global common goods. This is true for global warming, biodiversity, and systemic risk in banking. Others are transnational, such as international lake and river protection, and one is transboundary relating to jurisdictions at a lower level than the nation-state, namely the example of the siting of unwanted facilities. The Scope of Common Goods The presence of global, transnational, and transboundary common goods implies the incongruence of the functional scope of political problems and the political territories. This incongruence has undesirable consequences. First, as a result of the transboundary structure of the problem, policies cannot be implemented effectively. Second, spatial externalities lead to losses in collective welfare. Furthermore, democratic legitimacy is questioned, as the democratic condition of the identity of the citizens effected by a political decision and the decision makers is violated (Zürn 1996: 39). Therefore, it is desirable that the congruence of functional scope and territorial jurisdictions be restored. Basically, there are two possibilities for adjustment: Either the functional scope of a common good is adjusted to the territorial jurisdictions, or vice versa. At first sight it seems easier to adjust functional areas to political territories, as political territories are historically given and cannot easily be changed. However, there are limits to the adjustment of the functional scope, because that scope cannot be influenced by politics in each case. Empirically, the condition of congruence is often violated. The basic reason for this is the transboundary mobility of factors, such as wind, biomass, humans, goods, capital, and information. The
166
T R A N S N AT I O N A L C O M M O N G O O D S
mobility of these factors is responsible for the creation of single functional areas—natural, technical, economic, or cultural—which transcend political territories. Air, water, and pollutants have their own reach, which cannot easily be influenced by political decisions. The mobility of persons creates coordination problems, for example, with transport systems or with the recognition of school or university certificates. The mobility of goods, services, capital, and information creates coordination problems between territories that use different regulatory standards. Whenever it is difficult for those reasons to adjust the functional areas to jurisdictions, the only way of restoring congruence is by adjusting jurisdictions to the functional problems. It is thus necessary to clarify under which conditions the scope of a common good can be flexibly shaped, and which factors influence the adjustability of a common good. Two aspects of functional scope are to be differentiated: the spatial reach and the (optimal) size of functional areas. These two aspects are not independent. The size of a functional area can only flexibly defined if the spatial reach is adjustable at all. Most common goods will be situated on a scale between the two extreme possibilities, that is, of a policy that is completely independent of space and size, on the one hand, and a policy that is completely determined in its spatial scope and size by factors that cannot be influenced by humans, on the other. The adjustability of the functional scope of various common goods comes in degrees, because the factors that determine the scope are manifold. Sometimes, the scope of a good is adjustable in the short run; sometimes they can only be shaped over the long run. The spatial scope of common goods can be fixed, as they are in water protection, or they can be variable, as they are in the protection against air pollution, or they can be unpredictable, as in the case of epidemic diseases. In many cases, the scope of the good is influenced by more than one factor. For the following list of factors the degree of adjustability through political action increases from (1) to (5). 1. The functional scope of a common good is not adjustable when it is determined by natural, for example, geological, physical, or biological conditions. The reach of an earthquake cannot be influenced by humans. The same is true for many environmental common goods. For example, the pollution and the protection of a body of water pertain to a certain area and a certain population around it, such as the riparians of the Great Lakes or the Rhine.
M U LT I - L E V E L S Y S T E M S
2.
3.
4.
5.
167
The scope of air pollutants is determined by the exact nature of the pollutant, the directions of winds, and geological conditions. The functional scope of a problem can also be determined by technical facts. The reach of air pollutants is influenced by the height of smoke stacks. Another example is television: scope and adjustability is different for antenna, cable, or satellite technologies. There are other common goods where the scope cannot be arbitrarily shaped as a consequence of the self-perception of the collective concerned, that is, the ethnic, cultural, or linguistic community. For example, the reach of the contents of media is not only limited by technology, but also by language. The reform of the German orthography concerns more than one nation-state— Germany, Austria, and parts of Switzerland. In these cases, there are historical path dependencies, which cannot be changed in the short run. The scope of common goods can also be determined by earlier political decisions. Examples are the regulation of the heights of smoke stacks, or the prohibitions against settling in areas that are endangered by earthquakes. In this context political decisions are particularly important that affect the mobility of factors, such as the decision in favor of the “four freedoms” in the European Union (EU). At the end of the scale are common goods that are fully spatially independent. The regulation of capital markets provides an obvious example. The flow of capital is not determined by anything other than the will of its owners. Capital markets can be arbitrarily shaped in size and spatial scope by political decisions.
Even if the spatial scope of a common good can be flexibly adjusted to territorial jurisdictions, this does not imply that the size of the territory and the respective collective of citizens can be chosen arbitrarily. There are common goods that can have a collective of any size. This is valid for criminal law, for example. In many cases, however, there is something like an optimal size for the good. Sometimes distances may play a role in determining the optimal size of a common good. This would be true for primary schools, which should be evenly distributed within the territory of a jurisdiction. In other cases, there may be economies of scale that are correlated with the size of the area or the size of the collective. For those common goods the size of jurisdictions has to be optimized. Waste management facilities or other public utilities are examples. Another example is the markets for goods and capital: Larger markets promise welfare gains, which is the
168
T R A N S N AT I O N A L C O M M O N G O O D S
reason for the creation of common markets, such as the EU or the North American Free Trade Agreement (NAFTA). The examples of common goods that have been used for demonstration in this book stem from two policy fields: environmental policy and financial markets. These fields belong to the two extreme categories of the scale of adjustability of functional scope to territorial jurisdictions. Whereas the scope of environmental common goods cannot be adjusted to territories in most cases (one exception is waste abatement), the scope of financial markets can be chosen arbitrarily. Thus, in the case of environmental problems, the restoration of congruence requires that jurisdictions be adjusted to the functional scope of the problems. In the case of financial markets, in principle, the spatial extension of the markets can be adjusted to the scope of territorial jurisdictions. The area in which capital is mobile can be limited to the nation-state, and thus the nation-state can regulate its capital market so that the common good is achieved and systemic risk is contained. However, the area can also be chosen to expand worldwide, over the territory of the EU, or else. The difference in the adjustability of the functional scope is the only systematic difference between the two policy fields chosen. This difference has no consequences for the strategic constellations or the types of collective action problem, as has become obvious in part I of the book. The property of (non) adjustability implies different solutions to the problem of incongruence. While for transboundary and nonadjustable common goods the cooperation of several jurisdictions is necessary, for transboundary and adjustable goods it is possible to “close borders” to avoid the problems of incongruence. In practice, however, the difference is less relevant than it might seem. In the financial markets, this is a consequence of political decisions. The globalization of the markets for goods, services, and capital is a result of a long series of political decisions of nations-states (Deeg and Lütz 2000: 374−375). The liberalization of capital markets was a process that arose from economic insights and political will. Obviously, the expected welfare gains from worldwide capital markets were judged by the governments to count more than the danger of negative externalities spreading around the world. The liberalized market, however, with its increased mobility, changed the functional scope of the common good and led to new regulatory challenges for the international community. If the nation-states do not want to fall back on national capital markets and sacrifice the potential welfare gains of a worldwide market, cooperation among the nation-states is required to guarantee the common good—exactly as for environmental common goods.
M U LT I - L E V E L S Y S T E M S
169
Therefore, if the scope of a common good is not adjustable to political territories, or if the scope of a common good transcends given political borders owing to the political will of the concerned jurisdictions, cooperation among these jurisdictions is needed to restore congruence. Differences in regulations in the jurisdictions lead to transboundary externalities in the subterritories within the functional area. Consequently, some of the territories can free ride at the cost of others. Some form of collective action and of common regulation is thus required to ensure that the transnational common goods can be provided. This presupposes the existence of a minimal governance structure for the whole functional area. 7.2 Theories of Multi-level Systems and Multi-level Provision of Common Goods This section starts with a brief account of the many theories related to political multi-level systems. This is not meant as a review of the literature. It rather points to whether and how some aspects that are important in the context of multi-level provision of common goods are dealt with in these theories—the problem of congruence and the problem of heterogeneity and distortion in the aggregation of political will over several levels. The section concludes with a definition of a multi-level system that will be employed in the following chapters. Theories of Multi-level Systems The terms “multi-level system” and “multi-level governance” are very often used in connection with the EU, so much so that “the European multi-level system” is almost a collocation now.1 In a wider view, however, the terms multi-level system and multi-level governance are not restricted to the political system of the EU, but are also used for federal systems (Kern 2000) and for international regimes (Putnam 1988). In the following a wide understanding of multi-level systems is employed. This includes all systems in which more than one level of governance structure is present and in which the lower levels have some say in the decisions of the upper levels of governance. It encompasses the EU, federal states, and international regimes. Three families of theories relate to political multi-level systems: First, federalism is the classical manifestation of a multi-level system. The literature on federalism is huge, both regarding theory and comparative research. Only very few theories are of relevance to the questions asked in this book. Second, there are a number of political
170
T R A N S N AT I O N A L C O M M O N G O O D S
science theories that are interested in the effects of multi-level systems (in particular, the EU) on politics and policy outcomes. Third, the economic theories of federalism are relevant, because they start from the idea of a common good. The literature on federalism encompasses both normative and positive theory and is interested in the function of federalism as well as in the description of its institutions. There are few explanatory approaches, however, and mostly these are concerned with the development of federalism. Even fewer theories deal with the effects federalism has on politics and policy making. Important exceptions are legislative federalism (Riker 1964; Rose-Ackerman 1981) and the theory of Politikverflechtung (Scharpf, Reissert, and Schnabel 1976; Scharpf 1985). Rose-Ackerman develops a formal model in which she shows that policy choices can in fact change as a consequence of a federal structure—as compared to a unitary system and given the same policy preferences of the citizens. Politikverflechtung describes the necessities of vertical and horizontal cooperation in a federal system. Scharpf (1985) has coined the term “joint decision trap” to describe the negative consequences of Politikverflechtung. Joint decision making in a two- or more level institutional decision-making structure can lead to systematically inefficient and inadequate decisions and to the incapability to change the institutional conditions that lead to these results. The joint decision trap implies the hypothesis that policy change and constitutional change is difficult to achieve in multi-level systems. Putnam’s theory of two-level games (1988) deals with multi-level interactions in the sphere of international relations. In two-level games, negotiators find themselves simultaneously engaging in both international and domestic bargaining. The two levels are linked through the necessity of ratification of international agreements by the domestic constituencies. Putnam derives a number of conjectures on the relationship between domestic actors, the negotiators, and international agreements. Two aspects are of importance here: First, the task of the negotiators at the international level is much easier if their constituency is homogenous than if it has heterogeneous preferences. Second, the negotiators interest may diverge from that of its constituency and this may place additional constraints on the possibility of international cooperation. More recently, a number of theories of multi-level governance in the EU have been developed. The state-centric and intergovernmental theories of European integration have gradually been complemented by the analysis of multi-level governance (Marks, Hooghe, and Blank
M U LT I - L E V E L S Y S T E M S
171
1996). A number of effects of multi-level governance—namely, on the opportunities to act, on the policy process, and on outcomes—have been described in the literature (Grande 2000): Actors can strategically use the constraints placed on them by their constituencies to gain bargaining power in the game at the upper level; they can, however, also use the constraints placed on them by their partners at the European level to reject demands from domestic actors (Putnam 1988). The preferences of their domestic constituents are dominant for the governments: They primarily represent the interests of their home country, and they neglect collective welfare at the European level (Benz 1994: 187–190). The multi-level environment provides actors with many more opportunities to form “advocacy coalitions” (Sabatier 1998). Multi-level systems allow problems to be shifted between the different levels of action (“cuckoo game”: Wassenberg 1982); they also make it possible to shift responsibilities between them (“blame avoidance”). Finally, it is typical for multi-level systems to have a large number of veto points. The more consensus is required, the more difficult it is to find an agreement. The joint-decision trap is thus also a risk for multi-level systems of the European type (Scharpf 1985). Economic theories of federalism are important in the context of this work because they explicitly refer to the efficient provision of common goods by the various levels in a federal system in the presence of heterogeneous preferences. There are three subspecies of economic approaches: fiscal federalism, competitive federalism, and functional federalism. The theory of fiscal federalism is concerned with the incongruence of the spatial scope of common goods and the territorial scope of political jurisdictions. It provides a normative solution by the principle of fiscal equivalence (Olson 1969; Oates 1972), telling us that the provision of a public good is efficient, if the users of the good, the taxpayers and the decision makers are identical. This implies a static allocation of legislative responsibilities to the adequate level of government. Whereas this theory allows for a high degree of centralization, the theory of competitive federalism (Tiebout 1956; Oates and Schwab 1988) maintains that decentralized common goods provision is the best solution. According to this theory, common goods are efficiently provided by competition between jurisdictions (systems competition). The approach of functional federalism gives up the principle of territorial jurisdictions altogether. It suggests the establishment of purely functional jurisdictions, independent from space and political borders (Straubhaar 1995; Frey 1996, 1997; Eichenberger 1996). Functional federalism takes the idea of avoiding
172
T R A N S N AT I O N A L C O M M O N G O O D S
spatial externalities from fiscal federalism and combines it with the principle of competition, taken from competitive federalism. These theories have one aspect in common that is important in the context of the research question posed here: The effects they describe presuppose the heterogeneity of preferences among actors within the multi-level system. The effects would not arise if the actors analyzed by the respective theory had homogeneous preferences. Heterogeneity may be present horizontally, among actors at the same level—for example, at the level of individual citizens or of member states in a federal state or in the EU. Heterogeneity may also be present vertically, between the levels, if there is a divergence of interests between a collective and its representatives at the next level above. If there were complete homogeneity of preferences, both horizontally and vertically, the effects of multi-level systems described by the various theories would simply not occur. The connection to heterogeneity is most obvious in the economic theories of federalism, where homogeneity with respect to preferences, costs, and exposition to externalities, serves as a criterion for the formation of territorial or functional jurisdictions. Fiscal and functional federalism require a centralized solution for each homogeneous collective; competitive federalism predicts that a centralized solution for each homogeneous collective will arise from systems competition. However, homogeneity and heterogeneity of interests are also present in the other theories. In Rose-Ackerman’s model, federalism matters just because the status quo in the member states may imply heterogeneous preferences between the lower level jurisdictions. The theory of Politikverflechtung is only relevant if the interdependent actors within the multi-level system have heterogeneous interests. Only if there is conflict of interests, will some actors veto agreements and will blockades arise in consensual negotiation systems. Finally, Putnam’s theory of two-level games includes two conjectures about heterogeneity. Also many of the strategic effects described in the literature on the European multi-level system rest on the assumption that the relevant actors have heterogeneous preferences and/or heterogeneous capabilities. Examples are the sidestepping of the national governments, strategic commitment on the basis of constraints by the constituency, blame avoidance, or the cuckoo game. In fact, any effect that a multilevel system can have on the provision of common goods—if compared to a single-level system—is based on the heterogeneity of preferences. Therefore, horizontal and vertical heterogeneity will be examined in more detail in chapters 8 and 9.
M U LT I - L E V E L S Y S T E M S
173
Definition of a Multi-level System The terms “multi-level system” and “multi-level governance” will be used here in a general and abstract sense. In particular, they will not be restricted to the EU polity. They are used for all systems, which consist of a pyramid-shaped hierarchy of multiple levels of governance structures where the lower levels participate in the decision making of the upper levels. The upper levels include the jurisdictions at the lower level, while jurisdictions at the same level have mutually exclusive territory and citizens. This corresponds to the notion of “hierarchical federalism” or “territorial federalism.” It is distinct from this notion, however, in that it is not restricted to polities, in which the jurisdictions at all levels have state quality. It includes international regimes, which are not states, but governance structures (Zürn 1997). The term “multi-level system” is also distinct from the notion of “multi-arena governance”: it is restricted to levels of governance structures, governmental actors, and territorial jurisdictions, and it does not include nonterritorial and nongovernmental arenas and actors (Héritier 2002b). What counts as a level? Because the aggregation of preferences over the levels shall be examined, the basic “unit” is the citizen. The level of citizens, however, does not count as a level of governance. It will thus be called level L0, while the levels of government will be called level L1, L2, and so forth. Low numbers signify low levels of government. A typical pyramid encompasses all or some of the following governance structures, to which a citizen belongs: the local community; the district or county; the state, Land, Kanton, département, province, or region; the nation-state; the regional organizations, such as the EU or the NAFTA; finally, “global” international organizations or regimes. Not all of these levels of governance, however, count as levels in a multi-level system as defined earlier, as some of them do not participate in the decision making of the next higher level. This is usually true for communities and districts, and also for regional organizations, which usually do not participate, as such, in international organizations or negotiations—although the EU increasingly represents its members at the international level. For the analytical purposes of the following chapters, it is sufficient to distinguish two levels of governance structures. Thus, a two-level system will be analyzed that includes the levels • L0 citizens, • L1 subnational jurisdictions or nation-states, and • L2 nation-states or international structures.
174
T R A N S N AT I O N A L C O M M O N G O O D S
The kinds of jurisdictions or collective actors that form levels L1 and L2 depend on the example. There are three levels of actors: individuals at L0, and collective actors at L1 and L2. 7.3 Multi-level Analysis: The Role of Homogeneity and Heterogeneity The question of whether multi-level provision of common goods matters in terms of the strategic constellation implies that there is a point of reference. Compared to which situations might multi-level provision differ? As transnational common goods by definition concern more than one jurisdiction, the obvious point of reference is single-level provision within the total collective comprised of the subcollectives in the jurisdictions. Single-level provision could mean the results of noncooperative provision among the collective as a whole, that is, all individual actors concerned by the reach of the common good. It could also mean a cooperative, single-level solution, that is, a centralized solution. Accordingly, two-level provision might be cooperative at L2, which implies that there is a governance structure at two levels. It might also mean a noncooperative solution at L2, which implies mutual adjustment or the competitive solution. The cooperative solutions at the upper level L2 are not examined in this chapter. However, the presence of a cooperative solution at L2 in a multi-level system may affect the actual outcome and change it, vis-à-vis a single-level outcome, for two reasons: First, the cooperative structure may change the preferences of the representatives at L2. Second, aggregation mechanisms used at level L2 may alter the outcome. These effects are analyzed in chapter 9. The idea of a multi-level system presupposes a cooperative governance structure at L1. The presence of such a governance structure implies that there are actors at L1 who represent the L0 constituency at L2. Thus, if the single-level strategic constellation is to be compared to the two-level strategic constellation, three groups of actors have to be considered: the total collective of L0 actors, the subgroups of L0 actors within each jurisdiction in the two-level structure, and the L1 representatives of the subgroups who interact at L2. For the following points, I will stick to the two-level system as defined in section 7.2. To keep the analysis simple, one can additionally imagine that there are only two jurisdictions at level L1 and thus only two representatives at L2. The individual actors at L0 can belong to two types, X and Y, each with a different preference for the common
M U LT I - L E V E L S Y S T E M S
175
good. If all actors in a group belong to the same type, the group is said to be homogeneous, if the actors are distributed over both types the group is said to be heterogeneous. The total collective at L0 is called “the collective,” the L0 actors within each jurisdiction are called “the subgroups,” and the actors belonging to the governance structures at L1 are called “the representatives.” To find out whether multi-level provision matters, the strategic constellations among the actors within the three groups will be examined and compared: • the strategic constellation among the L0 actors within the collective (that is, the single-level game), • the strategic constellation among the L0 actors within each subgroup (that is, the L1 game), and • the strategic constellation among the L1 representatives (that is, the L2 game). As was claimed in the earlier chapters and will be demonstrated in the following, it is the heterogeneity of actors, which may change the strategic constellation in a multi-level, compared with a single-level system of provision. Each of the three groups can be homogeneous or it can be heterogeneous regarding their preferences for the common good. Homogeneity or heterogeneity within the three groups can come in different combinations across the three groups. The presence of homogeneity or heterogeneity in the three groups is not fully independent. For example, even if all subgroups are internally homogeneous, the collective as a whole need not be homogeneous, as there can be heterogeneity between the subgroups. On the other hand, if the subgroups are internally heterogeneous, the total collective cannot be homogeneous. Furthermore, the representatives can be homogeneous, although their constituencies are heterogeneous, whenever the aggregation mechanisms used in the jurisdictions leads to the same preference at the level of L1 actors (e.g., when a majority in each subgroup belongs to the same type and the decisions are taken by majority rule). Finally, the representatives may not truly represent their constituencies because of overriding self-interest or because of distortion through the aggregation mechanisms. The possible constellations are explored more systematically, before I show, for each constellation, whether the strategic structure changes in two-level games vis-à-vis single-level games. With three groups there are 23 (5 8) possible combinations of heterogeneity and homogeneity. Table 7.1 shows all combinations. Two of them must be excluded for logical reasons: It is logically
176
T R A N S N AT I O N A L C O M M O N G O O D S
Table 7.1 Possible Combinations of Homogeneous and Heterogeneous Actors Case Subgroups
Collective
Representatives
Example
1 2 3 4 — — 5 6
homogeneous homogeneous heterogeneous heterogeneous homogeneous homogeneous heterogeneous heterogeneous
homogeneous heterogeneous homogeneous heterogeneous homogeneous heterogeneous homogeneous heterogeneous
Lake Pollution EU tax coordination European Monetary Union Global Warming logically impossible logically impossible Siting of LULUs River Pollution
homogeneous homogeneous homogeneous homogeneous heterogeneous heterogeneous heterogeneous heterogeneous
impossible that the actors are heterogeneous within the subgroups, but homogeneous within the collective as a whole. Thus, the respective lines must be deleted and we are left with six constellations: • In constellation 1, the actors within all three groups are homogeneous: All actors at L0 and the representatives are in the same position. The pollution of a lake provides the example for constellation 1. • In constellation 2, the L1 actors are heterogeneous, although all L0 actors are homogeneous. There is a divergence of interests between the subgroups and their representatives. This could be a result of aggregation or of self-interested representatives. An example for the latter possibility is tax coordination in the EU, because the interest of governments is different from the interest of taxpayers in this constellation. • Similarly, in constellation 3, there is a discrepancy of representatives at L2 and the collective. Although L0 actors are homogeneous within the subgroups, the subgroups belong to different types of actors. The representatives, however, are again homogeneous, which may result from aggregation or from organizational self-interest on the part of L1 actors. The creation of the European Monetary Union serves as the example here. • Constellation 4 is similar to constellation 3 in that there is heterogeneity of homogeneous subgroups of L0 actors. It is different, however, as the L1 actors truly represent the heterogeneity at L2. This situation can be illustrated by global warming or by many constellations of environmental regulation. • In constellation 5 both the subgroups and the collective consist of heterogeneous actors. The representatives are homogeneous in their preferences, however. Again, this may be a result of aggregation or
M U LT I - L E V E L S Y S T E M S
177
self-interested L1 actors. The siting of locally unwanted land uses (LULUs) provides the example. • In constellation 6 the actors are heterogeneous within all three groups. Examples are the protection of international rivers or global biodiversity protection. In the constellations 1, 4, and 6, the homogeneity or heterogeneity of L0 actors within the collective is mirrored within the L1 actors. Thus, the L2 game can ceteris paribus be expected to have the same strategic structure as the single-level game. Moreover, in constellations 1 and 6 the L1 game can be expected to be the same, as well. The effect of multi-level provision in these constellations will be demonstrated using the examples of lake and river pollution (section 8.1) and of global warming (section 8.2). In the constellations 2, 3, and 5, the homogeneity or heterogeneity of the collective is not represented in the L2 game by the L1 actors. Thus, the L2 game can be expected to have a different structure than the single-level game. In constellation 3, there should be a difference between the L1 games in different subgroups, as well. In these constellations, the multi-level structure in fact matters, meaning that it may lead to a different outcome compared to the single-level system. Two factors can account for this. First, deviation of the L1 representatives from their constituencies’ preferences can explain the change in the strategic structure. Deviation may be a consequence of overriding organizational self-interest of the governance structure or of overriding personal motives of the representatives at level L1. In these cases, the representatives do not truly represent their constituencies. Second, aggregation mechanisms may lead to a change in the strategic constellation in the L2 game. This can be a consequence of distortion created by aggregation mechanisms, but is not necessarily so. For example, if both the subgroups and the total collective are heterogeneous and preferences are aggregated according to the majority rule, it may well be that the majority in each subgroup belongs to the same type of actors; thus truly representative L1 actors are homogeneous in the L2 game (constellation 5). This cannot be called distortion, as it is the intended consequence of the majority rule. Still, the strategic structures of the L1 games and the L2 game are different, as are the L2 game and the single-level game. The effects of self-interested L1 actors and aggregation mechanisms are demonstrated in chapters 9 and 10, using the examples of tax coordination (constellation 2), the European Monetary Union (constellation 3), and the siting of LULUs (constellation 5).
