Ruling the World
Ruling the World POWER POLITICS AND THE RISE OF S U P R A N AT I O N A L I N S T I T U T I O N S
LLO...
82 downloads
1150 Views
3MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
Ruling the World
Ruling the World POWER POLITICS AND THE RISE OF S U P R A N AT I O N A L I N S T I T U T I O N S
LLOYD GRUBER
P R I N C E T O N P R I N C E T O N ,
U N I V E R S I T Y N E W
J E R S E Y
P R E S S
Copyright © 2000 by Princeton University Press Published by Princeton University Press, 41 William Street, Princeton, New Jersey 08540 In the United Kingdom: Princeton University Press, Chichester, West Sussex All rights reserved
Library of Congress Cataloging-in-Publication Data Gruber, Lloyd, 1964– Ruling the world : power politics and the rise of supranational institutions / Lloyd Gruber. p. cm. Includes bibliographical references and index. ISBN 0-691-01040-4 (cl. : alk. paper) — ISBN 0-691-01041-2 (pb : alk. paper) 1. International relations. 2. International cooperation. 3. Supranationalism. 4. Balance of power. I. Title. JZ1308 .G78 2000 327.197 21—dc21 99-04063
Sources for figures: 6-1. IMF, Direction of Trade Statistics Yearbook, 1989; 6-2. IMF, Direction of Trade Statistics Yearbook, 1989 and World Bank, World Development Indicators, CD-ROM; 6-3. OECD, Economic Outlook—Historical Statistics, various years; 7-1. Adapted from Hart 1990, 71, table 12; 7-2. Adapted from Conroy and Glasmeier 1992/93; 7-3. Banco de México; 8-1. IMF, International Financial Statistics; 8-2. IMF, Direction of Trade Statistics Yearbook, 1984; 8-3. IMF, International Financial Statistics; OECD, Historical Statistics; 8-4. OECD, Economic Outlook—Historical Statistics, various years; 8-5. OECD, Economic Outlook—Historical Statistics, 1960 – 1990 (1992); 8-6. OECD, Economic Outlook—Historical Statistics, 1960–1990 (1992); 9-1. OECD, Economic Outlook—Historical Statistics, various years; 9-2. IMF, International Financial Statistics; 9-3. IMF, International Financial Statistics, CD-ROM; 9-4. OECD, Economic Outlook—Historical Statistics, various years; 10-1. Adapted from Riker 1996, 19, fig. 2-1. This book has been composed in Postscript Sabon The paper used in this publication meets the minimum requirements of ANSI/NISO Z39.48-1992 (R1997) (Permanence of Paper) www.pup.princeton.edu Printed in the United States of America 1
2
3
4
5
6
7
8
9
10
1
2
3
4
5
6
7
8
9 10 (pbk)
FOR DELIA
C O N T E N T S
LIST OF FIGURES
ix
LIST OF TABLES
xi
PREFACE AND ACKNOWLEDGMENTS
1. Introduction: From Anarchy to Organization PART I: WHY DO NATIONS COOPERATE?
xiii 3 13
2. Institutions, Collective Action, and the Prospect of Mutual Gain
15
3. Winners and Losers: The Case for Theoretical Reorientation
33
PART II: HOW DO NATIONS COOPERATE?
59
4. The Efficiency Rationale for Supranational Governance
61
5. Broadening the Debate: The “Power Politics” of Institutional Design
81
PART III: NORTH AMERICAN TRADE 6. Jump-Starting the Free Trade Bandwagon 7. NAFTA and Beyond: Is Free Trade Contagious? PART IV: EUROPEAN MONEY
93 95 122 169
8. Winners and Losers in the European Monetary System
171
9. Rigging the System: Why Did the EMS Take the Form It Did?
213
PART V: CONCLUSIONS
249
10. Rethinking International Cooperation
251
BIBLIOGRAPHY
279
INDEX
307
L I S T
O F
F I G U R E S
2-1.
The Positive-Sum Logic of Neoliberalism
18
2-2.
The Collective-Action Paradigm: A Utility Representation
30
2-3.
The Collective-Action Paradigm: Integrating Power into the Analysis
31
3-1.
Power as Coercion
35
3-2.
An Alternative Model of Power Bargaining
36
3-3.
Go-It-Alone Power
39
3-4.
Cooperation as a Game of Winners and Losers
44
3-5.
The Cooperation Game, Stage Two
49
3-6.
Go-It-Alone Power as a Bargaining Tactic
51
4-1.
The Prisoners’ Dilemma
64
4-2.
The “Battle of the Sexes” Game
77
6-1.
Geographic Distribution of U.S. and Canadian Trade, 1987
110
6-2.
Trade as a Share of U.S. and Canadian GDP, 1987
111
6-3.
Indicators of Canadian Economic Performance, 1984–1993
118
7-1.
U.S. Merchandise Imports from Canada and Mexico, 1979–1987
133
Ratios of Real Income per Capita in Mexico, the United States, and Western Europe, 1960–1988
149
Mexico’s Attractiveness as a Site for Foreign Investment, 1987–1996
150
8-1.
Monthly Exchange Rates in Germany, 1970–1980
176
8-2.
France and West Germany’s Major Trading Partners, 1978
185
8-3.
UK-ERM Interest Rate Differentials, 1979 –1993
188
8-4.
Consumer Price Inflation in Italy, 1977–1990
195
8-5.
Exports as a Percentage of GDP in Germany and Italy, 1960–1990
206
A Comparison of Growth Rates in Germany and the United Kingdom, 1978–1990
208
7-2. 7-3.
8-6. 9-1.
Inflation in France and Its Major European Trading Partners, 1974–1993 231
9-2.
The D-Mark/Franc Exchange Rate, 1974–1994
235
x
L I S T
O F
F I G U R E S
9-3.
Interest Rates in France and Germany, 1976–1994
238
9-4.
Annual Unemployment Rates in France and Other Major European Countries, 1976–1994
239
Riker’s Spatial Representation of Alternative Possible Constitutions
267
10-1.
L I S T
6-1.
O F
T A B L E S
U.S. Antidumping and Countervailing Duty Investigations Initiated against Canada, 1979–1986
102
7-1.
Foreign Direct Investment in Mexico, 1989
130
7-2.
Mexico’s Major Trading Partners, 1989
131
7-3.
U.S. Antidumping and Countervailing Duty Investigations Initiated against Mexico, 1979–1990
132
Areas of Intense Competition between Mexican and Canadian Manufactured Exports to the United States, 1987
134
Areas of Greatest U.S. Import Growth from Mexico, 1985–1987
137
Popular Vote in Italian Parliamentary Elections, 1953–1983
198
Election Results Immediately before and after Italy’s Entry into the EMS: A Closer Look
199
The Symmetry of Business Cycles in the European Community, 1965–1979
202
Cumulative Change in Bilateral Real Exchange Rates of Italy, France, and West Germany, 1979–1990
205
The Symmetry of Business Cycles in the European Community, 1965–1990
209
7-4.
7-5. 8-1. 8-2. 8-3. 8-4. 8-5.
P R E F AC E
A N D
A C K N O W L E D G M E N T S
The last few decades have witnessed an outburst of international cooperation and institution building unlike anything seen in previous years. While monetary and trade institutions have received the lion’s share of attention, the trend actually extends well beyond the economic arena, with new regional and multilateral initiatives proliferating in areas as diverse as foreign aid, health policy, human rights, the environment—even national security. But that’s really only the beginning, for many of these new initiatives do not look at all like the ones we are used to seeing. The difference is not just that they tend to be “wider,” often (like the North American Free Trade Agreement) spanning countries at vastly different levels of economic development. Nor is it simply that they bear more directly upon matters of domestic policy, their “deeper” provisions covering everything from antitrust law and pollution standards to basic questions of fiscal, monetary, and industrial policy. As striking as these differences may be, the most important departure from the regional and multilateral agreements of past years is that, with few exceptions, today’s agreements come equipped with their own built-in governance structures (cf. Williamson 1985). The result has been an extraordinary—perhaps unprecedented—transfer of policymaking prerogatives from individual nation-states to a host of new, higherlevel political institutions. If you think this is cause for celebration, you are not alone. Within the academic community (and not only among specialists in international relations), the notion that political institutions are mutually beneficial arrangements—that they would never come into existence, much less grow in size, stature, and assertiveness, were they not “Pareto-improving”—is today’s conventional wisdom. But is it really true? This book has a few basic purposes. The first is to suggest why I think this emphasis on interstate cooperation’s positive-sum consequences may be leading scholars of international relations down the wrong theoretical path. To be sure, history records few examples of multilateral institutions being foisted upon unwilling signatories. But the fact that membership in a cooperative arrangement is voluntary does not mean that that arrangement necessarily works to everyone’s advantage. Some cooperators, be they individuals or nations, may have a strong preference for the original, noncooperative status quo. The question, of course, is why these participants—the losers—would ever be willing to assume membership in such an arrangement. After all, it is not as if their partners—the arrangement’s beneficiaries—could prevent them from opting out. If the losers do not withdraw, one must conclude that it is because they choose not to withdraw. So why, then, would they ever make this (voluntary) choice? Stripped to its essence, my explanation is that the losers know the supra-
xiv
P R E F AC E
A N D
A C K N O W L E D G M E N T S
national arrangements they dislike are perfectly capable of functioning without them. While they themselves would like nothing more than to see these arrangements disappear, they realize that (1) there are other actors for whom supranational decision-making structures really are beneficial and (2) these other actors would continue to benefit from these structures even if they, the winners, were their only members. The fact that the winners would still realize welfare gains relative to the initial status quo affords them a kind of “go-it-alone power,” the importance of which—indeed the very existence of which—has been overlooked by previous scholars. This is not a trivial oversight. If I am right, a great deal of the supranational activity we have seen in recent years—the continuing frenzy of interstate deal making, treaty signing, and institution building—is a product of just this sort of go-it-alone dynamic. Not that go-it-alone power is always present, of course. When, exactly, is it most likely to operate? And what, precisely, are its consequences? How, in short, does go-it-alone power work? After introducing the concept, I go on to explore its inner logic, clarifying how it differs from that of other types of power and suggesting how these differences in turn point the way toward a broader, more balanced perspective on international cooperation. What emerges from this discussion is a new way of thinking not only about cooperation’s consequences—who wins and who, if anyone, loses—but also about the form it takes as well as the distinctive paths by which it occurs. Why have the relatively “thin” intergovernmental treaties and agreements of earlier years been giving way to “thicker” supranational arrangements? How can we account for these arrangements’ (differing) internal structures? Why does the process of institutional enlargement so often seem to take on a life and momentum of its own? My claim is that go-it-alone power has implications for how we go about addressing all of these questions, and others as well. If elucidating the theory of go-it-alone power is my first purpose, demonstrating how it works in practice is my second. To that end, I take a fresh look at the political origins and structures of the North American Free Trade Agreement (NAFTA) and European Monetary System (EMS). When it comes to global economic relations, the forging of these two interstate arrangements—the first an historic and, to most observers at the time, quite unexpected trade pact between Canada, Mexico, and the United States; the second an equally historic (and no less surprising) exchange rate regime involving all of the major economic powers in Western Europe—would have to rate among the postwar period’s most important international developments. The prevailing view is that they were also among the most successful: Nearly everyone seems to think that the launching of NAFTA (in the early 1990s) and the EMS (in the late 1970s) made it possible for their respective member governments to overcome collective-action problems in the pursuit of “gains from exchange.” My own view is that this line of thinking fundamentally misunderstands what NAFTA and the EMS were about. For if the criteria for “gains” are de-
P R E F AC E
A N D
A C K N O W L E D G M E N T S
xv
fined in Pareto terms—did the inaugurations of NAFTA and the EMS improve each of their members’ welfare relative to the state of affairs that existed beforehand?—matters are not so simple. While it is true that the creation of these arrangements did further the economic and political objectives of their respective “enacting coalitions,” other member states acceded to them, I will argue, only because they lacked a better alternative. Though eager not to be left behind, the latter states (or, more precisely, the individuals and parties presiding over their governments) nonetheless fully expected to incur utility losses as a result of their participation. If, in other words, it had been within their power to restore the noncooperative status quo— which, unfortunately for them, it was not—these NAFTA and EMS signatories would have gladly, indeed enthusiastically, done so. As will become clear as you move through the chapters that follow, this is not simply a study of international economic relations. While my substantive analysis focuses on the politics of North American trade and European money, my theoretical arguments are much broader, touching as they do upon issues of long-standing interest to students of domestic politics, organization theory, public policy—even political philosophy. In short, this is not a book about the “power politics” of international institutions. It is a book about power, cooperation, and institutional design in general. ?
?
?
This study is the culmination of a research project that I began . . . well, let’s just say several years ago and leave it at that. Along the way, I have benefited from the extraordinary generosity (not to mention patience) of a great many friends, colleagues, and students. Within this group, I owe a special debt to my three dissertation advisers, Geoffrey Garrett, Stephen Krasner, and Terry Moe. My acquaintance with Geoff dates back to 1986 when, as fate would have it, the Oxford college in which I had just taken up residency appointed him to be its new tutor in Politics. I took a tutorial from Geoff that autumn and, though I had originally come to England to pursue a graduate degree in sociology, immediately changed my field of study. Had it not been for that tutorial—and for Geoff’s strong encouragement later that summer—I doubt that I would have gone on to pursue a Ph.D. in political science upon returning to the United States, and I might not have pursued an academic career at all. My other two advisers, Steve and Terry, also deserve special thanks. Always prompt and incisive, their feedback was especially valuable in the early stages of this project when, in between (and sometimes during) sets of tennis, they helped me grapple with the ideas and arguments that have since made their way into this book. Among the other Stanfordians upon whom I inflicted early drafts of this work, Delia Boylan, John Ferejohn, Brian Gaines, Judith Goldstein, and Barry Weingast were especially giving of their time and seemingly boundless creative energy. As if my accumulated debts were not already large enough, I have, since
xvi
P R E F AC E
A N D
A C K N O W L E D G M E N T S
moving to the University of Chicago, incurred even more. I am particularly indebted to Charlie Glaser, the best foil a woolly-headed “institutionalist” could ever hope to have as a colleague—and also, I hasten to add, the most loyal and supportive of friends. Another of my debts, one I gladly continue to accrue, is to PIPES, Chicago’s Program on International Politics, Economics, and Security workshop. With Charles Lipson and my Harris School colleague Duncan Snidal at the helm, PIPES has made studying international relations at Chicago a thoroughly enjoyable experience, offering a weekly reminder of why I got into this business and why (most of the time, anyway) I still love it. Others around the Chicago campus who deserve special mention are Delia Boylan (again), Sven Feldmann, Thomas DeLeire, Howard Margolis, Brian Portnoy, Alex Thompson, and Stephen Walt, all of whom have taken the trouble to read parts of the manuscript and offer their suggestions for making it better. Among the many institutions that have supported my research, I want to single out the Massachusetts Institute of Technology, whose political science department was kind enough to offer me a visiting research fellowship in 1996–97. I am particularly appreciative of the encouragement I received that year from Stephen Ansolabehere, Eugene Gholz, Jeff Lewis, Richard Locke, Kenneth Oye, and James Snyder—as well as from Harvard scholars Jeffry Frieden and Lisa Martin, who were no less willing to answer my questions and suggest ways of clarifying particular points. Others who have commented on various aspects of my work over the years include James Caporaso, Stephanie Golob, Miles Kahler, Peter Lange, Helen Milner, Thomas Oatley, Ronald Rogowski, Arthur Stein, Ronald Wonnacott, and several anonymous reviewers, all of whom went well beyond the call of duty. Generous financial support for this project has been provided by the National Science Foundation, the MacArthur Program on International Peace and Security, the George P. Schultz Canadian Studies Program, the Andrew W. Mellon Foundation, and the North America Forum of the Institute for International Studies at Stanford University. The financial assistance of all of these organizations is greatly appreciated. In drafting the final version of the manuscript, I have had the distinct pleasure of working with Malcolm Litchfield, the editor in chief at Princeton University Press, as well as Peter Meyerhoff, Elizabeth Rourke, and my wonderful copy editor, Cindy Crumrine. For Cindy’s help in smoothing out the rough edges of my prose, I am deeply grateful—and believe me, dear reader, you should be too! Finally, I cannot even begin to express my gratitude to my wife, Delia Boylan, without whose support and companionship—and, yes, nudging—this book would never have been finished. Lest there be any doubt, Delia, this book is for you.
Ruling the World
C H A P T E R
1
Introduction: From Anarchy to Organization Without cooperation, we will be lost. Without institutions there will be little cooperation. And without a knowledge of how institutions work— and what makes them work well—there are likely to be fewer, and worse, institutions than if such knowledge is widespread. —Robert O. Keohane
A CALL FOR THEORETICAL REORIENTATION If supranational structures like the World Trade Organization, the European Monetary Union, and NAFTA have recently been “spreading like wildfire,”1 it is because they make it easier for states to cooperate—and in a world of rapidly increasing interdependence and complexity, the more interstate cooperation, the better. Inasmuch as there is a conventional explanation for the rise of supranational institutions, surely this is it. It’s true that the last several years have seen questions about the role of power, distributional conflict, bargaining, and coercion rise to the top of the cooperation literature’s analytical agenda.2 So far, however, few of the theorists responsible for this shift have thought to challenge the view, most closely associated with the “neoliberal” scholar Robert Keohane, that international cooperation—and hence international institutions—are fundamentally good things. Thus, while “neorealists” complain that Keohane’s theoretical orientation minimizes the importance of interstate disputes over the distribution of collective benefits (see esp. Grieco 1993 and Krasner 1991), they, too, assume that the cooperative equilibria that emerge once these distributional conflicts are resolved leave all participants, strong and weak alike, at least as well off as they would have been had the original, precooperation status quo remained in effect. In many ways, this convergence in liberal and realist opinion is a sign of theoretical progress. The underlying explanatory logic of the new liberalrealist research agenda has the virtue of being firmly anchored in rational 1
See, e.g., Whalley 1993. See, among many others, Grieco 1990; Krasner 1991; Garrett 1992; Knight 1992; Oye 1992; Morrow 1994; and Fearon 1998. 2
4
I N T R O D U C T I O N
choice foundations, and there is little question that its emphasis on possibilities for (and impediments to) mutual gain captures a great deal of what is going on.3 As productive as it has been, however, the line of theory developed by Keohane and subsequently elaborated by Stephen Krasner, Joseph Grieco, and other scholars has led to an overemphasis on the mutual-gains side of international cooperation. For while the prospect of collective gain may well be driving some of the action, there is a winners-and-losers dimension to recent developments that may, in many cases, be even more important. Might a number of the supranational arrangements we keep hearing about—even ones as “successful” as NAFTA and the newly reinvigorated European Monetary System (EMS)—end up working to the disadvantage of some of their participants? Although this conclusion runs directly counter to the positive-sum view of international cooperation that dominates today’s theoretical mainstream, the model of cooperation set forth in this book raises the possibility that, indeed, only some—perhaps a small fraction—of the supranational institutions that get created and endure actually benefit all of their signatories. To be clear, the issue here is not simply one of relative gains and losses. The NAFTA and EMS losers were indeed made worse off in relative terms—they did not benefit as much as the winners—but they were also worse off (or so I shall argue) in a more fundamental absolute sense. Had it been within their power to return the world to the original “noncooperative” status quo, they would have been only too delighted to do so. As we will see, the task of adducing evidence for this winners-and-losers perspective is complicated by a number of factors, not least of which is the need to imagine a world in which the cooperative arrangements in question had never gotten off the ground. What would the “utilities” of each cooperation partner have been if, for some reason, these arrangements had never materialized? Might some of these partners have been better off in this noncooperative (albeit hypothetical) situation? Counterfactual questions of this sort are inherently problematic, making it difficult to offer a definitive answer. That said, these methodological difficulties do not explain—and they certainly do not justify—the literature’s enduring preoccupation with the positive-sum consequences of international cooperation. Insofar as previous scholars have misconstrued the driving forces behind such far-reaching cooperative arrangements as NAFTA and the EMS (as I believe they have), the problem has far more to do with deficiencies in our theoretical understandings of cooperation and institution building than it does with any empirical or purely methodological limitations. If we really want to understand why supranational institutions are “ruling the world,” we will first have to broaden our current theories of collective action, extending their logic in ways that acknowledge and incorporate the possibility of both winners and (absolute) losers. That is what I hope to do in this book. 3
Cf. Olson 1965.
F R O M
A N A R C H Y
T O
O R G A N I Z A T I O N
5
Power Politics and the Institutionalization of International Relations So what, then, is the logic of (international) collective action? Few questions in the study of international politics are more basic or, indeed, more relevant to the concerns of today’s heads of state. And yet even after years of research, and despite the glaring policy significance of the issues involved, the truth is that we are still some distance from a completely satisfactory answer. It is tempting to think that this is because our empirical base is too limited, as if the deficiencies in our theoretical understanding of recent international developments could be overcome by simply piling up more facts or locating new sources of data. In fact, scholars of international relations and comparative politics have already accumulated a vast amount of information about interstate cooperation and institution building in Western Europe and the Americas, the two regions of the world in which current trends toward supranational governance have made greatest headway. Much is known, too, about the emergence of global arrangements like the General Agreement on Tariffs and Trade (GATT), whose member states recently established a new World Trade Organization (WTO) for the express purpose of mediating and resolving their future trade disputes. But while the literature is quite good at describing these momentous achievements,4 it has been markedly less successful at explaining why they occurred, and thus at telling us what sort of developments we ought to expect in the future. To this point, the newfound enthusiasm for supranational institutions—and for interstate collaboration more generally—has remained something of a mystery. This book traces the problem to a widely overlooked conceptual bias in the academic literature. Until now, most scholars have taken the view that international institutions, whatever their distinctive purposes or politics, are fundamentally welfare improving. Underlying this view is a simple two-step logic wherein institutions facilitate international cooperation (step 1), and cooperation in turn facilitates “utility gains” for each participating state or government (step 2). For well over a decade now, this positive-sum analytical orientation— most forcefully articulated in Robert Keohane’s groundbreaking book After Hegemony (1984)—is what has guided empirical inquiry in the study of international cooperation. Not that all has been sweetness and light, of course. Over the past few years, in particular, the disagreements among contributors to the “cooperation under anarchy” literature have been numerous, their internecine controversies heated and intense. Thus far, however, the debate inspired by After Hegemony has focused almost exclusively on the first leg of the institutions → cooperation → mutual-gains two-step: How, exactly, do institutions help their signatories overcome obstacles to collective action? Is it by weakening their incentives to free ride (as Keohane himself proposed), 4
See, e.g., Kahler 1995.
6
I N T R O D U C T I O N
or do institutions work their magic in other ways—by providing “focal points,” for example, or by altering the identities, and hence the underlying preferences, of the state actors who participate in them?5 But what about the second link in Keohane’s causal chain? Must international cooperation always leave each actor better off than it was in the initial, pre-cooperation status quo? While most scholars see international institutions as efficient, Pareto-improving responses to collective-action problems, this is not, it turns out, the only logical possibility. To demonstrate this point, the next chapter begins laying out a very simple theoretical model in which two states—the ultimate beneficiaries of cooperation—encounter an exogenous shock of some kind, be it a sudden shift in their relative “structural” positions or, more likely (since structural shifts tend to occur rather slowly), a sudden change in one of the state’s preferences, perhaps itself a consequence of a recently held election. In the wake of this shock, our two prime movers—neither of which had previously been inclined to coordinate its policies with the other—observe that they now have certain interests in common and thus could benefit, potentially a great deal, by structuring their future interactions in accordance with a common set of mutually agreed upon rules and procedures. In moving toward a more cooperative relationship, the two states (or, more accurately, their respective leaders) believe they will be able to pursue their joint interests more effectively than would otherwise be the case.6 What are their joint interests? Expanding the total pool of customers available to each other’s domestic producers is one possibility. Alternatively, the two states—let’s call them A and B—may be hoping to reduce the probability of their subsequently becoming embroiled in a costly military conflict. For present purposes, however, what matters is less the specific nature of A and B’s common interests than the fact that the cooperative arrangement they devise—their new multilateral regime—permits these interests to be realized. Having identified the motivations of the regime’s prime movers, I then take a look at the loser’s side of the story: Why might other actors in the international system choose to take part in the new regime established by A and B, its “enacting coalition”? Note that the motivations of these other cooperators may be very different from those of the regime’s original sponsors: although the prime movers have an understandable interest in complying with what is, after all, “their” regime, that same regime need not further the interests of other states. In contrast with A and B, who comply because they anticipate utility gains relative to the status quo, these other states—call them C and D—may actually prefer the original, noncooperative state of the world. But why, then, would C and D ever agree to come onboard? The an5
The preference-altering “transformational” impact of EU membership is explored in RisseKappen 1996, and Sandholtz 1993. On the logic of focal points, see Garrett and Weingast 1993. 6 Although conceptualizing the beneficiaries of cooperation as states or nations is fine for present purposes, I will argue later on that it is better to think of the prime movers not as states but as governments; cf. Moravcsik 1998.
F R O M
A N A R C H Y
T O
O R G A N I Z A T I O N
7
swer “because A and B make them” is too simple. What we would like to know, of course, is how A and B do this. How, precisely, do our two enactors succeed in getting other states or governments—some of whom would be perfectly content with the status quo—to agree to cooperate? The plot thickens. One possibility is that A and B resort to force, threatening these other actors (i.e., C and D) with economic sanctions or other forms of direct coercive pressure. These days, however, one rarely finds states being forced to take part in new multilateral ventures. To the contrary, states often line up to be admitted into cooperative institutions whose current members do not even want them. Are we therefore to conclude that international cooperation and institution building really are mutually beneficial affairs, that (absent blatant acts of coercion) everyone involved really does expect to derive utility gains relative to the status quo ante? Not necessarily. The fact that cooperation and institution building are matters of voluntary choice does not imply that the institutions in question raise (or even that their signatories expect them to raise) their baseline utilities. Given the losers’ preference for a world in which these arrangements did not exist, theoretical inquiry properly begins with the disruption of the initial, pre-cooperation status quo. How might the winners succeed in removing this status quo from the choice sets of the losers? While engaging in coercive diplomacy is one way, there is often, I will argue, an easier alternative: the new regime’s beneficiaries can simply “go it alone.” By pursuing this line of argument, this book shows how institutionalized cooperation by one group of actors (the winners) can have the effect of restricting the options available to another group of actors (the losers), altering the rules of the game such that members of the latter group are better off playing by the new rules despite their strong preference for the original, precooperation status quo. Once the winners seem likely to establish their new cooperative structure—something they would do even if they were destined to be its only members—the losers conclude that being left out would be even worse than joining. Their entry decisions are, in this sense, purely voluntary; indeed, the losers may well plead to be admitted into the new arrangement. And yet, despite their unabashed enthusiasm for inclusion, the losers would actually be much happier if this arrangement had never been created. While everyone who studies international cooperation is aware that power “matters,” most of the time it is either bargaining power (e.g., Fearon 1998; Krasner 1991; Oye 1992) or coercive power (e.g., Martin 1992a, O. Young 1983) that they have in mind. What I am talking about here, by contrast, is an altogether different kind of power, one that, unlike bargaining power, yields absolute—not just relative—losers and yet does not, like coercive power, require the winners to force their opponents into submission. At no point does anyone in the power-politics model I put forward coerce, or even attempt to coerce, anyone else. Nor, for that matter, do the beneficiaries of
8
I N T R O D U C T I O N
politics.”7
cooperation engage in “linkage The winners do not get the losers to acquiesce by compensating them (via side-payments) in some other sphere of interaction. The losers acquiesce because they know that the winners are in a position to proceed without them. In a myopic sense, of course, the losers really are better off cooperating. Yet—and this is the important point—that is only because the winners have presented them with a fait accompli. The resulting cooperative regime is not, then, the Pareto-improving outcome it outwardly appears. On the contrary, even when all cross-issue deals and side-payments are taken into account, the losers would still be happier if the original, noncooperative status quo had never been disrupted. Faced with a choice between joining the winners’ new cooperative system or being completely shut out, the losers enthusiastically submit their applications for membership. They do so, however, only because the winners’ actions have had the effect of removing the status quo from their choice sets, leaving them with what they view as a bad option (cooperating with the winners) and an even worse alternative (incurring the costs of exclusion). While this perspective offers a new way of thinking about why nations cooperate, it also provides new insight into how they cooperate. Among other things, for example, it suggests that the institutional preferences of certain cooperators (the governing officials who make up the enacting coalition) are likely to count for a great deal—and those of other cooperators, and of the collectivity as a whole, for practically nothing. Given the enacting coalition’s ability to proceed unilaterally, its members do not really need to take anyone else’s preferences into account at all. Whichever forms of cooperation they prefer will be the ones adopted in the end, whether or not these forms are conducive to the interests of other regime participants. But of course this simply moves the analysis back a step: Which kinds of institutional structures would the first group of participants—the enactors—tend to find most attractive? The obvious answer is the kinds that are most likely to endure. But this begs the question: endure what? What future uncertainties or threats might the enactors be particularly concerned about, and therefore try especially hard to mitigate, as they go about the task of designing “their” institutions? One might think that their greatest concern would be the possibility that one (or more) of the governments that made up the enacting coalition might refuse to hold up its end of the deal. Yet—as I am by no means the first to point out8 —the international relations literature has tended to exaggerate the vulnerability of cooperative arrangements to free riding and other forms of opportunistic behavior. If the rules of the game are clear, if compliance is easily monitored, and, most importantly, if the future stream of cooperative benefits or “payoffs” is expected to be substantial, members of the enacting coalition would presumably have strong incentives to honor one another’s commitments from the outset. 7 8
Cf. Frieden 1994; Garrett 1994; Martin 1992a and 1994; and Sebenius 1983. See, e.g., Grieco 1988.
F R O M
A N A R C H Y
T O
O R G A N I Z A T I O N
9
In fact, when it comes to the issue of regime durability, the real questions have less to do with the incentives of the enacting coalition than with the incentives of its domestic successors. Although the aforementioned benefit stream may be large for the political elites who make up the enacting coalition, it may be small, or even nonexistent, for the officeholders who succeed them. Consequently, the latter might wish not just to “defect” from the cooperative regime, but to destroy it altogether. While this would deny the beneficiaries of cooperation the utility gains they might otherwise have enjoyed, in the end cooperation is like beauty: its value is in the eye of the beholder. Insofar as their domestic opponents have their own conception of what it means to be “better off,” the governments that comprise the enacting coalition have a stake in finding a way to discourage one another’s internal successors from abandoning the international arrangement that they, the prime movers, are about to create. Why, it might be asked, wouldn’t the enacting coalition’s go-it-alone capabilities afford sufficient protection against this possibility? Wouldn’t the fear of exclusion suffice to keep the newly empowered “beholders” from sabotaging international regimes they dislike and whose initial creation they opposed? The answer is no—at least not necessarily—for unlike the enacting coalition’s external enemies (the losers in other countries), the coalition’s internal enemies (its domestic opponents) would be in a position to single-handedly bring down its new cooperative structures. Should they wish to do this, their withdrawal from these structures is all that would be necessary, since without their ongoing participation, the regimes in question would effectively cease to exist. And because these regimes would no longer exist, the “fear of exclusion” could hardly be expected to operate as a reliable deterrent. This is not to say that the losers would incur no costs in opting out of (and thereby destroying) the regimes they inherited upon taking office, particularly if these arrangements had already been up and running for some time. Indeed, the prime movers could be counted upon to try to make these costs— the price that their (potentially) hostile successors would pay for exercising their “exit options”9 —as high as possible. The flip side is at least as important, however, for one would also expect the enactors to try to reduce the costs to their regime’s would-be destroyers of staying in the regime. In short, the go-it-alone coalition’s interest in co-opting their domestic opponents might well influence the types of cooperative institutions they devise, inclining them toward “thicker” arrangements whose terms are not fully specified at the outset and thus leave the losers some (albeit limited) scope for ex post changes in the initial rules of the game. This, it seems to me, is exactly what logic would suggest. The problem with fully specifying the terms of cooperation ex ante is not that it would be more time-consuming, and hence less efficient, than allowing for subsequent revisions or “clarifications” over time. The problem is that it would deny the enacting coalition’s domestic successors any opportunity to moderate these 9
The phrase is Hirschman’s (1970).
10
I N T R O D U C T I O N
terms, reformulating—or simply reinterpreting—them in ways intended to make their continued participation in the regime somewhat less burdensome than it would otherwise be. It is for this reason, I will argue, rather than out of the (narrowly construed) efficiency considerations emphasized by previous scholars, that the contractual terms embodied in many of today’s international treaties take the supranational forms they do. Delimited though it is, the flexibility afforded by such arrangements works to co-opt the enacting coalition’s domestic opponents, reducing their propensity to mount a serious challenge to the new regime if, in future years, they should ever get the opportunity. All of this presumes, of course, that the utility consequences of interstate cooperation may differ radically depending on which parties, or even which individuals, happen to control domestic public authority. Allowing that internal leadership changes and government turnover can bring about abrupt shifts in “state” preferences certainly complicates matters, requiring as it does that the analyst peer inside the black box of domestic politics. In this case, however, the analytical payoff is well worth the effort. For if I am right, the realist view of states as internally undifferentiated units10 threatens to obscure what may well be the central motivating force behind the ongoing supranationalization of international cooperation: the vicissitudes of domestic politics inside the pivotal “enacting” states. In sum, power may permeate and infuse international cooperation just as it permeates and infuses international conflict. In the case of conflict, however, relations among nations are not strictly voluntary, and so the means by which the strong exercise power over the weak—outright bullying, brute force, and the like—are more transparent. Make no mistake: there are important differences between go-it-alone power and what might be termed structural or institutional power. Whereas the victims of go-it-alone power are “outsiders”—public officials in countries located outside the enacting coalition—the losers in the case of institutional power are “insiders”—the enacting coalition’s domestic successors. Other differences have to do with the powerful actors’ intentions. In one case (go-it-alone power) the pressure that is brought to bear on the losers is inadvertent; in the other case (institutional power) the enacting coalition goes out of its way to alter the losers’ decision-making calculus, deliberately moderating the costs of cooperation on actors who, were they to exit the cooperative arrangement in question, would effectively bring about its demise. As important as these distinctions are, however, my main concern is with the more fundamental point that power’s influence on international affairs may be far more pervasive than many scholars, including even many hard-nosed realists, currently appreciate. Whether one is talking about go-it-alone power or institutional power, the end result is the same: Without in any way being bullied or coerced, governments choose to participate in cooperative international arrangements they genuinely, and intensely, dislike. 10
See esp. Waltz 1979.
F R O M
A N A R C H Y
T O
O R G A N I Z A T I O N
11
The Plan of This Book Are all of the participants in today’s international institutions truly better off than they were in the pre-cooperation status quo, or can cooperative institutional arrangements spread like wildfire even when their most ardent “supporters” fully expect to get burned in the process? How much of the current trend toward collective decision-making, institutional flexibility, and delegated authority is being driven by efficiency considerations, and how much by power politics? The first four chapters of the book provide the theoretical groundwork necessary for answering these questions. Chapters 2 and 3 offer a new perspective on the question “why do states cooperate?” After sketching out the conventional explanation for the recent proliferation of cooperative initiatives, I introduce a simple model to demonstrate how an apparently clearcut case of voluntary cooperation may in fact represent an instance of go-italone power. In this model, the weak cooperate only because the strong present them with a fait accompli. This basic framework is further developed in chapters 4 and 5, where the focus shifts from the question of why nations cooperate to that of how they cooperate: Is there a power politics of institutional design and, if so, how might its internal logic differ from that of the efficiency-centered perspectives we have grown accustomed to seeing in the literature? The next four chapters supplement this theoretical discussion with analyses of two of the most important developments in the history of regional trade and monetary relations: the inaugurations of NAFTA (chapters 6 and 7) and the European Monetary System (chapters 8 and 9) during the mid1990s and late-1970s, respectively. In offering these case studies, I hope to show more concretely how theories of international cooperation that fail to appropriately integrate power into their larger explanatory edifice can end up obscuring more than they reveal. At the same time, however, these case studies also serve an additional purpose: They allow me to explore some of the many interesting nuances and subtleties in the book’s larger theoretical argument. Before launching directly into this argument, I should alert the reader to some of the empirical difficulties that it presents. One of these has already been noted: the outside observer must determine how well off a particular member government would have been if, counterfactually, that government had not joined the cooperative arrangement in question. To perform this sort of welfare comparison, the analyst needs to know (among other things) what that government’s leaders were originally seeking to accomplish during their tenure in office. At some point, therefore, the analyst is forced to specify the underlying preferences of the relevant government actors. This is admittedly a complicated task. On the other side of the coin, however, those scholars who embrace the conventional wisdom on international cooperation—that, as Keohane puts it in After Hegemony (1984, 88), “[international] regimes are developed . . . because actors in world politics believe that with such
12
I N T R O D U C T I O N
arrangements they will be able to make mutually beneficial agreements that would otherwise be difficult or impossible to attain”—seem to have no difficulty assuming (i.e., without even attempting to specify the underlying state or government preferences) that collective action, when it occurs, is in fact “mutually beneficial.” In any event, the fact that the book’s core hypotheses are difficult to test does not mean they are impossible to test. To the contrary, parts III and IV offer a whole series of tests, albeit ones that are more suggestive than definitive. The NAFTA and EMS stories I flesh out below may not qualify as “crucial cases” for my theoretical argument. At the same time, however, it would certainly have been possible for my investigations of these cases to turn up no evidence whatsoever either that defensive considerations—rather than any expectation of positive-sum gain or fear of coercive sanctions—were what motivated some of the participants in NAFTA and the EMS to take part, or that these signatories would actually have been better off, in the sense that their governments would have been less disadvantaged politically, had the pre-cooperation (i.e., pre-FTA or pre-EMS) states of the world remained in effect. The fact that I do turn up such evidence, while it may not prove the empirical validity of my theoretical claims beyond any doubt, does at least lend them plausibility—which, given the counterfactual and preference-specification problems noted earlier, may be all one can reasonably expect at this stage of the game. Let me note, finally, that in mapping out the logic of go-it-alone power, I am not seeking to account for international cooperation in any particular issue-area or geographic region. To be sure, I believe the theoretical approach developed here offers important new insights into the origins and structural foundations of NAFTA and the EMS. In singling out these two cases, however, I do not mean to suggest that this approach is applicable only to North America and Western Europe or only to cases of economic cooperation. In fact, though I dwell here upon the NAFTA and EMS experiences, there is every reason to think that the book’s larger analytical framework can also be used to shed light on other (e.g., non-Western, noneconomic) kinds of supranational institutions, and even other kinds of domestic institutions.11 In much of what follows, however, I deliberately set aside these “generalizability” concerns so as to keep the analysis as clear and simple as possible. 11 The latter strikes me as a particularly promising avenue for future research (see chapter 10). Indeed, much of what I will have to say on the subjects of power, cooperation, and institutions was foreshadowed in the work of an “Americanist,” the great E. E. Schattschneider. See esp. Schattschneider 1960. That the concept I term go-it-alone power bears a strong family resemblance to Schattschneider’s notion of agenda power will become apparent as we go along. Nevertheless—and for reasons I discuss at some length in chapter 3—these are not the same concepts. Quite the contrary, the differences between go-it-alone power and Schattschneider’s earlier conceptualization are as noteworthy, it seems to me, as their similarities.
P A R T
I
Why Do Nations Cooperate?
C H A P T E R
2
Institutions, Collective Action, and the Prospect of Mutual Gain THE NOTION that international institutions are worthy of close examination first gained broad acceptance during the early postwar years. Over the course of the 1960s and 1970s, however, the consensus among scholars shifted 180 degrees. As structural realism came to supplant the largely descriptive institutionalist traditions of the 1940s and 1950s, the supranational organizations that had once been the major focus of attention—principally, the United Nations—were deemed irrelevant or “epiphenomenal.” For realists, power and interests were the predominant forces driving world politics. Institutions were seen as mere facilitators, translating the interests of powerful states directly and unproblematically into international outcomes.1 It would be natural to suppose that this movement toward an institutionfree, balance-of-power theory of international relations arose as a reaction to theoretical deficiencies in the institutionalist literature of the 1940s and 1950s. And, indeed, much of this literature was deficient, with scholars analyzing organizational rules and procedures indiscriminately, almost never pausing to ask where they came from or how, if at all, they influenced or constrained the behavior of member states.2 Only recently, however, have realists felt the need to justify their neglect of institutionalist scholarship.3 From an empirical standpoint, supranational institutions just did not seem to matter very much, as states often violated their institutional obligations with impunity. Nor were realist scholars alone in thinking that institutions could be safely omitted from the analysis of politics. Proponents of pluralist theory (e.g., Truman 1951) also tended to leave institutions out of their larger explanatory framework. By the mid-1980s, though, institutions had once again begun to emerge as a subject of interest among students of international affairs. And this time things were different: Unlike their intellectual forebears, proponents of the “new” institutionalism recognized the importance of grounding their analysis in an overarching theoretical framework. Robert Keohane stands as the leading figure in this effort to go beyond the legalistic, institution-by-institution approach characteristic of earlier institutionalist thinkers. Initially dubbed “regime theory,” the voluminous litera1 The locus classicus of realist thought is Kenneth Waltz’s Theory of International Politics (1979). Other major works in the realist tradition include Carr 1946; Morgenthau 1958; and Waltz’s earlier book, Man, the State, and War (1959). 2 For a more charitable assessment, see Martin and Simmons 1998. 3 See, e.g., Mearsheimer 1994/95.
16
W H Y
D O
N A T I O N S
C O O P E R A T E ?
work4
ture spawned by Keohane’s took direct aim at Waltz and other realists who, in their preoccupation with international conflict and the balance of power, had failed to conceive of interstate cooperation as anything more than the spontaneous formation of status-quo-preserving, marriage-of-convenience alliances. Since rechristened “neoliberal institutionalism” (Grieco 1988), regime theory has become the theoretical point of departure for the bulk of all academic writing and research on the institutional developments of the 1980s and 1990s.5 The purpose of the next four chapters is to subject this literature to critical theoretical scrutiny. While others have engaged in similar efforts, much of their criticism has focused on what I regard as secondary issues. Thus, proponents of the new institutionalist paradigm are often taken to task for failing to adequately integrate the role of ideas, identity formation, and culture into their larger explanatory framework.6 While a source of concern, however, this is not, it seems to me, a crippling defect. Nor is the problem that international institutions do not really “matter,” in the sense of having a major (independent) impact on the flow of world events.7 Insofar as the institutionalist paradigm is deficient, the fault lies primarily, I will argue, in its theoretical one-sidedness: Why assume that multilateral institutions always and everywhere facilitate mutual gains? Though it is true that states throughout the world are scurrying to join such institutions—or, when shut out, are rushing to create new, smaller-scale institutions of their own—one cannot assume that the resulting “progression” from anarchy to organization is therefore welfare enhancing. Undergirded by rational choice assumptions about state behavior, the model of international cooperation I will introduce in this part of the book raises the possibility that the past several years’ frenzy of supranational activity, rather than fostering peace and prosperity, will end up working to the great disadvantage of some (perhaps most) participating nations.
Back to Basics: Why Do States (Sometimes) Engage in Collective Action? The publication of After Hegemony (1984) could not have been better timed. Emboldened by their success in revolutionizing the study of economic organization, rational choice theorists were about to launch a series of forays into other disciplines. In the field of international relations, Keohane’s justly celebrated book paved the way. 4
Esp. by Keohane 1984. Among the major works within this tradition are Keohane 1984; Kupchan and Kupchan 1991; Lipson 1984; Martin 1992a; Milner 1997; Moravcsik 1998; Snidal 1985b; and Yarbrough and Yarbrough 1992. See also the contributions to Krasner 1983, and Oye 1986a. 6 See, e.g., Goldstein and Keohane 1993; Katzenstein 1996; and Wendt 1994. 7 Cf. Mearsheimer 1994/95. 5
C O L L E C T I V E
A C T I O N
A N D
M U T U A L
G A I N
17
The starting point for the book’s argument is a simple observation: Policies and events originating in one nation often have large—albeit indirect— effects on the welfare of citizens and governments in other nations.8 Sometimes these spillover effects are mutually beneficial: If German economic policy makers succeed in boosting Germany’s domestic living standards, German consumers have more money to spend on goods and services, and French workers who specialize in producing exports for the German market enjoy a windfall. Other times, however, these spillovers are unwanted, as when a currency devaluation in the United States reduces the competitiveness of U.S. imports from Europe, prompting European countries to devalue their own currencies, thereby precipitating a further devaluation by the United States, and so on. In situations where independent policy making would give rise to what economists call “negative externalities,” Keohane drew the not unreasonable inference that nations would be better off actively coordinating their policies. Thus, the political leaders of country A could agree not to pursue policies harmful to country B so long as their political counterparts in country B refrained from undertaking policies harmful to country A, and vice versa. At bottom, then, cooperation is about reciprocity. As Keohane (1988, 380) writes: “When cooperation takes place, each party changes his or her behavior contingent on changes in the other’s behavior.”9 The result of this exchange—of what Helen Milner (1997, 7–8) characterizes as “the adjustment of one state’s policies in return for, or anticipation of, the adjustment of other states’ policies”—is that all parties end up better off. But if all parties benefit, why haven’t state leaders always been as willing to cooperate as they are today? Drawing on Keohane’s work, neoliberal theorists give two answers. The first is that nations today are more interdependent than they used to be, and thus the potential for negative externalities is greater—as, it follows, are the prospective benefits of multilateral cooperation—than was the case in previous years.10 That said, these benefits may not be sufficient, at least not in themselves, to induce the kinds of mutually beneficial policy adjustments that would leave all states better off. While there are many reasons for this, Keohane and his followers emphasize the potential for ex post defections: A state that suspects its cooperative overtures will not be reciprocated will rationally avoid making these overtures in the first place.11 This line of analysis suggests a second reason for the recent proliferation of new multilateral arrangements. Part of the explanation may be that the “knowledge of how institutions work—and what makes them work well” (Keohane 1988, 393) is becoming more prevalent. By using this knowledge to design “better” institutions, national leaders have ensured that the one8
Cooper 1980 and 1986. Emphasis in the original. See also Keohane 1984, 12 and 51– 52. 10 See, e.g., Milner 1997, 59. 11 Cf. Olson 1965, as well as Hardin 1982. 9
18
W H Y
D O
N A T I O N S
C O O P E R A T E ?
Figure 2-1. The Positive-Sum Logic of Neoliberalism
shot payoff available to opportunistic defectors amounts to something less than the ongoing stream of collective benefits available to long-term cooperators, the result being utility gains for everyone involved. Neoliberal theorists can be credited with putting forward a cogent explanation for recent developments, one that is grounded in clear, deductive reasoning. Before we accept that explanation’s universal validity, however, it is worth taking a moment to consider the second link in its implicit causal chain (Figure 2-1). Is it not conceivable that institutionally-mediated cooperation, rather than resulting in the realization or “capture” of collective gains, works exclusively to the benefit of the actors who set the process in motion, increasing their power and prosperity at the expense of other actors who participate only because they have no better option? Not only is this conceivable—it is often, I would argue, exactly what happens. Thus far, however, neoliberal theory’s supposedly tough-minded realist critics have taken a different line of attack. Instead of questioning whether multilateral cooperation does in fact help states “manage” globalization, realists have devoted their energies to showing that—the presence of international institutions notwithstanding—cooperative interstate relations are next to impossible to sustain, at least on a voluntary basis, for any significant length of time. Realists, in other words, have taken issue with the first arrow shown in Figure 2-1, not the second.12
The Realist Assault on Neoliberal Theory In the years following the publication of After Hegemony, Keohane’s claims about the role of institutions in facilitating cooperation among nations 12 In concentrating here on neoliberal and realist perspectives, I necessarily give short shrift to other approaches within the study of international cooperation. Two that deserve mention are critical theory (Finnemore 1996; Kratochwil and Ruggie 1986; Wendt 1987 and 1992) and neofunctionalism (E. Haas 1964; Nye 1971; Keohane and Hoffmann 1991). Proponents of these approaches criticize neoliberals and realists alike for (among other things) taking the underlying preferences of international actors as fixed and exogenously given. I will discuss this objection, as well as my own views on the sources and consequences of “preference instability,” in the analysis to follow; see esp. chapter 5.
C O L L E C T I V E
A C T I O N
A N D
M U T U A L
G A I N
19
emerged as the subject of intense controversy, inspiring a lively, sometimes rancorous debate between neoliberal scholars, who were broadly sympathetic to Keohane’s thesis, and realist scholars, who took the view that international institutions were essentially irrelevant. At first, as mentioned earlier, those in the latter camp felt little need to justify their position. Again, realists have always been more interested in understanding what states can do to preserve the status quo than in exploring how, by devising appropriate multilateral institutions, they might be able to make things better (and, as an empirical matter, none of the multilateral arrangements created in the immediate aftermath of the Second World War seemed to be having an especially large impact anyway).13 While numerous countries participated in institutional arrangements like the GATT, the European Community, and the UN, realists doubted that any of these arrangements would ever be permitted to encroach upon their member states’ freedom of maneuver—at least not in any serious or meaningful way.14 By the mid-1980s, however, these and other multilateral institutions had begun playing a significant role in the flow of regional and global developments, and realists seemed increasingly out of step with the world around them.15 At that point, realist critics of neoliberal theory could have retreated from their hard-line, anti-institutionalist position in favor of a more nuanced perspective. After all, the acknowledged father of neorealism, Kenneth Waltz, had never seen any reason why the logic of anarchy should prevent countries from behaving in neoliberal fashion, even if the resulting “rules, institutions, and patterns of cooperation [were all likely to be] limited in extent and different from what they might otherwise be.”16 But engaging in the messy business of distinguishing institutions that effectively promoted cooperation from those that were merely epiphenomenal was not what most realists had in mind. Instead, they continued to insist that international institutions had little (independent) impact on world politics, and that liberals who argued otherwise were relying upon “faulty logic.”17 Where, exactly, did liberals go wrong? According to one line of realist criticism, it was in neglecting the importance of “relative gains.”18 So much attention has been lavished on the relative-gains critique that the liberal-realist debate of the past several years is sometimes equated with the controversy over relative gains.19 Casting the debate in this way is a mistake, however, for this was but one flank in a larger realist offensive. 13
Jackson 1989; P. Taylor 1983. See, e.g., Waltz 1979, 115 –16. 15 Cf. Ruggie 1992. 16 Waltz 1986, 336. 17 Mearsheimer 1994/95. 18 See esp. Grieco 1988 and 1990. 19 See, e.g., D. Baldwin 1993. 14
20
W H Y
D O
N A T I O N S
C O O P E R A T E ?
Are International Agreements Enforceable? For many realists, an even bigger problem with neoliberal theory was its failure to explain how cooperating nations overcome the practical difficulties of treaty enforcement and compliance.20 According to Keohane, multilateral cooperation actually required little more than that each partner be able to distinguish fellow cooperators from free riders. Once these distinctions were clear, the assumption was that each participant would take it upon itself to reward the former and penalize the latter, all in a strictly decentralized fashion.21 In fact, this assumption was crucial, for if prospective free riders did not expect to be punished, the whole cooperative system would break down, preventing everyone—free riders and cooperators alike—from enjoying the ensuing benefits of policy coordination. To be effective as a deterrent, however, sanctions (against free riders) often need to be leveled by a large number of (cooperating) countries. Naturally, this raises the possibility that each member of the coalition of potential sanctioners will attempt to “pass the buck,” thereby impeding the decentralized enforcement mechanism so fundamental to the success of long-term cooperation. And the danger of buck-passing aside, what incentive would goodfaith cooperators have to impose retaliatory sanctions anyway? By the time their coalition partner had defected, it would already be too late; the damage would already have been done. This line of reasoning is consistent with one of the central tenets of realist theory, namely, that the kinds of enforcement problems so easily addressed in a domestic setting are intractable in an international one. In most domestic societies, impediments to enforcement created by collective-action and “time inconsistency” problems are mitigated by the coercive power of the state. When citizens or private corporations violate the law or fail to live up to contractual obligations, various state agencies (local police, national law-enforcement bodies, the courts) spring into action and the malefactors are made to pay for their crimes. By contrast, the anarchic setting of world politics has no police force, no judicial system, and no overarching legal order to ensure that those who violate the terms of their agreements will be prosecuted. Such violations may be “against the law,” and the nations that commit them may even face a serious risk of detection, but given the difficulties in organizing international sanctions,22 retaliation is rarely either swift or certain. Not only might deterring free riders in the international environment prove difficult, but the consequences for a cooperating nation of being victimized could be catastrophic. Here, too, realists draw a sharp distinction between the realms of domestic and international politics. In addition to safeguarding contracts, property rights, and the like, a primary function of 20
Cf. Morgenthau 1958. Chapter 4 provides a more detailed discussion of this point. 22 See, e.g., Gilpin 1984; Hufbauer and Schott 1985. 21
C O L L E C T I V E
A C T I O N
A N D
M U T U A L
G A I N
21
any domestic legal order is to penalize anyone who attempts to inflict bodily harm against his or her fellow citizens. One consequence is that people entering into purely domestic transactions need not fear for their physical safety. By contrast, when transactions occur across national boundaries (in other words, in the absence of a higher legal authority), a determined free rider could resort to military action, threatening the very survival of its transaction partners: not content to usurp their wealth and resources, an outlaw nation could encroach upon their sovereignty as well. The high risks and serious consequences of exploitation are both recurrent themes in realist thought. Realists do not claim that international cooperation is impossible, only that in a world where states cannot be compelled to keep their promises and free riders can wreak havoc, the incentives for states to enter into reciprocal international arrangements are weak at best. As for the multilateral schemes that have been springing up in recent years, these are hollow shells: If countries join, argue realists, it is only because they know they will not be asked to do anything they would not be doing anyway—and if they ever were, they could extricate themselves from their supposedly “binding” treaty commitments at little or no cost. Realists have adduced numerous historical examples in support of these claims. Among the most widely cited is the International Energy Agency.23 Though the IEA was the main empirical referent for the neoliberal arguments put forward in After Hegemony, during the 1974–81 period on which Keohane’s analysis focused, the cooperative regime over which the organization presided actually had little independent impact on the behavior of its members, a fact that (as realists have delighted in pointing out) Keohane himself concedes at various places in the book.24 Similar questions have been raised about the ostensible behavioral impact of membership in higher-profile organizations such as the United Nations and International Labor Organization.25 Finally, there are several prominent cases of states withdrawing from cooperative institutions altogether, as when, in 1984, an overreaching International Court of Justice led the United States to temporarily withdraw its prior acceptance of ICJ jurisdiction in disputes over U.S. policy toward Central America.26 On the other hand, not all regional and multilateral organizations are as costless to exit as the ICJ, ILO, and IEA. Granting that appearances are often deceiving, an inspection of recent history suggests that government leaders do sometimes allow themselves to become enmeshed in collaborative international arrangements from which they can withdraw only at what they (or their successors) perceive to be very high political costs.27 The fact that this 23
Cf. Grieco 1993, 308; Mearsheimer 1994/95, 25. Keohane 1984, 16 and 236. 25 See, e.g., Krasner 1985. 26 Abbott and Snidal 1998, 23. 27 As further evidence that “multilateralism matters,” some analysts would cite the enormous amounts of time and attention that go into “preparing, drafting, negotiating, and moni24
22
W H Y
D O
N A T I O N S
C O O P E R A T E ?
is so—that the anarchic nature of the international system has apparently not prevented states from negotiating enforceable long-term cooperative agreements—suggests that realists greatly exaggerate the domestic-international distinction.28 Certainly, the police powers of the state do make the domestic setting a more hospitable place to do business. But the fact that domestic polities have laws against breaching contractual agreements only partly explains why domestic transactions are safer. Though the threat of state coercion may deter some would-be free riders, it cannot be expected to deter all forms of disorderly behavior. As an enforcement mechanism, coercion just does not work very well—or not, at any rate, well enough to produce the degree of stability and cooperation one finds in most domestic societies. Notwithstanding numerous imperfections in the administration of justice, domestic citizens and firms routinely comply with long-term relational contracts that would be difficult or impossible to enforce in a court of law. They comply, in some cases, even when a flagrant violation by one party would spell catastrophe for the others. In short, the international system’s absence of a higher legal authority does not appear to be as significant as realists make it out to be. This observation points to a major weakness in the realist critique of decentralized enforcement: realists are simply wrong to assume that the risks of international cooperation necessarily outweigh the benefits. In theory, the accumulated gains that two or more nations expect to derive from their cooperative relationship may be great enough to offset—and could themselves serve to mitigate—the risks of exploitation by an opportunistic free rider, particularly if states that renege on their promises can be easily identified. Not wishing to forfeit the long-term benefits of cooperation, a prospective free rider might rationally forgo its opportunity to defect, thus giving rise to a sort of virtuous circle: the lower the perceived risk of being double-crossed by an opportunistic free rider, the higher the value each party will place on its ongoing relationship, and hence the lower each party’s propensity to free ride, and so on. Realism’s Second Line of Attack Issues concerning the fragility of decentralized enforcement are brushed aside by Grieco and other scholars in the vanguard of realism’s other, more recent line of attack. It is not that they believe the enforcement of international agreements is easy, but rather that it is, in their view, a second-order toring treaty obligations” (Chayes and Chayes 1993, 186– 87). Why, if states could costlessly terminate their multilateral obligations, would one find so much pushing and hauling among national leaders and their designated negotiators? Of course, realists could counter that so much attention is lavished on the design of multilateral arrangements precisely because each signatory wants to make absolutely certain that its obligations impose few if any meaningful constraints on its freedom of action. 28 Milner 1991; Powell 1992; Snidal 1996.
C O L L E C T I V E
A C T I O N
A N D
M U T U A L
G A I N
23
problem. Before state leaders get around to considering whether an agreement would be “self-enforcing,” they first ask whether the cooperative outcome produced by that agreement would work to their advantage. If they conclude that it would not—that the new arrangement would not in fact leave them better off—they will steadfastly refuse to take part and, consequently, the arrangement will never see the light of day.29 The thrust of Grieco’s argument is that international cooperation is far less extensive than neoliberals have made it out to be because states have concluded that participation in cooperative regimes—even perfectly enforceable ones—would do more to undermine than to promote their national interests. What are these interests? What would it take for a cooperative arrangement to be truly Pareto-improving? According to Grieco, two conditions would be necessary. First, the arrangement would have to afford each of its member states “absolute” gains—that is, gains vis-à-vis the initial, precooperation status quo. Second, the gains would need to be distributed in a manner that did not interfere with any member’s efforts to maintain its relative position within the international distribution of power. It is this latter, “relative-gains” constraint that Grieco emphasizes in his own research, and that he faults Keohane and other neoliberals for having omitted from theirs.30 The methodology employed in Grieco’s analysis has been subjected to a barrage of criticism. To capture the notion that security concerns can lead countries to forgo agreements that seem mutually beneficial on other grounds, Grieco treats nations as maximizers of a combination of absolute and relative gains.31 Critics argue that Grieco is wrong to incorporate states’ concerns about their relative positions (i.e., relative to the positions of their cooperation partners) directly into their utility functions, and that states should instead be analyzed as if they wish to maximize absolute gains alone. This is a reasonable suggestion. After all, Grieco is not claiming that states care about their relative status for its own sake. If states worry about their partners’ absolute payoffs, it is because they fear that by falling too far be29 It is worth emphasizing that the two realist critiques, though distinct, are not mutually exclusive. Thus, for example, John Mearsheimer’s (1994/95) study draws more or less equally on both. When criticizing proponents of collective security, Mearsheimer relies primarily on arguments about the theoretical and practical limitations of decentralized enforcement. In contrast, his critical analysis of neoliberalism (which he sees as a distinct school of institutionalist theory) is essentially a restatement of the relative-gains argument discussed here. 30 Waltz was among the first to raise the possibility that because states care greatly about the distributional allocation of gains, international transactions are less frequent than they might otherwise be. See esp. Waltz 1979, 115 –16. The same theme was later expanded upon by Joanne Gowa (1986, 172–79) and Oran Young (1986, 118 –19). Joseph Grieco (1988 and 1990) then picked it up, developing it into a full-blown critique of neoliberal theory. Echoing Waltz, Grieco sees the anarchic nature of the international system as predisposing all states to “worry at the extreme that today’s friend may be tomorrow’s enemy in war, and thus . . . that achievements of joint gains that advantage a friend in the present might produce a more dangerous potential foe in the future.” Grieco 1990, 28 –29. 31 Cf. Stein 1990, chap. 5.
24
W H Y
D O
N A T I O N S
C O O P E R A T E ?
hind they run the risk of incurring more than offsetting absolute losses at some point in the future. In this view, Grieco’s argument is really about “a tradeoff between short term absolute gains (i.e., immediate payoffs from cooperation) and long run absolute gains (i.e., security over the long haul).”32 These modeling deficiencies are rather easily corrected, at least in principle. The real problem with the relative-gains critique of neoliberalism is not that its basic logic is unsound. The problem, rather, is one of empirical robustness: Are relative-gains fears widespread enough to justify Grieco’s sweeping indictment of neoliberal theory? Grieco himself concedes that relative-gains considerations do not necessarily impede cooperation. In fact, logic suggests they would do so only when the disproportionate gains enjoyed by the relative winners could be easily put to use in ways that threatened the political and territorial integrity of the relative losers. How often will this be the case? There are, of course, some circumstances in which relative-gains considerations would pose no obstacle whatsoever. The onus is thus on proponents of the relative-gains critique to show that such instances are exceedingly uncommon. Yet this is exactly what proponents do not do. In fact, Grieco goes out of his way to show how built-in side-payments and various other institutional devices can be incorporated into international agreements from the outset to ensure that their fruits are distributed in an equitable fashion, with no one signatory deriving a disproportionately large share of the absolute gains these agreements make possible.33 If using compensatory mechanisms to equalize the distribution of benefits is as easy as Grieco suggests, it is hard to see how relative-gains considerations could pose a particularly serious impediment to cooperation. For the sake of argument, however, let us suppose that relative-gains differentials were altogether impervious to these kinds of equilibrating devices. In that case, the prospects for cooperation would depend, among other things, on the would-be relative losers’ sensitivity to these differentials. In Grieco’s game-theoretic analysis, this sensitivity is indicated by the parameter k: the higher a state’s k-value, the greater its concern for avoiding relative losses, and hence (holding other variables constant) the lower its propensity to cooperate. But while proponents of the relative-gains critique have an obvious stake in showing high ks, the reality is that many international agreements have little direct impact, either positive or negative, on the security interests of the states that enter into them. There are two reasons for this. One has to do with the particular kinds of issues they address. As a rule, for example, international agreements concerning matters of economic or environmental policy tend to produce fewer security externalities—and are therefore perceived as “safer”—than agreements concerning matters of stra32
Snidal 1991, 704. See also Powell 1992 and 1993. One of Grieco’s favorite examples is the “divergence indicator” of the European Monetary System. See chapter 9 below. 33
C O L L E C T I V E
A C T I O N
A N D
M U T U A L
G A I N
25
tegic or military policy (Lipson 1984). The second, more important reason is that many states, even ones whose own military arsenals are relatively modest, have the luxury of taking their security more or less for granted, a luxury they enjoy because their territories just happen to be of strategic importance to a larger superpower. Also important in lessening security-related concerns about relative losses is the stabilizing impact of nuclear weapons, a factor inexplicably left out of Grieco’s otherwise careful analyses of the GATT and EMS (see below).34 Beyond the “Relative-Gains” Debate There are many reasons for dismissing the relative-gains critique of neoliberal theory as an empirical irrelevance. Few would argue with Grieco’s assertion that interstate cooperation is a tricky business. But to attribute this fact to states’ concerns about relative gains is to make a much stronger claim, one that requires the analyst to demonstrate (albeit counterfactually) that a great many perfectly enforceable collaborative interstate ventures never come into existence because (1) at least one member fears it would incur a relative loss and (2) the relative losers would suffer a marked decline in their ability to defend against potential foreign aggressors. While one can make a compelling case that this logic applies to (the dearth of) trade initiatives between Eastern- and Western-bloc nations during the Cold War years,35 Grieco would push the argument further. In his view, relative-gains considerations were the major reason negotiators for the United States and the European Community failed to agree to a coordinated reduction in nontariff barriers during the Tokyo Round of the GATT.36 This is a most peculiar claim. When the Tokyo Round was completed in 1979, the United States was certainly not in any danger of being attacked by its European allies, at least not at any point in the foreseeable future. And European countries (the relative losers in Grieco’s analysis) had even less reason to fear an imminent assault by their partner across the Atlantic.37 34 See, e.g., Grieco 1990, 222–23. As numerous scholars have noted, the existence of nuclear weapons gives defensive strategies a major advantage over offensive ones. See, among others, Glaser 1990; Jervis 1978; and Waltz 1990. 35 Gowa 1994. See also Mansfield 1994. 36 Grieco 1990. 37 When it comes to evaluating Grieco’s empirical claims, even John Mearsheimer, whose own debunking of neoliberal theory draws its ammunition from Grieco’s theoretical arguments (see, e.g., Mearsheimer 1994/95, 12–13 and 19 –24), can be counted among the skeptics. In the course of examining the prospects for peace in post-Cold War Europe, Mearsheimer considers, but then explicitly rejects, the view that relative-gains fears were responsible for undermining transatlantic economic cooperation prior to the collapse of the Soviet Union (Mearsheimer 1990, 45–47). With the United States self-interestedly using its power to guarantee political and economic stability throughout the Western alliance, member states of the EC had the luxury, in Mearsheimer’s view (cf. Walt 1987), of focusing single-mindedly on the pursuit of wealth and other (absolute) objectives. To be sure, Mearsheimer believes all of this could change now that the Cold War is over and the U.S. strategic interest in creating a bulwark against the Soviet threat
26
W H Y
D O
N A T I O N S
C O O P E R A T E ?
To be fair, Grieco’s is not the only study purporting to adduce concrete evidence for the relative-gains position. Proponents also point to Michael Mastanduno’s detailed examination of U.S. responses to Japanese industrial targeting during the late 1980s. Mearsheimer, for example, cites the Mastanduno study (1991) as evidence that “concern about relative gains . . . was an important factor in shaping American policy towards Japan in three cases: the FSX fighter aircraft, satellites, and high-definition television.”38 Here too, however, there are good reasons to be skeptical that security-related relative-gains concerns are what really drove U.S. policy. To make a credible case, one would need to argue, first, that America’s ability to prevail in a future Japanese-American confrontation was in doubt and, second, that Japan would have been able to launch a discernibly more effective campaign against the United States if it had been permitted to co-develop a new fighter jet, acquire a market niche in satellites, or win the race to develop a better television. Each of these claims stretches the imagination. What, then, accounts for Mastanduno’s positive findings?39 In fact, all three cases were part of a broader American effort aimed at counteracting Japanese strategic trade policy for purely economic reasons. Mastanduno is explicit on this point: The immediate concern was not military security, but economic well-being. Some U.S. officials feared that certain patterns of economic interaction with Japan, even though mutually beneficial in absolute terms, would bring relatively greater economic benefits to Japan and over time work to the detriment of America’s competitive position in the development and application of advanced technology. In response to this concern, a series of “strategic” policies emerged for consideration, including technology restrictions in aircraft, the disruption of Japan’s home market sanctuary in satellites, and the promotion of American firms in the competition to develop [high-definition television].40
In effect, members of the U.S. Congress and their supporters in the Bush administration invoked “national security” as a pretext for torpedoing a series of U.S.-Japanese agreements that might have cost jobs, reduced profits, or otherwise proved damaging to the short-term economic interests of their constituents.41 is no longer as transparent. If, in the future, relative-gains fears do become salient, he, following Grieco, would predict a precipitous decline in economic collaboration between the United States and the EU. Not all scholars agree with this assessment, however. The obvious criticism is that it fails to give sufficient weight to the logic of mutually assured destruction. On the continuing dominance of defensive weaponry and its relevance to post-Cold War European security, see, among others, Glaser 1994/95; Levy 1990; and Van Evera 1990/91. 38 Mearsheimer 1994/95, 21. 39 Strictly speaking, only some of Mastanduno’s findings were “positive.” In the case of highdefinition television, for example, his study concluded that relative-gains considerations “affected U.S. policy only modestly, if at all.” Mastanduno 1991, 75. 40 Mastanduno 1991, 109. 41 It is perhaps revealing that the agencies most directly responsible for “defending the na-
C O L L E C T I V E
A C T I O N
A N D
M U T U A L
G A I N
27
So what remains of the realist assault on neoliberal theory? Why, exactly, is “cooperation under anarchy” so much more problematic than neoliberals have led us to believe? The answer offered by Mastanduno is that it poses a domestic-political rather than a security risk: Government leaders refuse to take part in multilateral initiatives for fear of incurring the wrath of politically influential domestic constituencies. After years of searching for the fatal flaw in neoliberal theory’s causal structure, is this the best that realists can do? After all, the new realist critique is hardly devastating. On the contrary, it would appear to lay the foundations for a new higher-order, liberal-realist consensus. Multilateral cooperation is possible, so the argument would run, if (but only if) three basic conditions are satisfied. First, to ensure that all of the relevant cooperative exchanges are “self-enforcing,” each participant must be able to clearly distinguish between cooperators and defectors. Second, any institutions established for this purpose must be designed in ways that safeguard their member states’ vital security, including, most importantly, that of any “relative losers” who would otherwise find it too risky to participate. And finally, the entry into force of these institutions must afford each participating government greater domestic political gains than domestic political losses.