178
T R A N S N AT I O N A L C O M M O N G O O D S
Two caveats are appropriate. First, the assumptions of complete homogeneity within the collective or the subgroups, and of heterogeneity being reduced to two types, X and Y, are highly stylized. In most real cases, there will be some heterogeneity, it will be present in varying degrees, and it will be distributed across the whole collective. The examples show, however, that it may still be justified to distinguish between cases where heterogeneity within countries is most important and cases where the heterogeneity between countries is most important for the strategic structure of a situation. Second, the effects of multi-level systems in real-world cases are much more complicated. If there are more levels, there are also more combinations and thus more possibilities of change of strategic constellations. Real-world aggregation procedures and real-world selfinterest of political actors can take many forms and, consequently, the opportunities for multi-level effects are multiplied.
Chapter 8
Case Studies 4: Homogeneity and Heterogeneity
I
n this chapter, three combinations of homogeneity and heterogeneity are examined none of which lead to relevant changes in the strategic structure through the introduction of a multi-level system. Section 8.1 examines the effects on the strategic constellation if actors are either homogeneous or heterogeneous at all levels. The examples of the pollution of an international lake and an international river are used to illustrate. In section 8.2, a situation is analyzed where the actors at L0 are homogeneous within the jurisdictions, but are heterogeneous between the jurisdictions. Consequently, the L1 actors are also heterogeneous—provided that there is no divergence between the representatives and their constituencies. This is illustrated by the example of global warming. 8.1 Homogeneity and Heterogeneity of the Groups: Lakes, Rivers, and Biodiversity In this section, the effects of multi-level systems on the strategic constellation of common goods provision is examined for constellations 1 and 6. The pollution of an international lake and of an international river serve as examples. In constellation 1, the actors are homogeneous; in constellation 6, they are heterogeneous within the subgroups of L0 actors, as well as within the total collective. These combinations imply that there are no distortion effects through aggregation or self-interested representatives at L1, so that L1 actors truly represent the strategic situation at L0. It is assumed that the representatives take the same position as all the constituents in the
180
T R A N S N AT I O N A L C O M M O N G O O D S
homogeneous case and as the majority of constituents in the heterogeneous case. Pollution of an International Lake In chapter 3, the pollution of an international lake was introduced as an example of an open-access common pool resource. The Great Lakes in North America or Lake Constance in Europe served as real-world examples; similar problems are provided by the pollution of oceans or seas, such as the Mediterranean. It was assumed that the riparians of the lake use the water for many different purposes such as extraction of drinking water, bathing and fishing, or the absorption of household and industrial discharges. The discharges are distributed in the water and deteriorate the water quality for all other purposes. It was further assumed that all riparians are homogeneous with respect to these uses. Thus, the discharges of one riparian create negative externalities that are shared by all riparians. On the other hand, protection measures by any one user would create positive externalities for the others. The externalities—positive or negative—are bidirectional in this case, and thus the riparians are homogeneous in their preferences for pollution or protection, respectively, of the lake. The actors were assumed to be states A and B who play a prisoners’ dilemma in the problematic second phase of exploitation of the resource. These “states” will now be “disaggregated” and I will examine the three groups of actors in a two-level system. First, the subgroups, the individual users of the lake within the two states, are considered. As all riparians are assumed to be homogeneous with respect to their uses, the individual riparians within state A, as well as within state B, face the same strategic constellations as outlinedearlier, depending on the actual phase of exploitation. The same is true for the collective as a whole, namely, the inhabitants of state A plus the inhabitants of state B. Finally, as we assumed that there are no self-interested representatives, the governments of states A and B are also in a symmetrical position and homogeneous in their preferences. As modeled in chapter 3, they play a prisoners’ dilemma in the second phase. In sum, the strategic structure is the same in the L1 games of the subgroups, the L2 game of the representatives, and in the single-level game of the total collective. For the comparison of the two-level system and the single-level system, the difference or equality of the L2 game and the single-level game is crucial, as it is the L2 game in a two-level system that finally determines the outcome. Therefore, it
H O M O G EN E I T Y A N D H E T E R O G E N E I T Y
181
can be said that multi-level provision has no effect in this constellation, because the L2 game and the single-level game are identical. Although there is no effect of the presence of a multi-level system on the strategic constellation, and thus on the noncooperative solution of the common goods provision problem, there may still be an effect of the multi-level system on a cooperative solution. A cooperative solution may be easier to find in a multi-level system than in a single-level system comprising the whole collective. This is due to the number and anonymity of actors. In cases where the number of actors is very large, such as the Great Lakes or the Mediterranean Sea, it would be difficult for all the individual actors to come together and to find an agreement, due to the sheer mass of individuals involved. Thus, representation at L1 has an aggregative function—a fact that is almost self-evident, but that I nevertheless want to emphasize. Furthermore, as the game is a prisoners’ dilemma, each individual is tempted to defect. Defection would be difficult to monitor and sanction in cases where the number of actors is very large and anonymity prevalent, if an agreement has been achieved in the singlelevel structure. As the governance structures at L1 can be assumed to possess a monopoly on legal powers, these structures can be made responsible for the implementation and enforcement of any agreement found at L2 among the L1 representatives. (This provides no solution to the temptation to defect for L1 governance structures, however, as long as there is no governance structure at L2 with similar enforcement power). In general, a multi-level system can facilitate the solving of common goods problems where the number of actors is very large and anonymity prevalent, because lower level governance structures have positive aggregation and enforcement functions. Pollution of International River Basins Does the earlier result—that there is no effect of multi-level provision on the strategic constellations with completely homogeneous collectives— change if the collective is heterogeneous? This is examined using an example from chapter 4, where the actors were assumed to be heterogeneous. The case of international river pollution is most similar to the case of international lake pollution. The uses of the river are the same as with the lake. The situations differ only relating to one characteristic: The riparian countries of a river are heterogeneous in their capacities to cause and receive externalities. There are two types of actors, namely, upstream riparians and downstream riparians. The upstream users are not affected by the actions of downstream users of
182
T R A N S N AT I O N A L C O M M O N G O O D S
the river, but the downstream users will suffer from the polluting discharges of upstream users. The upstream users have an advantage, because the externalities caused by users downstream cannot reach them. In this situation, both types of actors have a dominant strategy to pollute. They play a rambo game that poses a distributional problem. The model in chapter 4 was based on states. What about the strategic constellation of L0 actors within each state? Again, there are individual upstream and downstream users within the states. Each pair of upstream and downstream actors plays the same rambo game, if we assume that each upstream user releases polluting discharges and each downstream user extracts water for other purposes. This is true both within each subgroup or country, and within the total collective of countries. The representatives at L1 face the same situation: all their inhabitants are either upstream or downstream users vis-à-vis the inhabitants of the other state. Thus, the representatives also belong to two types: either upstream users or downstream users. As in the case of homogeneity, with heterogeneity in all groups the strategic constellation is identical in the L1 game, the L2 game, and the single-level game. That is, the presence of heterogeneity as such does not lead to a change in strategic structure in the multi-level provision of common goods. Whenever heterogeneity is distributed similarly across the subgroups as well as the total collective, and the representatives at L1 mirror the heterogeneity of L0 constituents, there is no difference in strategic constellations of the single-level game and the L2 game. What about the aggregation and enforcement functions in the case of a cooperative solution at L2? In chapter 4, it was mentioned that the downstream users have a strong incentive to negotiate. They try to induce the upstream user to stop polluting by offering compensation for the losses that accompany refraining from the polluting activity. These negotiations are much easier if they can be carried out collectively. It would not be possible for every individual downstream user to negotiate with each and every upstream riparian. If the L1 representatives of the downstream country negotiate successfully, they can achieve the reduction of pollution for their whole constituency. The upstream country has the power to enforce such an agreement among its inhabitants, and will do so if there is sufficient compensation. As in the case of lake pollution, the presence of a governance structure at L1 can thus help to find and enforce a cooperative solution, which would be much more difficult among a large number of anonymous and nonorganized individual actors at level L0.
H O M O G EN E I T Y A N D H E T E R O G E N E I T Y
183
However, on closer examination, it seems that the strategic constellations at the single-level game and the L2 game are not identical. Whereas it is certainly true that each pair of upstreamdownstream actors, and also each pair of upstream-downstream countries, face a rambo game, the situation is a little different when the whole collective at L0 is considered. The riparians live along the river in a row and only the first user upstream and the last user downstream are “pure” upstream or downstream users. All the others in-between play both roles; they represent a mixture of both types— upstream user and downstream user. Compared with all other users upstream, they are downstream users; but, compared with all other users downstream, they are upstream users. They suffer from upstream pollution but release discharges downstream themselves. The situation is similar to that of the lake, and thus the in-betweens are more homogeneous than it seemed at first glance. There is still a notable difference, however: The pollution of all upstream users is cumulative in the river, and the more downstream a riparian is located, the greater is the negative externality for him or her. The particulars of the river example, however, do not invalidate the argument about heterogeneity in all three groups of actors. To substantiate this, another example is briefly introduced. The protection of biodiversity was treated in chapter 3 as a symmetric game, that is, as if the actors were homogeneous. This is surely not the case; in general, there will be two types of actors. There are beneficiaries from biodiversity protection (type X), but there are also actors who suffer losses from it (type Y). The beneficiaries from the protection of, say a certain species or ecosystem, are spread all over the world. Practically everybody benefits from the protection of a species, but the individual benefit is very small and might not even be felt for most actors. Benefits are small and diffuse. Costs of protection, however, are usually concentrated and may be very high for some actors. The protection of a species or ecosystem, for example, the elephant or the tropical rainforests, often endangers the economic activities of certain people who earn their living with these activities. These people will have a strong preference against the protection of the endangered species or the rainforest, while most beneficiaries will only weakly favor protection. It is not necessary to model the exact strategic constellation. It is a heterogeneous actor constellation, and it is identical for the subgroups of L0 actors living in a jurisdiction, in which such a species shall be protected, as well as for the collective as a whole, because there will always be some actors belonging to type X and others belonging to
184
T R A N S N AT I O N A L C O M M O N G O O D S
type Y. The L1 representatives will be heterogeneous, if there are some countries where the species to be protected is not present, and other countries where the species forms the basis for an important economic sector. In this situation, the strategic constellation will be the same in the L1 game, the L2 game, and in the single-level game. Multi-level provision does not matter if the heterogeneous types of actors are distributed over the whole collective, such that there is heterogeneity within the subgroups and between the subgroups. 8.2 Heterogeneity of Homogeneous Constituencies: Global Warming This section examines the effects of multi-level systems on the strategic constellation for constellation 4. The subgroups are homogeneous, but the collective as a whole and the representatives of the subgroups are heterogeneous. This combination implies that there is heterogeneity between the subgroups. An example that matches these conditions is global warming. Again, it is assumed that L1 actors truly represent the strategic situation at L0. Global warming was introduced in chapter 3 as an example of a summation technology and as a symmetric game among nation-states. It was mentioned, however, that this model is not quite correct, because asymmetry is a very important factor of the strategic constellation in the global warming case. The heterogeneity of actors in this case has natural and economic causes. Global warming is expected to have serious consequences, which will not be equally felt all over the planet. Among the expected consequences are increases in heavy storms, the flooding of low lands and islands, and problems with world food supply. The latter two problems in particular are regional in character. Coastal lands, delta regions, and low-lying islands will suffer from the expected increase in sea level. Arid areas in Africa, South America, and Central and South Asia will suffer from higher temperatures (Benedick 1999: 13). Mostly third world countries will suffer from difficulties with food supply. Continental states in Europe and North America will neither be endangered by flooding nor by hunger. Thus the effects of global warming are distributed unevenly over the world: Some states will suffer first and much more severely, while others will experience only marginal effects at a later time (Tietenberg 1997, part II). Similarly, the contributions to the greenhouse effect are unevenly distributed: Some states, namely the developed economies in North America and Europe, have emitted (and continue to do so) much more greenhouse gases in the past than others. It is these gases from
H O M O G EN E I T Y A N D H E T E R O G E N E I T Y
185
the past that are responsible for the effects, which we are now just beginning to feel and which will be felt in the near future. For this reason, the less developed states would like the developed states to contribute more to the reduction of emissions. This distributional concern is disputed, however, as the highly populated less developed states, such as China and India, are expected to contribute at least as much to the greenhouse effect in the future (Benedick 1999). Another argument why the less developed countries should also contribute to the future reduction of emissions is that emission reduction is much cheaper for those countries than for the developed world. This is a consequence of increasing marginal costs of emission reduction and the fact that most of the developed world has already introduced a host of energy-saving measures. In these countries, the emissions can only be reduced by changing lifestyles. As the greenhouse gases mainly result from human production or consumption activities, these activities had to be given up or reduced to reach a substantial decrease in emissions. Thus, in these countries the costs of contributing to the restoration of the common good “climate” are felt to be very high. Thus, if the pure distributional disputes about who should contribute more, when and how, to the reduction of emissions are neglected, there are two types of countries. I will call those states “type X,” which will suffer less from the future consequences and at the same time have higher costs of reduction now. Those states, which will suffer more from the negative climate effects and have lower costs of emission reduction now, will be called “type Y.” There is a “double heterogeneity” among the states, which affects both the benefits and the costs of emission reduction. What about the individual L0 actors within the subgroups, that is within the states? Each individual is basically in the same situation as his or her respective subgroup or country. They live either in regions more endangered or in regions less endangered by the consequences of global warming, and they live either in regions that incur less immediate costs of emission reduction or in regions that incur more of such costs. Although individual actors within the subgroups may, in practice, be in different positions and have different preferences, depending on their personal economic position and lifestyle, these internal differences count much less than the differences between the subgroups. Thus, it can be said that the individual actors at L0 are homogeneous within their respective subgroups. The collective of L0 actors as a whole, however, is heterogeneous. It consists of heterogeneous subgroups that belong either to type X or
186
T R A N S N AT I O N A L C O M M O N G O O D S
to type Y. The representatives of the subgroups at L1, according to the assumptions, truly represent the preferences of their constituencies. Thus, there are two types of representatives at L1 and this group is heterogeneous. It was assumed in chapter 3 that the individual cost of one unit of emission reduction is higher than its individual benefit (c > b). Given this assumption about actors’ preferences, as well as nonrivalry and summation technology, the strategic structure was a prisoners’ dilemma. These assumptions are still correct for all actors of type X, that is, for the inhabitants and representatives of states that expect low benefits from emission reductions as a result of their geographical position and economic development. All subgroups of type X actors play in fact a prisoners’ dilemma. For the inhabitants and representatives of type Y states, however, costs of emission reduction are lower than expected benefits. The assumption for these subgroups has to be changed to b > c. Table 8.1 shows which game the type Y subgroups play among themselves. Table 8.1
Global Warming and Type Y Actors in the Level 1 Game
Assumptions
2b . b . c . 0 Strategy Combination
Actor A
Benefits from Emission Reduction
Costs of Emission Reduction
General Payoff
Ordinal
A: R
B: R
2b
c
2b 2 c
4
A: R
B: ~R
b
c
2
A: ~R B: R A: ~R B: ~R
b 0
0 0
b2c b 0
3 1
All factors are identical for actor B.
Game Matrix
Actor B
Reduction
Reduction
No reduction
2b 2 c, 2b 2 c 4, 4
b 2 c, b 2, 3
b, b 2 c 3, 2
0, 0 1, 1
Actor A No reduction
187
H O M O G EN E I T Y A N D H E T E R O G E N E I T Y
The game in table 8.1 is a harmony game. This implies that type Y actors will voluntarily reduce their climate relevant emissions. This model may exaggerate a little—in actual climate negotiations, many type Y countries demand instead that other countries reduce their emissions, rather than that they reduce their own. Also some of the type X states, in particular European Union (EU) member states, are obviously willing to reduce emissions. As mentioned earlier, in reality heterogeneity is not restricted to two types; there are many different interests represented in the climate negotiations (Benedick 1999). At the level of the L0 actors within the subgroups, there are two different L1 games in constellation 4: one type of subgroup plays a prisoners’ dilemma; the other type plays a harmony game. The L2 game of the representatives looks different from both of these L1 games. The representatives are heterogeneous, as some represent type X and others type Y. The strategic constellation is a hybrid of the prisoners’ dilemma and the harmony game (see table 8.2), namely a rambo game. The representatives of the Y subgroups have a dominant strategy to reduce emissions, while the representatives of X have a dominant strategy not to reduce. What would the single-level game played by all members of the whole collective look like? As the L0 actors within the collective are a mixture of type X and type Y players, the single-level game is the same as the L2 game. The homogeneous subgroups belonging to two types play different L1 games and, consequently, the representatives play another game at L2. This game is identical with the single-level game, wherein the same heterogeneity is represented. Therefore, in this case as well, the strategic constellation does not change through the introduction of a multi-level system. As in constellations 1 and 6, however, the presence of a multi-level system may ease the finding Table 8.2
Global Warming and the Representatives in the Level 2 Game
Game Matrix Representative Y
Reduction
Reduction
No reduction
2b 2 c, 2b 2 c 3, 4
b 2 c, b 1, 3
b, b 2 c 4, 2
0, 0 2, 1
Representative X No reduction
188
T R A N S N AT I O N A L C O M M O N G O O D S
and implementing of cooperative solutions through its positive aggregation and enforcement functions. In the case of global warming, the heterogeneity of homogeneous subgroups arises from the geographical position of the countries, as well as from their economic development. Economic differences very often account for situations that fit into the constellation of constellation 4. Whenever the subgroups (jurisdictions, countries) are more heterogeneous than the actors within the subgroup, the situation might be an example of the constellation 4 constellation—provided the governance structures truly represent the preferences of the subgroups. Environmental regulation within the EU, as treated in chapter 5, provides many examples. While the developed “green” countries have a high preference for strict environmental regulation, the less developed “nongreen” countries have a preference for further economic development without many environmental restrictions. These different general attitudes toward environmental protection between the European member states are stronger than the heterogeneities relating to environmental preferences within the countries. However, the context of the EU provides the heterogeneous L0 actors with the opportunity to form cross-boundary majority coalitions of type X or type Y actors, which is much less possible in an international negotiation system (Michaelowa 1998). Thus, although the multi-level system does not generally change the strategic structure in constellation 4, in a more detailed analysis the type of multi-level system might matter.
Chapter 9
Case Studies 5: Distortion Effects through Representation
This chapter examines the first two combinations of homogeneity
and heterogeneity that in fact lead to a change in the strategic constellation in a multi-level compared to a single-level system. The effect on the strategic structure arises from the distortion that may take place between the levels of a multi-level system. Sections 9.1 and 9.2 deal with situations where the change in strategic structure is a result of divergence of interests between the L0 actors and the representatives. There are many causes why the representatives of a collective, for example, the legislators within a multi-level governance structure or a government within international negotiations, deviate from the preferences of their constituencies. Just four are mentioned here: Divergence of interests may first be a result of personal preferences or convictions of a representative that are different from those of (the majority of) the constituents. This is the least interesting cause, as it can be the least generalized. Second, divergence may arise from organizational self-interest of the representatives, for example, the desire to get reelected. In general, collective actors in politics pursue different types of goals: the substantive goals relating to a certain policy or common goods problem, the official goals of the organization, and political-strategic goals. To form a position in a concrete instance of policy formulation, the different goals have to be brought in accord somehow. For example, in the problem of capital income taxation, governments have the official goal of “the state” to collect taxes on the one hand; on the other, they have political-strategic goals, namely enjoying the political benefits associated with a large capital market (section 9.1).
190
T R A N S N AT I O N A L C O M M O N G O O D S
Third, divergence of interests between the representative and their constituencies may be the result of a different perception of what best serves the collective welfare. Fourth, in a cooperative structure at the upper level, representatives communicate and may persuade one another that a certain solution to a policy problem is the best, which is not, however, the solution the constituency had opted for. Similarly, representatives may use bargaining techniques and, in this way, the position of some of the representatives may change. The introduction of the European Monetary Union (EMU) provides an example for a mixture of the third and fourth causes (section 9.2). The two examples of divergence of interest discussed in the following two sections represent two different motives for divergence. They also represent two different combinations of homogeneity and heterogeneity across the three relevant groups. In constellation 2, the actors at L0 are homogeneous both within and between the subgroups. Heterogeneity comes through the representatives’ interests (capital income tax coordination). In constellation 3, although there is heterogeneity between homogeneous subgroups, the representatives are homogeneous (EMU). 9.1 Distortion through Self-interested Representatives: Capital Income Tax Coordination Why is the problem of capital income taxation and its coordination in the European Union (EU) an instance of constellation 2, that is, a constellation where both the subgroups and the collective as a whole consist of homogeneous actors, while the representatives are heterogeneous? The common good in question is the collection of capital income taxes in a common European capital market. Capital income coordination in the EU was treated as a problem of common goods provision among the national governments in chapter 4. Thus, the strategic constellation analyzed was the L2 game at the level of the representatives. It was shown that the governments are heterogeneous regarding their preferences for tax coordination, respectively, tax competition. The heterogeneity was a result of different size of the capital markets and the different valuation of two elements in the utility function: tax revenue and political benefits from capital stock. Under these assumptions, the game played by the European governments is an asymmetric prisoners’ dilemma. What do the corresponding L1 games of the constituencies and the single-level game look like? What is the strategic constellation for the individual actors at level L0? The relevant individual actors are
D I S T O R T I O N T H R O U G H R E P R E S EN TAT I O N
191
the national taxpayers. The common good within the subgroups, that is, within the European member states, is the collection of capital income taxes, as the tax revenue can be used by the government to provide common goods. In fact, this game is the prototype of a common goods game: The government is the actual producer of common goods in a state and the taxpayers contribute to the common goods by paying taxes to the government. Thus, the strategies in the game of the individual taxpayers within the subgroups are to pay taxes or not to pay taxes. Although the taxpayers benefit from the goods provided by the government, they do not like to pay taxes. As long as these goods are nonexcludable, the taxpayers are able to free ride on them. It is sensible to assume that this “general common good” provided by the government and financed by the taxes follows a summation technology. The more taxes paid, the more goods can be provided. Thus, the associated game can be assumed to be a prisoners’ dilemma. The paying-taxes-for-common-goods game within nation-states has a cooperative solution: Citizens are obliged to pay taxes and, in general, national governments can be assumed to possess sufficient enforcement power so that tax evasion can be avoided, at least to a certain degree. This is less true, however, for capital income taxation. Although citizens are obliged to pay these taxes in the same manner, the monitoring capacity of the government is low. As a result of the liberalization of capital markets, it is easy for capital owners to evade taxes by investing their money abroad. Capital income taxpayers can free ride on national common goods. Thus, a prisoners’ dilemma is played among the taxpayers at L0 in case of capital income taxes, since there is no effective enforcement structure. This is not only true within the subgroups, but also for the collective as a whole, namely the capital owners in the EU. The scope of the problem of collecting capital income taxes is worldwide. All potential European capital income taxpayers face the same dilemma: They would like to receive the national common goods; however, it is individually rational for them to evade national taxes by investing the money abroad. There is no international cooperative solution yet. The strategic constellation is thus identical for the constituents within the member states and for the L0 actors within the total taxpayer collective of the EU. Both games are symmetric prisoners’ dilemmas, implying that there is homogeneity of the actors. This homogeneity will not be perfect: The incentive to evade taxes will differ between the individuals, depending on the tax rates in different states, personal marginal tax rate, and the share of capital income of
192
T R A N S N AT I O N A L C O M M O N G O O D S
one’s personal income. In principle, however, the incentive works in the same way for everybody. The benefits from national common goods and the paying of taxes can assumed to be “de-coupled” in individual perception. The problem of paying taxes for capital income is perceived as a separate game played not against the other citizens but against the government. This game between taxpayer and government is modeled, because it illustrates the divergence of the interests of taxpayers and governments. It is the official and most central task of governments, which creates this divergence. Governments have to collect taxes to be able to provide common goods. Governments and their constituencies have antagonistic goals regarding taxes. Although governments need to collect them—not only because it is their task to do so, but also because providing as many common goods as possible raises the prospects of reelection—the taxpayers want the common goods but do not want to pay for them. It is worth noting that, in this case, the main reason for the divergence of organizational self-interest on the part of the representatives from the interest of the individuals has to do with the basic tasks of governments and is in accordance with collective welfare. In chapter 4, a second goal of governments was introduced: namely, that they seek to enjoy the political benefits associated with a large capital stock and a well-functioning domestic capital market. Thus, governments not only want to collect taxes, they also want to avoid capital flight. This aggravates the problem for them: Nontaxation creates no revenue but taxation leads to capital flight and thus reduced revenue. The following assumptions are made for the game. There are two players, a taxpayer and the government. The government has two strategies, namely, to levy taxes (T) on capital income or not to do so (~T). The taxpayer’s two strategies are to invest the capital domestically (D) or to invest it in a foreign country (F). The government has two kinds of benefits, tax revenue (t) and political benefits from a large and flourishing capital market (b). Levying a tax implies transaction costs of introducing a tax scheme and collecting the tax (c). These costs, however, count less than tax revenue and the political benefits (t, b . c). The taxpayer earns the profits from the investments (i). His or her benefits shrink if he or she has to pay taxes (t) and if he or she has to invest his or her money abroad, as transferring money to a foreign country has some transaction costs (c). It is reasonable to assume that these transaction costs are much lower than the domestic taxes (t . c). As taxes are a share of the profits from investment, the profits are higher than the taxes (i . t).
193
D I S T O R T I O N T H R O U G H R E P R E S EN TAT I O N
Table 9.1 shows that the first preference of the taxpayer is to invest domestically and not get taxed by the government. His or her second preference is foreign investment, and his or her last preference is domestic investment, if the profits get taxed. The government’s first preference is to levy the tax while the capital is invested domestically. Its second preference is not to levy the tax while the capital is invested domestically. Its third preference is capital flight combined with no tax, and its last preference is to develop a tax scheme and to be confronted with capital flight as a result.
Table 9.1
Capital Income Taxation Game between Taxpayer and Government
Assumption G
t, b . c Strategy Tax Combination Revenue
Government G
Assumption I
G: T
I: D
t
b
c
t 1 b2c
4
G: T
I: F
0
0
c
1
G: ~T I: D G: ~T I: F
0 0
b 0
0 0
2c b 0
3 2
i . t.c Strategy Investment Combination
Taxpayer I
Political Transaction General Ordinal Benefits Costs Payoff
Tax
Transaction General Ordinal Costs Payoff
I: D
G: T
i
t
0
I: D I: F
G: ~T G: T
i i
0 0
0 c
i2c
3 2
I: F
G: ~T
i
0
c
i2c
2
i2t i
Game Matrix
Taxpayer
Tax
Tax
No tax
t 1 b 2 c, i 2 t 4, 1
2 c, i 2 c 1, 2
b, i 3, 3
0, i 2 c 2, 2
Government No tax
1
194
T R A N S N AT I O N A L C O M M O N G O O D S
The game in table 9.1 is a discoordination game. The preferences of the players are almost diametrically opposed. Whenever one player wants to choose a strategy, the other player wants to react by choosing another strategy. There is no intersection of optimal answers, and thus no Nash equilibrium. If the government levies the tax, the taxpayer invests the capital abroad; if the taxpayer invests abroad, the government does not want to collect the tax; in this case, the taxpayer prefers to invest domestically. However, in the case of domestic investment, the government would like to levy the tax, and so on. The divergence of interests of the constituency and the representatives is almost complete in this case. In sum, in the example of capital income taxation the strategic constellation changes through the introduction of a multi-level system. The L1 game among the constituencies—the taxpayers of each member state—is a prisoners’ dilemma. In a single-level system, the game among the L0 individual taxpayers is a prisoners’ dilemma, too. A cooperative solution to this problem is needed but faces the problem that the scope of the common good is actually worldwide. A cooperative solution at the EU level is not sufficient, as capital mobility is not restricted to the EU. The L2 game among the representatives—the EU member state governments—was shown to be an asymmetric prisoners’ dilemma, because of the heterogeneity of preferences of member state governments. The representatives’ interests deviate from that of the homogeneous individual actors at L0. Representatives’ interests do not simply aggregate and represent the preferences of their constituencies; representatives have their own interests. These organizational goals are tax revenue as well as political benefits capital markets. The heterogeneity of the representatives in the L2 game stems from the fact that different governments value these two goals differently, depending on the actual conditions in their respective countries. 9.2 Distortion through Cooperation: European Monetary Union Among the applications introduced in part I is no example that fits the constellation of homogeneous subgroups, heterogeneous total collective, and homogeneous representatives. I choose to use the introduction of the EMU as an example, which at least should fit, if certain assumptions about the individuals’ preferences are made. These assumptions do not fully match the empirical data. The EMU example is stylized in two respects. First, the L1 and the single-level
D I S T O R T I O N T H R O U G H R E P R E S EN TAT I O N
195
games are not real games; deciding on monetary union is a game among nation-states, not among citizens. Second, the homogeneity among the subgroups is somewhat counterfactual and, similarly, homogeneity among the representatives is not complete because there were opt-outs from the EMU. I return to these issues in the following text. Monetary Union as an Excludable Network Good Why is the EMU, or more precisely a single European currency, a common good for the European nations? As any standards, common currencies within countries or within certain currency zones are network goods. They are a means of saving the transaction costs of barter exchange, they allow for easy comparison of prices, and for the saving and accumulation of values. The larger the zone they can be used in, the greater the benefit for the individual user. A common European currency has a number of additional advantages for the participating nations. It eliminates the exchange rate risk in trade and financial transactions. It avoids the transaction costs of exchanging currencies within the EU common market. A common currency can become (and the Euro has in fact become) a major currency in international trade and financial markets and thus a competitor to the US dollar or the Japanese yen. A major internal benefit for most of the participating European member states is the expected price stability brought about by a single currency. The common currency and the associated common monetary policy implied for the participating governments is the most credible commitment to low-inflation policy (Sandholtz 1993: 14). A stability-oriented policy, in turn, is expected to create increased investment and, consequently, higher growth and employment. The Euro has been qualified as an excludable network good: “The Euro can be considered as an excludable network good to the extent that its utility grows with the number of participants and that nonparticipating countries are largely being excluded from these positive effects” (Taylor 1995: 26−27; Koelliker 2001: 136). The complementarity of benefits amounts to many of the benefits just mentioned. The more member states participate, the higher are the savings in transaction costs for the individual participant, the higher are the positive efficiency and growth effects from increased competition within the Euro zone, and the greater is the importance of the Euro as an international currency competing with the dollar and the yen. The effect on price stability, however, is excludable but not
196
T R A N S N AT I O N A L C O M M O N G O O D S
complementary. As nonparticipants can be excluded from the positive network benefits, the free riding of outsiders is very limited. The analysis of the effects of a multi-level system requires that the strategic structures be compared within the three groups as above. This is difficult in the EMU example, because deciding to have a common currency is a game that can be played by nation-states (governments) but not at the level of L0 actors. Individual actors do not play a strategic game in this case, as they cannot decide whether to introduce the common currency or not. However, equivalent games at the level of individual actors can be imagined. For example, a game could be modeled of an anarchic situation where no national currency exists and individuals use different means of payment. Under which conditions could we expect the common good of a common currency to come about? A more realistic scenario is based on the fact that individual actors within a state are not obliged to use just their own national currency. They have the choice to use foreign currencies and this, in turn, can have an effect on the currencies used within the overall collective. I shall briefly sketch a model of such a game, before I come back to the situation faced by individual actors in the advent of a common European currency. Noncooperative Common Currency Game The game is played among the L0 actors within two subgroups of a collective. Each subgroup has its own currency—one is weak and the other strong. Strong currency means that there is a strict disinflationary monetary policy and thus price stability. The weak currency is less stable, as there is inflation and devaluation, or loss in value against foreign currencies, respectively. What will the two players in each subgroup do? They have two strategies: to use their domestic or to use the foreign currency. There are different situations in the two subgroups: For the players in the weak-currency subgroup, the foreign currency is the strong one; for the strong-currency country, the foreign currency is the weak one. A single currency within each subgroup causes less transaction costs, that is, there are some costs of choosing the foreign currency. Under these conditions, the individuals in the weak-currency country play a harmony game where both players have a dominant strategy to choose the foreign, strong currency. This is valid as long as the individual value differential between the weak and the strong currency is greater than the transaction costs of having two
D I S T O R T I O N T H R O U G H R E P R E S EN TAT I O N
197
currencies within the subgroup. Correspondingly, the players in the strong-currency country play a harmony game where they stick to the domestic, strong currency. Consequently, a common currency evolves within the whole area or collective, namely, the strong currency. Only if both currencies were of the same strength would they coexist in the long run, as it would then be a harmony game sticking to its own currency. If transaction costs are higher than the value differential between the currencies players end up in an assurance game, where the Pareto-optimal equilibrium is again the strong currency. If the same game were to be played by the L0 actors within the whole area, the result would be the same. The individuals would choose between the strong and the weak currency, and, depending on the relative size of transaction costs and the value differential between the two currencies, either a harmony or an assurance game would be played. In any case, the strong currency would prevail. As the common currency is a network good, individual transaction costs of switching to the strong currency decrease as the number of other players who also switch increases. This scenario is realistic whenever there are very weak currencies. For example, in South America, the US dollar is used as a parallel currency in many countries. The same is true for Russia, and for some Middle and Eastern European states, where the US dollar, the German mark, and the Euro perform the function of strong parallel currencies. However, the domestic currencies must show very bad performance for considerable time, before such parallel currencies are widely used. In Argentina, for example, there had been economic decline for almost two decades during the 1970s and 1980s—including two major bank failures and two phases of hyperinflation—before Argentineans began privately, in the 1980s, to invest all of their money abroad and to use the US dollar as a means of payment in Argentina. In 1991, the currency board decided to fix the peso to the US dollar. As the strength of a currency is not a natural property, but is determined to a large degree by the economic and monetary policy of a country, the strategy of L1 actors equivalent to L0 actors’ choosing of the strong currency, is linking one’s own weak currency to a foreign, strong currency by a fixed exchange rate, or by mimicking the monetary policy of the foreign country, as in case of the European states just mentioned. Thus, if a noncooperative L2 game is played among governments, the strong-currency government will keep its own currency and maintain its monetary policy, while the weak-currency government will try to achieve stability by coupling its currency and monetary policy
198
T R A N S N AT I O N A L C O M M O N G O O D S
to the strong currency. The result is similar to the one at the level of L0 actors: Even if there is not really a single common currency, the strong currency becomes dominant. As the case of Argentina shows, the games at both levels may be played in parallel and may interact. Thus, the L1 market outcome and the noncooperative L2 game outcome lead to the same result: The strong currency becomes the common or dominant currency. Introduction of the Euro: Citizens’ Preferences As said, the decision whether to adopt a common currency is basically a cooperative decision among states. Nevertheless, the L2 game can be modeled as game of two member states, which have the two strategies either to “join the Euro” or “not to join the Euro.” As the EMU is an example of flexible integration, implying that member states can “opt out,” the EU is in this case an exit system. Only the L1 governments can play this game. L0 actors are not free to decide whether they join the Euro or not. There is neither a L1 game nor a single-level game. However, L0 actors can develop a preference order for the four possible outcomes formed by the strategies from which their countries can choose. I begin by examining the preferences of the actors at L0, within and between the subgroups. The L2 game will be modeled afterward. The individual actors within the subgroups develop their preference order on the basis of the four possible strategy combinations in the game played by their representatives. Their preference order can be interpreted as an answer to the question of what they would vote for: • • • •
not joining the Euro if the other countries also do not join, not joining the Euro in any case, joining the Euro in any case, and joining the Euro if the other countries do so as well.
How are individuals affected by the introduction of a single European currency? Which elements in their utility function determine their preference order? First, the individual actors also profit from the network benefits (b) of a common European currency. They save transaction costs when traveling or investing money abroad and they indirectly benefit from the positive effects on the economy of a large currency zone. During the phase of the actual changeover to the new currency, there are some transaction costs (t) of exchanging money and getting used to the new currency unit. Moreover, there
D I S T O R T I O N T H R O U G H R E P R E S EN TAT I O N
199
may be some costs for giving up a traditional national symbol, which I will call emotional costs (e). In general, it can be assumed that the introduction of the Euro is rated positively by the individual actors, as the expected benefits (b) should exceed the expected costs (t 1 e). With respect to these factors in the utility functions, the individual actors within each subgroup should be homogeneous. There are some qualifications to this claim: First, heterogeneity in the valuation of the Euro can be expected to be found between the general public and some business sectors, which are especially exposed to dealing with foreign currencies: the export industry and the financial sector. The benefits b will be much more directly felt by those business actors than by the general public. Thus, this group can be expected to value the introduction of the Euro even more positively than the general public on a purely rational basis. Second, as the valuation of the individual elements in the utility function is purely subjective, there may in fact be individuals who would vote against the Euro, for example, as a consequence of underestimation of the economic benefits or high valuation of the emotional costs. With respect to the analytical goal pursued here, it is important that there is heterogeneity between the subgroups and that, despite this heterogeneity, there is homogeneity among the governments. It is less important that the homogeneity of L0 actors within the subgroups is not perfect. It is sufficient here that, with respect to the elements of the utility function, there should be homogeneity among the actors from a perspective of economic rationality. However, I do not neglect the heterogeneities within the subgroups in the model; it is shown below empirically that the heterogeneities are in fact substantial in the EMU case. Thus, in table 9.2 the different preference orders for different types of actors within the subgroups are given (cases a and b). One element of the utility function of individual actors has not yet been mentioned: namely, the expected stability of the new common currency as compared to the stability of the traditional domestic currency. The issue of stability can be assumed to outweigh both the costs of the introduction of the Euro, if its stability is assessed positively, and the network benefits, if its stability is assessed negatively. Therefore, individuals who live in a weak-currency country can be expected to value the Euro much more positively than individuals who live in a strong-currency country. The first group can hope for more price stability including all its positive consequences, whereas the second group will fear that the former stability cannot be uphold with the new currency. Thus, there is heterogeneity between the
200
T R A N S N AT I O N A L C O M M O N G O O D S
Preferences of Level 0 Actors toward the Introduction of the Euro
Table 9.2
Assumption, case a Assumption, case b
b1s .t1e b1s,t1e
Strategy Benefit Stability Transaction Emotional Combination Costs Costs
Weak
Payoff
0 0
Ordinal
a
b 3 3 1 2
D D E
D E D
0 0 0
0 0 s
0 0 t
0 0 e
s2t2e
1 1 2
E
E
b
s
t
e
b 1 s2t2e
3
Assumption, case a b . s 1 t 1 e Assumption, case b b , s 1 t 1 e Strategy Benefit Stability Transaction Emotional Combination Costs Costs
Strong
D D E
D E D
0 0 0
E
E
b
0 0
Payoff
0 0
Ordinal
a
b 3 3 1 2
2s
0 0 t
0 0 e
2 s2t2e
2 2 1
2s
t
e
b2s2t2e
3
subgroups. Inhabitants of weak-currency countries will ceteris paribus favor the introduction of the Euro, while inhabitants of strong-currency countries will oppose it. Table 9.2 shows the preferences of individual actors in weak and strong currency countries. The possible strategies of their governments are to keep the domestic currency (D) or to join the Euro (E). Individual actors from weak-currency countries, who value the benefits of the Euro higher than the costs, prefer the Euro to their domestic currency, even if the other country does not join in. Their first preference is that both countries choose the Euro; their last preference is keeping the domestic currency. Those actors, who value the benefits less than the costs, prefer the domestic currency to a situation in which both states switch to the Euro. Their last preference is that their country chooses the Euro while other countries do not. Individual actors from the strong-currency states, who assess the network benefits higher than the costs, prefer that both states choose the Euro. Next for them comes keeping the domestic currency; their
D I S T O R T I O N T H R O U G H R E P R E S EN TAT I O N
201
last preference is a situation where their country switches to the common currency while the other countries do not. If individuals in strong-currency states assess the stability risk and the other costs higher than the benefits, then their first choice would be to stay with the national currency. Their second preference would be that both states adopt the Euro, and their last preference is that their state adopts the Euro unilaterally. If citizens in both weak and strong-currency countries attach similar values to network benefits, transaction costs, and emotional costs, then the only difference will be the expected change in stability of the currency in use. Let us assume that the transaction costs and the emotional costs are valued lower than the benefits from a common currency (which is economically reasonable, but not necessarily in line with subjective valuations). Under these assumptions, the L0 actors from subgroups with relatively weak currencies will favor their country’s adopting the Euro, whereas L0 actors from subgroups with relatively strong currencies will oppose the introduction of the Euro in their country. This is, in fact, how the majorities of citizens in the respective EU member states responded to the question of whether they favored the EMU, at the time of the Maastricht Treaty negotiations. While the citizens of states with weak (or small) currencies voted in favor of a single common currency, the citizens of states with strong (or large) currencies voted against it. As table 9.3 shows, this is especially true for the four largest members: Whereas the citizenry in Germany and the United Kingdom (with relatively strong currencies) did not support the common currency, the French and Italians (with relatively weak currencies) favored its introduction. The weak-currency states— Greece, Spain, and Portugal—also supported a common currency. Belgium, Luxembourg, Denmark, the Netherlands, and the Irish Republic did not have weak currencies at that time. This was to a large extent a result of the linking of their monetary policies to the policy of the German Bundesbank. These states did not have strong currencies before, however, and were dependent on the German mark. Moreover, their currencies were small currencies that were not significant on the international markets, so these countries thus had a lot to gain from a common currency. For just these reasons, the citizenry in these states could be expected to support a single European currency. As can be seen in table 9.3 there are two exceptions, namely, Luxembourg and Denmark, where a single currency enjoys the least support in Europe. In Luxembourg, the single currency is favored by almost 50 percent; that there is no majority support could be a result
202
T R A N S N AT I O N A L C O M M O N G O O D S
Table 9.3 Percentage of Citizens Favorable to a Common Currency Member State Belgium Denmark France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain United Kingdom European Union
1990
1991
62 35 61 49 64 58 72 46 61 55 53 38 55
62 35 64 45 61 54 69 48 58 53 58 40 54
Source: Commission of the European Communities: Eurobarometer No. 35 (A23) and No. 36 (A37).
of fears that a common currency might somehow threaten Luxembourg’s sovereignty in other financial matters. The lack of public support in Denmark probably has other reasons than economic considerations. It may be a consequence of the general euroscepticism of the population. Thus, there is evidence for the heterogeneity between subgroups relating to the distinction of weak and strong currencies. A Gallup survey of European business managers in 1989 found that their support for a single European currency was much higher than the support of the general public. Overall, 83 percent of business managers favored a common currency, ranging from 94 percent in Italy to 65 percent in the United Kingdom (Sandholtz 1993: 24). This is in line with the conjectures made earlier. Still, large minorities were against a single currency in weak-currency states, or favored its introduction in strong-currency countries. Thus the distinction represented in the a and b cases in table 9.2 has an empirical background. The following model, however, is based on the majorities: In weak-currency countries, the benefit of joining the Euro is greater than its costs (case a), and in strong-currency countries the costs are valued higher than the benefits (case b). The governments represent heterogeneous subgroups in the L2 game.