The Emerging Liberal-Realist Consensus Although neoliberals and realists often seem to revel in controversy, both camps have recently begun moving toward something like the consensus position I have just outlined. The result is that international institutions are increasingly seen—even by many realists—in neoliberal terms. As Grieco declares in the closing pages of his book, “international institutions do matter for states as they attempt to cooperate.”42 As for why cooperation itself matters, Grieco takes the answer to be self-evident: Multilateral cooperation helps states further their common interests (recall the second arrow in Figure 2-1). Indeed, were it not for the prospect of realizing utility gains, state tional interest” tended to be the least, not the most, sensitive to relative gains. In the case of the April 1989 aircraft agreement, for example, the State and Defense Departments were strongly in favor of Japan’s being allowed to co-develop the FSX fighter jet with General Dynamics, a U.S.-based corporation. Indeed, officials in these departments were the ones who drafted the memorandum of understanding that those who were sensitive to the relative-gains issue, including interest groups, lobbyists, and members of Congress, subsequently fought to overturn (Mastanduno 1991, 86). The story is much the same with satellites and high-definition television: in each case, opposition to the proposed arrangements came not from the parts of the government that had reason to take an interest in the welfare and security of the country as a whole—the State Department, the Department of Defense, the Office of Management and Budget, the Council of Economic Advisors—but from officials at the Department of Commerce and the Office of the United States Trade Representative, as well as from individual members of Congress. 42 Grieco 1990, 233 – 34, emphasis in the original.
28
W H Y
D O
N A T I O N S
C O O P E R A T E ?
leaders would never join multilateral institutions, much less take an active role in creating them—or so Grieco, echoing much of the literature, seems to assume. Since no one can force a state to coordinate its policies, if two states cooperate it must be because their leaders believe that doing so will ultimately work to their mutual advantage. Of course, some analysts are more strongly wedded to this way of thinking than others. Regarding the use of force, for example, the legal scholar Oran Young (1983) draws a distinction between voluntary and “imposed” (i.e., nonvoluntary) regimes, and even Keohane allows that international cooperation may not always be “benign.”43 Rarely, however, do the members of multilateral institutions in the real world attempt to coerce unwilling signatories into joining. To the contrary, states frequently seek admission into institutions whose current members do not even want them—an observation that sits rather uneasily with Young’s emphasis on imposition, coercion, and involuntary action.44 As for Keohane’s assertion that cooperation is not always benign, the losers he seems to have in mind are states located outside the cooperative regimes in question. One example here would be the OPECsponsored oil cartel of the 1970s, a regime that imposed untold losses on the United States and other nonparticipating countries.45 Note, however, that from the standpoint of the oil regime’s insiders, membership was most assuredly welfare enhancing. In this context, Keohane’s remarks are quite consistent with his larger claims about the role of institutions in lowering transaction costs and mitigating problems of market failure.46 43
See, e.g., Keohane 1984, 72. Donald Puchala and Raymond Hopkins (1983, 66) are particularly emphatic on this point: “[Regimes] distribute rewards to the advantage of some and disadvantage of others, and in so doing they buttress, legitimize, and sometimes institutionalize international patterns of dominance, subordination, accumulation, and exploitation.” A similar line was taken by the original proponents of hegemonic stability theory, who viewed the multilateral arrangements established in the immediate aftermath of World War II as devices for promoting the interests of the United States and not, or at least not necessarily, those of weaker, nonhegemonic states (Gilpin 1975; Krasner 1976). In recent years, however, a more “benevolent,” public-goods variant of hegemonic stability theory has come to supplant the “malevolent” rendering put forward in the mid-1970s by realists Robert Gilpin and Stephen Krasner. Today, the hegemon’s role is seen as one of establishing and maintaining institutions that advance the interests of all (or nearly all) states in the system, not just its own. See, e.g., Keohane 1980; Snidal 1985b; and Martin 1992a. 44 To be sure, Young’s discussion of “imposed orders” does not concentrate only on coercion and imposition. He is well aware that co-optation and the manipulation of incentives, phenomena that are clearly related to the argument I develop here, also play a role. See esp. O. Young (1994). 45 “Indeed,” writes Keohane in After Hegemony (1984, 73), “some regimes (such as alliances or cartel-type regimes) are specifically designed to impose costs on nonmembers.” Similarly, Helen Milner (1997, 6 n. 1) suggests that “cooperation may just be collusion, a way of fixing prices to increase one’s profits.” 46 Cf. Coase 1937 and Williamson 1985. It is not surprising that After Hegemony’s discussion of the potentially welfare-reducing effects of international cooperation is as brief as it is, nor that these passages have been all but ignored in the extensive body of research inspired by Keohane’s work. After all, the emphasis throughout After Hegemony—and echoed in Keohane’s
C O L L E C T I V E
A C T I O N
A N D
M U T U A L
G A I N
29
As these larger claims continue to take root, the debate between neoliberal and realist scholars appears to be entering a new, more conciliatory phase. No longer, it seems, are Keohane’s ideas about the possibility of “cooperation under anarchy”—or about the potential role of multilateral institutions in facilitating it—the hotly contested issues they once were. Instead, what is needed, virtually everyone now seems to agree,47 is a better understanding of where international institutions come from and why they take the distinctive forms they do. In working aggressively toward this goal, realists can be expected to place particular emphasis than neoliberals will on the influence and dynamics of state power. (A case in point is a much-discussed article by Stephen Krasner; see below.) But while realists may be especially attuned to the distributional struggles that attend the design and evolution of institutions, their insights in no way constitute a frontal challenge to neoliberal theory. Rather than seeking to destroy neoliberalism’s theoretical edifice, realists are now themselves actively building upon it. To see this more clearly, consider Figure 2-2, a stylized utility representation of what I term the “collective-action paradigm.” The vertical axes of the three diagrams in this figure indicate the preferences of actor Y, with points (utility imputations) located toward the top of each diagram corresponding to Y’s most preferred outcomes or states of the world. The preferences of a second actor, X, are arrayed along the horizontal axes, with X’s utility rising as one moves toward the right of the diagrams. Proponents of the collective-action paradigm hold that institutions facilitate cooperation—in this case, between X and Y—the end result of which is an improvement in the utilities of each cooperator (if not necessarily those of other actors) relative to the initial noncooperative status quo, SQ. But the fact that cooperation would help X and Y escape a plainly suboptimal situation does not tell us how X and Y will cooperate. The diagrams displayed in Figure 2-3 are meant to capture the distributional conflicts that can arise when actors are confronted with numerous Pareto-improving possibilities.48 Insofar as efficiency considerations militate against frequent renegotiation of cooperative rules and procedures, high stakes may be involved in the resolution of seemingly trivial matters of equilibrium selection and institutional design. In the end, however, it is reasonable to assume that powerful states will exert greater influence over this process than weaker states will and that,
subsequent writings (see, e.g., Keohane 1988)—is on how, by devising appropriate institutions, actors in the international system may be able to overcome collective action problems and thus succeed, the famous skepticism of realists notwithstanding, in securing peace and prosperity. Had Keohane instead emphasized the point I am making—namely that cooperation and institution building may, in certain circumstances, leave some insiders worse off—it is fair to say it would have weakened the thrust of his larger theoretical enterprise. 47 Abbott and Snidal 1998; Koremenos, Lipson, and Snidal 1999; Martin and Simmons 1998. 48 Cf. Krasner 1991.
Figure 2-2. The Collective-Action Paradigm: A Utility Representation
C O L L E C T I V E
A C T I O N
A N D
M U T U A L
G A I N
31
Figure 2-3. The Collective-Action Paradigm: Integrating Power into the Analysis
consequently, any institutional arrangements that emerge will conform more closely to the preferences of the powerful than to those of the weak.
Conclusion A key assumption in all of this is that the weaker actor, though it fails to obtain its most-preferred equilibrium, is still made better off. Even if X does not do nearly as well as its cooperation partner, Y, the onset of cooperation nonetheless generates welfare gains for both actors relative to the initial status quo. The implication of Krasner’s work is that all regime participants, no
32
W H Y
D O
N A T I O N S
C O O P E R A T E ?
matter how strong or weak, end up better off as a result of having coordinated their policies. So what is wrong with this line of analysis? Nothing, if one shares the dominant view that institutions arise to solve collective-action problems. Yet the fact that the members of a particular institution are willing to coordinate their policies should not be taken as an unambiguous sign that all of these members see the institution as an improvement on the status quo ante. It is time to consider another possibility.
C H A P T E R
3
Winners and Losers: The Case for Theoretical Reorientation AS THE INTELLECTUAL AGENDA begins to transcend the liberalism-versusrealism debate of the past several years, there is for the first time the promise of a truly comprehensive “grand theory” of international cooperation. So far, however, the scholarly analysis of supranational institutions—like that of international institutions more generally—has been framed as a debate between those who view these institutions as vehicles for mitigating collective action problems (the “liberal” position) and those who dismiss them as little more than window dressing (the “realist” position). And yet logically speaking there is, and has always been, a third alternative. Having previously suggested as much, I now want to make this alternative conception explicit, clarifying its core theoretical foundations and drawing out what I consider to be its most important logical implications for the study of international cooperation. A good place to begin is by taking a closer look at what scholars mean when they speak of “power” as a motivating force in international relations. Even if it is true that power’s importance derives from its role in determining which point along the Pareto frontier is chosen,1 that still leaves open the question of how, exactly, power does this. Or to put it in terms of Figure 23: How does actor Y (the more powerful actor) get its partner, actor X, to agree to a cooperative equilibrium that is closer to Y’s ideal point than it is to X’s?
Power and Bargaining The last decade has witnessed an explosion of research on the nature and dynamics of interstate bargaining.2 Virtually all of this work draws, either explicitly or implicitly, upon Albert O. Hirschman’s groundbreaking study, National Power and the Structure of Foreign Trade (1945). Although Hirschman focused on trade relations, his central variable—the differential opportunity costs of noncooperation—plays a central role in all bargaining relationships. 1
Krasner 1991. Among the major contributions are those of Fearon 1998; Friman 1993; Krasner 1991; Milner 1997; Morrow 1994; Oye 1992; Putnam 1988; and Wagner 1988. Of particular interest has been the influence of bargaining on the evolution of the European Community. See, among many others, Garrett 1992; Moravcsik 1991 and 1998; P. Taylor 1982. 2
34
W H Y
D O
N A T I O N S
C O O P E R A T E ?
The underlying logic is straightforward: The lower a negotiating partner’s potential gains from cooperation relative to the status quo, the greater is that partner’s capacity to “hold out” for its most-preferred cooperative equilibrium.3 By this logic, negotiations over matters of institutional design should result in structural biases that favor a new institution’s least enthusiastic participants. If a signatory is willing to incur a high risk of non-agreement, its input in bargaining over the structure of a proposed institution will tend to be greater, as will its share (relative to the shares accruing to other signatories) of any cooperative benefits made possible by the institution’s creation. This line of argument is widely invoked to explain why nearly all multilateral institutions display a built-in bias toward certain signatories. Why are some member states entitled to a substantially larger share of the cooperative gains than others? Following Hirschman, Stephen Krasner and other scholars would direct our attention to differences in each state’s initial “tolerance” for the status quo ante, the intuition being that a government with a high tolerance for the pre-cooperation status quo would be able to assume a higher risk of non-agreement. Conversely, a nation whose leaders were impatient for change would be likely to acquiesce rather quickly, since to reject the proposals put forward by its negotiating partners would be to perpetuate a status quo it was eager, if not desperate, to overturn. Yet this is not the only possibility. In some cases, Y’s bargaining strength may owe less to its being happy with the status quo (or less unhappy than X is) than to the fact that Y, unlike X, can alter that status quo unilaterally. What if, at the onset of negotiations, Y were to establish a credible “threat point,” as in Figure 3-1? In this figure, Y’s unilateral move results in an outcome that benefits all cooperators with respect to the initial status quo: the utilities of X and Y are both higher than they were at the onset of negotiations. Logically, though, there is no reason to think that coercive power plays of this kind will always result in Pareto improving outcomes. Coercive cooperation (cf. Martin 1992a) is one possibility; but so, too, is the Pareto-inefficient outcome depicted below, in Figure 3-2. Here, Y’s threat to impose sanctions or other penalties against noncooperators induces an equilibrium situated in the upper-left-hand quadrant: an improvement on the status quo for Y but a substantial setback for X.
Power, Bargaining, and Coercion There was a time when questions concerning the potential role of force and intimidation as instruments for inducing compliance with international regimes were at the forefront of scholarly debate among students of international cooperation. At the center of that debate were the three “Bretton 3
Cf. Raiffa 1982; Rubinstein 1982 and 1985; Sutton 1986.
W I N N E R S
A N D
L O S E R S
35
Figure 3-1. Power as Coercion
Woods” institutions—the World Bank, the International Monetary Fund, and the GATT—whose creation in the late 1940s and early 1950s laid the foundations for the open trading system that flourished in the decades following World War II. By creating foreign markets for U.S.-produced goods, this system furthered U.S. interests and values. In many countries, however, political elites took a more negative attitude toward free trade, fearing that shocks emanating from the global economy would give rise to social unrest and, in so doing, further undermine their already tenuous hold on domestic political authority. Why, then, did these elites agree to take part in the Bretton Woods institutions? Of crucial importance for some scholars was the ability of the United States to apply coercive pressure against “smaller” (nonhegemonic) states in the system.4 The fact that the United States emerged from the war as the world’s dominant economic power afforded it tremendous leverage. As the world’s largest consumer and producer of traded goods, the United States 4
See, e.g., Gilpin 1975 and Krasner 1976.
36
W H Y
D O
N A T I O N S
C O O P E R A T E ?
Figure 3-2. An Alternative Model of Power Bargaining
could threaten to close off its market to nations that did not share its commitment to multilateral norms and principles. U.S. officials also had recourse to other forms of leverage, such as denying other countries vital military or economic assistance.5 The strategic use of coercion has also been explored in recent work on 5 The original proponents of this “malevolent” strand of hegemonic stability theory were not always clear about the source of the hegemon’s leverage. Krasner, for example, devoted as much attention to the strategic manipulation of positive inducements (e.g., the offer of foreign aid for postwar reconstruction) as he did to negative sanctions, thus implying that regime compliance may not have been so harmful after all. Krasner 1976, 52; cf. Haggard and Simmons 1987, 502, and Lake 1993, 469. For analytical purposes, positive inducements—or, in common parlance, bribery—must be distinguished from threats. The implementation of a threat worsens the recipient’s welfare relative to the status quo ante. In this respect, threats are inherently coercive, even if the fear of being subjected to them leads actors to coordinate their behavior in mutually beneficial ways. This stands in marked contrast to positive inducements: the delivery of a promised side-payment enhances the recipient’s welfare relative to the status quo. While threats and side-payments
W I N N E R S
A N D
L O S E R S
37
linkages.”6
“strategic cross-issue Consider the above-mentioned study by Lisa Martin. In contrast with Krasner and Gilpin’s “malevolent” hegemon, Martin’s hegemon supplies a public good (see Figure 3-1 above).7 The hegemon’s interests are nonetheless in conflict with those of its smaller partners, for the hegemon is better off if subordinate states help defray the cost of supplying the public good, whereas the latter are better off letting the hegemon bear the entire burden. Because the hegemon places a high value on the public good it provides, it cannot credibly threaten to withhold its own contribution as a way of inducing its partners to pay their fair shares. Martin’s insight is that there may be other, more effective (but equally coercive) means of eliciting contributions that would not harm the hegemon’s own interests as much. If the hegemon would like other states to contribute to public good P1, it can threaten to withhold its contribution from public good P2, where P2’s provision is less important to the hegemon and hence could be given up more easily. As Martin concedes, however, efforts to link the provision of disparate goods may themselves raise questions of credibility, since they would require the hegemon to explicitly coordinate its policies across different issue areas. In fact, credibility problems can thwart virtually any attempt to exert one’s will by deploying sanctions or other types of coercive pressure. This is because the application of sanctions nearly always imposes high political and economic costs on the international actors doing the sanctioning.8 While it is true that these costs could be lower than the benefits the sanctioner expects to derive from its act of coercion, the sanctioner, in calculating the costto-benefit ratio, would need to take explicit account of the target’s likely response. And if the target is intent on standing firm, the costs of applying coercive pressure against it could easily exceed the would-be sanctioner’s expected benefits—a point noted long ago by John Harsanyi (1965), Anatol Rapoport (1960), and Thomas Schelling (1960), among others. In short, scholars have known for quite some time that, to be effective, “compellent” threats must be credible. And yet we also know that coercion is costly—sometimes hugely so—for coercer and coercee alike. In the end, are often treated as two sides of the same coin (cf. Martin 1993, 416–17), it is only the former that entail the exercise of power by one actor over another. For more on the distinction between “threats” and “offers,” see Wertheimer 1987, 204 – 6. 6 See particularly Frieden 1994; Martin 1992a. 7 The role of coercion as an instrument of cooperation is no less important to proponents of the “benevolent” variant of hegemonic stability theory (like Martin) than it is to proponents of the earlier malevolent school. The difference is that the benevolent hegemon coerces its subordinates—by imposing penalties for protectionist practices, enforcing international property rights, and the like—for their own good. See the useful discussions in Snidal 1985b and Lake 1993. 8 A good example here is the disruption of normal trade flows that occurred in the wake of the U.S. grain embargo of the Soviet Union during the 1970s. One consequence of the boycott was a rise in short-term unemployment among a small but politically powerful segment of the U.S. population.
38
W H Y
D O
N A T I O N S
C O O P E R A T E ?
therefore, coercion has major limitations as an instrument of power. This might seem an obvious point, but many studies of “power politics” nonetheless fail to appreciate its significance.9
How Go-It-Alone Power Differs from Coercion This, then, is the big lesson to be drawn from Figure 3-2: If actor Y’s only means of obtaining an outcome outside the Pareto set were to apply coercive pressure against actor X, such outcomes would probably not occur very often, for the act of applying this pressure would reduce Y’s own utility (note the positioning of the threat point below the initial status quo SQ). This, of course, reduces the credibility of Y’s coercive threats. And if Y’s threats are not credible, X will continue to regard SQ—and not Y’s threat point—as the appropriate baseline when deciding whether Y’s institution would leave it “better off.” At the same time, however, it would be wrong to conclude that because coercion is problematic—and because, as an empirical matter, most cooperative arrangements are entered into voluntarily—these arrangements are necessarily welfare improving. Consider the following alternative. In Figure 3-3, circumstances unique to actor Y are altered—due, perhaps, to a sudden change in Y’s domestic political leadership—such that this actor (the ultimate winner) would now be better off changing its initial behavior regardless of how actor X (the ultimate loser) chooses to respond. In this case, actor Y’s threat is wholly credible precisely because it is not really a “threat,” that is, Y’s action is not directed against X. Indeed, if X were to reject Y’s proposed form of cooperation, Y would simply “go it alone.” And yet the final outcome in this figure is precisely the one that Y sought to obtain in the previous figure. In the end, X is induced to participate in a coop9 On the limitations of economic sanctions as instruments of foreign policy, see Bayard, Pelzman, and Perez 1983; Gilpin 1984; Hufbauer, Schott, and Elliott 1990; and Pape 1997. Each of these studies factors in the potential costs of sanctions to the sanctioning country. These are not, however, the only determinants of success or failure. Whether a trade boycott is likely to result in the discontinuation of the practices that provoked it will also depend upon the vulnerability of the target country, which is itself a function of the demand elasticity of any goods previously imported by that country and the availability of alternative sources of supply. Also important are noneconomic considerations: the degree to which public officials are insulated from civil unrest, the domestic political costs of backing down, national pride, and so on. Two target countries may show differing degrees of resolve in the face of equally strong external pressures. For a discussion of why governments continue to resort to economic coercion despite copious empirical evidence that sanctions seldom have a major impact on the policies of their intended targets, see Kaempfer and Lowenberg 1992 and esp. D. Baldwin 1985. Most explanations focus on the influence of domestic interest groups. In mobilizing support for sanctions, these groups may be more interested in transmitting signals to political authorities in their own countries than in altering the policies of target countries.
W I N N E R S
A N D
L O S E R S
39
Figure 3-3. Go-It-Alone Power
erative arrangement that sharply lowers its utility relative to the status quo ante. Why? Because it recognizes that Y’s unilateral actions have removed that status quo—as well as the cooperative equilibrium that X itself prefers—from the set of feasible alternatives. What distinguishes go-it-alone power from coercive power (i.e., where the winners directly threaten the losers) is that, in the former case, the winners possess a “dominant strategy,” meaning simply a choice that is optimal regardless of the choices made by others.10 Given Y’s go-it-alone potential, how does X decide what to do? The answer is that it compares the disutility of accepting Y’s offer with the disutility of SQ9, the new status quo that would ensue if Y were to proceed unilaterally (as it surely would given its new incentive structure). Confronted with a pair of unattractive alternatives, 10
Luce and Raiffa 1957, 77–79.
40
W H Y
D O
N A T I O N S
C O O P E R A T E ?
X rationally chooses the option with the lowest disutility and agrees to coordinate its policies in accordance with Y’s preferred set of rules and procedures.11 Because the actions undertaken by Y, the initiator and beneficiary of this process, are not contingent upon the actions of X, the loser, the voluminous literatures on strategic interaction and threat credibility may capture only a small part of what is actually going on. In a sense, the credibility issue does not even arise, since Y’s unilateral actions impose no costs on Y itself. Even if X were to “stand firm,” Y would still be better off proceeding on its own. Given that X can do nothing to prevent this, there may not in fact be any need for Y to link issues or otherwise continually deploy (or threaten to deploy) coercive measures against it.12 Of greater potential relevance here are arguments dating back to the 1950s and 1960s about the exercise of influence through “agenda control.” An important contribution to this earlier debate was the simple—in retrospect, too simple—definition of power postulated by Robert Dahl in his famous 1957 article, “The Concept of Power.” Actor A exercises power over actor B, argued Dahl, to the extent that A succeeds in getting B to do something that B would not otherwise do. Attention immediately turned to the question of means: How, exactly, is A able to pull this off?13 On one side of the debate were early proponents of pluralist theory, including Dahl himself. At the risk of oversimplification, pluralists held that the key to A’s getting B to alter its behavior was A’s greater ability to mobilize a coalition in support of its preferred policy alternatives and, in so doing, obtain their passage into law.14 Other scholars criticized pluralists like Dahl for concentrating exclusively upon A’s ability to defeat B in head-to-head 11 It is worth emphasizing that the moves displayed in Figure 3-3 are not the only possible explanation for cooperation. I will return to the question of my argument’s “conditionality” later on in this chapter and, indeed, at many points throughout this book. See esp. chapter 10. 12 To be clear, Y’s go-it-alone potential would not deny X a choice. Strictly speaking, the losers in a power play always have a choice. The problem is that, if the winners’ threats are credible, all of the available options—as in the proverbial “your money or your life”—are worse than the status quo ante. By the same token, declining to participate in Y’s proposed arrangement would remain an option for X. The problem is that the utility consequences of nonmembership would now be considerably worse for X than before Y—whether by successfully engaging in coercive diplomacy or, as in the present case, by simply acquiring the capacity to go it alone—engineered a shift in the status quo. On the general point that some actors in the international system may be able to achieve their will by removing the status quo from the choice sets of other actors, see also Oatley and Nabors 1998. While I am (obviously) sympathetic to this line of inquiry, Oatley and Nabors do not develop its underlying logic in the systematic fashion I do here and, perhaps as a consequence, fail to appreciate that their “redistributive” outcomes may be a product of behavior that is inadvertent rather than intentional. 13 For a brief but incisive review of the ensuing debate, see Lukes 1974. 14 Hence Dahl’s method of inquiry in Who Governs? was “[to] determine for each decision which participants had initiated alternatives that were finally adopted, had vetoed alternatives initiated by others, or had proposed alternatives that were turned down. These actions were then tabulated as individual ‘successes’ or ‘defeats.’ The participants with the greatest propor-
W I N N E R S
A N D
L O S E R S
41
contests between each other’s preferred alternatives. The best known argument for broadening Dahl’s conception of power remains that of Peter Bachrach and Morton Baratz.15 According to Bachrach and Baratz, a less visible, but no less pervasive, way of exercising influence is for A to deny B the opportunity to vote for alternatives that would undermine A’s interests were they to be adopted: Of course power is exercised when A participates in the making of decisions that affect B. But power is also exercised when A devotes his energies to creating or reinforcing social and political values and institutional practices that limit the scope of the political process to public consideration of only those issues which are comparatively innocuous to A. To the extent that A succeeds in doing this, B is prevented, for all practical purposes, from bringing to the fore any issues that might in their resolution be seriously detrimental to A’s set of preferences.16
In raising the possibility that actors may be able to exert their will without directly threatening their opponents, the model of go-it-alone power set forth here bears a strong resemblance to the notion of agenda control elaborated by Bachrach and Baratz.17 There are differences, to be sure. In practice, for example, scholars sympathetic to Bachrach and Baratz’s approach have tended to focus on the ways in which agenda setters limit the choice set by changing the underlying preferences of other actors—or, as Bachrach and Baratz put it in the passage quoted above, by “creating or reinforcing social and political values.”18 In contrast, the model of go-it-alone power depicted in Figure 3-3 holds fundamental preferences (i.e., interests) and payoff structures constant.19 tion of successes out of the total number of successes were then considered to be the most influential.” Dahl 1961, 336. 15 Bachrach and Baratz 1962. Cf. Schattschneider 1960. 16 Bachrach and Baratz 1962, 948. 17 Although Bachrach and Baratz’s insights are applicable to a variety of settings—including, as I suggest here, the setting of international relations—subsequent investigation of the “two faces” of power has tended to concentrate on decision making (and non-decision making) within national legislatures. The idea that domestic political actors can wield power not only by mustering broad-based support for legislative initiatives they wish to see enacted but also by altogether excluding certain issues from the legislative agenda has played an especially prominent role in the analysis of congressional committees in the United States. That said, Shepsle and Weingast (1981) would argue that the extensive powers enjoyed by each committee chair within her committee’s policy domain are in the collective interest of all members of Congress. Given the potential for indeterminacy in majority-rule settings with unconstrained voting, all members are better off ceding agenda control powers to the chair—and, by implication, acquiescing in the passage of the chair’s preferred legislation—than insisting upon the right to vote for any proposal they wish and (potentially) setting off a never-ending majority-rule cycle. In this view, it was precisely the specter of “chaos” that led the U.S. Congress to set up the committee system in the first place. 18 See, e.g., Baumgartner and Burns 1975; Baumgartner, Buckley, and Burns 1975; Gaventa 1980; Lukes 1974. 19 In questioning whether fundamental preference change accounts for the loser’s decision to cooperate, I do not mean to imply that “national interests” are somehow fixed or immutable.
42
W H Y
D O
N A T I O N S
C O O P E R A T E ?
Another difference concerns the purposes for which agenda-setting powers are put to use. For many years, conservative “outliers” in Congress (mostly southern Democrats) used their control over key congressional committees to ensure that the legislative status quo would persist in the face of strong outside resistance, for example, by preventing liberal initiatives from coming up for a vote. In contrast, the agenda setters in Figure 3-3 achieve their will not by preserving the status quo but, to the contrary, by removing it from the set of feasible alternatives.20 At a more fundamental level, however, the logic of go-it-alone power is quite compatible with more traditional, Bachrach-and-Baratz-like arguments about the strategic manipulation of alternatives. One similarity is that changes in the loser’s behavior are viewed, in both cases, as a function of actions taken by the winners: the losers see themselves as victims of a “power play.”21 But the parallels run deeper still, for these power plays are not overt, as they are in the pluralist models criticized by Bachrach and Baratz, E. E. Schattschneider, and others. The winners achieve their objectives neither by outvoting their opponents nor—of greater relevance here—by engaging in blatant acts of coercion or intimidation. As the agenda-control literature reminds us, there are other, less transparent (but potentially no less effective) means of exercising leverage. A similar theme runs through Marxist theories of exploitation.22 When workers submit to capitalist institutions, it is not because anyone holds a gun to their heads. It is because their employers exercise exclusive control over the resources necessary for human survival, a privilege they enjoy as a result To the contrary, the fact that these interests can shift radically over time, often in response to electoral outcomes, offers important clues into why the winners might go to the trouble of establishing institutions in the first place. The relationship between political turnover and institutionalization is explored in chapter 5. For empirical applications to NAFTA and the EMS, see chapters 6 and 9, respectively. 20 Relatedly, the utility of the losers in many agenda control arguments is presumed to be lower than it might have been in some hypothetical—typically idealized—state of the world; cf. Biersteker 1993. Here, by contrast, the status quo that is removed from consideration is one that actually existed—and, indeed, would have persisted but for the winners’ go-it-alone capabilities. 21 This is not, of course, the only possibility. In theory, the losers could also be victimized by broader technological or “systemic” changes that the winners themselves played little or no direct role in bringing about. While such explanations may be valid in particular cases, it should be noted that they do a poor job of accounting for the two cases of cooperation that are examined later in this book. In the case of the EMS, for example, government support for the system is sometimes ascribed to an uncoordinated yet nonetheless dramatic increase in crossnational capital flows (Frieden 1991; Andrews 1994). The reality, however, is that the transition to a world of highly mobile capital was only just beginning to take place when, in 1978, Italy “surrendered” control over its exchange rate. Had it not been for the imminent prospect of a Franco-German exchange rate pact, it is doubtful that the Italian government would have deviated from its previous policy of using currency depreciation to stimulate output and employment. See chapter 8 below. 22 See, e.g., G. Cohen 1979, and Elster 1983; cf. Perrow 1986, 224 – 36.
W I N N E R S
A N D
L O S E R S
43
of having earlier, and in some cases quite fortuitously, accumulated sufficient quantities of capital. For Marxists, workers are indeed better off selling their labor power, but that is only because the alternatives with which they are presented—unemployment, impoverishment, starvation—would be even worse. The notion that political actors can impose their will on others unwittingly, without resort to brute force, is also central to arguments about the effects of U.S. market power during the nineteenth century. Scott James and David Lake (1989, 7–8) have observed that when the U.S. government unilaterally reduced its tariffs on cotton in 1846, cotton exporters located in countries outside the United States experienced a windfall. Newly enriched, these exporters in turn put pressure on their own governments to open up their economies. According to James and Lake, the resulting alteration in the economic policies pursued by America’s trading partners ended up working to the advantage of the U.S. officials who, without foreseeing the consequences of their actions, set the process in motion. Public officials also enjoy significant market power domestically. Given the vast sums spent by U.S. federal, state, and local authorities, even small changes in public sector purchasing priorities (e.g., the adoption of new laws requiring all government-procured automobiles to be equipped with special safety features) can lead private providers to implement costly changes in the profile of goods and services they produce.23 If the public sector were not such a large consumer, shifts in the status quo brought about by its unilateral—but, again, noncoercive—actions would be less likely to induce private firms to make these expensive adjustments.
Go-It-Alone Power as a Simple Three-Player Game To illustrate the logic of go-it-alone power somewhat more formally, Figure 3-4 depicts the first stage of a rudimentary extensive-form game, its payoff structure deliberately configured such that deviating from the status quo constitutes a dominant strategy for two of the three participants. (The first number indicates player A’s payoff, the second number B’s payoff, and so on. The higher the number, the larger is that player’s payoff relative to the status quo.) JUMP-STARTING THE COOPERATION BANDWAGON
In the story that corresponds to the game tree in the figure, there are in fact two winners, players A and B. Player C is the loser. At the beginning of play, an exogenous development of some kind—the abrupt departure from office of one of A’s major ruling parties, for example—causes the original (noncooperative) status quo to deteriorate in the eyes of A’s political leadership. 23
See, e.g., Vogel 1995.
44
W H Y
D O
N A T I O N S
C O O P E R A T E ?
Figure 3-4. Cooperation as a Game of Winners and Losers
To redress the problem, A tries to elicit B’s entry into a new cooperative arrangement.24 Because the actual emergence of this arrangement does not occur unless and until B accepts A’s new offer, how B chooses to respond in the initial period is crucial. (The term initial is used loosely here. If each actor is fully rational and enjoys perfect information about the strategic constraints under which others make decisions—if, in other words, each actor can anticipate exactly how others will behave at every given moment—then any decisions about whether to cooperate should occur instantaneously.) A’s inability to shift the status quo on its own is reflected in the game’s payoff structure: if B were to refuse A’s entreaty, (1) the path of action would follow the uppermost branch of the game tree, (2) each of the three players would continue to receive its status quo payoff, and (3) the state of the world would remain unchanged. As the game has been constructed, however, it would be irrational for B to refuse A’s offer. Assuming that player C lacks sufficient resources to “buy off” player B—that is, by offering it side-payments in an effort to prevent its moving first with A (cf. Sutton 1986)—B’s entry would afford utility gains not only for A, but for B as well. In this myopic sense, the establishment of the cooperative regime would represent a clear Pareto improvement: cooperation would improve both A’s and B’s utilities relative to the preexisting status quo. 24 Some readers may find my use of the term go-it-alone power a bit misleading in this threeplayer context, since it is not a single winner (e.g., a United States or a Japan or a Germany) but a collection of winners—what I earlier referred to as the “enacting coalition”—that goes it alone. I recognize the potential for confusion here and would be open to substituting a better, more precise formulation. I should note, however, that the expression go-it-alone power is appropriate so long as one conceives of the winning coalition as just that: a coalition.
W I N N E R S
A N D
L O S E R S
45
In fact, B would be better off agreeing to A’s request regardless of the move taken by actor C. If C were to join A’s proposed arrangement, A and B would derive their highest payoffs: 3 and 5, respectively. Yet even if C were to decline, A and B would continue to receive positive payoffs—1 and 3, respectively—and hence A and B would still be better off (albeit somewhat less so) than if B had scuttled the whole affair by declining A’s initial offer.25 Inasmuch as it would be rational for B to join irrespective of C’s decision, B may be said to possess a dominant strategy. Indeed, for B, the prudent course of action is to enter A’s preferred arrangement regardless of C’s decision. Now consider the decision-making calculus undertaken by actor C . However greatly C may dislike the new cooperative arrangement, the outcome that C prefers—a perpetuation of the status quo—has ceased to exist as a feasible alternative. The reversion point that C must take into account in reaching its decision is not SQ but SQ9, the outcome that would ensue if C were to opt out. Under this new status quo, C would obtain a payoff of 23. Henceforth, it is this payoff, rather than the status quo payoff of zero, that would constitute the appropriate baseline for comparison. The fact that C was happy with the previous status quo, SQ, no longer has any practical relevance. Instead, the choice that C faces is between two—and only two—alternatives: enduring the new status quo (i.e., settling for a payoff of 23) or joining the new regime (i.e., reducing its losses by accepting a payoff of 21). While C would be foolish not to enter the regime under these circumstances, its decision to join would in no way constitute evidence of collective-gains-seeking behavior. If C were allowed to choose between the noncooperative status quo, on the one hand, and the cooperative arrangement that emerges as the game’s equilibrium, on the other, it would gladly “settle” for the former. After all, a zero payoff, if obtainable, would be better than a loss of 1. C’s acquiescence is best viewed, then, as the outcome of a choice between two evils. Rather than seeking collective gains, C merely wishes to avoid a fate even more unpleasant than participating in A’s and B’s regime— namely, being excluded from it. IF A AND B BUILD IT, WILL C COME?
Nothing I have said thus far should be taken to mean that cooperation always looks like this. Whether it does or not will depend on a number of contextual considerations, the four most important of which are the following. 25 Implicit in the payoff matrix is an assumption that actor B enjoys a bargaining advantage over actor A—an advantage that B enjoys by virtue of A’s greater dissatisfaction with the situation that exists at the beginning of play—and that B uses this advantage to obtain a cooperative arrangement that produces greater gains (relative to the status quo) for itself than it does for A, the demandeur. In this way, the model takes into account the possibility that the gains from institutionalized cooperation may flow disproportionately toward those parties who were least unhappy with the noncooperative status quo and who could, as a result, hold out longest for their most-preferred outcomes. Cf. Hirschman 1945, and Krasner 1991.
46
W H Y
D O
N A T I O N S
C O O P E R A T E ?
Degree of Divergence in the Actors’ Underlying Preferences. Do actors in the system disagree about the desirability of the good that the cooperative arrangement under consideration is meant to provide? For the payoff structure depicted in the above matrix to correspond to empirical reality, the good in question must look “bad” to certain actors in the system. While this might not be true for all goods, coming up with examples of goods whose provision would benefit all actors in the system is harder than one might think (see below). Spillover Effects. Does provision of the good produce externalities for actors not participating in the cooperative arrangement and, if so, are these externalities negative or positive, large or small? For the argument to obtain, these externalities must be negative and large, which is why the losers are so eager to participate. That is not the only possibility, of course. Others include no externalities (private goods) and positive externalities (public goods); in addition, the spillover effects on outsiders—Kenneth Oye (1992) labels these “divertable” externalities—could be negative, but too small to induce entry. Holdup Potential. Are the actors who benefit from the good’s provision capable of supplying it by themselves? The argument above requires that the beneficiaries be able to provide the good on their own, that is, without any contribution from the losers. If, alternatively, the losers were able to form a blocking coalition or “counter-union,” the good would not be supplied at all. Expansion Potential. Are the beneficiaries of the good better off expanding “their” cooperative arrangement or keeping it exclusive? For my argument to apply, the prime movers cannot be worse off if the good they produce is consumed by additional actors. Alternatively, one can think of a “club good,” where the beneficiaries would prefer to keep other actors from enjoying the fruits of their cooperation (i.e., the privileges of membership) lest they dilute their own gains by spreading them too widely.26 Each of these variables is prior to, and therefore independent of, the particular cooperative arrangement or system that emerges, and there is no inherent reason why the values they assume cannot differ from good to good. Nor should one be surprised to find their values fluctuating over time as (for example) the domestic political priorities or capabilities of the actors in the system are altered for “exogenous” reasons. While the argument I have been making is a conditional one, however, it should be noted that none of the necessary conditions is especially restrictive or demanding. This is certainly true for the holdup and expansion variables. As for the spillover variable, Oye (1992, 17) contends that divertable externalities are already “pervasive 26
For further analysis of club goods, see Buchanan 1965.
W I N N E R S
A N D
L O S E R S
47
in international commercial and financial affairs.” Finally, cases in which actors disagree over the desirability of the “good” in question are also quite common, even when the good is something as seemingly noncontroversial as a coordinated reduction in nontariff trade restrictions or the stabilization of intraregional exchange rates.
Power Politics and the Dynamics of Exclusion The purpose of my discussion thus far has been to ask whether cooperative regional or multilateral arrangements necessarily leave their member states better off. My claim is that, in certain circumstances, some states may find it rational to take part in such arrangements even when they would greatly prefer the original, pre-cooperation status quo. What keeps these states— the losers—from opting out? The answer, put simply, is their fear of being left behind. What tips the balance for these participants—what motivates them to integrate their policies with those of the winners—is not the prospect of mutual gain so much as the absence of any better alternative. Loath to exclude themselves from the regional and multilateral regimes appearing around them, the losers decide it is in their own interests to cooperate. But they do so despite, not because of, these regimes’ utility consequences. Given this line of reasoning, we should not be at all surprised to find national governments “enthusiastically” joining institutional arrangements that leave them worse off than they would have been had those same institutions never been created. Nor would it be suprising to find governments continuing to participate in these institutions even after—perhaps long after—their deleterious consequences have become common knowledge. But that is not all. If the power-politics perspective elaborated here is on the right track, one would also expect to find the process of institutional enlargement taking on a life of its own. As it happens, the process by which international institutions take on new members is nearly always characterized by a “bandwagoning” dynamic rather than by a slower, more incremental evolution. The fact that this is so—that the enlargement process seems to naturally build momentum as it proceeds—is also, I would argue, firmly rooted in the logic of go-it-alone power.27 To see this more clearly, return to the extensive-form game presented in Figure 3-4 and consider what happens once C agrees to enter into a cooperative arrangement with A and B. By jumping on the bandwagon, C triggers another shift in the reversion point, thus raising the possibility that a fourth actor— 27 The contrast between balancing and bandwagoning behavior is developed in Waltz 1979, 125–28; Walt 1987; Christensen and Snyder 1990; G. Snyder 1991; and Martin 1992a, 41– 43. I follow Martin in using the term bandwagoning to refer to a process or dynamic. Other authors employ the term to indicate an outcome, e.g., the kind of alignment structure that occurs when—in violation of classical balance-of-power theory—weak states form alliances with their chief military rivals.
48
W H Y
D O
N A T I O N S
C O O P E R A T E ?
one who may be even more strongly averse to membership in the proposed regime—would also be better off coming on board. If involvement in the new regime were limited to A and B, this fourth actor might rationally refuse to join it. Now, however, it is confronted with the prospect of exclusion from an arrangement whose members either already or will soon include not just A and B, but A, B, and C. Moving along this evolutionary path-dependent trajectory, A and B’s capacity to remove the status quo ante from a third actor’s menu of alternatives could limit the strategic alternatives available to a fourth actor such that it too would be better off joining than remaining on the outside. Nor would this process cease with the fourth actor’s entry, as its decision to vest authority in the new regime would shift the status quo yet again, thus conceivably tipping the balance for a fifth actor, and so on. This simple dynamic is captured in Figure 3-5, a one-stage extension of the game tree presented in the preceding figure. In this second stage of the game, a fourth actor, D, must decide whether to submit to A and B’s preferred institutional arrangement, a decision it reaches by weighing the likelihood of participation by C and the attendant costs of exclusion (D’s payoffs are indicated by the fourth number enclosed in parentheses). As before, active support and participation from A and B are all that is required for the new regime to become established. C’s behavior has no bearing on the issue, as the regime could survive perfectly well even if C elected to remain an outsider. In another respect, however, C’s response is crucial. Suppose there is a fourth actor, D, with which C has a preexisting bilateral relationship. Consider D’s strategic options. In responding to the actions taken by A and B, D must take into account the expected utility of two different scenarios: (1) C stands firm or (2) C acquiesces. If C stands firm and refuses to join A and B’s regime, D would be wise to refuse entry as well. In that case, D would incur a loss of 1 relative to the status quo—a decrease in its utility, but not as severe a cost as the 2-unit decrease it would incur by joining in C’s absence. As we have seen, however, C’s ordering of preferences disposes it toward entry. Indeed, C would be inclined to join irrespective of D’s choice. If C stayed out, it could do no better than 23, whereas if it entered it could do no worse than 22, the payoff it would receive if both C and D—rather than C alone—were to enter. The bottom line is that joining the new regime is now a dominant strategy not just for B, but for C as well.28 28 Note that Figure 3-5 has C suffering a loss upon D’s entry. To examine the case in which C’s utility is enhanced by D’s participation, one could simply invert C’s preference ordering such that the cooperative stage-two equilibrium was now preferable to SQ0 (i.e., C would be better off if D took part in the proposed arrangement than if only A, B, and C did). While this modification would not alter D’s choice (see below), it does raise an interesting possibility: If each additional signatory reduces C’s losses by, say, one unit, then—assuming A and B are willing to admit several new members into “their” institution—C might well be able to recoup all of its losses relative to the status quo ante.
Figure 3-5. The Cooperation Game, Stage Two
50
W H Y
D O
N A T I O N S
C O O P E R A T E ?
That D would have been able to obtain a better (but still negative) payoff under SQ9 is quite beside the point. From the moment C acquires a dominant strategy, D must take as its point of reference SQ0; that is, it must assume that membership in the new institutional arrangement will include at least A, B, and C. Given this certainty, D faces mounting pressure to settle for a cooperation payoff of 23, thus avoiding the even lower exclusion payoff of 24. Although the interaction depicted in the figure involves only four actors, it is easy to see how additional actors might get swept up in the process as D’s entry into A and B’s preferred institutional structure triggers yet another shift in the status quo. With each such move—from SQ to SQ9, from SQ9 to SQ0, from SQ0 to SQ-, and so on—the “attractions” of regime membership become stronger. They become so strong, in fact, that actors whose political and economic priorities are wholly different from A’s or B’s (or for that matter C’s or D’s) may conclude that they too are better off joining the regime in question. While the resulting cooperation would likely further the interests of the actors who set the bandwagon in motion (in this case A and B), other signatories—those whose chief motivation in joining was the fear of being left behind—could end up worse off than if SQ, the pre-cooperation status quo, had never been altered.
Meeting Some Objections Determining the extent to which the model of cooperation presented here provides an accurate accounting of recent developments is no easy task. As noted earlier, much of the challenge stems from the methodological difficulties involved in counterfactually deducing what the reaction of the losers would have been if the winners had been unable, or perhaps just unwilling, to proceed on their own.29 Go-it-alone power is usefully conceptualized as a two-step process. In step one, depicted by the arrow in Figure 3-3 pointing from SQ to SQ9, the Y’s commit themselves to establishing a new cooperative regime. In step two, the action shifts to the X’s, who request (and are granted) admission into this new arrangement. These actors do not view their accession as a move toward the Pareto frontier. Quite the contrary, if they could abolish the regime and restore the original status quo, they would gladly do so. From the standpoint of the new entrants—the X’s—accession is simply the lesser of evils. The utility consequences of accession deserve close scrutiny at this point. For while my power-politics model highlights the downside potential of goit-alone power, there is no reason why the final outcome must fall within the upper-left-hand quadrant. Perhaps by jumping on the cooperation band-
29
On the prevalence of counterfactuals in theory testing, see Fearon 1991.
W I N N E R S
A N D
L O S E R S
51
Figure 3-6. Go-It-Alone Power as a Bargaining Tactic
wagon, X, the would-be loser, more than compensates for the negative externalities of Y’s initial move. Go-it-alone power is, in this view, a fairly innocuous bargaining device, one that the Y’s deploy in order to obtain a cooperative equilibrium as close as possible to their ideal point (Figure 3-6). As in previous examples, the construction of the new arrangement benefits Y, the prime mover; hence, all the action takes place above the horizontal axis. Note, however, that the cooperative arrangement that is finally established represents a utility gain for the “losers” as well. Were it not for the winner’s initial power play, the transition to this Pareto-improving equilibrium would not have been possible at all. Figure 3-6 thus depicts a second, more circuitous path to the distributional outcomes analyzed by Krasner, Grieco, and others (cf. Figure 2-3). As is the case for go-it-alone power, the winner’s unilateral action creates a negative externality; the establishment of a new reversion point—SQ9—diminishes the loser’s utility relative to the initial status quo. In this case, however, the externality is completely eliminated, not just reduced or mitigated. Once the
52
W H Y
D O
N A T I O N S
C O O P E R A T E ?
prospective loser cedes the demands of the winners, the outcome moves from the upper-left-hand to the upper-right-hand quadrant, and all participants benefit. Thus, one can make explicit use of go-it-alone logic and yet still view interstate cooperation in positive-sum terms. Consider the oft-heard claim that Britain would never have agreed to the 1986 Single European Act (SEA) but for its fear of being denied a voice in collective decisions regarding the future of Europe. Even if this is correct, it does not mean that the SEA necessarily left British authorities worse off in terms of the pre-SEA status quo ante. After careful scrutiny, for example, Andrew Moravcsik concludes that the institutional reforms embodied in the SEA actually worked to the advantage of the British, whose demands were satisfied on matters ranging from voting procedures (qualified majority on “internal market” decisions only) and economic principles (mutual recognition) to the scope of the agreement itself (nonincorporation of the Social Charter). In opting out of the agreement, Moravcsik writes, the Thatcher government would have been forgoing “substantive gains.”30 From this perspective, the exclusionary threats issued by Britain’s European partners were a welcome development, since the country’s Conservative government would otherwise have continued to hold out for an even better (i.e., even more market-oriented) agreement, the result being no agreement—and hence no “substantive” benefits— at all.31 To be sure, in his book’s theoretical discussion of bargaining Moravcsik allows for the possibility that an EU member like Britain, when confronted with an exclusionary threat, could end up signing an agreement that made it absolutely worse off.32 Moravcsik seems to think that this line of argument is consistent with, and indeed can be subsumed under, his “liberal intergovernmental institutionalist” model of state behavior. Yet what most clearly distinguishes that model from other approaches—and on this Moravcsik is quite explicit—is the causal weight it ascribes to state preferences. What accounts for the variation we observe in the willingness of states, in Europe and elsewhere, to integrate their policies? The answer given by intergovernmental institutionalists is straightforward: States integrate when— but only when—the individuals who preside over their governments believe that closer cooperation will advance their underlying interests. As intergovernmental institutionalism’s leading proponent, Moravcsik is theoretically 30
Moravcsik 1998, 376. Was the British government’s fear of exclusion really the decisive factor in its (ostensibly) Pareto-improving accession to the internal market? The jury, it seems fair to say, is still out. On one interpretation, British authorities not only did not feel threatened with exclusion; they actually drove the negotiations. That is, the UK’s Conservative government skillfully used its lack of enthusiasm toward integration—and resulting tolerance for the pre-SEA status quo—to extract concessions from its European partners. This line of analysis is consistent with Moravcsik’s assessment that the British were even more satisfied with the negotiated outcome than were the French and the Germans. See esp. Moravcsik 1991, 49. 32 Moravcsik 1998, 61– 64. See also Moravcsik 1991, 26. 31
W I N N E R S
A N D
L O S E R S
53
committed to defending this view. For if the forward progress of European integration were not congruent with the underlying interests of each EU member government—if, to the contrary, the process of European integration prevented the leaders of some EU states from achieving the ideological, economic, or political goals they had originally set for themselves—it would be hard to argue that the underlying preferences of ruling governmental coalitions were in any sense fundamental, let alone decisive. Empirically, then, Moravcsik has a strong stake in showing that the British authorities who signed the SEA, though faced with the prospect of exclusion, nonetheless succeeded in realizing their distinctive (market-oriented) preferences. But how confident can we be that he is right? In keeping with his liberal orientation, Moravcsik would have us situate the threat-induced SEA within the upper-right-hand quadrant of any utility mapping. But for all of the reasons discussed earlier, we cannot simply assume that the establishment of an institutional threat point will translate every time into a Pareto-improving equilibrium. Perhaps the level of economic integration necessitated by the SEA-mandated internal market so far exceeded Europe’s “lowest common denominator” (in Moravcsik’s terminology) that some government signatories—including the British Conservatives—would actually have preferred the original, pre-SEA status quo. Perhaps, in other words, the internal market equilibrium properly belongs in the upper-left-hand quadrant, with the utilities of the equilibrium’s Franco-German beneficiaries arrayed along the vertical axis, and those of the British government (and of any other absolute losers) along the horizontal axis. Unfortunately, the task of distinguishing positive-sum from non-positivesum equilibria, though straightforward in theory, may be quite difficult in practice. Before an outside observer can say definitively whether a particular member government is better or worse off because of its involvement in a particular institution, he or she must know how well off that government would have been if, counterfactually, that institution had never been made it into existence. To perform this sort of welfare comparison, however, one would first need to know (among other things) what that government’s leaders had originally set out to accomplish during their tenure in office—and the task of specifying the “underlying” preferences of government actors is rarely a simple matter. But that does not mean it is impossible. The temptation, of course, is to rely on what government leaders tell us: If they declare themselves to be better off, why is that not sufficient evidence of the salutary effects of membership? There are many reasons one should be wary of taking such declarations at face value, and not just because the rhetoric of political discourse often fails to mirror the reality. A government may want desperately, and understandably, to participate in a cooperative arrangement from which it is in danger of being excluded, even if that same government (and its domestic constituencies) would prefer that the arrangement in question had never been established. If there is a lesson here, it is that public outpourings of sup-
54
W H Y
D O
N A T I O N S
C O O P E R A T E ?
port from a cooperative institution’s leading enthusiasts can be both genuine and deeply misleading at the same time. But surely, it might be argued, if a member government complains bitterly about its institutional affiliations, it cannot possibly be better off. Even here, however, one must be careful lest a government’s public pronouncements be mistaken for its true beliefs. Why might a winner find it advantageous to portray itself as a loser? One possibility is that its misrepresentation is simply a ploy to strengthen its initial bargaining position. If a winner can persuade its negotiating partners that it is actually a loser (i.e., would prefer the status quo ante), then, assuming its partners are incapable of going their own way, they are more likely to offer it a better bargain, modifying institutional structures or piling on concessions so as to tip the balance in the winner’s favor. By this logic, all parties to the negotiations—winners and losers alike—should downplay the benefits they expect to derive from any proposed agreement, exaggerating the strength of domestic opposition to particular provisions, if not to the agreement as a whole. This is nothing more than basic bargaining strategy.33 Note, however, that if the incentives to dissemble were strictly for bargaining purposes, the advantage to be gained in falsely misrepresenting oneself as a loser would disappear as soon as a deal were struck. Once the terms of the agreement had been fleshed out and the new institutional arrangement entered into force, the logical move would be for each participant to proclaim victory, taking credit for the hard-won “concessions” it had extracted from its fellow signatories during prelaunch negotiations. And yet government leaders often speak derisively about institutions in which they have been participating for years. Why might government leaders be inclined to exaggerate the true costs of their current institutional affiliations? The reason, say skeptics, is to deflect responsibility for policies they privately desire—and, indeed, were planning to pursue anyway—but for which they fear there may be little domestic support. In this view, international institutions are actually political assets, for though they do not dictate policy, their detractors can use them politically as scapegoats. The European Monetary System was used in precisely this way, the economist Paul Krugman (1994, 185) has argued, as justification for “the 33 Schelling 1960. Whether and how internal political constraints may affect the outcome of international negotiations is a central question in the literature on “two-level games” (Putnam 1988; see also Evans, Jacobson, and Putnam 1993 and Milner 1997). Of primary interest since the publication of Putnam’s study has been the issue of credibility: Government negotiators might swear up and down that their hands are tied by domestic pressures “back at home,” and yet, if this were not actually the case, wouldn’t their negotiating partners know it? One solution would be for governments to deliberately tie their own hands. Thus, domestic leaders could go before their home constituencies to declare not only that they opposed the terms of an international agreement under consideration, but that to sign the agreement in its current form would be an act of gross incompetence. On how public officials increase their bargaining leverage internationally by deliberately incurring “audience costs” domestically, see Martin 1993 and Fearon 1994.
W I N N E R S
A N D
L O S E R S
55
swallowing of unpleasant economic medicine.” The reality, according to Krugman, is that European governments would have consumed this medicine—raising interest rates and fighting inflation to keep the value of their currencies from falling against the German D-mark—even if the EMS had not “forced” them into it.34 But rather than take the political heat for the ensuing unemployment, government officials in countries like Italy and France sought to pin the blame on the exchange-rate regime. A similar story is sometimes told about the International Monetary Fund: Government leaders are said to invoke IMF conditionality as a way of justifying the hardship imposed by economic policies they would have had to pursue anyway.35 And like the IMF, the Group of Seven (G-7) and its offshoots are said to afford similar opportunities for blame avoidance, a fact that may account for international summitry’s enduring appeal among political elites in countries around the world.36 Yet while all public officials have an interest in deflecting blame for unpopular policies, the institutions-as-scapegoats thesis goes one step further. If proponents of the thesis are correct, these officials do not just try to deceive their constituents; they actually succeed! This is a most curious claim, for the notion that citizens are somehow tricked into thinking the institutions are responsible for their hardship presumes that they are not clever enough to realize what their leaders are doing—namely, ducking responsibility for the unpleasant side effects of their preferred policies. If European members of the EMS had been trying, as Krugman (1994, 185) says, to “[cloak] a policy of high unemployment . . . under a less brutal-seeming label,” their domestic constituents surely would have caught on at some point and, just as surely, would have penalized them for it. Not only does Krugman offer no reason for supposing otherwise, but his depiction of government supporters as dupes sits uncomfortably with a vast body of empirical research demonstrating that citizens, like markets, are not easily fooled.37 34 There is little reason for accepting this claim. If the EMS had not been established, it is highly unlikely that inflation fighting would have been the “dominant priority” (Krugman 1994, 187) of the center-left and socialist governments that held office in Italy and France during the 1980s. Chapters 8 and 9 offer an in-depth discussion of government priorities in both countries, as well as in Germany and the United Kingdom. 35 E.g., Kahler 1992. 36 Cf. Putnam and Bayne 1986. Consider the situation facing Helmut Schmidt, the host of the 1978 gathering of G-7 leaders in Bonn, West Germany. As the standard-bearer for the German Social Democratic party, Schmidt was naturally inclined toward growth-oriented policies benefiting German workers. Nevertheless, he was also aware that his preferred budgetary stimulus would impose a substantial financial burden on his country’s famously inflation-averse business and banking sectors. The timing of the Bonn Summit was, in this sense, quite fortuitous, for it allowed the chancellor to cast his stimulus package as the product of unrelenting outside (especially American) pressure, when in fact it was a straightforward act of redistribution. In this view, “the chancellor ‘let himself be pushed’ into a policy he privately favored.” Putnam 1988, 429. 37 On the difficulty of putting the wool over the public’s eyes, see, among many others, Popkin 1994.
56
W H Y
D O
N A T I O N S
C O O P E R A T E ?
There is, then, at least some basis for thinking that government leaders who complain of being “forced” to pursue policies that leave them worse off may actually be telling the truth. Even if complaints of this sort are not always justified, we would be unwise to dismiss them out of hand. To assume that governments criticize international institutions solely for the purpose of scoring political points is to suggest that their constituents are incapable of seeing through their leaders’ self-serving attempts at obfuscation. But the justification for taking the losers’ protestations seriously is not just that the microfoundational underpinnings of the institutions-as-scapegoats thesis are murky, to put it mildly. One can also question its implicit assumption that the protesting governments’ institutional commitments do not really matter—that from a practical (as opposed to a public relations) standpoint the institutions at issue are irrelevant. Will this always be the case? Perhaps the institutions being complained about really do (1) limit their member governments’ policy options and (2) do so in ways that are fundamentally at odds with the policy priorities of the governments doing the complaining. What are these priorities? While there is no simple method for answering this question, one can safely assume that the partisan attachments, ideological beliefs, and electoral interests of the individuals who hold office at a particular point in time will all play a role in the “formation” of their underlying preferences.38 Identifying the ideological and partisan loyalties of government officials is often straightforward. The difficulty comes in ascertaining their electoral interests. All too often, scholars interested in deriving an objective measure of “state” welfare bias their analyses toward long-run, national-welfare considerations. True, a government that presides over a steadily deteriorating economy is a government in deep political trouble.39 But while governing elites cannot afford to completely disregard the long-term effects of their actions on society as a whole, incumbent politicians are likely to be especially attuned to the short-term interests of their “core” constituencies.40 Whatever their interest in increasing the long-term welfare of all citizens, democratically elected officials can retain office only so long as they successfully respond to the demands of those segments of the wider public upon whom they rely for electoral and financial support. Not only will these individuals and groups often have short time horizons, but they are also likely to be pre38
Frieden and Rogowski 1996; Ikenberry, Lake, and Mastanduno 1988; Milner 1997; Moravcsik 1998; Simmons 1994. 39 On the relationship between aggregate economic performance and electoral success in the advanced industrialized democracies, see Kramer 1971; Hibbs 1987; Tufte 1978; Kiewiet and Rivers 1984; and Alesina and Rosenthal 1995. That such a powerful relationship should exist is not surprising: in no stable democracy has the traditional constituency of either the left or the right ever been large enough to guarantee reelection on the basis of “pure” partisan strategies. On this last point, see Przeworski and Sprague 1986. 40 See, e.g., Alesina 1989; Garrett and Lange 1995; Hibbs 1987.
W I N N E R S
A N D
L O S E R S
57
occupied with increasing their individual shares of society’s aggregate wealth and resources.41 Much the same is true for authoritarian societies: Authoritarian rulers are likely to put the interests of certain societal actors—namely, those with the potential to stage palace coups, organize grass-roots revolutions, or otherwise jeopardize ruling elites’ personal security—above the interests of the country as a whole.42
Conclusion It follows that to accurately assess the utility consequences of a country’s participation in a given agreement or institution, we may first need to learn something (perhaps a great deal) about what motivates that country’s public officials. Who, in particular, are their most influential constituencies? And what, exactly, do these groups and individuals—be they corporate managers and directors, small business owners, labor unions, farmers, or residents of particular geographic regions—want and expect from “their” government? Only rarely can governing elites afford to sacrifice the immediate material well-being of their core constituencies for the sake of longer-run or more diffuse societal gains. If these constituencies suffer costs as a result of their government’s involvement in an international accord, pact, treaty, or organization—whether because of high transition costs, greater susceptibility to macroeconomic shocks, or other reasons—it is reasonable to assume that the officeholders who represent them will suffer as well. 41 42
Olson 1982. Bates 1981; Linz 1975.
P A R T
I I
How Do Nations Cooperate?
C H A P T E R
4
The Efficiency Rationale for Supranational Governance IF THE MODEL of cooperation I have just set forth has any merit, we should not be surprised to find cooperative multilateral arrangements imposing absolute utility losses on some member states—even ones whose leaders are outwardly enthusiastic about joining. None of this would be particularly controversial were it not for the mainstream literature’s continuing preoccupation with the mutual-gains side of international cooperation. If, however, we were to shift the theoretical focus to the losers’ side of the story, I am confident we would quickly arrive at a more subtle—and ultimately more convincing—answer to the question “why do states cooperate?” Having argued as much, I now want to suggest how the winners-andlosers approach I have been developing might be used to shed light on the related, albeit analytically distinct, question of how states cooperate. As noted in the introduction, many of today’s international agreements look fundamentally different from those we are used to seeing. It is not simply that, as Robert Lawrence (1996) and others have pointed out, they impinge more directly upon matters of domestic policy, their terms covering everything from environmental and worker safety standards to basic questions of fiscal, monetary, and industrial policy. At least as important is the fact that many, if not most, of the interstate agreements signed in recent years have enhanced the power and prerogatives of supranational institutions. What are supranational institutions? As the literature has not yet settled upon a precise definition, let me offer one of my own. For the purposes of this book, supranational regimes can be understood as having two defining features. First, the substantive provisions of their underlying agreements— the basic terms of cooperation embodied in the founding charters of the WTO, the EMS, and NAFTA, for example—are considerably more openended than those found in ordinary agreements, the provisions of which may have been fully specified (inasmuch as this is possible) during the initial round of negotiations. The second defining attribute of a supranational regime has to do with procedural matters. Rather than leaving it up to each individual signatory to determine how the terms of cooperation apply in a given situation, the parties to a supranational regime agree ahead of time to render these determinations in accordance with a well defined—and prespecified—collective decision-making procedure. It is this procedure that the regime’s member states adhere to as they go about the task of “filling in” their agreement, interpreting its (broadly worded) provisions in light of specific cases and unforeseen contingencies.
62
H O W
D O
N A T I O N S
C O O P E R A T E ?
Not all supranational regimes adopt the same procedure, of course. Some require all member states (or their representatives) to enter directly into collective deliberations; others delegate the task of resolving disputes to an impartial third party.1 In the larger scheme of things, however, these differences—though certainly worth exploring—are of secondary importance. It is not the fact that supranational governance structures exhibit variation that is puzzling; it’s the fact that they exist at all. For every time a party to a supranational agreement submits to a higher-level decision-making process, it is forfeiting the right to decide for itself (i.e., unilaterally) how that agreement is to be interpreted and applied over time. Why would a “sovereign” state— or, more precisely, the collection of individuals who exercise governmental authority within it—ever be willing to relinquish that prerogative? This part of the book proposes an answer that, while firmly grounded in rational choice foundations, nonetheless differs quite markedly from the literature’s traditional way of thinking about how nations cooperate. In this chapter, I take a detailed look at what international relations scholarship has had to say on the subject of institutional design, and on the “politics of structural choice” (Moe 1990b) more generally. Then, in chapter 5, I put forward my own perspective, one that flows directly out of the theoretical logic I began fleshing out in the first part of the book. While students of international relations have tended to emphasize the collective efficiency rationale for supranational governance, delegated authority, and the like, the analysis presented in chapter 5 raises an altogether different possibility: that the actors responsible for setting the cooperation bandwagon in motion—the enactors—may be less concerned that “their” regime maximize the collective good than that it be able to survive once they, its (sole?) beneficiaries, no longer hold power domestically. After developing this line of argument, I lay out some broad hypotheses to be examined in subsequent chapters.
Supranational Institutions as Vehicles for Collective Action In recent years, the stature and assertiveness of international institutions have increased to the point where supranational authorities now perform many decision-making tasks once undertaken directly and exclusively by individual states. This shift has been both sudden and dramatic. Throughout much of the postwar period—and indeed throughout much of recorded history—supranational entities were afforded little or no independent role in the formulation of state policy.2 Starting in the late 1970s, however, the dis1 The question of what may account for this variation is taken up below, first at a theoretical level, here and in chapter 5, and then empirically, with explicit reference to North American trade and European money, in chapters 6 and 9 respectively. For other recent work in this line of research, see Abbott and Snidal 1998; Kahler 1995; Yarbrough and Yarbrough 1990 and 1992; and the contributions to Koremenos, Lipson, and Snidal 1999. 2 Jackson 1989; P. Taylor 1983.
T H E
E F F I C I E N C Y
R A T I O N A L E
63
cretionary powers enjoyed by supranational institutions began to increase beyond anything seen in previous years. The result has been a marked trend toward the collective management and supervision of interstate relations. Already, according to some accounts, the “anarchy” that once characterized the international system has given way to a complex web of overlapping supranational networks and a nascent “community of nations.”3 Even if this is overstated—we are still a long way from anything remotely resembling a World Parliament—few political analysts, let alone diplomats or public officials, would dispute that a fundamental change is underway. And unlike the “exogenous” constraints on state action imposed by the integration of world financial markets (where a country’s leaders can plausibly claim to be swimming in currents whose direction and strength they do not control), the constraints imposed by the proliferation of supranational decision-making structures are a product of political choice. Each time government officials agree to abide by the determinations of a higher international body, they are making a conscious decision to give up direct control over the policy-making process. Why would they ever be willing to do this? Having long been interested in the nexus between institutions and cooperation, Robert Keohane and other so-called neoliberal institutionalists believe they are well equipped to explain why supranational institutions have emerged as today’s preferred instrument for managing interstate relations.4 In their view—one neoliberals share with other proponents of the “new” institutionalism5 —institutional choice is ultimately guided by efficiency considerations, with groups of actors (here, states or governments) struggling to choose whichever institutional forms will enable them most effectively to respond to market failures, mitigate collective-action problems, and generally further their common interests. This line of analysis begins with the observation that certain policy objectives—even seemingly “domestic” ones like generating economic prosperity—are difficult for nations to achieve on their own. From this it is but a short step to the conclusion that, by acting in concert, national governments can significantly improve their collective welfare; if all make the necessary behavioral adjustments, all benefit. True, these behavioral adjustments might not look so beneficial to outsiders. But for those on the inside—that is, for the cooperators themselves—the move to a coordinated outcome would represent a clear Pareto improvement, leaving each of them at least as well off as it was under the previously existing (noncooperative) status quo. So far, so good. The problems come when cooperation fails to emerge 3
See, e.g., Lebow 1994. Proponents of realist theory make no such claims. As noted in chapter 2, realists spent many years insisting that international institutions had no appreciable impact on the prospects for cooperation among nations or, for that matter, on international politics generally. 5 March and Olsen 1990. 4
64
H O W
D O
N A T I O N S
C O O P E R A T E ?
Figure 4-1. The Prisoners’ Dilemma
spontaneously. To explain why the prospect of collective gain may not be sufficient to induce the kinds of mutually beneficial policy adjustments that would leave everyone better off, theorists invoke the well-known “prisoners’ dilemma.” Two suspects imprisoned by the police—both implicated in the same crime (and both guilty)—must decide whether to lighten their sentences by testifying against each other. Interrogated separately and therefore unable to communicate, the prisoners are told that if they both defect (i.e., turn state’s evidence), both will be sentenced to three years behind bars. If instead they cooperate (i.e., neither prisoner “rats out” the other under questioning), the authorities will be forced to convict them on a lesser charge, punishable by only one year in jail. By cooperating, then, the prisoners can avoid the longer sentence—read: lower payoff—they would receive were they both to defect (contrast the 5–5 and 0–0 payoffs depicted in Figure 41). On the other hand, if only one of the prisoners defects, that prisoner will immediately be set free (thereby obtaining the game’s highest possible individual payoff), while his cooperating partner—the “sucker”—will have his sentence increased to ten years (the game’s lowest payoff). Now imagine that you are one of the two prisoners. If you believe your partner will betray you, you are surely better off betraying him as well; that way you avoid the possibility of a ten-year sentence, that is, the sucker payoff. In fact, it would make sense for you to give evidence against your partner even if you thought
T H E
E F F I C I E N C Y
R A T I O N A L E
65
he was likely to keep mum, for in that case you would secure your immediate release, as opposed to having to spend a year in jail. Assuming your partner reasons as you do, he too will defect “no matter what,” thus guaranteeing that the two of you end up spending three years behind bars. And yet, had you both cooperated, you and your partner could have obtained a lighter (one-year) sentence. Hence the dilemma: Your individually optimal strategies lead you to a collectively suboptimal outcome. By precisely the same logic, the fear of being “suckered” can make it difficult for a group of individually rational states (or governments) to coordinate their policies even when doing so would leave each member of the group unambiguously better off.6 Logic suggests that such fears would be particularly intense in cases where, for cooperation to succeed, one of the group members would have to make the first move, altering its policies before its partners were called upon to do so. All things equal, one would also expect to find nations behaving in a collectively suboptimal fashion when the cooperative outcome required some parties to invest in specialized, or “dedicated,” assets that could not easily be converted to other uses if their partners were later to renege. Consider an example from trade: Even if two neighboring countries could improve their collective welfare by establishing a free trade zone, the leaders of each country might nonetheless be reluctant to take the first step since, were their counterparts to reintroduce trade restrictions ex post, any private firms that had arisen exclusively to service the other country’s market would be forced to make the costly switch to other markets or sources of supply. Though the owners, managers, and workers employed in these firms would presumably be hardest hit, the rise in transitional unemployment caused by the unilateral defector’s opportunism could also have distinctly political costs for the leaders of the exploited partner— costs these leaders could avoid by refusing to liberalize in the first place. To be sure, the officials who anticipated being exploited could threaten to retaliate against their partners should the envisaged cooperative gains fail to materialize. In principle, then, the expectation of future reprisals and loss of reputation could in itself be sufficient to keep everyone in line.7 That’s the good news. The bad news is that this “Axelrodian” path to cooperation can only work if each partner is able to distinguish the opportunistic behavior it seeks to deter from the cooperative behavior it wishes to encourage. This is where formal institutions enter the story.
The Information Argument Of all those institutional functions that neoliberal theorists have identified as having a salutary effect on the prospects for cooperation, those involving 6 7
See, e.g., Conybeare 1984; Gowa 1986; and Lipson 1984. Axelrod 1984. See also M. Taylor 1987.
66
H O W
D O
N A T I O N S
C O O P E R A T E ?
the collection and distribution of information are considered to be the most important.8 Why? Because it is only when actors are able to distinguish cheaters from cooperators that they can be expected to appropriately dole out punishments (to opportunists) and rewards (to fellow cooperators) in subsequent rounds of the “game.”9 It is fair to say that Axelrod initially underestimated the importance of these informational requirements. Thus, after conceding that “recognition and recall” are both critical to the success of collective action, his book The Evolution of Cooperation quickly dismisses their practical significance, noting that the informational demands of strategies like tit-for-tat, which “respond only to the recent behavior of the other player,” are so limited that even bacteria can fulfill them. “And if bacteria can play games,” writes Axelrod (1984, 174), “so can people and nations.” But while its informational requirements may be low compared to those of other strategies, even tit-fortat requires the players who deploy it to be able to identify whether their 8
See particularly Keohane 1984, chap. 6. Collecting and distributing information are not, however, the only tasks that institutions perform. Perhaps the best-known proposition to emerge from the literature spawned by Keohane’s work is that international institutions increase the frequency of diplomatic interactions (see, e.g., Axelrod and Keohane 1985, 248 – 50). There are problems with this line of analysis, however. It’s true that if the participants in a collective endeavor doubt that they will meet again, the threat of retaliation will not be very useful as a deterrent to free riding. For this reason, the expectation of future interactions may well be essential for the maintenance of cooperation along the lines suggested by Axelrod and others (cf. Kreps et al. 1982; Kreps and Wilson 1982). On the other hand, it is hard to see why international institutions would be necessary for enhancing the iterativeness of international exchange. While institutional rules frequently stipulate that state leaders meet at regular intervals, states would surely continue to interact even if their leaders were not required to engage in the kinds of institutionalized, face-to-face encounters that make front-page news. Another standard refrain in the regimes literature is that institutions promote cooperation by facilitating cross-issue linkages (e.g., Tollison and Willett 1979). The greater the dimensionality of a cooperative relationship, the greater will be each side’s capacity to reward compliance and penalize noncompliance. But while it is surely better to have more avenues for making threats or offering bribes than fewer, the people who routinely invoke this argument usually have something more in mind. What they claim is that issue linkage is itself largely a consequence of institutions—that “without international regimes linking clusters of issues to one another” (Keohane 1984, 91), the task of negotiating side-payments is simply too difficult for groups of states to manage effectively on their own. But is this really true? The most enduring postwar example of the linkage concept is the EC’s bundling of the Common Market, a customs union for trade in industrial goods designed to appease German manufacturers, with the Common Agricultural Policy (CAP), an agricultural price support scheme that has worked primarily to the advantage of French farmers (see, e.g., Sandholtz and Zysman 1989). Few would deny that the striking of this deal during the 1950s improved the credibility of each side’s retaliatory threats. If Germany were to renege on a promise to cooperate with France in another policy area, France now had the option of excluding German manufactured goods from its market. Conversely, if France failed to deliver in a Franco-German cooperative venture, Germany could choose to withhold its contributions to the CAP. But France and Germany might well have struck the same deal outside the context of the EC, thus raising the question of whether the EC’s role in the linkage was really all that significant. 9
T H E
E F F I C I E N C Y
R A T I O N A L E
67
partners have cooperated or defected in their most recent move of the game. This might not be difficult for bacteria, but for people and nations it can be quite a complicated matter. In fact, virtually all transactions between human agents entail some degree of privately held information. That said, imperfect (or asymmetric) information is likely to pose a greater impediment in some situations than in others. Thus, for example, it is more likely to pose a problem if the actors involved in the transactions at issue have failed to go beyond strictly informal agreements or understandings. Unless acceptable standards of behavior have been clearly defined at the outset, it may be difficult for each participant in a reciprocal relationship to determine at any given point in time whether its transaction partners are complying with the strategies they are meant to be following.10 Or take the case where the realization of collective gains would require the compliance of a large number of diverse actors. In multiplayer games, there is always a chance that one player will misinterpret defection by a second player as cheating when in fact the second player is merely retaliating against a third player for committing an unwarranted defection in a previous round.11 Translating to the world of international relations, this suggests that multilateral cooperation efforts could come to a grinding halt even— depressingly enough—when all of the nations involved are faithfully and consistently adhering to their assigned strategies.12 And similar difficulties could arise in a strictly bilateral interaction. Suppose, for example, that effective control over the policy realm in question were to change hands within one of the parties to a bilateral agreement at some point after the “game” has commenced. In that event, the other party might find itself lacking vital information about how its new transaction partner had behaved in its prior dealings, and thus whether that new partner was likely to prove as reliable a cooperator as its domestic predecessors had been before their ouster.13 In “noisy” environments like these, neoliberals suggest that international institutions can play a useful role in formalizing the initial terms of cooperation, monitoring subsequent behavior, and efficiently transmitting information about each party’s past and present records of compliance. Once it is clear that opportunists will be detected and their identities revealed, the tit-for-tat logic of decentralized enforcement then takes over. Rather than having the institution punish the defectors directly, each individual partner, acting in its own self-interest, takes that task upon itself, thus securing the 10
Lipson (1991) takes the opposite view, arguing that formalization impedes rather than promotes cooperation. 11 Bendor 1987; M. Taylor 1987. 12 Cf. Oye 1986b, 18 –20. 13 Downs and Rocke (1995) discuss some of the strategies that mutual-gains-seeking cooperators might use to surmount this particular source of uncertainty. The next chapter offers further analysis of the relationship between domestic politics and the formation of international institutions, a topic that, as Downs and Rocke correctly note, does not fit comfortably within the unitary-actor framework of most international relations theory.
68
H O W
D O
N A T I O N S
C O O P E R A T E ?
fruits of long-term cooperation in highly uncertain situations where collective action would not normally be possible. The upshot of this line of analysis is that where international institutions exist and operate as intended— clarifying expectations via written charters, accords, pacts, and the like; keeping each member state apprised of how its partners are behaving; and should there be an unwarranted defection, clearing up any ambiguity about the identity of the true culprit(s)—international cooperation may not in fact be so difficult to achieve (or to sustain) after all.14 At first blush, this perspective would seem sufficient to account for the institutional developments to which I referred earlier. Delve beneath the surface, however, and the standard account quickly runs into problems. For even if one accepts the basic thrust of neoliberal institutionalist theory—that formal institutions facilitate collective gains by helping states overcome obstacles to cooperation—the powers delegated to today’s international institutions often go well beyond what many of the theory’s original proponents had in mind. In fact, neither Keohane nor his followers ever developed anything that can be called a theory of delegation. Nor, when one considers their objective, is this especially surprising. In a sense, neoliberals have always been more interested in doing just the opposite—that is, in demonstrating that international institutions could significantly improve the prospects for interstate cooperation despite their lacking the kinds of discretionary powers enjoyed by domestic-level institutions. Indeed, for neoliberals to attribute such powers to international institutions would have been to suggest that the interstate system was not truly anarchic—a position that, following Keohane, these theorists had explicitly disavowed (partly, perhaps, out of deference to their tough-minded realist “opponents,” who were likely to dismiss it out of hand). In any event, most neoliberal scholars took the view that institutions could promote cooperation without doing much more than expediting the flow of information. Prospective cooperation partners needed some assurance that major defections would be detected and the identity of their perpetrators exposed. But as long as appropriate surveillance structures were established to meet these (minimal) requirements, it would be in everyone’s own interests to cooperate. For now, any transaction partner who violated the rules would face a considerable risk of retaliation, and thus of ending up worse off than if it had chosen to obey these rules (i.e., to cooperate) all along.15 In this view, the vesting of authority in supranational bodies and con14 Along these lines, a provocative study by Milgrom, North, and Weingast (1990) suggests that purely informational mechanisms may be able to sustain cooperation even in the extreme case where none of the players has ever encountered its current partner before or expects to do so again in the future. 15 The idea that international cooperation ultimately requires little more than the collection, processing, and transfer of information may be one reason for neoliberal theory’s enduring appeal among American academics and policymakers. For a related argument, see Mearsheimer 1994/95, 47–49.
T H E
E F F I C I E N C Y
R A T I O N A L E
69
comitant loss of national autonomy are wholly unnecessary. The fact is, however, that very few of today’s international institutions operate merely as collectors and transmitters of information. In addition to serving as watchdogs, passively monitoring compliance with whatever rules of the game their member states have agreed to uphold, most of these institutions also empower supranational entities of one kind or another to modify these rules—or to clarify their “true” meaning—once they have taken effect. In some cases, it is only a slight exaggeration to say that these entities dictate what their members can and cannot do. At first, this trend toward supranational governance was primarily a regional phenomenon.16 In recent years, however, the trend has begun to encompass a number of global regimes as well. The General Agreement on Tariffs and Trade is just one recent example. Since the Uruguay Round was “completed” in 1994, virtually every major trading nation in the world has committed itself to lowering tariffs and eliminating quantitative trade restrictions.17 The 1994 treaty also went a considerable distance toward removing nontariff impediments to trade, such as those resulting from differing national health and safety standards and lax enforcement of intellectual property laws. While these were important departures, however, the new supranational decision-making structure established by the treaty is surely its most remarkable achievement.18 The institutional details of the GATT’s new World Trade Organization (WTO) have been well described elsewhere.19 In addition to serving as a forum for regularized biennial meetings of top trade officials, the WTO also provides for a new secretariat, director general, and other institutional trappings originally intended for the GATT but delayed indefinitely when the U.S. Congress refused to support the creation of the so-called International Trade Organization (ITO) in the late 1940s. Most importantly, though, the WTO is entrusted with the task of adjudicating trade disputes among its member states.20 While the pre-Uruguay GATT provided similar mecha16
Two excellent analyses are those of Haggard (1997) and Kahler (1995). Formally signed in April 1994, the Uruguay Round represented the eighth, and most ambitious, expansion of the GATT since its establishment in 1947. The cornerstone of the regime is the most-favored-nation (MFN) principle: any trade concession offered to one signatory, such as a reduced tariff or the privilege of opening a bank, must be granted to every other signatory. As of 1996, more than 120 nations have agreed to this rule, leaving China the only major country denied the benefits of the GATT’s unconditional MFN status. Participants in the Uruguay Round have agreed to cut their tariffs by an average of one-third, and many quantitative restrictions on imports have been abolished outright. (The quotas that remain are scheduled to be replaced by tariffs, which are to be reduced or phased out over time.) While the vast bulk of world trade falls under the rubric of the new GATT, some important sectors—most notably shipping, telecommunications, and financial services (e.g., banking, insurance, and securities)— were permitted to operate outside the system’s unconditional most-favored-nation framework. 18 See, e.g., Whalley and Hamilton 1996. 19 Jackson 1994; Lowenfeld 1994. 20 The provisions concerning dispute settlement are contained in Annex 2 of the Uruguay treaty. For the text, see General Agreement on Tariffs and Trade 1994. 17
70
H O W
D O
N A T I O N S
C O O P E R A T E ?
nisms, custom dictated that the recommendations of GATT mediators take effect only when approved unanimously by the executive committee of the body’s full membership.21 By granting the would-be loser a veto over the final decision, this de facto consensus rule prevented the GATT from playing the “governance” role currently enjoyed by the WTO. Now, if a WTO arbitration panel determines that a particular signatory is engaging in GATTillegal activities, that signatory’s only practical recourse is to appeal the decision to an appellate body whose decisions cannot be overturned except by the unanimous consent of all WTO member states. Barring this unlikely development, the original panel’s recommendations take effect, the objections of the GATT transgressor notwithstanding.22 While most apparent within the realms of trade and finance, the movement toward supranational governance has also begun to make inroads into national security policy. A case in point is the once-obscure Conference on Security Cooperation in Europe. At its founding in 1975, the Vienna-based Conference was little more than a forum for East-West dialogue structured by a collection of loosely connected security and human rights agreements. Beginning in the early 1990s, however, members began calling upon their institution to play a more active role in mediating European conflicts.23 The result was the signing of a 1994 agreement aimed at bolstering the dispute resolution apparatus of the regime, which has since been upgraded from a “conference” to a full-fledged “organization.”24 The OSCE and WTO are just two manifestations of a wider trend toward international agreements whose initial terms and provisions are intended to be somewhat flexible. The qualifier somewhat is important here, for while these terms and provisions may not be as rigid as those embodied in previous agreements, they are not totally flexible either. Instead of letting individual member states alter (or reinterpret) the terms of cooperation unilaterally, many of today’s international agreements set forth explicit rules—procedures negotiated and incorporated into the agreement ex ante—for changing these terms ex post. While some international treaties provide for more elaborate collective decision-making bodies than others, the very fact that 21
Jackson 1994; Pescatore 1993. Does the absence of sanctioning power mean that WTO member states can freely ignore negative rulings against them, and hence that these rulings “do not matter” after all? Not necessarily. While it is true that the losing party is perfectly free to disobey the recommendations handed down by the organization, such a flagrant violation would warrant the suspension of concessions by other WTO member states. As these other states raised their tariffs against the exports of the defector, the costs for the defector in terms of reduced market access could quickly begin to outweigh the benefits of standing firm. See, e.g., Whalley and Hamilton 1996, 115. 23 Heraclides 1993. 24 Greenhouse 1994. The negotiating partners at the 1994 Budapest summit also authorized the OSCE to supply peacekeepers on an as-needed basis to areas of brewing hostilities in the former Soviet Union—a measure explicitly designed to reduce the peacekeeping burden of the United Nations, an international organization that is currently in the midst of its own transformation into a full-fledged supranational governance structure. See, e.g., Durch 1996. 22
T H E
E F F I C I E N C Y
R A T I O N A L E
71
such bodies exist poses something of a problem for Keohane and other neoliberal theorists. How could it possibly be in a state’s interests to submit to a “higher” supranational authority? The remainder of this chapter tells the story of how some scholars, sympathetic to the logical structure and normative thrust of neoliberal theory yet disappointed by its inability to account for the rise of full-blown supranational governance structures, went in search of new theoretical tools. What they discovered is exactly what analysts of American politics found when they went on a similar quest in the 1980s: the new economics of organization.25 For those on the frontiers of this research effort, the new economics of organization holds the promise of bringing the study of international institutions back into line with empirical realities. At the same time, however, it also vindicates the core intuition of neoliberal theorists—and, if one takes their rhetoric seriously, that of many policy makers as well—that institutions are somehow fundamental to the pursuit of international cooperation.26
The Incomplete Contracting Argument It might be instructive at this point to distinguish between two different types of contractual arrangements. First are those in which the parties—be they buyers and sellers transacting in a marketplace or nations interacting within the international system—agree ahead of time to a fixed set of rules and, if necessary, establish appropriate surveillance mechanisms to discourage opportunism. These arrangements are the bread and butter of neoliberal theory. A second type of contract is nearly identical to the first except that the contracting parties explicitly allow for subsequent changes in the initial terms of cooperation, specifying that any disagreements concerning the nature of these changes be settled by a higher authority rather than, as would ordinarily be the case, through an essentially open-ended, extra-institutional process of direct bilateral bargaining and negotiation. In both types of contracts, signatories agree to limit their freedom of action. Yet while neoliberal theorists have been singularly attentive to the first type, it is the second type that, in today’s international arena, is fast becoming the more prevalent. Recognition of the yawning gap between (neoliberal) theory and reality has led in recent years to an exciting new round of theoretical innovations and refinements. As has long been the case within the study of American politics, questions about the formation of collective decision-making institutions—and about the creation of governance structures more generally—are thus now emerging as a focus of lively discussion within the study of inter25
Moe 1984; Shepsle 1986; Weingast and Marshall 1988. Realists, it seems fair to say, are skeptical. While they concede that regional and multilateral governance structures may be flourishing in some areas, they insist that states will never allow this new trend—if indeed it is a trend—to extend to security structures like NATO and the UN. 26
72
H O W
D O
N A T I O N S
C O O P E R A T E ?
well.27
national relations as Often using the European Union as a reference point, contributors to this body of work suggest that cooperation under conditions of rapid political, economic, and technological change may in fact necessitate a transfer of national policy-making prerogatives to supranational entities. This work thus shares neoliberal theory’s overarching emphasis on efficiency as the central motivating force behind institution building. The difference is that it seeks to explain why national governments are often willing to “pool” their sovereign authority—a possibility that neoliberal theorists never seriously entertained, much less sought to incorporate into their larger explanatory model. To explore this phenomenon, contributors to the new literature draw extensively upon economic theories of hierarchy, organization, and firm structure, none of which take the institutional requirements for cooperation to be as easily satisfied as Keohane and others had envisaged.28 True, institutional agents may well be necessary to perform the tasks of monitoring compliance and identifying defectors. But as long as each cooperation partner is clear about what is permitted and forbidden, these tasks can be carried out fairly easily. To proponents of the new economics of organization, the real challenge lies in ensuring that the partners really are clear about what is permitted and forbidden—and at all times, not just at the beginning of the process.29 In international relations, questions about what it means to “cooperate” or “defect” are endemic. In part, this is because the individuals who negotiate international agreements are often guided by time-sensitive domestic political concerns, and so rush into deals without taking the time to set forth their terms as carefully as might be the case in less political environments. Under pressure to accommodate a multiplicity of conflicting domestic interests, these negotiators may also view the incorporation of vague language and loopholes as a necessary evil without which their new accord would stand little chance of being ratified back at home. There is, in short, good reason to think that international “contracts” would be particularly susceptible to differing interpretations of each party’s rights and obligations. But some degree of imprecision or incompleteness is inherent in all agreements, at least all those intended to endure for more than a very short period of time. In a world of rapid political, economic, and technological change, it is simply not possible to determine ahead of time which types of conflicts and questions will arise over the lifetime of a long-term contractual 27
Garrett and Weingast 1993; Moravcsik 1998; Yarbrough and Yarbrough 1992. See, e.g., Williamson 1985; Milgrom and Roberts 1992. While economists address issues of structure and control as they arise in the private sector, proponents of the new economics of organization note that these same issues also arise in political settings, even if the institutional forms taken by political organizations tend to differ, in some cases quite fundamentally, from those of organizations (e.g., firms) located outside the public sector. See Dixit 1996; Lipson 1985; Moe 1984; North and Weingast 1989; and Weingast and Marshall 1988. 29 Cf. Grossman and Hart 1986; Kreps 1990; and Milgrom and Roberts 1990. 28
T H E
E F F I C I E N C Y
R A T I O N A L E
73
relationship between two states. And even if it were, the contracting parties (i.e., the individuals who preside over the governments of these states) would not necessarily have the information, let alone the time, to specify appropriate responses for each one. Before deciding on an appropriate response, these parties would first have to consider whether the particular rule violation in question was the product of deliberate malfeasance rather than an inadvertent, and thus innocent, misreading of their agreement. Only in the former case—and perhaps not even then—would retaliation against the defector be warranted. By the same token, there would be little point in retaliating against a party whose failure to adhere to the letter of an agreement was due to limited resources rather than to a lack of political will.30 Lacking any independent authority, neoliberalism’s watchdog agencies would be unable to render such distinctions, leaving the door open for each party to the agreement to read its (ambiguous) provisions as it pleased. Alternatively, the parties to an international pact or treaty could agree ahead of time that a particular set of collective decision-making procedures would be followed whenever a dispute over a particular clause or provision in the agreement needed adjudicating. Otherwise, an unstructured and openended—hence time-consuming—bargaining process would be necessary each time there arose a new set of circumstances not explicitly covered by the (unavoidably indeterminate) language of the agreement.31 In the course of trying to resolve these contractual ambiguities to everyone’s satisfaction, international cooperation could quickly devolve into chaos. As in traditional neoliberal theory, the institutions envisaged by proponents of the new economics would play a crucial role in bolstering decentralized enforcement mechanisms based on reputation and reciprocity. But 30 This issue often arises in the case of environmental treaties, where compliance may require a higher level of technical sophistication and financial or bureaucratic wherewithal than the treaty’s poorer signatories actually possess. See Chayes and Chayes 1993, 188 – 95. 31 The problem with open-ended bargaining is not only that it is time-consuming. There are in fact a whole range of costs associated with unstructured (e.g., market) interactions—costs that may account for the fact that so much economic activity takes place within large, vertically integrated firms (Coase 1937; Klein, Crawford, and Alchian 1978; Williamson 1985). Of particular interest in the contracting literature are the inefficiencies that arise when certain parties to a transaction would be required to invest in relation-specific assets, thus making themselves vulnerable to the “holdup” problem ( Joskow 1985). In the international relations literature, this idea lies behind much of Albert O. Hirschman’s work (e.g., Hirschman 1945 and 1970); see also Waltz 1979, 106–7; Frieden 1991, 434– 40; and Yarbrough and Yarbrough 1990 and 1992. Note that the issue here is not asset specificity per se but its (potential) asymmetry. Suppose that party A’s ability to exit the transaction were to decline faster than party B’s. Party B would then be in a position to decide unilaterally how unforeseen contingencies would be dealt with, a prospect that might lead A to refuse to cooperate in the first place. If, on the other hand, A were assured ahead of time that all future disputes would be handled “fairly”—that is, according to a set of previously agreed upon rules and procedures rather than through an essentially open-ended bargaining process—this holdup problem (and concerns about “exit options” and “sunk costs” more generally) would no longer pose as great an obstacle to successful longterm contracting.
74
H O W
D O
N A T I O N S
C O O P E R A T E ?
rather than passively relaying information between members, these institutions would serve as full-fledged governance structures, their mandates extending beyond fact-finding to encompass evaluation (determining whether an ambiguous provision or rule has been violated and, if so, whether the breach was committed inadvertently or by design) and recommendation (deciding what each party must do to make amends). In the interests of cooperation, national leaders would transfer their discretionary authority to a predesignated supranational intermediary or decision-making structure, giving it license—within certain contractually specified boundaries—to act as their agent, resolving any unforeseen developments that might happen to arise over the lifetime of their (i.e., the government-principals’) ongoing relationships.32 Thus far, empirical applications of this perspective have been limited, with Europe’s construction of a supranationally administered internal market (its so-called 1992 agenda) receiving by far the most attention.33 That the empirical focus of this literature should be so narrow is curious given that the emergence of statelike international institutions has occurred in other domains of state interaction besides trade. Later on I will try to suggest what a “new economics” account of the Canada-U.S. free trade pact and European Monetary System might look like (see chapters 6 and 9, respectively). For now, I merely want to underscore the point that the new economics of organization has broad applicability to a much wider class of phenomena. It is sometimes argued, for example, that trade liberalization on a global scale requires a hegemon to ensure compliance.34 From the perspective of the new economics, however, it is hard to see why the Axelrodian reputational mechanisms discussed earlier—if bolstered by a well-crafted supranational governance structure—could not support broad-based compliance with multilateral trade liberalization schemes nearly as well as smaller-scale arrangements involving two or three countries. While proponents of the new economics have yet to extend their analysis to the prospects of the WTO, it is a safe bet that when they do they will place little if any weight on the organization’s role in oversight and monitoring. If the authority of the new trade organization were limited to this—if it permitted each member state 32 The authority exercised by the agent over—but ultimately for the benefit of—its principals is always assumed to be limited. As Coase writes in his classic 1937 essay, “For [a] series of contracts is substituted one . . . whereby the factor, for a certain remuneration . . . agrees to obey the directions of an entrepreneur within certain limits. The essence of the contract is that it should only state the limits to the powers of the entrepreneur. Within these limits, he can therefore direct the other factors of production.” Coase 1937, 391, emphasis added. Coase’s idea here has obvious parallels with Herbert Simon’s (1947) “zone of acceptance”; see also Barnard 1938. 33 While the geographical scope of Beth and Robert Yarbrough’s work (see esp. Yarbrough and Yarbrough 1992) is somewhat broader—their book has a brief discussion of the CanadaU.S. Free Trade Agreement, for instance—they, too, concentrate on the politics of European trade. 34 Gilpin 1975 and 1981; Kindleberger 1973; Krasner 1976.
T H E
E F F I C I E N C Y
R A T I O N A L E
75
to render its own interpretation of the rights and obligations of the parties— the prospects for multilateralism would be little better under the new trade regime than they were under the GATT regime it supplanted. With no mechanism to clarify the terms of an enormously complicated treaty, disagreements over what constituted cheating would be likely to erupt into spirals of retaliation and counter-retaliation, threatening each time to send the world back to a more protectionist trade environment. And differences over matters of interpretation were indeed, on many accounts, what finally stalled progress toward liberalization under the GATT regime that preceded the WTO. The problem with the GATT was not, primarily, the absence of a hegemonic enforcer. The idea all along was that the regime would be self-enforcing—and had the appropriate structural arrangements been in place, there is little reason to believe that it would not have been. Nor can GATT’s failures be attributed to defective monitoring— though, to be sure, there was considerable room for improvement in the regime’s Trade Policy Review Mechanism (TPRM), which was established in 1989 to gather data on “the general effects of national trade policies” rather than on compliance per se.35 Rather, the main shortcoming of the GATT, as Finger (1991) and other trade analysts point out, was that its various codes, despite their considerable length and detail, remained highly ambiguous. As such, they afforded ample room for differing interpretations and selective utilization, the most prevalent form of shirking being the erection of nontariff barriers. (GATT codes covering nontariff barriers were necessarily more ambiguous than those regulating tariffs.) Now that an elaborate dispute resolution system has been established to resolve confusion over the meaning of “less than fair value,” “serious injury,” and other vague but crucial terms, proponents of the new economics should be predicting an unprecedented expansion in the flow of goods and services throughout the global marketplace.36 35
Abbott and Snidal 1998, 17; General Agreement on Tariffs and Trade 1994, Annex 3. The inauguration of the WTO would thus seem to contradict the traditional view that multilateralism can occur only in the presence of a global hegemon. That two leading advocates of the new economics of organization would continue to subscribe to this view seems odd (cf. Yarbrough and Yarbrough 1992, 103 – 4). As the quintessential multilateral trade organization appears to be taking on the attributes of a very large “minilateral” regime, the Yarbroughs’ distinction between mini- and multilateralism no longer seems apt: Virtually all significant steps toward liberalization are presently taking place within the context of nascent supranational— i.e., minilateral—institutions. By the same token, one could question the utility of the authors’ distinction between minilateral and bilateral arrangements. Invoking standard neoliberal arguments, the Yarbroughs suggest that watchdog or monitoring institutions may be sufficient to ensure compliance with free trade pacts involving no more than two or three nations, thus implying that bilateral agreements can be effective even when they do not incorporate governance structures for dispute settlement. This line of reasoning does not mesh with the new economics perspective the Yarbroughs claim to embrace, however. It may also be empirically unwarranted, as efforts to liberalize trade bilaterally—but not minilaterally (i.e., with no ex ante provision for the adjudication of disputes)—have rarely endured or made much difference to world trade flows. Notwithstanding their obvious importance, questions about the practical implications of 36
76
H O W
D O
N A T I O N S
C O O P E R A T E ?
The Multiple Equilibria Argument Although the analytics and implications of the prisoners’ dilemma continue to fascinate theoretically oriented students of international cooperation, the last several years have seen the theoretical spotlight shift toward the less well understood (or at any rate less familiar) coordination dilemma. Ironically, this shift in focus was prompted by the work of a self-proclaimed realist, Stephen Krasner, whose 1991 study of the global telecommunications industry called attention to a new kind of collective-action problem, one that Keohane and other neoliberal researchers had tended to relegate to secondary status.37 Historically, as was noted at the beginning of chapter 2, much of the controversy between neoliberals and realists centered around the issue of enforcement, with neoliberals proposing—and realists disputing—various solutions to the problem of how to keep nations from reneging on international agreements once they have taken effect. That the problem of enforcement emerged as a hot topic of debate was only natural, for international cooperation is possible only insofar as the parties involved trust each other to keep their promises. Strictly speaking, however, the types of difficulties that arise in trying to ensure that all parties do in fact follow through are second-order problems. The initial problem is getting a group of would-be cooperators to make these promises in the first place. Why, in practice, might this prove to be a major hurdle? To simplify matters, let us assume that any agreement a group of actors might enter into would be enforceable—that is, there is no possibility that anyone will be suckered once the agreement comes into force—and that each party to the negotiations knows this to be the case. Why, despite this knowledge, might the group of would-be cooperators still find it difficult to strike a deal? One reason may be that some parties to the negotiations believe their future cooperation partners would reap a disproportionate share of any ensuing joint gains.38 The world envisioned by Grieco and other realists is one different institutional arrangements—how the forms they take might influence whether, when, and where trade liberalization occurs—receive hardly any attention in the Yarbroughs’ treatment. 37 The question of whether to classify coordination problems as a particular type of collective-action problem or as a distinct analytical category raises a host of knotty conceptual issues. The defining attribute of a collective-action problem is that individual and collective rationality are at odds: the utility-maximizing course of action for each member of a collectivity produces an outcome that is patently suboptimal from the standpoint of the collectivity as a whole. In some coordination games (e.g., “chicken”), however, it is unclear what utility maximization for each player would entail. Within the international relations literature, Stein (1983) and Snidal (1985a) offer two early and illuminating treatments of these and related issues; Stein, for example, draws a sharp distinction between “dilemmas of common interests” (collective-action problems) and “dilemmas of common aversions” (coordination dilemmas). 38 Grieco 1990.
T H E
E F F I C I E N C Y
R A T I O N A L E
77
Figure 4-2. The “Battle of the Sexes” Game
in which these kinds of “relative-gains” concerns are both serious and pervasive, and in which (as a result) security-conscious states encounter few opportunities for mutually beneficial deal making. Yet careful analysis of the coordination dilemmas highlighted in recent work by Krasner and others raises a different possibility: Perhaps reaching agreement is difficult because states encounter too many, not too few, opportunities for collective gain. Since the publication of Krasner’s 1991 study, a number of scholars have begun taking a closer look at this issue.39 Most begin by noting that firstorder obstacles to cooperation are most severe in cases where the available Pareto-improving equilibria from which the group of would-be cooperators must choose are distributionally distinct. For this reason, scholars have tended to view the problem in terms of a “battle of the sexes” game.40 The story behind this game, one rendering of which is depicted in Figure 4-2, runs as follows. A man and a woman must decide how to spend their upcoming vacation. One possibility would be to go hiking together in the mountains (see the upper-left-hand cell of the figure). This alternative would be the man’s most-preferred outcome but the woman’s second choice. Conversely, a vacation spent together at the beach (the lower-right-hand cell) would be the woman’s most-preferred outcome but the man’s second choice. And, fi39 40
See, e.g., Fearon 1998; Garrett and Weingast 1993; Martin 1992b; and Morrow 1994. E.g., Krasner 1991, 339 – 42.
78
H O W
D O
N A T I O N S
C O O P E R A T E ?
nally, either of these outcomes would be preferable to vacationing separately, that is, to ending up in one of the other two cells of the matrix.41 Because both players prefer to spend time in each other’s company, there would be little point in either one’s defecting unilaterally. If, for example, the man were to head for the mountains after he had promised to meet the woman at the beach, he would simply be lowering his own payoff, effecting a switch from his second-most-preferred outcome (the lower-right-hand cell) to his third (the cell located in the lower left). But the fact that there is no incentive to cheat in a battle of the sexes game and that, as a result, monitoring and enforcement issues do not arise does not mean that cooperation between the two parties is therefore unproblematic. To the contrary, cooperation could fail to emerge even if neither player faced the slightest prospect of being double-crossed. However badly the man and woman in Figure 4-2 might wish to share each other’s company, they would first need to address—and resolve—the thorny issue of how to share each other’s company, whether by joining each other at the beach (the woman’s first choice) or by hiking together in the mountains (the man’s first choice). If both players were to hold out for the equilibria they most strongly preferred, the result would be an outcome desired by neither player, with each ending up worse off than needed to be the case. Having belatedly come to appreciate the potential seriousness of this “multiple equilibria” impediment to cooperation, scholars of international relations are now moving energetically to explore its implications for supranational governance. Might supranational arrangements foster collectively desirable outcomes by limiting the number of potential equilibria?42 One possibility is that institutions provide “focal points” along the lines previously suggested by Thomas Schelling.43 Common sense suggests that this focusing role would be of greatest importance in situations similar to the battle-of-the-sexes-type sce41 While most scholars use the battle-of-the-sexes game as a springboard for exploring the causes of suboptimal behavior (see below), Krasner uses the game to make the point depicted in Figure 2–3: namely, that when the distributional conflicts engendered by interstate cooperation are finally resolved, the resulting equilibrium is likely to benefit powerful states a lot and weak states only a little. For Krasner, the issue is not how to get to the Pareto frontier—at no point does he entertain the possibility that states confronting multiple equilibria might, for this reason, fail to coordinate their activities—but rather, as discussed in chapter 2, which point along that frontier is selected. 42 I should point out that the difficulty of striking agreement when multiple paths to cooperation are possible is not unique to the battle-of-the-sexes game. Although Axelrod did not emphasize the point in The Evolution of Cooperation (1984), game theorists have known for some time that the iterated prisoners’ dilemma analyzed by neoliberal analysts has a potentially infinite number of stable solutions, ranging from permanent defection by all parties to permanent cooperation (see, e.g., Fudenberg and Maskin 1986). Even together, of course, the prisoners’ dilemma and battle-of-the-sexes games hardly capture the full range of strategic settings that nations encounter, and many important political interactions conform to neither game’s basic structure. For a useful survey of additional games and their relevance to international politics, see Martin 1992b. 43 Schelling 1960, esp. chap. 4; cf. O. Young 1994, 110 –11.
T H E
E F F I C I E N C Y
R A T I O N A L E
79
nario just described, in which the underlying preferences of each would-be cooperator differ. Yet even if what these actors confronted were a “pure” coordination dilemma (i.e., even if they placed the same value on each of the Paretoimproving equilibria they were considering), it is still possible that they would be better off ceding agenda-setting powers—that is, the right to designate the initial terms of agreement—to a third party. The point is sometimes illustrated with reference to the agenda-setting powers exercised by the European Commission, most transparently during negotiations over the landmark Single European Act of 1986. By the mid1980s, all twelve member states of the European Community had recognized the importance of reinvigorating the integration process. Attitudes about what the “new” EC should look like nonetheless varied from one government to the next. Thus, while British Prime Minister Margaret Thatcher strongly favored the minimalist form of integration enshrined in the principle of mutual recognition—any good or service that can be sold in one country must be available for sale in all member countries—her more interventionist German counterpart, Chancellor Helmut Kohl, emphasized the need for increased harmonization of EC-wide health and safety standards, tax policy, and political decision making. On some accounts, it was in order to ensure that their competing visions would not prevent the process of integration from moving forward that Thatcher, Kohl, and other European leaders allowed the commission to dictate the integration agenda.44 The result, thanks in large measure to the catalyzing effects of the commission’s detailed directive on the internal market (the 1985 “White Paper”), was a pro-integration agreement that, while pleasing no one perfectly, was better than ceaseless intergovernmental wrangling and no agreement at all. To be sure, a supranational third party was not strictly necessary in this case. Instead of granting such wide latitude to the EC Commission, the governance structure established by the twelve EC nations might have assigned a single nation—a France or a West Germany, for example—the task of deciding which vision(s) of European integration would be voted upon by the group as a whole. But of course that is not the governance structure the EC’s member states chose to create. Like many international agreements, the Treaty of Rome (the founding charter of the European Community) vested agenda-setting powers in a set of largely autonomous supranational decision-making institutions and agencies. Given the alternative—granting one’s negotiating partner total control over the process of equilibrium selection— it is not hard to see why. Conclusion While the (ex ante) agenda-setting tasks performed by international institutions have only recently emerged as a topic of serious scholarly investigation, 44
See, e.g., D. R. Cameron 1992b; Garrett 1992; Pollack 1997.
80
H O W
D O
N A T I O N S
C O O P E R A T E ?
there is a common thread linking this work with the earlier, albeit still relatively recent, tradition of analyzing supranational institutions as referees in (ex post) interstate disputes. Both approaches suggest that successful cooperation often requires something more than the existence of well-functioning information clearinghouses. What it requires, particularly in situations characterized by high uncertainty, is the construction and empowerment of higher-level, quasi-sovereign governance structures. Once established, these structures do the necessary work of clarifying the rules of the game in cases where the provisions contained within the international agreements at issue are ambiguous or “incomplete.” And where states run the risk of getting bogged down in battle-of-the-sexes-like coordination problems (and thus of forfeiting the gains from cooperation altogether), governance structures may also be permitted, indeed encouraged, to directly formulate the terms of the agreements themselves. Or so, at any rate, the argument runs.
C H A P T E R
5
Broadening the Debate: The “Power Politics” of Institutional Design
THE INTERWEAVING of neoliberal theory and the new economics of organization is a relatively recent phenomenon, and as with all research endeavors in their early stages of development, a great deal of work remains to be done. That said, contributors to the literature are confident that they are at least on the right track. I am less convinced. For while these contributors can be credited with putting forward a cogent explanation for the recent proliferation of supranational regimes, theirs is not the only explanation. In what follows, I lay out an alternative perspective, one in which the actor or (more commonly) actors who set the cooperation bandwagon in motion are free to impose their own institutional choices on other actors in the system. Because the “enacting coalition” can shift the institutional status quo unilaterally, its members do not need to take account of their current partners’ institutional preferences. When it comes to matters concerning institutional design, the enacting coalition is in the driver’s seat, and everyone else better look out.
The Importance (to the Winners) of Entrenching the New Status Quo For the sake of argument, let us assume that this is the case—that, in other words, certain states would be likely to join virtually any institutional arrangement that their more powerful neighbors might choose to establish. Now the theoretical task becomes one of explaining why, despite their holding a dominant position, the actors who constitute the enacting coalition might want to do their bidding through supranational structures that afford other actors, including their opponents (i.e., the losers), an opportunity to voice their complaints. In fact, such structures do not just permit the losers to voice their disagreements; they also ensure that the losers get to play some role, however institutionally delimited or circumscribed, in determining how these disagreements are ultimately resolved. By doing this—granting actors who prefer the pre-cooperation status quo some scope to pursue their own (very different) interests—the members of the enacting coalition would seem more likely to diminish than to enhance their own gains from cooperation. Why grant their regime’s opponents any input at all? Expanding upon the argument of chapter 3, let me suggest that the win-
82
H O W
D O
N A T I O N S
C O O P E R A T E ?
ners’ motivation for offering the losers some (albeit limited) means of addressing their concerns is the belief that this will weaken the losers’ “natural” inclination to abandon the arrangement at the first available opportunity. Because the A’s and B’s in Figure 3-5 want the arrangement they create to endure, they purposely structure it in a way that mitigates the adverse consequences of participation for would-be defectors. Of course, some losers—the C’s and D’s, for example—would pose relatively little risk of defection. As long as the alternative of nonmembership would be worse for these actors than cooperation, they could be relied upon to comply with A and B’s new regime. And even if the C’s and D’s were to go ahead and exit, the A’s and B’s would still be better off with respect to the initial status quo. In short, even if their cooperation partners in other countries disliked the arrangement that was being “forced” upon them, this fact alone would probably not be viewed by the prime movers as a major threat to their new regime’s long-term survival. After all, these partners are caught between a rock and a hard place (as the enactors would surely appreciate), since opting out of the regime would be even worse for them than allowing the enactors to go it alone. Suppose, however, that a member of the enacting coalition itself were suddenly turned out of office by a political party that strongly disapproved of the cooperative arrangement it had inherited. In that event, the beneficiaries of the arrangement, including the original coalition of governments responsible for creating it, would have a serious problem on their hands. For now the list of “losers” would include an actor whose future departure from the new regime would be sufficient in and of itself to severely disrupt its operation, if not single-handedly bring about its demise. That being the case, the regime’s original sponsors would presumably have a strong interest in finding some way to mitigate the suffering experienced by their internal enemies, particularly, again, those who might one day be in a position to destabilize and subvert their collective achievement.1 The new enthusiasm for supranational governance is, in this view, a direct outgrowth of increasing domestic political instability. Even if two or more “states” have been cooperating with each other for an extended period, there is always a possibility that one of them will one day find itself confronted with an unfamiliar partner—a country presided over by an entirely new group of individuals. Not only might these new individuals (the incoming domestic elites) be playing the game for the first time; they would also be 1 Not that these beneficiaries would need to drop the costs incurred by their domestic opponents all the way to zero (cf. Moe 1990b). Once the enactors’ new decision-making structure was up and running, the costs to the losers of abandoning it would have a natural tendency to rise over time. As a result, the A’s and B’s would not need to offset the burden of participation to the point that their successors ended up better off than before the initial (noncooperative) status quo had ever been removed from their choice sets. The next section offers a more complete discussion of this point.
B R O A D E N I N G
T H E
D E B A T E
83
bringing to the table their own set of political priorities—policy objectives that might differ substantially from those embraced by the governments that held power when the regime was first proposed and its terms ironed out. Two additional points are worth making here. First, the prospect of new leadership—whether precipitated by a peaceful election or a violent coup d’état—would be a source of concern even if the prime mover behind a new international arrangement were to consist of a single state or government. If, instead of a multinational coalition (A and B), the prime mover were a singular entity (A or B), that entity’s desire to shield its policies from future shifts in its internal political orientation could itself be sufficient inducement for “supranationalizing” the policy process. Perhaps this can explain why domestic officeholders in the United States have often been willing to exert their power at one remove—through multilateral institutions such as the United Nations and the World Bank—rather than through more direct forms of control.2 Second, although government turnover is the “internal” threat upon which I have been focusing thus far, within-government preference shifts are certainly also possible, as it is not uncommon for governments to abandon the policy objectives their leaders espoused upon first taking office.3 Whatever the underlying reason for this—whether government policy reversals are driven mostly by external developments (e.g., changing global economic conditions) or internal ones (e.g., a party revolt that culminates in the ouster of a standing government’s top officials)—the very possibility of such swings is likely to weigh heavily on the minds of the original prime movers. Like the threat of government turnover, then, the possibility of radical policy U-turns is something the institution’s creators would surely want to take into account. In tailoring new supranational institutions to meet these threats, the winners would be taking the long view, adjusting their current choices in light of domestic developments that would not yet have occurred. In this sense, their behavior is farsighted, oriented less toward their current political opponents than toward the (domestic) losers of the future. As for altering the behavior of the losers in other countries, this is less important since, given their constrained choice sets, these losers—the C’s and D’s—would be foolish to exit the institutional arrangements engineered by the prime movers, and waging an all-out war against them would be pointless. In sum, one should not be surprised to find the institutional choices made in the initial period being driven primarily by expectations about subsequent periods, and specifically about what might happen were the members of the current enacting coalition to be supplanted by a group of actors whose pol2
A different argument is that the United States is attracted to quasi-autonomous multilateral institutions because they veil or “legitimate” its domination of the postwar international system. Krasner 1985, 62 and 77; see also Abbott and Snidal 1998; Baumgartner and Burns 1975; Block 1977. 3 Kitschelt 1999; Stokes 1999.
84
H O W
D O
N A T I O N S
C O O P E R A T E ?
icy objectives differed radically from their own. By lowering the costs of compliance, the original coalition members would be hoping to delay—and ideally prevent—tomorrow’s dominant actors from abrogating any preexisting treaty obligations or, even worse, forming a new enacting coalition of their own.
Implications for Institutional Design How do the winners’ interests in “freezing” their achievements influence the types of cooperative arrangements they create? Might these interests predispose them toward more elaborate—and more flexible—supranational structures than efficiency considerations alone would dictate? This, I have just argued, is exactly what one ought to expect. By delegating decision-making responsibilities to a set of mediating governance structures, the enacting coalition reduces the burden of participation in its new regime, thus ensuring (it hopes) that the regime will continue to operate long after the power held by the coalition’s individual members has waned or been usurped by internal challengers. To entrench the new cooperative status quo, the winners are actually better off providing their domestic successors some means of influencing—presumably to their own advantage—the original terms of cooperation. And this, I have argued, makes for cooperative agreements that are considerably more flexible than would otherwise be efficient. That is not the end of the story, though, since the most flexible agreement would be one in which each party could interpret the rules however it wished. Indeed, dropping the costs of compliance to zero—by, say, making it possible for signatories to renegotiate the terms of a treaty from the ground up—would provide the enacting coalition with the greatest protection against its regime’s would-be destroyers. At the same time, however, this would come at the price of completely eliminating the benefits that accrue to members of the enacting coalition itself. In fact, the enacting coalition would not need to pay this price. In most cases, simply moderating the costs of participation should suffice. The reason is that there are often costs to “opting out” of well-established cooperative arrangements, even for participants who would have preferred not to have these arrangements in the first place. This assertion is, of course, somewhat controversial. According to Susan Strange (1983, 345), “All those international arrangements dignified by the label regime are only too easily upset when either the balance of bargaining power or the perception of national interest (or both together) change among those states who negotiate them.” The implication of Strange’s remarks— that international arrangements are inherently different from domestic arrangements—would seem consistent with a vast body of research on the durability of domestic regulatory regimes within the United States. As numerous scholars have pointed out, the prime movers behind such regimes
B R O A D E N I N G
T H E
D E B A T E
85
had the luxury of embedding them within an overarching legal framework. Henceforth, future officeholders who wished to abolish them had to engineer the passage of new legislation. And in the United States—a separation of powers system characterized by a multiplicity of veto points—that is never an easy task.4 But even if legal injunctions against the destruction of newly established routines and institutions are unlikely to work as well in the nonhierarchical, anarchic environment of international politics, that does not mean that international regimes are as “easily upset” as Strange would have us believe. While it may be harder to insulate international structures, the evidence suggests that it is not impossible.5 This should not be surprising. After all, domestic political institutions often display a remarkable degree of staying power even in countries characterized by highly majoritarian systems of government. The impressive durability of an institution like Britain’s National Health Service cannot be attributed to formal checks on the exercise of public authority, since, if it so desired, a newly installed British government could readily obtain passage of new legislation calling for wholesale changes in the institutional status quo. In Westminster-style democracies, legal prohibitions against tampering with or destroying existing bureaucratic structures are a weak deterrent at best.6 And yet even in a country like the United Kingdom—where current power holders are free (in a legal sense) to abandon or replace institutional arrangements they dislike—domestic structures routinely outlive the constellation of actors who created them.7 What this suggests is that new organizational structures can be shielded from the vicissitudes of domestic politics even when they cannot be embedded in “legislative concrete” (to use Moe and Caldwell’s terminology). But what, precisely, is the source of this tendency toward inertia, this institutional stickiness? In some cases, the mere passage of time may be sufficient to protect a new institutional arrangement from its would-be destroyers. As Oliver Williamson and others have noted, the longer an ongoing cooperative relationship has been in existence, the greater—holding other factors constant—will be the cooperation partners’ mutual interests in keeping the relationship going.8 Not only does length of association breed familiarity; it also affords opportunities to those in the relationship to adapt to nuances in individual styles and characteristic modes of operation. Long-standing transaction partners are likely to have a difficult time locating suitable replacements. This basic principle applies in the present context, as well. Whatever the distributional consequences of a particular international agreement or treaty, the longer it has been in existence, the higher the value will government sig4
Cowhey 1993; Moe 1990b; Tsebelis 1995. See, e.g., Goldstein 1993; Keohane 1984; Krasner 1976; and Stein 1983. 6 Shugart and Carey 1992; Weaver and Rockman 1993. 7 Moe and Caldwell 1994. 8 See esp. Williamson 1985. 5
86
H O W
D O
N A T I O N S
C O O P E R A T E ?
natories—as well as private actors—typically place on its continuation, and hence the lower the losers’ propensity to terminate or otherwise disrupt it if they are ever given the chance. Altering long-established procedures and routines would require the losers to make costly behavioral adjustments. Entirely new sets of procedures and routines would need to be formulated and executed, along with corresponding changes to domestic legislation, wage contracts, product standards, and the like. Any time and decision costs that had been invested in figuring out how to behave under the old governance arrangements would be irretrievably lost. But the costs to the losers of uprooting long-standing cooperative systems could be greater still. After all, the longer a cooperative arrangement has been in effect, the more likely it is to have spawned powerful new actors who either did not exist prior to the arrangement’s entry into force (e.g., new government agencies created specifically to oversee a member country’s involvement in the arrangement) or who did exist but lacked political clout (e.g., industries and investors who benefit directly from the arrangement). The emergence of these new stakeholders could cause even the most stalwart opponents of the enacting coalition’s cooperative initiative to think twice about tampering with it if and when they were to someday take power themselves. For once the initiative had generated its own independent constituency—as, for example, the GATT generated a constituency among export producers9 —political opponents would have to consider whether the benefits of destroying the new arrangement (or removing its bias in favor of the original enacting coalition) would offset the costs of alienating the new breed of stakeholders it had acquired since coming into existence. Over time, then, there is a natural tendency for all cooperative arrangements to develop new, ever-expanding bases of support. In the absence of major interventions, however, constituency building typically proceeds rather slowly, occurring over decades rather than years or months. This, of course, poses something of a problem for the prime movers, who would like their new institutional creation to be surrounded as quickly as possibly with a broad-based and growing coalition of supporters. Given the immediacy of their concerns—even in a highly stable democracy like the United States, major partisan swings can occur (at least potentially) every two years—the prime movers would surely want to pursue every available means of expediting this “natural” constituency-building process.10 9
Destler 1992; Destler and Odell 1987; Oye 1992. While the “neofunctionalist” perspective on regional integration also emphasizes the importance of constituency building (Haas 1964; Nye 1971), neofunctionalists are inclined to dismiss the role of purposive action on the part of integration’s key beneficiaries (see, e.g., Streeck and Schmitter 1991). Much debated (see esp. Moravcsik 1998) has been the question of whether the French and German leaders who spearheaded the initial drive toward European unification after the war shared the view—held at the time by a number of prescient European intellectuals (see, e.g., Mitrany 1966; Monnet 1978)—that the integration process upon which Europe was soon to embark would develop its own momentum. A similar dispute concerns the recent development of a pro-integration constituency among domestic lawyers and judges within many 10
B R O A D E N I N G
T H E
D E B A T E
87
Other Strategies for Co-opting a New Regime’s Would-Be Destroyers One way of doing this would be to deliberately hold down the initial costs of the new arrangement, preventing the bulk of the costs from kicking in until after the dust has settled and expectations have begun to adjust. To this end, new treaties could be designed to take effect gradually rather than all at once. Trade agreements often take this form, with allowances for step-bystep implementation and phased-in concessions granted to especially sensitive sectors or even whole countries. The prime movers could also design their regime specifically to ensure that its domestic beneficiaries included a wider-than-usual cross-section of each signatory’s population. A widely employed method for broadening a regime’s clientele involves the interweaving of numerous sector- or industry-specific agreements into a single comprehensive package. In addition to economizing on transaction costs,11 the linking of disparate agreements in this fashion has the effect of enlarging the potential numbers and geographic diversity of new stakeholders. Henceforth, though the enacting coalition’s political successors might complain—quite legitimately—of being made worse off, any attempt to wipe out the entire package and start anew would likely be greeted by strong and vocal opposition. As a further means of accelerating the constituency-building process, the prime movers could make use of the “bully pulpit,” launching an aggressive publicity campaign directly linking their regime to principles and values embraced by their societies and, even better, by the global financial community. Here, too, the purpose would be to enlarge the pool of potential stakeholders, thereby tempering the destructive zeal of regime opponents who might one day be in a position to subvert the new institutional status quo.12 Yet another means to this end would be to impose a higher-than-usual legal threshold for amending the core rules and principles embedded within the enacting coalition’s original agreement. These higher “renegotiation European countries. On one side are neofunctionalists, who stress the role of accident and happenstance in promoting this convergence of opinion (Burley and Mattli 1993; Mattli and Slaughter 1995). On the other side are “neorationalist” critics of neofunctionalism, who insist that the emergence of support for integration within Europe’s legal community was foreseen, and possibly engineered, by the government leaders who launched the integration process back in the 1950s (Garrett 1995; Garrett, Kelemen, and Schulz 1998). 11 Keohane 1984, 91; Martin 1993; Yarbrough and Yarbrough 1992, 76 –78. 12 Note that for a publicity campaign of this kind to succeed in mollifying the new ruling elites who assume power after the original prime movers have departed, it would not need to alter these elites’ core identities or values. If the enacting coalition’s political successors refrain from destroying the institutional system they inherit, it may be that they simply do not wish to excite the financial markets or appear on the “wrong side of history,” not that they have somehow been indoctrinated or socialized into believing that the establishment of the new system was a good idea after all. The contrasting view—that membership in institutions generates normative and ideational consensus—is discussed in chapter 10.
88
H O W
D O
N A T I O N S
C O O P E R A T E ?
costs” would have the effect of restricting the prime movers’ successors to a dichotomous choice between outright annihilation of the agreement or total acceptance. This was almost certainly on the minds of the U.S. diplomats who crafted the United Nations in the aftermath of the Second World War. It is important to recall that at the UN’s inception in the mid-1940s, there was strong opposition to the organization within all five of the nations that were to become permanent members of the Security Council. While the “internationalists” were setting policy in 1945, their tenuous hold on power meant that the political situation had the potential to change dramatically within a matter of years, if not months.13 It was at least partly for this reason that the internationalists were willing to encumber their new organization with a high threshold for fundamental institutional reform.14 Finally, and perhaps most importantly, the enacting coalition could make a special effort to hasten the launch of its new organizational creation, particularly if the enactors had reason to believe their dominant position would be short-lived. There would certainly be costs to hasty action, as, indeed, there would be costs to all of the insulation devices discussed here. Setting an early date for passage and ratification of a new treaty would mean limiting the amount of time available for scrutinizing alternative proposals, and hence for allowing the winners to determine exactly which scheme would stand the best chance of advancing their interests. The schemes adopted would also be more likely to enter into force before all of the relevant procedural details and contingencies had been clearly specified and rendered internally consistent. These costs, however, would need to be set against the benefit to the prime movers of getting something “out there” as quickly as possible—even if that something did not accord quite as closely with their underlying preferences as another institutional arrangement that might have been chosen. The sooner their new structure was up and running, the sooner would citizens, interest groups, and members of the global financial community begin developing a vested interest in its perpetuation—and the greater, therefore, its prospects of withstanding a future decline in the enacting coalition’s power. 13
Cf. Keohane 1984, 118. Encumber is the right word here, for though it might work well as an insulation device, a high threshold for institutional reform does not come without costs. In the case of the UN, the effect of raising barriers to full-scale change in the UN’s basic architecture has been to reduce the organization’s capacity to adapt to fundamental shifts in the geopolitical balance of power. A nice illustration is the UN’s response to the emergence of Japan and Germany as major economic and (to a lesser extent) military powers. Although one can make a strong case that awarding Japan and Germany permanent seats on the UN Security Council would enhance the organization’s ability to serve as an engine for genuine collective security, any reallocation of Security Council seats would require a supermajority vote (two-thirds) in the General Assembly, ratification by national parliaments, and, most importantly, the support of the council’s existing members. Whatever the merits of the case for Japanese or German accession, the formal rules of the organization grant numerous actors—ranging from the five nations currently in possession of permanent council seats to the long list of aspiring council members from the developing world—a legally entrenched right to say no. 14
B R O A D E N I N G
T H E
D E B A T E
89
Power Politics and Institutional Variation If my analysis to this point is correct, power may well turn out to be the most important driving force not just behind the recent profusion of cooperative international arrangements, but also behind the profusion of particular kinds of cooperative arrangement: those in which permanent supranational governance structures give all signatories a voice, either directly (via supranational working groups and other collective decision-making bodies) or indirectly (through the creation of independent dispute resolution mechanisms and the like). To be sure, one would not always expect power to be the crucial explanatory variable; nor, even in cases where a power-politics approach did capture the essence of what was going on, would efficiency considerations be altogether irrelevant. Nevertheless, insofar as certain actors in the international system possess what I have termed go-it-alone power, one should not be surprised to find these actors establishing supranational governance structures for the express purpose of locking in or “congealing” their distinctive tastes or preferences.15 And if this is right, one would also expect the prevalence of such arrangements to vary from case to case depending (most importantly) on the strength of the original enactors’ domestic opposition. If, for example, the enactors’ cooperative agenda initially enjoyed widespread support “back home”—or if support for it were weak but all of the countries involved were autocracies16 —one would not expect the resulting cooperative arrangement to afford much scope for collective decision making or provide avenues for quasiindependent dispute settlement. Yet even where the cooperative structures established by the winners did afford such flexibility, we would still predict the losers to be worse off than they were under the original, pre-cooperation status quo. True, the prime movers would have incentives to accommodate the interests of the losers in various ways. But particularly if the winners were successful in their efforts to bolster the inertial tendencies of the cooperative arrangements they created, they would not need to raise the losers’ utilities by very much. From this, one can hypothesize that the supranational structures engineered by the winners, though designed to appease the losers, would nonetheless circumscribe the latter’s ability to modify the terms of the initial agreement. As noted earlier, formal rules making it difficult to dislodge or overturn an international agreement would be attractive to the agreement’s beneficiaries insofar as these rules discouraged future losers from abrogating the terms of the agreement prematurely (or from establishing an entirely separate arrangement more to their liking). In addition to serving as an insulation device, however, a high threshold for fundamental institutional reform would also benefit the agreement’s initial sponsors in another way: by limiting the 15 16
The term “congealing” is borrowed from Riker 1980: 445. Cf. Przeworski 1985.
90
H O W
D O
N A T I O N S
C O O P E R A T E ?
losers’ ability to alter the terms of this agreement in more than a superficial fashion. Thus, even when supranational mechanisms afford each party to an agreement some input into organizational decision making, this input is likely to be restricted to secondary issues or matters of interpretation—or so, again, the analytical framework set forth above would lead us to expect. Note, finally, that the purpose of supranational governance in all of this would not be to deter “cheaters” but, rather, to deter the enacting coalition’s successors from mounting a serious challenge to the new regime if they should ever get the opportunity. As we have seen, neoliberals and realists alike have tended to exaggerate the vulnerability of cooperative institutions to free riding. As long as the stream of “benefits” (relative to the available alternatives) is sufficiently large, participants will find it in their own interests to comply with institutional rules and norms. In fact, when it comes to institutional fragility, the real problem is that this benefit stream may be large for the political elites who make up the enacting coalition, but not for the officeholders who succeed them. To the contrary, the same institutional arrangement that is advantageous for the former may inflict untold damage upon the latter, in which case the latter—the enacting coalition’s domestic successors—will want not just to “defect” from the arrangement, but to destroy it altogether. By sabotaging the regime (or redirecting it toward very different purposes), these successors would be forfeiting neither individual nor collective gains. Like beauty, the value of cooperation is ultimately in the eye of the beholder.
Supranational Governance in Trade and Money: Two Cases that “Matter” To argue that go-it-alone power may be the driving force behind international cooperation and institutional choice is one thing; to demonstrate that it is the driving force is, of course, quite another. In the next four chapters I take up this challenge by examining the origins and structural characteristics of two particularly important recent multilateral initiatives—the North American Free Trade Agreement and the European Monetary System. Relying as they do upon counterfactual analysis and after-the-fact assessments of “state” utilities, these case studies do not constitute a definitive proof of my theoretical arguments. Unfortunately, the theoretical propositions introduced in this book do not easily lend themselves to testing via quantitative or other, more “definitive” techniques. That said, it would be wrong to conclude that my arguments are somehow impervious to empirical verification. In fact, aptly chosen and well-constructed case studies can go a long way toward compensating for these limitations.17 With this in mind, I have decided 17
See, e.g., Bates et al. 1998.
B R O A D E N I N G
T H E
D E B A T E
91
to concentrate the book’s empirical spotlight on cases where the potential payoff is highest. NAFTA and the EMS are two such cases. Among other things, they are both examples of economic cooperation, and—to this point, at least—the depth and overall policy impact of economic arrangements have far exceeded the depth and impact of cooperative arrangements in other, noneconomic spheres of interstate interaction.18 By focusing on two cases where cooperation clearly “mattered,” I can move directly to an examination of what these cases tell us about the theoretical questions at issue—that is, about how interstate cooperation emerged and why it took the particular institutional form it did.19 Taken together, the EMS and NAFTA also afford my analysis considerable breadth in terms of regional and issue coverage. In this respect, my study departs from the customary practice among students of international relations of examining interstate interactions within a single policy realm (e.g., trade or money, but not both) or a single geographic region (e.g., Europe or the Americas, but not both). Finally, while I would submit that my model of cooperation and institutional choice does at least as good a job of explaining NAFTA and the EMS as the more conventional neoliberal and realist models we are used to seeing in the literature, I want to be clear at the outset that I do not believe this qualifies as grounds for discarding these models, much less for abandoning the attempt to anchor the study of international cooperation in rational choice foundations. The conclusion I draw is not that I am right and “they” (i.e., 18
Lawrence 1996; Shulman and Jacobson 1998; Young and Osherenko 1993. Few observers would dispute that the EMS and NAFTA have had profound effects on state behavior, as public officials were compelled to initiate far-reaching reforms in their domestic monetary and exchange rate policies, in the first case, and in their trade and investment policies, in the second. That the EMS and NAFTA have had a corresponding impact on the behavior of nongovernmental (e.g., market) actors cannot be taken for granted, however. The fact that a supranational arrangement exists does not guarantee that it “matters,” in the sense of influencing economic or other outcomes. Thus, for instance, one cannot assume that the construction of an exchange rate regime will necessarily lead to low inflation, since the latter is influenced by pricing decisions made by private individuals whose behavior governments do not fully control. Likewise, even if one were to observe an increase in the amount of cross-border trade taking place within a given region, that increase could be due to population growth or rising incomes rather than necessarily to the formation of a government-initiated preferential trading arrangement. Be that as it may, there is a consensus among nearly all observers that the EMS and NAFTA did in fact have a sizable impact on macroeconomic aggregates (inflation, exchange rate volatility, etc.) as well as on intraregional flows of trade and investment, which explains why each case is widely—and, I think, correctly—regarded as among the three or four most far-reaching and important examples of international cooperation since the end of World War II. References to the empirical literatures on the two cases will be provided in the appropriate chapters below. For a more general discussion of how regional economic institutions may be contributing to the redirection of commercial and financial flows (“economic regionalism”), see Frankel 1997. 19
92
H O W
D O
N A T I O N S
C O O P E R A T E ?
proponents of the mainstream paradigms) are wrong, but that we should all begin working aggressively to supplement these paradigms, elaborating their rational choice logics in ways that can encompass the kinds of power dynamics that would appear to be at work in the real world and yet have, to this point, been almost entirely absent from scholarly analysis and debate.