203
D I S T O R T I O N T H R O U G H R E P R E S EN TAT I O N
Introduction of the Euro: The Game among the Governments What would the L2 game look like if the governments would truly represent the preferences of their constituencies? There would be two types of governments, the weak-currency (type W) and the strongcurrency (type S) governments. The type W governments have the preferences of case a, the type S governments have the preferences of case b. The type S governments have a dominant strategy to stick to the domestic currency, while the type W governments have a dominant strategy to adopt the Euro. As a consequence, they play a rambo game (see table 9.4). There is a unique and Pareto-optimal equilibrium that gives the type S governments their first preference and the type W governments their second preference. As a result, there is no common currency: Domestic currencies and the Euro coexist, with only the weak-currency states joining the Euro zone. Such a “split outcome” was in fact what happened in the EU, although some strong-currency countries joined, as well. However, a number of countries opted out, namely, the United Kingdom, Denmark, and Sweden (when it joined the EU in 1995). Heterogeneity among the EU governments was indeed strong enough to bring about this result. At the time of the Maastricht Treaty negotiations, however, it was only the UK government, which actually resisted the introduction of the Euro. The governments of Denmark, Germany, and Luxembourg were in favor of the Euro and thus deviated from their constituencies’ preferences. I will come back to this subsequently. The fact that a common European currency is a network good has not yet been accounted for. Each additional state adopting the Euro increases the potential benefit of joining the Euro zone for each other player. States will adopt the Euro if their individual benefit of joining Table 9.4
The Euro Game among Heterogeneous Governments
Game Matrix
Government W Domestic currency Domestic currency
0, 0 3, 1
Adopt the Euro
2 s 2 t 2 e, 0 1, 1
Government S
Adopt the Euro 0, s 2 t 2 e 3, 2
b 2 s 2 t 2 e, b 1 s2t2e 2, 3
204
T R A N S N AT I O N A L C O M M O N G O O D S
the zone outweighs the costs. Thus, for each additional state switching to the Euro, the states compare the costs and benefits of adopting the Euro with the costs and benefits of keeping their domestic currency. Costs and benefits of keeping the domestic currency are simply 0 (see table 9.2). The stability, transaction, and emotional costs (s, t, e) will be summarized here as costs of adopting the Euro (CE). The network benefits are 0 for a state, if no other state switches to the Euro, and they are b, if all the states adopt the Euro. If a linear network benefit is assumed, the utility of switching to the Euro is thus
()
k uE = n b – CE where n is the number of other states and k is the number of states actually switching to the Euro. A state will choose the Euro, if its costs equal its share of the benefits CE =
(1n)b.
Table 9.5 shows the calculation of state A in our example. When state A switches to the Euro will depend on the size of the parameters. If CE = 10 and b = 12, for instance, it will switch only after five other states have joined the Euro zone. Obviously most European member states assessed their share of the network benefits to outweigh the costs, since, by the end of the Maastricht negotiations, only one government, the United Kingdom, had rejected the option to join a single common currency in Europe. Denmark’s opting out was a result of the rejection of the Maastricht Treaty by popular vote, not of the position of the Danish government, which strongly favored the EMU. Despite the heterogeneity at the level of their constituencies, the governments showed a surprising homogeneity of preferences toward the EMU. This happened because the utility function of the governments was different from the utility function of the citizens in some cases. Table 9.5
Network Effect in the Euro Game Number of States Adopting the Euro Other than State A 0
State A
Euro 2 CE Domestic
0
1
2
3
4
5
6
1 – 6b 2 C E 0
2 – 6b 2 C E 0
3 – 6b 2 C E 0
4 – 6b 2 C E 0
–5 6b 2 C E 0
b 2 CE 0
D I S T O R T I O N T H R O U G H R E P R E S EN TAT I O N
205
The utility functions of the governments were probably, in most countries, different from the utility function of the general public with respect to four goals: • It can be assumed that the emotional and symbolic factor was not particularly significant for the respective governments and that they were much more pragmatic in this respect. • The goal of European integration as such is valued much higher by political elites and governments in Europe than by the public. Introducing a common currency can be considered a symbol of and driving force behind further integration. Thus, it can be assumed that most governments had some additional integration benefit (i). This was probably not true for the UK government. • The stability concern was still more important for most governments than for their general publics. At that time, a stable currency was believed to have many positive consequences for the state of the economy. Therefore, most of the governments were highly interested in finding a strategy of committing themselves to a disinflationary monetary policy (Sandholtz 1993). The expected positive consequences on the economies enhanced the prospect of reelection for the respective governments. This was least the case for Greece, Spain, and Portugal, which, at that time, did not adhere to the above mentioned macroeconomic philosophy. The initial reservations of these countries against the EMU (Koelliker 2001) can possibly be explained by the fact that they were perhaps not keen on being forced into the disinflationary policies implied by a common monetary course. The concerns of the strong-currency countries, on the other side, about the stability of the future common currency were even more important for their respective governments than for their citizenries. • Finally, some of the states pursued a foreign policy power goal. The German mark was the dominant currency in the EMS. France, Italy, and the Benelux countries hoped to gain more influence in the common policy of a monetary union—which would be required if there were to be a single European currency—(Sandholtz 1993). There was thus an additional benefit of the EMU for the weakcurrency countries. In sum, most governments valued the potential benefits of a currency union higher than their constituencies. They had less costs and additional benefits. Some governments, however, had good reasons to oppose the single European currency. In the process of
206
T R A N S N AT I O N A L C O M M O N G O O D S
negotiating the EMU, most of this opposition could be overcome. This was achieved through compromise and package deal. Thus, not only different utility functions of the governments but also elements of the cooperative solution of the L2 game accounted for the final homogeneity of governments and for the divergence of some of them from the preferences of their respective constituencies. First, there was a solution that accommodated the concerns expressed by Germany, Luxembourg, and the Netherlands about the stability of the future common currency. European monetary policy is legally committed to price stability, the European central bank is independent, and the criteria states have to meet to be able to participate fully in the EMU are very strict. This has helped to overcome the “stability costs” of stability-oriented countries. For these countries, the insistence on strict criteria of stability represented no divergence from the goals of their constituencies. This result of the negotiations can be qualified as a gain from the cooperative solution of a common goods problem. Second, the issue of the EMU was linked with German unification. Since World War II, there has been a broad policy goal aimed at by the Western countries to bind Germany to its neighbors and thus to preclude it from further aggression. France felt that this underlying goal of European integration was threatened by the prospect of German unification in 1990. The French and other European governments desired to tie Germany irrevocably to the Union, and the EMU was an important means of doing so. Thus, the EMU and German unification formed a package deal. As Sandholtz puts it, “German support for monetary (and political) union was a bargain, the other one-half of which was French assent to rapid German unification” (1993: 33). With respect to the constituency’s preferences toward the EMU, it was a divergence on the part of the German government. Third, there was obviously no way of compensating the UK government and its constituency for their reservation against the EMU. The arrangement found to avoid a breakdown of the entire EMU project altogether was the possibility for individual member states to opt out. Thus, the transition from a veto to an exit system allowed for the introduction of the common currency. The British government did not diverge from the will of its population. In agreeing to the introduction of the EMU, three governments opted against the preferences of the majority of their constituents: Germany, Denmark, and Luxembourg. Two of these governments were compensated for potential losses by Treaty provi-
207
D I S T O R T I O N T H R O U G H R E P R E S EN TAT I O N
sions satisfying their demands and by a package deal. The Danish government was in favor of the common currency rationally: It was a weak-currency country and its expected utility from the Euro was clearly positive in economic terms. However, the majority of its population thought differently. As the Maastricht Treaty failed to achieve the required majority in the Danish popular vote, Denmark finally had to opt out. Table 9.6 summarizes how the almost complete homogeneity of governments came about. It gives the payoff function and the preference order for five relevant groups of countries: weak-currency countries (France, Italy, Belgium, Denmark, Ireland), cohesion countries (Greece, Spain, Portugal), weak-currency countries with some reservations (Netherlands, Luxembourg), Germany, and the United Kingdom. The utility function includes network benefits (b), integration benefits (i), stability benefits or costs (s), transaction costs (t), political costs of being forced to disinflationary economic policy (p), and compensation in the form of increased cohesion funds (c), concessions with respect to stability of the Euro (s), or German unification (u). Table 9.6
The Euro Game among Homogeneous Governments
Assumptions Strategy Combinations
b, i, s . t;
b 1 s1 i 1 c . p;
u . s, t
F, I, B, Dk, Irl
GR, SP, P
L, NL
Germany
United Kingdom
Payoff
Payoff
Payoff
Payoff
Payoff 0
D
D
0
0
0
0
D
E
0
0
0
0
E
D
k/n (b 1 i) 1 s2t
E
E
b 1 i + s2t
Strategy Combinations
k/n (b 1 i) 1 s 2 t k/n (b 1 k/n (b + i) 1 p2c i) 2 s 2 t 1 s 2 s 2 t 1 s 1 u b 1 i 1 s2t 1 p2c
b 1 i2s 2t 1 s
b + i2s2t 1 s 1u
0 k/n (b 1 i) 2 s 2 t b2s2t
F, I, B, Dk, Irl
GR, SP, P
L, NL
Germany
United Kingdom
Ordinal
Ordinal
Ordinal
Ordinal
Ordinal
D
D
1
1
1
1
3
D
E
1
1
1
1
3
E
D
2
2
2
2
1
E
E
3
2
3
3
2
208
T R A N S N AT I O N A L C O M M O N G O O D S
The following assumptions about the size of the parameters can be reasonably made. The transaction costs are small compared to network benefits, integration benefits, and stability benefits or costs (b, i, s . t). Political costs of disinflationary politics are lower than the gains from the Euro plus increased cohesion funds (b 1 s 1 i 1 c . p). Benefits from agreement to German unification are higher than any costs considered to be associated with the Euro (u . s, t). The network and integration benefits are k (b 5 i) whenever only k other countries join the Euro. n Under these assumptions, the utility of adopting the Euro is positive for all countries except for the United Kingdom, while the utility of keeping the domestic currency is zero. If not all member states adopt the Euro, the utility for all other members is somewhat smaller than in the case all member states do. The preferences of the first four groups of states are perfectly homogeneous now. These states play a harmony game to introduce the Euro. Only the United Kingdom prefers to stay out. This result is an ex post rational reconstruction, of course, not an empirical explanation. The introduction of the EMU represents a case where the subgroups are heterogeneous and the governments homogeneous. In most member states, the majority of the citizenry was in favor of the EMU, whereas in four member states the majority rejected a single common currency. Three out of these four governments, however, agreed to the introduction of the Euro against the will of the majority of their constituents. This was a result of different valuations of benefits and costs of the EMU, different goals on the part of governments, and compromise, compensation, and package deals within the framework of a cooperative solution at the EU level (L2). Some governments obviously had a different perception of the public welfare than the public itself. Others did not differ so much in their perception of the value of the common goods, but received compensation through the cooperative procedure.
Chapter 10
Case Studies 6: Distortion Effects through Aggregation
I
n this chapter another source of distortion effects in multi-level provision will be examined. Preferences of L0 actors can become distorted in the aggregation process as a result of effects of the aggregation mechanisms employed or as a result of the different size of lower level jurisdictions. One of the six constellations of homogeneity and heterogeneity distinguished in section 8.1 has not yet been examined. Constellation 5 represents a combination of heterogeneous subgroups, a heterogeneous collective, and a homogeneous set of representatives. It shall be illustrated by the example of the siting of locally unwanted land uses (LULUs) that shows both the distortion effects of aggregation mechanisms and of size effects. This problem was introduced in chapter 3 when it served as an example for best-shot aggregation technology. There are two jurisdictions, which want to build a commonly used facility. The facility produces negative neighborhood externalities. For the following model it is assumed that there are also some positive neighborhood effects, because the facility creates jobs and this increases employment within the jurisdiction eventually hosting the site. The problem is a volunteer’s dilemma (chicken game): it is sufficient if only one jurisdiction hosts the site. This jurisdiction has higher costs than the other, but only slightly more benefits. Consequently, both would like to have the facility, but both would prefer that the other jurisdiction provide the site.
210
T R A N S N AT I O N A L C O M M O N G O O D S
10.1 Strategic Constellations in the L1, L2, and Single-level Games: Locally Unwanted Land Uses The strategic constellations in this common goods problem are first compared for each of the three groups. Although the L0 actors are heterogeneous, the representatives are homogeneous if they truly represent the will of the majorities of their constituencies. The assumptions of a LULU problem usually imply that the majority in each jurisdiction is negatively affected by the facility. Thus, the homogeneity of the L1 representatives is a result of the aggregation mechanism used, namely, the majority rule. The effects of aggregation mechanisms shall be examined more closely later in this section. Four aggregation mechanisms will be compared on the basis of a fictitious LULU example: dictatorship, direct democracy, bargaining, and representative democracy. That is, the cooperative solution of the common goods problem at L1 and at L2 will be taken into account. The comparison no longer relates only to the basic strategic constellation, but to the outcomes of cooperative games. What is the position of the individual actors within the two subgroups? They are heterogeneous regarding their preferences for the locally unwanted facility. While most of them are negatively affected, if the facility is built within their own territory, some of them are positively affected, as they might find jobs there. For the model, it will be assumed that there are benefits (b) from the facility for everybody living within the reach of the facility’s services. For the disadvantaged neighbors (D) of the site, however, there are negative externalities (e). For a small group of advantaged citizens (A), there are additional benefits (a) from the facility. The standard benefit from the common good is less than its externalities to each of the disadvantaged (b , e). The sum of the standard and the additional benefit is larger than the externalities for the advantaged neighbors (b 1 a . e). As a result, the advantaged citizens favor the locating of the facility within their own territory, while the disadvantaged favor the building of the facility on the other jurisdiction, strongly rejecting the option to locate it at home. The preference orders are given in table 10.1. The individual actors in each of the subgroups would play a harmony game, if they had the power to really decide whether to offer the site. As the individuals in the Euro game, they cannot really act on their own; they can just voice preferences. The disadvantaged citizens will vote against the offering of the site by their own
211
D I S T O R T I O N T H R O U G H AG G R E G AT I O N
Table 10.1
The Siting Game at the Level of L0 Actors
Assumptions
A: b 1 a . e; D: b , e
Strategy Combinations
Benefits
A A, D
A: O
B: O
b1a B: ~O b 1 a A: ~O B: O b A: ~O B: ~O 0 A: O
Externalities
D
A
D
b
e
e
b
e
e
b 0
0 0
0 0
Payoffs
A
Ordinal
D
A
D
b 1 a2e b2e
3
1
b 1 a2e b2e b b 0 0
3
1
2 1
3 2
Game Matrix
Individual D
Offer site
Offer site
Do not offer site
b 1 a 2 e, b 2 e 3, 1
b 1 a 2 e, b 3, 3
b, b 2 e 2, 1
0, 0 1, 2
Individual A Do not offer site
jurisdiction, while the advantaged citizens will vote in favor of offering the controversial facility. The harmony game has an equilibrium where one individual would vote for offering the site and the other would not. If the game is played among the representatives at L2, that is, the governments of the two jurisdictions, the strategic structure changes. These actors do in fact have the power to decide whether to offer a site in their own territory or not. They truly represent the view of the majority of their constituents, that is, the view held by disadvantaged citizens. The L2 game is a chicken game, as modeled in chapter 3. From the perspective of governments, collective costs and benefits are different than from the perspective of individual citizens. The overall benefits of the facility outweigh its externalities (b 1 a . e), but additional benefits are valued less than the external costs (a , e). The first preference of the governments is, not to offer the site and hope that the other jurisdiction will do so. Their last preference is that the facility is built nowhere.
212
T R A N S N AT I O N A L C O M M O N G O O D S
The situation is different for the whole collective of L0 actors. There are advantaged and disadvantaged actors in both jurisdictions. The advantaged actors in jurisdiction I will vote for building the facility in jurisdiction I together with the disadvantaged actors of jurisdiction J. The disadvantaged actors of jurisdiction I will vote for locating the facility in jurisdiction J together with the advantaged actors there. There are pro-I and pro-J actors within the whole collective. Thus, the strategies are different: “offer site in J” or “offer site in I.” The game would be a harmony game with the equilibrium of (voting for) offering site J for the pro-J actors and offering site I for the pro-I actors. The interesting point relating to the strategic situation of the whole collective is that there are cross-jurisdictional coalitions of pro-I and pro-J actors. Therefore, in this case, the strategic constellation changes in a multi-level system. What would be a harmony game to pursue split strategies at the level of L0 actors, both within the subgroups and within the collective as a whole, is a chicken game at the level of the jurisdictions. The change is owing to the effect of the aggregation of L0 actors’ preferences by the L1 representatives. The noncooperative outcomes are no solution to the actual problem, however, as one cannot build a facility (in I, J) and not build it at the same time. In this situation, a coordinated strategy is needed, namely, either to build the facility in I or to build it in J. The noncooperative harmony outcomes do not lead to the locating of the facility, as the actors are not unanimous in which strategy to choose. The chicken game implies a similar risk: The jurisdictions may “miss each other,” meaning that either no facility will be built or that two of them will—whereby neither outcome is efficient. 10.2 Aggregation Effects in Cooperative Multi-level Solutions: Locally Unwanted Land Uses In the remainder of this section, I deal with cooperative solutions to the problem of siting noxious facilities. Such solutions can be found in different kinds of aggregation mechanisms, namely, dictatorship, direct democracy, bargaining, and representative democracy. Both cooperative solutions at L1 and at L2 are taken into account. To this end, the example of siting a LULU needs to be further articulated. I work with a numerical example, which can reveal a number of effects of aggregation mechanisms on cooperative solutions. All assumptions of the fictitious example are listed in table 10.2 and described in the following text.
213
D I S T O R T I O N T H R O U G H AG G R E G AT I O N
Table 10.2 Group
Number Benefit Costs Net Total Total Net p.c. p.c. Benefit Net Benefit J p.c. Benefit I
1 Neighbors I
Employees I
Neighbors J
Employees J
Other Citizens
Total
Costs and Benefits of the Facility
2
3
4
5
2,000
1
2
21
500
1,000
500
6,000
10,000
3
1
3
1
---
0
4
1
0
---
3
23
2
1
---
6 22,000
1,500
1,000
500
6,000
7,000
Voters of Government Opposition Case a
Case b
7
8
9
2,000
1,200
1,800
800
200
300
400
200
100
500
23,000
1,000
6,000
6,500
600
100
400
900
300
100
200
400
3,600
3,600
2,400
2,400
6,000
6,000
4,000
4,000
The structure of the siting problem is as follows. There is a jurisdiction at the upper level, L2, and two jurisdictions at the lower level, L1. A certain locally unwanted facility is to be built in L2, the general desirability of which is uncontested. Both jurisdictions, I and J, are capable to offer a site for the facility. Potential neighbors of the site suffer from negative externalities. However, there are some other citizens who gain from the existence of the facility. Finally, there are a number of citizens who are not directly affected by externalities or additional benefits, but profit from the general benefit the facility provides. • The L2 jurisdiction has 10,000 inhabitants altogether. In jurisdiction I, 2,000 neighbors would be negatively affected by the facility, while 500 persons could expect to be future employees in the facility. In jurisdiction J, only 1,000 neighbors would suffer from the negative externalities, while 500 would be future employees. The remaining 6,000 inhabitants have no special costs or benefits; each jurisdiction has 3,000 of these individuals (column 2).
214
T R A N S N AT I O N A L C O M M O N G O O D S
• Each citizen has a benefit of 1 from the facility. The potential employees have an additional benefit per person of 2 in both jurisdictions (column 3). • The neighbors in jurisdiction I have external costs of 2 per capita from the facility. For the neighbors in jurisdiction J, the external costs are much higher for ecological reasons, namely, 4 per capita. The potential employees in J also suffer from the externalities, but only have costs of 1 per capita. The employees in I, as well as all other citizens, do not suffer from the externalities (column 4). • As one of the aggregation mechanisms explored is representative democracy, assumptions about the distribution of voters within the whole area must be made. The governing party at L2 won 60 percent of the votes in the last election; the opposition party won 40 percent. With respect to voter distribution between the jurisdictions, two cases are distinguished. In case a, voters of the governing and the opposition party are proportionally distributed over the jurisdictions. In case b, the governing party has a stronghold in jurisdiction I, and the opposition party has a stronghold in jurisdiction J (columns 8 and 9). • In column 5 the net benefit per capita, in columns 6 and 7 the collective net benefit for each group and for the two jurisdictions are given. For jurisdiction I, the collective net benefit is 7,000, for jurisdiction J it is 6,500. The welfare maximizing decision would be to select the site in jurisdiction I for the facility. Which site will be chosen if each of the four aggregation mechanisms examined here is employed? I start by briefly defining dictatorship, direct democracy, bargaining, and representative democracy. • Aggregation by dictatorship means that there is a single decisionmaking center in the jurisdiction that can decide without being constrained by the constituency in any way. L0 actors have no say in the decision. • Aggregation by direct democracy means that the decision is taken by a simple or qualified majority of the actors at L0. • Aggregation by bargaining means that the different groups of L0 actors are each represented by a leader in a bargaining process. The decision is taken unanimously among these leaders. The representatives are bound by the preferences of their constituent groups. • Aggregation by representative democracy means that in each jurisdiction a government is elected by a simple majority of voters. As
D I S T O R T I O N T H R O U G H AG G R E G AT I O N
215
there are only two parties, the governing party and the opposition party, simple majority implies more than 50 percent of the votes. The government takes the decision and, in doing so, it is bound by the preferences of the majority that voted for it. In a two-level system, different kinds of aggregation mechanisms can appear together, as there may be one kind of mechanism at level L1 and another one at level L2. If we assume that all L1 jurisdictions use the same aggregation mechanism, there are 16 possible combinations. However, it might well be that the jurisdictions at L1 use different mechanisms. For example, there might be representative democracies or dictatorships at L1 combined with a bargaining system at L2, as often in international negotiations. In this case, the number of combinations rises quickly and if we additionally admit for more than two levels, the problem with this casuistic approach is aggravated. Thus, I examine only a number of very basic combinations, where either both levels use the same mechanism, or where the combinations are widespread. I stick to the two-level scenario and, moreover, assume that all jurisdictions at L1 indeed do use the same mechanism. One more general remark has to be made before the analysis can begin. The welfare maximizing solution is reached by each of the four mechanisms, if the decision makers—be they dictators, voters, representatives, or governments—are assumed to be motivated solely by sincere interest in the welfare of the public. In the case that the decision makers have complete information about all the facts given in table 10.2 and that their aim is to maximize public welfare, they will choose the site in jurisdiction I. The same is true for a fair distribution. If the decision makers are solely motivated by equity concerns and completely informed, they will decide in favor of a fair outcome. Dictatorship What decision would a dictator take, who is fully informed, but cannot be assumed to be interested in maximizing public welfare? The result will depend on the motives and the situation of the dictator. In a single-level system, a dictator might decide to build the facility in J, just because he or she happens to live in the neighborhood of the site in jurisdiction I. If he or she lives in the neighborhood of J, then he or she would probably prefer that the facility be built in I, and so on. Thus, no general prediction of the dictator’s choice can be made. It is obvious, however, that dictatorship cannot guarantee a welfare maximum, as long as the dictator is not purely benevolent.