P A R T
I I I
North American Trade
C H A P T E R
6
Jump-Starting the Free Trade Bandwagon
As basic as it is, the question of why nations sometimes give up or pool their sovereignty is only now beginning to emerge as a topic of serious debate among scholars of international relations. Outside the academy, however, the voluntary “retreat of the state”1 has been the subject of intense discussion—and controversy—for some time now. This is certainly the case in Western Europe, where a single European currency, the euro, has officially supplanted the mark, the franc, and the lira, and seems destined one day to subsume the British pound as well (see chapters 8 and 9). Not that this has been the only, or even necessarily the most controversial, case of countries’ surrendering their sovereign authority. Take the World Trade Organization, for instance. In the United States, the WTO’s inauguration in 1994 was fraught with political conflict, with opponents of the new institution—most notably the conservative columnist and perennial presidential contender Pat Buchanan—accusing President Clinton, a strong WTO supporter, of having ceded U.S. power to “faceless” international bureaucrats. In reality, of course, much of the world trading system had already become supranationalized—albeit on a smaller, regional scale—by the time the WTO appeared on the scene. That said, some of today’s regional trading arrangements do have a stronger supranational dimension than others.2 Of recent initiatives, for example, the European Union’s internal market has what is by far the most intricate and well-articulated higher-order governance structure. And yet, even in North America, the governance of intraregional trade and investment has now been institutionalized to the point that comparisons with Western Europe and the EU no longer seem inappropriate. This is so even if, at present, two of the signatories to the North American Free Trade Agreement—the United States and Canada—have no plans to eliminate their existing restrictions on the inflow of workers from Mexico, their NAFTA partner. Nor does anyone envisage the establishment of a common North American tariff (at least not anytime soon), much less the creation of a common North American currency.3 But while North America may never establish a single currency or a full-fledged common market, Canada, Mexico, and the United States have nonetheless developed, and are 1
Strange 1996. De la Torre and Kelly 1992; Fieleke 1992; see also Haggard 1995, 112–13. 3 That said, episodes of extreme exchange rate volatility will almost certainly be dealt with through concerted action on the part of Canadian, Mexican, and U.S. officials, as they were following the sudden devaluation of the peso in 1994. 2
96
N O R T H
A M E R I C A N
T R A D E
already managing their affairs in accordance with, a remarkably sophisticated EU-like system of supranational governance. Presiding over this system is the Free Trade Commission, a body comprised of cabinet-level officials or their designated representatives. Permanently headquartered in Mexico City, the commission operates as a North American analogue to the European Council of Ministers.4 Acting on the basis of consensus (unless otherwise agreed), the commissioners are charged with the formulation of policies affecting trade in manufactured goods, agricultural products, textiles, and financial services. The supranational body is also responsible for developing common rules of origin, customs procedures, transportation regulations, sanitary measures, and labeling standards. To assist the commissioners in performing these tasks, the NAFTA accord explicitly provided for numerous standing committees and subcommittees, as well as over a dozen working groups and advisory panels. Members of the commission, who supervise the day-to-day activities of all these bodies, have the power to reorganize them—or even, if they prefer, to create entirely new ones.5 NAFTA, then, does not simply call for the establishment of common rules, regulations, and standards; it also specifies how, procedurally, this harmonization effort is supposed to occur. In addition, NAFTA’s three signatories also established an overarching system for the settlement of intraregional trade and investment disputes. When invoked, this system calls for “panels of experts” to exercise what are, in essence, powers of judicial review. One might draw a useful (albeit imperfect) analogy here to the European Court of Justice, the determinations of which are understood to take precedence over the rulings issued by the national courts of individual EU member states.6 This part of the book uses NAFTA as an empirical touchstone for rethinking the rational foundations of international cooperation and supranational governance. While it is true that the regime’s three participants— Canada, the United States, and Mexico—have made bold strides toward the goal of freeing up the North American market, the implementation of their trilateral pact in 1994 has not been an unqualified success. The experience 4 Just as the Council of Ministers is assisted by the European Commission, the Free Trade Commission receives logistical support from a Secretariat. Indeed, the formal recognition of this permanent bureaucracy is one of the few institutional differences between NAFTA and the bilateral Canada–United States agreement it superseded; see Winham 1993, 257. 5 Weintraub 1994, 29; Winham 1993, 256. Operating alongside NAFTA’s umbrella Free Trade Commission are two separate commissions, one created to promote environmental cooperation, the other to promote labor cooperation. Located in Montreal and Dallas, respectively, these bodies were established by NAFTA’s side agreements; see Hufbauer and Schott 1993, 178–80, for an early assessment. 6 See, e.g., Burley and Mattli 1993; Garrett 1995; Weiler 1991. There are differences, to be sure. In cases concerning the imposition of antidumping or countervailing duties, for example, the task of the NAFTA panelists is to ascertain whether the trade-distorting practices in dispute are consistent with the alleged offender’s own domestic legislation, not (as would be the case in the EU) with the “constitution” embodied in the NAFTA charter.
J U M P - S T A R T I N G
T H E
B A N D W A G O N
97
of Mexico’s ruling party, a “loser” under the NAFTA regime, will be discussed in chapter 7. Rather than delve immediately into the loser’s side of the story, however, I will begin my account with an analysis of NAFTA’s two leading beneficiaries. I begin, in other words, with the story of the regime’s “enacting coalition,” the Canadian and U.S. sponsors of NAFTA’s institutional forebear, the Canada-U.S. Free Trade Agreement (FTA) of 1988. Many students of the earlier agreement trace its origins to the rising tide of protectionist sentiment that began sweeping across the United States in the early 1980s. On this account, the Canadian government’s interest in negotiating the 1988 pact was motivated by its concern that, without such an agreement, Canadian producers would find it ever more difficult to penetrate their traditional markets in the United States. If there is a “root cause” for the FTA, however, it surely resides more in the vagaries of Canadian domestic politics than in the U.S.-centered changes on which, curiously, so much of the previous literature has focused. Pursuing this line of analysis, I will argue below that, more than anything else, it was the Canadian general election of 1984 that finally tipped the scales in favor of the agreement. Although free trade was not a central issue in the election itself, the leadership of the Conservative party that won the contest had little stake in preserving the protectionist policies favored by the Liberal party it defeated. In short, the Canada-U.S. Free Trade Agreement would never have become a reality had it not been for a pro–free trade (Conservative) party’s success in wresting the Canadian polity away from a long succession of anti–free trade (Liberal) governments. The importance of the Conservatives’ electoral coup cannot be underestimated, for if the agreement between Canada and the United States had not become a reality, it is doubtful that the Mexican government would have taken the historically unprecedented step of requesting a trade deal of its own (see chapter 7). The Canada-U.S. accord is also implicated in the increasingly charged debate over whether NAFTA “violates” the sovereignty of its member states. Given that most of NAFTA’s controversial dispute settlement procedures were simply carried over from the earlier agreement—the institutional details of which were first ironed out in 1987—it is only natural to ask why the Canadian and U.S. negotiating teams who engineered that agreement decided upon the particular dispute resolution system they did, and indeed why they deemed it necessary to include such a system at all. It is with the aim of shedding new light on these issues that I now turn the analytical spotlight back to the earlier Canada-U.S. accord, leaving it to the next chapter to explain how—for better or worse—the resulting “shift in the status quo” precipitated the FTA’s metamorphosis into the larger continental agreement we see today.
Canada-U.S. Cooperation and the Transition to a New Status Quo Canada first informed the United States of its desire to relax bilateral trade and investment restrictions in March of 1985. By December of that year,
98
N O R T H
A M E R I C A N
T R A D E
Canada’s new head of government, Brian Mulroney, and his U.S. counterpart, Ronald Reagan, had formally committed themselves to establishing a bilateral free-trade zone, with Reagan requesting—and receiving—approval from the Congress to negotiate under the “fast-track” provisions of the 1974 Trade Act. This authorization gave the president two years to reach an agreement, at which point Congress had forty-five days either to accept the agreement in its entirety (i.e., without amendment) or to reject it. Bilateral negotiations commenced in Ottawa on May 21, 1986. The details of the negotiating process need not concern us here. Suffice it to say that the talks initially proceeded at a snail’s pace, the low point coming on September 23, 1987, when the chief Canadian negotiator, Simon Reisman, angered at the recalcitrance of his American partner, Peter Murphy, refused to return to the bargaining table. After higher-level officials in the two governments interceded, the talks resumed, culminating in a last-minute flurry of negotiations at the beginning of October. With an agreement finally in hand—the official signing ceremony took place on January 2, 1988—Reagan and Mulroney set about the task of getting their new pact enacted into law.7 In the United States, this process went smoothly, with both houses of Congress voting to ratify the FTA by overwhelming margins (366–40 in the House of Representatives, 83 –9 in the Senate). In Canada, by contrast, the agreement’s implementing legislation met with strong opposition in the upper house, where the governing Conservative party lacked a majority of the seats. Making use of the Senate’s powers of delay, John Turner, the leader of the Liberal party and an outspoken opponent of the accord, threatened to hold up the implementing legislation in committee unless Mulroney agreed to “let the people decide.” With assurances that the Senate would allow the legislation to go forward if the Conservatives retained their majority in the lower house, Mulroney agreed to Turner’s request, and a general election was held on November 21, 1988. Although, as we shall see, it was not a foregone conclusion that Mulroney’s party would prevail, in the end the Conservatives did hold onto their parliamentary majority, thus clearing the path for Canadian ratification of the accord on December 31, 1988. The CanadaU.S. Free Trade Agreement went into effect the very next day. The forging of a free trade pact between Canada and the United States was a stunning achievement. In terms of its coverage and overall economic impact, the free trade zone established by the agreement was rivaled at the time only by the EU’s internal market. Yet while the launching of the EU’s “1992” agenda has sparked a wide-ranging scholarly debate,8 there have been few systematic analyses of the FTA’s political and economic origins. What literature there is on the agreement focuses rather narrowly on the negotiations 7 For the text of the agreement, see Canada-U.S. Free Trade Agreement, H.R. Doc. No. 216, 100th Congress, 2d Session 2977 (1988). Useful overviews are provided in Schott 1988 and Steger 1988. 8 See, e.g., D. R. Cameron 1992a; Garrett 1992; Moravcsik 1991; and Sandholtz and Zysman 1989.
J U M P - S T A R T I N G
T H E
B A N D W A G O N
99
themselves (Campbell and Pal 1989; Doern and Tomlin 1991; Goldstein 1996; Hart 1994) or, taking the FTA’s existence as a given, assesses its impact from the perspective of specific countries or economic sectors (Morici 1990; Schott and Smith 1988).9 It may be that scholars have not studied the FTA in depth because the free trade area it established fell short of a common market. Unlike members of the EU, Canada and the United States continue to set their own tariffs; nor, with a few exceptions, are citizens from Canada allowed to seek work in the United States, or vice versa. Insofar as the FTA has escaped close scrutiny, however, I would argue that the reason is not the pact’s (relative) lack of importance so much as the fact that scholars fail to see its passage as in any way problematic or paradoxical. In most studies, the Canada-U.S. agreement is treated simply as the culmination of two nations’ long-standing efforts to overcome collective-action problems in “classic” neoliberal fashion.10 A closer look at the formation of the FTA provides a rather different picture.
From Free Rider to Free Trader: The Decisive Shift in Canadian Preferences Why did the United States and Canada take so long to permit free trade across their common border? The answer turns out to have relatively little to do with collective-action problems. On the contrary, a strong case can be made that, prior to the mid-1980s, the governments that presided over Canada actually preferred the existing (pre-free trade) bilateral arrangement. To be sure, Canada-U.S. relations during this period may have looked “suboptimal” to authorities in the United States. In Canada, however, it was only in the mid-1980s that the noncooperative status quo began to lose its appeal in the eyes of government officials. What changed? Canada-U.S. Relations and the Politics of “Fair Trade” Most analysts would emphasize the increasing threat of U.S. retaliation.11 Throughout the 1960s and 1970s, the United States showed extraordinary restraint in the face of aggressive Canadian screening of foreign direct investment, forced divestiture of foreign holdings in the Canadian energy sector, and numerous other protectionist barriers. As a result, bilateral relations during this earlier period were strongly biased in Canada’s favor.12 Busi9 There have also been a number of ideologically charged studies, chiefly by opponents of the pact; see, e.g., Hurtig 1991. 10 Cf. Yarbrough and Yarbrough 1992, 99 –103. 11 See Chapman 1988, 7; Doern and Tomlin 1991, 250. 12 Prior to the FTA, for example, Canadian tariffs averaged 9.9 percent, approximately triple U.S. rates. At least as important were other, subtler forms of Canadian discrimination, such as a 1973 law establishing a Foreign Investment Review Agency and requiring all American-owned enterprises in Canada to satisfy a number of conditions regarding export performance, the use
100
N O R T H
A M E R I C A N
T R A D E
nesses in the United States were hardly pleased about this, and allegations about Canadian exporters dumping their products in U.S. markets, receiving unfair government subsidies, or otherwise causing serious harm to American industries were common. Only rarely, however, were Canada’s “unfair traders” taken to task.13 The state of Canada-U.S. relations during the 1960s and 1970s was thus not one of mutual defection, with each nation refusing to open its markets to the other, but rather one of unilateral defection: the Canadians erected protectionist barriers while the Americans, for the most part, acquiesced. Why was Canada able to get away with a trade regime skewed against the interests of American industry? It is not as if the United States was incapable of retaliating against its northern neighbor. Because of Canada’s dependence upon trade in general—and upon U.S. trade in particular (see below)—the imposition of countervailing duties or other restrictions on access to the American market had the potential to do enormous damage to Canada’s domestic economy. That being the case, the United States was in an excellent position to demand that Canada open its markets. Until the 1970s, however, the United States was exceedingly reluctant to exploit Canada’s more vulnerable situation. After all, America’s trade policy was at that time still largely in the hands of the president, and deploying coercive trade sanctions against surplus countries (like Canada) was simply not in the president’s interest.14 Why not? The most obvious explanation is of Canadian inputs (i.e., domestic content), and other criteria. In addition, Canada “cleverly manipulated auto production and trade in its favor” (Gandhi 1990, 430 – 31). In this case, the chief culprit was the 1965 Canada-U.S. Auto Pact. Under the pact, American producers who wanted duty-free access to the Canadian automotive market had to agree to assemble one car in Canada for every car that they sold there. Canadian car manufacturing (as opposed to car assembly) also benefited from the pact, which stipulated that at least 60 percent of the value of the automobiles being assembled in Canada be of Canadian origin. Given the enormous size of Canada-U.S. automobile trade—trade in cars and car parts comprised approximately one-third of all bilateral merchandise trade—the economic costs of these measures to the United States were substantial; see Johnson 1993, and R. Wonnacott 1991c. 13 Rugman and Anderson 1987. Throughout this period, a U.S. business seeking to retaliate against a Canadian competitor’s (ostensibly) unfair trading practices faced an uphill climb. Although it was free to file antidumping cases against Canadian companies suspected of pricing their products in the U.S. market at “less than fair value,” the department in charge of assessing such claims was the Treasury, a stronghold of the presidency that, for this reason (see below), was wary of undertaking diplomatically sensitive micro-interventions on behalf of individual import-competing firms. Likewise, an American firm that believed its Canadian competitors were receiving unfair subsidies could ask the U.S. government to impose countervailing duties sufficient to compensate for the artificial difference in price, but here, too, the Treasury had the last word in deciding what would qualify as unfair foreign government assistance. See Robert Baldwin 1985, chaps. 2 and 3. 14 By contrast, members of Congress were quite prepared to retaliate against countries perceived to be free riding. Until the 1970s, however, fair traders in Congress were hamstrung by earlier laws granting the president enormous discretionary authority over the formulation and execution of U.S. trade policy. See Haggard 1988; O’Halloran 1994; R. Pastor 1980; and Schattschneider 1935.
J U M P - S T A R T I N G
T H E
B A N D W A G O N
101
geopolitical: Trade restrictions would have undermined the economies of America’s Cold War allies.15 Then, too, presidents also had strong domestic incentives for going easy on foreign protectionists. By sanctioning U.S. trading partners, the individuals who held the office during the 1960s and early 1970s knew they ran some risk of triggering an all-out trade war, and thus of disrupting economic growth and prosperity at home. So long as U.S. trade policy remained at the discretion of the president, retaliatory threats against Canadian infractions lacked credibility. With the onset of hard times in the early 1970s, this state of affairs began to change. Reeling from the effects of the 1973 oil crisis, many U.S. citizens began to view protectionism as a legitimate tool for stemming the continuing flood of cheap foreign imports and thereby “leveling the playing field.” As public support for a get-tough approach mounted, Congress attacked the Ford and Carter administrations for making too little use of America’s powerful arsenal of trade sanctions. Whether or not applying pressure on surplus countries was an economically appropriate means of reversing the (at that time) burgeoning U.S. trade deficit, members of Congress seized upon the idea for political reasons and, after winning passage of two sweeping pieces of trade legislation, began taking matters into their own hands.16 With members of Congress increasingly in the driver’s seat of U.S. policy, it could be argued that Canadian authorities had little choice but to reorient their basic economic priorities in the direction of greater openness. This line of argument was given its most coherent public statement in a report issued by a special Royal Commission in the summer of 1985.17 In justifying its endorsement of closer bilateral ties, the commission emphasized the specter of U.S. protectionism. “Only by negotiating a free trade deal with the United States,” wrote the commission chair Donald S. Macdonald, “can we assure [Canada’s] future as a trading nation. Without free trade, we will be contin15
Robert Baldwin 1976; Verdier 1994. See Bhagwati and Patrick 1990; Destler 1992; and Goldstein 1988 and 1993. The 1970s saw two landmark trade bills enacted into law, one in 1974 and another in 1979. Passage of the second law was particularly significant, for, among other things, the 1979 Trade Act shifted responsibility for determining whether foreign competitors were dumping their products or benefiting from unfair subsidies to the International Trade Administration of the Department of Commerce. In comparison with the laissez-faire-minded Treasury, Commerce was more likely to be receptive to congressional demands for contingent protection (Goldstein 1996; see also Robert Baldwin 1985, chap. 2). The 1979 law also introduced a new injury test—to be decided by the U.S. International Trade Commission, not the Treasury—for some countervailing duty cases, where injury was defined as “harm which is not inconsequential, immaterial or unimportant” rather than in the more demanding terms that had been used for escape clause cases (“a substantial cause of serious injury”). 17 Established in 1982, the commission was initially charged with helping the then-governing Liberal party develop a long-term strategy for addressing Canada’s regional and economic problems. Before its work was complete, however, the general election of 1984 intervened, and the Liberals lost their parliamentary majority. Its partisan mandate rendered obsolete, the commission nonetheless continued to function, finally releasing its recommendations (Canada 1985) amid great fanfare in 1985. 16
102
N O R T H
A M E R I C A N
T R A D E
Table 6-1 U.S. Antidumping and Countervailing Duty Investigations Initiated against Canada, 1979–1986 Antidumping Cases
1979 1980 1981 1982 1983 1984 1985 1986 Total
Countervailing Duty Cases
Total
Against Canada
Total
Against Canada
26 21 15 65 46 74 66 71
3 2 1 2 2 3 4 3
40 14 22 140 22 51 43 27
1 0 2 4 0 1 3 2
507
20
392
13
Source: Compiled from Destler 1992, Apps. B and C.
ually harassed by protectionist measures against our steel, fish, lumber, and, next perhaps, our metals and minerals.”18 Extrapolating from this line of reasoning, one might conclude that U.S. “harassment” is what ultimately prompted Canada to seek a free trade agreement with its southern neighbor. In reality, however, Canada-U.S. relations during the late 1970s and early 1980s were not nearly as hostile as the authors of the Macdonald report made them out to be. For example, Table 6-1 reveals that over the seven-year period between the passage of the 1979 Trade Act and the commencement of bilateral negotiations in 1986, the United States filed a grand total of only twenty antidumping and thirteen countervailing duty cases against Canadian concerns. Nor were these large concerns: the principal targets were Canadian producers of softwood lumber—an industry that contributed barely more than 5 percent of Canada’s total export earnings.19 Political Developments within Canada: The Neglected Side of the Story While the specter of U.S. protectionism may have played some role in Canada’s mid-1980s free trade conversion, it was hardly the major impetus. Far more important was a purely internal development: the election in 1984 of the first Conservative government since 1962 to command a majority of seats in the Canadian House of Commons. To be sure, the fact that the Conservative party’s stance toward the United States was less confrontational than that of the Liberals was not the decisive factor in the Conservative victory. Of greater importance were the internal divisions that emerged within the governing Liberal party during the run-up to the 1984 election, as well 18 19
Quoted in McQueen 1985, 6. Hart 1990, 6.
J U M P - S T A R T I N G
T H E
B A N D W A G O N
103
as the personal unpopularity of John Turner, the Liberal leader who succeeded Trudeau as prime minister less than a year before Canadians went to the polls. In short, one should be cautious about reading into the 1984 electoral results a popular mandate for free trade. But while it may not have been popular with all Canadians, the argument for free trade did resonate with the new Conservative government’s core political constituencies. Large internationally competitive Canadian corporations were one such Conservative stronghold, one that (not surprisingly) wanted the new government to move rapidly ahead with the free trade initiatives spelled out in the Macdonald report.20 The Conservatives also had a stake in satisfying the demands of their new support base in Quebec. In 1984, the party nearly swept the province, picking up 58 of its 75 total seats (compared with only 1 of 75 in the previous election). Although some of the party’s success among residents of Quebec may have been due to the fact that Mulroney, himself a bilingual native of the province, convinced a number of Quebec nationalists to run as Conservatives, the Conservative party’s softer position on relations with the United States also appears to have been an important factor. Perceiving (correctly) that most Canadians outside Quebec were opposed to closer ties with the United States, many Quebecois may have regarded their vote for the pro–free trade Conservatives as yet another means of asserting their independence from the rest of the country.21 And indeed—the Macdonald report notwithstanding—much of the Canadian public did oppose closer ties, viewing the proposed trade pact as a blatant “sellout” to the United States.22 From a strategic standpoint, however, the lack of broad-based public support for the forging of closer bilateral ties was counterbalanced by several other factors. One was the size of the Conservatives’ electoral cushion. Mulroney’s landslide victory in 1984 had given his party an enormous legislative majority—almost three-quarters of the seats (211 out of 282) in the Canadian parliament. True, relatively few voters in Canada hold long-term, ideologically rooted attachments to particular parties, as evidenced by their predilection for voting for different parties at the federal and provincial levels of government, and for turning out once-popular incumbents.23 Inasmuch as late-breaking events, leadership 20 On the pro–free trade consensus among Canadian big business, see McQuaid 1991. Canadian corporations supportive of the free trade pact were represented by the Business Council on National Issues (an umbrella organization composed of over 150 of the largest corporations operating in the Canadian market), the Canadian Alliance for Trade and Job Opportunities, and numerous other lobbying groups and trade associations, including the Canadian Federation of Independent Businesses, the Canadian Chamber of Commerce, and the Canadian Manufacturer’s Association. 21 Russell 1992, 144; see also Gaines 1994. 22 Public opinion polls conducted throughout the mid-1980s consistently found majorities opposed to greater liberalization. Even in Quebec, only 41 percent of those polled in a Gallup survey taken in the run-up to the “free trade election” of 1988 expressed support for the proposed FTA, and the percentages in Ontario (30 percent), British Columbia (32 percent), and the country as a whole (34 percent) were lower still. Toronto Star, November 19, 1988. 23 Canadian voters unseated five of twelve federal governments, including two led by Diefen-
104
N O R T H
A M E R I C A N
T R A D E
changes, and the emergence of new issues had a greater potential to influence electoral outcomes in Canada than would have been the case in other advanced democracies, Mulroney’s electoral constraints may have been somewhat tighter than his large majority in the House of Commons made it appear. There is, however, another side to the electoral volatility argument. The fact that dramatic swings in voter sentiment were commonplace in Canadian politics meant that Mulroney did not need to be as concerned about the poor reception with which his free trade overture to the United States had initially been greeted by the Canadian electorate: he could more easily afford to take risks, pursuing policies consistent with his partisan and ideological interests (but opposed by a majority of the population at large) in the expectation that voter sentiment would turn in his favor by the next election. And in 1985, when Mulroney first announced his intention to pursue a free trade accord with the United States, it was not at all farfetched to suppose that a pro–free trade party would be able to win reelection a few years down the road. As one study notes, “Pursuit of an agreement at least gave the government the appearance of purpose and political courage; these were electoral assets in their own right.”24 As for free trade’s drawbacks—its inevitable economic dislocations, the increased dependence on market conditions in the United States, and the like—the Conservative government went to great lengths to avoid broaching these in public. On the contrary, fearing that the Canadian electorate might be put off if the terms of the proposed agreement were to become widely known, the prime minister tended to steer clear of the free trade issue altogether.25 Of course, this did not stop other party leaders from broaching the issue. Doing so, however, did not necessarily work to their advantage, since neither of Canada’s two anti–free trade parties—the Liberals and the New Democrats—were likely to garner all of the country’s anti–free trade vote. In the event of a “free trade” election, this division would play into Mulroney’s hands: as the sole recipient of the pro–free trade vote, his party would have an opportunity to capture seats even in majority anti–free trade districts that would otherwise have gone to one of the country’s two (anti– free trade) opposition parties.26
baker and Trudeau, between 1957 and 1993. Two of the remaining seven governments, while able to remain in office, were unable to obtain a majority of seats and continued in office as minority governments. A further indication of this volatility is the fact that nearly one-fifth of all Canadian voters did not decide which party to support until the last week of the election campaign. LeDuc 1991, 357. 24 Johnston et al. 1993, 241. 25 An internal memorandum—subsequently leaked to the press (see Duncan Cameron 1986, 3–10)—had explicitly called for this strategy, urging Mulroney to distance himself as far as possible from the FTA while at the same time directing the public’s attention toward his government’s positive accomplishments: his achievements in housing policy, his concern for the environment, and so on. 26 Johnston et al. 1992.
J U M P - S T A R T I N G
T H E
B A N D W A G O N
105
By the mid-1980s, the collective gains afforded by liberalizing trade across the Canada-U.S. border were thus being touted not only by the U.S. administration but also, at long last, by its (new) Canadian counterpart. It is now time to consider why the resulting trade pact—the Canada-U.S. Free Trade Agreement of 1988—assumed the particular form it did.
Explaining the FTA’s Supranational Architecture Among the 1988 agreement’s many important innovations were the application of the “national treatment” principle to many sectors of Canada-U.S. trade and the lowering of the public sector procurement threshold below which bids tendered by contractors in one country had to be accepted by authorities in the other.27 The FTA also expressly prohibited Canada from requiring U.S. investors to meet specific performance targets.28 To many observers, however, these substantive provisions were less important than the purely procedural provisions outlined in Chapters 18 and 19 of the agreement. Taken together, these chapters laid the foundations for a new supranational governance structure whose dispute resolution machinery could be brought into play in areas where Canada and the United States were likely to read the terms of the FTA differently. With respect to Canadian complaints concerning the application of U.S. antidumping and countervailing duty law, for example, Chapter 19 of the FTA called for the Canadian government—or, upon petition, individual Canadian firms—to take their disputes to a rotating “panel of experts” made up of two Canadians, two Americans, and one person chosen jointly.29 Chapter 18 of the agreement established a separate supranational body— the Canada–United States Commission—comprised of cabinet-level officials (or their designates) from each side. In the event of a dispute over broader issues related to the functioning, interpretation, and future evolution of the agreement, this body was given thirty days to arrive at a mutually acceptable solution. Should this prove impossible, the matter was to be referred to “binding arbitration.” Using a similar process to the selection of the Chapter 19 panels (but with individuals chosen from a separate roster), the commission would empower a five-member panel to ascertain whether new leg27 In deference to Canadian concerns, the FTA did not extend national treatment—that is, the requirement that Canada treat goods, services, and investors from the United States no less favorably than its own goods, services, and investors—to transportation, health care, or cultural industries (e.g., the national media and film industries). Carried over into NAFTA, the cultural exemption was only partial, however, and an American purveyor of culture facing discrimination could retaliate with measures of equivalent commercial effect. 28 Many of the these (and other) provisions were also incorporated into NAFTA and will be discussed in that context. See chapter 7. 29 All panelists were to be drawn from a fifty-person roster. By stipulation, the twenty-five people nominated by each country had to include a majority of lawyers. Chairs were to be chosen from among the lawyers on each panel by majority vote or, in the absence of a majority, by lot.
106
N O R T H
A M E R I C A N
T R A D E
islation being considered in either country constituted a violation of the agreement. Once formed, this panel would meet in secret and set its own procedures, holding at least one hearing and permitting each side to argue its case orally and in writing. Both sides would then be obliged, by law, to accept the panelists’ recommendation.30 So why, then, did the FTA’s U.S. and Canadian “enactors” find it in their interests to devise such an elaborate system of supranational governance? And just what were their (institutional) interests, anyway? The Efficiency Rationale One possibility is that the free trade enthusiasts who crafted the FTA’s dispute resolution mechanism viewed it as a tool for reducing transaction costs and mitigating problems of collective action. While the emergence of a pro–free trade consensus among political elites in Canada and the United States may have been a necessary condition for trade liberalization, it was not (so this argument would go) a sufficient one. For however great were the mutual benefits of reduced trade barriers, the short-term or domestic political benefits for each country of free riding—that is, of gaining unhampered access to the markets of its trading partner and then raising protectionist barriers—would have been greater still. As rational actors, the two countries participating in the arrangement would thus have retained, or perhaps even raised, their existing trade restrictions, yielding a worse outcome than would have been possible if both countries had simply cooperated with each other instead.31 Take the Canadians, for example. Although Canada’s Conservative government was itself a strong supporter of free trade, its leaders might have questioned whether the U.S. Congress would be able to resist the temptation to enact new protectionist measures at some point in the future. And if Congress ever did succumb, the costs to Canada of being “suckered” would have been substantial, since the imposition of new restrictions on Canadian access to American markets would have forced a large segment of the Canadian manufacturing sector to reorient its production—potentially at great expense in terms of short-run economic dislocations—toward new overseas markets. It is true that members of the U.S. Congress, when deciding whether to “free trade” or “free ride,” would have had to take into account the possibility of Canada’s retaliating in some future round of the U.S.-Canada trade game.32 Then again, if Congress (or any other potential FTA noncooperator) knew that it could escape detection—that it could cheat on the agreement and still, in all likelihood, get away with it—the temptation to defect might well have been irresistible. Enter institutions. It was precisely to reduce this likelihood that the framers of the FTA integrated a formal dispute resolution system into their agree30 For a detailed account of the FTA’s dispute resolution procedures, see Horlick, Oliver, and Steger 1988. 31 Cf. Conybeare 1984. 32 Cf. Axelrod and Keohane 1985; Magee, Brock, and Young 1989.
J U M P - S T A R T I N G
T H E
B A N D W A G O N
107
ment—or so, at any rate, the logic of this argument would suggest. True, even before the inauguration of the FTA, it was quite easy for interested parties in Canada and the United States to obtain information about their neighbor’s trade practices. But while the task of monitoring each side’s behavior may not have necessitated the establishment of a formal institutional structure, the task of interpreting each side’s behavior—of determining whether a defection had or had not occurred—did. Why? Because the terms of the FTA were ambiguous, as indeed they had to be, given that ambiguity is an inherent feature of all contracts, international and domestic alike. Cognizant of the problems created by the (unavoidable) incompleteness of their agreement and yet not wishing to forfeit the collective benefits of free trade, the negotiators for Prime Minister Mulroney and President Reagan devoted their energies to ironing out a mutually agreeable set of conflict resolution procedures ahead of time. This—or so, again, the argument would run—is something they would never have done had they been able to draft a comprehensive set of rules spelling out appropriate behavior for every situation that might possibly arise over the course of their two countries’ future interactions. As for why their new adjudication system afforded such a large role to “outside” (i.e., nongovernmental) agents, perhaps this too was motivated by efficiency considerations: better to delegate the task of resolving future conflicts over the interpretation and application of the FTA to a set of more or less impartial, quasi-independent institutional intermediaries than to endure the delays and inefficiencies associated with adjudication through direct intergovernmental bargaining. A Closer Look What explains the FTA’s institutional centerpiece? The answer given above starts with the assumption that the original parties to the agreement—Mulroney and Reagan—were incapable of formalizing the terms of their new relationship in a precisely worded treaty. That being the case, each stood to benefit from the creation of a wholly new authority structure, for if it were clear that disputes over matters of treaty interpretation would be handled by a set of politically independent mediators, free riders from either country would find it that much more difficult to pass themselves off as good-faith cooperators. No longer would they be able to justify their actions, and thus escape retaliatory measures, by claiming that the bilateral accord was unclear or ambiguous about what “free trade” in a particular context really meant. So, how does this explanation stack up against the actual facts of the case? Are the kinds of efficiency-based considerations emphasized in the contracting literature what really drove the institutional choices of the FTA’s U.S. and Canadian framers? There are at least two reasons to be skeptical. First, the dispute resolution system these framers adopted was less conducive to the interests of free traders in one of the two countries—the United States—than was the previous dispute resolution system that it replaced. Second, the Canadian and U.S. negotiating teams chose not to incorporate a subsidies
108
N O R T H
A M E R I C A N
T R A D E
code directly into their agreement, even though doing so would have obviated many of the “incompleteness” problems we have just been discussing. Let us take each of these points in turn. WHY WOULD THE UNITED STATES WANT TO REPLACE ITS EXISTING DISPUTE SETTLEMENT MACHINERY?
Under the earlier (pre-FTA) system for resolving U.S.-Canadian trade disputes, Canadian complainants had been required to seek redress through the American legal system. Ultimately, the task of determining whether or not U.S. antidumping and countervailing duty actions were “unfair”—as American industry’s Canadian competitors frequently alleged—fell to the U.S. Court of International Trade, which in turn answered to the U.S. Court of Appeals for the Federal Circuit. Neither Mulroney nor Reagan sought to perpetuate this well-established arrangement. In Mulroney’s case this reluctance is not surprising, as the FTA’s new system for resolving U.S.-Canadian trade disputes was considerably less biased in favor of U.S. industry. What is curious, at least if one thinks of institutional outcomes as arising from a bargaining process in which each side tries to improve upon its original position, is Reagan’s “caving in” to Canadian pressure. For though his Canadian counterpart may have had a preference for a more neutral dispute system—one in which U.S. courts would no longer be allowed to serve as the sole judge and jury—the United States enjoyed an overwhelming bargaining advantage throughout the FTA negotiations. This advantage derived from at least three distinct sources. First, Canada initiated the trade talks, thereby signaling its greater stake in reaching agreement.33 If we want to explain why Reagan was in a better position to extract concessions from Mulroney than vice versa, the fact that the Canadian head of state initiated the negotiations is not, however, the first place we should look. At least as consequential was the earlier-noted shift within the United States in the balance of power on trade-related issues, a shift that meant the president could credibly claim his hands were tied.34 Although protectionist forces were not yet dominant within the U.S. Congress, their eventual triumph could certainly not be ruled out. One ominous sign was the near defeat of Reagan’s request for fast-track authorization in 1986.35 The swelling 33
See, e.g., D. R. Cameron 1988. The same “two-level games” argument (cf. Milner 1997; Putnam 1988) can be used to explain Germany’s success in negotiations over the structure of the European Monetary System: by tying the hands of the German government, the Bundesbank’s unilateral offensive against price instability during the mid-1970s made it easier, not harder, for the German chancellor to obtain concessions from his European counterparts. See chapter 9. 35 On April 23, 1986, the Senate Finance Committee, though it was controlled by the president’s own party, rejected a resolution of disapproval by the narrowest of margins: a 10 –10 tie. It did so, moreover, only after two pivotal members of the committee, Senators Steven 34
J U M P - S T A R T I N G
T H E
B A N D W A G O N
109
of protectionist sentiment within the U.S. legislature was also evident in the early drafts of the “omnibus” trade bill that had begun wending its way through the House and Senate in 1985.36 In addition to benefiting from Congress’s increasingly aggressive stance toward Canada, the president also had the advantage of presiding over an economy that could withstand even a major reduction in bilateral trade. As noted above, Canada was far more dependent upon trade with the United States than the other way around. Indeed, though Canada was the largest foreign market for American-produced goods, accounting for just under a quarter of total American merchandise exports during the 1980s, Canada sent nearly three-quarters of its total exports to the United States (Figure 61). Adding to Canada’s vulnerability was its greater reliance on trade in general. Whereas changes in the global competitiveness of U.S. products had relatively little impact on the performance of the overall U.S. economy, this was not the case for Canada. In 1987, exports to the United States alone comprised 17.3 percent, or nearly a fifth, of Canadian gross domestic product. By contrast, American exports to Canada contributed a mere 1.3 percent of U.S. GDP (Figure 6-2). This basic asymmetry was a third factor contributing to the American side’s strong bargaining position. Because the failure to strike a deal would have produced greater political problems on the Canadian side of the Canada-U.S. border, the Americans enjoyed greater “exit options” in the negotiations.37 In turn, the fact that the American negotiators could hold out longer for their most preferred outcome meant that if the chief negotiator for the United States, Peter Murphy, had sought to perpetuate the United States’ existing dispute resolution machinery, his Canadian counterpart, Simon Reisman, would almost certainly have acquiesced. WHY NO SUBSIDIES CODE?
Casting further doubt on the FTA structure’s efficiency rationale is the fact that the Canadian and American negotiating teams refused to incorporate a subsidies code directly into the text of their new agreement.38 By clarifying Symms (R-Idaho) and Robert Packwood (R-Oregon), received assurances that the president would renew his efforts to resolve a long-standing dispute over Canadian lumber subsidies (O’Halloran 1994). Although Congress ultimately acquiesced in granting Reagan’s request, the creation of a free trade area was clearly not a top priority for many U.S. senators. 36 It took a full three years—until 1988—for the bill to be enacted, by which time many of the legislation’s more egregious protectionist provisions (e.g., a stipulation that would have required the executive to retaliate against U.S. trading partners manifesting “excessive” trade surpluses) had been watered down or dropped entirely. 37 Cf. Hirschman 1970. 38 Rather than introducing a subsidies code into the FTA itself, Mulroney and Reagan weakly committed themselves to negotiating a new system of rules on government subsidization within some seven years after their agreement entered into force. NAFTA dropped even the aspiration of ironing out such a code or, for that matter, of negotiating a new set of rules on unfair pricing. See Huntington 1993, 440; Lipsey, Schwanen, and Wonnacott 1994, 130.
Figure 6-1. Geographic Distribution of U.S. and Canadian Trade, 1987
Figure 6-2. Trade as a Share of U.S. and Canadian GDP, 1987
112
N O R T H
A M E R I C A N
T R A D E
from the outset exactly which sorts of subsidies would be deemed unacceptable, the formulation of a new system of rules setting forth detailed criteria for determining when exporters in either country could (and could not) file countervailing duties against their competitors would have prevented many disputes from arising in the first place. And yet at no point was the idea of introducing a subsidies code into the agreement ever seriously entertained by either the U.S. or the Canadian negotiators.39 Let me suggest a simple explanation. The reason the negotiators left the terms of their agreement incomplete is that they knew any subsidies code acceptable to the U.S. Congress would have placed severe constraints on the ability of future Canadian governments to employ subsidies for the purposes of regional development and social policy. To be sure, this would not have been a great loss to Canada’s present government.40 But it would have been a loss to that government’s most likely successor, for, as we will see in a moment, the Conservative government’s Liberal opposition was an ardent defender of precisely those kinds of interventionist policies that a subsidies code (particularly one engineered to meet with approval in the U.S. Congress) would have prohibited. This is why the United States did not present Canada with a complete, or even a partial, subsidies code. By choosing instead to delegate the task of reviewing Canadian subsidies and pricing practices to a supranational mediator, U.S. officials were deliberately seeking to reduce the burden of FTA compliance for the accord’s Canadian losers. The United States conceded as much as it did, then, not because it hoped to entice the Conservatives into signing a deal—Mulroney’s acceptance of the U.S. offer was a foregone conclusion—but because it wanted to ensure that the FTA would be able to survive a possible return to power of the Conservatives’ anti–free trade opposition.41 39 Nor did the two sides consider formulating a bilateral competition (i.e., antitrust) policy, though doing so would have gone a long way toward reducing the impediments to bilateral trade caused by the seemingly arbitrary application of U.S. antidumping law. Another solution to this problem—and another road not taken—would have been for the United States to extend the principle of national treatment to its domestic antitrust and competition laws. See, e.g., Trebilcock and York 1990. 40 By the time the negotiations had gotten underway, the Conservative government had already begun opening Canada’s vast energy sector to foreign ownership, easing the Auto Pact’s restrictions on Canadian content, discontinuing the practice of subsidizing Canadian exports of softwood lumber, and phasing out Canada’s special review procedures for foreign investment. With respect to the latter, the Mulroney government renamed the aforementioned Foreign Investment Review Agency “Investment Canada” and enlarged its mandate to include attracting (not just screening) foreign capital. 41 Although the focus here is on the choice of institutional structure, the same motivations may also have driven other aspects of treaty design, including the decision to allow the FTA to take effect gradually, over a nine-year period, rather than (as the pro–free trade architects of the agreement might themselves have preferred) all at once. The effect of this decision was to smooth the process of adjustment, thereby moderating the agreement’s adverse consequences for workers employed in “sensitive” sectors of the Canadian economy, the vast majority of
J U M P - S T A R T I N G
T H E
B A N D W A G O N
113
This is not to suggest that the United States, by forgoing a subsidies code, sought to eliminate the FTA’s burden on future Canadian governments. If completely eliminating the agreement’s costs had been the goal of U.S. policy, officials from the Canadian government would have been accorded a more direct role in the dispute settlement process.42 The fact that the dispute settlement provisions in the agreement were designed to operate outside normal government channels (and, ideally from the U.S. standpoint, to take on a life of their own) reflected the Americans’ interest—shared by Canadian Conservatives—in limiting a future Liberal government’s room for maneuver. In sum, neither Reagan nor Mulroney had reason to fear that their accord’s built-in adjudication mechanism would drop a future Liberal government’s costs all the way down to zero. Nor, for reasons I discuss below, did the FTA’s two framers need to go that far to achieve their goal of insulating bilateral openness. For both knew that once it was up and running, the new bilateral regime would be extraordinarily difficult to abandon, even for a Canadian government whose leaders had strongly opposed the signing of the original agreement. MAKING THE DEAL STICK: OTHER STRATEGIES FOR LOCKING IN THE “GAINS” FROM LIBERALIZATION
Over time, international agreements have a way of acquiring their own independent constituencies. Much of this is a product of simple inertia: the longer a given treaty or regime has been in existence, the greater, in most cases, are the costs of overturning it.43 That said, Mulroney and Reagan had an obvious stake in accelerating this “natural” constituency-building process. Moderating their free trade accord’s burden on its opponents was one strategy for accomplishing this objective, but there were others as well. One was the interweaving of numerous sectoral agreements into a single, overarching treaty. By linking separate agreements into a “package deal,” the prime movers behind the FTA ensured that its Canadian beneficiaries would include a wide cross-section—though not necessarily a majority—of Canada’s population. The FTA’s enacting coalition also stipulated that the whom were supporters of the opposition Liberal party. 42 The process itself might also have been “stronger,” in the sense of requiring the United States to immediately discontinue its FTA-illegal trade practices in the not unlikely event that it should find itself on the losing side in a dispute. As the system was designed, however, a Chapter 19 panel that determined a particular U.S. antidumping or countervailing duty action to be in violation of the agreement was authorized only to send the case back to the relevant U.S. agency and request that it be reconsidered (Moyer 1993, 718 –19). And although the dispute resolution mechanism set forth in Chapter 18 of the FTA (i.e., for disputes on broader questions concerning the functioning, interpretation, and evolution of the agreement) was termed “binding,” here, too, the Americans were under no formal obligation to comply with an arbitration panel’s recommended settlement. The architects of the FTA explicitly permitted the United States to ignore the panelists if it deemed its “fundamental rights” to be at risk. 43 Cf. Keohane 1984 and Stein 1983.
114
N O R T H
A M E R I C A N
T R A D E
accord could not be amended in piecemeal fashion, thereby making it difficult for FTA opponents to simply nullify those parts of the agreement they disliked. Short of launching an entirely new round of negotiations, a newly empowered Liberal government would be left with a choice between abrogating the accord (including those provisions of which they approved) or taking no action at all.44 It is also significant that Mulroney and Reagan chose to label the product of their collaborative effort a “free trade” pact. To members of the global financial community, free trade is something of a mantra. Just as the French and German prime movers behind the European Monetary System sought to make participation in that regime a litmus test of a government’s commitment to fighting inflation (see chapter 8), the U.S. and Canadian beneficiaries of the FTA presented ratification of the accord as a testimony to their countries’ abiding faith in the free market.45 Henceforth, any derogation of the pact would take on a larger symbolic significance. Finally, the FTA’s enacting coalition was inclined to launch the new regime as quickly as possible (even if prematurely), for the sooner it came into existence, the sooner it would begin shaping the expectations of private actors—groups and individuals who could be relied upon to object vociferously to any efforts by a future set of officeholders to reinstate the pre-FTA status quo. The Importance of Entrenching Canada’s New Commitment to Bilateral Openness If the sponsors of the FTA had known that future Canadian governments would share the Conservatives’ interest in dismantling the protectionist legacy of their Liberal predecessors, much of this would have been unnecessary. But the beneficiaries of the FTA did not know this. In fact, free traders 44 Interestingly, the Uruguay Round of the GATT exhibited many of the same all-or-nothing characteristics. What explains this departure from the earlier “GATT à la carte” arrangement? No doubt the member states were motivated, in part, by the goal of achieving greater simplicity. In addition, however, the negotiators from the United States and EU may also have been hoping to deter future power holders from opting out of “their” treaty, the conclusion of which they had worked so hard to bring about. Prior to the Uruguay Round, members of the GATT had been able to pick and choose among different GATT codes. Now that these side texts have been integrated into a single legal structure—one to which, with very few exceptions (see Jackson 1994), all members of the regime must adhere in its entirety—a government that believes itself to be suffering under the burden of a particular code may decide to comply with it anyway. As of 1995, the only alternative would be for the complainant to take the rather extreme step of disaffiliating entirely from the GATT and, in so doing, forfeiting any offsetting concessions it might have won in other codes. 45 Technically, of course, the FTA did not establish “true” free trade between Canada and the United States. The agreement was silent on intellectual property, for example, and it exempted the entire transportation sector. And even with the incorporation of the dispute resolution procedures outlined in the agreement, the FTA did not entirely foreclose the possibility that American exporters would continue creatively applying U.S. trade remedy legislation as a way of artificially raising the price of Canadian exports.
J U M P - S T A R T I N G
T H E
B A N D W A G O N
115
on both sides of the Canada-U.S. border had deep concerns about the domestic political situation in Canada. Although by the mid-1980s the process of liberalizing the Canadian economy was well underway, the Conservative government’s public approval ratings had declined precipitously during their first years in office, raising doubts about whether the government would be able to survive the next election.46 As Mulroney’s political fortunes waned, it seemed only a matter of time before Canada’s Liberal party reclaimed its position of power. From the standpoint of free traders, this was not an appealing scenario. Historically, the practice of using selective policy interventions to exclude foreign goods from Canadian markets had been popular among Liberals for much of the party’s existence. Upon winning the election of 1867 and taking office as the country’s first prime minister, Sir John A. Macdonald (a Conservative) launched a program calling for heavy government involvement in the building of a transcontinental railway, free access to western lands, and a steep tariff wall to protect fledgling Canadian industries from U.S. competitors. Macdonald hoped this “National Policy,” when coupled with his practice of using patronage appointments to broker compromises between competing wings of his party, would help prevent north-south trade from undermining the development of a unified nation. In the elections of 1878 and 1891, the Liberal party attempted to defeat the Conservatives with the promise of continental free trade. Each time, however, the Conservatives prevailed. In 1896 the Liberal party finally unseated the Conservatives, but by then it, too, had come to see high tariff barriers and patronage politics as conducive to nation building and electoral success. Any lingering doubts about the National Policy’s popularity with voters were put to rest in 1911. In that year, the Liberals—then under the leadership of Sir Wilfrid Laurier—suffered a scarring defeat after fighting an election on a platform calling for “unrestricted reciprocity” (i.e., free trade) with the United States. The Liberal party returned to power in 1920, governing more or less continuously until the 1980s. Except for a few sporadic, behindthe-scenes initiatives in the 1930s and 1940s, however, it never again sought to reverse Canada’s protectionist legacy—that is, until it took Canada into NAFTA in 1993 (see chapter 7).47 Well into the 1980s, there continued to be good reasons for the Liberals to oppose free trade. Whereas many Americans equate free trade axiomatically with equal opportunity and prosperity, Canadian citizens continued to be deeply ambivalent about the issue, and Canada’s close ties with the United States were routinely blamed for everything from the poor health of the 46 Although Gallup forecasted a Conservative victory in the 1988 election, most pollsters were predicting a much closer election right up until the end, and few were prepared to rule out the possibility that neither the Liberals nor the Conservatives would muster a majority. See, e.g., Campbell and Pal 1989, 345. 47 The enduring impact of the National Policy on U.S.-Canadian trade relations is discussed in Simpson 1988 and White 1988.
116
N O R T H
A M E R I C A N
T R A D E
Canadian economy to the country’s declining sense of national identity.48 In addition, though many respected analysts believed that the free trade agreement would improve Canada’s economic performance in the long run, few doubted that Canadians employed in sheltered sectors—particularly auto workers and farmers (both Liberal strongholds)—would be forced to endure high adjustment costs in the short term.49 Thus, while the long-run macroeconomic consequences of the movement from protectionism to free trade may have been favorable, its political consequences for the Liberal party ran in precisely the opposite direction. Nor were there any illusions among the party’s elite that persuading voters with relatively short time horizons to support policies whose benefits were likely to accrue only in the long term would be easy. To the contrary, critics of the FTA had largely succeeded in convincing Canadian citizens that the lower costs of doing business in the United States—a legacy of its weaker unions and less generous social policies—would induce Canadian firms to relocate south of the border, and that free trade would “either force Canadian workers to accept lower wages in order to save their jobs or so gut Canada’s manufacturing sector that the country would have to rely even more heavily on exports of natural resources to sustain its standard of living, thereby entrenching its role as the ‘hewer of wood and drawer of water’ for the United States and thwarting future economic development.”50 Among the staunchest opponents of the free trade accord were residents of Ontario, the Canadian province most heavily engaged in auto production and assembly and a bastion of Liberal support. Although the 1965 CanadaU.S. Auto Pact was formally left intact, the actual operation of the FTA eroded the pact’s safeguards for Canadian production and assembly.51 As in the past, an American automobile manufacturer seeking to qualify for duty-free status was still required to assemble as many cars in Canada as it exported. Once the FTA took effect, however, the content requirement was changed so that production could be of Canadian or U.S. (versus strictly of Canadian) origin. In practice, this meant that Auto Pact companies could continue receiving duty-free status even if they produced nothing whatsoever in Canada.52 And 48
During the 1960s and 1970s, prominent Canadian nationalists such as Kari Levitt began calling attention to the dramatic growth in American foreign direct investment. For Levitt, whose Silent Surrender (1970) became a Canadian best-seller, Canada-U.S. relations posed an even greater threat to Canada’s economic independence and identity than the country’s historical ties to the British Empire. As hostility toward the British came to be supplanted by anti-U.S. sentiment, Canadians began looking to trade barriers (such as those erected by the Trudeau governments of the 1970s) as a way of insulating Canadian industry and culture from U.S. domination. 49 Precisely how large these adjustment costs were likely to be was a matter of intense debate. Among those analysts who believed Canada could weather the storm relatively easily were Lipsey and York (1988) and Schott (1988). Others were less optimistic. See, for example, the studies by Bowker (1988); Duncan Cameron (1988); and Henderson (1987). 50 Lipsey, Schwanen, and Wonnacott 1994, 147– 48. 51 R. Wonnacott 1991c, 109. 52 Mackenzie 1988, 128; Johnson 1993, 92. Nor was Canada able to offset the loss of U.S.
J U M P - S T A R T I N G
T H E
B A N D W A G O N
117
because the FTA phased out all tariffs on bilateral trade in auto products and parts, critics of the FTA argued that the value of duty-free status itself—which is what had given U.S. auto companies the incentive to continue assembling cars in Canada—would evaporate. Even if Canadian auto workers had shared the view that the FTA would increase national income, they also expected there to be a considerable lag, extending for several years if not decades, before the anticipated gains materialized. In the meantime, American car manufacturers would be afforded greater scope to circumvent the stringent Canadian content restrictions embodied in the 1965 pact. And as competition from the United States increased, those workers lucky enough to retain their jobs would be forced to settle for ever-shrinking paychecks. The Liberals’ stance against the FTA was also in keeping with a view that was becoming increasingly prevalent among members of Canada’s foreign policy community. Published in 1981, a widely discussed article by Allan Gotlieb, the Undersecretary of State for External Affairs and later the Canadian ambassador to the United States, argued that Canada was in danger of becoming too reliant on the U.S. market. According to Gotlieb, the best way for Canada to stave off American domination was through the geographic diversification of Canadian foreign economic relations, a pattern whose virtues had first been extolled in the wake of President Nixon’s August 1971 imposition of a 10 percent surcharge on Canadian imports. Like earlier proponents of the “Third Option,” Gotlieb believed that Canada would continue to be victimized by U.S. unilateralism until it decided once and for all to scale back its ties to its southern neighbor.53 A Successful Case of Institutional Engineering On October 25, 1993, Canada’s Conservative government suffered a shattering general election defeat and the Liberals returned to power. For the FTA’s enacting coalition, it was their worst nightmare come true. It was also, for that matter, one of the most striking electoral reversals ever recorded in the history of democratic politics. Having entered the election holding 153 out of 295 seats in the legislature, the Conservative party emerged with just two seats. The Conservatives even lost their status as an “official party,” a designation reserved for parties with more than nine members of parliament.54 Although it would be tempting to identify the FTA as a cause of the production and assembly by luring Japanese and Korean auto manufacturers with the promise of duty-free access to the Canadian market. The FTA expressly prohibited Canada from extending the terms of the Auto Pact to foreign producers that did not already enjoy it. It also eliminated another provision that had once prompted Asian car companies to set up shop in Canada—a loophole allowing non-U.S. auto producers who satisfied the terms of the Auto Pact to export duty-free to the American market (see R. Wonnacott 1991c, 108 – 9). Finally, the FTA also called for Canada to end its embargo against American used cars by 1993. 53 The idea behind the Third Option was first introduced into Canadian government circles in a foreign policy position paper issued by Canada’s External Affairs Ministry during the federal election campaign of 1972. See Mace and Hervouet 1989. 54 The decline in the party’s share of seats in the House of Commons was not merely an ar-
118
N O R T H
A M E R I C A N
T R A D E
Figure 6-3. Indicators of Canadian Economic Performance, 1984–1993
Conservatives’ pummeling at the polls—and disenchantment with the agreement certainly did not help the Conservative cause55 —other factors played the decisive role in the government’s defeat. TRACING THE ROOTS OF THE CONSERVATIVES’ ELECTORAL DEMISE
The Conservatives’ reputation as the party responsible for the FTA was not the primary reason for their electoral setback, just as their support for free trade was not the only reason for their electoral success in the 1988 campaign. Whereas in 1988 a significant portion of the Liberal party’s anti-free trade vote had been siphoned off by a third party, in 1993 it was the proFTA vote that was divided. Voters sympathetic to the FTA and inclined to predicate their vote on the issue now had several pro–free trade parties from which to choose. Both of the two upstart parties in the Canadian political system—the Bloc Quebecois, a separatist party formed in 1990 by francophone-nationalist MPs from Quebec, and the populist Reform party led by Preston Manning—were staunch supporters of the bilateral accord. tifact of Canada’s first-past-the-post electoral system. The Conservatives’ share of the popular vote also dropped, although less dramatically. 55 In the immediate aftermath of the 1988 election, Canadian public opinion on the agreement, though still divided, shifted in a pro–free trade direction. By 1990, however, attitudes toward the accord had moved dramatically in an unfavorable (that is, anti–free trade) direction. Public opinion polls taken prior to the 1993 election registered even stronger opposition to the FTA than was recorded at the time of the “free trade” election of 1988. See Clarke and Kornberg 1992, 41.
J U M P - S T A R T I N G
T H E
B A N D W A G O N
119
An even greater political liability for the Conservative government was the state of the Canadian economy. Almost immediately after the FTA went into effect, Canada plunged into a recession deeper and more intractable than the one that was to hit the United States. Having governed continuously since 1984, the Conservative party could not easily avoid blame for Canada’s mounting fiscal deficit and double-digit rates of unemployment (Figure 6-3). Declining confidence in the Conservative government’s ability to manage the Canadian economy was the single most important factor in the deterioration of Mulroney’s public approval ratings after 1988.56 One month after the 1988 election, nearly half of those surveyed in a Gallup poll supported the government. By the spring of 1990, fewer than one in six respondents (16 percent) did so. During the same interval, Mulroney’s approval rating declined from an average score of 50 (out of a possible 100) to only 28.57 With his popularity rating the lowest of any prime minister since polling in Canada began, Mulroney announced in the spring of 1993 that he intended to retire from politics. At their annual conference in June, Conservatives selected Kim Campbell, the former defense minister, to lead them into the upcoming federal election. Campbell, however, turned out to be an incompetent campaigner—aloof, not particularly charismatic, and prone to gaffes. In fact, Mulroney’s personal unpopularity, exacerbated by his last-minute Senate packing and a lavish farewell world tour, had already eroded support for the Conservative party even before Campbell took the helm. To be sure, some commentators did blame the FTA for the dismal state of the Canadian economy.58 Yet much of the advanced industrialized world suffered an economic downturn during this same period, and to the extent that Canada’s problems were more serious, blame more clearly rests with the coincident decline in worldwide commodity prices—as a large producer of raw materials, Canada was especially vulnerable—and the anti-inflation zeal of the Bank of Canada governor, John Crow. Crow’s restrictive monetary stance resulted in higher interest rates than those found in many other Western countries. In turn, high interest rates caused a sharp rise in the value of the Canadian dollar, further depressing Canadian manufacturing earnings and employment.59 The dismantling of long-standing trade barriers thus seems to have been no more responsible for Canada’s economic woes after the implementation of the FTA than protectionist practices had been for its earlier economic success. WHY CANADA’S DOMESTIC POLITICAL SHOCK FAILED TO DISRUPT THE NEW STATUS QUO
Subsequent political developments provide a striking indication of how entrenched the FTA had become since making its first appearance on January 1, 1989. Despite commanding a comfortable majority in the Canadian House of 56
Clarke and Kornberg 1992. Clarke and Kornberg 1992, 30. 58 See, e.g., Barlow 1990 and Hurtig 1991. 59 Gaston and Trefler 1994. 57
120
N O R T H
A M E R I C A N
T R A D E
Commons, the new Liberal prime minister, Jean Chrétien, never once repeated his party’s earlier pledge to “tear up the agreement.”60 Instead, his platform— published just over a month before the 1993 election—called for “renegotiating” the FTA.61 When asked during the campaign to elaborate, Chrétien explained that this simply meant that a Liberal government, if elected, would seek to obtain a “more effective” dispute resolution mechanism, and that this could be accommodated under the framework of the existing agreement. The question of how a Liberal government would respond if the Americans were to reject this “accommodation” was left deliberately vague. Why did the new Liberal government decide not to abrogate the FTA upon taking office? Was it because Liberal voters regarded the agreement as an efficient solution to a bilateral collective-action problem? The answer is almost certainly no, for the Liberal party’s support base continued to look upon Canada’s initial entry into the pact as a monumental mistake (see chapter 7). That they continued to prefer the pre-FTA status quo did not mean, however, that they were prepared to pay absolutely any price to restore it. Why not? Let me suggest two answers. The first—and most salient from an institutional design standpoint—is that the new post-FTA status quo, as unappealing as it looked to traditional Liberal constituencies, was not intolerable. It had, after all, been specifically designed with (some of ) their interests in mind. Most importantly, the agreement provided these constituencies— and the FTA’s Canadian “losers” more generally—an institutional means of fending off the insistent demands of U.S. fair traders. Although it is true that the adjudication procedures set forth in Chapters 18 and 19 of the FTA failed to shield Canadian opponents of the agreement from a number of highly publicized U.S. trade actions, the very existence of these procedures—and the apparent success enjoyed by Canadian industries in cases involving Canadian raspberries, pork, and other products62 —no doubt caused many American businesses to think twice about using antidumping or countervailing duty actions to pressure the Canadians into opening up their markets beyond what they (as opposed to the Americans) deemed fair.63 The second answer is, in many ways, the flip side of the first: just as the original beneficiaries of the FTA made sure to lower the participation costs 60 In a much-quoted line during the 1988 campaign, John Turner, the Liberal party leader at that time, vowed he would “tear up the agreement” were he to succeed Mulroney as the head of government. See chapter 7. 61 Simon 1993b. 62 Goldstein 1996; Horlick and DeBusk 1992; Boddez and Trebilcock 1993. 63 One exception was an action filed by the United States against Canadian beer-selling practices, with U.S. breweries alleging that laws restricting imported beer sales in Canada constituted an unfair trade barrier. Additional frictions were caused by the Americans’ continuing objections to the Canadian practice of subsidizing its lumber industry through low tree-cutting fees. In the case of the lumber dispute, Canada protested that it had already overhauled its “stumpage” fee schedule in response to earlier American complaints, yet neither these reforms nor the institutional innovations in the FTA seemed to diminish the combativeness of U.S. logging companies threatened by cheap Canadian imports.
J U M P - S T A R T I N G
T H E
B A N D W A G O N
121
for nonbeneficiaries whose political representatives might one day be in a position to abandon it, the FTA’s original sponsors also took steps to raise the costs of not participating in the agreement—that is, the costs (to the losers) of withdrawing from it. And here, too, the FTA’s open-ended structure and innovative system for dispute settlement played an important part in the story. For by the time the Liberal party scored its decisive victory over the Conservatives, the FTA had already been in effect for several years, during which time its binational dispute settlement panels, in the course of handing down decisions over time, had begun promulgating a new body of “common law.”64 For an anti–free trade government to have abandoned the agreement at that point would thus have been a major, and no doubt highly disruptive—move, one sure to provoke the ire of businesses, public agencies, and private citizens throughout Canada (and the United States as well) who had begun, however reluctantly, adapting to the new institutional reality.65
Conclusion As attractive as the Liberals’ anti-FTA platform had been in 1988, that same platform would have met with strong resistance in 1993. This is in part, I have suggested, because the perceived costs of withdrawing from the Canada-U.S. Free Trade Agreement had steadily increased over the course of the late 1980s and early 1990s. Once Canadian citizens—as well as members of the international financial community—had come to accept the FTA as a part of the economic landscape, it began to seem that the only prudent course for the new Liberal government was to cut its losses and leave well enough alone. But that is really only part of the story, for the architects of the pact also had a stake in moderating these losses, and thus in endowing their new arrangement with a sufficiently accommodating institutional structure. And this, as we have seen, is exactly what they did. In this respect, Reagan and Mulroney’s insulation strategy may be judged a success. By interweaving a multitude of sector-specific liberalizing reforms into a single unalterable package, hastening that package’s entry into force, and, perhaps most importantly, lessening its burden on those American and (especially) Canadian citizens who, if given a choice, would have opted for the initial “noncooperative” status quo, the politicians who engineered the pact ensured that their fundamental overhaul of Canada-U.S. relations would persist long after they themselves had left office. 64
Goldstein 1996; Horlick and DeBusk 1992; Moyer 1993. See, e.g., B. Simon 1992. In addition to withstanding the vicissitudes of domestic politics in Canada, Reagan and Mulroney’s shared free trade agenda also survived the turbulence of domestic politics south of the Canada-U.S. border, including the election in 1992 of a Democratic president—the first in more than a decade. 65
C H A P T E R
7
NAFTA and Beyond: Is Free Trade Contagious?
CONTINUING OUR EXAMINATION of North American economic integration, this chapter looks at how, only a few years after its inauguration in 1989, the ambitious Canada-U.S. trade and investment regime we have just been discussing came to be supplanted by an even more ambitious, continental regime.1 Remarkably, this transformation took less than five years from start to finish. Meeting in secret, cabinet officials from Mexico and the United States first broached the issue in March of 1990, just fourteen months after the original Canada-U.S. pact entered into force.2 Three months later, the presidents of the United States and Mexico—George Bush and Carlos Salinas de Gortari, respectively—issued a joint statement declaring their mutual interest in forging a separate Mexico-U.S. accord. The next act in the unfolding drama came in January of 1991, when Brian Mulroney, the Canadian prime minister whose pro–free trade government was shortly to be ousted from office, announced that he too wished to take part in the negotiations. Trilateral talks did not formally begin for another six months, as Bush was unable to obtain congressional approval of his request for a two-year extension of “fast-track” procedures until the end of May. Once negotiations were underway, however, they progressed rapidly, concluding in just over a year on August 12, 1992. In the United States, final ratification of the agreement did not come until November 18, 1993. Nevertheless, a last-minute lobbying blitz by President Bush’s pro-NAFTA successor, Bill Clinton, succeeded in mobilizing the requisite congressional support, and with NAFTA’s affirmative vote in Congress easing the way for final approval by the Canadian and Mexican governments, the accord entered into force as originally scheduled, on January 1, 1994. To date, the decision-making calculus of the United States has received far more scrutiny than the political and economic forces that propelled NAFTA’s other two signatories into the agreement. This is unfortunate, for in many ways the Mexican and Canadian perspectives offer even greater insight into 1 Just how ambitious was this new regime? If attention is restricted to the Western Hemisphere, the “runner-up” to the North American Free Trade Agreement (NAFTA) in terms of economic scope and coverage is the Common Market of the South—otherwise known as Mercosur—which encompasses the vast bulk of South America’s economic activity. Mercosur’s population, however, is less than two-thirds that of NAFTA, and its combined gross domestic product is less than one-tenth. Indeed, when NAFTA was first put into effect, the combined GDP of its three member states exceeded that of the entire EU—$7.6 trillion compared with $6.8 trillion—and accounted for nearly 30 percent of total world output (World Bank 1997). 2 Wall Street Journal, March 27, 1990.
N A F T A
A N D
B E Y O N D
123
the political well-springs of international collective action. This chapter will, I hope, make the case. My discussion unfolds as follows. After briefly setting the stage in the next section, I consider the case of Mexico and its “scramble for inclusion.” Then, having inquired into the Mexican government’s motivations for embracing NAFTA, I turn in the third section to the (analytically distinct) question of how the trilateral pact may have altered the utility of Mexico’s ruling elite. Rather than viewing NAFTA as a political asset, Mexico’s political establishment would in fact have preferred and, had it been within their capacity, would have gone to great lengths to restore North America’s original (i.e., more protectionist) status quo—or so, in this section, I will argue. As of 1989, however, that status quo appeared to be a thing of the past. For though that year’s newly enacted Canada-U.S. FTA was certainly not intended to discriminate against Mexico, the bilateral pact’s very existence—by affording U.S. imports from Canada a preference over U.S. imports from Mexico— was to have precisely that effect. In essence, then, the Canada-U.S. agreement upped the ante for Mexico, which as an FTA outsider now had to liberalize further (and faster) than even its relatively pro-free trade president would have liked. Shifting the spotlight from Mexico to Canada, the fourth section of the chapter takes a closer look at the calculus underlying Canada’s equally “enthusiastic” embrace of NAFTA in the early 1990s. Often overlooked in discussions of Canada’s involvement in the agreement is the fact that, when Canada ratified the accord in 1993, its government was led not by Mulroney, whose pro-free trade Conservative party had just been relegated to the political sidelines, but by newcomer Jean Chrétien, the Liberal prime minister whose historically anti-free trade party had only a few years earlier fought tooth and nail to prevent NAFTA’s precursor, the bilateral Canada-U.S. pact, from taking effect. To be sure, the Chrétien government did eventually agree to bring Canada into the trilateral regime. Its decision was motivated less by any newfound appreciation for the political or economic benefits of free trade, however, than by the fear of what might happen if Canada were to stand aside and allow its southern neighbor to become the hub in a U.S.-dominated hub-and-spokes system. The chapter concludes by asking whether go-it-alone power might be propelling the extraordinary spread of free trade initiatives that has occurred—and is still taking place—throughout much of the world since NAFTA’s entry into force in 1994.