216
T R A N S N AT I O N A L C O M M O N G O O D S
Similarly, in a two-level system, if there are two dictators in the jurisdictions, it is unclear whether they would provide the site or not. For example, if the dictator of I does not suffer from negative externalities, he or she would probably offer to provide the site; if he or she lives in the neighborhood to the site, he or she would reject the building of the facility in his or her own jurisdiction. Thus, depending on the actual circumstances, the two dictators might agree on site I or site J, or they may each favor a different site. If the decision at L2 were to be the result of bargaining by the two dictators, then, in the first two cases, a bargaining agreement could be found easily. In the third case, it cannot be predicted which bargain would be found, as this depends on the possibility of package deals or compensation (for a deeper analysis Raiffa 1985; Holzinger 1997). If the decision at L2 were to be taken by a dictator, as well, the votes of the L1 dictators would simply not count, and the result would be the same as earlier. Dictatorship is the least predictable aggregation mechanism. Direct Democracy and Majority Voting What would be the result of the decision on the site if the single-level system were a direct democracy? We assume that all citizens listed in table 10.2 are entitled to vote. The voters act rationally and in accordance with self-interest. Under those conditions, the suboptimal site in jurisdiction J would be selected in the single-level, direct democracy. The 6,000 other citizens would rationally abstain from voting, as they are indifferent to both two sites. Only the 4,000 citizens who are directly concerned with the facility would vote. This should suffice to reach a usual quorum in direct democracies. One thousand and five hundred citizens vote in favor of I, namely the neighbors of the site in J and the potential employees in I. Two thousand and five hundred citizens vote in favor of J, namely the neighbors of the site in I and the future employees in J. Thus, there is a clear majority for the suboptimal site in J. Direct democracy in the two L1 jurisdictions would lead to the acceptance of the site in both jurisdictions. The “other citizens” can be assumed to vote in favor of the facility. Then, in jurisdiction I, 2,000 would vote against the facility and 3,500 in favor of it. In J, there are 3,500 supporters vis-à-vis 1,000 citizens who reject the facility. This presumes that there would be a decision on whether to offer the site or not. If the decision were about which site should be chosen, I or J, the “other citizens” would abstain, and in both jurisdictions the vote would be that the other jurisdiction should offer the site. Thus,
D I S T O R T I O N T H R O U G H AG G R E G AT I O N
217
in direct democracy, the result depends on which alternatives are voted on: either to offer a site or not to offer it; or to support site I or site J. What happens at L2 in the two-level game? As discussed earlier, dictatorship at L2 leads to an unpredictable result. The possible results of bargaining, that is, a unanimous decision of the representatives of the majorities in the L1 jurisdictions, will be discussed in the following text. Direct democracy is not possible at L2—there is either direct democracy in a single-level system or at the L1 level in a multi-level system. L0 actors do not directly participate in L2 decision making in a multi-level system as defined in chapter 7. The equivalent of direct democracy at L2 is majority voting among the representatives. There are two ways to set the parameters for voting at L2: Either a “one-jurisdiction-one vote” rule is applied (the US senate principle), or there is proportional representation according to the size of the jurisdictions (the principle used in the German Bundesrat, or in the European Union (EU) Council so far). If there is perfect proportional representation and a simple majority rule is applied, the result is the same as in direct democracy in the singlelevel system. If each country has the same weight, or if there is imperfect proportional weighting of votes, the result will be distorted compared to direct democracy. Moreover, instead of a simple majority rule, all kinds of qualified majority rules may be applied. This creates room for further distortion. The closer the rule comes to unanimity, the more will decisions be distorted, given the different sizes of the jurisdictions. If the unanimity rule were to be applied in our example, the assent of both jurisdictions to one of the two sites would be needed. If both jurisdictions would in fact offer the site, it is unclear what kind of agreement would result. The same is true if both jurisdictions prefer that the other jurisdiction provide the site. Only if the unanimity requirement would imply a shift away from the logic of a “onejurisdiction-one-vote” system to the logic of bargaining, where the amount of costs and benefits is taken into account, does a solution become predictable (see the following section on Bargaining ). In our example, the “one-jurisdiction-one-vote” rule combined with simple majority rule is difficult to apply, as there are only two jurisdictions. In the two-jurisdiction case, majority decision-making converges toward unanimity. If there are more than two jurisdictions, coalitions can be formed. For example, there might be three jurisdictions in our illustration (I, J, and K), where K consists of the 6,000 citizens who are not directly concerned with the siting of the facility. K cannot offer a site as a result of natural conditions in its territory. In this case,
218
T R A N S N AT I O N A L C O M M O N G O O D S
two possible coalitions can be formed: either I and K outvote J and decide to build the facility in J or J and K outvote I, and select I as the site. Depending on which jurisdiction, I or J, offers K the most attractive coalition contract, the outcome might be that I or J must offer the site. There is no guarantee that the optimal site I will be chosen, or that the distribution associated with the outcome can be assessed fair. We saw that majority voting and unanimous decision making at level L2 can lead to distortions of the outcome compared to the outcome of majority voting in a single-level system. If the jurisdictions at L1 differ in size, distortion through representation will take place whenever there is no perfectly proportional representation in a system of weighted voting; whenever a one-jurisdictionone-vote system and whenever unanimity as a decision rule is used. In systems that apply weighted voting further distortion of the actual voting power compared to the share of votes may be created by the combination of the distribution of votes and the majority requirement. The exact nature and size of these distortions depend on the particulars of the situation. Aggregation of preferences over several levels by majority voting creates many opportunities for distortion. The outcomes of various combinations of decision rules (unanimity, simple, or qualified majorities) and representation rules (one state one vote, proportional, or some kind of weighing) can be analyzed by using voting power indices, such as the Shapley-Shubik or the Banzhaf index. Those indices measure the power one actor has in a committee to influence the decision. As they imply the prediction of “who wins” and thus an outcome of the decision, they can also be read as measures of the distortion of preferences that happens through aggregation in the presence of certain rules. I will not go deeper into the theory of coalition formation and power distribution in systems of weighted voting. However, this field shows how important decision rules in combination with representation rules are as sources of distortion through aggregation. The interested reader may have a look at Holler and Owen (2002). Bargaining Bargaining was defined as a way of finding unanimous agreement among the representatives of certain groups. These groups can be the subgroups of L0 actors in the jurisdictions or they can be groups that are formed on the basis of a common interest. It is assumed that each individual L0 member of the collective has to be represented in the
D I S T O R T I O N T H R O U G H AG G R E G AT I O N
219
bargaining process. How the groups select someone to represent them at the bargaining table need not be specified here. In the single-level game, the representatives decide unanimously in which jurisdiction the facility will be built. Which of the groups concerned will send representatives? Given the structure of the siting conflict, there are three possible bargaining positions: an actor can support the site in jurisdiction I, he or she can support the site in jurisdiction J, or he or she can be indifferent to the choice. The position pro-I is supported by the individuals advantaged by the facility in I and the neighbors to the site in J. The position pro-J is favored by the advantaged actors in J and the neighbors in I. The other inhabitants support the position I-or-J. Even if all six groups would send representatives to the bargaining table, coalitions between supporters of the same position would be formed quickly. Thus, it can be assumed that there are at least three actors present at negotiations—one pro I, one pro-J, and one I-or-J representative— irrespective of how many persons actually participate. The representatives or leaders are assumed not to diverge from the preferences of their constituents. The pro-I leader represents 1,500 persons who would have a benefit of 2,500 if the facility is built in I, and who would have costs of 2,500 if it is built in J. The 2,500 persons who favor J, would have a benefit of 3,000 from site J, but costs of 1,500 from site I. The other inhabitants would have a benefit of 6,000 regardless of the jurisdiction in which the facility is built. A unanimous decision for either I or J can only be achieved if the persons negatively affected by each decision are compensated for their losses. Unanimity guarantees a Pareto-optimal result. An agreement presupposes compensation, that is, redistribution. Rational bargainers will decide to build the facility in I, as the collective net benefit is greatest in this case. Consequently, the pro-J faction must be compensated. The leader of this faction will demand compensation of at least 1,500. However, the leader could also demand more, since the other groups have some positive benefit from building the facility in I, and the pro-J faction would like not only to avoid costs but also to participate in the benefits. Therefore, there is some leeway regarding the exact amount of compensation demanded. Similarly, it is open to negotiation, which of the other groups pays what share of the compensation. Both groups will be willing to pay at best a sum that guarantees their benefit to remain positive. In the example, this would amount to 8,500 for both groups together. Within these limits—at least 1,500 and at best
220
T R A N S N AT I O N A L C O M M O N G O O D S
8,500 of compensation payment for the pro-J faction—every possible distribution of the collective benefit can be achieved. Thus, the bargaining process leads to the selection of the welfaremaximizing and Pareto-optimal solution. No group will be worse off after the bargaining than it was before. There is one qualification, however: In real-world bargaining, information is usually incomplete. The parties may not really know which solution maximizes welfare. Moreover, under conditions of incomplete information they may prefer a solution that is inferior relating to collective welfare, because they can achieve a higher individual benefit with this solution. For example, parties in a bargaining process can achieve just such a result if they strategically withhold private information (cf., for the so-called negotiation dilemma, Scharpf 1997b: Chapter 6). On the basis of the assumptions made, it is not possible to predict how the exact distribution of the gain from cooperation will look.1 There is no guarantee that there is equal or fair distribution. What would be the result in a two-level game? Bargaining at the level L1 in the two jurisdictions would work like bargaining in the single-level game. The three groups in jurisdiction I would agree to build the facility in I, as there is a collective net benefit of 2,500 from offering the site. The neighbors in I had to be compensated for their costs of 22,000. This would be possible because the two other groups together have a benefit of 4,500. In J the collective net benefit is much smaller but still positive. The neighbors in J would demand compensation of at least 3,000. As the other groups have a benefit of 4,000, they would be able to compensate the neighbors. Consequently, both jurisdictions would be prepared to offer the site in negotiations at level L2. Bargaining at L2 in a two-level system would look different from single-level bargaining, because the different groups of L0 actors would no longer be represented at the bargaining table, but rather the jurisdictions. In our example, the two representatives of the jurisdictions would agree to build the site in jurisdiction I because the collective net benefit is highest there. The representative of jurisdiction J, however, should demand compensation because the net benefit for the L0 actors living in J is lower (1,000) than the net benefit for collective in J (2,500). Again, it is not possible to predict how the actual compensation scheme would look—either between the jurisdictions, or internally. An equal distribution rule would imply that each L0 actor gets the same benefit, which would be 0.7 in the example. However, it is not obvious that two-level bargaining would lead to this kind of equality argument. The representative of J might demand
D I S T O R T I O N T H R O U G H AG G R E G AT I O N
221
an equal split of benefits between the jurisdictions, neglecting the fact that they are different in size. Bargaining leads thus to the same solutions in both a single-level system and a two-level system. Perfectly rational actors will choose the Pareto-optimal solution. Even if this implies the necessity for compensation, it is nevertheless impossible to predict the actual distribution of gains from cooperation. With respect to distribution, there might be a difference between the single-level and the multilevel solution to a common goods problem. Representative Democracy The decision maker in a representative democracy is the government. The government is assumed to be self-interested; its goal is reelection. For the model, the government’s decisions are determined by how many votes it can gain or loose from a certain action. The distribution of voters for the governing and the opposition parties is assumed to be independent of the decision on the siting of the noxious facility. This distribution is known from a timely poll. If the government or the opposition party opt for one of the two sites in I and J, the voters positively or negatively affected by this decision will change their voting behavior in the next election and vote for the other party. Therefore, the decision of the government depends on what the opposition chooses, as well. The actual decision is a result of strategic interaction—a game between the government and the opposition party. Which of the two sites will be chosen by the government in a single-level system depends on the distribution of the voters. Two cases are distinguished in table 10.2. In case a, the voters are evenly distributed throughout the whole area. The decision to vote by indifferent inhabitants—that is, those who are not concerned with the facility—is not influenced by this issue. The government will get these 3,600 votes anyway (table 10.2). If it opts for J, it will also get the votes of the neighbors in I (1,200) and the prospective employees in J (300). However, the government will lose the votes of prospective employees in I (300) and the neighbors in I (600). If, at the same time, the opposition chooses I, then the government gets additional votes from former opposition supporters among the neighbors in I (800) and prospective employees in J (200). In this case the government wins 6,100 votes. If the opposition chooses J, as well, these 1,000 voters remain supporters of the opposition. In this case, the government would get only 5,100 votes. The votes can be calculated in the same way, if the government opts for
222
T R A N S N AT I O N A L C O M M O N G O O D S
Table 10.3 The Siting Game in a Representative Democracy: Even Distribution of Voters Game Matrix
Opposition Site I
Site J
Site I
4,500; 3,000 1, 1
5,100; 4,900 2, 4
Site J
6,100; 3,900 3, 2
5,100; 3,400 2, 2
Government
jurisdiction J as a site for the facility. Votes for the opposition party are calculated in the same fashion. The results are given in the game matrix (table 10.3). The game is a chicken game. Either the government opts for J and the opposition favors I or the government opts for I and the opposition favors J. If voters are evenly distributed, it pays for the government and the opposition to opt for just the opposite of the other side. The government prefers the first equilibrium (opting for J), while the opposition prefers the second equilibrium (also opting for J). However, in this example, there is a unique solution, since to opt for J is, for the government, a weakly dominant strategy. If the opposition favors J, the government is indifferent to both I and J. Therefore, the government will decide that the facility will be built in jurisdiction J under any circumstances (and the opposition will opt for I). Thus, if the voters of the governing and the opposition party are distributed evenly over the jurisdictions, the result is the same as it would be in a direct democracy: the suboptimal site J is chosen. Which site does the government select in case b, where it has a stronghold in jurisdiction I, and the opposition has a stronghold in jurisdiction J? For each strategy combination, the votes that can be gained are given in the game matrix (table 10.4). This game is a harmony game; it has a unique and Pareto-optimal equilibrium. The government has a dominant strategy to choose J for the site: J is the stronghold of the opposition party and the government has least to lose there. The opposition chooses jurisdiction I, because I is the governments stronghold. In case b, as well, the suboptimal site J will be chosen by the government. This result changes only if we assume that there is a stronghold of the government in jurisdiction J and a stronghold of the opposition
223
D I S T O R T I O N T H R O U G H AG G R E G AT I O N
Table 10.4 The Siting Game in a Representative Democracy: Strongholds of Parties in the Jurisdictions Game Matrix
Opposition Site I
Site J
Site I
4,100; 3,400 1, 2
5,100; 4,900 2, 4
Site J
6,100; 3,900 4, 2
5,500; 3,000 3, 1
Government
in jurisdiction I. Under these assumptions, the government decides to build the facility in I and the opposition opts for J as a site. The welfare-maximizing site would thus be selected. The examples show that a democratically elected government will prefer to build a locally unwanted facility in a stronghold of the opposition and never build it in its own stronghold. In general, the decision of the government depends on the distribution of the voters within the lower level jurisdictions. If we assume the government wants to maximize votes, representative democracy will not be able to guarantee the selection of the welfare-maximizing site for a locally unwanted facility. What would happen in a two-level system if representative democracy is used at level L1? The two governments would calculate the effect of their decision on voter support in their respective jurisdictions. Their decisions—and thus the solution for each jurisdiction— depend on the distribution of the supporters of both parties (governing and opposition) over the three groups: neighbors of the site, potential employees, and other citizens. There is an additional element, however, that leads to even more variation. As with direct democracy at the level of L1 jurisdictions, the solution also depends on which alternatives are exactly decided upon: either to offer the site or not to offer it, or to support jurisdictions I or J as sites for the facility. In the first pair of options the “other citizens” also have a preference, namely, that the facility is built; thus, they must be taken into account in the government’s calculus. In the second pair of options, these voters are indifferent and therefore will not change their party support as a result of this particular political issue. Depending on the particulars in each jurisdiction, the governments at L2 have different positions: either they both would like to offer the site, or both would
224
T R A N S N AT I O N A L C O M M O N G O O D S
prefer that the other jurisdiction offers the site, or one is willing to offer the site and the other not. What could representative democracy look like at L2? The only way to have representative democracy at L2 in a multi-level system is that there be double representation of L0 actors at L2: If there were only one representative body at L2—a parliament or a government— where L0 actors are directly represented, this would be equivalent to a single-level representative democracy (centralization). In a two-level system, a second body is needed where the L1 jurisdictions as such are represented. The L0 actors are represented according to their party preferences in the first body, and they are represented according to their lower level citizenship in the second body. This corresponds empirically to bicameral systems in federal states. Basically, there are two ways in which constituents may be represented in the second chamber: L0 actors may be represented by directly elected representatives or by the L1 governments. Decision making at L2 requires interaction of the two representation bodies. Because there are not only different ways in which the constituents may be represented in the second chamber, different election systems for both chambers, and different voting rules within the chambers, but also different rules of interaction of both bodies; the outcome of the procedure (and thus the solution to the siting problem) is determined by so many aggregation rules that no general prediction is possible. I will not attempt to single out a particular system by making assumptions and to find out what the effects are in this system. There is only one general statement that can be made for the solution of common goods problems in multi-level representative democracies: The solution found in the single-level system may be (very) different from the solution found in the multi-level system, but this need not necessarily be so. There are many factors in the aggregation process creating many opportunities for distortion. The effects of distortion, however, need not necessarily be cumulative: they may also neutralize each other. In sum, it has been shown that in case of a double dictatorship at the two levels the outcome of the political process is unpredictable and that the outcome may differ between a single-level and a multilevel system. In direct democracy and simple majority voting at the upper level, there is no guarantee that the welfare maximizing solution is chosen and the outcome may also differ between a single-level and a multi-level system, depending on the concrete rules. The same is true for representative democracy, where it is most likely that distortion through aggregation will be created. Only in case of a double
D I S T O R T I O N T H R O U G H AG G R E G AT I O N
225
bargaining system the welfare maximizing solution will be chosen at both levels of government, even if the strategic constellation looks different at the different levels. I finish this section with a few remarks on the empirical relevance of the different combinations of aggregation mechanisms at the different levels in a multi-level system. Usually, the member states in federal systems or in international organizations are representative democracies. Some of them have a high degree of direct democracy in their respective political systems, for example, Switzerland or California. For most of them, bargaining also plays a vital role in the aggregation of political preferences and the transformation into governmental positions. In general, however, representative democracy is the most prominent feature of the aggregation system. There is also dictatorship at L1 jurisdictions: Parties of international regimes are often autocracies. At the upper level, dictatorship is less common. To give an example, in the former Soviet Union, there was dictatorship of the communist party at the upper level of the union. There is no dictatorship in international regimes, in the EU, and in regional organizations in the Western world—at least not in the formal definition of dictatorship used here. In international regimes, the aggregation mechanism used at the upper level is usually bargaining or unanimous voting. In federalist states, the mechanism used at the upper level is representative democracy. In US dual federalism, the second chamber is directly elected; in German cooperative federalism, the second chamber is a representation of the L1 governments. The element of representative democracy, that is, representation via political parties and majority voting, is much stronger in dual federalism. Aggregation in dual federalism, with its vertical separation of power, is in effect close to a single-level system, at least relating to decisions within the realm of the powers allocated to the upper level. In cooperative federalism, the elements of intergovernmental decision making and bargaining are stronger. Aggregation at the upper level in the EU is a complex mixture of bargaining, unanimous or majority voting, and elements of representative democracy.