An Introduction to the Puzzle In 1984, Mexico’s governing party—the Partido Revolucionario Institucional, or PRI—began aggressively reorienting its trade policy toward the goal of boosting manufactured exports. Over the next few years, Mexico’s tariffs fell from 23.5 to 11.0 percent, its direct export subsidies were almost com-
124
N O R T H
A M E R I C A N
T R A D E
pletely eliminated, and the country’s elaborate system of import licenses, which formerly protected more than 90 percent of its domestic production, was scaled back to only about 23 percent.3 It is true that Mexican President Miguel de la Madrid initially sought to delay this process of trade and investment liberalization—or at least to cushion its effects—by engineering a depreciated exchange rate. In 1988, however, the peso was allowed to appreciate and this period of “exchange rate protection” came to an end.4 Many scholars have come to see NAFTA as a necessary and integral part of de la Madrid’s earlier apetura to world markets. Rather than a radical disjunction in the history of Mexico-U.S. relations, NAFTA is widely viewed as the “logical culmination” of policies the PRI had been pursuing unilaterally—and, for the most part, successfully—since the mid-1980s.5 The reality could not be more different. In fact, Mexico’s political establishment was deeply ambivalent about its initial trade opening. For moderate though it was by NAFTA standards, President de la Madrid’s unilateral initiative had greatly distressed many of the PRI’s traditional constituencies. Mexican corporations oriented toward the domestic market, the highly paid employees of Petroleos Mexicanos (PEMEX) and other large state-owned enterprises, large landowners, small agricultural producers, organized labor, the government bureaucracy—all had developed a strong stake in the maintenance of high barriers to trade with the outside world. Indeed, it would be hard to overstate just how deeply entrenched the PRI’s inward-looking economic policies had become since the 1930s, and thus how wrenching their removal stood to be for their long-time beneficiaries.6 Take, for instance, the case of small agricultural producers, a crucial element in the PRI’s political base. Prior to de la Madrid’s liberal reforms, Mexican producers of corn and beans (campesinos), the two main staples of the Mexican diet, had enjoyed hefty government subsidies and near-total protection from import competition. Starting in the mid-1980s, however, Mexican authorities set about the task of dismantling this long-standing arrange3
Bosworth, Lawrence, and Lustig 1992. Emblematic of the PRI’s trade opening was Mexico’s entry into the GATT in 1986. As a developing country, however, Mexico enjoyed certain privileges and exemptions under the GATT, and it refused to sign the code on government procurement. The differences between GATT- and NAFTA-based liberalization are discussed later on in this chapter. 4 For a comprehensive overview of Mexican economic policy making during the 1980s, see Lustig 1992. 5 Baer 1991, 131. On this point, Mexican specialists often try to have it both ways. On the one hand, NAFTA is described as a natural outgrowth of previous policies. On the other hand, it is suggested (sometimes by the same authors) that NAFTA was a U-turn of monumental proportions. The 1991 article by Delal Baer is a case in point. Only a few sentences after depicting NAFTA as the logical culmination of de la Madrid’s policies, Baer describes the decision to pursue a free trade agreement as “a stunning political act—heretical in some Mexican intellectual and political circles—and [one that] defies Mexico’s long history of defensive nationalism.” 6 See, e.g., Cook, Middlebrook, and Molinar 1994b; Weintraub 1990b.
N A F T A
A N D
B E Y O N D
125
ment.7
This, of course, enraged campesinos throughout the country, but that was really only the beginning of the PRI’s political problems. With more than a quarter of Mexico’s total labor force employed in agriculture, the real fear was that by opening the agricultural sector to foreign competition, the government would trigger a mass migration from the Mexican countryside to the country’s already overcrowded urban areas.8 Were this to happen, the PRI’s political standing would have been undermined not only within the agricultural community but also—an even bigger concern—among the country’s urban, largely blue-collar population. Fearing a massive influx of displaced rural workers, many Mexican workers, though ostensibly among the long-term beneficiaries of liberalization, were skeptical. The fact that these workers had already seen their real wages decline by some 40 percent since the onset of Mexico’s debt crisis in 1982 did not help matters. In 1990, three-quarters of all Mexican citizens received less than U.S.$300 monthly.9 As if that were not bad enough, Mexican workers also had to contend with a growing income gap between rich and poor.10 Between 1984 and 1989, the share of income accruing to the poorest 50 percent of Mexican households fell from 19.0 percent to 16.3 percent, while the income received by the wealthiest 10 percent of households rose from 34.3 percent to 39.0 percent.11 Although by no means the only source of Mexican workers’ deteriorating economic situation, trade liberalization was the most consequential politically. It was, after all, the most visible. Because de la Madrid’s foray into world markets had coincided with a dramatic decline in living standards, there was, as Sidney Weintraub (1990b, 108) notes, “a post hoc, ergo propter hoc tendency to blame the economic disaster on . . . the trade opening, the push to export in order to service external debt, and the shift from a statedominated to a more private-sector-oriented economic model.” In the past, public attitudes of the kind that Weintraub describes would have posed little threat to the PRI’s dominant political position.12 By the late 1980s, however, Mexico’s authoritarian rulers were becoming increasingly vulnerable to expressions of societal disapproval and dissent. Fueled by the PRI’s bungled relief effort after the September 1985 earthquakes, its resort to election fraud in the legislative elections of 1986 and the presidential election of 1988, and, perhaps most of all, its persistent failure to improve the economic conditions of ordinary Mexican citizens, pressures for democrati7 It is important to emphasize that the introduction of markets into Mexico’s long-sheltered agricultural sector was already well underway by the time Salinas took over the political stage. See Josling 1992, 146 – 49. 8 Ramirez de la O 1993; Weintraub 1993. 9 Castañeda 1993b, 65. 10 See, e.g., Castañeda 1993b; Weintraub 1990b, 27. 11 Percentages computed from Pastor and Wise 1997, 425, table 2. 12 Cook, Middlebrook, and Molinar 1994a; Cornelius 1996, 51–75.
126
N O R T H
A M E R I C A N
T R A D E
intensify.13 Indeed,
zation had begun to had it not been for a well-timed computer failure, it is doubtful that Salinas, President de la Madrid’s handpicked successor, would have emerged victorious in the 1988 election.14 While Cuauhtémoc Cárdenas, the head of the left-wing National Democratic Front, officially received 31.1 percent of the popular vote, most observers believe his true vote share exceeded Salinas’s 50.7 percent, which was in any case the lowest vote share ever before registered for a PRI candidate. By the time Salinas took office in 1988, the PRI thus could no longer afford to incur the wrath of Mexico’s newly “dealigned” electorate.15 This is why the news that Salinas was prepared to introduce Mexico into a formal free trade agreement with the United States came as such a surprise. Even veteran observers of Mexican politics were taken aback.16 Only one year earlier, the notion that Mexico would be willing to commit itself to a comprehensive agreement with the “imperial power” to its north had been dismissed as an idea too radical to merit serious attention. “A U.S.-Mexico FTA or Mexican adherence to the Canada-U.S. FTA is not feasible in the short to medium term,” a well-respected study by Jeffrey Schott (1989, 44) had concluded. “A NAFTA will therefore remain an ideal, not a real policy option, for the next decade or more.” At the time it was offered—Salinas’s presence at the helm of the Mexican polity notwithstanding—this was not a particularly controversial assessment. How could so many intelligent observers have failed to predict Mexico’s 1990 overture to the United States? Part of the explanation lies in the absence of any historical precedent.17 At least as important, however, was the widespread perception that a free trade agreement, by accelerating a liberalization process whose economic payoff had yet to materialize, would spell trouble for Mexico’s ruling elite. As the PRI’s new leader, Salinas had to be especially concerned about maintaining support, or at least passive acquiescence, from his party’s core constituencies, many of which had disapproved of de la Madrid’s earlier trade initiative, and all of which had the power to bring down the government—if not through the ballot box, then through labor strife, capital flight, demonstrations, even revolution. What the PRI wanted above all else, then, was time: although willing to leave de la Madrid’s reforms intact, it was not about to escalate them—or so, again, nearly everyone (incorrectly) assumed. 13
Morris 1992; Reding 1991; Weintraub and Baer 1992. Cook, Middlebrook, and Molinar 1994a, 21. 15 Klesner 1994. 16 See, e.g., Bueno 1988, 118; Diebold 1988, 160 – 63; Schott 1989, 41– 45; and Weintraub 1990b, 15–23. 17 By way of comparison, Weintraub (1993, 71) notes that “it took the U.S. several decades after the infamous Smoot-Hawley tariff of 1930 to open its market as thoroughly as Mexico did in one decade.” Prior to NAFTA, formal trading arrangements between Mexico and the United States were limited to a series of modest framework agreements and “understandings.” 14
N A F T A
A N D
B E Y O N D
127
Mexico’s Scramble for Inclusion Two years after assuming office, Salinas embraced a trade accord that went far beyond the liberalizing measures that President de la Madrid—or, for that matter, Salinas himself—had previously introduced. Something, therefore, must have changed between the final years of de la Madrid’s tenure and Salinas’s 1990 announcement, for otherwise Mexico would not have been willing to take the hitherto unthinkable step of signing a comprehensive free trade agreement with its northern neighbor.18 Contending Explanations There are a number of plausible explanations for this “stunning political act.”19 One possibility is that Mexico’s internal political coalitions and structures shifted in ways that made liberalization suddenly seem more attractive.20 Those who take this view note that it was Salinas, not Bush or Mulroney, who initiated the NAFTA negotiations. From this they conclude that the origins of Mexico’s trade opening—the catalyst that ultimately triggered NAFTA—must be found “inside the Mexican political economy.”21 Often included in the list of possible suspects are the collapse of Mexico’s long18 Note that my claim here presumes that Salinas and other top Mexican officials knew exactly what they were getting themselves into. In contrast, Beth and Robert Yarbrough (1992, 45) attribute NAFTA to a combination of ignorance and wishful thinking on the part of the Mexican government: “The deliberate fostering of dependence that can be exploited requires an asymmetry of information between the parties about the relation. The intended victim must be either unaware of the danger being plotted or so desperate to trade with the opportunist that the potential danger is ignored.” Likewise, Manuel Pastor and Carol Wise, although sharing my view that NAFTA exacted a heavy toll on Mexico, maintain that Mexican policy makers were taken by surprise; hence, the stated purpose of their study (1994) is to demonstrate “[how] ‘rational’ actors and policymakers could settle on strategies that subsequently produce unexpectedly high political costs.” Pastor and Wise 1994, 471, emphasis added. 19 Baer 1991, 134. 20 What follows is not intended to be a comprehensive survey of possible motivations for NAFTA. Among the candidate explanations I will not be addressing are (1) that Mexico’s NAFTA decision was motivated by its belated “discovery” of East Asia, as well as its continuing fascination with Chile, a country whose leaders embarked on the path of market reform in the 1970s (Edwards 1995); and (2) that Mexico’s ruling elites reassessed their long-standing aversion to free trade agreements after reflecting upon Canada’s experience. Regarding the first explanation, however, it is worth noting that neither the Chilean nor the Asian “miracles” were the result of free trade agreements, and that each country’s political conditions (e.g., military rule in Chile) were fundamentally different from those found in Mexico. As for the second explanation, it too seems highly implausible. After all, free trade with the United States aroused intense opposition in Canada—this despite the fact that Canadian citizens (unlike their Mexican counterparts) had the benefit of a long-standing, highly developed social safety net. If anything, then, the lesson to be gleaned from the Canadian experience was that free trade agreements could be political bombshells even when their signatories were at similar levels of economic development, as Mexico and the U.S. were most definitely not. 21 Pastor and Wise 1994, 471, emphasis added.
128
N O R T H
A M E R I C A N
T R A D E
standing corporatist policy-making structure and its replacement by what Manuel Pastor and Carol Wise have described as “a new alliance between a highly insulated executive, his technocratic corps, and representatives of the country’s largest companies and conglomerates.”22 There is, of course, some truth to this domestic-politics explanation for NAFTA. The financial wing of the Mexican state apparatus did in fact grow, both in size and, more importantly, in influence, over the course of the 1980s—a decade that saw a coterie of ambitious young economists and administrators emerge “as a major, well-trained business-oriented and outward-looking force in Mexican politics.”23 Many of these individuals played a direct role in formulating de la Madrid’s liberalization program and so, presumably, wished to see it continue. But while Salinas could not afford to ignore the preferences of his so-called técnicos,24 even less could he afford to disregard the demands of Mexico’s anti–free trade “popular” sectors, without whose support (or at least passive acceptance) his party stood little chance of retaining office in the coming years. A different and, I think, more compelling account of Salinas’s behavior would put the spotlight on large-scale changes in his country’s external environment. More than any purely domestic development, the real impetus for Mexico’s decision to seek a free trade agreement with the United States was his party’s concern that were Mexico to continue on its present (preNAFTA) trajectory, ongoing shifts in the world economy could play themselves out in ways that would end up diminishing the country’s access to long-term investment capital.25 But this merely moves the analysis back a step: Why did Mexico’s leaders believe that negotiating a free trade agreement with the United States would help their country attract the requisite amounts—and right kinds—of foreign investment? While the question admits no simple answer, much of the evidence, including a policy paper written at the time by Jaime Serra Puche, the PRI official who would later become Mexico’s chief representative at the NAFTA negotiations, as well as a similar analysis drafted by the president himself, points to one reason in particular: a free trade agreement with the United States would mitigate the negative trade and investment discrimination that Mexico might otherwise experience as a result of its exclusion from the large regional trading blocs that were just then beginning to crop up in Asia, Eu22
Pastor and Wise 1994, 480. See also Poitras and Robinson 1994. Castañeda 1990, 408. 24 Centeno 1994. 25 Lustig 1992, 184; Salinas de Gortari 1990. Attracting foreign capital had been one of the PRI’s major priorities when Salinas took office—as indeed it had been throughout much of the 1980s. It was with this goal in mind that Mexican officials engineered a debt restructuring agreement with the commercial banks in 1989. When it went into effect the following year, the so-called Brady Plan precipitated an enormous—and, to observers at the time, quite surprising—inflow of foreign capital. As I point out later, however, a disappointingly small share of this capital consisted of new, long-term investments; see also Castañeda 1993b, 63– 65; Peres Núñez 1990, 20; Ros 1992, 67. 23
N A F T A
A N D
B E Y O N D
129
America.26
rope, and North This desire not to be shut out of the world economy, dominated as it was (or so it seemed likely soon to become) by a handful of outwardly protectionist regional “fortresses,” has emerged as a central explanation for the regional initiatives that have appeared since NAFTA made its debut in 1994.27 So far, however, there has been relatively little attention paid to the role of defensive considerations in the formation of NAFTA itself.28 This is surprising given that, as just noted, Salinas himself cast his 1990 decision in defensive terms. As the president reflected at the time, “The changes in Europe and East Asia and an apparent reliance on blocs convinced me that we should also try to be part of an economic trading bloc with the United States and Canada.”29 This rationale is consistent with the observations of Jesus Silva-Herzog, Mexico’s Minister of Finance from 1982 to 1986 and a leading contender for the presidential nomination in 1988: A year ago [1990] everybody in Mexico rejected the idea of a free-trade agreement with the U.S. on the grounds of disparity—in per capita income, size of economy and degree of development. Today, almost everybody accepts the idea of a freetrade agreement, but mainly because they see the challenges faced by individual nations as regional economic blocs begin to form in Japan and Asia, Europe, and North America. Not to join with other nations in trade agreements would be to risk marginalization, to remain spectators on the world scene.30
But while Silva-Herzog and Salinas both refer to the construction of regional blocs in Asia and Europe, the emergence of a regional bloc in Mexico’s own backyard was surely a more immediate cause for concern, for as Serra Puche (1990, 529) noted at the time, Mexico’s commercial ties with Canada and, particularly, with the United States had “a special importance.” Why? For one thing, well over half of Mexico’s new and accumulated for26
Salinas de Gortari 1990; Serra Puche 1990. See, e.g., Bouzas and Ros 1994a; Frankel 1997. 28 There is one important exception: a prescient study by Sidney Weintraub, a leading authority on Mexico-U.S. relations. Writing less than a year after the Canada-U.S. Free Trade Agreement had taken effect, Weintraub predicted that the formation of the Canada-U.S. accord would have a marked impact on Mexican policy. “Over time,” he wrote, “as the effects of the Canada-U.S. FTA are more fully felt, and if they do prejudice Mexico’s own trade expansion (as I believe they will), the pressure to join in a wider free trade agreement in North America will grow” (Weintraub 1990a: 120; cf. Bueno 1988, 116). At the time, however, Weintraub believed that such an agreement was still many years away: “Mexico’s conversion from a stateregulated economy, looking inward, is too new to be pushed very far for now—to a free trade agreement with the very country whose dominant economic status it most fears.” Weintraub 1990a, 110. 29 Quoted in R. Pastor 1990, 32. 30 Silva-Herzog 1991, 29. After acknowledging that joining with the United States in a trade agreement was necessary to keep Mexico from becoming a “spectator on the world scene,” Silva-Herzog continues: “That said, we must also keep in mind that the initial argument against a U.S.-Mexico free trade agreement, the disparity argument, is still valid. That is why we still need deeper analysis and open dialogue before moving too fast.” 27
130
N O R T H
A M E R I C A N
T R A D E
Table 7-1 Foreign Direct Investment in Mexico, 1989 Accumulated
United States Great Britain West Germany Japan Switzerland France Holland Other Total
New
(U.S. $, millions)
(% of total)
(U.S. $, millions)
(% of total)
16,740 1,797 1,661 1,344 1,175 n.a.a n.a. 3,844
63.0 6.8 6.3 5.1 4.4 n.a. n.a. 14.5
1,783 n.a. 78 n.a. 171 49 46 349
72.0 n.a. 3.2 n.a. 6.9 2.0 1.9 14.1
26,561
100.0
2,476
100.0
Source: Lustig 1992, 125. aNot
available for specific country, but included in “other.”
eign direct investment came from U.S. sources (Table 7-1). Nor is it any great mystery why U.S. investors were attracted to Mexico. Particularly after the mid-1980s downturn in the oil industry, Mexico’s most promising long-term investment prospects were to be found in its burgeoning (non-oil) export sector. As promising as they were, however, the long-term profitability of Mexico’s new export-oriented businesses—their expected earnings potential— was almost entirely contingent on their ability to attract customers within the United States. Table 7-2 offers striking testimony to Mexico’s overwhelming dependence on the United States both as a supplier of inputs to Mexican industry and, more important for present purposes, as a purchaser of Mexican output.31 In short, Mexico’s access to markets in Europe and Asia was not nearly as consequential, in terms of luring new foreign direct investment, as was its access to the U.S. market. So what, then, was threatening this access? It was almost certainly not the increasing use of administrative trade remedies by the United States, though some scholars have suggested as much.32 It is true that authorities in the United States did (sometimes) retaliate against U.S. trading partners deemed to be using “unfair” means of penetrating the American market (see chap31 If anything, the 69.3 and 68.0 percent figures in the table understate the United States’ true importance to Mexico. Any new infusions of foreign capital were likely to come from investors who hoped to profit from an expansion of Mexico-U.S. trade in manufactured products, an area in which the U.S. share in Mexico’s exports and imports was even greater than its share in Mexico’s total exports and imports. Weintraub (1990c, 120) estimates that in 1987 Mexico’s dependence on the United States for trade in manufactured goods and services was somewhere between 80 and 85 percent. 32 See, e.g., Kahler 1995, 21.
N A F T A
A N D
131
B E Y O N D
Table 7-2 Mexico’s Major Trading Partners, 1989 Exports (%)
Imports (%)
Western Hemisphere United States Canada Latin-American Integration Associationa Argentina Brazil Venezuela Others Central American Common Marketb Others
78.5 69.3 1.2 3.3 0.5 0.9 0.3 1.7 1.9 2.8
73.9 68.0 1.8 3.0 0.6 1.5 0.2 0.6 0.4 1.0
Western Europe European Communityc European Free Trade Aread Others
12.5 11.4 0.8 0.3
16.9 14.3 2.6 0.0
Asia Japan Others
8.4 5.8 2.6
7.4 4.5 2.8
Rest of World Eastern Europe Others
0.6 0.1 0.5
1.8 0.3 1.5
Countries
Source: Lustig 1994, 55, table 2.3. Data originally compiled from Banco de México, The Mexican Economy: 1991 (Mexico City, 1991), 223. a LAIA (Argentina, Brazil, Bolivia, Columbia, Chile, Ecuador, Paraguay, Peru, Uruguay, and Venezuela) bCACM
(Costa Rica, El Salvador, Guatemala, Honduras, and Nicaragua)
cEC
(West Germany, Belgium-Luxembourg, France, Italy, the Netherlands, the United Kingdom, Denmark, Ireland, Greece, Spain, and Portugal) dEFTA
(Austria, Finland, Norway, Sweden, Switzerland, and Iceland)
ter 6). In the case of Mexico, however, U.S. antidumping and countervailing duty actions—never very significant to begin with—had been on the decline for several years prior to 1990, the year the PRI went public with its decision to begin negotiating a free trade pact with the United States (Table 7-3). The Impact of the Canada-U.S. Agreement Not on the decline, by contrast, was Mexico’s concern that, as Serra Puche (1990, 530) himself put it, “our exports to [Canada and the United States]
132
N O R T H
A M E R I C A N
T R A D E
Table 7-3 U.S. Antidumping and Countervailing Duty Investigations Initiated against Mexico, 1979-1990 Antidumping Cases
1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 Total
Countervailing Duty Cases
Total
Against Mexico
Total
Against Mexico
26 21 15 65 46 74 66 71 15 42 23 43
1 0 0 0 1 1 3 2 0 0 2 1
40 14 22 140 22 51 43 27 8 11 7 7
0 1 1 9 6 7 5 0 0 0 0 0
507
11
392
29
Source: Compiled from Destler 1992, Apps. B and C.
will be shut out as a result of the preferential treatment they give each other.” Indeed, the Canada-U.S. Free Trade Agreement reduced Mexico’s access to the U.S. market, and thus Mexico’s attractiveness as a site for foreign investment, by its very existence. It did so by (among other things) lowering the tariffs on nearly all U.S. imports from Canada. True, these tariff reductions posed a threat to Mexico only insofar as U.S. imports from Mexico competed directly against those from Canada. And yet, as the following tables and figures make clear, head-to-head competition between Canadian and Mexican suppliers to the U.S. market was remarkably intense, and getting more so all the time. Consider the sector-by-sector breakdowns displayed in Figure 7-1. The big lesson to take away from this figure is that by 1987 Mexico’s export “structure” had largely converged with that of Canada, with both countries relying upon just one product category—machinery and transportation equipment—for nearly half of their total U.S. exports. Table 7-4 presents a more fine-grained picture. U.S. tariffs in many of the product lines depicted in this table were initially in the range of 4 or 5 percent.33 After the Canada-U.S. free trade pact entered into force, however, the tariffs on Canadian exports 33 Trade in some of these product categories was relatively trivial. Yet even if one restricts attention to product categories for which U.S. imports from Mexico and Canada exceeded $100 million, the proportion of Mexican exports to the United States facing tough competition from Canadian concerns falls only slightly, from 87.7 percent to 61.8 percent (United Nations 1989).
Figure 7-1. U.S. Merchandise Imports from Canada and Mexico, 1979–1987
5000 5122 5133 5135 5141–43 5811 5812 6000 6291 6318 6412 6419 6429 6516 6569 6612 6647 6651 6742 6748 6783 6811 6821 6822 6851
SITCa
Chemicals Alcohols, phenols, etc. Inorganic acids, etc. Metallic oxide for paint Metal composed of inorganic acid Product of condensation, etc. Product of polymerization, etc. Manufactured Goods Classified Chiefly by Material Rubber tires, tubes Wood simply worked, nesd Other printing paper, nes Other paper, etc., nes, bulk Paper, etc., articles, nes Yarn of synthetic fibers Other textile products Cemente Safety glass Bottles, etc., of glass Iron, steel medium plate Iron, steel thin-coated, nes Iron, steel tube, pipe, nes Silver unworked, partly worked Copper alloys, unwrought Copper alloys, worked Lead, alloys unwrought
Product 98 69 51 140 99 314 530 70 514 230 76 28 41 159 157 39 247 133 250 123 356 88 62
38 60 32 22 183 26 70 157 109 23 36 54 48 277 56 45 29
U.S. Imports from Canada (millions U.S.$)
22 61 25 34 43 73
U.S. Imports from Mexico (millions U.S.$)
Table 7-4 Areas of Intense Competition between Mexican and Canadian Manufactured Exports to the United States, 1987
0–5; median 4 0–5 0–17; median 5 0–17; median 5 0–17; median 5 9–15; median 12 0–30 0–$0.22/ton 5–6 6–38 6.5 6.5 1–8 0–6 1–6 1 3.5
0–20 0–4 nac 0–17; median 3 0–12.5 20
Range of U.S. MFN Tariff b
6861 6922 6952 6981 6986 6989 7000 7115 7143 7149 7184 7191 7192 7193 7195 7199 7221 7222 7231 7241 7242 7249 7250 7291 7292 7293 7294 7295 7299
Zinc, alloys unwrought Mil transport boxes, etc. Tools, nes Locksmiths’ wares Springs and leaves Other base metal manufactures Machinery and Transportation Equipment Piston engines, nonair Statistical machines, computers Office machines, nes Construction mining machinery, nes Heating, cooling equipment Pumps, centrifuges Mechanical handling equipment Powered tools, nes Machine parts, accessories, nes Electric power machinery Switchgear, etc. Insulated wire, cable Television receivers Radio broadcast receivers Telecomm. equipment, nes Domestic electric equipment Batteries, accumulators Electric lamps, bulbs Transistors, values, etc. Automotive electrical equipment Electrical measuring, control equipment Other electrical machinery
302 37 55 246 187 291 1,639 477 674 225 132 291 262 81 330 286 313 223 32 15 567 111 30 32 533 105 223 193
42 23 24 73 36 54 973 150 317 30 124 103 76 25 64 538 516 1,001 293 604 813 156 56 27 318 107 55 299
(continued)
0–4; median 3.7 na 0–5; median 3.7 0–3.7 2–4 0–3 0–6 2.5–4.5 0–6; median 4 3 5–6 5–8; median 5.3 6–15 4–8; median 5 5–8; median 5 0–5 5.1–5.3 0–8; median 3.7 0 0–8; median 3 4–10; med. near 10 3–5; median 4
2–19 0–6; median 5 4–8 2–7.5; median 6 4–6 0–6
Passenger motor vehicles, excl. buses Lorries, trucks Motor vehicles parts, nes Miscellaneous Manufactured Products Lighting equipment Furniture Textile clothes, not knit Clothing, accessories knit Footwear Measuring, controlling instruments Sound recording tapes, discs Articles of plastic, nes Toys, indoor games Outdoor sporting goods, nes Works of art, etc.
Product 10,257 3,700 5,143 27 1,031 96 35 35 202 57 555 27 62 136
38 311 307 66 105 106 74 80 88 31 89
U.S. Imports from Canada (millions U.S.$)
1,178 89 651
U.S. Imports from Mexico (millions U.S.$)
0–8; median 3.7 2.5–9; median 6 median 17 above 20 2–48; median 37.5 0–8; median 5 4.2 2–14; median 5 0–12; median 7 0 0–11; near 11
2.5–8.5; near 8.5 na 0–4; median 3
Range of U.S. MFN Tariff b
The MFN tariff was the tariff normally levied on imports from Mexico and Canada in 1987—figures are percentages.
not elsewhere specified
Boldface denotes product category for which imports from Mexico and Canada were, in both cases, in excess of $100 million.
e
d
not applicable.
c
b
a Standard International Trade Classification category. Table reports data only on product categories for which U.S. imports from Mexico and Canada exceeded $20 million.
Source: Weintraub 1990a, 112–14, table 1. Trade data originally compiled from tapes of United Nations trade statistics. The original source for tariff and nontariff measures was U.S. International Trade Commission, Harmonized Tariff Schedule of the United States, 1st ed., supp. 3 (Washington, D.C.: United States Government Printing Office, 1988).
7321 7323 7328 8000 8124 8210 8411 8414 8510 8619 8912 8930 8942 8944 8960
SITCa
Table 7-4 Continued
N A F T A
A N D
137
B E Y O N D
Table 7-5 Areas of Greatest U.S. Import Growth from Mexico, 1985–1987 U.S. Imports from Mexico (millions of U.S.$) SITCa
Product Category
7143
†Statistical machines, computersb †Passenger motor vehicles, excl. buses Copper, alloys, worked †Rubber tires, tubes †Other printing paper, nesc †Iron, steel medium plate †Heating, cooling equipment †Automotive electrical equipment Television receivers †Iron, steel thin-coated, nes †Copper, alloys unwrought †Pumps, centrifuges †Locksmiths’ wares Real jewelry, gold, silver Medical instruments, nes Clothing, accessories knit †Domestic electric equipment †Furniture †Cement †Insulated wire, cable
7321 6822 6291 6412 6742 7191 7294 7241 6748 6821 7192 6981 8971 8617 8414 7250 8210 6612 7231
Avg. Annual Growth Rate (%)
1985
1987
6
150
1,177.4
111
1,178
479.8
6 6 5 8 30 27
45 38 32 36 124 107
319.9 283.2 258.5 179.7 157.5 146.1
88 16 19 38 29 26 47 29 81 168 87 618
293 54 56 103 73 64 110 66 156 311 157 1,001
116.0 113.9 96.8 86.6 77.9 71.7 67.2 63.6 46.8 42.7 39.7 31.0
Sources: Adapted from Weintraub 1990b, tables 1 and 2. Data originally compiled from tapes of UN trade statistics. aStandard
International Trade Classification category
bSITC
items marked by a dagger (†) are those for which, in 1987, U.S. imports from Canada exceeded $100 million. cnot
elsewhere specified
to the United States began to decline (eventually to zero), while tariffs on Mexican exports remained unchanged. But if the tariff preferences embedded in the FTA posed a threat to established export-oriented firms in Mexico, they posed an even greater threat to Mexico’s emerging markets. Table 7-5 displays data on Mexico’s twenty fastest-growing manufactured U.S. exports between 1985 and 1987. By 1987, many of these industries had already begun to make inroads into the United States. In several cases, however, these product lines had only recently
138
N O R T H
A M E R I C A N
T R A D E
gained a foothold in the American market and (with very few exceptions) these industries, including fledgling upstarts such as tires and newspaper, faced stiff competition from Canadian suppliers.34 And that was in 1987. With the inauguration of the FTA, the duties imposed on U.S. imports from Canada began to decline, thus further weakening the price competitiveness of (dutiable) imports in some of Mexico’s most promising export-oriented industries.35 This is not to suggest that the discriminatory effects of the FTA were only a product of its tariff provisions; also important was the agreement’s dismantling of numerous nontariff impediments to Canada-U.S. trade. To be sure, the nontariff impediment of greatest concern to the Canadians—the “harassment” potential created by America’s administrative trade remedy legislation—was of relatively little concern to Mexican exporters, who, as noted earlier, were infrequent targets of U.S. antidumping and countervailing duty actions. The fact that Chapter 19 of the FTA provided authorities in Canada with a vehicle for challenging the legality of U.S. trade remedy actions was thus of lesser consequence to the Mexican side than were the supranational dispute settlement procedures set forth in Chapter 18 of the agreement (the provisions for which were subsequently “carried over” into NAFTA). The FTA’s formalizing of these procedures had already gone some distance toward mitigating the nuisance caused by U.S. border controls, quotas, technical standards, health and safety measures, preferential public procurement policies, and the like, all of which had served to disrupt CanadaU.S. trade in the years before the FTA came into force and were threatening to do the same for Mexico-U.S. trade in the years to come. Was Mexico a Victim of Coercion? To claim that Mexico sought a free trade agreement with the United States in order to mitigate the potential diversionary effects of the Canada-U.S. FTA is not to say that Mexico was coerced into taking this action. To describe NAFTA as a case of “deep integration through leverage” (Haggard 1995) is to fail to appreciate the voluntary nature of Salinas’s decision. At no point was Mexico’s ruling elite ever subjected to coercive pressure. Quite the contrary, neither of the Canada-U.S. Free Trade Agreement’s two sponsors had any intention of modifying Mexico’s behavior, even if their go-it-alone initiative was to have precisely that effect. 34 SITC items marked by a dagger in Table 7-5 are those for which, in 1987, U.S. imports from Canada exceeded $100 million. The major exceptions were television receivers, jewelry, medical instruments, and insulated wire. In each of these sectors, the value of U.S. imports from Canada was less than half that of competing imports from Mexico. 35 Mexican exporters also stood to lose market share in Canada, in this case because of increased post-FTA competition from duty-free American exports. Most analysts rightly downplay the importance of this second diversionary effect, however, owing to Mexico’s small share in Canada’s total imports (1.0%) and exports (0.4%). See Figure 6-1.
N A F T A
A N D
B E Y O N D
139
Does this mean that authorities in Canada and the United States were indifferent to Salinas’s overture? Not at all. In the United States, NAFTA was immediately hailed by Reagan’s successor, George Bush, as a way of insulating the market-oriented reforms of de la Madrid and Salinas from the vagaries of Mexican politics.36 And by providing the U.S. with a “regional fallback,” NAFTA may also have strengthened Bush’s bargaining position in the Uruguay Round negotiations of the GATT.37 But to say that the president welcomed NAFTA is not to say that he pressured the Mexican government into signing the accord. Though Bush viewed the pact as an improvement over the earlier bilateral agreement, he did not regard it as a top priority. Nor, given Mexico’s marginal role in the overall U.S. economy, is this particularly surprising. To be sure, Mexico’s large population—nearly three times that of Canada—meant that a future economic takeoff had the potential to dramatically expand export opportunities for American producers. In 1989, however, Mexico’s total market was less than half the size of Canada’s and only about 3.6 percent of the combined Canada-U.S. market.38 Moreover, while Mexico enjoyed the distinction of being the third largest U.S. trading partner, American trade with Mexico in 1989 contributed only 1.0 percent of total U.S. GDP.39 As for Congress, more than a few U.S. representatives and senators were wary of supporting an agreement that would allow low-wage Mexican labor to compete directly against businesses in their districts.40 In short, though NAFTA was ultimately signed by the U.S. president and (after a prolonged and stormy debate) ratified by both houses of Congress, the impetus for the accord did not come from the U.S. side. Nor did it come from Canada, which was decidedly unenthusiastic about the whole affair. As I argue later in this chapter, it is virtually inconceivable that Prime Minister Jean Chrétien, whose Liberal government ratified the pact shortly after taking office in 1993, would have been willing to “trilateralize” the Canada-U.S. agreement had it not been for the prospect of a separate Mexico-U.S. bilateral agreement. In the end, therefore, Salinas’s 1990 decision to seek a free trade agreement was made freely—the product of voluntary, albeit constrained, choice rather than of “bullying” by the United States or (least of all) Canada. 36
While officials in the Bush administration were heartened by the PRI’s abandonment of the protectionist policies it had embraced during much of the postwar period, there was no guarantee that President Salinas’s successor—or even, for that matter, Salinas himself—would remain committed to the new approach. After NAFTA, however, Mexican authorities would presumably find it more difficult “to reverse . . . course and turn Mexico back toward economic isolationism” (Bosworth, Lawrence, and Lustig 1992, 1). See also Haggard 1995, 94; Kahler 1995, 21; Krugman 1993b; and Tornell and Esquivel 1995. 37 Hurrell 1995, 271. 38 World Bank 1997. 39 Lustig et al. 1992, 71, table 1. 40 See, e.g., Wiarda 1994; O’Halloran 1994.
140
N O R T H
A M E R I C A N
T R A D E
Mexico’s Preference for the “Noncooperative” Status Quo Suppose I am right that defensive considerations—specifically, the desire to mitigate the discriminatory effects of the FTA—weighed more heavily in Mexico’s decision-making calculus than did the more “proactive” motivations we are used to reading about in the cooperation-under-anarchy literature. Even if this is correct, it does not automatically follow that the inauguration of the Canada-U.S. agreement, by prompting Mexico to seek a free trade agreement of its own, undermined Mexico’s interests. To the contrary, the notion that NAFTA was somehow “bad” for Mexico might seem to fly in the face of basic economic theory. Assessing the Economic Effects of Mexico’s Participation in NAFTA The question of how regional trading blocs affect the long-term welfare of their members has been the focus of considerable work within economics.41 Much of this work grows out of a now-classic study by Jacob Viner, The Customs Issue (1950), which poses the essential question: Is within-bloc liberalization a good thing or a bad thing? TRADE CREATION VERSUS TRADE DIVERSION
On the one hand, noted Viner, the removal of tariff barriers by members of a customs union causes trade to increase (except where huge distances and high transportation costs render trade prohibitively expensive). Part of this increase occurs when consumers switch from their own high-cost domestic producers to lower-cost producers in other members of the union. The replacement of inefficient domestic production by low-cost imports, called “trade creation,” enables consumers throughout the union to immediately enjoy the benefits of lower-priced goods and services. Over the longer term, as production is concentrated in those areas in which each partner holds a comparative advantage, trade creation also generates gains in efficiency and productivity. On the other hand, Viner went on to argue, part of the increase in trade between members of such a union may result from “trade diversion,” the displacement of more efficient sources of supply outside the union by less efficient suppliers within it. Although NAFTA is not a customs union, the same principle would apply. Thus, upon gaining preferential access to American and Canadian markets, Mexican consumers might be inclined to purchase goods and services from the United States and Canada that they had previously imported from a lower-cost producer in a fourth country, such as Japan. Even if duty-free U.S. and Canadian imports were cheaper than dutiable Japanese imports, a switch to American and Canadian suppliers could still impose net costs on Mexican citizens, since the duties the Mexican Treasury formerly collected 41 See, among many others, Frankel 1997; Krugman and Obstfelt 1994, chap. 6; and P. Wonnacott and Lutz 1989.
N A F T A
A N D
B E Y O N D
141
on Japanese imports would have to be made up somewhere else, presumably by increasing the burden on Mexican taxpayers. To predict whether a country like Mexico would benefit from an expansion of trade, Viner would have us take the expected gains from trade creation (minus any transitional adjustment costs associated with the reallocation of resources) and subtract the anticipated losses from trade diversion. IMPLICATIONS FOR MEXICO’S ACCESSION TO NAFTA
Mexico’s entry into NAFTA passes this test with flying colors.42 Given the complementarities between the Mexican economy, on the one hand, and the U.S. and Canadian economies, on the other, the potential for trade creation is (obviously) quite large. As Mexico focuses on the production of low-wage, labor-intensive goods and services while the United States and Canada concentrate on high-wage, research-intensive production, all three countries should eventually be better off. Nor are the resulting gains for Mexican citizens likely to be offset by the effects of trade diversion from non-NAFTA countries, since Mexico was already obtaining well over half of its imports from Canada and (especially) the United States before NAFTA was ever instituted.43 Furthermore, many economists now believe trade diversion is not such a bad thing after all.44 Trade diversion presents problems when a country replaces a low-cost supplier outside the free trade zone with a higher-cost supplier inside. But if NAFTA makes it possible for producers in United States and Canada to manufacture on a larger scale, the resulting savings could eventually reduce the costs of American and Canadian suppliers down to, or possibly even below, those of the (initially) lower-cost suppliers outside the region. On this basis, NAFTA stands to improve the long-term economic health of all three partners.45 In fact, neoclassical trade theory would predict that Mexico, having the smallest of the three participating economies, would gain most. Whereas NAFTA’s impact on the U.S. and Canadian economies should be relatively small, the post-NAFTA alteration in Mexico’s preexisting price structure, and hence the potential for major efficiency gains, should (in principle, at least) be dramatic.46 42 Studies highlighting the potential for mutual economic gain include Globerman and Walker 1993; Helleiner 1993; Hufbauer and Schott 1992 and 1993; Kehoe 1994; Lustig, Bosworth, and Lawrence 1992; Weintraub 1990c; and R. Wonnacott 1991b. 43 Trade diversion would be a greater concern in the case of Chile’s accession to NAFTA. Unlike Mexico, which obtained less than a third of its total imports from outside the United States, imports from the U.S. constituted only about a quarter of Chile’s total imports. 44 See, e.g., Corden 1972; R. Wonnacott 1996. 45 Ros (1992) suggests that the anticipation of these traditional trade gains was less important to Mexican decision makers than were the agreement’s anticipated effects on capital flows. Yet, as I discussed earlier, inflows of long-term investment capital into Mexico were unlikely to be forthcoming unless prospective investors could be assured of maintaining their access to customers in the United States. In practice, therefore, the two aims were inextricably linked. 46 This conclusion would be warranted even if regionalism were but a “second best” arrangement. Krugman 1991b. The first best alternative would be a true multilateral trading regime,
142
N O R T H
A M E R I C A N
T R A D E
BROADENING THE ANALYSIS
In recent years, however, mainstream economists have begun to question whether the task of distinguishing free trade winners from free trade losers is as straightforward as the line of analysis sketched out above would suggest.47 One problem is that NAFTA does more than eliminate “at-the-border” barriers; as noted earlier, it also provides for the establishment of an elaborate system of supranational governance. This is problematic insofar as the rules and standards that issue forth from this new arrangement may not work to the advantage of all three member states.48 NAFTA’s side accords on labor issues and the environment are a case in point. Inasmuch as the labor accord calls for Mexico to align its standards on wages and working conditions with those of the United States and Canada, its principal effect might well be to raise production costs in the export-oriented industries in which Mexico would otherwise enjoy a comparative advantage.49 By compelling Mexico to raise its domestic environmental standards (or to more strictly enforce those it has already adopted), the environmental accord could have much the same effect. Another problem with conventional economic arguments is their neglect of “economic geography.”50 The fact that NAFTA promotes economies of scale is normally seen as a desirable feature of the agreement. Yet, as businesses move to exploit the cost advantages facilitated by larger production runs, they may also decide to shut down their Mexican operations and concentrate production in the United States, the member of NAFTA with the largest domestic market. As Roberto Bouzas and Jaime Ros (1994b, 12) point out, the size and direction of these dynamic effects “will ultimately rest upon the uncertain balance of ‘virtuous’ as opposed to ‘vicious’ circles of investment, productivity, and growth, about which economic analysis provides all but precise answers.” which would ostensibly be better not just for members of NAFTA but for the global economy as a whole. In practice, however, multilateral arrangements are frequently plagued by large-N problems, special exceptions, and other impediments. Bhagwati (1991) argues that even with its problems, the current multilateral system of nondiscriminatory rules and disciplines is preferable to the new regionalism, which he sees as a distant second; see also Bhagwati and Panagariya 1996a. 47 Bhagwati and Panagariya (1996b) provide an overview of recent research. For critiques focused specifically on NAFTA, see Grinspun and Cameron 1993; Koechlin and Larudee 1992; Randall 1992; and Robinson 1993. 48 Haggard 1995; Lawrence 1996. 49 Even before NAFTA’s entry into force, some analysts (e.g., Castañeda 1993a; Robinson 1993) expressed concern that Mexican wages were not low enough! While the average hourly compensation of the typical American production worker in 1991 was eight times that of her Mexican counterpart—$15.60 versus $1.95 in U.S. dollars—the productivity of American labor was more than eight times that of Mexican labor (as measured by the value added per employee in U.S. manufacturing and Mexican maquiladora operations, respectively). See Hufbauer and Schott 1993, 173. 50 Krugman 1991a.
N A F T A
A N D
B E Y O N D
143
Most importantly, however, the real question is not whether NAFTA will encourage a more efficient allocation of Mexico’s economic resources; it is whether Mexico’s opening to world markets might have been accompanied by less economic and political upheaval if the country’s leaders had not been “forced” to deviate from their pre-NAFTA trajectory. All else equal, the Mexican government, faced with a growing risk of social unrest, would have been inclined to pursue a more gradual approach to market liberalization. But though this approach might have limited the burden of adjustment for Mexico’s newly enfranchised citizens, a gradualist strategy was fundamentally incompatible with Mexican participation in NAFTA.51 A Big Deal: How NAFTA “Forced” Mexico’s Accelerated Liberalization Had it not been for the world’s devolution into regional trade blocs, a gradual transition to open markets would have been well within Mexico’s reach. Since 1986, the country’s rulers had been liberalizing under the auspices of the GATT.52 For Mexico to continue down this multilateral path, however, would have been to risk losing access to its all-important U.S. market. Instead, Mexico entered into negotiations for a free trade agreement with the United States, thereby committing itself to speeding up—not merely preserving—the trade reforms it had initiated in accordance with its GATT obligations.53 No one would dispute that certain elements of Mexican society stood to benefit from an acceleration of the liberalization process. That NAFTA was championed by Mexico’s internationally competitive export producers comes as little surprise, for instance.54 The fact is, however, that Mexico’s 51 Even Adam Smith, the world’s most celebrated proponent of market-based reform, believed that nations could make too hasty a transition to free trade. “It may sometimes be a matter of deliberation,” Smith wrote in The Wealth of Nations ([1776] 1965, 435), “how far, or in what manner it is proper to restore the free importation of foreign goods . . . when particular manufacturers, by means of high duties or prohibitions upon all foreign goods which come into competition with them, have been so far extended as to employ a great multitude of hands. Humanity may in this case require that freedom of trade should be restored only by slow graduations, and with a good deal of reserve and circumspection.” 52 Aside from permitting members to protect infant industries through higher tariffs, the GATT’s multilateral trade regime—which Mexico joined on August 24, 1986—afforded certain exports from developing countries special tariff preferences in the markets of the United States and other advanced countries (Sapir and Lundberg 1984). In 1989, Mexico became the largest beneficiary of this Generalized System of Preferences (GSP), which eliminated duties on 9.3 percent of its total exports (Krueger 1993, 230 n. 52). In addition to enjoying “special and differential treatment” under the GSP, Mexico also benefited from provisions in the U.S. tariff code allowing commodities assembled abroad using U.S. parts and components to be imported back into the United States with duty paid only on the value added during assembly As a vehicle for Mexican trade reform, this maquiladora program had the advantage of concentrating the effects of liberalization around border areas far from Mexico City. 53 The specific terms of Mexico’s commitments under NAFTA are discussed below. 54 Cf. Poitras and Robinson 1994, 14. Peak business organizations in support of the accord included the Mexican Employers’ Confederation (Confederación Patronal de la República Mex-
144
N O R T H
A M E R I C A N
T R A D E
resort to import substitution as a strategy for economic growth—a strategy that had oriented Mexican policy for much of the previous fifty years—had spawned a large number of businesses whose very survival depended upon steady infusions of state subsidies and the maintenance of protectionist barriers to foreign competition. And as one might have expected, these businesses were strongly opposed to NAFTA. Thus, “as Mexico’s signing of the NAFTA drew nearer, and the prospect of increased competition became imminent, representatives from the more vulnerable economic sectors (textiles, trucking, and government procurement) openly were complaining about the short time line set for further trade exposure and about the lack of adequate financing to facilitate economic adjustment.”55 In response, it might be argued that NAFTA, though it posed a threat to Mexico’s more vulnerable economic sectors, was unambiguously positive in its net effects. Analysts who take this position often view the agreement, along the lines suggested earlier, as a device for binding future Mexican governments to the free trade policies that the country had been pursuing since the late 1980s. Had it given “credibility, permanence, and prominence” (Lawrence 1996, 67) to Mexico’s domestic economic reforms, NAFTA would have markedly increased the propensity of multinational corporations and the global financial markets to invest in Mexico over the long haul, and it is quite likely that the society’s net welfare under the agreement would have been enhanced. A massive infusion of foreign direct investment could have stimulated economic growth and facilitated the creation of thousands of new jobs, more than compensating for the losses that Mexican citizens—and, more important for present purposes, their “elected” representatives in the PRI—might otherwise have incurred. There is little evidence, however, that officials within the PRI expected NAFTA’s lock-in effects to precipitate a dramatic inflow of long-term investment.56 Nor, in the end, did they actually do so (see below). This should not be surprising inasmuch as Mexico’s ability to attract foreign direct investment (FDI) was contingent on the maintenance of domestic political staicana, or COPARMEX), the National Association of Chambers of Commerce (Confederación de Cámaras Nacionales de Comercio, or CONCANACO), and the National Association of Importers and Exporters (Asociación Nacional de Importadores y Exportadores de la República Mexicana, or ANIERM). Mexico’s largest association of small- and medium-sized businesses, the Confederación de Cámaras Industriales (CONCAMIN), also came out in favor of NAFTA, although some members broke ranks with the CONCAMIN leadership, complaining that the organization had not accurately represented their interests during the NAFTA negotiations. Pastor and Wise 1994, 465. 55 Pastor and Wise 1994, 465. The notion that Mexico’s trade opening severely damaged the interests of large segments of Mexican society is not inconsistent with traditional economic theory. Few economists would dispute that free trade often generates substantial costs of adjustment in the short term, and, even in the long run, may exacerbate preexisting disparities in the distribution of income within the countries that engage in it. For evidence that income inequality rose sharply in Mexico between 1984 and 1994, see Pastor and Wise 1997. 56 Ros 1992, 72.
N A F T A
A N D
B E Y O N D
145
bility, and even if Mexico’s political establishment remained committed to the accord, the possibility of such instability could not be ruled out. In fact, rather than quelling popular opposition to the government’s market-based reforms, NAFTA seemed more likely to exacerbate Mexico’s political turmoil, introducing new tensions into an already volatile political environment. In this sense, liberalization through the GATT, precisely because it was less sweeping, might indeed have provided a more effective device for entrenching Mexico’s outward-looking economic policies. As noted earlier, Mexico’s ability to lure FDI was also dependent on the promise of continued access to the U.S. market. Yet, whatever its effects on Mexico’s political climate, NAFTA offered Mexican exporters little protection against the sudden closure of their largest foreign market. Such a closure would have had grave results: if at some point in the future the United States were to erect new barriers to Mexican imports, many of the costs that Mexican exporters and their financial backers had incurred in purchasing new plant and equipment to service the U.S. market would be suddenly, and irretrievably, lost.57 To be sure, U.S. dependence upon Mexico had increased over the course of the 1980s; insofar as the Mexican economy afforded U.S. producers a source of cheap labor close to home, there was less danger of the United States putting up barriers to trade with Mexico than there might otherwise have been. Still, the opportunity costs of closure for the United States remained relatively low, in part because Mexico-U.S. economic relations were still in their infancy, but also, more importantly, because the existence of numerous lower-wage economies elsewhere in Latin America vitiated Mexico’s attractiveness as a source of cheap labor. NAFTA’s very existence did provide some insurance against “reneging” by the United States.58 Most importantly, NAFTA, like the earlier Canada-U.S. pact, incorporated a sophisticated (and essentially identical) set of dispute resolution mechanisms, one purpose of which was to discourage parties to the agreement from using ambiguities in the accord as a pretext for engaging in unwarranted protectionism.59 By themselves, however, these mecha57
Krasner 1992; Yarbrough and Yarbrough 1992. See, e.g., Kahler 1995, 21. 59 Like the Canada-U.S. Free Trade Agreement that preceded it, NAFTA provides for different arbitration procedures depending on the nature of the dispute. Disputes concerning the passage of new trade legislation are covered by Chapter 20, the subject material of which is essentially equivalent to Chapter 18 of the earlier bilateral agreement. The provisions outlined in Chapter 19 of NAFTA specify the earlier-mentioned arrangements for handling disputes over the application of U.S. antidumping and countervailing duty laws, while (in a departure from the FTA) Chapter 14 spells out formal dispute settlement procedures for trade in financial services. In addition, procedures for addressing disputes that arise between a private investor and a NAFTA government—e.g., because the government or a government-run monopoly has failed to offer proper compensation in the event of an expropriation—are set forth in Chapter 11. Finally, the NAFTA side accords spell out rules for handling disputes concerning labor issues and the environment. For further discussion of NAFTA’s institutional arrangements, see Horlick and DeBusk 1993; D. Huntington 1993; Winham 1993. 58
146
N O R T H
A M E R I C A N
T R A D E
nisms were not sufficient to deter U.S. officials from disrupting Mexico-U.S. trade at a politically more opportune moment in the future. In the case of disagreements concerning the application of U.S. antidumping and countervailing duty legislation, the juridical constraints on U.S. action were actually weaker under NAFTA than they had been under the original bilateral agreement.60 Coming to Terms: To What, Exactly, Did Mexico Agree? Even if NAFTA provided only a modest safeguard against future disruptions in Mexico-U.S. trade, one might argue that the agreement nonetheless afforded Mexico numerous trade concessions and, for this reason, represented an improvement over the preexisting status quo. In fact, a close inspection of the NAFTA accord’s provisions reveals that in areas ranging from trade in autos to the treatment of foreign direct investment, Mexico was required to liberalize its economy far more aggressively than it had done in the past.61 Among other things, for example, NAFTA called upon Mexico to abolish virtually all of its preexisting duties on U.S. capital equipment, high-tech goods, and other products not already covered by GATT guidelines. Was this a major change? Absolutely, for when Mexico agreed to the accord (in 1990), its tariff rates on dutiable U.S. imports averaged 13 percent, or “only” 11 percent if weighted by production.62 What about the requirement that Mexico dismantle its elaborate system of nontariff barriers? Was that also a significant departure? To argue that it wasn’t, one would have to assert that by the end of President de la Madrid’s term of office, Mexico’s traditional practice of requiring import licenses had already been substantially scaled back. There is, of course, some truth to this. According to one study, the proportion of Mexican domestic production protected by licensing requirements declined from more than 90 percent in 1985 to only about 23 percent in 1988.63 That said, when Mexico agreed to NAFTA, licenses continued to be required for about one-fifth of all Mexican imports in terms of value.64 Consider, too, that in joining NAFTA, Mexico agreed to treat Canadian and American investors exactly the same as it did its own investors. In the lexicon of trade policy, the measures outlined in Chapter 11 of the accord extended the principle of “national treatment” to most new investments (as was also true for the FTA) as well as to most existing investments (an innovation in NAFTA).65 In short, Mexican authorities have relinquished the 60
Lipsey, Schwanen, and Wonnacott 1994, 130 – 32; J. Smith 1994. A chapter-by-chapter analysis of the accord can be found in Hufbauer and Schott 1993. See also Lipsey, Schwanen, and Wonnacott 1994. 62 Bosworth, Lawrence, and Lustig 1992, 2– 3, 8. 63 Bosworth, Lawrence, and Lustig 1992, 8. 64 Weintraub 1990a, 103. 65 Chapter 11 also narrowed the criteria that the Mexican government could use in disal61
N A F T A
A N D
B E Y O N D
147
right to subject Canadian and U.S. investments to “special scrutiny,” a common practice during pre-NAFTA days.66 The agreement’s text also included a number of seemingly narrow but, upon closer inspection, quite sweeping provisions. Annex 300-A is a good example. Specifically designed to pry open Mexico’s highly protected automotive market, the terms set forth in the annex called upon Mexico to gradually reduce—and eventually eliminate—the industry’s long-standing domestic-content and trade-balancing requirements.67 It is true that negotiators from Canada and the United States agreed to lengthen the phase-out time for the scheduled elimination of Mexican tariffs on particularly sensitive items (the implementation schedule incorporated into the FTA continues to apply between Canada and the United States). Another concession to Mexican concerns was the treatment of PEMEX, Mexico’s enormous state-owned oil industry.68 In a departure from the terms of the original FTA, Annex 602.3 of the trilateral agreement allowed Mexico to retain its ban on foreign ownership of oil and natural gas reserves. Moreover, unlike Canada’s nationalized oil and gas corporation, Petro-Canada, PEMEX was not required to offset the burden imposed on North American energy importers should another OPEC-like short-supply situation arise at some point in the future.69 lowing foreign acquisitions, removed previous limits on the amount of equity that Canadian and American nationals could hold in Mexican enterprises, and prevented Mexico from seizing assets held by Canadians or Americans—or “implicitly appropriating” them via huge royalty increases—without full compensation. Canadian and U.S. investors who believed they were being held up to a higher standard could now seek binding arbitration from a trinational review panel. 66 See particularly Ros 1992. Chapter 12 of the agreement extended the national treatment principle to competition in services, and Chapter 14 applied it to Mexico’s highly sheltered financial sector. As a result, U.S. and Canadian banks and securities firms are now permitted to compete in the Mexican market as long as they adhere to the rules and regulations that apply to Mexico’s own banks and brokerages. 67 The liberalizing measures of the 1980s had temporarily waived the Mexican auto industry, which continued to be governed by the (protectionist) Mexican Industry Automotive Decree of 1962 (Peres Núñez 1990, 17). For a description of this earlier arrangement, see Hufbauer and Schott 1992, 215–19. The auto decree’s domestic-content requirements have now been replaced by a regional (i.e., North American) content requirement. Thus, NAFTA stipulates that 62.5 percent of the value added in cars, light trucks, and buses must originate within North America, though not necessarily within any particular country. As a result, the Mexican import market in autos and auto parts has been effectively reserved for U.S. suppliers. The move to a regional content requirement also has the effect of excluding much of the U.S. transplant and Canada-based production from the Mexican market. Whalley 1993, 357– 60. 68 In 1990, PEMEX accounted for some 20 percent of the Mexican government’s total revenues. 69 While allowing Petro-Canada to continue to operate, the FTA had prohibited Canada’s federal government from using its control over the Canadian energy industry to restrict energy exports to the United States. Specifically, the FTA had guaranteed American consumers (at least) the same share of total Canadian production they had received over the previous three years. If for some reason (e.g., an unexpected energy shortfall) Canadian exports to the United States were to decline, the FTA obligated Canadian consumers to reduce their own consumption.
148
N O R T H
A M E R I C A N
T R A D E
Conciliatory though these measures were, it is important to keep things in perspective. While NAFTA did lengthen the phase-out time—in some cases up to fifteen years—for the elimination of some tariffs, these were the exceptions. In fact, the schedule was to be “virtually completed in ten years, with the bulk of the tariffs . . . substantially reduced or eliminated in earlier years.”70 And while PEMEX may have been allowed to retain its monopoly on oil exploration, NAFTA opened the bidding on contracts for energyrelated goods and services to all North American firms, thereby ensuring that Canadian and American drilling companies would have an opportunity to share in any future profits the Mexican oil industry might generate.71 The fact that Canada and the United States did not substantially tone down the liberalizing thrust of their 1988 agreement is all the more striking when one considers that Mexico, as a developing country, was likely to experience far greater dislocations as a consequence of its entry into NAFTA than would Canada or the United States, each of which had already undergone the transition from a mostly agrarian to an industrial and service-based economy. Developing countries typically start out with higher trade barriers, and so must undergo a more extensive reallocation of resources. In most cases, moreover, such countries have a particularly hard time adjusting to “exogenous” shifts in supply and demand.72 There are a number of reasons for this, ranging from poor physical infrastructure to structural rigidities in Third World labor markets. The lower mobility of labor in developing countries means that workers employed in import-competing sectors cannot as readily switch to jobs located in export-producing firms. And then, too, there is the problem that for individuals who live near the level of subsistence, even temporary economic “disturbances” can be life-threatening.73 After all, few developing countries have the institutional or financial resources to compensate displaced workers. In the case of Mexico—and notwithstanding the fact that the PRI developed a National Solidarity Program (PRONASOL) in 1989—the country’s social safety net remained full of holes.74 But though the per capita income gap between the United States and Mexico was substantially greater than the gap between the wealthiest and poorest members of the European Union (Figure 7-2), the United States never once seriously considered the idea of establishing a North American equivalent of the EU’s Structural Funds. The contrast with German unification is also instructive: Although East Germany’s per capita income was nearly 60 percent of West Germany’s prior to unification—for comparison purposes,
70
Lipsey, Schwanen, and Wonnacott 1994, 26. Watkins 1993. 72 Krasner 1985. 73 Scott 1976; Popkin 1979. 74 Designed to transfer wealth to disadvantaged groups, PRONOSAL was never considered more than a stopgap measure. See Castañeda 1993b, 60; Kaufman, Bazdresch, and Heredia 1994; and Morris 1992, 32– 34. 71
N A F T A
A N D
B E Y O N D
149
Figure 7-2. Ratios of Real Income per Capita in Mexico, the United States, and Western Europe, 1960 –1988 (PPP-adjusted; population-weighted)
Mexico’s per capita income in 1990 was barely more than 10 percent of that of the United States—the West German government agreed to spend in excess of U.S.$100 billion a year in order to provide adjustment assistance for East Germany’s 17 million inhabitants.75 Yet when President Salinas announced in 1992 that he would be asking the incoming Clinton administration to establish a (considerably smaller) economic support fund, the U.S. response was tepid at best. To this day, aside from devoting resources for infrastructural projects and environmental clean-up in the Mexico-U.S. border region, the United States has shown relatively little interest in defraying the costs of Mexico’s more onerous adjustment burden.76 From NAFTA to Chiapas and the Demise of One-Party Rule While it is too early to say precisely what effect NAFTA will have on Mexico over the long run, the years between the announcement of the agreement in 1990 and its entry into force in 1994 were marked by a rapid deterioration in the Mexican political economy.77 This is significant insofar as the lib75
Grunwald 1993. Castañeda 1993a, 75; Hufbauer and Schott 1993, 165. The United States and Canada did establish a North American Development Bank to facilitate borrowing from private capital markets. 77 Not only did the Mexican economy fail to take flight during this period, but the benefits of what little growth there was—about 2.5 percent a year on average—were very unevenly dis76
150
N O R T H
A M E R I C A N
T R A D E
Figure 7-3. Mexico’s Attractiveness as a Site for Foreign Investment, 1987–1996
eral reforms undertaken during this period were relatively moderate by NAFTA standards. And while the news that Mexico was prepared to participate in a free trade agreement with Canada and the United States may have mitigated the diversionary consequences of the Canada-U.S. agreement, long-term investors—as opposed to those buying pesos to acquire Mexican stocks and bonds that could be sold again quickly and easily—continued to tributed: “Developing country statistics on both unemployment and income distribution are fairly unreliable,” notes Krugman (1995, 41), “but there is not much question that even as Latin American stock markets were booming, unemployment was rising, and the poor were getting poorer.” On the distributional effects of Mexico’s trade opening, see Pastor and Wise 1997.
N A F T A
A N D
B E Y O N D
151
be circumspect.78 Revealingly, net long-term foreign investment showed only a slight increase after 1990 (Figure 7-3). Already tenuous, Mexico’s economic and political situation deteriorated further in 1994, the year in which NAFTA was officially inaugurated. It is not at all clear that either of the PRI’s two main rivals—the Democratic Revolutionary Party (the Partido de la Revolución Democrática, or PRD) or the National Action party (Partido de Acción Nacional, or PAN)—would have handled things differently had they, and not the PRI, been in power.79 But that did not stop NAFTA from becoming a rallying point for popular opposition to the Mexican government’s new market-oriented approach; indeed, by 1994, opposition demands for radical political reform had begun to resonate with large elements of the Mexican public.80 Not that the economic uncertainties and political controversy engendered by the agreement were totally unanticipated by the ruling party. An indication of the PRI’s sensitivity to anti–free trade attitudes within the Mexican electorate was its unwillingness to openly debate the potential effects of NAFTA.81 If the PRI had been expecting a political windfall, it is doubtful that it would have gone to such great lengths to hide its initial overtures to the United States from ordinary Mexican citizens. In the words of one observer, “The Mexican government [gave] NAFTA negotiations the equivalent status of a national security affair, keeping information almost a state secret, preventing any meaningful public debate, maintaining a close vigilance on its opponents, and transmitting only general propaganda messages to the public.”82 A major blow to the Mexican government came when—on the very day NAFTA was scheduled to take effect—a peasant insurgency erupted in the southern province of Chiapas. Precipitating the uprising was a group calling itself the Zapatista Army of National Liberation, or EZLN. While EZLN fed 78
Castañeda 1993b, 63 – 65; Peres Núñez 1990, 20; Ros 1992, 67. The PAN, though supportive of free trade in principle, “was quick to point out the dangers of negotiating such a historic agreement without ensuring that it contained sufficient guarantees of not only the rights of workers but also the principles of national sovereignty.” Poitras and Robinson 1994, 20 –21. 80 Cook, Middlebrook, and Molinar 1994a; Cornelius 1996. 81 There are few reliable indicators of public attitudes on NAFTA. Some polls suggest that, notwithstanding their deep-rooted suspicions of the United States, most Mexican citizens were complacent about the prospect of increasing economic integration; see, e.g., Hellman 1993. In contrast, a poll conducted by Mexico’s largest private bank, Banamex, found that popular support for NAFTA declined sharply between November of 1990, when 67 percent of those surveyed said they believed the pact would be “beneficial,” to April of 1993, when only 33 percent of the survey’s respondents expressed this view. Heredia 1994, 20 n. 7. 82 Aguilar Zinser 1993, 207– 8. Aguilar Zinser notes that the government kept the lower house of the Mexican legislature, where opposition deputies had controlled 45 percent of the seats since 1991, from having any input into the trade negotiations. Instead, deliberations were confined to the Senate, a body in which the PRI held 61 of the 64 seats. On the closed nature of the NAFTA debate within Mexico, see also Heredia 1994, 17–20, and Pastor and Wise 1994, 480. 79
152
N O R T H
A M E R I C A N
T R A D E
off the anxieties and frustrations that the government’s previous agricultural reforms had generated among the province’s largely rural population, NAFTA itself also played a role. Even if the NAFTA negotiations were not themselves responsible for the transformation of Mexican agriculture—as noted earlier, the introduction of markets into Mexico’s previously sheltered agricultural sector was underway before the NAFTA talks began—the agricultural provisions found in Chapter 7 of the accord took these reforms one step further.83 Once the accord entered into force, for example, all nontariff barriers to agricultural imports were converted to tariff-rate quotas (TRQs), many of which, including, most importantly, those on imported corn and beans, were scheduled to be phased out over a fifteen-year period. Even more important than these (or any other) specific provisions, NAFTA’s very existence provided a convenient target for Mexican peasants concerned about their economic future. Hoping to exploit these concerns and, in so doing, fan the flames of rural discontent, the EZLN military leader, Subcommandante Marcos, declared NAFTA a “death certificate for the Indian peoples of Mexico.”84 From the standpoint of the Mexican government, the upsurge of revolutionary activity in Chiapas was a major problem—less because of Chiapas itself than because of the possibility of the unrest spreading. With Mexico’s pre-NAFTA liberalization efforts having failed to bring about an immediate improvement in the lives of ordinary Mexican workers (not just rural peasants), long-term foreign investors began to fear that Chiapas was only the beginning. Already apprehensive about the prospect of a resurgence in U.S. protectionism, the financial markets now had to factor in the possibility of domestic upheaval and social unrest on an unprecedented scale.85 Adding to these fears was the assassination a few months after the outbreak of the Chiapas rebellion of Luis Donaldo Colosio Murrieta, the cabinet official whom Salinas had tapped to be his successor. Not surprisingly, inflows of foreign direct investment into Mexico began to decline at that point—and the picture has not improved much since then. After Chiapas, the government faced mounting pressure both to raise yields on peso-denominated assets (by raising interest rates) and to reduce their price (by lowering the peso’s value relative to the U.S. dollar). In a risky gamble, the government delayed pursuing either strategy to full effect, instead drawing upon its foreign reserves to finance the trade deficit that Mexico had accumulated since the early 1980s.86 If Mexico’s central bank had 83 NAFTA incorporates three separate bilateral agreements on agriculture. For a detailed discussion of these agreements, see Josling and Barichello 1993. 84 Cook, Middlebrook, and Molinar 1994a, 25 n. 34. 85 Cook, Middlebrook, and Molinar (1994a, 25) note that the Zapatista movement received a “groundswell of support . . . from across the social spectrum.” 86 By 1994, Mexico’s trade deficit had reached 8 percent of the country’s total economic output. Bradsher 1995.
N A F T A
A N D
B E Y O N D
153
kept pushing up rates on treasury bonds, which were already hovering around 18 percent, the result would have been a slowdown in Mexico’s overall rate of economic growth. If the government had chosen instead to devalue, this would have further reduced the purchasing power of Mexico’s already-disgruntled electorate in the run-up to the 1994 presidential election. (It would also have tarnished the international reputation of President Salinas, who was at that time a candidate to head the new WTO.) As the Mexicans were to discover in December 1994, however, there were limits to this strategy of avoidance. By that time, the Mexican central bank was dangerously close to running out of reserves. So on December 22, the government bowed to the inevitable, allowing the peso to float freely against the dollar. The result was an abrupt 50 percent depreciation of the peso and the further impoverishment, over the ensuing months and years, of millions of Mexican citizens. From the perspective of the government, things could have been even worse. The PRI was successful in its attempt to delay the post-Chiapas correction until after the election. It also benefited from the strengthening of the PAN, which had the effect of dividing the anti-PRI vote. In the end, Ernesto Zedillo Ponce de Léon, Colosio’s replacement as the PRI’s presidential candidate, pulled out a victory that many had thought impossible, the PRI winning 50.2 percent of the valid votes cast in the presidential election, compared with the PAN’s 26.7 percent and the PRD’s 17.1 percent. On the other hand, the government emerged from the election badly weakened. Amid the post-election political uncertainty, José Francisco Ruiz Masieu, the secretarygeneral of the PRI, was assassinated on September 28, 1994. Then, coming just three weeks after Zedillo took office, the peso crisis erupted with a vengeance. In retrospect, the political turmoil following the Chiapas uprising and the peso crisis of 1994 thus appears to have had a direct—and deleterious—impact on the political standing of the country’s ruling elite. Although at this point it remains to be seen whether Salinas’s successor, Ernesto Zedillo, will be forced to hand over the presidential sash to a member of another party after his six-year term expires in the year 2000, one thing is clear: The oncehegemonic PRI is no longer invulnerable.