This page intentionally left blank
Chapter 11
Conclusion
T
his book has dealt with the provision of transnational common goods. Inevitably, transnational or global common goods have to be provided by international multi-level systems of governance. International multi-level systems are not states. In classical public goods theory this poses a particular challenge. In this view public goods have to be provided by the state, as the market is not capable of efficient public goods provision. The problem of common goods provision is equated with the strategic structure of a prisoners’ dilemma. To resolve this dilemma, a state-like structure is necessary that is able to monitor and sanction free riders. The point of departure of this book was the claim that this traditional perspective must be qualified. Neither is the state always capable of, or necessary for, the efficient provision of common goods. Nor is the strategic constellation posed necessarily a prisoners’ dilemma. Therefore, the provision of transnational common goods in the absence of the state is at least not in general doomed to failure. The question that remained to be answered was whether international multi-level provision is strategically different from common goods provision in single-level systems of governance. Theory and empirical research on common goods have already shown that the idea of a common good is not equivalent to providing it by the state. There are at least five reasons for this. First, the state may not be able to provide the good more efficiently than the market, as it faces an information problem with respect to the preferences of potential consumers. Second, in some cases the character of the good is such that private provision is possible, as with club goods or marketable public goods. Third, whenever binding agreements and enforceable contracts are the only solution to a common goods
228
T R A N S N AT I O N A L C O M M O N G O O D S
provision problem, the power to enforce the solution need not necessarily be a nation-state. Fourth, empirical research has shown that self-governance is in fact possible. Finally, in cases of transnational common goods, there is no single nation-state that can provide the good. The cooperation of states is required. This book set out first, to show that problems of common goods provision are so manifold and complex that they can pose all kinds of collective action problems, and second, to inquire in which respect international multi-level provision makes a difference. It thus addressed the following questions: (1) What are the strategic constellations and incentive structures created by different characteristics of the goods, the actors concerned, and their social and institutional environment? Which types of collective action problem do these strategic constellations imply? (2) Do the strategic constellations change if transnational common goods must be provided by multi-level systems? Under which conditions will multi-level provision change the strategic constellation compared to single-level provision? Part I of the book was devoted to the first set of questions, part II to the second set. Context Characteristics and Collective Action Problems It has been demonstrated in chapters 3 through 5 that common goods problems may result in very different kinds of collective action problems. A great number of features of the concrete situation influence the strategic constellation. This may even lead to no collective action problem at all, although the respective common good clearly exhibits the classical properties of nonrivalry or nonexcludability. Important factors affecting the strategic structure apart from these two defining properties of common goods have proved to be: the cost-benefit constellations, that is, the valuation of the goods by the concerned actors, the supply-side properties or aggregation technologies, the heterogeneity of the actors, and prevalent institutional or social rules. The exact strategic constellation is a result of a combination of these (and other) properties. As has been demonstrated with empirical cases, the strategic constellation depends very much on how strongly the problem in question is simplified and which characteristics are modeled. From this analysis, some general conclusions can be drawn if certain factors are held constant. These conclusions are valid for pure public goods that follow a summation technology, if the actors are homogeneous and if there are no specific social or institutional factors
CONCLUSION
229
present that might change the game. Given these conditions, variation of properties leads to the following results: • The variation of cost-benefit configurations yields harmony games (1) if the individual costs are higher than the total benefits a player can derive from the good, provided that all players make a contribution or (2) if the individual benefits are higher than the individual costs of a contribution. In the first case, the good will not be provided, while in the second case it will be. If the individual costs are higher than the individual benefits of one contribution but less than the benefits from all contributions, the game is a prisoners’ dilemma. The good will not be provided. • In the case of marketable public goods, the associated game is an assurance game. Pure public goods lead to a prisoners’ dilemma. The same is true for common pool resources—at least in a phase of exploitation of the resource in which collective marginal benefits are negative while individual marginal benefits are still positive. • The strategic constellation is a prisoners’ dilemma if the aggregation technology follows the summation rule. In the case of a weakestlink technology, an assurance game, and in the case of a best-shot technology, a chicken game results. In general, nonadditive goods tend to render coordination games rather than dilemmas. • The heterogeneity of players may lead to any of the strategic constellations. In many cases it will result in asymmetric games with inequality in Pareto-optimal outcomes. However, other games are also possible, depending on which kind of preference orders are combined. In the case of the standard conditions assumed earlier, the games will probably combine prisoners’ dilemma preferences with harmony, chicken, or assurance preferences. This leads to rambo games and asymmetric dilemmas. Thus, many different games may appear in common goods provision, not only the prisoners’ dilemma. Therefore, a systematic examination of strategic constellations and collective action problems was undertaken. Seven types of collective action constellation can be distinguished that represent combinations of four problematic aspects: problems of efficiency, coordination, distribution, and stability. They can be described as follows: • Harmony games do not pose any collective action problem. There are neither efficiency, coordination, distribution, or stability problems.
230
T R A N S N AT I O N A L C O M M O N G O O D S
• Mere distributional problems are posed by rambo games and games of pure conflict. These games lead to efficient and stable outcomes. Their only problem is that their Pareto-optimal outcomes result in inequality, which may lead to manifest conflict. • Pure and impure coordination games entail the problem that the actors may not be able to coordinate their strategies at the desirable equilibrium. Some of these games, such as assurance games, also have an efficiency aspect, since they have optimal and suboptimal equilibria, and since there is a risk that the optimal equilibrium will not be achieved. • Defection problems are more severe than mere distribution and coordination problems. Dilemma games have a suboptimal equilibrium, in some cases combined with inequality in the Pareto-optimal outcome. Their most unpleasant feature is that even if the parties agree to play the cooperative strategies, they have an incentive to defect afterward. • Disagreement problems combine aspects of both coordination and distribution problems. Typical games in this group are chicken and battle of the sexes. They have two equilibria, each one preferred by one of the actors. Even if the coordination aspect can be overcome by communication or other means, the parties find it difficult to attain agreement. • Instability problems are posed by games that have no equilibrium in pure strategies. There will be no stable solution to these problems, because the actors try to discoordinate their strategies. Multi-level Provision of Transnational Common Goods Part II of the book was devoted to the question whether multi-level provision leads to changes in strategic structures as compared to the provision of common goods in a centralized system of governance. A review of the literature on multi-level systems and a theoretical discussion of the multi-level solution to common goods provision indicated that the heterogeneity of actors across different jurisdictions and across different levels of government seems to be important for strategic changes. Therefore, the analysis of multi-level provision concentrated on the effects of possible combinations of homo- and heterogeneity of different collectives of actors. Two-level provision was compared to single-level provision within the total collective that is concerned by the scope of the good. For such a comparison, it was necessary to consider the strategic constellation
CONCLUSION
231
within three groups of actors: the individual actors within the whole collective in a single-level system; the individual actors within the jurisdictions at the lower level in a multi-level system (the subgroups); and the representatives of the jurisdictions at the upper level. Each of the three groups of actors can be homogeneous or heterogeneous in its preferences for the common good. There are six logically possible combinations of heterogeneous or homogeneous subgroups, representatives, and collectives. Three of the combinations imply that the presence of multi-level systems does not alter the strategic constellation. This was demonstrated by continuation of the analysis of some examples from part I, which match the conditions for the three constellations: international lake pollution, international river pollution, biodiversity and global warming. In the cases where all three groups of actors are homogenous or heterogeneous the strategic constellation remains the same at all levels. In the case of internal homogeneity of the subgroups at the lower level and heterogeneity across subgroups the two-level constellation yields the same result as the single-level constellation would (although the game within the subgroups is different). For the remaining three combinations, the effects of multi-level provision make a difference not only in the strategic constellation but also in the political decisions taken. Tax coordination in the European Union served as an example for the case in which the subgroups and the whole collective are homogeneous, but the representatives are heterogeneous. The strategic constellation is identical in the singlelevel game and the game of the individual actors at the lower level. It is different, however, from the strategic constellation in the game of the representatives at the upper level. The introduction of the European Monetary Union provided the illustration for the case in which the subgroups are internally homogeneous, but there is heterogeneity between the subgroups. The representatives, however, were homogeneous in their preferences. The strategic constellation in the upper level game was thus different from the strategic constellation in the single-level game. The siting of locally unwanted facilities exemplified the case in which the individual actors are heterogeneous within the subgroups and the whole collective, whereas the representatives are homogenous. Again, the strategic constellation in the upper-level game is different from the single-level game. In all three cases occurs the change of the strategic constellation as a consequence of the divergence of the representatives from the individual actors’ preferences. The divergence may have two causes. First, the representatives may pursue goals that are different from the
232
T R A N S N AT I O N A L C O M M O N G O O D S
goals of their constituency; either they have official organizational goals or organizational self-interests such as reelection, or any other motives, which imply a utility function different from the one of the constituency. Second, the divergence may be a result of the aggregation mechanisms used. There are effects of aggregation that are a consequence of the mechanisms themselves and effects that result from size differences of the subgroups. Whereas the former are also present in single-level systems, the latter are typical for multi-level systems. The tentative examination of the effects of four different aggregation mechanisms on the solution of common goods problems in multi-level systems led to the following results: • The solutions chosen in dictatorship are solely determined by the self-interest of the dictator. As long as the utility function of the dictator is not specified, no general predictions are possible. This is true for both single-level and multi-level systems. Welfaremaximizing solutions or fair distribution will only be achieved by chance. • The mechanism of bargaining implies that the intensity of preferences is taken into account. The benefits and costs for each individual or each group represented at the bargaining table play a role in the decision. If bargainers are perfectly rational, they will select a Pareto-optimal and welfare-maximizing solution. Unanimity forces the parties to compensate other parties, in case these are worse off with a given solution compared to the status quo. Therefore, the bargainers have a strong incentive to “distribute the biggest cake.” However, it cannot be predicted how the gain from cooperation will be distributed among the parties. As long as the scope of the problem actually concerns all jurisdictions involved at the lower level, a bargained solution should be the same, regardless of whether it is achieved in a single-level system or in a multi-level system. What is not necessarily the same, however, is the distribution of the gain from cooperation. • The solutions in a direct democracy as well as in a representative democracy depend very much on how voters are distributed among the jurisdictions and among the heterogeneous interest groups, and which political party they support. For the aggregation process, only the numbers of the concerned subgroups are important, not their benefits and costs. Neither direct nor representative democracy can guarantee that welfare maximizing solutions or fair distributions will be achieved.
CONCLUSION
233
• In multi-level systems, there are further determinants of distortion resulting from aggregation, whenever majority voting is involved, be it in a direct or a representative democracy. First, which alternatives are actually presented for decision might be important. Second, in a multi-level system, the weight of the jurisdictions at the upper level can be based on equal or proportional representation. Equal representation of jurisdictions that differ in size can lead to massive distortion as compared to single-level decision making in a direct or representative democracy. Third, the exact weighing of votes in combination with majority requirements at the upper level can further distort the weight of the individual vote in a multi-level system. Overall, it has been shown that the strategic constellation in the provision of transnational common goods und thus the policy output will most likely change in a multi-level system, as compared to a single-level system, as a consequence of the self-interest of representatives at intermediate levels and of the aggregation mechanisms used. This presupposes, however, that there is some form of heterogeneity of the actors’ preferences across the different levels of government.
This page intentionally left blank
Notes
3
Case Studies 1: Attributes of the Goods
1. There is a third equilibrium in mixed strategies that is Pareto-inferior. In general, I will neglect randomization in the analysis of the examples. It is briefly discussed in chapter 6. 2. Haveman (1973) gives a more detailed analysis, which includes costs, revenue, and willingness to pay for the good produced by the polluting activity.
4
Case Studies 2: Attributes of the Groups
1. This section relies heavily on an excellent case study by Genschel (2002). A more detailed analysis along the lines pursued here is provided by Holzinger (2005). 2. As all models in this book, this model is extremely stylized. A more differentiated model can be found in Durth (1996: 46–48). The noncooperative version of his model yields the same result, however: Both players pollute and this is a Pareto-optimal solution.
5
Case Studies 3: Attributes of the Social and Institutional Context
1. Among the OECD countries, Germany is surely the country that has been least affected by disintermediation as a result of its traditional universal banking system (Vitols 1998; Sinclair 2002: 271). There are less than 100 companies, which have an international rating (Strulik 2002: 320). 2. It is similar to what is often called a zero-sum game. However, strictly spoken, in a zero-sum game the sums of all cells must add up to zero. This presupposes that a game is formulated cardinally.
236
NOTES
3. The metaphor of a race to the top (or bottom) implies a dynamic aspect: Many states adjust their national standards successively to the level of the highest (or lowest) standards, or they apply ever higher (or lower) standards. There is a move toward an equilibrium. In the logic of 2 x 2 matrix games, there are only two levels of standards and only two countries, which adjust behavior with a single step. The system is in equilibrium after two simultaneous moves. The dynamic aspect of the metaphor is thus lost as a result of analytical reduction.
6
Strategic Constellations and Collective Action Problems
1. In fact, there are twelve possible combinations, as the players’ positions can be reversed. This, however, does not change the basic strategic structures, although in case of rambo games it is now the other player who has the distributive advantage. 2. My work was greatly facilitated by the fact that these 78 games are provided by Rapoport, Guyer, and Gordon 1976: 23–30. 3. There may be games with four Pareto-optimal outcomes ranked completely oppositional and lying on a nonlinear curve.
7 Transnational Common Goods and Multi-level Systems: Analytical Framework 1. See, for example, the titles of König, Rieger, and Schmitt 1996; Marks, Hooghe, and Blank 1996; Benz 1998; Grande and Jachtenfuchs 2000; Scharpf 2000b.
10 Case Studies 6: Distortion Effects through Aggregation 1. Predicting a solution would only be possible if an axiomatic model were applied (e.g., the Nash bargaining solution; Nash 1950), or if more specific assumptions were made.
References
Adams, Roy and Ken McCormick (1993). The Traditional Distinction between Public and Private Goods Needs to Be Expanded, Not Abandoned. Journal of Theoretical Politics 5: 109–116. Aggarwal, Vinod K. and Cedric Dupont (1999). Goods, Games, and Institutions. International Political Science Review 20: 393–409. Aghion, Philippe, Patrick Bolton, and Mathias Dewatripont (2000). Contagious Bank Failures in a Free Banking System. European Economic Review 44: 713–718. Al Baz, Ismail, Volkmar Hartje, and Waltina Scheumann (eds.) (2002). Co-operation on Transboundary Rivers. Baden-Baden: Nomos. Allen, Franklin and Douglas Gale (2000). Financial Contagion. Journal of Political Economy 108: 164–174. Arce, M., Daniel G. and Todd Sandler (2001). Transnational Public Goods: Strategies and Institutions. European Journal of Political Economy 17: 493–516. Bagnoli, Mark and Michael McKee (1991). Voluntary Contribution Games: Efficient Private Provision of Public Goods. Economic Inquiry 29: 351–366. Banks, Jeffrey S., Charles R. Plott, and David P. Porter (1988). An Experimental Analysis of Unanimity in Public Goods Provision Mechanisms. Review of Economic Studies 55: 301–322. Barkin, Samuel J. and George E. Shambaugh (1999). Anarchy and the Environment: The International Relations of Common Pool Resources. New York: State University of New York Press. Barrett, Scott (1998). On the Theory and Diplomacy of Environmental Treaty-making. Journal of Resource Economics 11: 317–333. ——— (1999). Montreal versus Kyoto: International Cooperation and the Global Environment. In Global Public Goods: International Cooperation in the 21st Century (Inge Kaul, Isabelle Grunberg, and Marc A. Stern, eds.), pp. 192–219. New York-Oxford: Oxford University Press. Basel Committee on Banking Supervision (2000). Credit Ratings and Complementary Sources of Credit Quality Information. Basel Committee on Banking Supervision Working Papers No. 3. Basel: Bank for
238
R EF E R EN C E S
International Settlements, August 2000 (http://www.bis.org/publ/ bcbs_wp3.htm, July 18, 2008). Benedick, Richard E. (1999). Contrasting Approaches: The Ozone Layer, Climate Change, and Resolving the Kyoto Dilemma. WZB Discussion Paper FS II 99–404. Berlin: Wissenschaftszentrum für Sozialforschung. Benz, Arthur (1994). Kooperative Verwaltung. Funktionen, Voraussetzungen und Folgen. Baden-Baden: Nomos. ——— (1998). Politikverflechtung ohne Politikverflechtungsfalle Koordination und Strukturdynamik im europäischen Mehrebenensystem. Politische Vierteljahresschrift 39: 558–589. Benz, Arthur, Fritz W. Scharpf, and Reinhard Zintl (1992). Horizontale Politikverflechtung: Zur Theorie von Verhandlungssystemen. FrankfurtNew York: Campus. Blankart, Charles B. and Günther Knieps (1992). Netzökonomik. Jahrbuch für Neue Politische Ökonomie 11: 73–87. Blatter, Joachim (2000). Entgrenzung der Staatenwelt? Politische Institutionenbildung in grenzüberschreitenden Regionen in Europa und Nordamerika. Baden-Baden: Nomos. Blümel, Wolfgang, Rüdiger Pethig, and Oskar von dem Hagen (1986). The Theory of Public Goods: A Survey of Recent Issues. Journal of Institutional and Theoretical Economics 142: 241–309. Bohm, Peter (1972). Estimating Demand for Public Goods: An Experiment. European Economic Review 3: 111–130. Bonus, Holger (1980). Neue Politische Ökonomie und Öffentliche Güter. In Neue Politische Ökonomie als Ordnungstheorie (Erik Boettcher, Philipp Herder-Dorneich, and Karl-Ernst Schenk, eds.), pp. 153–172. Tübingen: Mohr. Bräuninger, Thomas and Thomas König (2000). Making Rules for Governing Global Commons. The Case of Deep-Sea Mining. Journal of Conflict Resolution 44: 604–629. Brennan, George and James M. Buchanan (1980). The Power to Tax: Analytical Foundations for a Fiscal Constitution. Cambridge: Cambridge University Press. Buchanan, James M. (1965). An Economic Theory of Clubs. Economica 32: 1–14. ——— (1968). Demand and Supply of Public Goods. Chicago: RandMcNally. Buck, Susan J. (1998). The Global Commons. An Introduction. Washington, DC: Island Press. Bucovetsky, Sam and John D. Wilson (1991). Tax Competition with Two Tax Instruments. Regional Science and Urban Economics 21: 333–350. Burghof, Hans-Peter (1998). Eigenkapitalnormen in der Theorie der Finanzintermediation. Berlin: Duncker & Humblot. Camerer, Colin (1997). Progress in Behavioral Game Theory. Journal of Economic Perspectives 11: 167–188.
R EF E R EN C E S
239
Cantor, Richard and Frank Packer (1994). The Credit Rating Industry. Federal Reserve Bank of New York Quarterly Review 19: 1–26. Chen, Lincoln C., Evans, Tim G., and Cash, Richard A. (1999). Health as a Global Public Good. In Global Public Goods. International Cooperation in the 21st Century (Inge Kaul, Isabelle Grunberg, and Marc A. Stern, eds.), pp. 284–305. New York-Oxford: Oxford University Press. Cheung, Stephen N. (1970). The Structure of a Contract and the Theory of a Non-exclusive Resource. Journal of Law and Economics 13: 49–70. Coase, Ronald H. (1960). The Problem of Social Cost. Journal of Law and Economics 3: 1–44. Cornes, Richard and Todd Sandler (1994). The Comparative Statics of the Impure Public Good Model. Journal of Public Economics 54: 403–451. ——— (1996). The Theory of Externalities, Public Goods, and Club Goods. New York: Cambridge University Press. Cosmides, Leda and John Tooby (1994). Better than Rational: Evolutionary Psychology and the Invisible Hand. American Economic Review 84: 327–332. Dasgupta, Partha, Karl-Göran Mäler, and Allessandro Vercelli (eds.) (1997). The Economics of Transnational Commons. Oxford: Clarendon Press. Dawes, Robin, Jeanne McTavish, and Harriet Shaklee (1977). Behavior, Communication, and Assumptions about Other People’s Behavior in a Commons Dilemma Situation. Journal of Personality and Social Psychology 35: 1–11. De Bandt, Olivier and Philipp Hartmann (2002). Systemic Risk in Banking: A Survey. In Financial Crises, Contagion, and the Lender of the Last Resort. A Reader (Charles Goodhart and Gerhard Illing, eds.), pp. 249–297. New York: Oxford University Press. Deeg, Richard and Susanne Lütz (2000). Internationalization and Financial Federalism. The United States and Germany at the Crossroads? Comparative Political Studies 33: 374–405. Deheija, Vivek H. and Philipp Genschel (1999). Tax Competition in the European Union. Politics & Society 27: 403–430. Demsetz, Harold (1967). Toward a Theory of Property Rights. American Economic Review 57: 347–359. Diekmann, Andreas (1992). Soziale Dilemmata. Modele, Typisierungen und empirische Resultate. In Theorie, Daten, Methoden: Neuere Modele und Verfahren in den Sozialwissenschaften (Hans-Jürgen Andress, ed.), pp. 176–203. Rieden: WB-Druck. Durth, Rainer (1996). Grenzüberschreitende Umweltprobleme und regionale Integration. Zur Poltischen Ökonomie von Oberlauf-UnterlaufProblemen an internationalen Flüssen. Baden-Baden: Nomos. Edwards, Jeremy and Michael Keen (1996). Tax Competition and Leviathan. European Economic Review 40: 113–134. Ehrlich, Paul R. and Anne H. Ehrlich (1997). The Value of Biodiversity. In The Economics of Transnational Commons (Partha Dasgupta, Karl-Göran Mäler, and Allessandro Vercelli, eds.), pp. 97–117. Oxford: Clarendon Press.
240
R EF E R EN C E S
Eichenberger, Reiner (1996). Eine «fünfte Freiheit» für Europa: Stärkung des politischen Wettbewerbs durch «FOCJ». Zeitschrift für Wirtschaftspolitik 45: 110–130. Epiney, Astrid (2000). Flexible Integration and Environmental Policy in the EU – Legal Aspects. In Environmental Policy in a European Union of Variable Geometry? The Challenge of the Next Enlargement (Katharina Holzinger and Peter Knoepfel, eds.), pp. 39–64. Basel: Helbing & Lichtenhahn. European Commission (2001). Proposal for a Council Directive to Ensure a Minimum of Effective Taxation of Savings Income in the Form of Interest Payments Within the Community. COM/01/0400/final – CNS01/0164, OJ C 270E, 25.09.01. ——— (2002). Report on the negotiations with third countries about the taxation of interests. Communication by the Commission to Council Ecofin, 03.12.02 (internal document). European Council/Ecofin (2000). Report on the tax package. Santa Maria da Feira, European Council (Doc. 1355500 FISC 190). Fehr, Ernst and Karl M. Schmidt (1999). A Theory of Fairness, Competition, and Cooperation. Quarterly Journal of Economics 114: 817–868. Fehr, Ernst, Erich Kirchler, Andreas Weichbold, and Simon Gächter (1998). When Social Norms Overpower Competition: Gift Exchange in Experimental Labour Markets. Journal of Labour Economics 16: 324–351. Feldman, Allan M. (1997). Welfare Economics and Social Choice Theory. 12th edition. Boston et al.: Kluwer-Nijhoff. Frey, Bruno S. (1990). Intergovernmental Tax Competition. In Influence of Tax Differentials on International Competitiveness (Charles McLure, Hans-Werner Sinn, and Richard Musgrave, eds), pp. 87–114. Deventer: Kluwer Academic Publishers. ——— (1996). A Directly Democratic and Federal Europe. Constitutional Political Economy 7: 267–279. ——— (1997). Ein neuer Föderalismus für Europa: Die Idee der FOCJ. Tübingen: Mohr Siebeck. Frey, Bruno S. and Iris Bohnet (1996). Cooperation, Communication and Communitarism: An Experimental Approach. Journal of Political Philosophy 4: 322–336. Frey, Bruno S. and Reiner Eichenberger (1996). To Harmonize or to Compete?: That’s Not the Question. Journal of Public Economics 60: 335–349. Gautschi, Thomas (2000). History Effects in Social Dilemma Situations. Rationality and Society 12: 131–162. Gavin, Michael and Ricardo Hausmann (2000). Preventing Crisis and Contagion: Fiscal and Financial Dimensions. In Wanted: World Financial Stability (Eduardo Fernández-Arias and Ricardo Hausmann, eds.), pp. 77–89. Washington, DC: Johns Hopkins University Press. Genschel, Philipp (2002). Steuerharmonisierung und Steuerwettbewerb in der Europäischen Union. Frankfurt a.M.: Campus.