Sideswiped: Canada’s New Leadership Faces a Dilemma Although it has not been subjected to close scrutiny by scholars of international politics, the decision by Canada’s newly elected Liberal government to ratify NAFTA in 1993 presents another explanatory conundrum for proponents of international relations theory’s collective-action paradigm. As we have seen, the Liberals vehemently opposed the implementation of the Canada-U.S. agreement in 1988, their leader, John Turner, swearing he would “tear up” that agreement as soon as the Canadian electorate returned
154
N O R T H
A M E R I C A N
T R A D E
his party to power. By 1993, it would have been economically and politically imprudent for the Liberals to follow through on Turner’s earlier pledge; by the time the Liberals reclaimed a majority of seats in the Canadian House of Commons, the FTA had already been in effect for several years (see chapter 6). NAFTA, however, was a different story. Because it had not yet been ratified when the new Liberal leader, Jean Chrétien, took office as prime minister, the Liberals had an opportunity to veto NAFTA and just stick with the FTA. Instead, they ratified the larger agreement just in time for it to take effect on January 1, 1994. The question is why. Canadian Accession to NAFTA: A Collective Good? One possibility is that by the time the Canadian Liberals recaptured a parliamentary majority in 1993, their deep-rooted attachment to protectionism had been supplanted by a newfound enthusiasm for free trade. In this view, the underlying preferences of Liberal constituencies shifted dramatically over the course of the late 1980s and early 1990s: the same Liberal party whose leadership had once regarded support for free trade as an electoral liability now viewed it as a political asset. If such a sea change in Liberal voters’ attitudes had occurred, it would have begun with the 1988 election—the so-called free trade referendum—in which the Liberal party was crushed by the pro-free trade Conservatives.87 As superficially compelling as it is, however, the “enlightenment” explanation for the Liberals’ support of NAFTA requires us to believe that the Liberals lost the 1988 election because of widespread opposition to the FTA within the Canadian electorate. The reality is much less straightforward, for if the 1988 election had been a true referendum on the free trade question, the governing Conservatives would almost certainly have been defeated—and rather soundly at that. In this respect, the Conservative 1988 victory was achieved “as much in spite of the free trade issue as because of it.”88 It is, of course, quite likely that the government’s earlier support for the FTA bolstered the Conservative party’s support among its “core” constituencies. In the general population, however, the enthusiasm for free trade was not nearly strong enough to cause members of the Liberal party to rethink their long-standing aversion to it. On the contrary, whenever the 1988 campaign addressed the free trade issue directly, the Liberals encroached on the Conservatives’ head start in the polls. 87 When the dust finally settled after the 1988 campaign, Mulroney’s Progressive Conservative party controlled 170 of the 295 available seats (compared with the Liberals’ 82 seats and the NDP’s 43 seats), making Mulroney the first Canadian prime minister since 1953 to achieve two consecutive parliamentary majorities, and the first Conservative prime minister to do so since the nineteenth century. Not only did the Conservatives solidify their position in Quebec, but they even held their own against the left-wing, anti–free trade opposition in the industrial province of Ontario. 88 LeDuc 1991, 363; see also Doern and Tomlin 1991, 238 – 40.
N A F T A
A N D
B E Y O N D
155
To be sure, during the campaign’s first couple of weeks the Liberal party leader toned down his fiery anti-free trade rhetoric, fearing a single-minded focus on the free trade question might split his already badly fractured party. Once he resumed his attack, however, Turner’s standing improved dramatically. With less than a month to go before the election, supporters and opponents of the FTA were at a standoff, with 40 percent on each side and 20 percent still undecided.89 Yet, as subsequent events were to bear out, what support there was for the pact was extremely shallow. Seeing an opportunity to capitalize on the Canadian public’s ambivalence, Turner began looking for the right occasion to thrust the free trade question forcefully onto the campaign agenda. He found it during the second of his three party-leader debates with Mulroney. In an outburst replayed countless times over the next several weeks in Liberal campaign spots and on the national news, Turner accused his rival of betraying the country. “I happen to believe that you’ve sold us out,” he said to the prime minister. “With one signature of a pen, you’ve . . . thrown us into the north-south influence of the United States and will reduce us, I’m sure, to a colony of the United States” (Toronto Globe and Mail, October 26, 1988, A1). Within a matter of days, the Liberals saw their public support soar from 32 to 43 percent, moving their party from last to first place in the polls. In the history of its polling in Canada, the Gallup organization had never before recorded such a large week-to-week shift.90 With the emergence of the free trade question as the campaign’s central issue, the Conservatives were thrown off stride. And yet, in the end, the Conservatives did manage to pull out a victory, ultimately capturing 58 percent of the seats in the Canadian House of Commons. The fact that the Conservatives were able to recover as quickly as they did has been attributed to a host of factors.91 In retrospect, however, two were particularly important. One is that Mulroney successfully downplayed the FTA in the last few weeks of the campaign. Fearing that his party’s pro– free trade position would cost it the election, he shifted the focus after the party leaders’ debates to personality issues, raising questions about the character and competence of the Liberal party leader who had emerged since the debates as his chief rival.92 Yet despite the emergence of Turner’s credibility 89
Johnston et al. 1992, 147. LeDuc 1991, 362. 91 Some analysts stress the importance of the intensive, last-minute advertising campaign that large Canadian corporations waged in support of the government. As long as they refrained from making explicit partisan recommendations, “nonpolitical” entities like the pro-FTA business-sponsored Canadian Alliance for Trade and Job Opportunities could spend their money freely, unconstrained by campaign finance laws. 92 Duncan Cameron 1986, 3 –10; Campbell and Pal 1989. Did the negative assault launched against the Liberal leader succeed in deflecting attention from the FTA? Although it is hard to know for certain, the fact is that large numbers of voters never formed an opinion about the accord. When asked about their attitudes toward free trade, a full quarter of the respondents in the last Gallup poll taken before the 1988 election claimed to be undecided, and fewer than half 90
156
N O R T H
A M E R I C A N
T R A D E
as a major campaign issue, the Liberals still would probably have captured a majority if Canada’s New Democratic Party (NDP) had not siphoned off some of the anti–free trade vote.93 In this respect, Mulroney’s recovery owed a great deal to the strategic problems it posed for the two main opposition parties. After all, the Conservatives won an outright majority with only 43 percent of the popular vote. Even if we dispose of the notion that the Conservatives owed their 1988 victory to a nationwide resurgence of enthusiasm for free trade, that is hardly the end of the story. Another possibility is that there was in fact such a resurgence, but that it came after the free trade “referendum” of 1988, that is, during what was to become the Conservative government’s last term in office. So what does the evidence look like? Was there a groundswell of popular support for free trade during the five-year period before the Conservatives were sent packing? Most certainly not. While calls for abandoning the Canada-U.S. FTA continued to enjoy little political support during these years, many Canadian voters, particularly those on the left of the political spectrum, became progressively less, not more, enamored of free trade. Even if the FTA was not solely responsible for job losses and lower earnings in Canada’s manufacturing sector (see chapter 6), a large segment of the Liberal party’s support base—and, indeed, of the Canadian electorate as a whole—nonetheless continued to feel victimized by the pact. And not entirely without reason: the free trade pact precipitated a surge of U.S. imports in some sectors, as well as the highly visible departure of a number of U.S.-owned plants built to service the Canadian market.94 Many Canadian-owned factories, especially those producing car parts and equipment, also relocated south of the border.95 Perhaps most importantly, even if the rise in joblessness and overall decline in Canadian living standards were mostly due to other factors, the notion that there was a direct cause-and-effect relationship between the FTA’s entry into force in 1989 and the sorry state of the Canadian economy thereafter proved extremely difficult to dispel. It is doubtful, then, that the newly elected Liberal government’s willingness to participate in an enlarged free trade zone can be attributed to a shift in the underlying preferences of its leadership or key political supporters. Though no one in the new cabinet was prepared to abrogate Canada’s ear(48 percent) of those surveyed in another major exit poll reported that the free trade issue affected their vote “a great deal.” 93 Johnston et al. 1992 and 1993. 94 A 1995 study reports that Canadian manufacturing industries with high tariffs in 1988 experienced a disproportionate share of job losses over the 1988 – 93 period compared with lowtariff industries, as did more heavily unionized industries (Gaston and Trefler 1995, 8). The same study finds little evidence of FTA-induced effects on Canadian earnings, however. See also Schwanen 1993. 95 B. Simon 1992.
N A F T A
A N D
B E Y O N D
157
lier bilateral agreement, the underlying anti-free trade preferences of the Liberal party elite were not substantially different from what they had been during the 1988 campaign or, indeed, during virtually the entire period since 1911. The party’s supporters within the Canadian electorate continued to suffer—or at least perceived themselves to be suffering—from free trade with the United States. Nor, finally, can the Liberals’ receptivity to NAFTA be ascribed to the fact that the trilateral accord substantially improved Canada’s position in the U.S. market relative to the original FTA. To be sure, Canadian negotiators did not return from the NAFTA negotiations empty-handed. Among other things, for example, NAFTA clarified the earlier agreement’s “rules of origin” provisions, thus ensuring against a repetition of the Honda dispute that erupted between Canada and the United States in 1992.96 Canada’s negotiating team also trumpeted a provision in the NAFTA side agreements permitting the Canadian government, and only the Canadian government, to refer labor and environmental disputes to Canada’s own domestic courts as a last resort.97 For Liberal constituencies, however, NAFTA was if anything a step backward. As in the United States, there was widespread concern in Canada, particularly on the part of Canadian workers, that Mexico’s low wages would cause manufacturing jobs to head south.98 A more legitimate concern would have been the impact of the agreement’s more restrictive rules of origin (relative to those found in the original FTA) in automobiles, clothing, and footwear. In the case of autos, for example, the tightening of the FTA’s rules of origin removed the incentive for car companies based outside North America to locate “transplants” in Canada.99 Most importantly, though, NAFTA’s entry into force eliminated Canada’s special tariff and nontariff preferences vis-à-vis Mexico in the U.S. market (though Canada retained its privileged position with respect to fourth countries). Whatever its political drawbacks in the eyes of Liberals, the FTA at least had afforded Canadian exporters preferential access to American consumers and suppliers. Under NAFTA, by contrast, Mexican and Canadian exports both enter the United States duty-free. Insofar as Canadian and Mexican industries square off against each other in the U.S. marketplace, NAFTA has thus had the effect of diluting Canada’s competitive advantage. 96 E.g., Lipsey, Schwanen, and Wonnacott 1994, 60. In 1992, the U.S. Customs Service imposed a 2.5 percent duty on Honda Civics manufactured in Ontario on the grounds that they contained insufficient North American content. In Mulroney’s view, the American action against Canada was simply a ploy designed to induce the Japanese to build more cars in the United States. Canada, he complained, was “getting sideswiped by American Japan-bashing.” Quoted in Farnsworth 1992. 97 B. Simon 1993a. 98 B. Simon 1992. 99 Whalley 1993, 357– 59; R. Wonnacott 1991a, 80 – 81.
158
N O R T H
A M E R I C A N
T R A D E
Choosing the Lesser Evil In the end, however, the Liberals concluded that they were better off ratifying NAFTA anyway. And, indeed, they were not wrong about this: if Chrétien had gone ahead and rejected the accord, the United States would have been only too happy to negotiate a separate Mexico-U.S. agreement. That, after all, had been Salinas’s initial request. In effect, the Canadian prime minister found himself presented with a fait accompli. However much his Liberal government may have preferred the original pre-FTA status quo, neither that status quo (SQ) nor even the postFTA status quo (SQ9) could be considered a viable alternative. Given Canada’s inability to prevent the United States from signing a separate free trade pact with Mexico, the only relevant question was whether Canada would be worse off taking part in a trilateral pact or enduring a new, even more unattractive status quo, SQ0. Why would settling for SQ0 have been even more undesirable for Canada? The reason is not that Canada would have lost the preference it enjoyed over Mexico in the U.S. market. Given that Mexico and the United States were prepared, if necessary, to “go it alone,” the elimination of that preference— a legacy of the earlier Canada-U.S. accord—was inevitable. Why? Because a separate Mexico-U.S. bilateral agreement would have extended to Mexican producers the same rights and privileges that Canadian producers had received under the FTA. In fact, such an agreement might well have afforded Mexico better access to the American market than was enjoyed by Canada itself. How could this have been possible, given that duties on Canadian imports to the United States were already being reduced to zero? Ronald Wonnacott (1991b, 58–59) has identified two scenarios. First, if Canada had refused to take part in the NAFTA negotiations, the United States might have agreed to let the special in-bond privileges of Mexico’s maquiladora producers remain intact. Under the maquiladora program, Mexico could import U.S. inputs in bond, process or further manufacture them, and then export them back to the United States without paying duties on anything except the value added in Mexico.100 Second, Canada also had to ensure against Mexico’s winning greater access to America’s lucrative public procurement market, worth an estimated $750 billion if both federal and state purchases are included.101 While the FTA had lowered the government procurement threshold from $171,000 to $25,000, the renegotiated terms still left ample room for discrimination against Canadian bidders by agencies of the U.S. government.102 If the Canadians had opted out of the trade 100
Once Canadian duties reached zero, however, this temporary advantage would disap-
pear. 101
Doern and Tomlin 1991, 84. The FTA-imposed threshold did not apply to purchases by state governments or municipalities, for example, nor to government purchases in such product areas as telecommunications, power-generating equipment, services, and urban mass transit equipment. Large public 102
N A F T A
A N D
B E Y O N D
159
talks between Mexico and the United States, it is quite possible that the Americans would have afforded better treatment to Mexican contractors than that which Canadian contractors had been getting under the FTA. Yet even if the Americans had not afforded Mexico any better treatment, it would still have been in Canada’s interest to agree to NAFTA. By refusing to trilateralize the trade talks, the Canadian government would have given the United States a green light to establish itself as a hub in a new, U.S.-centered hub-and-spokes system. If this had been allowed to happen, economic theory tells us what the result would have been—namely, an immediate decline in the competitiveness of Canadian exports to the United States (and, less importantly, to Mexico), followed in all likelihood by an outflow of Canadian capital to the newly advantaged American hub.103 Why, exactly, would Canada’s “demotion” to a spoke economy have produced such deleterious effects? The answer is that Canada’s U.S.-based competition would have been able to acquire duty-free (and therefore lowerpriced) inputs not just from Canada, as had been the case since 1989, but from Mexico as well. Canadian producers would also have been at a competitive disadvantage because of their restricted access to Mexican consumers. Only U.S. producers would enjoy preferential (i.e., duty-free) access for their exports in both the Canadian and Mexican markets. And if the Liberals had opted out of NAFTA and the FTA, Canada’s export competitiveness—and its attractiveness as a site for long-term productive investment— would have deteriorated even further. Why? Because now, instead of Canada’s sharing a preference with Mexico, Mexico alone would get special treatment. And if the United States had then gone on to sign separate bilateral agreements with other countries, Canada’s competitive position within its American and overseas markets would have eroded even more. Ultimately, of course, the Canadian Liberals did come out in “favor” of NAFTA. But it was not, as proponents of the collective-action paradigm would have it, because they expected to benefit from the new agreement. In fact, Chrétien did not like NAFTA any more than he liked its precursor, the Canada-U.S. agreement signed by the Conservatives in 1988. Indeed, there was a good chance that the new trilateral accord would actually worsen his party’s political fortunes relative to the pre-NAFTA (but post-FTA) status quo. If it was not the prospect of “gains from exchange” that prompted Chrétien to ratify NAFTA, why, then, did he do it? The answer presented here is quite simple: Unable to stop Mexico and the United States from striking a separate free trade agreement, Chrétien became the victim of a kind of subterranean power play. Although his new government was never forced to accept the particular offer with which it was presented, it knew that rejecting that offer would have meant settling for a new, and in all probability even works projects funded by the U.S. federal government could also continue to be closed off to outside bidding for national security reasons. 103 Lipsey 1990; Lipsey, Schwanen, and Wonnacott 1994; R. Wonnacott 1991b.
160
N O R T H
A M E R I C A N
T R A D E
worse, status quo. Faced with a choice between two unattractive alternatives, Chrétien and his Liberal government opted for the lesser evil, jumping on the NAFTA bandwagon so as not to relegate their country to the position of a spoke in a U.S.-dominated hub-and-spokes trading system.
After NAFTA: The Global Response One might think that Mexico’s economic and political travails in the years following NAFTA’s implementation would have put an end, at least temporarily, to any consideration of a more expansive Free Trade Agreement of the Americas (FTAA). Rather than losing momentum, however, the NAFTA bandwagon gained important new ground during these years. Today, as a consequence, the notion of a free trade area spanning the entire hemisphere no longer seems, as it did before NAFTA, “too eccentric even to merit consideration.”104 This concluding section examines the spread of free trade initiatives throughout the Western Hemisphere, and indeed throughout much of the world, following NAFTA’s entry into force in 1994. What accounts for all of this activity? Is it being driven by the anticipation of mutual gain? Or are countries jumping on board the free trade bandwagon out of defensive considerations, because (and only because) they know that their neighbors are prepared, if necessary, to proceed on their own? My conclusion is that in many cases—and in direct opposition to the collective-action-based theories of international cooperation we are used to seeing in the literature— it is the latter motivation that is actually driving much of the action. Latin America’s Newfound “Enthusiasm” for North-South Integration As we have seen, the launching of the Canada-U.S Free Trade Agreement in the mid-1980s put Mexico under enormous pressure to negotiate a bilateral accord of its own. In turn, the prospect of a separate Mexico-U.S. bilateral agreement compelled the Canadian government to trilateralize the arrangement, lest Canadian exporters lose their competitive advantage in the U.S. marketplace. But Mexico and Canada were hardly the only born-again free traders in the region. Just as the earlier Canada-U.S. agreement served to limit Mexico’s choice set, so, too, has NAFTA “shifted the status quo” for the export-oriented nations of Latin America and the Caribbean. Not wishing to be left behind, and concerned in particular about NAFTA’s restrictive rules of origin concerning trade in automobiles and textiles, a number of these countries responded to North America’s free trade initiative by clamoring to be next in line. “All we are asking is to be put on a level playing field with Mexico,” is how Jamaica’s Minister of Industry, Investment, and Commerce put it. “We are not seeking a handout, but only the opportunity not 104
Bouzas and Ros 1994b, 1.
N A F T A
A N D
B E Y O N D
161
to be prevented from taking full advantage of the North American market.”105 This would suggest that defensive considerations were in fact what motivated Jamaica, along with virtually every other country in Latin America and the Caribbean, to embrace President Bush’s Enterprise for the Americas Initiative in 1990.106 Note, however, that this outpouring of Southern support for a new partnership with the North is perfectly consistent with the go-italone logic discussed earlier. Indeed, it would have been surprising if Bush’s proposal had not been embraced by the Southern (i.e., Latin American and Caribbean) countries, many of whose exports competed directly against Mexican exports for market share in the United States. Even so, the EAI’s warm reception, like Mexico’s earlier embrace of NAFTA, came as a shock to many observers, including, once again, a number of the world’s foremost authorities on Latin American politics and economics.107 In 1994, the end-of-the-century deadline for launching the hemispheric free trade zone that Bush had envisaged was postponed until the year 2005. Significantly, however, the impetus for extending the timetable came not from the Latin American or Caribbean side, but from the United States and Canada.108 In the meantime, NAFTA outsiders have been working to offset the costs of exclusion. How? In part by striking separate bilateral deals with Mexico, Canada, and (especially) the United States. At the same time, however, a number of Latin American countries, led by Chile, Brazil, and Colombia, have also been making a concerted effort to secure new markets and sources of supply closer to home. As a result of these twin developments, the Western Hemisphere has come to resemble a “patchwork quilt” of regional—and, increasingly, subregional—trade arrangements.109 Aside from NAFTA, the two most prominent such arrangements are the Common Market of the South (Mercosur), an evolving customs union and free trade area whose members include Argentina, Brazil, Paraguay, and Uruguay; and the Andean Pact, a similar 105
Quoted in Rohter 1997. The Enterprise for the Americas Initiative would have created a hemispheric free trade area by the year 2000. 107 See, e.g., Bouzas and Ros 1994b and Saborio 1992. 108 Given that NAFTA already covered approximately 80 percent of all U.S. exports to the hemisphere (IMF 1992), the economic stakes for the United States in an immediate southward extension of NAFTA were relatively low. Developments since the signing of the Enterprise for the Americas Initiative in 1990—most importantly, the completion of the Uruguay Round in 1994—also played a role in weakening U.S. support for the turn-of-the-century deadline that President Bush had originally proposed. Just as the Canada-U.S. free trade agreement served as a stepping-stone toward NAFTA, the prospect of a North American trade pact led to the reinvigoration of the GATT (see below). This is significant insofar as the “new” GATT secured for the United States a high degree of access to low-wage economies in Latin America—and just about everywhere else—thus reducing the urgency (from the U.S. standpoint) of moving quickly to solidify a hemispheric free trade pact. See Feinberg 1997. 109 Whalley 1992, 126. 106
162
N O R T H
A M E R I C A N
T R A D E
arrangement involving Bolivia, Ecuador, Colombia, Peru, and Venezuela. Chile, though not a full-fledged member of either grouping, has a separate free trade agreement with Mercosur.110 In the spring of 1997, the hemisphere’s trade ministers announced that negotiations for a Free Trade Area of the Americas (FTAA) would begin in earnest at their next summit, in Santiago, one year later. Why the sudden urgency? While there is no single explanation, concerns about losing access to the Northern markets appear once again to have played a pivotal role. In Brazil, for example, the idea of accelerating the process of hemispheric liberalization received strong support from the country’s private sector, which “has decided, however reluctantly, that, though the FTAA would be uncomfortable, that is a price worth paying to get into the United States market.”111 Thus far, as the Latin American political economists Roberto Bouzas and Jaime Ros have noted, “Most concerns regarding the effects of negative discrimination have been related to the U.S. market.” In this respect, Brazil’s experience is typical. On the other hand, if (as seems likely) Chile and other countries actively involved in intra–Latin American trade eventually become incorporated into North-South agreements with the NAFTA “hub,” the exclusion problems confronting Brazil and its neighbors may begin to transcend the U.S. market. This, write Bouzas and Ros, “may turn out to be a powerful defensive motivation for countries that have ‘natural’ trade partnerships with neighbor countries (i.e., the Southern Cone countries).”112 The dynamic to which Bouzas and Ros are referring here is, of course, quite consistent with the extensive-form cooperation game discussed earlier (Figures 3-4 and 3-5). Recall that in Figure 3-5, player C’s decision to join the cooperative regime initiated by A and B (think here of Mexico’s “entry” into the Canada-U.S. Free Trade Agreement) alters the strategic calculus of player D (Chile for example) such that D is now (and only now) “better off” entering the regime as well. Let us assume that D goes ahead and joins. Would this bring the free trade bandwagon to a halt? Not necessarily. As I explained in chapter 3, the fourth player’s decision to participate might well increase the exclusion costs incurred by a fifth player—a Brazil, for example, or an Argentina or Columbia. In this scenario, player E’s cost-benefit analysis would have to incorporate the cumulative effects of negative discrimination not only in the markets of A, B, and C (Canada, the United States, and Mexico), but also in those of D (Chile), its “natural” trading partner. To be sure, this does not mean that a hemispheric free trade zone is inevitable. There is always a chance that the post-NAFTA promise of greater 110
For surveys of these and other “safe-haven” arrangements established during the early 1990s, see Bouzas and Ros 1994a; Frankel 1997; and Hufbauer and Schott 1994. On the internal politics of Mercosur, see Manzetti 1993/94. 111 Pan-American Trade 1997, 42. 112 Bouzas and Ros 1994b, 24, 24 –25.
N A F T A
A N D
B E Y O N D
163
North-South cooperation will ultimately fail to be realized. If a Free Trade Agreement of the Americas does fail to materialize, however, it will not be due to resistance from the South. As painful as a “deepening” of NorthSouth economic integration would be for the nations of Latin America and the Caribbean, the alternative—economic marginalization—would, for most of these countries, be even more painful. In the end, therefore, it is the North that poses the greater impediment to hemispheric free trade. If Canada, Mexico, and the United States ever decide they would benefit from a Southward extension of NAFTA, rest assured their Southern neighbors will do whatever it takes to qualify for membership. There is, in fact, a general pattern here. Some would argue that we are moving rapidly toward a three-bloc world, with preferential trading arrangements centered around the United States (the American bloc), Germany (the European bloc), and Japan (the Asian bloc). What is striking about this transformation is not simply that new regional trade blocs are expanding to the point that they now cover much of the globe; it is that, in contrast with the empires of ancient times, membership in these new “unipolar formations” has been strictly voluntary.113 So why have the United States, Germany, and Japan—the three nations that stand to benefit most directly from today’s empire-building frenzy—not made greater use of their coercive powers? The answer is that, so far at least, these powers have not been needed. Why should the United States, Germany, or Japan apply brute force against their regional trading partners when, for the reasons outlined above, their partners have been willing to open their markets voluntarily? Initially, of course, some of their partners were more “willing” than others. Over time, however—and, again, for reasons that flow directly out of the theoretical analysis elaborated in the first part of this book—the pressures on outsiders have intensified. And with more and more countries rushing to join the nearest bloc (so as to avoid losing access to the markets of other countries in their region), the empire-building process has started to take on a sort of organic momentum of its own. It thus comes as little surprise that nearly one-third of all regional integration agreements notified to the GATT between 1948 and the end of 1994 were signed after 1990.114 Bandwagoning in Europe: The Ever-Expanding EU It is important to emphasize that this free trade contagion has not been confined to the Western Hemisphere. In addition to the spread of subregional agreements throughout the Americas (e.g., Mercosur), the 1990s also witnessed the accessions of Austria, Finland, and Sweden to the European Union. Why did this expansion happen? What, exactly, did the EU have to offer its three new entrants? 113 114
Kupchan 1998. Frankel 1997, 4.
164
N O R T H
A M E R I C A N
T R A D E
The conventional answer is a “for members only” basket of benefits, both material and nonmaterial. Note, however, that this depicts EU expansion as a positive-sum game, the “widening” of the organization boosting the utilities of the new entrants no less (and perhaps even more) than those of the EU’s already-affiliated member states. But are matters really so simple? After all, the citizens who populate the newly admitted countries bring with them distinct—and some would say incongruous—values and traditions. Not only do most Austrians, Swedes, and Finns disagree with their European neighbors on matters of security policy, but for both historical and cultural reasons, they also maintain a deeper commitment to social equality.115 As for the strictly economic “privileges” of membership—the material benefits that will accrue to Austria, Sweden, and Finland now that they have joined the EU’s internal market, legal system, and monetary union—these, too, are open to debate.116 Having decided to come aboard, for example, Austria, Sweden, and Finland are now required to support the poorer regions of the EU through what strike many observers as disproportionately large financial contributions. Why then, if not in the anticipation of individual or collective gains, did the governments of Austria, Sweden, and Finland decide to break ranks with Norway, Switzerland, and other members of the ever-shrinking European Free Trade Area (EFTA)?117 In keeping with the “free trade bandwagoning” idea presented above, let me suggest a simple explanation: Just as the countries of Latin America and the Caribbean felt they could not afford to be excluded from NAFTA, so, too, did the Austrians, Swedes, and Finns feel they could not afford to be excluded from the EU. In this view, EFTA’s incorporation into the EU represents yet another case in which defensive motivations outweighed the offensive considerations that take center stage in collectiveaction-based accounts. More than the prospect of mutual gain, what prompted the Austrian, Swedish, and Finnish defections was the pending “completion” of the EU’s internal market—a market whose very existence would raise the relative costs of doing business within the union for non-EU firms, including those located in countries that were at that time members of EFTA. Though he does not draw out the “absolute” utility implications of EU membership as I do here, Richard E. Baldwin (1995) has made a persuasive case that Austria’s application to the EU in 1989 was indeed prompted (provoked?) by the EU’s Single Market initiative.118 In turn, Baldwin argues, the prospect of Austrian accession amplified the already high costs of exclusion 115 Antola 1991; Luif 1991; Michalski and Wallace 1993; Stålvant and Hamilton 1991. This is not to say that social equality is altogether irrelevant to citizens of the EU. On the vicissitudes of the EU’s “Social Chapter”—a list of proposals that would commit each member state to similar workplace standards and employment practices, including an EU-wide minimum wage— see Lange 1993. 116 See, e.g., Alho 1995 and Anderson and Tyers 1995. 117 Of EFTA’s remaining member states, Norway is the only one not currently seeking admission into the EU. See below. 118 See also Krugman 1988 and Wijkman 1994.
N A F T A
A N D
B E Y O N D
165
for EFTA’s remaining members, thus tilting the balance in favor of EU membership for countries like Finland and Sweden, whose exporters were less heavily dependent on the EU market. Although Austria, Finland, Sweden, Norway, and Switzerland all put in applications, Baldwin likens the process to a “domino effect,” with Austria as the first mover. To put the point in terms of our earlier cooperation game (Figure 3-5), one could say that Austria played the part of actor C: Austria’s decision in favor of entry was decisive for actor D (Finland), whose entry shifted the status quo for actor E (Sweden), whose leaders would not otherwise have joined, and so on. Because “each EFTA nation individually faced the prospect of losing out in the EU-12 markets and in the markets of those EFTAns acceding to the EU,”119 Europe’s free trade bandwagon continued until virtually every EFTA country had decided to get onboard—even Norway, a country whose exports (predominantly oil) were threatened by the EU’s Single Market initiative only after it appeared that the internal market would encompass Sweden and Finland.120 A similar dynamic is now playing itself out in Central and Eastern Europe. In preparation for entry into the EU sometime after the turn of the century (the earliest possible entry date is 2002), the governments of Hungary, Romania, and Slovakia recently agreed to settle a series of long-standing disputes over minority issues, while in 1995 Slovenia committed itself to joining the Czech Republic, Hungary, Poland, and Slovakia in a Central European Free Trade Area (CEFTA). Were governing parties in these formerly Communist countries happy about all of this? Might their political “utilities” have been higher before the EU’s (unilateral) action than after? These, too, are questions ripe for further theoretical and empirical study. Defensive Motivations for APEC Although interventionist government policies were certainly not the only factor, it would be hard to explain the stellar economic performance enjoyed by many Asian countries over the past two decades without acknowledging the role played by targeted state subsidies, controlled credit, and the like, in stimulating export-led growth.121 At their 1993 summit meeting in Seattle, however, the leaders of these same countries resolved to dramatically scale back the degree of government involvement in their economies. The centerpiece of what came to be called the Asia-Pacific Economic Cooperation (APEC) forum, this agreement commits all APEC members—whose ranks include (among others) Japan, China, Taiwan, and Malaysia in the East; Canada, Mexico, Chile, and the United States in the West—to eliminating virtually all barriers to trade with one another by no later than the year 2020. In seeking to explain this unprecedented achievement, a number of ana119
Baldwin 1995, 34, emphasis in the original. The Norwegian government applied for admission to the EU in 1992. Two years later, however, the country’s voters rejected EU membership in a nationwide referendum. 121 See, e.g., Haggard 1990 and A. Young 1994. 120
166
N O R T H
A M E R I C A N
T R A D E
lysts point to Asian concerns about the potential diversionary effects of NAFTA and, to a lesser extent, the European Union.122 More than anything else, say these analysts, it was the specter of a future NAFTA-induced decline in their terms of trade that attracted APEC’s Asian “enthusiasts” to the new scheme. In this sense, the politics of the APEC initiative can be said to offer yet another case of go-it-alone power at work: by “shifting the status quo,” the formation of a comprehensive regional trading arrangement by governing elites in one set of Pacific Rim nations (namely Canada and the United States) prompted governing elites in other Pacific Rim nations (Mexico, Japan, China, Malaysia, etc.) to pursue a course of action that they would not otherwise have taken. Although membership in APEC was strictly voluntary—none of its signatories was actively coerced into joining—some APEC participants might well have preferred the noncooperative (but no longer available) protectionist status quo.123 From Regionalism to Globalism: The Reinvigoration of the GATT Might the passage of NAFTA and, in turn, APEC have played a role in reviving multilateral trade talks in 1993, thus paving the way for the completion of the Uruguay Round the following year? Suggestive evidence to this effect comes from Lorenz Schmerus, the European trade negotiator who, when asked in 1994 why the European Union had decided to abandon its long-standing, go-slow approach toward GATT negotiations, responded: “The chief determination of the successful conclusion of the Uruguay Round was the APEC summit in Seattle; they sent us a clear message. You [North America] had an alternative, and we did not.”124 The notion that Western Europe’s concerns about being shut out of regional trade groupings in other parts of the world are what prompted its turnaround argues for viewing these groupings, in the words of Robert Lawrence (1991, 27), as “the building blocks of an integrated world economy [rather] than stumbling blocks which prevent its emergence.” For Lawrence, it is only a matter of time before the process of regional fragmentation gives way to “an expansionary dynamic.”125 Was Schmerus correct in thinking that the European Union would be better off pushing for multilateral free trade than suffering a deterioration in its preferences within the U.S. and Asian markets? Almost certainly yes. Will 122
Crone 1993, 519 –22. See also Krueger 1993 and Saxonhouse 1993. Exactly how this regional dynamic may be altering the economic policies pursued by particular Asian governments—and how these policy changes may in turn be affecting the region’s prospects for democracy—is an interesting question for future research. Whatever the answer one gives, however, the existence of such a dynamic would seem beyond dispute. For a sampling of critical reaction to APEC’s institutional forebear, the Pacific Economic Cooperation Conference (PECC), see Crone 1993, 516 –17. 124 Quoted in Funabashi 1995, 107; see also Wolf 1996, 135 and 138. 125 Lawrence 1991, 27; cf. de Melo and Panagariya 1993. 123
N A F T A
A N D
B E Y O N D
167
EU welfare be higher under the newly installed GATT regime than it was under the “old” (i.e., pre-NAFTA, pre-APEC) status quo? Here, the issue is much less clear, and more extensive empirical work would be needed before one could confidently make such a judgment. There is, at any rate, no reason to assume that because the new WTO was embraced by governing parties throughout Europe, its creation was therefore conducive to their distinctive political and economic interests. Nor, by the same token, is there any logical or necessary reason for assuming that the demise of the pre-1994 GATT—with its exclusions for poorer countries in the area of services, intellectual property, trade-related investment measures (TRIMs), and technical standards—was in the interests of the developing world.126 Once the advanced countries of Europe and North America had resolved to reform the GATT, however, there were strong incentives for developing nations to take part in the new regime. After all, the countries of the developing world would only further undermine their positions vis-à-vis the United States and other advanced member countries were they not to take part. 126
See, e.g., Kahler 1995, 35 – 42.
P A R T
I V
European Money
C H A P T E R
8
Winners and Losers in the European Monetary System
This part of the book explores the European Monetary System (EMS) of the 1980s and 1990s. One of the boldest and most comprehensive efforts at international monetary cooperation ever undertaken, the EMS regime did not, I will argue, benefit all of its member states. To the contrary, authorities in Italy and the United Kingdom, two of the regime’s largest and most prominent signatories, preferred the pre-EMS world of freely floating exchange rates. So what, then, prompted the Italians (in 1979) and the British (in 1990) to relinquish their monetary autonomy? The answer, in a word, is power— specifically, the go-it-alone power exercised by the EMS regime’s French and German beneficiaries. To be sure, this rendering of the EMS story is not uncontroversial, for while the distributional conflicts engendered by the regime have been widely noted and discussed, the consensus of scholarly opinion— at least among political scientists—is that the EMS represented a Pareto-improvement over the noncooperative monetary status quo it supplanted.1 Focusing on the entry decisions of Italy and Britain, this chapter presents a different view. In chapter 9, I then turn from the issue of membership in the new arrangement to the question of institutional choice: Why did the EMS regime take the particular (supranational) form it did?
Rethinking European Monetary Integration Formally inaugurated in March 1979, the EMS has been the longest-lasting, most encompassing exchange rate arrangement among industrial countries since the collapse of the Bretton Woods regime in the early 1970s. While exchange rate agreements were not new to Western Europe, previous efforts to create a common exchange rate policy had enjoyed little success.2
1 See, for example, the recent study by Oatley (1997), who is most explicit on this point. The EMS, he writes, was “a Pareto-improving bargain; weak-currency policymakers gave the Bundesbank control of the system’s single monetary policy instrument in exchange for the right to use the exchange rate to draw on the Bundesbank’s commitment to price stability and facilitate disinflation at home.” Oatley 1997, 49. A similar view is implicit in Moravcsik’s description of the European Community, of which the EMS was an integral part, as “the most successful international institution of modern times.” Moravcsik 1998, 501. 2 See esp. D. R. Cameron 1990 and 1992a.
172
E U R O P E A N
M O N E Y
Some Background While many observers interpreted the failure of these earlier initiatives as a sign that progress toward monetary integration in Europe would have to await the creation of a European common market, others disagreed, maintaining that the introduction of a single European currency would promote—because its day-to-day operation would necessitate—the further dismantling of trade restrictions and, ultimately, the establishment of an integrated European market. Foremost among these dissenters was the president of the European Commission, Roy Jenkins. During a visit to Italy in October of 1977, Jenkins delivered a speech exhorting European heads of state to renew their earlier commitment to monetary integration precisely as a way of accelerating the process of trade integration. Six months later, these European leaders took his advice, embracing a bold (albeit initially rather vague) proposal for a new, exclusively European “zone of monetary stability.” The next eleven months, from April until December of 1978, were spent filling in the details.3 While the monetary arrangement that emerged from this process incorporated a broader set of agreements, its centerpiece was the so-called exchange rate mechanism, or ERM. When a country joined the ERM, it was assigned a central exchange rate, or “parity,” denominated in a common unit of account known as a European Currency Unit (ECU). The ERM signatory then promised to adjust its interest rates, buy and sell reserves in the currency markets, or take whatever other measures might be necessary to keep the value of its currency within a given range—either 2!f or 6 percent—of this central rate. Eight countries—France, Germany, Italy, Denmark, Ireland, the Netherlands, Belgium, and Luxembourg—agreed to these rules from the outset. These eight were subsequently joined by Spain (1989), the United Kingdom (1990), and Portugal (1992). In 1992, less than two years after the United Kingdom had become a full member of the system, a tidal wave of speculation forced the British government to withdraw the pound from the ERM.4 The European currency crises of September 1992 and August 1993 also saw the departure of the Italian lira, devaluations of the Portuguese escudo, Spanish peseta, and Irish punt, and a dramatic widening of all ERM bands (except the one between the German D-mark and the Dutch guilder) to plus or minus 15 percent. In 3
The EMS proposal was first announced on April 7, 1978, at a European Council summit meeting in Copenhagen. A more detailed version of this proposal was approved at a second European Council summit held three months later in the German city of Bremen. Negotiations came to a close with the ratification of a formal resolution shortly after the conclusion of yet another summit, this one in Brussels, on December 15, 1978. 4 Technically speaking, the United Kingdom had been a member of the EMS since its inception in 1979. Prior to 1990, however, that membership was essentially meaningless, as the British government had steadfastly refused to introduce the pound into the regime’s all-important exchange rate mechanism.
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
173
addition, a number of European countries—both inside and outside the system—began reimposing capital controls. At the time, the European Monetary System seemed to be in disarray. In retrospect, however, the crisis was but a temporary setback, as members of the European Union have now committed themselves to creating a fullfledged monetary union (EMU) by the year 2002. By that time, and in keeping with the timetable initially set forth in the 1991 Treaty on European Union, most EU signatories will have completely abandoned their separate currencies in favor of the “euro,” the new appellation for the EMS regime’s ECU.5 There is of course no guarantee that the euro will endure. Even if things do go awry, however, the fact remains that the EMS, whether or not it ever gives way to a (sustainable) monetary union, has already altered the course of events in Western Europe, profoundly influencing the economic and political trajectories of countries throughout the region. The next section reviews the conventional wisdom on why this achievement—the setbacks of 1992 and 1993 notwithstanding—worked to the benefit of the regime’s two largest members, France and West Germany, without whose initiative and support the EMS would never have become a reality.6 The Enacting Coalition: France and West Germany Why did the ambitious monetary agenda that Jenkins laid out in his 1977 address resonate so strongly with West Germany’s chancellor at that time, Helmut Schmidt, and his French counterpart, President Valéry Giscard d’Estaing? A variety of different answers have been suggested. Preventing World War III. Perhaps it was because Schmidt and Giscard believed an EMS would prevent the kind of erratic exchange rate movements that preceded, and may well have precipitated, the outbreak of World War II.7 Unifying Europe Politically. Relatedly, the German and French leaders may have viewed Jenkins’s proposal as a vehicle for achieving political unification. 5
I will discuss the larger theoretical significance of Europe’s transition from the EMS to EMU at the end of chapter 9. 6 That the French and German governments were the two prime movers behind the EMS proposals of the late 1970s is made abundantly clear in Ludlow’s (1982) detailed account of the negotiations. On the politics surrounding the creation and early years of the regime, Ludlow’s book remains the definitive work. Also see De Cecco 1989; Frieden 1994; Goodman 1992; Guerrieri and Padoan 1989; Heisenberg 1999; Kruse 1980; McDonald and Zis 1988; McNamara 1998; Oatley 1997; Spaventa 1980; Woolley 1992. 7 It is debatable whether the volatility of European currency markets during the 1920s did in fact exacerbate Germany’s hyperinflation, thus indirectly precipitating the rise of German fascism. Be that as it may, the view that such a causal link existed was widespread in French and German political circles at the time (Giavazzi and Giovannini 1989, 5; Eichengreen and Wyplosz 1993, 117). For statistical evidence that the EMS reduced exchange rate uncertainty over the course of the 1980s, see Artis and Taylor 1988; Fratianni and von Hagen 1992, chap. 7; Giavazzi and Giovannini 1989, chaps. 2 and 5; McDonald and Zis 1989; Weber 1991.
174
E U R O P E A N
M O N E Y
In his memoirs, Schmidt recalls: “I had always regarded the EMS not only as a mere instrument to harmonize the economic policies of the EC member countries, but also as part of a broader strategy for the political self-determination of Europe” (quoted in Fratianni and von Hagen 1992, 17–18). Stimulating European Trade and Investment. Some analysts suggest that Schmidt and Giscard’s interest in mitigating the uncertainties created by volatile exchange rate movements stemmed from their belief that these fluctuations were impeding intra-European trade and investment.8 Because France and West Germany both had large domestic markets, their stake in maintaining a conducive climate for the flow of goods, services, and capital throughout Europe was not as great (at least not on strictly economic criteria) as it was for smaller, more heavily trade-dependent EC members such as Belgium and the Netherlands. Be that as it may, in 1978 France and West Germany sent nearly one-tenth of their total output to their EC partners— 7.8 percent and 9.5 percent, respectively.9 Safeguarding the CAP. Finally, some observers have speculated that the leaders of Germany and France were concerned about the possibility of future disruptions in the operation of the EC’s agricultural price-support program. Over the course of the 1970s, European leaders instituted a number of changes to their Common Agricultural Policy, or CAP, which had proved exceedingly difficult to operate under the de facto floating exchange rate system that emerged in the wake of Bretton Woods. The most important of these reforms was the establishment of new agricultural pricing rules—known as monetary compensatory amounts—whose introduction was designed to mitigate the rise in food prices, and resulting inflationary pressures, experienced by devaluing countries. In practice, however, the new compensatory arrangement had the perverse (not to mention politically awkward) effect of benefiting German agriculture at the expense of the French farmers whose interests the postwar founders of the CAP had had in mind when they initially conceived the system.10 Avoiding a recurrence of the pre-World War II rise in exchange rate volatility, paving the way toward a European superstate, creating a more conducive environment for European trade and investment, providing a new lease on life for the CAP—all of these considerations may have played some role in Schmidt and Giscard’s thinking. Although these motivations are widely cited in the literature, however, their importance should not be overstated. To begin with, there is little evidence that the erratic exchange rate movements 8
Giavazzi and Giovannini 1989, 2– 5. IMF 1984a and 1984b. 10 For a detailed discussion of the new agricultural pricing system and its defects, see Giavazzi and Giovannini 1989, 12–19; Loriaux 1991, 255 – 57; and McNamara 1998, 98 –104. 9
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
175
of the 1970s were having much, if any, retarding effect on intra-European flows of trade and investment.11 And while there is no question that the increase in post–Bretton Woods exchange rate volatility undermined the operation of the Common Agricultural Policy, the EMS itself did little to solve the problem. Indeed, the combination of Britain’s initial exclusion from the exchange rate mechanism and Italy’s wider fluctuation band may have made the calculation of monetary compensatory accounts for the CAP even more cumbersome than it had been before the system entered into effect.12 Finally, in neither France nor Germany were governing officials in a position to dwell on lofty concerns about Europe’s collective future. It is true that Giscard and Schmidt both claimed to support the idea of political union— the creation of a European superstate—as a long-term goal of foreign policy. For neither leader, however, did the issue promise a large immediate payoff in terms of votes won or lost. And yet it was precisely such a payoff that Giscard and Schmidt were hoping to obtain when they began formulating their monetary proposal in 1977. The underlying reasons for this will be discussed below. Suffice it to say that when the EMS was initially conceived, considerations about “European unity” may not have been as politically salient to decision makers in France and Germany as some accounts (including those offered years later by the protagonists themselves) would suggest.13 THE VIEW FROM WEST GERMANY
For Chancellor Schmidt, the most attractive feature of Jenkins’s 1977 proposal was its promise of a new, more durable link between the deutsche mark and the French franc.14 Global demand for the D-mark had been increasing ever since the abandonment of the gold standard in 1971 (Figure 8-1). With the United States liberated from the burden of pursuing fiscally responsible policies—a burden it had been shirking, in any case, since the late 1960s— currency speculators increasingly looked to the D-mark as a bulwark against the rising inflationary tide. The mark’s chief attraction was that the Bundesbank, not the government, controlled German monetary policy, and Bundesbank officials had a reputation for standing up to German governments they considered too soft on inflation.15 11 Eichengreen 1992, 7– 8. See also the discussions in De Grauwe 1992 and Frankel and Wei 1995. 12 Loriaux 1991, 257; van Ypersele and Koeune 1985, 65. 13 Moravcsik 1998, chap. 4, reaches a similar conclusion. 14 Between 1966 and 1972, Germany was governed by a “grand coalition” comprising all three of the country’s major political parties: the Social Democrats on the left of the spectrum, the Free Democrats in the center, and the Christian Democrats (a coalition of the Christian Democratic Union and the Bavaria-centered Christian Social Union) on the right. In the fall of 1972, however, the Social Democrats, then under the leadership of Willy Brandt, won an electoral landslide. The end result was the formation of a left-of-center coalition encompassing the Free Democrats but excluding the Christian Democrats. Although Schmidt succeeded Brandt as chancellor in 1974, that government, with Schmidt as its leader, remained in power until 1982. 15 Goodman 1992; Marsh 1992.
176
E U R O P E A N
M O N E Y
Figure 8-1. Monthly Exchange Rates in Germany, 1970 –1980
As the 1970s wore on and the financial markets’ thirst for mark-denominated securities showed few signs of abating, the resulting appreciation of the D-mark put tremendous strain on Germany’s export-driven economy. Even more problematic than this appreciation’s effects on the price competitiveness of German exports in the United States was its damage to Germany’s economic position within Europe.16 Although the currencies of France, Italy, and Britain all tended to track the (depreciating) dollar instead of the (appreciating) deutsche mark, German authorities were most preoccupied with the exchange rate between the D-mark and the French franc. In 1977, France alone consumed 12.3 percent of all German exports, making France Germany’s number one trading partner.17 Given that the franc’s depreciation was posing problems for Germany’s economy, it may seem surprising that the German government did not simply enter the foreign exchange market directly. In theory, the government could have used its foreign exchange reserves to correct the situation unilaterally, that is, by purchasing French francs in exchange for D-marks. In practice, however, it would have been pointless for Schmidt’s governing coalition to try to drive down the value of the D-mark in this fashion. Had it done so, the Bundesbank would have responded by engineering an offsetting con16 De Cecco (1989, 89) reports that German manufacturers of steel, chemicals, machine tools, automobiles, and textiles were all at risk of losing market share to European competitors. See also Gros and Thygesen 1992. 17 IMF 1984a. Germany’s next largest export market was the Netherlands (10.1 percent), followed in descending order by Belgium (7.8 percent), Italy (6.8 percent), the United States (6.1 percent), and the United Kingdom (5.9 percent).
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
177
traction of Germany’s money supply. By the late 1970s, the Bundesbank was already threatening to raise domestic interest rates, a threat aimed at keeping the social democratic-liberal coalition in power at that time from undertaking its preferred high-growth, export-oriented program.18 Recent history dictated that this threat be taken seriously. The last time the government had passed a budget calling for a large fiscal stimulus—immediately after the OPEC price hike of 1973—Germany’s monetary authorities had promptly clamped down on the money supply, precipitating a recession in which over a million German workers lost their jobs. It was, in short, the prospect of strengthening the franc (and thereby countering the Bundesbank-induced appreciation of the D-mark) that most strongly attracted the German government to the monetary proposal that EC commissioner Roy Jenkins put forward in 1977. “It is no accident,” noted an observer at the time, “that the rediscovery of the advantages of fixed exchange rates . . . is occurring just at the time when the end of the era of export-led growth is beginning to be perceived as a reality by German policymakers. . . . It is not a problem of economic stability (because this is not threatened but strengthened by the upward trend of deutsche mark); nor is it a problem of the European or of the world economy. It is, rather, a problem of saving what one can of the traditional foundations of Germany’s export-led economic growth.”19 THE VIEW FROM FRANCE
The prospect of closer monetary collaboration was also appealing—albeit for different reasons—to governing officials in France. In linking the franc to the D-mark, French politicians were primarily hoping to lend credibility to France’s most recent economic stabilization initiative, the so-called Barre Plan. From the moment he took office in 1974, stabilizing prices had in fact been President Valéry Giscard d’Estaing’s central preoccupation. This is not surprising, for the core constituencies of conservatives like Giscard (managers of small firms and others members of the French middle class) had felt the sting of inflation more than the constituencies of the left (blue-collar workers, public sector employees, etc.), whose wages generally rose along with prices. The conservative president’s concerns about double-digit inflation were also shared by the National Assembly, where delegates from Giscard’s 18
In the aftermath of the first oil crisis, the United States began looking to West Germany and Japan to act as “locomotives of growth,” pulling the rest of the world to recovery. As important as these international pressures were, however, much of Schmidt’s interest in a dramatic fiscal expansion stemmed from purely domestic considerations. In the German federal elections of 1976, the conservative Christian Democratic party emerged as the single largest political grouping within Germany’s three-party system. Hoping to prevent further losses in the Social Democrats’ working-class base, Schmidt turned from the path of austerity in 1977, reducing taxes and implementing a deficit-financed program of public investment. See Scharpf 1987, 143–46. 19 Hankel 1980, 30. For similar views, see Heisenberg 1999, 51– 53; Moravcsik 1998, 245 – 48; Triffin 1979, 69.
178
E U R O P E A N
M O N E Y
own party, the Union pour la Démocratie Française (UDF), and the neoGaullist Rassemblement pour la République (RPR) together controlled a majority of seats. During Giscard’s first two years as president, his administration succeeded in returning inflation to single digits. Still not satisfied, Giscard appointed Raymond Barre, a technocrat with unblemished inflation-fighting credentials, to take over the duties of prime minister. Immediately following Barre’s inauguration in 1976, the administration announced that it would begin implementing a new program of fiscal and monetary contraction—the aforementioned Barre Plan—and that French citizens could expect it to be even more stringent than the plans that had preceded it.20 In keeping with the “disciplinary” theory of the EMS, Giscard and his advisers hoped that an exchange rate agreement between France and Germany would speed the continuing downward adjustment of inflationary expectations.21 If Giscard were to forswear the use of currency depreciation as a tool for cushioning French workers from the effects of counterinflationary policies, these workers would be more inclined to moderate their wage demands—or so Barre and others assumed—for fear of losing their jobs. And this would make France’s transition to the low-inflation equilibrium desired by the UDF and RPR’s middle-class supporters less wrenching than might otherwise have been the case.22
The “Attractions” of Monetary Integration for Italy and the United Kingdom By 1977, it had become apparent that governing elites in France and Germany had a mutual interest in coordinating the franc-mark exchange rate. Put simply, the French and German politicians who launched the new monetary regime did so because they thought it would make them better off. But what about the Italian and British governments? Did their leaders also expect to benefit? This section examines the evidence. First, however, I want to briefly revisit the issue of why assessing the utility implications of membership in multi20 A good source on French economic policy during the 1970s is Lauber 1993. See also Oudiz and Sterdyniak 1985. 21 The disciplinary theory of the EMS is discussed in Fratianni and von Hagen 1992, 48– 53. Classic statements include Giavazzi and Pagano 1988; Giavazzi and Giovannini 1988 and 1989. 22 It might be argued that money supply targets, such as those introduced a couple of years later in the United Kingdom, could serve equally well as an indication of a government’s commitment to lowering inflation. From a credibility standpoint, however, the problem with using money supply targets is that central bankers can use their private information about economic developments to justify unnecessary monetary expansions. In countries (such as France) where central banking institutions are highly permeable to political influence, exchange rate targeting is assumed to provide greater credibility; see, e.g., Canzoneri 1985.
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
179
lateral instititutions in so difficult empirically. At first glance, one might think that the question of whether Italy and the UK were EMS “losers” could be resolved by simply evaluating how each of their governments fared politically. If that were the case, the empirical task would be a relatively easy one. As I discuss later on, the political fortunes of both countries’ governments did in fact decline after their EMS entries—in Italy’s case quite precipitously. Does that mean that governing elites in Italy and the UK would have preferred a world in which the EMS did not exist? Not necessarily, for even if the Italian government did fare worse after joining the EMS than before, no one would claim that the EMS deserves all of the blame for this pre- versus post-entry differential. Nor, by the same token, was the EMS the only salient issue on the British political agenda when the UK joined in 1990. How, then, to proceed? Given the abundance of cross-national and time-series data on various indicators of economic performance, there is a strong temptation to orient the empirical analysis around economic issues—as if one could answer the “utility” question by simply comparing Italy’s or the UK’s economic performance before the EMS came into existence with its performance after the regime was inaugurated. Here, too, this temptation should be resisted. One reason has to do with the multiple causality problem just noted. Just as it would be a mistake to assign the EMS all of the blame for the political difficulties experienced by governing elites in Italy or the United Kingdom, it would also be a mistake to blame it for Italy’s or the UK’s economic misfortunes. There is, however, a second, even more fundamental reason for rejecting the statistical approach: the methodologically appropriate baseline—the “before” in the before-and-after comparison—is not how the Italian or British governments were doing before the EMS was launched; it is how they would have done if the pre-EMS status quo had continued after 1978. The counterfactual problem looms large. Strictly speaking, of course, what we want to know is not whether things would have been better for the Italians and British if (counterfactually) the EMS regime had never come into existence. Rather, we are concerned with motivation: Did members of Italy’s or the UK’s political establishment believe that things would have been better without the EMS? Now, one might suppose that because we are more interested in what these countries’ political leaders were thinking at the time they decided to join the regime than in what actually came to pass, one could simply go out and interview the protagonists. But, once again, matters are not more complicated. There are wellknown reasons why one cannot always infer the truth from politicians’ retrospective renderings of their motivations and thought processes. In this case, moreover, the policy decisions at issue were taken many years ago. Although many of the protagonists no longer hold official positions and so should be less inclined to dissemble for political purposes, the passage of time can also make it more difficult for the relevant parties to recollect why, years earlier, they made the choices they did. Then, too, there is the problem that
180
E U R O P E A N
M O N E Y
a number of the principals—including the Italian prime minister who took the lira into the EMS in 1979—have since been found guilty on charges of political corruption, thus casting considerable doubt on their trustworthiness, to put it mildly. For all of these reasons, it is important to use caution when gathering information from (or assessing first-person accounts written by) the individuals who were directly involved. In the end, however, I would submit that my description of Italy and the United Kingdom as EMS “losers” holds up despite these methodological and practical hurdles. Demonstrating this requires showing that the elected leaders responsible for introducing the Italian lira and British pound into the new regime would actually have preferred the initial, pre-EMS status quo—that, in other words, these leaders would have breathed an enormous sigh of relief if France and Germany’s EMS initiative had gone nowhere. Let us begin with the case of Italy. Why Did Italian Authorities Give Up Control of the Lira? It was not until April 7, 1978, several weeks after Giscard and Schmidt had begun formulating the specifics of their monetary proposal, that Italian Prime Minister Giulio Andreotti learned of its existence. The firsthand accounts cited by Ludlow (1982, 91–92) suggest that when Giscard and Schmidt finally broached the issue—the occasion was an informal after-dinner gathering of European heads of state at a European Council summit meeting in Copenhagen—the Italian prime minister was caught completely off guard. In time, however, Andreotti became a strong proponent for introducing the lira into the regime, a move he hoped to undertake (the vagaries of Italian domestic politics permitting) at the same time that France and Germany introduced their own currencies. Contemporary accounts suggest that deciding to give up control of the lira was not at all difficult for the Italian leader. Why was Andreotti’s decision such an easy one? HIGH POLITICS
A proponent of realist theory might call attention here to the larger geostrategic context of Italian politics during the last years of the Cold War: At the end of the day, it could be argued, Italy’s economic situation took a backseat to its “prior” security interest in keeping Germany firmly embedded in the Western Alliance. While there may be something to this explanation, finding hard evidence to support it is not easy. Not only are there few reports of Italy’s political leaders having invoked geopolitical considerations as justification for their actions, but it is unclear, in any case, why a still-divided Germany would have posed a major security threat to the Italians during the late 1970s. After all, Europe’s monetary agenda was developed during the height of the Cold War, a period when, as Jeffry Frieden (1998, 35) has pointed out, “German fealty to the alliance was not in question.” Even if fears of German expansionism
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
181
had been the motivation for introducing the Italian lira into the ERM, one wonders why British authorities saw no need to take similar countervailing measures. Insofar as West Germany posed a security hazard to Italy—or to any country in Europe, for that matter—the safety and economic well-being of citizens in the United Kingdom would have been threatened as well. And yet successive British governments resolutely refused to give up control of the pound until 1990. CREDIBILITY
Another possibility is that Andreotti thought Italy’s entry into an exchange rate arrangement centered around the German mark would lend credibility to his government’s recently enacted program of fiscal and monetary retrenchment, thereby accelerating the downward adjustment of wages.23 Initiated at the beginning of 1978, the Pandolfi Plan (named after the minister of the Italian Treasury) combined monetary contraction and fiscal consolidation with calls for greater wage restraint on the part of Italian workers. The “disciplinary” rationale for Andreotti’s decision will be discussed at greater length below. For now, suffice it to say that even if Andreotti himself had shared French President Giscard d’Estaing’s view of the EMS as a device for accelerating the downward adjustment of wages, Italy’s decision to join the EMS occurred during a period in which the Italian Communist party (PCI), although denied ministerial portfolios in the government itself, was nonetheless a crucial component of Andreotti’s governing coalition. Why does this matter? It matters because there is no evidence that the leaders of the PCI believed Italy’s participation in an exchange rate regime anchored on the D-mark would work to the advantage of their party’s predominantly working-class supporters. To the contrary, the view within Italian Communist circles at the time was that “the government would ‘use’ the EMS as an excuse to push through domestic policies that would be deflationary and therefore damaging to the interests that they represented.”24 Ludlow’s description of a speech given by Luciano Barca, the PCI’s main spokesperson during this period, conveys the sense of frustration felt by many politicians on the left of the Italian political spectrum: His tone was sharp from the beginning. He understood the German chancellor’s anxiety to create the system, and [that] it was only proper . . . that Mr. Schmidt should seek to defend his country’s interests, but it was important to realize that Germany’s interests were not identical with Italy’s and that entry into the EMS on the terms that the Germans appeared to be insisting on would involve placing the country under Germany hegemony. It would also reinforce the position of those who were bent on domestic policies that would increase unemployment and lower living standards.25 23
See, e.g., Giavazzi and Pagano 1988. Ludlow 1982, 212. 25 Ludlow 1982, 214. 24
182
E U R O P E A N
M O N E Y
Despite its belief that Italy’s entry into the EMS would lead to higher unemployment and a lower standard of living for its constituencies, the Italian Communist party ended up going along with Andreotti’s decision. My contention below is that the PCI did so because it feared a “no” vote would lead to an (even) worse outcome. By the spring of 1978 it was clear that the French government was going to peg the franc to the D-mark irrespective of Italy’s decision. As a result, the status quo that the Communists (and even perhaps Andreotti’s own Christian Democratic party) would have preferred—a de facto laissez-faire system in which Italy, France, Germany, and the United Kingdom set their monetary policies more or less independently— had ceased to be a viable option. Had it not been for this dramatic change in Italy’s external environment, Italian domestic politics would have strongly militated against placing the lira in the new monetary regime. If not, in other words, for this shift in the status quo, the notion that preserving the value of the lira was about to become the touchstone of Italian policy would have struck most observers as too farfetched to warrant serious consideration. THE COSTS OF EXCLUSION
This puts the explanatory emphasis where it belongs—on the decision by governing officials in France and Germany to establish an explicit link between their two currencies. While the Italian government could still have chosen to retain its monetary independence, this “noncooperative” strategy would have entailed (or so nearly everyone at the time believed) extraordinarily high costs. Indeed, over the course of the previous three decades, the idea that Italy had to maintain its European “credentials” in order to prosper economically had come to be accepted as an article of faith by both the left and right of Italy’s political establishment.26 While one can fairly question the rational basis for this view, the fact remains that politicians across the Italian political spectrum put a premium on their country’s European connection. That being the case, it is not hard to understand why they would have been reluctant to exclude themselves from the new Franco-German regime. By directly undermining Italy’s claim to being an integral part of “Europe,” a refusal to join this regime would have cast a shadow over a government that was not (as we shall see) all that stable to begin with.27 Nor did it help matters that the EMS regime, unlike the less comprehensive and ultimately short-lived “Snake” regime that preceded it, “was publicly related to other aspects of European integration, in ways that implied that a country not in the ERM would become a second-tier member of the Community.”28 26
See, e.g., Spaventa 1980. Ludlow 1982, 147– 50; see also De Cecco 1989, 90. 28 Frieden 1994, 33. Founded in 1972, the Snake required its members to keep their currencies from fluctuating against other member currencies by more than 2!f percent in either direction. The original idea was that Snake members would limit fluctuations against one another’s currencies by plus or minus 2!f percent yet remain within the wider (64!s percent) limits 27
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
183
In short, the announcement made by the French and German heads of state at the Copenhagen summit served to limit the range of possibilities for Italy. Once President Giscard and Chancellor Schmidt had resolved to link their currencies, public officials in Italy concluded that the price of exclusion— Italy’s being relegated to the European periphery, with all of the symbolic and economic costs that this implied—would have been even greater than the price of entry. To be sure, the risks of joining were considerable: a further deepening of austerity and the acceptance of new restrictions on Italy’s ability to use currency depreciation as a cushioning device during periods of depressed demand (see below). Given the alternatives, however, Italy would simply have to take its chances. This line of argument can explain why, from the very outset of the 1978 negotiations, Andreotti was predisposed to join the new Franco-German monetary scheme “if it was in any way possible to do so”29 —a sentiment echoed in the run-up to the Brussels summit by Andreotti’s foreign minister, Renato Ruggiero, who expressed his position as follows: I believe that it is very important for us, and there has been the clearest political will of the Government in this sense, to participate in the EMS since its inception, negotiating the conditions in the best possible way. Today we need the Community more than ever. If we did not participate in the EMS, this would show that we are unwilling to accept the challenge . . . of being a fully European country.30
Even the Communists, whose supporters were likely to be among the hardest hit by the unemployment effects of Italy’s new exchange rate commitment, were reluctant to keep Italy out of the system. Finally on the threshold of national power, the Communist leadership did not wish to appear “anti-Europe.”31 Although the PCI refused to vote in favor of the EMS, its leaders worked out an arrangement whereby the Italian Socialist party would abstain, thus ensuring that Andreotti would have enough votes to introduce the lira into the ERM and averting an all-out government crisis.32 This is not to say that Italy’s political leadership welcomed the new arrangement. What Giscard and Schmidt did, in effect, was to raise the deffixed by the gold standard; hence, the notion of a “snake in a tunnel.” After the collapse of the Bretton Woods system, the tunnel disappeared and the currencies of Snake members were permitted to float freely against those of non-participants. In 1972, Italian authorities agreed to participate in this system. Less than six months later, however, Italy suffered a severe economic downturn, prompting the government to withdraw from the arrangement and chart an independent course. Kruse 1980, chap. 6. By 1976, Britain and France had also dropped out of the system, rendering the Snake little more than a “deutsche mark zone” comprised of Germany and its smaller neighbors. 29 Ludlow 1982, 115. 30 Quoted in Spaventa 1980, 69. 31 Emblematic of its desire not to appear reckless or irresponsible was the PCI’s public opposition to left-wing terrorism, which reached its height with the 1978 kidnapping and subsequent slaying of the Christian Democratic leader and former prime minister Aldo Moro. 32 Ludlow 1982, 270 –73.
184
E U R O P E A N
M O N E Y
initional threshold for being considered a “good European.” Up until that point it had sufficed for a country to participate as a member in good standing of the EC, a standard that Italy could maintain without great difficulty.33 Now, however, Italy was being asked to meet a higher standard. In order to maintain its European connection, Italy would now have to accelerate its disinflationary trajectory to a degree that was neither economically nor, from the point of view of the center-left majority governing Italy at that time, politically optimal. The anxieties created by all of this were intertwined with a pervasive sense of resignation, a mixture nicely captured at the time by the official spokesman for Italian industry, Guido Carli. A former governor of the Banca d’Italia, Carli observed: “If Italy is too weak to participate in the EMS, it is also true that she is too weak not to participate.”34 A VICTIM OF COERCIVE PRESSURE?
Too weak, yes; coerced, no. At no point did the French and German beneficiaries of the EMS use force or intimidation to “bully” Italy into joining. Nor, for that matter, did Giscard and Schmidt resort to subtler, indirect forms of coercive pressure—promising, for example, to exclude nonsignatories from the benefits of cooperation in other areas. One version of this coercion-through-linkage argument holds that Italian political elites consented to the terms of the EMS only because the arrangement’s French and German sponsors had been threatening to shut Italy out of the larger process of EC integration. That Italian politicians did not wish to be perceived as having “missed the European train” is certainly true. The question is whether this train’s departure was part a deliberate strategy on the part of the new monetary regime’s French and German beneficiaries to manipulate Italy’s decision-making calculus. In fact, President Giscard and Chancellor Schmidt seemed relatively unconcerned with whether or not Italy opted out of “their” regime—perhaps because in 1978 Italy’s economy was less than two-thirds the size of France’s (58.5 percent) and less than half that of West Germany (43.5 percent), and trade with Italy represented a fairly small proportion of both countries’ total trade (see Figure 8-2). Further belying the “strategic issue-linkage” explanation is the fact that Giscard and Schmidt let the United Kingdom opt out of the ERM without demanding that the country give up its affiliation with the European Community. British Prime Minister Jim Callaghan’s request for special status, which was made at an early stage in the negotiations (and the granting of which was never in doubt), made it highly unlikely that Italy would be singled out for punishment in the event that it, too, were to opt out of the ERM framework.35 33 When Andreotti pledged to join the EMS in 1978—and even when Mitterrand finally made up his mind to stay in the system in 1983 (see chapter 9)—the Single European Act and the broader “1992” agenda were still years away. 34 Quoted in Ludlow 1982, 210. 35 During the early stages of the EMS negotiation process, Callaghan had expressed concern
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
185
Figure 8-2. France and West Germany’s Major Trading Partners, 1978
Britain versus Italy: A Natural Experiment What, exactly, did members of the Italian political establishment stand to lose by Italy’s being relegated to the European periphery? If the argument I have been making thus far is correct, their desire to avoid severing Italy’s European connection was at least partly motivated by their expectations about that Britain’s limited participation would prevent it from exercising influence over the character of the regime as it evolved over time. Ludlow 1982, 246. To allay these concerns, Callaghan made his support for the EMS resolution contingent upon the inclusion of a proviso (Article 3.1) granting any EC country that initially refused to join the exchange rate mechanism the right to join it later on. Callaghan also demanded that all EC members be given the right to participate in important decisions over parity realignments and other aspects of exchange rate policy (Article 3.2).
186
E U R O P E A N
M O N E Y
how the financial markets would react to such an occurrence. Once the envisaged “zone of monetary stability” had become a reality, the expectation was that EC members located outside this zone would find themselves at a relative—and growing—disadvantage in attracting foreign capital. Underlying this prediction were three basic assumptions: first, that the explicit link between their own currencies and the D-mark would afford ERM member states greater protection against future inflation; second, that countries lacking this protection (e.g., the nonparticipating “laggards”) would face the specter of massive capital outflows; and third, that this would translate into higher interest rates, and ultimately slower growth and higher unemployment, in the non-ERM countries than in their ERM-participating neighbors.36 But did all of these assumptions really make sense? How can we know what Italy’s interest rate premium would have been if Andreotti had refused to introduce the lira into the ERM? The short answer is we cannot know, at least not for certain. But while there is no sure-fire method for determining how Italy would have fared as an ERM outsider, all is not lost. In this case, history has conspired to provide an appropriate point of reference—the United Kingdom. THE EU’S MOST RELUCTANT REGIONALIST
British authorities might well have entered the pound into the ERM at the inception of the European Monetary System in 1979. Their country was, after all, a member in good standing of the European Community. For a variety of idiosyncratic reasons, however, the British government delayed becoming a full-fledged member of the system until 1990. The Italian case thus invites comparisons with Britain’s experience as an ERM outsider, making it a logical choice for further investigation. Numerous considerations entered into Britain’s decision to remain on the sidelines of the EMS throughout the 1980s.37 Part of its hesitation had to do with the inherent incompatibilities between the targeting of monetary aggregates, the centerpiece of Thatcher’s macroeconomic agenda during her first term in office, and the targeting of the exchange rate necessitated by membership in the ERM. Upon taking office in 1979, the new Conservative prime minister announced that her government would strictly adhere to a set of previously announced five-year projections of the rate of growth in the money supply. It would not waver from these targets, Thatcher insisted, even if it meant sharply raising taxes in the midst of an economic recession.38 If 36 For evidence that foreign exchange controls had, by the 1970s, lost whatever efficacy they may once have had, and hence that European countries were highly permeable to capital movements across borders, see Gros and Thygesen 1992, 114 –26, and Marston 1993. 37 As in Italy, the announcement of the Franco-German proposal provoked a strong—and arguably well justified—negative reaction from power holders in the UK (see below). My discussion of British behavior draws on Dennis and Nellis 1986; Harmon 1994; Lomax 1987; Miller and Sutherland 1992; Moravcsik 1998; D. Smith 1993; and Tsoukalis 1989. 38 The underlying rationale was the “tying-hands” argument we encountered earlier. Pro-
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
187
there had ever been any doubt about the Conservative government’s commitment to monetary targeting, these were put to rest in 1981 (albeit, as it turned out, only temporarily) when, in the depth of a severe recession, Thatcher defied the advice of numerous British economists and issued a budget calling for higher taxes and sharp reductions in public expenditure. Guided as it was by a strict monetarist agenda, British macroeconomic policy during the early 1980s was fundamentally incompatible with membership in the ERM. Insofar as the pound during this period was appreciating rapidly against the D-mark and other European currencies, the British government sooner or later would have been called upon to loosen its monetary policy, at which point the Conservatives’ monetary and exchange rate commitments would have come into direct conflict. This fundamental opposition between Thatcher’s monetarist principles and participation in the ERM was explicitly acknowledged in a speech delivered in 1981 by Financial Secretary Nigel Lawson: “The present Government has no exchange rate policy as such—for the simple reason . . . that the attempt to have such a policy greatly complicates (if it does not actually make it impossible) the difficult enough task of pursuing a sound monetary policy.”39 Complications such as these certainly played an important role in the Conservative government’s (initially) tepid response toward the EMS. Of even greater importance, however, were the perceived symbolic costs of British entry. Had Thatcher relinquished control of the pound upon taking office— or if Callaghan, her Labour predecessor, had done so when the EMS entered into force in 1979—a large segment of the British electorate would have regarded the act as a blatant and wholly unwarranted encroachment on their country’s long tradition of political independence. Whatever the precise motivation, the fact that Britain’s leaders delayed introducing the pound into the ERM enables us to conduct a “natural experiment.”40 Did Britain pay a price for its self-imposed isolation—a price that Italy, too, would have had to bear had it chosen not to participate in the ERM? Let us begin by looking at the interest rate differentials displayed in Figure 8-3. MAINTAINING CREDIBILITY OUTSIDE THE ERM
If attention is restricted to the first part of the 1980s, Britain’s relatively low interest rates might seem to belie the claim that the UK suffered a credibility loss as a result of its nonparticipation in the ERM. Between 1980 and 1985,
ponents of this argument believed that Thatcher’s uncompromising adherence to preannounced monetary targets would encourage wages to adjust quickly, if not instantaneously, to the new, low-inflation reality. And if for some reason they did not quickly adjust, the resulting increase in unemployment would still have the beneficial effect (from Thatcher’s standpoint) of weakening the British labor movement. 39 Quoted in D. Smith 1993, 47; see also Harmon 1994. 40 Cf. King, Keohane, and Verba 1994, chap. 4.
188
E U R O P E A N
M O N E Y
Figure 8-3. UK–ERM Interest Rate Differentials, 1979–1993
British interest rates were typically lower, not higher, than those of other ERM member states. The absence of a persistent interest rate differential over this period comes as little surprise, however, given that—initially, at least—the Thatcher government seemed intent on wringing every last drop of inflation out of the British economy, even, in fact, if that meant presiding over several straight years of double-digit rates of unemployment. That being the case, the financial community had reason to be confident that Britain’s inflation gap with France and Germany would continue to decline even if the Tories were to insist on maintaining exclusive control over the value of the pound. In short, any misgivings the financial markets might have had about the UK’s refusal to submit to the anti-inflation discipline of the ERM would appear to have been offset—perhaps more than offset—by Thatcher’s unparalleled, and seemingly unflappable, commitment to price stability. As the 1980s wore on, however, officials within the Thatcher cabinet— everyone, it seemed, but the prime minister herself—began to see Britain’s self-imposed monetary isolation as a major political and economic liability. By the start of Thatcher’s third term in 1987, there were few illusions that the presence of a fiscal conservative at the helm of British politics would be sufficient, given lingering rigidities in the British labor market and an unprecedented expansion of consumer credit, to induce a rapid, much less an instantaneous, decline in domestic inflationary pressures. Though Thatcher had been able to hold the line against inflation during her first two terms,
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
189
she had been forced to do it the old-fashioned way—by swelling the ranks of the unemployed—and it had not been easy. To be sure, large-scale unemployment was not as great a political liability for the Tories as it would have been for a Labour government. If they voted at all, unemployed workers were highly unlikely to vote Conservative. But that did not mean that unemployment was politically unproblematic for the government. On the contrary, though they themselves may not have been at risk, most British voters nonetheless rated unemployment “the major problem facing the country.”41 Nor, by the start of its third term, did the Conservative government have the luxury of using North Sea oil or privatization revenues to cover the costs of unemployment compensation. Finally, and perhaps most importantly, the opposition Labour party, though still greatly outnumbered in the House of Commons, had moved aggressively to the center of British politics, divesting itself of its militant faction (the “loony left”) in an effort to gain credibility among the broader electorate. For the first time since coming to power, the Tories’ once-formidable lock on power appeared to be dissipating. THE ECONOMIC CONSEQUENCES OF BRITAIN’S SELF-IMPOSED ISOLATION
As her government’s earlier willingness to bear the brunt of double-digit rates of unemployment began to decline, Thatcher found herself facing a predicament not unlike the one that had confronted Italian policymakers in 1978. As much as she disliked the idea of relinquishing control of the pound, she recognized that there was simply no other way—short of continuing to drive up Britain’s already high rate of unemployment—of demonstrating her government’s anti-inflation resolve. Even so, the prime minister refused to give in until October 5, 1990. In the meantime, and as Thatcher’s highly regarded Chancellor of the Exchequer, Nigel Lawson, never tired of pointing out, Britain’s special status as the only major EC member not to participate in the ERM caused the risk premium attached to sterling-denominated investments to increase relative to the premium on investments denominated in other European currencies. In turn, this rising differential translated into higher interest rates—and thus a slower rate of growth—than might have been necessary if Thatcher had followed Lawson’s advice in favor of immediate entry.42 Indeed, Lawson had been urging the prime minister to give up control of the pound since 1985. By the time Thatcher took Lawson’s his advice, Britain had endured nearly six straight years in which money market rates in the UK exceeded those of its ERM-participating neighbors. Nor are we talking here about trivial differences: during the two years between September 1988 and September 1990—a period characterized, not coincidentally, by a dramatic slowdown in Britain’s overall rate of growth—interest rates in the UK were consistently 41 42
Garrett 1993, 538. See Lawson 1992, 111.
190
E U R O P E A N
M O N E Y
more than four points higher than those found in ERM member states (Figure 8-3).43 Britain’s exclusion from the ERM was not the only reason for these disparities. Nevertheless, the view was widespread that British entry into the ERM, by calming investors’ fears about an imminent rise in inflation, would go a long way toward alleviating the problem. And indeed, when Thatcher did finally acquiesce (after countless hours of lobbying by Lawson and others), the result was a dramatic, overnight decline in British interest rates.44 By the summer of 1992, interest rates in the United Kingdom were running more than four points lower than those in other ERM countries, suggesting that British entry did in fact lower the perceived risk to international lenders of holding sterling-denominated securities. The lesson to draw here, then, is not that Britain’s exclusion from the ERM had no adverse credibility effects. Quite the contrary, its exclusion came at a price so high that, in the end, not even the “Iron Lady” herself was willing to pay it. The Costs of Nonparticipation: A Comparison of the Two Cases In justifying the UK’s entry into the EMS, the Thatcher government placed greatest weight on what might be termed “defensive” considerations—that is, on the costs of not participating in the Franco-German regime rather than on the merits of the regime itself. After Thatcher’s ouster as party leader in 1990, the defensive argument for continuing Britain’s participation was taken up by Norman Lamont, the Chancellor of the Exchequer appointed by Thatcher’s successor as prime minister, John Major. According to Lamont, a decision to “suspend” Britain’s membership would have been nothing short of disastrous. The following passage, excerpted from a speech that Lamont delivered before the European Policy Forum in July of 1992, nicely articulates the defensive case against British withdrawal: Many who advocate floating know full well what the consequences would be. They intend a devaluation of the pound. And they would certainly achieve it. For the result of leaving the ERM, combined with large cuts in interest rates, would be 43
It was because of Thatcher’s firm opposition to British entry during the early 1980s that Lawson began unofficially “shadowing” the deutsche mark in 1987. Insofar as this surreptitious policy restored (somewhat) the confidence of private investors, Lawson’s interventions may have temporarily diffused upward pressure on British interest rates. In the absence of a tacit currency link, it is likely that the risk premium on British investments, and hence the output and unemployment costs of deflation, would have been even higher than it was. 44 The government entered the ERM with a central parity of 2.95 deutsche marks to the pound, agreeing to limit fluctuations around this basic rate to 6 percent in either direction (the same band width that Italy enjoyed when it entered the system in 1979). For a detailed analysis of Britain’s accession, see Harmon 1994 and D. Smith 1993. Most contemporary accounts give credit to John Major, then the Chancellor of the Exchequer, for finally persuading the prime minister to change her position. See, e.g., Stephens 1990. Lawson 1992 provides an excellent firsthand account by one of the main protagonists.
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
191
a fall in the pound probably unprecedented in the last forty years. It’s the cut and run option; cut interest rates and a run on the pound. . . . Markets would see that, unlike all our major European competitors, Britain lacked the will to pursue the goal of permanently low inflation. And they would be right. We would have given up after less than two years. They would conclude that we were back to our bad old ways; that given the chance we would always delude ourselves by thinking that with a little more inflation we could get a little more growth.45
As this statement makes clear, British policy makers viewed the costs of nonparticipation mostly in terms of how the “markets” would react. In Italy, the costs of exclusion were viewed somewhat differently. As sensitive as he was to Italy’s reputation within the global financial community, Andreotti was also motivated, as noted earlier, by a separate, more diffuse set of concerns: by opting out of the ERM, Italy would have been relegating itself to the sidelines of “Europe.” Even if this opt-out option had had no financial implications whatsoever, its negative symbolic implications—the notion that Italy had somehow been “left behind”—would probably have been enough to tip the scales in favor of entry. This was not the case in the UK. Though Britain’s status within Europe’s emerging architecture mattered a great deal to some British citizens—and even, as we will see, to a few members of the Thatcher cabinet—the prime minister herself viewed Britain’s European connection as more of a political liability than an asset. Indeed, Thatcher’s wing of the Conservative party would have been only too happy to see Britain opt out of the “new” European Union, a body that, except perhaps for the liberalizing reforms enshrined in the EU’s Internal Market, they did not look upon with anything remotely resembling enthusiasm.
Were the EMS Losers Really Worse Off? Even if the fear of “missing the European train” was more salient in the case of Italy than Britain, the basic point—that the high costs of exclusion rather than the regime’s (ostensible) benefits are what prompted both countries to take part in the EMS—would seem beyond dispute. But does this mean Italy and Britain were worse off? Even if I am right that defensive considerations are what induced the Italians and, eventually, the British to take a seat on the fixed exchange rate bandwagon, it does not follow that their participation in the new monetary regime was a bad thing. Is it really true that the Italian and British governments would have preferred the pre-EMS world of freely floating currencies? The answer isn’t obvious. In the case of Italy, for example, one could conceivably make the following sort of argument. Even if it is true that (1) France and Germany’s go-it-alone initiative removed the ex ante floating-rate status 45
Quoted in D. Smith 1993, 237.
192
E U R O P E A N
M O N E Y
quo and (2) this in turn “forced” Andreotti to do something he would not otherwise have done, it does not necessarily follow that (3) Andreotti’s decision undermined his government’s economic or political welfare. The current conventional wisdom is that Andreotti’s decision to surrender control of the lira actually improved his government’s utility, though, to be sure, the fact that Italy’s political establishment secretly delighted in their monetary “predicament” was not something they wished to broadcast to the Italian public. As it happens, this view is in keeping with a well-established—though much-criticized46 —body of economic theory according to which membership in a fixed exchange rate regime makes it possible for countries to reduce their rates of inflation without at the same time increasing their rates of unemployment. If this had been the case, the Italian officials who introduced the lira into the EMS would have had nothing to worry about. But it was not the case, and here’s why. Italy’s Preference for the Pre-EMS Status Quo Proponents of the positive-sum interpretation make two critical assumptions. The first is that membership in the EMS acted as a kind of low-inflation commitment device.47 This is reasonable enough, since high-inflation countries could remain in the system only by working aggressively to reduce their inflation differentials with Germany, the regime’s low-inflation anchor. Less plausible is the second assumption—namely, that wage setters in a highinflation country like Italy, anticipating their government’s new, harder-line disinflationary stance, would moderate their wage demands in advance, thus obviating the “need” for high unemployment during the country’s post-entry adjustment period. ASSESSING THE “DISCIPLINARY” RATIONALE FOR PEGGING THE LIRA
At the time that Andreotti made his decision, the expectation of most observers was that Italian labor markets would in fact be quite slow to adjust to the new, post-EMS (i.e., deflationary) reality. This expectation was well justified. Not only does the disciplinary argument presume that Italian wage setters trusted their government to stay in the system (i.e., not to resort to “unwarranted” depreciations); it also requires that the institutional framework governing the wage-setting process be free of major distortions, allowing labor contracts to be quickly and easily renegotiated in response to sudden changes in inflationary expectations. Over time, however, Italy’s labor market had in fact grown more, not less, distorted and inflexible. In 1975, for example, though Italy was in the mid46 See, e.g., Blanchard et al. 1993; Dornbusch 1989 and 1993b; Eichengreen and Wyplosz 1993; Feldstein 1997; and Krugman 1993a. 47 See, e.g., Giavazzi and Pagano 1988.
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
193
dle of its deepest recession since the end of World War II, Italian labor unions nonetheless succeeded in getting the employers’ association to accept an automatic wage indexation scheme. Known as the scala mobile, or “escalator,” this scheme was intended to keep workers’ wages from being eroded by future inflation. Although the government announced its opposition to the scheme in 1977, the combination of double-digit rates of inflation and strong labor support for indexation severely undermined the short-term prospects for wage moderation. The success of the campaign initiated by Italian unions in behalf of the scala mobile underscores the comparative strength of organized labor in Italy. Encompassing approximately 45 percent of the total labor force, unionized workers in Italy were in a strong position to disrupt the operation of market forces in the wage-setting process.48 The structure of Italy’s labor movement thus made it likely that any EMS-induced downward adjustment in wages would take considerably longer than in most other EMS-participating countries. As for the prospects of obtaining a coordinated, “corporatist”-style wage reduction, current research suggests that two conditions—both noticeably absent in Italy—would have been required.49 The first requirement is that workers expect to be compensated for restraining their wage demands. When it comes to Italy, however, even if Andreotti had wanted to offer such compensation (the PCI certainly did), his capacity to increase public expenditures on the most direct form of compensation—welfare—was severely constrained by Italy’s fiscal situation. By the late 1970s, the Italian government was laboring under a national debt whose size rivaled that of the country’s total annual economic output. In this context, sharply increasing deficit spending would have meant adding to Italy’s existing interest burden and, in so doing, sparking a potentially lethal deficit-interest spiral. The second prerequisite for sustained wage restraint—a labor movement that is sufficiently centralized and encompassing to ensure that workers’ good-faith efforts to restrain their wage demands are not undermined by collective-action problems within the labor movement—was also lacking. The obstacle here was not the density of the Italian labor movement; as just noted, nearly half of Italy’s workers were unionized, a fairly high proportion by Western European standards of that time. The obstacle was the Italian labor movement’s lack of centralization. Organized workers in Italy were divided among three competing camps: the Confederazione Generale Italiana del Lavoro (CGIL), Confederazione Italiana Sindacati Lavoratori (CISL), and the Unione Italiana del Lavoro (UIL). Not only were there factional rifts between these confederations; there were also conflicts within them, with Communists competing against Socialists in the CGIL, moderate Catholics 48 In France, by comparison, organized labor encompassed only about one-fifth of the total workforce. 49 See, e.g., Alvarez, Garrett, and Lange 1991; Lange and Garrett 1985.
194
E U R O P E A N
M O N E Y
against militant left-wing Catholics in the CISL, and Socialists against Social Democrats and Republicans in the UIL. Given these internal divisions, it comes as little surprise that Andreotti’s efforts to promote “institutionalized collaboration” between Italian workers and employers went nowhere: as unions intensified their militancy, unemployment and inflation threatened to run completely out of control.50 CONCERNS ABOUT THE PROSPECT OF EMS-DRIVEN UNEMPLOYMENT
To Andreotti, the fact that Italy’s EMS-induced wage adjustment was likely to occur relatively slowly, with Italian workers experiencing considerable economic upheaval during a lengthy post-entry transition period, was a serious cause for concern.51 While the prospect of higher short-term unemployment would have struck fear within any Italian government, it was particularly alarming to Andreotti’s government, beholden as it was to the Communist party. Not that everyone in Italy was fearful of the EMS: for those on the right, the appearance of the regime was like manna from heaven. As the chart in Figure 8-4 makes clear, the inauguration of the EMS was followed by a precipitous decline in Italy’s rate of inflation.52 While the most obvious beneficiaries of this drop were Italian-based multinational corporations and banks heavily engaged in international transactions, economic theory suggests that companies specializing in the production of nontradable goods and services would have benefited as well.53 Be that as it may—and this is the important point—the officials who held positions within Italy’s government did not rely exclusively upon these economic sectors for their electoral, or even their financial, support. Far from being the handmaiden of any one constituency, the dominant member of Andreotti’s governing coalition—the prime minister’s own Christian Democratic party (DC)—was in fact highly factionalized, its divisions revealed in heated internal debates during the course of the negotiations leading up to the Brussels summit in 1978.54 Moreover, about half of the DC’s electoral support came from rural voters, many of whom lived in southern Italy, a 50
Flanagan, Soskice, and Ulman 1983, 555 – 61; see also Posner 1978. For evidence that Italy’s “sacrifice ratio” during the 1980s was consistently higher than that of many other OECD countries, see Dornbusch 1989 and De Grauwe 1990. 52 Before crediting Italy’s success in combating inflation to its 1979 EMS commitment, one should note that the disinflation records of EMS and non-EMS member states did not differ substantially (Collins 1988; De Grauwe 1990; Dornbusch 1989; Fratianni and von Hagen 1992; Weber 1991; Ungerer et al. 1986). That said, many of the econometric studies purporting to show a weak or statistically insignificant relationship between a country’s membership in the regime, on the one hand, and changes in its price structure, on the other, fail to take into account the fact that inflation in some EMS signatories (e.g., the Benelux countries) had already converged on Germany’s by the time the regime entered into force. Statistical tests of the disinflationary effects of the regime also include Austria and Switzerland in the non-EMS control group even though their currencies were informally linked to the deutsche mark throughout the 1970s and 80s (Frieden 1994, 28 n. 6; Grilli 1994, 74). 53 See, e.g., Frieden 1991. 54 Ludlow 1982, 205 –17. 51
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
195
Figure 8-4. Consumer Price Inflation in Italy, 1977–1990
much poorer region than the thriving, more cosmopolitan north.55 As one account put it: the DC saw itself more as a populist party than one whose main appeal was to the middle classes and business. It thus found itself acting . . . as an aggregate of interest groups—small-holding peasants, the traditional middle classes, small-scale industry, and the Catholic church, as well as important sections of the industrial working class and big business—mimicking at the national level the “clientelistic,” or patronage-based, structure of decisionmaking at the local level.56
Finally, even if the interests of Italy’s internationally oriented businesses had held sway within the Christian Democratic party, Italy’s multiparty system made it impossible for the Christian Democrats to govern except as part of a broad center-left coalition. And, as noted earlier, Italy’s decision to join the EMS occurred at a time when the domestic balance of power was shifting toward the left, not the right. The origin of this shift can be dated to Communist party secretary Enrico Berlinguer’s 1973 announcement that the PCI was no longer interested in overthrowing the government. The prospect of a “historic compromise,” wherein the Italian Communists would agree not to disrupt the operation of democratic institutions in exchange for some degree of influence over national policy, in turn produced a dramatic upsurge in the PCI’s electoral standing. After the party’s breakthrough success in the national legislative elections of 1976—the Communists’ share of seats in the Chamber of 55 A 1975 survey conducted by Giovanni Sartori and Alberto Marradi (cited in Spotts and Wieser 1986, 34) found that 48 percent of the Christian Democratic party’s supporters resided in rural areas. 56 Flanagan, Soskice, and Ulman 1983, 498 – 99.
196
E U R O P E A N
M O N E Y
percent57 —it
Deputies went from 27 to 34 was clear that any government formed by the Christian Democratic party would have to be acceptable to the PCI as well. In fact, three of the parties on whom the Christian Democrats relied for their parliamentary majority (the Socialists, Social Democrats, and Republicans) had all announced that they would refuse to serve in any government that did not accommodate the PCI’s policy preferences.58 Thus, even though the Communists were never awarded ministerial portfolios within the government itself, their support was vital to the success of Andreotti’s governing coalition. A NOTE ON SIDE-PAYMENTS
At their July 1978 gathering in Bremen, the European heads of state resolved to undertake measures “to strengthen the economies of the less prosperous [EMS] member countries.” From that point on, it was clear that some sort of financial aid package would be incorporated into the new monetary agreement. The question, of course, was what sort of package: How much aid would the regime’s wealthier participants be willing to sacrifice for the benefit of their poorer partners? The original plan was to settle the matter in a special series of intra-Community negotiations to be held—and concluded—prior to the upcoming European Council summit in Brussels. While representatives of the Irish government approached this forum with a detailed request for assistance amounting to approximately 2 percent of Ireland’s gross domestic product each year for five years, Italy’s representatives initially refused to put a precise figure on their own demands, saying only that (1) grants as well as loans should be considered and (2) substantial increases in expenditures allocated through the Community’s existing redistributive arrangements, such as the European Regional Development Fund, would be necessary.59 After the relevant negotiations ended in deadlock, attention turned to Brussels. How much would France, Germany, the Benelux countries, and Denmark be willing to dispense? The answer was soon made apparent when, in a statement leaked to the press, Giscard declared: “France cannot upset her own financial arrangements in order to ensure the adhesion of those for whom membership ought to be an act of political will rather than a question of cash.”60 In the end, the offer to Ireland amounted to only slightly more than a third of its request. Italy fared even worse. Although Andreotti had reportedly asked for assistance totaling approximately 360 billion lira, to be divided roughly equally between grants spread out over three years and interest rate subsidies, he was 57 Regarding the magnitude of the Italian Communist party’s electoral upsurge, Sassoon (1981, 227) notes that “while it took 26 years to increase its percentage by eight points (19% in 1946 to 27% in 1972), it took only another four years to jump by another seven points.” 58 Sassoon 1981, 227–28. 59 Ludlow 1982, 174 –75; see also Spaventa 1980, 80 – 81. 60 Quoted in Ludlow 1982, 264 – 65.
W I N N E R S
A N D
L O S E R S
I N
T H E
197
E M S
year.61
granted only 73 billion lira worth of subsidies per Roughly a fifth of what Italy had been seeking, this sum was equivalent to less than 0.03 percent of Italy’s GDP in 1978.62 Deeply disappointed, Andreotti joined with Ireland’s prime minister, Jack Lynch, in requesting a “pause for reflection.” Less than two weeks later, however, both countries accepted the financial assistance offers made at Brussels—with some slight modifications in the Irish case—and jumped on the EMS bandwagon. (The offer to Ireland was increased slightly from 45 to 50 million pounds over two years and was changed from subsidies to grants.63 Italy’s aid package was not renegotiated.) In the end, therefore, though some analysts have emphasized the role of side-payments in reducing Italy’s burden of austerity,64 the amount of financial assistance the EMS regime’s wealthier member governments were ultimately willing to make available for the purposes of equalization did not come close to compensating for the wrenching economic dislocations that countries like Italy, which had previously relied upon currency depreciation to create jobs and stimulate growth, seemed likely to endure. Even if one acknowledges that Italian representatives had an incentive to overstate their true needs, the disparity between the amount of financial support that Italy demanded and the amount that its more affluent EC partners agreed to dispense is striking.65 This should not be taken to imply that side-payments were completely inconsequential. If nothing else, they made it possible for Italian negotiators to claim credit for having won concessions from their French and German partners. In reality, however, the offer of financial assistance was little more than a symbolic gesture. Not only was the amount of concessionary finance pitifully small, but its provision was also contingent on each would-be recipient’s ability to do the necessary paperwork ahead of time—a task that turned out to be beyond the administrative capacities of many of the Italian provincial authorities who qualified for (but never actually received) financial assistance.66 THE PARLIAMENTARY ELECTIONS OF 1979
Not surprisingly, politicians on the left of Italy’s political spectrum exhibited the greatest discomfort with Andreotti’s decision to take part in France and Germany’s EMS initiative. The ambivalence of the PCI was particularly understandable: members of the Italian working class were already somewhat apprehensive about supporting Communist candidates for national office. Nor, given Italy’s multiparty system, was there much chance of stopping dis61
Ludlow 1982, 266 – 67. Author’s own calculations. 63 Ludlow 1982, 269. 64 See esp. Grieco 1990. 65 P. Taylor 1980, 381– 83. 66 Alter 1993. 62
198
E U R O P E A N
M O N E Y
Table 8-1 Popular Vote in Italian Parliamentary Elections, 1953-1983 1953 (%)
1958 (%)
1963 (%)
1968 (%)
1972 (%)
1976 (%)
1979 (%)
1983 (%)
PR
40.1 (263)b 22.6 (143) 12.8 (75) 4.5 (19) 1.6 (5) 3.0 (13) —
42.4 (273) 22.7 (140) 14.2 (84) 4.5 (22) 1.4 (6) 3.5 (17) —
38.3 (260) 25.3 (166) 13.8 (87) 6.1 (33) 1.4 (6) 7.0 (39) —
39.1 (266) 26.9 (177) 14.5 (91) —c — 2.0 (9) 5.8 (31) —
38.8 (267) 27.2 (179) 9.6 (61) 5.1 (29) 2.9 (14) 3.9 (21) —
DP
—
—
—
—
—
— — 5.8 (29) 9.6 (43)
— — 4.8 (24) 6.5 (30)
— — 5.1 (27) 3.0 (12)
— — 4.4 (24) 7.3 (32)
— — 8.7 (56) 4.0 (4)
38.7 (262) 34.4 (228) 9.6 (57) 3.4 (15) 3.1 (14) 1.3 (5) 1.1 (4) 1.5 (6) — — 6.1 (35) 0.8 (4)
38.3 (262) 30.4 (201) 9.8 (62) 3.8 (20) 3.0 (16) 1.9 (9) 3.5 (18) 0.8 (—) 1.4 (6) 5.3 (30) 2.7 (6)
32.9 (225) 29.9 (198) 11.4 (73) 4.1 (23) 5.1 (29) 2.9 (16) 2.2 (11) 1.5 (7) —d — 6.8 (42) 3.2 (6)
Partya DC PCI PSI PSDI PRI PLI
PdUP MSI Otherse
Source: Spotts and Wieser 1986, app. A; Istituto Centrale di Statistica (ISTAT) aDC (Democrazia cristiana), PCI (Partito communista italiano), PSI (Partito socialista italiano), PSDI (Partito socialista democratico italiano), PRI (Partito repubblicano italiano), PLI (Partito liberale italiano), PR (Partito radicale), DP (Democrazia proletaria), PdUP (Partito di unità proletaria per il communismo), MSI (Movimento sociale italiano). bFigures cPSI
in parentheses are the party’s number of seats.
and PSDI presented joint election lists.
dRan
on PCI lists.
eIncludes
Monarchist party, South Tyrol People’s party (SVP), Sardinian Action party (PSA), Valdôtaine Union (UV), and Socialist Party of Proletarion Unity (PSIUP).
affected working-class voters (i.e., voters who might have supported the PCI but for its having “sold out” to the right) from defecting to other viable leftwing parties. Analysis of the 1979 election suggests that many Communist voters did in fact jump parties. Table 8-1 reveals that the PCI, though the recipient of less than a quarter of the popular vote during the 1950s, steadily advanced over the course of the 1960s and early 1970s until, in 1976, it achieved a major
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
199
Table 8-2 Election Results Immediately before and after Italy’s Entry into the EMS: A Closer Look
Parties
1979 Results (% popular vote)
Difference with 1976
DC PCI PSI PSDI PRI PLI PR MSI PdUP
38.3 30.4 9.8 3.8 3.0 1.9 3.5 5.3 1.4
20.4 24.0 10.2 10.2 20.1 20.6 12.4 20.8 [no previous results]
breakthrough, capturing 34.4 percent of the vote and 228 of the 630 seats in the Chamber of Deputies, vote- and seat-shares that rivaled those of the Christian Democratic party. In the parliamentary election of 1979, however, the Communist party experienced a severe setback. The 4 percentage point decline in the PCI’s share of the vote—enormous by Italian standards—was the party’s first such drop since 1953. The biggest source of the Communist party’s electoral woes was the defection of PCI supporters to the upstart Radical party.67 A loose grouping of left-wing libertarians, the Radical party enjoyed a 2.4-percentage-point increase in its previous share of the popular vote (see Table 8-2) and captured eighteen seats—almost as many as the Social Democrats and two more than the Republicans. After having its previous electoral gains wiped out virtually overnight, it was to take the PCI another seventeen years to regain the trust of the Italian electorate.68 While the Communists saw their number of seats in the Chamber of Deputies fall from 228 to 201, the Christian Democratic party emerged from the 1979 election with the same number of seats (262) it had controlled since 1976. Nevertheless, the election results were a major blow to the Christian Democrats as well. More troubling than the decline in the DC’s vote share was the discrediting of the PCI, strange though that may seem to readers unfamiliar with Italy’s distinctive brand of partisan politics. In addition to drawing strong support from women, rural residents, and churchgoers, the Christian Democratic party had also drawn support from citizens who saw 67
Parisi and Pasquino 1980, 27. Under the leadership of Massimo D’Alema, the remnants of Italy’s Communist party, since renamed the Democratic Party of the Left, triumphed in national elections held on April 21, 1996, as a center-left coalition known as the Olive Tree won control of the Senate and captured the largest number of seats (284) in the 630-member Chamber of Deputies. 68
200
E U R O P E A N
M O N E Y
it as a counterbalance to the Communist party. After 1979, however, those citizens felt free to shift their allegiance to other parties (e.g., the Radicals). Although other factors were also at work—the decline of the DC’s near monopoly on patronage, for example—it is revealing that in the parliamentary election of 1983 the party’s vote declined from 38.3 percent to 32.9 percent, a drop of more than 5 percentage points. The party subsequently surrendered the prime ministership and invited Socialist leader Bettino Craxi to form the next government. In short, while some have speculated that Italian authorities took secret delight in joining the EMS—and, by inference, in relinquishing direct control over their country’s monetary policy—it is unclear why members of Italy’s governing majority would have taken this view. In fact, the political fallout generated by Italy’s entry into the system paints a rather sobering picture, one with which neither the centrist Christian Democrats nor, certainly, the left-wing Communists could have been particularly delighted. SHORT-TERM LOSSES FOR THE SAKE OF LONGER-TERM GAINS?
In response, it might be argued that the Italian officials who came out in favor of the EMS were simply taking the long view, putting their country’s long-term economic interests ahead of their own, more immediate political aspirations. Italy’s entry into the EMS was, in this view, a case of short-term political sacrifice—higher unemployment and slower growth during the initial adjustment phase, declining support (at least temporarily) for members of Italy’s political establishment, and so on—for the sake of longer-term gains. But was this really the case? As we have just seen, neither the Christian Democrats nor, least of all, the Communists enjoyed the kind of short-term electoral security that this sort of argument requires. And even if these parties had been preoccupied with advancing Italy’s long-term economic interests, it is doubtful that they would have chosen the path they did (surrendering control of the lira) had it not been for the go-it-alone “threat” posed by France and Germany. For while the inauguration of the EMS did promise certain long-term economic benefits for Italy—the stabilization of the lira-mark exchange rate, the end to double-digit rates of inflation, et cetera—it was not at all clear that these benefits would be worth the long-term economic costs. What were these costs? Once Italy joined the EMS, officials in the government lost the ability to cushion Italian industry and agriculture from temporary economic disturbances, known in the literature as supply and demand “shocks.” Ever since the first OPEC shock in 1973, Italian authorities had been using lira depreciation as part of a deliberate strategy to compensate for Italy’s uniquely vulnerable position (prior to OPEC Italy had been among Europe’s largest importers of foreign oil).69 As the Italian political economist Marcello De Cecco recalls: “The Italian monetary authorities had 69
Giavazzi and Spaventa 1989.
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
201
. . . been piloting the lira close to the wind, with the aim of gaining on the dollar and losing on the D-mark. This policy had been crowned by remarkable success. Italian exports, which had previously moved towards the American market, had invaded the northern European markets. Reserves had been rebuilt in a very short time and an attack was being launched on public debt and inflation.”70 In 1979, Italy’s entry into the EMS forced the government to abandon this devaluation-centered adjustment strategy, successful though it had proved to be during earlier years. Henceforth, any effort by Italian monetary authorities to compensate for “Italy-specific” shocks by temporarily driving down the value of the lira would have required the permission of the European Monetary Committee, any one of whose delegates, including the representative from the German Bundesbank, could withhold his consent. To be sure, the fact that EMS signatories could not readily alter their exchange rates to address country-specific shocks—except, as noted earlier, within certain predetermined (and relatively narrow) margins—did not much trouble the regime’s French and German cofounders. By the late 1970s, the French economy had already become closely linked to that of West Germany, whose economic ties with Belgium and the Netherlands, among other smaller European countries, were even more highly developed. Given the similarity in the underlying industrial structures of France and Germany, an external disturbance that injured French producers could also be expected to injure producers in Germany, and vice versa. Consequently, the chances were good that if France ever faced a legitimate need for a short-term exchange rate stimulus (to alleviate the burden imposed by a temporary downturn), so, too, would Germany. As recent research has shown, however, economic disturbances originating outside the EC, whether produced by sudden swings in the fiscal and monetary priorities of the United States or by fluctuations in the world price of oil, do not influence all EC members in the same way.71 By most accounts, Italy was indeed in a unique position. One reason—the Italian economy’s greater reliance upon foreign sources of oil—has already been noted. But the structural differences between Italy’s economy and those of countries located inside the EMS “core” ran much deeper.72 A revealing, albeit crude, indication of the distinctiveness of Italy’s industrial structure is the “symmetry” ranking presented in Table 8-3: Germany’s annual rates of GDP growth be70
De Cecco 1989, 90.
71 Renewed interest in European monetary unification has sparked a renaissance in economic
research on the relative merits of fixed versus floating exchange rates. Much of this literature draws on the pioneering work of Mundell (1961) and McKinnon (1963). For a lucid analysis dealing explicitly with European monetary integration, see Bayoumi and Eichengreen 1993. On the theory of “optimum currency areas” more generally, see De Grauwe 1992 and Ishiyama 1975. For an argument that the EC’s regions, as they become more specialized, will become more sensitive to externally generated, region-specific shocks, see Krugman 1993a. 72 See esp. Bayoumi and Eichengreen 1993.
202
E U R O P E A N
M O N E Y
Table 8-3 The Symmetry of Business Cycles in the European Community, 1965–1979 Country Germany Denmark France Netherlands Belgium [Greece]b United Kingdom [Austria] [Finland] [Portugal] [Spain] Italy [Sweden] Ireland
Correlationa
Symmetry Rank
— 0.77 0.71 0.67 0.66 0.59 0.58 0.56 0.53 0.49 0.468 0.467 0.29 0.16
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Sources: Garrett 1998, table 3; OECD, Historical Statistics, various years; and OECD, Economic Outlook 59 (1996), A4. aCorrelations between annual GDP growth rates in Germany and other European countries. bCountries in brackets were not members of the EC during the time period in question. (Greece was admitted into the EC in 1981. Spain and Portugal gained entrance in 1986.)
tween 1965 and 1979 were less correlated with Italy’s growth rates than with those of any other West European country except Sweden and Ireland. It is true that if Italy had had an effective substitute for currency depreciation—some possible candidates are discussed below—its vulnerability to Italy-specific macroeconomic disturbances would not have been especially problematic. If, in other words, Italy’s leaders had possessed some other means (i.e., other than devaluation) of temporarily restoring demand for Italian goods and services, the country could have continued to honor its ERM commitments without experiencing a protracted period of slow growth and high unemployment. In fact, however, Italy’s economy was anything but resilient. Consider what is perhaps the most visible substitute for devaluation: outward migration. In the United States, region-specific downturns have historically been followed by a major outflow of unemployed workers from the low-growth region (e.g., the rust belt) toward states located in higher-growth regions (e.g., the sun belt). This outflow serves as a kind of “natural” cushioning device, reducing the depressed region’s need for the temporary stimulus that currency depreciations can provide. In Europe, by contrast, even if a “re-
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
203
gion” (e.g., Italy) were to experience a prolonged period of stagnant growth, that region’s newly unemployed workers could not be counted upon to pick up and move to other, more prosperous regions (e.g., West Germany). By comparison with the United States, the mobility of labor within the EC remains quite low,73 a fact readily explained by the EC’s greater cultural and linguistic diversity. And if labor mobility is low today, after the “1992” agenda’s removal of all barriers to the free flow of persons, it was even lower when Italy first joined the EMS in the late 1970s.74 Outward mobility was not, of course, the only substitute for lira depreciation. If at some point after joining the EMS Italy were to suffer a sudden loss of competitiveness vis-à-vis the European “core,” Italy’s leaders could have offset the burden facing unemployed Italian workers by undertaking a fiscal expansion—which is to say, by spending more money. That, at any rate, is the theory. In practice, Italy’s potential for adjustment via fiscal policy was severely constrained by the public debt that successive Italian governments had run up over the course of the 1970s. If investors proved reluctant to purchase government debt (or were willing to do so only at very high interest rates), policies designed to artificially pump up demand were as likely to exacerbate Italy’s unemployment problems as to alleviate them.75 To be sure, things might have been different if, instead of having to spend its own way out of recession, Italy could have counted on major financial contributions from its European neighbors. In the United States, fiscal federalism provides an effective short-term palliative in emergency situations, cushioning region-specific shocks that cannot be dealt with through depreciation (since each U.S. region shares the same currency) and yet whose effects may be only partially offset by outward migration and deficit spending. In Europe, however, the structure of the EC’s federal funding mechanism—specifically, its inelasticity to temporary disturbances—did not facilitate its use as a stabilizing device. True, a European Regional Development Fund (ERDF) was established in 1975. But because the ERDF was targeted at regions where incomes were persistently, rather than only temporarily, below the EC average, it, too, was poorly equipped to address 73
Eichengreen 1992. Linguistic barriers are by no means the only reasons unemployed workers in Italy do not simply pack their bags and move elsewhere during periods of depressed demand. Eichengreen 1992, 18–20, shows that, within Italy itself, current rates of internal migration (e.g., from the poorer south to the wealthier north) are considerably lower than rates of intraregional migration within the United States. 75 Eichengreen 1992, 22–23. For analysis of why the loss of devaluation entailed in pegging their currencies to the deutsche mark was likely to be particularly painful for high-debt countries, see Sargent and Wallace 1981 and, more recently, Dornbusch 1989. In 1981, as part of its ongoing efforts to dampen inflationary pressures on the Italian economy, the Christian Democratic-led coalition released the Italian central bank from its previous obligation to purchase all unsold Treasury securities. For a discussion of the automatic financing of public deficits through money creation and its elimination in the so-called divorzio between the Banca d’Italia and the Treasury, see Goodman 1992 and Tabellini 1988. 74
204
E U R O P E A N
M O N E Y
short- or medium-term income fluctuations. In any event, the ERDF’s budget amounted to only 4.8 percent of the EC’s budget, which itself represented a tiny fraction—less than 5 percent—of the consolidated government spending of its member states.76 If devaluation, outward migration, deficit spending, and fiscal federalism are all ruled out, that still leaves adjustment through wages. But for wages to adjust rapidly to sudden swings in aggregate supply and demand, labor markets must be free of “structural rigidities”—which Italy’s were most certainly not. To the contrary, when Andreotti decided to join the EMS in 1979, the institutional framework governing the setting of Italian wages was among Europe’s most distorted, ossified, and inflexible. Compared with its ERM partners, then, Italy was both (1) more vulnerable to idiosyncratic, country-specific downturns and (2) less well equipped to respond to such shocks were they to occur. A prolonged period of lira “overvaluation” seemed inevitable. In the event, Italy’s participation in the regime did coincide with a significant real appreciation of the lira (see the first column of Table 8-4). Most damaging was Italy’s loss of competitiveness with respect to West Germany, its major trading partner. Between 1979 and 1990, the lira appreciated against the D-mark by some 42 percent. And that was after controlling for Italy’s higher rates of inflation. Bilateral real exchange rates are not, of course, the only indicators of “competitiveness.” Alternatively, one could look for EMS-induced effects on the lira’s real effective exchange rate. Because a currency’s real effective exchange rate (RER) is a trade-weighted average of each of its bilateral rates (controlling for inflation), an increase in the RER ratio between the lira and the D-mark would indicate a worsening of the Italian economy’s overall competitiveness vis-à-vis Germany. In surveying this data, the economists Giavazzi and Giovannini (1989, 89) find that the RER ratio between the lira and mark did in fact increase over the course of the 1980s, with Italy suffering about a 10 percent cumulative decline in its overall competitiveness. This finding is consistent with the work of another economist, Michael Artis, whose analysis uses movements in relative unit labor costs to determine how Italy’s involvement in the EMS may have affected its competitive position. On this criterion, Italy’s competitiveness deteriorated by as much as 50 percent.77 Findings like these do not sit comfortably with the view of the EMS as an unmitigated long-term “good” for the Italian economy, least of all for Italy’s export-producing firms and their employees. While Germany’s export sector grew rapidly over the 1980s—its contribution to total German output ris76 By 1987, the ERDF’s budget had grown from 258 million ECU to more than 3 billion ECU. Even then, however, its redistributive capacity remained limited, representing as it did only 9.1 percent of the EC’s total spending for that year (Marks 1992, 194). A major weakness in the EMS, the absence of an effective U.S.-like stabilization mechanism is also a concern for proponents of EMU. See Eichengreen 1994 for an in-depth discussion of this issue. 77 Artis 1987, 186.
W I N N E R S
A N D
L O S E R S
I N
T H E
205
E M S
Table 8-4 Cumulative Change in Bilateral Real Exchange Rates of Italy, France, and West Germany, 1979–1990 Italy (%) Germany (19.0)a France (14.3) Netherlands (4.1) Belgium (3.4) Denmark (0.8) Ireland (0.3)
France (%) 42.0 31.1 47.8 48.8 34.8 18.8
Germany (16.8) Italy (10.5) Belgium (9.9) Netherlands (5.3) Denmark (0.8) Ireland (0.4)
West Germany (%) 10.9 231.1 17.7 16.6 3.7 212.4
France (12.2) Netherlands (9.9) Belgium (8.3) Italy (6.8) Denmark (2.2) Ireland (0.5)
210.9 16.6 6.8 242.0 27.2 223.3
Sources: Fratianni and von Hagen 1992, 29; IMF 1984a. Note: A positive (negative) number indicates a real appreciation (depreciation) of the currency of the country shown in the column heading with respect to the currency of the country shown in the row heading. aFigures
in parentheses are percentage of column country’s total exports in 1978.
ing from 26.4 percent in 1980 to 32.0 percent in 1990—the 1980s saw no such growth in Italy: exports comprised 21.9 percent of Italian GDP at the beginning of the decade and 21.0 percent at the end.78 Would Italy’s export sector have performed any better (or Germany’s any worse) if the EMS had never been inaugurated? While there is of course no way to know for certain, one cannot help but be struck by the difference between the Italian export sector’s rapid rate of growth before the regime entered into force and its lackluster performance thereafter. Prior to 1979, Italian exports had been steadily rising as a share of Italian output, just as German exports had been increasing as a proportion of German output. After 1979, however, only Germany’s export sector continued to increase in size and importance (Figure 8-5). Does this mean the EMS was solely responsible for the economic misfortunes that befell Italy after the lira was introduced into the regime? Not at all. Although Italy’s rate of unemployment did increase sharply over the course of the 1980s, rising from 7.7 percent in 1979 to a high of 12 percent in 1987 and 1988, macroeconomic aggregates like unemployment and inflation are rarely caused by any one factor. In any case, the methodologically appropriate question here is not “Was Italy’s economic performance better after 1979 than before?” but “Was its performance better after 1979 than it would have been if the pre-EMS status quo had never been disrupted?” Although I have suggested a number of reasons for thinking that Italian economic performance would indeed have been better—not only in the short term but also, quite possibly, in the long term as well—the reality is that the 78
OECD 1992.
206
E U R O P E A N
M O N E Y
Figure 8-5. Exports as a Percentage of GDP in Germany and Italy, 1960–1990
pre-EMS status quo was disrupted, making it impossible for me (or, presumably, anyone else) to construct a definitive test. But while it might not be possible to disprove the conventional “positive-sum” account of Italian entry, the evidence considered here does, I think, cast serious doubt upon this interpretation. Would Italy’s “utility” really have been lower if the EMS had never made it off the ground—if, in other words, the non-system of freely floating exchange rates that began to emerge in the early 1970s had continued without interruption through the 1980s and 1990s? Granting that we may never know for certain, there is ample justification for thinking that just the opposite is true. The View from Thatcher’s Britain The same goes for the United Kingdom: it too would have been better off— or so one might plausibly argue—in a “noncooperative” world of floating exchange rates.79 Although Thatcher did (eventually) admit the pound into the ERM, she did so with little enthusiasm, viewing Britain’s participation 79
By concentrating here on Italy and Britain, I do not mean to suggest that they were the only two EMS “losers.” A more complete analysis would also have to include a discussion of Ireland, a country whose EMS-induced disinflation coincided with very high unemployment, massive emigration, and a rising public debt. See, e.g., Bradley and Whelan 1992 and Dornbusch 1989.
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
207
in the EMS—and in “Europe” more generally—as a serious threat to British sovereignty.80 The EMS also ran contrary to the Conservatives’ interest in eliminating government interference from private markets, including the market for foreign exchange. “You cannot buck the market!” is how the prime minister put it when, in March 1988, she publicly rebuked Lawson, her Chancellor of the Exchequer, for “shadowing” the D-mark at approximately £1:DM3—a policy that Lawson had been pursuing without Thatcher’s permission since 1987.81 WAS BRITAIN’S ANTIPATHY TOWARD THE EMS JUSTIFIED?
While Thatcher may have been the most vocal in her opposition to the EMS, the truth is that British authorities had been critical of the regime from its very inception. More than simply a product of “Thatcherism,” their reservations reflected legitimate concerns about the differing structures of the British and West German economies. This mismatch was immediately recognized by Thatcher’s predecessor, Labour Prime Minister Jim Callaghan, who refused to introduce the pound into the ERM for fear that it would force Britain to adhere to Germany’s contractionary policies even when they were not appropriate to economic conditions in the UK.82 Ken Couzen, the Treasury official who was Britain’s delegate to the earliest round of talks on the EMS, left little doubt as to the Labour government’s position on the issue: “There is,” he confessed, “considerable anger at the way in which the proposal was ‘sprung’ on the rest of the Community.” As for the terms of the “Franco-German scheme,” they were “vague and often confused,” whence came the British government’s “deep suspicion that the system is little more than a means of holding down the D-mark and imposing restrictive policies on Germany’s partners.”83 Couzen voiced these “suspicions” in 1978. One year later, British oil companies began extracting oil from the North Sea. With the prospect of huge windfalls from the second OPEC shock, the market’s demand for assets denominated in sterling grew rapidly despite Britain’s underlying economic problems. The pound thus took on many of the attributes of a petro-currency, rising sharply in value against the currencies of Britain’s European partners and depressing external demand for Britain’s non-oil-related ex80 As deep-seated as it was, Thatcher’s antipathy toward the EMS did not stop her from signing the Single European Act in 1986. In so doing, Thatcher agreed to accept qualified-majority rule in the European Council, albeit only on matters pertaining to the operation of the Internal Market. For analysis of this earlier decision, see D. R. Cameron 1992b; Garrett 1992; and Moravcsik 1991. 81 Notwithstanding the prime minister’s personal crusade on behalf of market forces, the British economy remained highly regulated in certain areas. For a discussion of Thatcher’s (failed) attempt to reform Britain’s National Health System, see Pierson 1995. 82 Dennis and Nellis 1986, 328; Jenkins 1989, 333– 34; Moravcsik 1998, 279– 81; Oatley 1997, 57–60; D. Smith 1993, 45– 46. 83 Quoted in Ludlow 1982, 112–13.
208
E U R O P E A N
M O N E Y
Figure 8-6. A Comparison of Growth Rates in Germany and the United Kingdom, 1978 –1990
ports.84 By driving the British and German economies even further out of alignment, these developments lent support to the argument, made by Labourites and Tories alike, that British entry into the ERM would serve merely to exacerbate the continuing “deindustrialization” of the British economy. By the mid-1980s, the price of oil had dropped by half, eliminating what had earlier been an idiosyncratic source of upward pressure on the pound. Even then, however, economic conditions in the United Kingdom continued to be out of sync with those on the continent (Figure 8-6). In 1980, Britain’s economy contracted by 2.2 percent while Germany’s grew by 1 percent. The next few years saw these trends reverse themselves: in 1982, for instance, the British economy grew by 1.7 percent while Germany’s contracted by 1.1 percent. Though both countries enjoyed positive growth between 1985 and 1988, Britain’s rate of growth exceeded the German rate by as much as 3.3 percentage points. Finally, in 1989 Britain began slipping into a deep and prolonged recession at a time when the German Bundesbank, in an effort to combat the inflationary problems created by German unification, was furiously raising interest rates—as, by necessity, were the central banks of Germany’s ERM partners. The central message of Figure 8-6 is that while growth in ERM-participating countries approximated growth in West Germany, the British and German economies spent the 1980s following very different trajectories. The same lesson can be gleaned from the correlations presented in Table 8-5. Legitimate as they were in the late 1970s, British concerns about the imposition of German economic priorities were, it seems, even more firmly 84
Alt 1985.
W I N N E R S
A N D
L O S E R S
I N
T H E
209
E M S
Table 8-5 The Symmetry of Business Cycles in the European Community, 1965–1990
Country
Correlation a
Germany Netherlands France Belgium [Austria]b Denmark Greece Spain Italy Portugal United Kingdom [Finland] [Sweden] Ireland
Symmetry Rank (Rank for 1965–79 period in parentheses)
— 0.76 0.68 0.67 0.66 0.60 0.58 0.55 0.55 0.48 0.44 0.41 0.39 0.35
1 2 3 4 5 6 7 8 9 10 11 12 13 14
(1) (4) (3) (5) (8) (2) (6) (11) (12) (10) (7) (9) (13) (14)
Sources: Garrett 1998, table 3; OECD, Historical Statistics, various years; and OECD, Economic Outlook, vol. 59 (1996), A4. aCorrelations between annual GDP growth rates in Germany and other European countries. bCountries in brackets were not members of the EC during the time period in question.
grounded in 1990, the year Thatcher finally agreed to introduce the pound into the ERM. THE EMS CONTROVERSY WITHIN THE BRITISH GOVERNMENT
Over and above its deleterious economic effects and symbolic associations, the birth of the EMS also precipitated years of internal bickering and controversy within the UK’s Conservative movement. While the question of whether Britain should actively engage itself in European affairs had long been a contentious issue among British Conservatives, the inception of the EMS thrust the Tories’ internal divisions into the public eye. For among those pushing for immediate entry was Thatcher’s own Chancellor of the Exchequer, Nigel Lawson, who had in fact been urging Britain to become a fullfledged member of the EMS since 1985.85 Thatcher herself, however, only belatedly accepted—and never fully embraced86 —the logic of Lawson’s po85 86
Lawson 1992, 111. See, e.g., Thatcher 1993, chap. 24.
210
E U R O P E A N
M O N E Y
sition. To the contrary, Thatcher remained dead set against the EMS until some time shortly before the Madrid Economic Council summit on June 26, 1989. At the Madrid meetings, the prime minister spelled out a set of conditions under which her government would at last be prepared to introduce the pound into the exchange rate mechanism. Even after she had resolved to take the pound into the ERM, however, Thatcher took a number of steps to ensure that the move not appear as the capitulation it was. Instructing Lawson to keep his European counterparts guessing as to the exact date of British entry was one such ploy, but neither that nor anything else the prime minister tried could dispel the impression that she had caved in to the “wets” in her government. Although the Conservatives’ political future appeared in doubt only a short time before the election took place, the party successfully eked out its fourth straight majority in 1992. Thatcher herself was not around to enjoy the occasion, however, having been ousted as prime minister at a Conservative party conference held less than a month after the ERM decision.87 Precisely how great a role the controversy surrounding Britain’s belated entry into the ERM played in Thatcher’s personal downfall is difficult to determine. Other factors also contributed to her political demise—for example, her somewhat bizarre intransigence in the face of mounting opposition to the “poll tax,” a highly regressive levy introduced in the late 1980s that many British citizens considered grossly unfair. Even if it played only a minor role in Thatcher’s ouster, however, the controversy surrounding the EMS is widely (and, I think, correctly) blamed for undermining the political fortunes of her government and, by implication, her party. Much of this was due to the controversy’s impact on public perceptions: the fissures opened up by the EMS decision gave the impression of a government in disarray. Throughout the 1980s, media reports of protracted and often acrimonious debates between ministers were frequent, as were cabinet reshufflings. The most important of these came in the summer of 1989, a month after the Madrid summit, when Thatcher demoted the two members of her cabinet who had been most outspoken in support of British entry, Lawson and Foreign Secretary Geoffrey Howe. Howe was asked to serve as deputy prime minister and leader of the House of Commons, a largely ceremonial post. A few months later, Lawson resigned after a dispute over who controlled monetary policy, he or the prime minister’s economic adviser, Sir Alan Walters, an outspoken critic of the EMS.88 By the time the pound officially entered the system in October 1990, pub87 The first round of voting at the Conservative party’s annual conference was held on November 20, 1990. On the first ballot, Thatcher came up four votes shy of the number needed to ensure her continued possession of the party leadership. Rather than put her name forward on the second ballot (held one week later), she withdrew from the race. 88 John Major served briefly as Lawson’s successor until winning the party leadership and being installed as the new prime minister.
W I N N E R S
A N D
L O S E R S
I N
T H E
E M S
211
lic opinion polls were giving the opposition Labour party a ten-point lead over the Tories. One month later, the Conservative party’s political standing received a near-fatal blow when, speaking before the House of Commons, Howe delivered a resignation speech that riveted the country’s attention.89 Sharply critical of Thatcher’s European policy, Howe declared at one point: “It is now, alas, impossible to resist the conclusion that today’s higher rates of inflation could well have been avoided had the question of ERM membership been properly considered and resolved at a much earlier stage.” These words echoed the sentiment expressed in another well-publicized speech to Parliament, this one delivered just a few weeks earlier by Thatcher’s erstwhile Chancellor of the Exchequer. In Lawson’s words, “The real tragedy is that we did not join the ERM at least five years ago. It was not for want of trying, as a number of my then colleagues can testify.”90 While Lawson and Howe may well have been correct that Britain was better off joining the ERM than maintaining its earlier policy of monetary isolation, the months following British entry were certainly no bed of roses for Thatcher’s successor, John Major. Prior to reasserting control over the pound in 1992, Major’s tenure was characterized by abysmal public approval ratings. Not all of the blame rests with the EMS, of course; the British economy was mired in recession throughout much of this period. Nevertheless, Germany’s restrictive monetary policies, combined with EMS-induced deflation in France and Italy, imposed an economic straitjacket on Britain, precluding its newly elected government from cutting interest rates as drastically as the country’s recessionary conditions warranted.
Conclusion Widely regarded as a paradigmatic case of states working together to achieve mutual gains, the EMS qualifies as a “hard case” for the rather less sanguine, winners-and-losers perspective on international cooperation that I have been developing in this book. Be that as it may, the evidence presented in this chapter raises a strong possibility that at least two EMS entrants—Italy and the United Kingdom—would have preferred the original, noncooperative status quo, and thus would never have joined (at least not voluntarily) had it remained a viable alternative. The case of Italy is particularly problematic for proponents of the standard, positive-sum account of European monetary integration, for Italy’s industrial relations system was more ossified than those of its EMS partners, its sensitivity to idiosyncratic, country-specific shocks was greater, and in times of stress it lacked adequate substitutes for devaluation. Because Italy was not part of an “optimum currency area” with France and West Germany, 89 90
Howe 1990. Lawson 1990.
212
E U R O P E A N
M O N E Y
the country’s 1978 decision to join a quasi-fixed exchange rate regime centered around the German mark was widely expected to exacerbate unemployment among blue-collar workers and other key constituencies of the Italian government. And yet, insofar as the financial markets were convinced that future inflation would be lower in countries that had fixed to the deutsche mark than in those that had not, and insofar as few Italian citizens wished to sever their country’s European connection, Italy’s government could ill afford not to join the new regime. Unemployment did rise after Italy joined the EMS, but there is good reason to think that the ensuing dislocations—both economic and political—would have been even worse had Italy opted out of the system. The story for Britain is much the same. If British authorities had had the power to prevent the Franco-German monetary regime from taking shape in the late 1970s, they most certainly would have done so. By the same token, few tears would have been shed in the Thatcher cabinet if President Mitterrand had gone ahead and floated the franc in the early 1980s, thus effectively abolishing the currency link his conservative predecessor had helped to establish in 1979. Finally, even during the short time that Britain participated in the system, John Major, though less stridently anti-European than Thatcher, was adamant that Britain’s abdication of monetary policy be seen as a defensive maneuver—a concession to economic realities rather than a product of unconstrained choice.
C H A P T E R
9
Rigging the System: Why Did the EMS Take the Form It Did?
The previous chapter asked why the major countries of Western Europe— with the exception, until 1990, of the United Kingdom—agreed to coordinate their monetary and exchange rate policies throughout the turbulent years of the 1980s and early 1990s. There my purpose was to show how the motivations at work did not always fit comfortably within the standard positive-sum framework of contemporary international relations theory. In this chapter, I move beyond the question of why European states cooperated to explore the related, albeit analytically distinct, question of how they cooperated. Of particular concern are the issues concerning formal institutions and delegation discussed in chapters 4 and 5. At the time the EMS was taking shape in the late 1970s, Giscard and Schmidt, the regime’s two prime movers, were fully prepared to “go it alone” (i.e., to establish a FrancoGerman regime of their own) should other European countries refuse to take part. That being the case, the two leaders enjoyed a great deal of discretion in structuring the monetary arrangement that was to become the European Monetary System. Had they wished, for instance, they could have made the EMS as restrictive and inflexible—and thus as tough on inflation—as the earlier Snake regime had been. All it would have taken was the insertion of a provision into the EMS charter prohibiting signatories from altering their exchange rate parities once they had entered the system. Despite the “virtues” of this sort of arrangement—its congruence with Giscard’s and Schmidt’s shared interest in reducing France’s double-digit rates of inflation as quickly as possible—the charter that came out of the EMS negotiations called for a much looser, more flexible institutional structure. What can account for this? Why, in short, did the regime’s French and German architects make the institutional choices they did? There is, I shall argue, a perfectly straightforward explanation: Though they themselves would have preferred a tighter (hence less inflationary) regime, Giscard and Schmidt had to take into account the fact that (1) not everyone shared their preferences and (2) among those who did not were members of the French Socialist party, a collection of EMS “losers” whose leader, François Mitterrand, stood a good chance of becoming France’s next president. By endowing the EMS with a more fluid institutional structure than they would normally have preferred, Giscard and Schmidt increased the likelihood of their new monetary regime’s surviving such an oc-
214
E U R O P E A N
M O N E Y
currence: better a loose EMS with (political) staying power than a tight EMS with none. As loose as it was, however, the EMS contract was not entirely forgiving. While it is true that the EMS could accommodate ex post changes in the ex ante configuration of exchange rate parities, the regime’s institutional superstructure ensured that requests for such changes would be relatively few and far between. Why? Because the French and German officials who drafted the EMS charter made all such requests subject to the approval of a supranational forum—the European Monetary Committee—whose membership included a representative from Germany’s hard-line central bank. That the regime’s French and German prime movers would choose to delegate this crucially important gatekeeper task to a supranational collective decision-making body may seem puzzling, for it was not their only option. Another possibility would have been to let each signatory of the regime decide for itself whether prevailing conditions warranted a change in the existing configuration of ERM parities. Why did Giscard and Schmidt reject this option? Here, too, there is a perfectly straightforward explanation. As we will see, however, it is not the explanation that other analysts have put forward. Whereas most studies emphasize the efficiency of supranational governance, the evidence presented below suggests that this emphasis may not in fact be any more justified in the case of European monetary institutions than it was in the case of North American trade institutions (see esp. chapter 6)— though, to be sure, exposing the shortcomings in the efficiency argument will take a good bit of work. Before getting started, I should note that the question at issue in this chapter—whether the EMS regime’s distinctive institutional structure is best explained by efficiency considerations or, as I argue below, by the fact that the winners (who designed the regime) had a stake in co-opting certain pivotal losers—has implications that go beyond the EMS per se. Ideally, for example, the kind of power-politics analysis I offer in this chapter could help us make sense of the otherwise quite perplexing institutional choices made by the founders of EMU, the new monetary union whose institutional centerpiece— a new Frankfurt-based European Central Bank (ECB)—began regulating the supply of “euros,” and hence fixing European interest rates, in January 1999. Understanding why the ECB took the form it did is a question ripe for further study.1 In any event, the EMS experience discussed below, though it predates EMU by a good twenty years, raises what turns out to be a very similar explanatory puzzle. This is not to say that the EMS and EMU are analytically equivalent. In contrast with the Treaty on European Union (which mandated the creation of a single European currency), the founding charter of the EMS allowed each signatory to retain at least some degree of 1 See the epilogue at the end of this chapter for my own, admittedly somewhat speculative explanation.