R EF E R EN C E S
241
Genschel, Philipp and Thomas Plümper (1996). Kommunikation und Kooperation in der internationalen Bankenregulierung. Zeitschrift für Internationale Beziehungen 3: 225–253. ——— (1997). Regulatory Competition and International Co-operation. Journal of European Public Policy 4: 626–642. Gibbons, Robert (1992). Game Theory for Applied Economists. Princeton, NJ: Princeton University Press. Giovannini, Alberto and Hines, James R. (1991). Capital Flight and Tax Competition: Are There Viable Solutions to Both Problems? In European Financial Integration (Alberto Giovannini and Colin Mayer, eds.), pp. 172–220. Cambridge: Cambridge University Press. Görg, Christoph and Ulrich Brand (2000). Global Environmental Politics and Competition between Nation-States: On the Regulation of Biological Diversity. Review of International Political Economy 7: 371–398. Grande, Edgar (2000). Multi-Level Governance: Institutionelle Besonderheiten und Funktionsbedingungen des europäischen Mehrebenensystems. In Wie problemlösungsfähig ist die EU? Regieren im europäischen Mehrebenensystem (Edgar Grande and Markus Jachtenfuchs, eds.), pp. 11–30. Baden-Baden: Nomos. Grande, Edgar and Markus Jachtenfuchs (eds.) (2000). Wie Problemlösungsfähig ist die EU? Regieren im europäischen Mehrebenensystem. Baden-Baden: Nomos. Group of Thirty (1997). Global Institutions, National Supervision and Systemic Risk. A Study Group Report. Washington, DC: Group of Thirty. Güth, Werner and Hartmut Kliemt (1995). Competition or Co-operation. On the Evolutionary Economics of Trust, Exploitation and Moral Attitudes. Working Paper. Berlin: Humboldt University. Güth, Werner, Steffen Huck, and Wieland Müller (2001). The Relevance of Equal Splits in Ultimatum Games. Games and Economic Behavior 37: 161–169. Güth, Werner, Rolf Schmittberger, and Bernd Schwarze (1982). An Experimental Analysis of Ultimatum Bargaining. Journal of Economic Behavior and Organization 3: 367–388. Hallerberg, Mark (1996). Tax Competition in Wilhelmine Germany and Its Implications for the European Union. World Politics 48: 324–357. Hardin, Garrett (1968). The Tragedy of the Commons. Science 162: 1243–1248. Hardin, Garrett and John Baden (eds.) (1977). Managing the Commons. San Francisco: Freeman and Company. Hardin, Russell (1971). Collective Action as an Agreeable n-Person Prisoners’ Dilemma. Science 16: 472–481. Hausken, Kjell and Thomas Plümper (1999). The Impact of Actor Heterogeneity on the Provision of International Public Goods. International Interactions 25: 61–94. Haveman, Robert H. (1973). Common Property, Congestion, and Environmental Pollution. Quarterly Journal of Economics 87: 278–287.
242
R EF E R EN C E S
Hellwig, Martin (1997). Systemische Risiken im Finanzsektor. In Finanzmärkte im Spannungsfeld von Globalisierung, Regulierung und Geldpolitik (Dieter Duwendag, ed.) pp. 123–151. Berlin: Duncker 1 Humblot. Héritier, Adrienne (2002a). Introduction. In Common Goods: Reinventing European and International Governance (Adrienne Héritier, ed.), pp. 1–13. Lanham, MD: Rowman & Littlefield. ——— (2002b). New Modes of Governance in Europe: Policy-Making without Legislating? In Common Goods: Reinventing European and International Governance (Adrienne Héritier, ed.), pp. 185–206. Lanham, MD: Rowman & Littlefield. Hirshleifer, Jack (1983). From Weakest-link to Best-shot: The Voluntary Provision of Public Goods. Public Choice 41: 371–386. ——— (1985). From Weakest-link to Best-shot: Correction. Public Choice 46: 221–223. Hofmann, Gerhard (ed.) (2001). Auf dem Weg zu Basel II: Konzepte, Modele, Meinungen. Frankfurt a.M.: Bankakademie Verlag. Holler, Manfred (1983). Collective Action, Rational Man and Economic Reasoning. Quality and Quantity 17: 163–177. Holler, Manfred und Guillermo Owen (eds) (2002). Power Measures II (Homo Oeconomicus 19). Munich: Accedo. Holzinger, Katharina (1994). Politik des kleinsten gemeinsamen Nenners? Umweltpolitische Entscheidungsprozesse in der EG am Beispiel der Einführung des Katalysatorautos. Berlin: edition sigma. ——— (1997). Kompensationen in alternativen Konfliktlösungsverfahren. Analyse & Kritik 19: 33–63. ——— (2000). Optimal Regulatory Units: A Concept of Regional Differentiation of Environmental Standards in the European Union. In Environmental Policy in a European Union of Variable Geometry? The Challenge of the Next Enlargement (Katharina Holzinger and Peter Knoepfel, eds.), pp. 39–64. Basel: Helbing & Lichtenhahn. ——— (2001). Aggregation Technology of Common Goods and Its Strategic Consequences: Global Warming, Biodiversity and Siting Conflicts. European Journal of Political Research 40: 117–138. ——— (2003). Common Goods, Matrix Games, and Institutional Solutions. European Journal of International Relations 9: 173–212. ——— (2005). Tax Competition and Tax Co-operation in the EU: The Case of Savings Taxation. Rationality and Society 17: 475–510. ——— (2008). Treaty Formation and Strategic Constellations. An Extension of Sandler, Illinois Law Review: 187–200. Holzinger, Katharina and Christoph Knill (2008). Competition, Cooperation and Communication: Theoretical Analysis of Different Sources of Environmental Policy Convergence and Their Interaction. Journal of Comparative Policy Analysis (forthcoming). Holzinger, Katharina, Christoph Knill, and Thomas Sommerer (2008). Environmental Policy Convergence? The Impact of International
R EF E R EN C E S
243
Harmonization, Transnsational Communication and Regulatory Competition. International Organization (forthcoming). Horowitz, Shale and Uk Heo (eds.) (2001). The Political Economy of International Financial Crisis. Interest Groups, Ideologies, and Institutions. Lanham, MD: Rowman & Littlefield. Isaac, R. Mark, Kenneth F. McCue, and Charles R. Plott (1985). Public Goods Provision in an Experimental Environment. Journal of Public Economics 26: 51–74. Isaac, R. Mark, James Walker, and Susan H. Thomas (1984). Divergent Evidence on Free Riding: An Experimental Examination of Possible Explanations. Public Choice 43: 113–149. Isaac, R. Mark, James Walker, and Arlington W. Williams (1994). Group Size and Voluntary Provision of Public Goods: Experimental Evidence Utilizing Large Groups. Journal of Public Economics 54: 1–36. Isaac, R. Mark, McCue, Kenneth F. and Plott, Charles R. (1985). Public Goods Provision in an Experimental Environment. Journal of Public Economics 26: 51–74. Kanbur, Ravi and Michael Keen (1993). Jeux sans frontières: Tax Competition and Tax Coordination When Countries Differ in Size. Economic Review 29: 81–85. Kapstein, Ethan B. (1989). Resolving the Regulator’s Dilemma: International Coordination of Banking Regulations. International Organization 43: 323–347. ——— (1992). Between Power and Purpose: Central Bankers and the Politics of Regulatory Convergence. International Organization 46: 265–287. ——— (1994). Governing the Global Economy. International Finance and the State. Cambridge, MA-London: Harvard University Press. Katz, Michael L. and Carl Shapiro (1985). Network Externalities, Competition, and Compatibility. American Economic Review 75: 424–440. ——— (1994). System Competition and Network Effects. Journal of Economic Perspectives 8: 93–115. Kaufman, George G. (ed.) (1995). Banking, Financial Markets, and Systemic Risk. Greenwich, CT: JAI Press. Kaul, Inge and Pedro Conceicao (2006). The New Public Finance: Responding to Global Challenges. Oxford: Oxford University Press. Kaul, Inge, Pedro Conceicao, and Katell le Goulven (2003). Providing Public Goods: Managing Globalization. Oxford: Oxford University Press. Kaul, Inge, Isabelle Grunberg, and Marc A. Stern (eds.) (1999). Defining Global Public Goods. In Global Public Goods: International Cooperation in the 21st Century (Inge Kaul, Isabelle Grunberg, and Marc A. Stern, eds.), pp. 2–19. New York-Oxford: Oxford University Press. Keen, Michael (1993). The Welfare Economics of Tax Co-ordination in the European Community: A survey. Fiscal Studies 14: 15–36. Keohane, Robert O. (1984). After Hegemony: Cooperation and Discord in the World Political Economy. Princeton, NJ: Princeton University Press.
244
R EF E R EN C E S
Keohane, Robert O. and Elinor Ostrom (eds.) (1995). Local Commons and Global Interdependence: Heterogeneity and Cooperation in Two Domains. London: Sage. Keohane, Robert O., Michael McGinnis, and Elinor Ostrom (eds.) (1993). Proceedings of a Conference on Linking Local and Global Commons, held at Harvard University, April 23–25, 1992, mimeo. Kern, Kristine (2000). Die Diffusion von Politikinnovationen. Umweltpolitische Innovationen im Mehrebenensystem der USA. Opladen: Leske 1 Budrich. Kerwer, Dieter (2002). Standardizing as Governance: The Case of Rating Agencies. In Common Goods: Reinventing European and International Governance (Adrienne Héritier, ed.), pp. 293–315. Lanham, MD: Rowman & Littlefield. Khor, Martin (2000). Why Capital Controls and International Debt Restructuring Mechanisms Are Necessary to Prevent and Manage Crises. In Global Finance. New Thinking on Regulating Speculative Capital Markets (Walden Bello, Nicola Bullard, and Kamal Malhotta, eds.), pp. 140–158. London-New York: Zed Books. Kim, Oliver and Mark Walker (1984). The Free Rider Problem: Experimental Evidence. Public Choice 43: 3–24. Koelliker, Alkuin (2001). Bringing Together or Driving Apart the Union? Towards a Theory of Differentiated Integration. West European Politics 24: 125–151. König, Thomas, Elmar Rieger, and Hermann Schmitt (eds.) (1996). Das europäische Mehrebenensystem. Frankfurt a. M.: Campus. Kreps, David M. (1990). Game Theory and Economic Modelling. Oxford: Clarendon Press. Laffont, Jean-Jacques (1988). Fundamentals of Public Economics. Cambridge: Cambridge University Press. Large, Sir Andrew (1998). The Future of Global Financial Regulation. Occasional papers No. 57. Washington, DC: Group of Thirty. Ledyard, John O. (1995). Public Goods: A Survey of Experimental Research. In The Handbook of Experimental Economics (John H. Kagel and Alvin E. Roth, eds.), pp. 111–194. Princeton, NJ: Princeton University Press. Libecap, Gary D. (1995). The Conditions for Successful Collective Action. In Local Commons and Global Interdependence: Heterogeneity and Cooperation in Two Domains (Robert O. Keohane and Elinor Ostrom, eds.), pp. 161–190. London: Sage. Lichbach, Mark I. (1996). The Cooperator’s Dilemma. Ann Arbor: University of Michigan Press. Liebowitz, S. J. and Stephen Margolis (1994). Network Externality: An Uncommon Tragedy. Journal of Economic Perspectives 8: 133–150. Lindahl, Erik (1919). Die Gerechtigkeit der Besteuerung. Positive Lösung. Lund: Gleerup. Lubell, Mark and John T. Scholz (2001). Cooperation, Reciprocity, and the Collective-Action Heuristic. American Journal of Political Science 45: 160–178.
R EF E R EN C E S
245
Lütz, Susanne (1997). Die Rückkehr des Nationalstaats? Kapitalmarktregulierung im Zeichen der Internationalisierung von Finanzmärkten. Politische Vierteljahresschrift 38: 475–497. ——— (1999). Zwischen “Regime” und “kooperativem Staat.” Bankenregulierung im internationalen Mehr-Ebene-System. Zeitschrift für Internationale Beziehungen 6: 9–39. Mäler, Karl-Goran and Aart de Zeeuw (1998). The Acid Rain Differential Game. Environmental and Resource Economics 12: 167–184. Malkin, Jesse and Aaron Wildavsky (1991). Why the Traditional Distinction between Public and Private Goods Should Be Abandoned. Journal of Theoretical Politics 3: 335–378. Marks, Gary, Liesbet Hooghe, and Kermit Blank (1996). European Integration from the 1980s: State-centric v. Multi-level Governance. Journal of Common Market Studies 34: 341–378. Martin, Lisa L. (1995). Heterogeneity, Linkage and Commons Problems. In Local Commons and Global Interdependence: Heterogeneity and Cooperation in Two Domains (Robert O. Keohane and Elinor Ostrom, eds.), pp. 71–91. London: Sage. Marwell, Gerald and Ruth E. Ames (1979). Experiments on the Provision of Public Goods I: Resources, Interest, Group Size, and the Free Rider Problem. American Journal of Sociology 84: 1335–1360. Marwell, Gerald and Pamela Oliver (1993). The Critical Mass in Collective Action: A Micro-Social Theory. New York: Cambridge University Press. Mazmanian Daniel A. and Michael Stanley-Jones (1995). Re-conceiving Lulus: Changing the Nature and Scope of Locally Unwanted Land Uses. In Lösung von Umweltkonflikten durch Verhandlung. Beispiele aus dem In- und Ausland (Peter Knoepfel, ed.), pp. 27–54. Basel: Helbing & Lichtenhahn. Mendez, Ruben P. (1999). Peace as a Global Public Good. In Global Public Goods: International Cooperation in the 21st Century (Inge Kaul, Isabelle Grunberg, and Marc A. Stern, eds.), pp. 382–416. New York-Oxford: Oxford University Press. Michaelowa, Axel (1998). Übertragung des Demokratiemodells der Neuen Politischen Ökonomie auf die Klimapolitik. Zeitschrift für Umweltpolitik und Umweltrecht 21: 463–491. Mitchell Ronald B. (1995). Heterogeneities at Two Levels: States, Non-state Sectors and Intentional Oil Pollution. In Local Commons and Global Interdependence: Heterogeneity and Cooperation in Two Domains (Robert O. Keohane and Elinor Ostrom, eds.), pp. 223–251. London: Sage. Moremen, Philip (1995). Negotiating the Biological Diversity Convention. Program on Negotiation at Harvard Law School, Working Paper Series 95–3. Cambridge, MA. Morrow, James D. (1994). Game Theory for Political Scientists. Princeton, NJ: Princeton University Press. Musgrave, Richard A. and Peggy B. Musgrave (1973). Public Finance in Theory and Practice. New York: McGraw-Hill.
246
R EF E R EN C E S
Myles, Gareth D. (2000). Wasteful Government, Tax Evasion, and the Provision of Public Goods. European Journal of Political Economy 16: 51–74. Nash, John (1950). The Bargaining Problem. Econometrica 18: 155–162. ——— (1951). Non-Cooperative Games. Annals of Mathematics 54: 286–295. Nida-Rümelin, Julian (1991). Practical Reason or Metapreferences? An Undogmatic Defense of Kantian Morality. Theory and Decision 30: 133–162. Oates, Wallace (1972). Fiscal Federalism. New York: Harcourt Brace Jovanovich. Oates, Wallace E. and Robert M. Schwab (1988). Economic Competition among Jurisdictions. Efficiency Enhancing or Distortion Inducing? Journal of Public Economics 35: 333–354. OECD (1996). Saving Biological Diversity: Economic Incentives. Paris: OECD. Olson, Mancur (1965). The Logic of Collective Action: Public Goods and the Theory of Groups. Cambridge, MA: Harvard University Press. ——— (1969). The Principle of “Fiscal Equivalence”: The Division of Responsibilities among Different Levels of Government. American Economic Review 59: 479–487. Orbell, John M. and Robyn M. Dawes (1991). A “Cognitive Miser” Theory of Cooperators’ Advantage. American Political Science Review 85: 515–528. Orbell, John M., Alphons van de Kragt, and Robyn M. Dawes (1988). Explaining Discussion-Induced Cooperation. Journal of Personality and Social Psychology 54: 811–819. Ostrom, Elinor (1990). Governing the Commons. The Evolution of Institutions for Collective Action. New York: Cambridge University Press. ——— (1998). A Behavioral Approach to the Rational Choice Theory of Collective Action. American Political Science Review 92: 1–22. ——— (2002). Property Rights Regimes and Common Goods: A Complex Link. In Common Goods: Reinventing European and International Governance (Adrienne Héritier, ed.), pp. 29–58. Lanham, MD: Rowman & Littlefield. Ostrom, Elinor, Roy Gardner, and James Walker (1992). Covenants with and without a Sword: Self-Governance Is Possible. American Political Science Review 86: 404–417. ——— (1994). Rules, Games, and Common Pool Resources. Ann Arbor: University of Michigan Press. Perman, Roger, Yue Ma, and James McGilvray (1998). Natural Resources and Environmental Economics. London-New York: Longman. Pigou, Arthur C. (1920). The Economics of Welfare. London: MacMillan. Plümper, Thomas (1996). Der Wandel weltwirtschaftlicher Institutionen. Regimedynamik durch ökonomische Prozesse. Berlin: Duncker & Humblot.
R EF E R EN C E S
247
Posner, Richard (1977). Economic Analysis of Law. Boston, MA: Little, Brown & Co. Putnam, Robert D. (1988). Diplomacy and Domestic Politics: The Logic of Two-level Games. International Organization 42: 427–460. Radaelli, Claudio M. (1998). Creating the International TaxOorder: Transfer Pricing and the Search for Coordination in International Tax Policy. EUI Working Paper RSC No. 98/28. Raiffa, Howard (1985). Creative Compensation: Maybe “In My Backyard.” Negotiation Journal 1: 197–203. Randow, Philipp v. (1996). Rating und Regulierung. In Handbuch Rating (Hans E. Büschgen and Oliver Everling, eds.), pp. 543–576. Wiesbaden: Gabler. Rao, J. Mohan (1999). Equity in a Global Public Goods Framework. In Global Public Goods: International Cooperation in the 21st Century (Inge Kaul, Isabelle Grunberg, and Marc A. Stern, eds.), pp. 68–87. New YorkOxford: Oxford University Press. Rapoport, Amnon and Ramzi Suleiman (1992). Equilibrium Solutions for Resource Dilemmas. Group Decision-making and Negotiations 1: 269–264. Rapoport, Anatol (1988). Experiments with N-Person Social Traps I: Prisoner’s Dilemma, Weak Prisoner’s Dilemma, Volunteer’s Dilemma and Largest Number. Journal of Conflict Resolution 32: 457–472. Rapoport, Anatol and Albert Chammah (1965). Prisoner’s Dilemmas: A Study in Conflict and Cooperation. Ann Arbor: University of Michigan Press. Rapoport, Anatol and Melvin J. Guyer (1966). A Taxonomy of 2 3 2 Games: General Systems. Yearbook of the Society for General Systems Research XI: 203–214. Rapoport, Anatol, Melvin J. Guyer, and David G. Gordon (1976). The 2 3 2 Game. Ann Arbor: University of Michigan Press. Raustiala, Kal (1997). Domestic Institutions and International Regulatory Cooperation. Comparative Responses to the Convention on Biological Diversity. World Politics 49: 482–509. Remmert, Hermann (1992). Ökologie. Ein Lehrbuch. Berlin et al.: Springer. Riker, William H. (1964). Federalism. Origins, Operation, Significance. Boston-Toronto: Little Brown. Rogers, Peter (1997). International River Basins: Pervasive Unidirectional Externalities. In The Economics of Transnational Commons (Partha Dasgupta, Karl-Göran Mäler, and Allessandro Vercelli, eds.), pp. 35–76. Oxford: Clarendon Press. Rose-Ackerman, Susan (1981). Does Federalism Matter? Political Choice in a Federal Republic. Journal of Political Economy 89: 152–165. Rosenau, James N. and Ernst-Otto Czempiel (eds.) (1992). Governance without Government: Order and Change in World Politics. Cambridge: Cambridge University Press.
248
R EF E R EN C E S
Rotberg, Eugene H. (1992). Risk Taking in the Financial Services Industry. In Risk Management in Financial Services (OECD, ed.), pp. 9–58. Paris: OECD. Roth, Alvin E. (1995). Introduction. In The Handbook of Experimental Economics (John H. Kagel and Alvin E. Roth, eds.), pp. 3–110. Princeton, NJ: Princeton University Press. Runge, Ford C. (1984). Institutions and the Free Rider: The Assurance Problem in Collective Action. Journal of Politics 46: 154–181. Sabatier, Paul A. (1998). The Advocacy Coalition Framework: Revisions and Relevance for Europe. Journal of European Public Policy 5: 98–130. Samuelson, Paul A. (1954). The Pure Theory of Public Expenditure. Review of Economics and Statistics 36: 387–389. ——— (1955). Diagrammatic Exposition of a Theory of Public Expenditure. Review of Economics and Statistics 37: 350–356. Sandell, Rickard and Stern Charlotta (1998). Group Size and the Logic of Collective Action. A Network Analysis of a Swedish Temperance Movement 1896–1937. Rationality and Society 10: 327–345. Sandholtz, Wayne (1993). Choosing Union: Monetary Politics and Maastricht. International Organization 47: 1–39. Sandler, Todd (1997). Global Challenges: An Approach to Environmental, Political, and Economic Problems. New York: Cambridge University Press. ——— (1998). Global and Regional Public Goods: A Prognosis for Collective Action. Fiscal Studies 19: 221–247. ——— (2004). Global Collective Action. Cambridge: Cambridge University Press. ——— (2008). Treaties: Strategic Considerations. Illinois Law Review: 155–180. Sandler, Todd and Keith Sargent (1995). Management of Transnational Commons: Coordination, Publicness, and Treaty Formation. Land Economics 71: 145–162. Scharpf, Fritz W. (1985). Die Politikverflechtungsfalle: Europäische Integration und deutscher Föderalismus im Vergleich. Politische Vierteljahresschrift 26: 323–356. ——— (1996). Politische Optionen im vollendeten Binnenmarkt. In Europäische Integration (Markus Jachtenfuchs and Beate Kohler-Koch, eds.), pp. 109–140. Opladen: Leske 1 Budrich. ——— (1997a). Introduction: The Problem-solving Capacity of Multi-level Governance. Journal of European Public Policy 4: 520–538. ——— (1997b). Games Real Actors Play: Actor-Centered Institutionalism in Policy-Research. Boulder: Westview Press. Scharpf, Fritz W., Bernd Reissert, and Fritz Schnabel (1976). Politikverflechtung. Theorie und Empirie des kooperativen Föderalismus in der Bundesrepublik. Kronberg/Ts. Schelling, Thomas (1960). The Strategy of Conflict. Cambridge, MA: Harvard University Press.
R EF E R EN C E S
249
——— (1978). Micromotives and Macrobehavior. New York-London: Norton. Schlager, Edella, William Blomquist, and Shui Yan Tang (1994). Mobile Flows, Storage, and Self-Organized Institutions for Governing CommonPool Resources. Land Economics 70: 294–317. Schroeder-Wildberg, Esther (2002). The Negotiation Process of the International Water Convention: Balancing Upstream-Downstream Interests. In Co-operation on Transboundary Rivers (Ismail Al Baz, Volkmar Hartje, and Waltina Scheumann, eds.), pp. 35–51. Baden-Baden: Nomos. Serageldin, Ismail (1999). Cultural Heritage as Public Good: Economic Analysis Applied to Historic Cities. In Global Public Goods: International Cooperation in the 21st Century (Inge Kaul, Isabelle Grunberg, and Marc A. Stern, eds.), pp. 240–263. New York-Oxford: Oxford University Press. Sinclair, Timothy J. (2002). Private Makers of Public Policy: Bond Rating Agencies and the New Global Finance. In Common Goods: Reinventing European and International Governance (Adrienne Héritier, ed.), pp. 279–292. Lanham, MD: Rowman & Littlefield. Sinn, Hans-Werner (1990). Tax Harmonization and Tax Competition in Europe. European Economic Review 34: 489–504. ——— (1997). The Selection Principle and Market Failure in Systems Competition. Journal of Public Economics 66: 247–274. Snidal, Duncan (1995). The Politics of Scope: Endogenous Actors, Heterogeneity and Institutions. In Local Commons and Global Interdependence: Heterogeneity and Cooperation in Two Domains (Robert O. Keohane and Elinor Ostrom, eds.), pp. 47–70. London: Sage. Spar, Debora L. (1999). The Public Face of Cyberspace. In Global Public Goods: International Cooperation in the 21st Century (Inge Kaul, Isabelle Grunberg, and Marc A. Stern, eds.), pp. 344–363. New York-Oxford: Oxford University Press. Spencer, Peter D. (2000). The Structure and Regulation of Financial Markets. Oxford: Oxford University Press. Stiglitz, Joseph E. (1999). Knowledge as a Global Public Good. In Global Public Goods: International Cooperation in the 21st Century (Inge Kaul, Isabelle Grunberg, and Marc A. Stern, eds.), pp. 308–325. New YorkOxford: Oxford University Press. Straubhaar, Thomas (1995). Ein Europa des funktionalen Föderalismus. Mehr als ein Denkmodel? ORDO – Jahrbuch für die Ordnung von Wirtschaft und Gesellschaft 46: 185–202. Strulik, Torsten (2002). Rating Agencies and Systemic Risk: Paradoxes of Governance. In Common Goods: Reinventing European and International Governance (Adrienne Héritier, ed.), pp. 317–335. Lanham, MD: Rowman & Littlefield. Sy, J. Habib (1999). Global Communications for a More Equitable World. In Global Public Goods: International Cooperation in the 21st Century (Inge
250
R EF E R EN C E S
Kaul, Isabelle Grunberg, and Marc A. Stern, eds.), pp. 326–343. New York-Oxford: Oxford University Press. Szego, Giorgio (1997). A Critique of the Basel Regulations, or How to Enhance (Im) Moral Hazards. In Risk Management and Regulation in Banking (Dan Galai, David Ruthenberg, Marshall Sarnat, and Ben Z. Schreiber, eds.), pp. 147–158. Boston: Kluwer Academic Publishers. Taylor, Christopher (1995). EMU 2000? Prospects for European Monetary Union. London: The Royal Institute of International Relations. Taylor, Michael (1976). Anarchy and Cooperation. New York: Wiley. ——— (1987). The Possibility of Cooperation. Cambridge-New York-Paris: Cambridge University Press. Tiebout, Charles M. (1956). A Pure Theory of Local Expenditures. Journal of Political Economy 64: 416–424. Tietenberg Tom (ed.) (1997). The Economics of Global Warming. Cheltenham: Edward Elgar. Ullmann-Margalit, Edna (1977). The Emergence of Norms. Oxford: Clarendon Press. Verweij, Marco (2000). Transboundary Environmental Problems and Cultural Theory: The Protection of the Rhine and Great Lakes. New York: Palgrave. Vicary, Simon and Todd Sandler (2002). Weakest-link Public Goods: Giving In-kind or Transferring Money. European Economic Review 46: 1501–1520. Vitols, Sigurt (1998). Are German Banks Different? Small Business Economics 10: 79–91. Vogel, David (1995). Trading Up: Consumer and Environmental Regulation in the Global Economy. Cambridge, MA: Harvard University Press. ——— (1997). Trading up and Governing across: Transnational Governance and Environmental Protection. Journal of European Public Policy 4: 556–571. Wassenberg, Arthur F. P. (1982). Neo-Corporatism and the Quest for Control: The Cuckoo Game. In Patterns of Corporatist Policy-Making (Gerhard Lehmbruch and Philippe Schmitter, eds.), pp. 83–108. Beverly Hills, CA: Sage. Weesie, Jeroen and Axel Franzen (1998). Cost Sharing in a Volunteer’s Dilemma. Journal of Conflict Resolution 2: 600–618. Wicksell, Knut (1896). Ein neues Prinzip der gerechten Besteuerung. Jena: Fischer. Woodward, F. I. (1994). How Many Species Are Required for a Functional Ecosystem? In Biodiversity and Ecosystem Function (Ernst-Detlev Schulze and Harold A. Mooney, eds.), pp. 271–291. New York: Springer. Wyplosz, Charles (1999). International Financial Instability. In Global Public Goods: International Cooperation in the 21st Century (Inge Kaul, Isabelle Grunberg, and Marc A. Stern, eds.), pp. 152–189. New York-Oxford: Oxford University Press.
R EF E R EN C E S
251
Yongding, Yu (2000). China: The Case for Capital Controls. In Global Finance. New Thinking on Regulating Speculative Capital Markets (Walden Bello, Nicola Bullard, and Kamal Malhotta, eds.), pp. 177–187. London and New York: Zed Books. Young, Michael D. (1992). Sustainable Investment and Resource Use. Paris: UNESCO. Young, Oran R. (1989). International Cooperation: Building Regimes for Natural Resources and the Environment. Ithaca, NY: Cornell University Press. Young, Peyton (1998). Individual Strategy and Social Structure. An Evolutionary Theory of Institutions. Princeton, NJ: Princeton University Press. Zacher, Marc W. (1999). Global Epidemiological Surveillance: International Cooperation to Monitor Infectious Diseases. In Global Public Goods: International Cooperation in the 21st Century (Inge Kaul, Isabelle Grunberg, and Marc A. Stern, eds.), pp. 266–283. New York-Oxford: Oxford University Press. Zintl, Reinhard (1994). Kooperation kollektiver Akteure: zum Informationsgehalt angewandter Spieltheorie. In Praktische Rationalität. Grundlagenprobleme und ethische Anwendungen des Rational choiceParadigmas (Julian Nida-Rümelin, ed.), pp. 239–258. Berlin: Walter de Gruyter. ——— (1995). Der Nutzen unvollständiger Erklärungen: Überlegungen zur sozialwissenschaftlichen Anwendung der Spieltheorie. MPIFG Discussion Paper 95/2. Köln. ——— (2001). Rational Choice as a Tool in Political Science. Associations 5: 35–50. Zürn, Michael (1992). Interessen und Institutionen in der internationalen Politik. Grundlegung und Anwendungen des situationsstrukturellen Ansatzes. Opladen: Leske 1 Budrich. ——— (1996). Über den Staat und die Demokratie im europäischen Mehrebenensystem. Politische Vierteljahresschrift 37: 27–55. ——— (1997). Positives Regieren jenseits des Nationalstaats. Zur Implementation internationaler Umweltregime. Zeitschrift für Internationale Beziehungen 4: 41–68.
This page intentionally left blank
Index
2 × 2 games, see matrix games aggregation, 169–79, 209, 212, 214, 218, 224–5, 232–3 functions, 69–70, 79, 181–2, 188 mechanisms, 174–5, 177, 209, 210, 212–25, 232–3 technologies, 6, 32, 38, 43, 69–82, 137–43, 209, 228–9 assurance game, 32, 52, 53, 58, 70, 77, 79, 82, 93, 96, 131, 134–5, 137, 143–8, 154, 157, 197, 229, 230 attributes of actors, 2, 36, 38–40, 83–109 of goods, 1–2, 36–9, 43–81, 83 of groups, 38, 40, 83–109 of social and institutional context, 2, 40, 111, 137 bank failures, 63–4, 66, 119, 197 run, 63 banking, 63–6, 165 crises, 1, 47, 63–5, 87 sector, 63, 94 bargaining, 79, 81, 84, 151, 170–1, 190, 210–21, 225, 232 definition, 214 Basle Accords, 1, 47–8, 67, 119 Basel Committee, 45–8, 119 battle of the sexes, 146, 148, 152, 154, 157, 230
best-shot technology, 43, 69–71, 79–80, 82, 143–4, 209, 229 biodiversity, 1, 6–7, 27, 43, 69, 72, 75–80, 138, 165, 177, 179, 183, 231, see also Convention on Biological Diversity Capital Adequacy Directive, 48, 67, 119 capital allocation, 86–8 capital income taxation, see capital income taxes capital income taxes, 6, 7, 84–8, 91–3, 98, 139, 189–91, 193 coordination, 6, 84, 85–102, 190–4 capital market, 5, 43–53, 62, 67, 85–99, 112–22, 167–8, 189–94 chicken game, 70, 81–2, 137, 143–8, 154, 157, 209, 229–30 classification, 12, 15, 17, 28, 55, 139 of matrix games, 149–53 of public goods, 15–21 see also taxonomy, typology climate, 1, 39, 72–5, 123, 185–7 club goods, 12, 15, 17, 19, 24, 28, 80, 143, 227 collective definition, 175
254
INDEX
collective action problem, 2–6, 11, 28–36, 43–4, 54, 82, 93, 96, 125, 137–59, 168, 228–9 definition, 29 collective consumption good, 12–14, 16 Commission (EU), 85–90 common goods definition, 28 experimental research, 21–3 field research, 23–5 and prisoners dilemma, 29–33 scope of, 35, 165–9, 171 common goods provision attributes of, 38–41 characteristics of, 37–42 social situation of, 36–7 common pool resources, 3, 14, 15, 19, 55, 58–9, 61, 84–5, 92, 143, 180, 229 competitive (dis) advantage, 123, 127–9, 131, 133–4 congestibles, 28 constant-sum game, 92, 151, 156 constituency, 170–4, 182, 190, 194, 206, 214, 232 contagion, 63–7 Convention on Biological Diversity, 1, 75 coordination game, 32, 79, 85, 92–6, 141, 143–8, 153, 157–8, 194, 229–30 pure, 143, 153, 157, 230 coordination problem, 29, 81, 135, 137, 157, 166, 230 cost-benefit configuration, 6, 43–58, 69, 74, 137, 140–2, 229 Council of Finance Ministers, 87 of Ministers, 89–90 CPR, see common pool resources credit rating, 43–58, 62, 111–22, 139–44 credit risk, 62–7, 118–19, 121 crisis(es), see banking crises
currency common, 195–208 strong, 196–205 weak, 196–207 defection problem, 98, 154, 156, 230, see also prisoners dilemma definitions bargaining, 214 collective, 175 collective action problem, 29 common good, 28 dictatorship, 214 direct democracy, 214 global common good, 164–5 multilevel system, 173–4 public good, 15–17 representative, 174 representative democracy, 214 subgroup, 175 transnational common good, 164–5 demand-side properties, 38, 43, 55–69, 140–4 dictatorship, 210–17, 224–5, 232 definition, 214 dilemma asymmetric, 98–9, 102, 137, 143–7, 156, 229 volunteers’, 32, 148 see also prisoners dilemma direct democracy, 210–14, 216–18, 222–5, 232 definition, 214 disagreement problem, 154, 157, 158, 230 discoordination game, 143–7, 157–8, 194 disintermediation, 45, 65 distortion through aggregation, 212–25 through cooperation, 194–207 effects, 7, 179, 189–225 through representation, 190–4, 218 distribution problem, 155–7, 230
INDEX
dominant strategy, 50–2, 57, 60, 62, 69, 73, 77–8, 98, 102, 106, 114, 117, 121, 131, 139, 141, 182, 187, 196, 203, 222 downstream countries, 103–9, 183 ecosystem, 75–7, 183 emission reduction, 73–4, 185–6 regulation, 135 standards, 123–4 EMU, see European Monetary Union environmental pollution, 6, 14, 21, 43, 55, 58, 60–2, 67, 85, 102–6, 143 regulation, 123, 132, 176, 188 standards, 41, 111, 122, 125, 127, 128, 131, 134–6 equilibrium, -ia, 5, 13, 23, 30, 32, 49, 52, 77, 81–2, 96–100, 102, 105–6, 114, 117–18, 120, 124, 127, 131–5, 141, 146–8, 194, 197, 203, 211–12, 222, 230 EU Council, see Council of Ministers Euro, see European Monetary Union Euro zone, 195, 203–4 European Commission, see Commission European Council, 85–90 European Court of Justice, 97, 108 European Monetary Union, 19, 89, 176–7, 195–208, 231 excludability, 11, 15, 18, 43–4, 55, 62 externalities bidirectional, 60, 103, 180 transboundary, 169 unidirectional, 85, 103 federalism, 169–73, 225 economic theories of, 170–2 legislative, 170
255
financial markets, 5–6, 27, 38, 46–7, 55, 62–9, 168, 195 free trade, 131–5 games assurance, 32, 52, 53, 58, 70, 77, 79, 82, 93, 96, 131, 134–5, 137, 143–8, 154, 157, 197, 229, 230 asymmetric dilemma, 98–9, 102, 137, 143–7, 156, 229 battle of the sexes, 146, 148, 152, 154, 157, 230 chicken, 70, 81–2, 137, 143–8, 154, 157, 209, 229–30 constant-sum, 92, 151, 156 degenerate coordination, 141, 142, 147, 153–4, 157 discoordination, 143–7, 157–8, 194 harmony, 49, 51, 53, 60, 62, 73, 78, 81–2, 105, 114–15, 118, 120–1, 124, 134, 139, 141, 143–7, 154–9, 187, 196–7, 208–12, 222, 229 harmony, pure, 105, 154–5 harmony, weak, 105, 154, 158 matching pennies, 153–4, 158 prisoners dilemma, 1–2, 22, 29–37, 44, 57–8, 62, 69, 73–4, 76, 82, 85, 96, 98–9, 117, 124, 132–5, 139–41, 143–8, 156, 159, 180–1, 186–7, 190–1, 194, 227, 229 pure conflict, 116, 118, 137, 139, 143–7, 151, 153, 156, 158, 230 pure coordination, 81, 143, 153–4, 157, 230 rambo, 105–6, 109, 121, 127, 131–9, 143–7, 149, 154–9, 182–3, 187, 203, 229–30 volunteers’ dilemma, 32, 148 zero-sum, 147–9, 151, 154, 156 genetic variability, 75 German unification, 206–8
256
INDEX
global common goods, 1, 2, 4, 25, 35, 77, 163–5, 227 definition, 164–5 global warming, 6, 7, 27, 72–6, 79, 103, 138, 165, 176–9, 184–9, 231 greenhouse effect, 72–3, 184–5 emissions, 73 gases, 72–3, 184–5 Group of Thirty, 63–6 harmonization advantage, 127–9, 131–2 harmony (game), 49, 51, 53, 60, 62, 73, 78, 81–2, 105, 114–15, 118, 120–1, 124, 134, 139, 141, 143–7, 154–9, 187, 196–7, 208–12, 222, 229 heterogeneity, 83–5, 174–8 of actors, 7, 27, 40, 62, 83, 85, 121, 124, 144–7, 163–4, 175–6, 184, 230 capabilities, 83–5, 99, 103, 144, 172 of constituencies, 175, 184–9, 204 of the groups, 175, 179–84 preferences, 84, 97, 99, 122, 125, 134, 144, 170–2, 194, 233 heterogeneous, see heterogeneity homogeneity of constituencies, 184–9 of the groups, 40, 175, 179–84 homogenous, see homogeneity information good, 15, 19, 27, 33, 43–4, 46, 48–50 instability problem, 29, 145, 149, 154–9, 230 international river basin, 6, 59, 85, 99, 102–8, 109, 181–4 International Water Convention (IWC), 104 jurisdiction, 35, 168–79, 209–33
locally unwanted facilities, see locally unwanted land uses locally unwanted land uses, 6–7, 43, 69, 80–2, 176–7, 209–25, 231 LULUs, see locally unwanted land uses Maastricht Treaty, 201–7 market risk, 48, 65, 119 market segmentation, 126–34 marketable public goods, 15, 19, 33, 49, 53, 55, 62, 141, 143, 227, 229 matching pennies, 153–4, 158 matrix games, 3–5, 30, 70, 73, 82, 92, 100, 139, 147–54, 158 monetary policy, 195–7, 205–6 Monetary Union, see European Monetary Union multilevel system, 172, 224 definition, 173–4 theories of, 169–73 Nash equilibrium,-ia, 31–2, 49, 52, 77, 81, 96–9, 105, 114, 117–20, 141, 145–6, 148–9, 150–2, 153–8, 194, see also equilibrium network effect, 204 network goods, 18–19, 28, 33, 195 nonexcludability, 1, 15–17, 19, 36, 38, 55–8, 124, 140–3, 228 nonrivalry of consumption, 1, 6, 12, 15–21, 36, 38, 49, 51, 53, 55, 58, 124, 140, 142–3, 147, 186, 228 open access resources, see common pool resources Pareto-optimal outcomes, 114, 146, 150–8, 229–30 Pareto-optimality, 29, 149–51, 157 Politikverflechtung, 170, 172
INDEX
pollution environmental, 6, 14, 21, 43, 55, 58, 60–2, 67, 85, 102, 103, 106, 143 of a lake, 59–62, 103, 105, 176, 180–1, 231 phases of, 60–1 of a river, 85, 99, 102–9, 139, 144, 177, 181–4, 231 price stability, 195–6, 199, 206 prisoners dilemma, 1–2, 22, 29–37, 44, 57–8, 62, 69, 73–4, 76, 82, 85, 96, 98–9, 117, 124, 132–5, 139–41, 143–8, 156, 159, 180–1, 186–7, 190–1, 194, 227, 229 private goods, 15–16, 20, 53, 55 problematic situation, see also collective action problem product standards, 127–35 production standards, 123–35 properties, see attributes public goods definition, 15–17 impure, 15–16, 19 pure, 15, 17, 19, 24, 28–9, 38, 53, 55, 58, 62, 69, 143, 228 theory, 12–21, 26–9, 33, 35, 227 pure conflict, 116, 118, 137, 139, 143–7, 151, 153, 156, 158, 230 race to the bottom, 91, 96, 100, 111, 123–4, 127, 131–5 race to the top, 123–5, 127, 131, 134–5 rambo game, 105–6, 109, 121, 127, 131–9, 143–7, 149, 154–9, 182–3, 187, 203, 229–30 rating agencies, 44–6, 50–7, 112, 115–18, 121–2 ratings, 4–6, 41, 43–62, 111–22, 139–44 regulation environmental, 123, 132, 176, 188
257
regulatory competition, 7, 41, 111, 122–36, 138, 144 representative democracy, 210, 212, 214, 221–5, 232–3 definition, 214 representatives, 7, 150, 159, 172–95, 209–20, 224, 231–3 definition, 175 riparian countries, 99, 102–4, 107–8, 181 rivalry of consumption, 6, 11, 14–21, 43, 55, 58, 62, 72, 143 single level game, 175, 177, 180–4, 187, 190, 198, 210, 219–20, 231 provision, 2, 163, 174, 228, 230 system, 6, 172, 175, 177, 180–1, 189, 194, 215–18, 221, 224–5, 227, 231–3 siting conflict, 72, 80–2, 219 species, 72, 75–9, 122, 171, 183–4 standards environmental, 41, 111, 122, 125, 127–8, 131, 134, 136 product, 127–35 production, 123–35 strategic constellations, 2–7, 11, 30–41, 43–4, 55, 58–60, 62, 69–71, 75, 82–5, 92–9, 111, 131–8, 139–47, 163, 168, 174–84, 187, 189–91, 210, 212, 225, 227–31, 233 strategic form games, 148–54 strong currency, 196–205 subgroup, 25, 174–208, 209–12, 218, 231–2 definition, 175 summation technology, 43, 69–71, 140, 143–4, 184, 186, 191, 228 supply-side properties, 38, 43, 69–81, 142–3, 228 systemic risk, 1, 6, 43, 55, 62–9, 119, 142–3, 165, 168
258
INDEX
tax base, 91–102 competition, 86, 91–102, 144, 190 coordination, 4, 6, 84, 86, 92–102, 138–9, 144, 176–7, 190 harmonization, 85–6, 91–2 haven, 91, 96 revenue, 84–7, 89, 91–102, 190–2, 194 taxation capital income, 7, 85–93, 139, 189–91 double, 86–8 savings, 85–93 taxonomy, 18–19, 148–9 of common goods, 18–19 of matrix games, 147–9 see also classification, typology trade barriers, 128, 129–31, 132, 133–4, 135 regime, 7, 41, 111, 122, 128, 134–5, 138 regime, free, 131–5
tragedy of the commons, 12, 14–15, 18, 24, 59, 62 transboundary common goods, 5, 122, 163, 165, 168 transnational common goods, 2–7, 25, 28, 163–5, 169, 174, 227–8, 233 definition, 164–5 field research, 25–8 two-level games, 170, 172, 175, 217, 220 typology of collective action problems, 139, 147–59 see also classification, taxonomy upstream countries, 85, 103, 105–7, 182 weak-currency, 196–207 weakest link technology, 43, 69–79, 85, 93, 96, 143–4 withholding taxes, 86–91, 93–4, 96 zero-sum game, 147–9, 151, 154, 156