R I G G I N G
T H E
S Y S T E M
215
national monetary independence. Specifically, currencies participating in the system’s exchange rate mechanism were permitted to fluctuate within a certain range—either 2!f or 6 percent—of their bilateral ERM exchange rates. On the other hand—and this is the important point—these ERM exchange rates, or central parities, were themselves matters for collective decision making, since EMS member countries could alter their individual parities (a possibility explicitly allowed for under the terms of the arrangement) only at the discretion of a “higher” collective decision-making body: the aforementioned European Monetary Committee. Though the euro’s entry into circulation effected a further upward shift in the locus of European monetary authority—from the European Monetary Committee to a full-fledged European central bank—the launching of the EMS in the late 1970s was, in fact, the decisive turning point. As the economist Niels Thygesen has observed, the EMS moved Europe “into the largely uncharted waters of discretionary collective management of exchange rates. . . . It [was] sui generis in its reliance on the ability of the participants to agree on realignments acceptable to both the initiating country and its partners.”2 Granting, then, that this earlier, pre-EMU transformation in the “management” of European exchange rates is worthy of study in its own right, how are we to account for it? In what follows I offer two quite different explanations. It is to the first of these—what I term the efficiency explanation—that we now turn.3
The Efficiency Argument If there is a single theme running through the rational choice branch of the “new institutionalism,” it is that the development of appropriate institutional structures may be a necessary precondition for the emergence of cooperation even when the prospective gains to each cooperation partner are large and transparent. Following this line of analysis, one might well conclude that the institutions of the EMS were efficient. That is, their existence and, more importantly, their particular form made it possible for signatories of the regime to enjoy the fruits of monetary cooperation (e.g., exchange rate 2
Thygesen 1984, 272. Much of the terrain I will be covering in this chapter is new, for, as voluminous as it is, the literature on European monetary unification has thus far paid relatively little attention to questions concerning institutional design and (especially) variation. This is particularly true of research on Europe’s transition from the Snake regime of the early 1970s to the European Monetary System of the 1980s; indeed, a number of analysts dispute that the institutional structures of these two monetary arrangements differed in any meaningful way, except perhaps for the fact that the EMS afforded the Italian lira wider margins than it had enjoyed under the Snake (Frieden 1994, 27; McNamara 1998, 32; see also Oatley 1997). What these accounts overlook—and it is not a trivial oversight—is the fact that parity realignments under the EMS were (1) possible and (2) subject to the consent of a supranational decision-making body, neither of which was the case under the Snake. 3
216
E U R O P E A N
M O N E Y
stability and low inflation), which had previously eluded the grasp of all but a handful of European countries. In this view, the institutional structure of the EMS solved a cooperation problem, facilitating utility gains for all involved. But what exactly was that problem? There is, it turns out, an enormous literature on the problems—for there are often more than one—that arise when monetary authorities in two (or more) countries attempt to coordinate their exchange rates. Most of these studies begin with the simple4 assumption that exchange rate cooperation is a good thing. Of course, the fact that it is good thing—that its emergence would be Pareto-improving—does not mean it will happen. Why not? Obstacles to Decentralized Cooperation Consider France and Germany. If the analysis I offered in chapter 8 is correct, sometime around 1977, Valéry Giscard d’Estaing, who was at that time presiding over the French government, and Helmut Schmidt, his counterpart in Germany, developed a mutual interest in stabilizing the franc-mark exchange rate. The fact that both leaders stood to benefit from a Franco-German exchange rate agreement did not mean that such an agreement would be struck, however, or, if struck, that it would necessarily be adhered to. Looking at the situation from the standpoint of France, some analysts have suggested that Giscard’s optimal strategy would have been to uphold such an agreement until—but only until—the inflationary expectations of French workers had begun to adjust to (what they perceived as) the new “franc fort” reality.5 Under standard assumptions (e.g., nominal wage contracts could not easily be renegotiated in the short term),6 a franc depreciation at that point would have permitted Giscard’s supporters to enjoy the benefits associated with an undervalued currency—increased demand for French exports, faster output growth, and the like—without at the same time having to endure the higher rates of wage inflation normally produced by a depreciating currency. In reality, of course, the relationship between governments and private wage setters is a continuous process, not a discrete, single-event interaction. Even if one allows that “surprise” depreciations do sometimes generate short-term advantages, the long-term costs would also need to be factored into the equation. Would these costs have been large enough to keep Giscard from violating France’s exchange rate commitments with Germany? The conventional answer is most assuredly yes. By breaking his promise with Schmidt, the theory goes, Giscard would have signaled to the French public that his future pronouncements, whether on economic policy or any other matter, were not to be trusted. Making matters worse, a surprise de4
Albeit not, in my view, entirely warranted; see chapter 8. See, e.g., Melitz 1988. 6 Bruno and Sachs 1985. 5
R I G G I N G
T H E
S Y S T E M
217
preciation in France would have had the effect of exacerbating inflationary pressures in Germany, thus prompting a retaliatory response from authorities at the Bundesbank (with whom Giscard was also involved in a “repeated game”). By provoking a Bundesbank-engineered economic slowdown in Germany, Giscard’s defection would have ended up dampening, not stimulating, foreign demand for French exports. Following this logic, one might conclude that exchange rate coordination among (sufficiently farsighted) European governments was not really so difficult to achieve after all. Given that beggar-thy-neighbor exchange rate policies such as those just described offered no lasting advantage to the wouldbe defector (i.e., France)—and they certainly did not benefit the exploited party (i.e., Germany)—a Franco-German exchange rate agreement would have been “self-enforcing.”7 Note, however, that by this logic a simple treaty would have sufficed. Why, then, did the French and German architects of the EMS go to the trouble of establishing a full-blown supranational governance structure? Supranational Governance as a Solution to Incomplete Contracting Problems Here, once again, proponents of the “new institutionalism” would invoke the iterated prisoners’ dilemma game discussed in chapter 4.8 Recall that for cooperation to emerge within the context of an iterated PD, each player has to believe that a defection in the current round of the game will be met with retaliation in some future round (or rounds). In practice, however, this condition is likely to hold only insofar as the players are familiar with the histories of their current partners—specifically, whether or not their partners have kept their promises in previous plays of the game—for it is only in light of such information that the appropriate strategy for each player (given the rules set forth, formally or informally, at the outset of the game) can be determined. If player A is confronted with a partner B whose past record is uncertain, there is a greater chance that A will adopt the “wrong” strategy: cooperating when it ought to be retaliating, retaliating when it ought to be cooperating. So where does that leave us? For the sake of argument, let us suppose that each of the European leaders who participated in the EMS negotiations—not just A and B (e.g., Gis7 A well-known study of economic summitry among the G-7 countries relies on similar logic. Members of the G-7 would have been foolish to renege on an agreement concluded by the body, the authors of the study contend, given the “visibility and solemnity of the commitment, its verifiability, the frequency of dealing among the contracting parties, the value that they attribute to those other deals, and the number of other parties ‘observing’ the game.” Putnam and Henning 1989, 102. 8 For applications of the theory of repeated games to issues concerning monetary and exchange rate policy, see Kydland and Prescott 1977; Barro and Gordon 1983. A comprehensive overview of this literature is provided in Blackburn and Christensen 1989.
218
E U R O P E A N
M O N E Y
card and Schmidt) but also C (e.g., Andreotti), D (e.g., Callaghan), E, F, G, and so on—had been concerned about satisfying the informational requirements for sustained collective action. How might these concerns have been reflected in the kinds of institutional structures these leaders (collectively) agreed upon? THE EMS AS AN INFORMATION CLEARINGHOUSE?
The first consideration to note here has to do with the monitoring and publicizing of ex post defections. Even before the inauguration of the EMS, information about daily fluctuations in the franc-mark exchange rate was both readily available and (for obvious reasons) subjected to intense scrutiny by financial analysts. This is not to say that informational impediments to European monetary cooperation were unimportant or irrelevant. Not seeing any need for an elaborate new information clearinghouse, the EMS negotiators did not create one. But while rule violations may have been easy to detect, the rules themselves were inherently ambiguous. Why did this matter? The answer given in the literature flows directly out of the point just mentioned—namely, that successful collective action cannot occur unless each “player” has the capacity to distinguish rule breakers from rule followers. Needless to say, this capacity is contingent on the rules themselves being clear. Otherwise, transgressors would have an easy time passing themselves off as good-faith cooperators, and good-faith cooperators would be viewed, at least some of the time, as transgressors. Pretty soon, every player, regardless of its “true” type, would conclude that it was better to transgress (i.e., knowingly violate what the player itself understood to be the rules of the game) than to cooperate (i.e., comply with these rules and risk being “suckered” by another player who interpreted them differently). THE ADVANTAGES OF FORMALIZATION
By this logic, all of the European governments that participated in the EMS negotiations had an interest in prespecifying, as clearly and precisely as possible, the standards by which their future behavior would be judged. Suppose that the new exchange rate agreement had been purely informal. What would have happened if one of the parties to the agreement had decided to allow its currency to fall against those of other participating countries? Would the latter have been justified in retaliating? What if the (ostensible) transgressor’s economy had been suffering from the effects of an exogenous shock produced by, say, a sudden rise in the price of foreign oil? Should other parties to the agreement have been willing to make an exception in this case? If so, what was there to prevent these other participants (including those whose economies had not been as hard hit by the shock) from depreciating their own currencies, justifying their actions by reference to the “spirit” of the original agreement? The clearer the initial rules of the game, the less likely were these kinds of issues to arise in the first place. Drafting the EMS treaty—officially, the Res-
R I G G I N G
T H E
S Y S T E M
219
olution of the European Council on the Establishment of the European Monetary System and Related Matters—was itself, in this view, a kind of “cooperative device.”9 By taking time at the start to formalize the terms of a mutually acceptable code of monetary “ethics,” the architects of the EMS made it possible for participants in the system to tell if their partners were behaving improperly. Formalization was thus a necessary first step toward the realization of collective gains. Though necessary, however, it was not sufficient, for the EMS charter was sure to be “incomplete.”10 Without a well-developed body of rules for dealing with unforeseen contingencies and special circumstances not specifically covered by the treaty—the onset of a recession in one member country but not in the others, for example—there would be nothing to stop each EMS signatory from interpreting the treaty differently, creating a crisis that could lead to a breakdown of the entire system. DETERMINING RESIDUAL RIGHTS OF CONTROL
To address this problem, the founders of the EMS could have stipulated that all disputes concerning matters of treaty interpretation be settled through open-ended intergovernmental negotiations. This is not, however, what the founders did. Instead, they specifically required that all such disputes be adjudicated by “a common procedure” (Article 3.2). The effect of this provision was to take these disagreements out of the hands of individual member governments and transfer them to a “higher” collective decision-making body. Which body? The one the founders had in mind was the European Monetary Committee, a supranational panel made up of the deputy governors from the central banks of each ERM signatory, senior representatives from member countries’ finance ministries, and two representatives of the European Commission. Once the EMS charter came into force, this panel assumed responsibility for determining whether prevailing economic conditions warranted a readjustment of exchange rates and, if so, which EMS signatories would be permitted to devalue and by how much. Technically, of course, an EMS member state whose request for a devaluation was denied by the European Monetary Committee could go ahead and devalue anyway; the committee’s decisions were authoritative only insofar as national governments chose to honor them. Refusing to comply, however, would have meant exiting the system and thus forgoing any benefits of participation or (of greater salience to governing officials in Italy and the United Kingdom) incurring the costs of exclusion. Either way, it was an extremely risky move—one that, right up until the system’s near collapse in 1993, member governments were strongly inclined to avoid if at all possible. Thus, while the committee’s determinations might not have been as binding as rulings issued by a domestic court, each of the system’s member states nonethe9
Fratianni and von Hagen 1992, 129. Cf. Lohmann 1992.
10
220
E U R O P E A N
M O N E Y
less had powerful incentives to comply. And comply they did: though the adjustment of central rates did not immediately result from truly collective decisions—two of the first three realignments occurred without any prior consultation with the committee11 —by the early 1980s it had already begun to play the “governance” role that its founders had envisaged.12 The fact that EMS signatories were willing to adhere to a predesignated set of collective decision-making rules and procedures has been described as “a revolutionary development,” one that “touches at the very heart of monetary sovereignty.”13 What inspired this dramatic departure from past practice? Now that I have provided a little background, the efficiency rationale for this “revolutionary” change should be clear. The actors who designed the EMS anticipated that a simple exchange rate agreement would not be fine grained enough to cover all possible contingencies and that, in all likelihood, signatories of the regime would use the resulting ambiguities—particularly during periods of crisis and instability—as a pretext for driving down the value of their currencies. To prevent this sort of thing from happening, the “principals” who drafted the treaty empowered a supranational “agent” to determine when, and by how much, member governments would be permitted to reformulate their exchange rate targets. Had it not been for the expectations-clarifying role of the European Monetary Committee, many of these governments would have taken every opportunity to free ride (knowing they could do so without patently violating the letter of the original treaty) and, wracked by compliance problems, the EMS would have met an early death. Weaknesses in the Efficiency Argument There are two major flaws in this account. The first is that it views the designing of the EMS as a collective endeavor when, in reality, it was dictated by two individuals—the president of France and the chancellor of Germany. True, the fact that Giscard and Schmidt did not feel the need to reconcile their institutional preferences with those of their European counterparts would not have mattered if their counterparts had held similar preferences. But in Italy and the United Kingdom, at least, their counterparts did not hold similar preferences. As the European leaders who stood to benefit the most from a stable regime, Giscard and Schmidt wanted to endow “their” EMS with the most efficient set of institutional arrangements possible (but see below). By contrast, governing elites in Italy and the UK would have been only too pleased if the new Franco-German scheme had been structured in 11 In November 1979, Danish representatives simply let it be known to a member of the committee that their government had decided to devalue the krone by 5 percent. In March 1981, the Italian coalition government followed suit, devaluing the lira by 6 percent. Here, too, consultations with the chairman of the committee were conducted after the fact. 12 Padoa-Schioppa 1986, 65 – 66. 13 Tsoukalis 1989, 63.
R I G G I N G
T H E
S Y S T E M
221
a way that made it highly vulnerable to ex post reneging, for the less stable the new monetary arrangement, the sooner would Europe revert to the original, pre-EMS (i.e., floating-rate) status quo that both the Italians and British really preferred. Granting that Giscard’s and Schmidt’s institutional preferences counted for a lot more than those of, say, the prime ministers of Italy or Britain, how well does the efficiency analysis do in explaining the former? Were the two EMS enactors as intent on lowering the informational impediments to successful collective action as the above account implies? Or might their institutional choices have been driven by other considerations altogether? Though the importance of overcoming information-related free-rider problems is what gets pride of place in the new institutionalist literature, the EMS presents a case in which none of the major players (that is to say, neither Giscard nor Schmidt) was particularly concerned about being doublecrossed. In fact—and this is the second big weakness in the efficiency argument—the French and German actors who configured the regime were not the least bit tempted to defect from their initial agreement. Why not? Because they both knew that any short-term boost generated by an unlawful, “surprise” depreciation would be followed by an offsetting rise in inflation, and higher inflation was the last thing that either of them needed. Take the German chancellor. With the Bundesbank maintaining its tightfisted control over Germany’s money supply, it would have been pointless for Schmidt even to try to deviate from the path of low inflation. For thus provoked, the Bundesbank would have been only too quick to raise German interest rates, as, indeed, it had done in 1973, the last time a German government had tried to enact a large fiscal stimulus.14 As for Giscard, his free-rider incentives were only slightly greater—he was, after all, a conservative. It is true that in 1976 his administration had been moved to withdraw the franc from the Snake, the forerunner to the EMS in which France had been (intermittently) participating since 1972.15 By the end of 1976, however, Giscard had come to appreciate the limitations of franc depreciation as a strategy for stimulating economic growth.16 As long as French workers remained unwilling to moderate their wage demands—a safe bet given the militancy of France’s labor movement and anti-labor orientation of the government—a continuously depreciating franc would have exacerbated the very inflation problem that Giscard and his newly appointed prime minister, Raymond Barre, had been working so urgently to redress. For barring a spontaneous “outburst” of wage restraint, a depreciation of the franc would have encouraged producers for France’s domestic market, now effectively shielded from foreign competition, to raise the prices they charged domestic consumers, in this way nullifying the otherwise desirable effects of a cheaper currency. 14
See chapter 8. France first dropped out of the Snake in 1974, a year and a half after the regime’s inception. It then rejoined in July of 1975, only to withdraw again nine months later. 16 De Boissieu and Pisani-Ferry 1998. 15
222
E U R O P E A N
M O N E Y
Given how little there was for France’s conservative government to gain— either economically or politically—by engaging in competitive devaluations, Schmidt did not need to worry that Giscard would use the inherent ambiguities in their treaty as an excuse for behaving opportunistically. Nor, as we have seen, would Schmidt have faced a strong temptation to defect from the treaty by opportunistically driving down the value of the mark. So why, then, couldn’t the regime’s French and German prime movers have drafted a “complete contract,” that is, one explicitly disallowing all ex post devaluations? One possibility is that Giscard and Schmidt did not think the flexibility afforded by the looser institutional framework they devised would generate any significant economic or political costs. There is no evidence of their having held this view, however, and indeed it would have been quite surprising if they had. For by allowing periodic realignments in the franc-mark exchange rate, Giscard and Schmidt undermined one of their new regime’s primary raisons d’être, which was to serve as a kind of commitment device, bolstering the credibility of the Barre Plan and thereby (it was hoped) accelerating the French economy’s ongoing process of stabilization. The decision to allow devaluations and resulting “realignment uncertainty” had the further drawback of fueling unwarranted speculation and instability within the financial markets.17 In short, there are reasons to doubt that efficiency considerations are what led the regime’s Franco-German enacting coalition to prefer a more openended institutional arrangement. As for this coalition’s decision to channel all devaluation requests to a European-level collective decision-making body, this, too, poses problems for the efficiency-based account. Rather than erecting a full-blown supranational governance structure, Giscard and Schmidt could have cut down on transaction costs by allowing German monetary authorities to serve as the sole arbiter of all such requests. Had the German Bundesbank been permitted to play this “hegemonic” role, the efficiency gains might well have been substantial.18 These potential gains notwithstanding, however, the architects of the EMS opted for a less demanding, more consensus-based arrangement. The question, of course, is why.
A Power-Politics Perspective If it is true, as I have just been arguing, that neither Giscard nor Schmidt had reason to fear being “suckered” or “double-crossed” by the other, how, then, can we explain their institutional choices? If the institutional structure they 17
Fratianni and von Hagen 1992, 146– 53. This last point is often cited as an argument for EMU; see, e.g., Padoa-Schioppa 1986. By removing the very possibility of intra-European devaluations, a monetary union reduces the risk of speculative attacks such as those that ignited the 1992–93 crisis. 18 Cf. Eichengreen 1989.
R I G G I N G
T H E
S Y S T E M
223
devised was not intended to mitigate free-rider problems, what was it intended for? In this section I offer a simple answer that flows directly out my earlier inquiry into the “power politics” of institutional design (see esp. chapter 5). What most concerned the EMS’s French and German architects, I will argue, was the possibility of their regime’s one day being destroyed by a newly empowered EMS loser. My discussion begins by identifying the French left as the enemy of the EMS most likely to bring about its demise. It was not, in other words, an external enemy—an Italy or a UK, for example—that most concerned the regime’s Franco-German enacting coalition. It was an internal enemy—namely, France’s newly resurgent Socialist party. Having identified the Socialists as the EMS loser whose potential to wreak havoc was of greatest concern to the regime’s French and German enactors, I then show how these concerns can help us make sense of the latters’ otherwise perplexing institutional choices. Had it not been for the growing threat posed by the Socialists, Giscard and Schmidt would never have preferred, let alone created, the open-ended and inclusive monetary structure with which, in 1978, they endowed the EMS. To claim (as I do here) that the real purpose of this structure was to discourage potential enemies of the regime from actively working to bring about its demise is not the same as saying that the institutions of the EMS were designed to discourage opportunism. As we will see, the task of counteracting a newly installed Socialist government’s “natural” propensity to abandon the EMS presented not just a different but, indeed, a more difficult challenge than that of circumscribing the behavior of a potential free rider. A prospective free rider knows that if it behaves opportunistically, it risks forfeiting a future stream of cooperative benefits. Under standard assumptions, this knowledge may itself be sufficient to keep would-be opportunists from shirking their responsibilities. Yet from the perspective of an EMS loser like the French Socialists, there were no cooperative benefits to be forfeited. The task of altering the Socialists’ behavior was thus analytically distinct from—and more demanding than—that of discouraging a prospective free rider. More demanding, yes; impossible, no. There were still steps, a number of which are discussed below, that the French and German officials who formulated the terms of the EMS charter could take to diminish the destructive zeal of their new regime’s would-be annihilators. Of these ex ante measures, the most important was the drafting of a deliberately incomplete treaty. While the French and German architects of the new regime could have explicitly precluded member countries from realigning their central parities, in this way bolstering Giscard’s credibility as an inflation fighter and expediting France’s transition from high to low inflation, this is not what they chose to do. Instead, they implemented a more flexible—albeit less efficient—institutional structure specifically engineered to reduce (without going so far as to eliminate) the burden of EMS compliance for all-out opponents of the regime like the French Socialists. Were these opponents to assume office in
224
E U R O P E A N
M O N E Y
the near future, the hope was that this new “softer” institutional structure would reduce their distaste for the regime, thus in turn diminishing their eagerness to destroy it. Though these EMS opponents would continue to prefer the initial, pre-cooperation status quo—at no point would the Socialists become EMS winners—the new, post-cooperation world would endure nonetheless. Internal Threats to the Enacting Coalition’s Position of Dominance This line of analysis implies that the institutional structure of the EMS would have looked substantially different if the anti-inflation policies that Giscard and Schmidt embraced in 1977 and 1978 had been certain to persist into the 1980s. If these individuals had been confident that their distinctive preferences (see chapter 8) would continue to orient Franco-German economic relations for years to come, they might well have chosen a more efficient set of institutions for “their” EMS. But because they could not be confident of this, Giscard and Schmidt favored the adoption of a more accommodating, albeit less efficient, supranational structure. In response, one might note (correctly) that Chancellor Schmidt’s lowinflation policies were at little risk of being overturned by a future German government. Even if the German chancellor’s own Social Democratic party were to be turned out of office, its coalition partners, the neoliberal Free Democrats, were likely to remain a vital part of any new governing coalition, as were the conservative Christian Democrats. The threat to the Social Democrats thus came from the right—not, as in France, from the left—and as Alesina (1987), Hibbs (1977 and 1987), and others have shown, centrist and right-wing parties tend to place a greater emphasis on low inflation than do parties of the left. Finally, the Bundesbank was firmly committed to keeping inflation in check irrespective of changes in the partisan complexion of the German government. In France, however, Giscard’s low-inflation orientation was anything but secure. Indeed, even before the outbreak of the second oil crisis in 1978–79, many observers were skeptical that the conservative president would be able to fend off his Socialist challenger in the next presidential election. In the short term, Giscard’s supporters could take solace from the results of the National Assembly elections of 1978. Not only did the president’s left-wing opposition suffer an electoral setback, but within Giscard’s parliamentary majority the UDF (Giscard’s own party) gained twenty seats at the expense of the more nationalist RPR. Nevertheless, public support for the French left, and particularly for the Socialists, had increased dramatically over the course of the 1970s. Given the personal popularity of their leader, François Mitterrand (whom Giscard had defeated in 1974 by only the slimmest of margins), and the steady rise in France’s rate of unemployment, the possibility of a Socialist victory in the next presidential election, which was scheduled for 1981, had to be taken seriously indeed.
R I G G I N G
T H E
S Y S T E M
225
Were Mitterrand to succeed in capturing the presidency, his natural inclination would be to end France’s participation in the EMS. True, in so doing he would be forgoing any credibility “bonus” he might have enjoyed by virtue of continuing to link the franc to the D-mark. On the other hand, Mitterrand shared the view of most other Socialist leaders at that time that the costs of obtaining this bonus would far exceed any potential benefits. Nor was this view entirely without foundation. Given the left’s well-known aversion to austerity, it was safe to assume that the “transitional phase” during which inflationary expectations in France converged to those in Germany would extend over several years, if not decades, forcing thousands of French workers out of their jobs (see below). Thus, it is hardly surprising that France’s two left-wing parties, the Communists and the Socialists, failed to support the EMS initiative when it was first proposed in 1977. Had these parties captured a majority of seats in the parliamentary elections held the following year (as pre-election polls predicted they would), it is quite likely that Giscard’s proposal would have been rejected—perhaps resoundingly so—in which case the European Monetary System might never have seen the light of day.19 It was because the French left had the capacity to return Europe to this pre-EMS status quo that Giscard regarded it as such a serious threat. Critics of the EMS could also be found in other countries—not least, as we have seen, in Italy and the United Kingdom—but only in France did opponents of the EMS have the potential to single-handedly bring about its demise: if the French franc had been pulled from the system, the Italian lira and British pound would have been certain to follow. While this was a source of acute concern to middle-class citizens in France who had benefited from Giscard’s stabilization policies, and from the Barre Plan in particular, the possibility of a French pullout was also a concern to Giscard’s German cooperation partner, Helmut Schmidt, who feared that a future return to a system of freely floating exchange rates would cause the deutsche mark to appreciate rapidly against the franc (as well as any other European currency that had formerly pegged to the D-mark), thus effectively raising the price of German-produced goods and services in some of Germany’s leading export markets. Implications for Institutional Design As they set about designing their new cooperative framework, Giscard and Schmidt thus saw their task as one of devising rules and structures that France’s left-wing opposition parties would be hesitant to abandon if, as seemed increasingly likely, they were one day to gain control of the French presidency, the National Assembly, or both. Why not simply revive the Snake? If the left had not been knocking at the door, the idea might well have caught fire. As things stood, however, few observers were confident that a 19
Ludlow 1982, 85.
226
E U R O P E A N
M O N E Y
new left-wing French government would be willing to take the harsh steps necessary to keep the franc from falling through the floor of its assigned Dmark parity—and the Snake strictly prohibited realignments. Moreover, if the newly elected government were to let the franc fall out of alignment, there would be essentially no chance of its rejoining the Snake at a later date (as Giscard’s administration had done in 1975). Once cooperation had broken down, it would be exceedingly difficult to put back together.20 If exchange rate cooperation between France and Germany were to endure, it was clear that it would need to take another, more flexible institutional form. Although there would be some loss in terms of disciplining inflation, a “looser” institutional structure would have the virtue of extending the new arrangement’s lease on life. Hence the view shared by governing officials in both France and Germany—and embraced (not surprisingly) by their counterparts from Italy and the United Kingdom—that “however strict the system might eventually become, flexibility, and more particularly provisions for changes in exchange rates, would have to be written into the arrangement from the beginning.”21 Designed ex ante to discourage a future Socialist government from pulling the franc out of the system ex post, the regime’s innovative “pegged but adjustable” framework stood at least some realistic chance of surviving the coming to power of the French Socialists. Indeed, the evidence suggests that had it not been for their fear of provoking West Germany’s powerful central bank (whose president, Otmar Emminger, and Karl Otto Pöhl, his deputy and eventual successor, were initially inclined toward the status quo), President Giscard and Chancellor Schmidt would have introduced even greater flexibility into the regime than they did.22 Initially, for example, the two architects of the system had proposed requiring strong-currency countries to offer unlimited and automatic credit to any signatory whose currency had reached its lower marginal threshold and thus was in need of immediate financial assistance.23 Yet as 20 Persistent doubts about the left’s commitment to the Snake had been an important reason for France’s ill-fated reentry into the system in June of 1975. By that time, Giscard had already begun contracting the French economy and imposing price controls. But with the French left experiencing an electoral resurgence, investors could not help wondering how much longer Giscard’s low-inflation priorities would continue to orient French economic policy. Things reached a head when the prospect of the Socialists’ scoring further electoral gains in an upcoming round of municipal elections led to a surge in speculation against the franc, prompting Giscard (reluctantly) to pull the franc out of the Snake on March 15. Ludlow 1982, 33, 82. 21 Ludlow 1982, 159; see also Padoa-Schioppa 1986 and Thygesen 1984. 22 Giavazzi and Giovannini 1989, 41– 42; Heisenberg 1999, chap. 3; Kaltenthaler 1998, chap. 3; Marsh 1992, 168; cf. Putnam 1988. The reservations expressed by Emminger and Pöhl stemmed from their fear that the new monetary arrangement would impinge on the Bundesbank’s previously unchallenged institutional autonomy. Not being particularly fond of the EMS to begin with, the officials at the Bundesbank were not especially concerned about the possibility that Mitterrand, were he to become the new president of France, would find himself so constrained by the new monetary rules and structures that he would see no choice but to pull France out of the new system, thus effectively destroying it. 23 In theory, unlimited financing had also been available to help Snake members experiencing localized recessions defend their margins. In practice, however, Snake members had had to
R I G G I N G
T H E
S Y S T E M
227
skeptical Bundesbank officials were quick to point out, these credit rules would have compelled the German central bank to expand the global supply of deutsche marks. Thus, instead of providing for unlimited and automatic credit, the EMS resolution signed at the European Council summit in Brussels imposed strict limits on both the absolute amount of debt that a weak-currency country could accumulate and the length of time that credits could continue to be drawn.24 This was not the only instance of Bundesbank pressure leading to a somewhat more rigid, less accommodating institutional framework than Giscard or Schmidt would have preferred. Also proposed during the months leading up to the Brussels summit was a scheme wherein intramarginal central bank interventions would be triggered by deviations in each member currency’s “ECU-weight.”25 In principle, a depreciation of the franc could have pushed the D-mark past its upper ECU-based marginal fluctuation limit without at the same time pushing the franc beyond its lower limit. Were this to occur, the scheme would have required the Bundesbank alone to take corrective measures—lowering interest rates, intervening directly in the foreign exchange markets, and so on—even if France’s expansionary policies had been responsible for pushing the D-mark over the edge. Already under intense pressure to expand the German economy, Chancellor Schmidt’s social-liberal government was sympathetic to this “maximalist” ECU-based intervention scheme. Although it would have been cumbersome to operate—the initial proposal was riddled with technical problems—it would have had the desirable effect (from an insulation standpoint) of easing France’s adjustment burden. To monetary officials in Germany, however, any proposal that threatened the Bundesbank’s direct control over the German money supply was a cause for alarm. Bundesbank officials favored a “minimalist” arrangement wherein the ECU would have served merely as a numeraire for the calculation of central rates, with no requirements for intramarginal central bank intervention, symmetrical or otherwise. If they did not get their way, Germany’s central bankers stood ready to sabotage the fledgling monetary regime, driving down the value of other member currencies to the point that no amount of financial intervention would be able to keep them from exiting the system. In the end, the final draft of the EMS charter did incorporate a set of resettle any debts they had incurred to the European Monetary Cooperation Fund (the agency that administered the loans) after only one month. 24 Articles 3 and 4 of the EMS charter stipulate that when two members reach their bilateral EMS parity, the one whose currency is in the stronger position is required to loan the other as much as the debtor wishes for up to forty-five days after the end of the month in which the intervention takes place. The debtor can renew its line of credit once this deadline is reached, but only for three additional months or until it uses up its maximum allowable quota, whichever comes first. 25 A currency’s weight is determined by the number of units of that currency in one ECU multiplied by the price of one ECU in terms of that currency. The relative proportions of the different currencies that make up the ECU—a “basket” of all EC currencies—is normally reviewed every five years; otherwise, the composition of the ECU remains constant.
228
E U R O P E A N
M O N E Y
quirements regarding intramarginal interventions. When a particular ECUbased indicator crossed a certain “threshold of divergence,” those countries whose currencies were most out of line were expected to take corrective measures to reduce the disparity. But—in deference to the Bundesbank’s hardline position—there was never more than a “presumption” (Article 3.6) that such measures would be taken. If members refused, the rules required only that the reluctant parties provide an explanation and, if necessary, consult with their partners in the appropriate Community bodies.26 The Bundesbank, then, cast a large shadow over the negotiations. Still, while members of the regime’s enacting coalition would have preferred a more flexible institutional framework than the one favored by Bundesbank officials, their interest in insulating the EMS from the vagaries of French domestic politics did not require them to go the extreme of eliminating their new scheme’s deflationary bias altogether. If that had been their goal, Giscard and Schmidt could have modeled their regime’s institutional structure on that of the more permissive Bretton Woods regime of the 1950s and 1960s. Whereas both arrangements allowed for periodic changes in exchange rates, under the Bretton Woods regime “the decision to change the parity was essentially unilateral.”27 By contrast, EMS signatories were explicitly required to settle all decisions regarding the timing and size of parity realignments collectively, under the auspices of a preexisting supranational structure composed predominantly of European central bankers. By “[moving into] the largely uncharted waters of discretionary collective management of exchange rates,” the architects of the EMS ensured that realignments would be relatively infrequent.28 But while the EMS was thus 26 If the divergence indicator had been fully binding, Bundesbank interventions would have been commonplace, loosening (though not alleviating) the pressure on the system’s inflationprone members to contract their economies. As noted above, however, the divergence indicator was riddled with technical problems, and there was in any case only a “presumption” that intramarginal interventions stimulated by the crossing of a country’s divergence threshold would occur. In practice, the divergence indicator was almost entirely ignored. It was not until the signing of the Basle-Nyborg agreement in 1987 that France and other weak-currency countries were able to get the Bundesbank to engage in intramarginal interventions. Insofar as the Bundesbank did intervene in the foreign exchange markets prior to 1987, it was by purchasing ERM currencies after they had already fallen—or been deliberately pushed (Tsoukalis 1989, 62)—to the floors of their parity fluctuation bands. 27 Giavazzi and Giovannini 1989, 40. In principle, the Bretton Woods system permitted parity realignments only to correct what the International Monetary Fund deemed a “fundamental disequilibrium.” In practice, however, each signatory evaluated its own economy, with the IMF receiving notification of parity changes after they had already occurred. A member seeking a first-time realignment of less than 5 percent was not required to notify the IMF at all. 28 Thygesen 1984, 272; see also Kees 1987. Although seven realignments, many involving multiple currencies, took place during the first four years after the EMS entered into force, only one realignment was implemented during the five and one-half years between early 1987 and the September 1992 crisis. Even during the years when parity adjustments were relatively common, however, the adjustments were never large enough to fully compensate for the inflation differentials between the system’s high- and low-inflation member states. As a result, the disinflationary pressures on France and other inflation-prone EMS signatories remained intense (see below).
R I G G I N G
T H E
S Y S T E M
229
rendered less flexible than it might have been, Giscard and Schmidt did not need to completely “emasculate” their regime in order to shield it from would-be destroyers like the French Socialists. For they knew that once the new regime went into effect, withdrawing from it would be costly, even for a left-wing government whose leaders had opposed its initial creation. Of course, Giscard and Schmidt had an obvious stake in these costs—and in the barriers to exit more generally—being as high as possible. This can account for the intensive public relations campaign the two leaders waged in behalf of their proposal. Much of this public fanfare was aimed at convincing individuals within the global financial community that membership in the new exchange rate system signaled a larger commitment to fiscal and monetary discipline. The more successful this effort to mold the expectations of private investors, the greater the chance that a future Socialist government’s withdrawal from the system would lead to an economically destabilizing (and politically embarrassing) departure of financial capital from the French economy. Insofar as shaping actors’ expectations was important, it was also crucial to have the regime up and running as quickly as possible. Less than a year elapsed between the European Council meeting in Copenhagen at which Giscard and Schmidt offered their preliminary proposal (April 7–8, 1978) and the EMS charter’s official entry into force (March 13, 1979). Representatives of the French and German leaders filled in the details in a number of secret meetings held prior to the Bremen summit in July of 1978. Yet because their goal was to reach a final decision at the next European Council summit, Giscard and Schmidt gave these negotiators only three months after the Bremen meeting to finalize a draft of the treaty. This accelerated timetable was explicitly set out in the preamble to the communiqué issued at the Bremen summit and “reflected the wish of the French and German leaders to progress at the greatest possible speed.”29 Moving the negotiations along at such a rapid pace meant limiting the amount of time available for ironing out the earlier-noted problems with the divergence indicator. Given a more relaxed negotiating calendar, Giscard and Schmidt might also have been able to establish a more comprehensive set of rules concerning the financing of intramarginal interventions. That said, there were clear advantages to getting something “out there” quickly. The sooner the EMS became a reality, the sooner would members of the global financial community begin acquiring a stake in its continued existence and, hence, the greater the regime’s chances of surviving within France’s highly volatile political climate. ? ? ? To summarize the argument to this point, the institutional structure favored by the new monetary arrangement’s two prime movers was intended to limit the burden of participation for other actors who might one day be in a po29
Ludlow 1982, 127.
230
E U R O P E A N
M O N E Y
sition to destroy it. To this end, Giscard and Schmidt made it possible for inflation-prone members experiencing balance of payments difficulties to request a realignment of their exchange rate parities. But while the EMS afforded its member states a degree of flexibility not enjoyed by signatories of the Snake, its institutional structure was not—because it did not need to be— so malleable as to permit a newly empowered Socialist administration to altogether escape the need for austerity. Rather than permit each member state to decide for itself whether it was deserving of special dispensation, Giscard and Schmidt transferred authority over all ERM realignment requests to a supranational body, the European Monetary Committee, whose decisions were meant to be arrived at by consensus. There was, then, no guarantee that a member state’s devaluation request would be granted. As we will see, the EMS opponents of greatest concern to the regime’s two architects did not fare as poorly under the arrangement as they might have under a newly reconstituted Snake. But though the French Socialists (wisely) refrained from withdrawing the franc from the system, in the end the deck remained firmly stacked against them.
Institutional Engineering and the Co-optation of the French Socialists The years between the signing of the EMS resolution in 1978 and the French presidential election of 1981 marked one of the most economically and politically disruptive periods in modern French history. The oil shock of 1978, followed as it was in 1979 by a tightening of monetary policy in the United States and West Germany, created severe economic dislocations in industrial democracies throughout the world. In turn, fuel shortages and deteriorating economic conditions prompted voters in many advanced democracies to lose faith in their elected leaders. France was no exception. Heavily dependent upon foreign sources of oil, France saw inflation shoot up dramatically in 1979 and again in 1980 (Figure 9-1).30 Politically, the effect of higher inflation was to cause voters on 30 Clearly, any effects the EMS may have had on the formation of inflationary expectations during 1979 and 1980 were overwhelmed by the fallout from the second oil shock. Still, the EMS can be given some credit for having limited the franc’s depreciation against the D-mark during the tumultuous years leading up to the 1981 presidential election. While currency devaluations were commonplace during the regime’s first few years of existence, they were, in most cases, too small to boost the competitiveness of individual ERM signatories. This was certainly true in France’s case, as the European Monetary Committee was unwilling to permit the franc to fall against the D-mark by more than a few percentage points at a time. (Between 1979 and 1980, the franc lost about 3 percent of its value against the D-mark, a depreciation not nearly sufficient to compensate for France and Germany’s 14.8 percent cumulative inflation differential over this period.) Insofar as the resulting overvaluation of the franc caused French exports to lose competitiveness in German markets, Giscard’s ERM commitment may have kept France’s rate of inflation from climbing even higher than it did. Of even greater importance—to Gis-
R I G G I N G
T H E
S Y S T E M
231
FIGURE 9-1. Inflation in France and Its Major European Trading Partners, 1974– 1993
the right of France’s political spectrum to become disillusioned with Giscard’s handling of the economy.31 In protest, members of the French middle class shifted their allegiance to the president’s leading challenger on the right, Jacques Chirac, the leader of the neo-Gaullist party who had served as Giscard’s first prime minister. On the left, meanwhile, Mitterrand’s campaign for the presidency was buoyed by the end of the disunity that had weakened the left in the 1978 elections, as well as by the seemingly inexorable rise in France’s rate of unemployment. By 1980, the unemployment rate had risen to 6.3 percent, or more than twice the rate at which it stood when Giscard first took office in 1974. In 1981, the left achieved its long-awaited victory, as Mitterrand, its presidential candidate, captured an impressive 51.4 percent of the popular vote in the second round of balloting.32 This was followed by a new round of parliamentary elections in which the left, receiving an even larger share of the popular vote, easily secured an outright majority of seats in the National Assembly. With these back-to-back victories, twenty-three years of conservative rule suddenly came to an end. card’s middle-class supporters if not to Giscard himself—were the disciplining effects of the EMS on the Socialist government that took office in 1981 (see below). 31 The precipitous decline in France’s rate of economic growth, which fell from an extraordinary 5.2 percent in 1976 to only 1.1 percent in 1980, was an additional cause for concern. 32 Mitterrand’s triumph was not a foregone conclusion. If Chirac, who emerged as the right’s second leading vote getter in the first round of the presidential election of 1981, had more actively encouraged his supporters to vote for Giscard on the second ballot (as the leader of the French Communist party encouraged his followers to support Mitterrand), there is a very good chance that Giscard would have been reelected. See, e.g., McCarthy 1990. As it was, the difference between Giscard’s and Mitterrand’s vote shares in the second round of balloting—48.6 percent and 51.4 percent, respectively—was fewer than three percentage points.
232
E U R O P E A N
M O N E Y
During its first eighteen months in office—a period known as the Socialist “experiment”—the left radically overhauled the French economy, increasing welfare provision, nationalizing strategic industries, introducing a minimum wage, shortening the workweek, and guaranteeing workers the right to collective bargaining. But because many of these policies were inflationary, they put severe downward pressure on the franc, making it difficult for Mitterrand to continue upholding the exchange rate commitments that his administration had inherited from its predecessor. Mitterrand was thus confronted with a choice between (1) continuing the Socialist experiment or (2) maintaining France’s commitments under the EMS. In late 1982, Mitterrand abruptly terminated the experiment, replacing his administration’s earlier reforms with a succession of stringent austerity programs.33 This raises an obvious question: Why didn’t Mitterrand destroy the EMS when he had the chance? Why the Socialists Agreed to Uphold France’s Commitments Under the New Regime The prevailing view is that Mitterrand left the franc in the ERM because he feared a pullout would shatter the global financial community’s confidence in his administration’s ability to manage the French economy.34 By the time Mitterrand took office, Europe’s new monetary regime had already come to symbolize a commitment to “sound” macroeconomic policies. That being the case, France’s departure from the regime would have sent massive shock waves throughout the financial markets. Seeking protection from the specter of financial collapse, investors would have been likely to withdraw large amounts of capital from the French economy, precipitating a further deterioration in the country’s already precarious balance of trade.35 THE EC-EMS LINKAGE
In addition, there was another, more subtle cost to exiting the EMS, one that members of the Snake would have been less likely to take into account. In 33 The harshest of these was announced on March 23, 1983. Named after Minister of Economics and Finance Jacques Delors, the “Delors Plan” included price and income restraints as well as budget cuts. For overviews of economic policy making during the Socialist government’s first several years in office, see D. R. Cameron 1989; Hall 1986, chap. 8; and Sachs and Wyplosz 1986. Sachs and Wyplosz (1986, 294) explicitly make the case that the EMS is what “tipped the balance towards austerity.” 34 See esp. D. R. Cameron 1989. 35 One study has Mitterrand’s Treasury minister, Michel Camdessus, predicting that interest rates in the range of 20 to 21 percent would be necessary to avert an all-out fiscal catastrophe (Camdessus envisaged the value of the franc dropping by more than 20 percent against the Dmark). The prospect of such a large increase in interest rates was sufficient to convince Laurent Fabius, the budget minister who had once been a leading proponent of floating the franc, that leaving the EMS would in fact be a serious mistake. In turn, Fabius convinced the president. See D. R. Cameron 1989.
R I G G I N G
T H E
S Y S T E M
233
contrast with the Snake, the EMS was explicitly linked to the broader process of European integration, as reflected in the fact that its membership was restricted (as the Snake’s was not) to member states of the European Community. Countries that abandoned the EMS thus ran some risk of being shut out of the larger integration process. Now, to be sure, the danger of France’s being relegated to the European “periphery” was far lower than that of, say, Italy. While it is easy to see why Italian politicians would have been concerned about “missing the European train,”36 French authorities were in a very different position: in addition to having Europe’s second largest economy, France also provided a geopolitically important counterweight to Germany. Even if France’s European partners had been threatening to exclude ERM nonmembers from debates over the future direction of the EC—and there is no evidence that they were—it is hard to imagine why Mitterrand would have taken these threats seriously. Though perhaps wary of endangering the Community’s agricultural price support scheme, Mitterrand knew that France stood little chance of being cast out of the European “core” simply for failing to uphold its ERM obligations. That said, Mitterrand’s proEurope inclination did perhaps add some weight to the case for continued EMS membership.37 BROADENING THE DEBATE
No doubt the fear of a possible market backlash entered into the French president’s decision-making calculus. So too, one suspects, did the prospect of France losing its status as a “good” European. The real issue, however, is whether either of these scenarios was sufficient to deter France from abruptly reasserting control over its currency. Evidence to the contrary is suggested by the experience of the earlier Snake regime, whose membership varied dramatically over its six years of existence. In its first year alone, the United Kingdom, Ireland, and Italy, their leaders’ fears of provoking the financial markets apparently notwithstanding, all dropped out of the system. As for France, it dropped out on two separate occasions, first in 1974 and then (after Giscard reintroduced the franc at a more competitive exchange rate the following year) in 1976. In each of these withdrawals, the reputational costs of exiting the system were deemed insufficiently large to justify the signatory’s continuing participation. This lesson was not lost on the architects of the EMS. Consider what would have happened if the two architects of the EMS had endowed their new arrangement with an institutional structure as inflexible as the one that had characterized the Snake regime of the early- and mid1970s. In that event, I doubt many observers would have been surprised if— the reputational costs notwithstanding—Mitterrand had gone ahead and yanked the franc out of the system. Indeed, if the economic straitjacket imposed by France’s involvement in the EMS had been as tight as the earlier 36 37
See chapter 8. Frieden 1994; Sachs and Wyplosz 1986.
234
E U R O P E A N
M O N E Y
straitjacket imposed by the Snake, the outcome would almost certainly have been different. As it was, however, the new monetary system had been specifically engineered to reduce, without necessarily eliminating, the costs it imposed on its more inflation-prone member governments. While Mitterrand did not particularly like the EMS, the regime’s in-built flexibility made life “on the inside” a great deal less onerous than it might have been. BENDING WITHOUT BREAKING
Thus, for example, the French franc was permitted three devaluations against the German mark during the Socialists’ first two years in office alone (Figure 9-2). The first two of these devaluations came on October 5, 1981 (8.5 percent), and June 14, 1982 (another 10 percent). The European Monetary Committee then granted another large devaluation of the franc’s central parity—this time by 8 percent—on March 21, 1983. Taken together, these realignments reduced the value of the franc by more than 25 percent.38 And yet in all of the research that has been done on Mitterrand’s EMS “nondecision,” the regime’s provisions for exchange rate adjustments have received only passing notice. Nor have other institutional features of the regime, such as its more accommodating provisions (relative to the Snake) for financial assistance, been accorded the explanatory role they deserve. EMS signatories provided the French with generous loans through the system’s medium-term financing facility, thus further defraying the costs of Mitterrand’s postexperiment stabilization policies.39 Of course, some might argue that the reason the new French administration felt it could live with the EMS had less to do with the regime’s institutional structure than with the absence of any (sustainable) left-wing alternative to the right’s policy of fiscal and monetary contraction.40 Once this “reality” had begun to sink in, Socialist leaders stopped viewing the EMS as a liability, though for blame-avoidance purposes this is how the government chose to characterize the regime in its public pronouncements. On the one hand, the notion that Mitterrand’s Socialist experiment never had much chance for success is consistent with recent research demonstrating that expansionary macroeconomic policies tend to work best in countries where organized labor is unified and a large proportion of the total workforce is unionized.41 If this is true, Mitterrand’s “U-turn” toward aus38 As large as they were, these devaluations were still too small to compensate entirely for the inflation differentials that had accumulated between France and Germany over this period (see below). For evidence that the EMS may have contributed to a substantial overvaluation of the franc—and a correspondingly large decline in the competitiveness of French exports—see Fratianni and von Hagen 1992. Adjusting for inflation, the authors find that between 1979 and 1990, the franc appreciated against the D-mark by some 11 percent in real terms. Fratianni and von Hagen 1992, 29, table 2.4. 39 See, e.g., Ungerer et al. 1986, 6. 40 Woolley 1992; cf. McNamara 1998. 41 See particularly Garrett and Lange 1991.
R I G G I N G
T H E
S Y S T E M
235
FIGURE 9-2. The D-Mark / Franc Exchange Rate, 1974–1994
terity was a foregone conclusion; wracked by internal divisions, the French labor movement encompassed only one in five French workers. On the other hand, neither Mitterrand nor his economic advisers seemed aware of these constraints at the time.42 Quite the contrary, the individuals who entered office in 1981 believed the rest of the world was on the verge of economic recovery—in which case a newly primed French economy would have been well positioned to satisfy the sudden burst in global demand. l’autre politique It is true that the envisaged surge in demand never materialized; because the world economy remained mired in recession, the new government’s domestic expansion had the unfortunate effect of precipitating a flood of new imports and worsening France’s trade deficit. Even after these effects had become clear, however, it was by no means inevitable that the Socialist government would use deflationary policies to redress its balance of trade difficulties. In place of the austerity packages advocated by Delors and Prime Minister Pierre Mauroy, a number of Mitterrand’s closest economic advisers proposed an alternative strategy (l’autre politique) wherein France would restrict the flow of imports, encourage wage restraint through collective bar42
Hall 1986, 196.
236
E U R O P E A N
M O N E Y
gaining, and, most importantly, withdraw the franc from the exchange rate mechanism. Supporters of l’autre politique—who included several key government ministers, as well as members of the CERES group, a highly influential left-wing faction within the Socialist party (see below)—believed that by freeing itself from the need to defend the franc, the government would be able to slash interest rates, thereby stimulating investment and creating thousands of new jobs. What kept Mitterrand from pursuing this alternative path? The answer is not that he believed the existing international environment necessarily rendered aggressively expansionary policies incompatible with long-term economic growth. His decision was not, in other words, “a foreordained and inexorable consequence of ‘the constraints of the international economy.’”43 Rather, the government took the path it did because, as political scientist David Cameron (1989, 53) notes, “it chose to.” As for why the government chose to, the reason offered by Cameron is that in pursuing the “alternative strategy,” France would have been required to abandon the EMS.44 But, of course, this simply moves the analysis back to the question with which this section began: Why, exactly, was Mitterrand so reluctant to violate the exchange rate commitments he inherited upon taking office? My contention is that his propensity to abandon the EMS would have been far stronger—and the balance far more likely to tip in favor of the alternative (i.e., expansionary) strategy—had the French and German officials who drafted the EMS charter not taken care to mitigate their new regime’s deflationary burden. Why the Socialists Should (Still) Be Counted among the EMS Losers To be clear, the architects of the regime did not go so far as to eliminate this burden. Nor, given the reputational and other costs of withdrawal (not to mention sheer inertia) did they need to do this. Although the institutional structure with which Giscard and Schmidt endowed the EMS was less rigid than that of the Snake, France’s EMS-induced fiscal and monetary retrenchment nonetheless exacted a heavy toll on the Socialist government’s traditional left-wing constituencies. This is why President Mitterrand, in reaffirming France’s commitment to the EMS on March 16, 1983, was moved to declare: “We do not control our policy. In staying in this system, we are indeed condemned to the policy of the dog that kills itself swimming against the current. Only to the profit of Germany.”45 A CREDIBILITY BONUS?
Coming as it did on the same day the president publicly renounced the “alternative strategy,” this outburst would seem to remove any doubt about 43
D. R. Cameron 1989, 52– 53. Sachs and Wyplosz (1986, 294) advance a similar explanation. 45 Quoted in D. R. Cameron 1989, 51. 44
R I G G I N G
T H E
S Y S T E M
237
Mitterrand’s continuing preference for the pre-EMS (i.e., floating-rate) status quo. Still, some might say this overlooks the added measure of credibility the EMS afforded his government by institutionalizing the link between the relatively “soft” franc and “hard” deutsche mark.46 In chapter 8, I invoked this tying-hands argument to explain why the regime looked so attractive to President Giscard in the late 1970s. Is it not conceivable that Giscard’s successor also enjoyed an EMS-related credibility bonus? There are two reasons to be skeptical. First, given the social composition of its political support base, a Socialist government’s commitment to keeping the franc closely tied to the D-mark (as required by the exchange rate mechanism of the EMS) was never likely to be as credible as that of a conservative government. Second, if the Socialists had viewed the EMS as a device for speeding the downward adjustment of inflationary expectations, their leader certainly did not show it. If Mitterrand had wanted to purchase credibility for his counterinflationary policies, he would have wasted little time in publicly forswearing the use of franc depreciation as a tool of economic policy. In fact, the president vacillated for well over a year, and even then, after it appeared he had rejected the advice of those within his administration who had been urging him to disengage from the EMS, the scathing terms he used to describe the regime, as typified by the passage quoted above, left ample room for doubt. The persistence throughout the 1980s of a significant differential between interest rates in France and Germany (Figure 9-3) indicates that the financial markets never completely trusted Mitterrand’s commitment to the EMS. Nor, it would appear, did French wage setters. Inflationary expectations did decline eventually; but because the franc’s continued participation in the ERM could not be taken for granted, they did so only slowly.47 As a result, the process of adjustment was marked not only by a sharp deceleration of the growth in real wages but also by rising unemployment. When the Socialists came to power in 1981, the French rate of unemployment stood at 7.3 percent (Figure 9-4). By 1987, it had risen from this relatively low level to 10.5 percent, or approximately the same level as that of France’s five most important trading partners: West Germany (8.9 percent), Italy (12.0 percent), Belgium (11.1 percent), the United Kingdom (10.2 percent), and the Netherlands (10.0 percent). Although the situation improved after 1987, French unemployment did not decline as much as it did in these other countries. Insofar as Mitterrand’s decision not to abandon the EMS bolstered his government’s reputation as an inflation fighter, these reputational effects were modest at best. Notwithstanding the supposed benefits of “tying one’s hands,” the Socialists’ participation in the regime coincided with a protracted period of low growth and high unemployment, the fallout from 46 47
Cf. Giavazzi and Pagano 1988. Blanchard and Muet 1993; de Boissieu and Pisani-Ferry 1998.
238
E U R O P E A N
M O N E Y
Figure 9-3. Interest Rates in France and Germany, 1976–1994
which was borne primarily by the left’s principal political constituency: wage earners.48 Even if the situation could have been worse—imagine what would have happened if EMS signatories had been prevented from devaluing their currencies under any circumstances—the Socialists never regarded Europe’s “cooperative” monetary regime as anything other than an undesirable imposition. Presented with a fait accompli, France’s first left-wing government in more than twenty years was forced to compromise its basic commitment to improving the lives of French workers and the poor. POLITICAL FALLOUT
Not surprisingly, this had the effect of alienating many of the Communist voters who had turned out to support Mitterrand in the second stage of the 1981 presidential election, after their own candidate had been eliminated from the race. Dominated as it was by traditional blue-collar workers, the French Communist party had embraced the expansionary macroeconomic agenda Mitterrand pursued during his first year and a half as president; indeed, they urged the government to go even further.49 By the same token, Mitterrand’s U-turn also caused resentment within his own party, preventing the president from solidifying his base of support among its powerful leftwing factions. Chief among these was the neo-Marxist CERES group led by Jean-Pierre Chévènement. CERES had played a pivotal role in helping Mitterrand obtain the party leadership at the 1979 Socialist Party Congress, tipping the balance against his longtime rival, Michel Rocard. As it turned out, France’s participation in the EMS did not prevent President Mitterrand from winning reelection in 1988. For Communists and So48 49
See, e.g., de Boissieu and Pisani-Ferry 1998. Kesselman 1980.
R I G G I N G
T H E
S Y S T E M
239
Figure 9-4. Annual Unemployment Rates in France and Other Major European Countries, 1976 –1994
cialists alike, however, it was a hollow victory. Two years earlier, a conservative RPR-UDF coalition had won a majority of the seats in the National Assembly, thus denying the president much-needed legislative support. In most cases, economic policy was set by Jacques Chirac, the RPR leader who served as the country’s prime minister from 1986 until 1988; Mitterrand was responsible only for the handling of foreign policy. Unemployment remained in double digits throughout this period of “cohabitation” between left and right, and with the Socialists no longer in control of the domestic economic agenda, the RPR-UDF coalition in the Assembly was able to enact a major privatization program. Mitterrand could not legally initiate policy on his own. What power he was able to exert over domestic economic policy was derived mainly from his ability to veto Chirac’s initiatives.50 Chirac’s frequent missteps and the October 1987 stock market crash (one effect of which was to undermine support for the prime minister’s justenacted privatization scheme) both played a role in the right’s failure to unseat Mitterrand in 1988. However, most analysts regard infighting within the French right as the decisive factor.51 In the first round of the 1988 presidential election, Raymond Barre, the UDF’s candidate, and Jean-Marie Le Pen, the candidate of the far-right National Front, polled 16.5 percent and 14.4 percent of the vote, respectively. Together with Chirac’s 19.9 percent, candidates of the right amassed 50.8 percent of the total vote. By contrast, Mitterrand received only 34 percent. If supporters of Chirac, Barre, and Le Pen had rallied around a single candidate on the second ballot, Mitterrand 50 The French constitution grants the president the right not to sign any legislative decree of which he or she disapproves. 51 See, e.g., McCarthy 1990.
240
E U R O P E A N
M O N E Y
would have been soundly defeated. In any event, Mitterrand’s personal success in the 1988 election was tempered by his party’s failure to regain the parliamentary majority it had enjoyed prior to 1986. In the National Assembly elections Mitterrand called upon being returned to office, disgruntled voters punished Socialist candidates by staying home—the 34.5 percent abstention rate was high by the standards of past elections—and by turning (or returning) to the French Communists, whose share of the vote increased from 9.7 percent to 11.2 percent. The next five years saw the Socialists’ electoral standing continue to deteriorate. In the National Assembly elections of 1993, a new RPR-UDF coalition received 40 percent of the popular vote and captured an astonishing 84 percent of the seats in the National Assembly. Although the right had previously allowed Mitterrand to select prime ministers from within his own party, after the 1993 election the president was forced to appoint a conservative, Edouard Balladur, and a second period of divided government ensued.52 This second “cohabitation” lasted only a couple of years. In 1995, the left surrendered the presidency to Mitterrand’s long-time nemesis, Jacques Chirac. This time, to absolutely no one’s surprise, the French right prevailed.
Looking Back: The System Worked The EMS treaty’s elevation of exchange-rate changes into matters for collective decision making and adjudication was to utterly transform European monetary relations in the 1980s and early 1990s. How are we to account for such a bold, if not “revolutionary,”53 act of supranational delegation? This chapter has laid out two fundamentally different explanations. One possibility is that the EMS charter’s provisions for supranational governance were incorporated for the purpose of deterring opportunists, for example, by clarifying expectations during currency crises. As we have seen, however, matters are not that simple: while this efficiency-based account may fit comfortably within the rubric of “new institutionalist” theory, it sits rather uneasily with the facts of the case. What, then, was really going on? The first point to note is that the institutional preferences of the French and German actors who spearheaded the EMS initiative, and who were prepared to go it alone if necessary, counted for a great deal more than the preferences of either the Italians or the British. If the leaders of France and Germany, President Giscard and Chancellor Schmidt respectively, had had to defer to their Italian and British counterparts—if, in other words, the act of designing the new arrangement had been a truly collective endeavor—there 52 A former minister of finance, Balladur had supervised Chirac’s denationalization program during the earlier period of “cohabitation.” 53 Tsoukalis 1989, 63.
R I G G I N G
T H E
S Y S T E M
241
probably never would have been a European Monetary System. As for why the system’s Franco-German enacting coalition rejected the idea of simply reviving the Snake—the simpler, nondiscretionary exchange rate regime in which France and Germany had both participated (the former only for a brief period) during the years immediately following the collapse of Bretton Woods—the explanation I have offered is quite straightforward: Giscard and Schmidt harbored serious doubts about whether such a regime, though preferable from an efficiency standpoint, would be able to withstand the destructive zeal of France’s left-wing Socialist party. In short, the point of incorporating so much flexibility into the EMS’s institutional framework was less to boost the gains from cooperation than it was to ensure that the new regime, even if less than maximally efficient, would be able to survive the coming to power of a (potentially) very dangerous EMS loser. And survive it did. Although the French Socialists who took power in 1981 suffered badly under the EMS, they were not made so much worse off that Mitterrand, the party leader, felt compelled to reassert France’s control over the franc upon taking office. Does this mean that the French Socialists were somehow transformed from losers into winners? Not at all. While the enacting coalition may have been successful in defusing the Socialist threat, it is safe to say that Mitterrand, had he been standing in Giscard’s shoes in 1978, would never have allowed the EMS—more accommodating though it was than the Snake—to come into existence. So who did benefit from all of this? The French right, most obviously. By 1987, French inflation had dropped from 10 percent, which is where it was running when the EMS was conceived in 1979, to just over 3 percent. By 1992, inflation was running at lower rates in France than in Germany (see Figure 9-1 above). In Germany, meanwhile, the beneficiaries of European monetary cooperation included citizens on both sides of the political spectrum. Invited by the Free Democrats to form a new governing coalition in 1982, Helmut Kohl’s Christian Democratic party was no less receptive to the EMS than its Social Democratic predecessors had been. And in retrospect, it is not hard to see why. By retarding the appreciation of the D-mark against the franc, the EMS reinvigorated Germany’s export sector, making it possible for Germany to go from a current account deficit in 1979–81 to a current account surplus that was to persist throughout the 1980s.54 The regime’s importance to the German economy was made especially apparent after the 1985 Plaza Accord. Although the next two years saw the German currency appreciate by more than 20 percent against the U.S. dollar, it appreciated relatively little against the franc and other ERM currencies. As a result, the 9 percent decrease in Germany’s effective real exchange rate was only half as large as the loss of German competitiveness that occurred during the late 1960s, when the deutsche mark appreciated against 54
Henning 1994, 196.
242
E U R O P E A N
M O N E Y
amount.55
the dollar by a similar While German industry performed poorly in U.S. markets, causing Germany’s trade surplus with the United States to decline from DM23 billion in 1985 to DM8 billion in 1989, Germany’s surplus with France, along with other European countries that pegged their currencies to the deutsche mark, more than made up for the difference.56
Looking Ahead: From Anarchy to Organization— and Back to Anarchy? Whatever its impact on the political fortunes of the EMS losers that presided over Italy, Britain, and France during the 1980s and early 1990s, the EMS was a virtually unmitigated success from the perspective of the center-right Franco-German coalition that conceived and designed the regime in the late 1970s. As for whether this “success story” will continue beyond the 1990s, the answer is much less clear. What follows is a brief discussion of European monetary relations in the run-up to the twenty-first century. My own view is that the pattern of developments during these years accords quite closely with this book’s larger theoretical analysis, although (as noted earlier) any assessment drawn from such recent events should be considered somewhat speculative. The Collapse of the EMS Examination of this period properly begins with the abrupt withdrawal of two major ERM currencies—the Italian lira and British pound—in September of 1992. The events leading up to this crisis, which culminated in the widening of most remaining members’ currency fluctuation bands from 2.25 to 15 percent, have been well chronicled elsewhere.57 Suffice it to say that the impetus for the departures of the lira and pound, and for much of the turmoil that characterized the ensuing eleven months, was the unification of Germany in 1990. Faced with the urgent task of rebuilding the former East Germany, Chancellor Helmut Kohl’s center-right government undertook an unprecedented fiscal expansion, transferring some 8 percent of West Germany’s total output to the former East Germany. Aggregate demand received a further boost when, shortly after unification, Kohl agreed to a one-to-one exchange rate (with some limitations) between the deutsche marks of East and West Germany at a time when the market rate was estimated at between four and seven Eastern D-mark for one Western D-mark. It was at this point that the Bundesbank, fearing a rise in Germany’s rate of inflation, began furiously raising short-term interest rates. By the summer 55
Giavazzi and Giovannini 1989, 60. Henning 1994, 217. 57 See, e.g., Branson 1994; Eichengreen 1994, 96 – 98; Henning 1994, 228 – 41. 56
R I G G I N G
T H E
S Y S T E M
243
of 1992, the discount rate in Germany had reached 8.75 percent, its postwar peak. With German interest rates climbing skyward, European countries whose currencies were tied to the deutsche mark were forced to raise their interest rates as well. This would not have been a problem if Italy and the United Kingdom had been experiencing their own postunification inflationary shocks. In fact, their governments were at that time struggling with high unemployment and stagnant (in Britain’s case, declining) growth, and so were desperate for a relaxation of monetary policy. Both Italy and Britain responded to these added deflationary pressures by withdrawing from the ERM on “Black Wednesday,” September 16, 1992.58 In analyzing the larger theoretical significance of these withdrawals, it is important to appreciate that the Italian and British governments were both extraordinarily reluctant to abandon the system. In the end, of course, both did choose to drop out. But that was only after a tidal wave of speculation in the foreign exchange markets had thoroughly depleted the reserves of their two central banks. The Italians and British pulled out, in other words, only because they could not continue paying the huge sums necessary to maintain the value of their currencies. In neither country did public officials think that devaluing would somehow translate into lower interest rates. Quite the contrary: In both Italy and the UK, most officials took the view that while it was unfair that their constituents be asked to bear the costs of German unification, the alternative—cutting loose from the ERM—would only make matters worse. Better for the Italian and British governments to maintain their current ERM parities, the thinking went, than to abandon them and, in so doing, raise questions about their larger commitments to “responsible” economic policies. This is, of course, perfectly in keeping with the power-politics interpretation of European monetary cooperation I began putting forward in chapter 8. Their mounting losses notwithstanding, the Italian and British governments were exceedingly reluctant to renege on their prior ERM commitments, and it was not until the bitter end—and, even then, mostly out of necessity—that they did so. That said, Britain’s jumping off the EMS bandwagon did not in fact result in the balance of payments difficulties that nearly everyone had been predicting. Three days after Black Wednesday, the Bank of England reduced interest rates from 15 to 9 percent. By January 1993, short-term interest rates 58 France nearly abandoned the system as well. Prior to German unification, the French and German economies had tended to move in tandem; an externally generated shock that caused a downturn in one country caused a similar downturn in the other (see, e.g., Bayoumi and Eichengreen 1993). German unification was different from earlier “exogenous” shocks, however, in that its adverse economic consequences were almost entirely concentrated on the German economy. France thus found itself in a position not unlike that of Italy and the United Kingdom. The mystery is why, notwithstanding this unification-induced divergence, France was unwilling to grant a revaluation of the D-mark (see, e.g., Henning 1994, 237– 44).
244
E U R O P E A N
M O N E Y
had fallen to 6 percent, a fifteen-year low. That the financial catastrophe predicted by Lamont and others never materialized came as a great relief to Britain’s Conservative government, which was able to relax its monetary policy and speed the country’s economic recovery. As for Italy, the economic and political fallout produced by its ERM “defection” was more serious, but here, too, life on the outside, if not exactly carefree and easy (see below), was at least manageable. Why did the two ERM drop-outs not suffer more as a result of their withdrawals and subsequent exclusion? To what did they owe their good fortune? In retrospect, three political-economic developments deserve much of the credit—or blame, depending on one’s point of view: The Easing of German Monetary Policy. The German Bundesbank sharply lowered interest rates immediately following the 1992–93 crisis, in part to ensure that France would not be forced off the bandwagon like Italy and the UK. A Political Corruption Scandal in Italy. Coming just months before the outbreak of the currency crisis, the Italian corruption scandal, involving as it did virtually every prominent member of Italy’s governing elite, paved the way for a series of previously unthinkable deflationary measures. Almost immediately after its decision to withdraw the lira from the ERM, the Italian political establishment (what was left of it) began clamping down on inflation harder than ever.59 A Fundamental (Albeit Temporary) Shift in Market Expectations. Over time, the global financial markets came to appreciate that the newly unified Germany was undergoing an unprecedented “country-specific” economic shock and that, under the circumstances, it was simply unrealistic to expect other ERM countries to honor their earlier commitments. Changing Money: The Why and How of European Monetary Union Notwithstanding the extraordinary turbulence that characterized European monetary relations during 1992 and 1993, it would appear that Europe’s monetary bandwagon has now regained much of its earlier momentum. There is, however, one important difference between Europe’s new monetary union and the “old” EMS it will fully supplant (barring a major crisis) shortly after the turn of the century. 59 Observing the situation in 1993, the economist Rudiger Dornbusch was struck by the “veritable and amazing range of reforms” that Italy’s government had undertaken since the toppling of the lira: “Pervasive privatization has been accepted politically and is actually moving forward; budget balancing is proceeding against all odds; the political system is being cleaned out; and, most importantly, interest rates have not been cut in half. In fact, . . . discipline about inflation is central to the Italian strategy. No one would question that today Italy looks far more serious than before the attack [on the lira].” Dornbusch 1993a, 133.
R I G G I N G
T H E
S Y S T E M
245
The difference is that the supranational architecture of the new monetary union affords its member governments considerably less scope for independent, discretionary decision-making. As we have seen, the EMS regime’s provisions concerning parity alignments were, by EMU standards, relatively lenient. This flexibility came in handy for the expansionist-minded government that presided over the French economy during the early 1980s, making it possible for France to run a somewhat looser fiscal and monetary policy than was being pursued at that time elsewhere in Europe. The adoption of the euro as France’s national currency has, for obvious reasons, removed this flexibility. In theory, of course, a left-wing French government hell-bent on stimulating aggregate demand could still (1) use its influence in the Governing Council of the new European Central Bank (ECB) to seek a reduction in European interest rates or (2) run a series of large budget deficits.60 In practice, however, France’s central bank governor would get to cast but one vote on the Governing Council, and the ECB’s six-member Executive Board (without whose support it would be difficult to gain the necessary majority) would probably not be very receptive to French proposals calling for a dramatic expansion in the European money supply. As for the prospects of the second option—using fiscal policy to stimulate France’s output and employment— the difficulty here is that participants in the new union are required to uphold a nondiscretionary “stability pact.” Championed by the former German finance minister, Theo Waigel, this pact imposes an automatic fine of up to 0.5 percent of GDP on any member country whose budget deficit exceeds 1 percent. The conservative French president Jacques Chirac was insistent that allowances be made for exceptional circumstances, but the numerical fiscal requirements established by the pact nonetheless remain highly restrictive. Can the theory of institutional design elaborated earlier help us make sense of any of this?61 60 There is in fact a third, albeit less direct, avenue for mitigating the “disciplining” effects of membership in the new monetary union: a member government may try to get the European Council of Economics and Finance Ministers (ECOFIN), which controls the euro’s exchange rate, to drive down its value relative to other, non-EMU currencies. See D. R. Cameron 1998, 208–11; Feldstein 1997, 38 – 39. 61 As noted in the text, the original impetus for the stability pact (which was not part of the original EMU treaty) came from the German finance minister, Theo Waigel. Waigel and others were concerned that the inauguration of a full-fledged monetary union would encourage profligate fiscal policies, as EMU members would now be able to increase their budgetary deficits without fear of triggering a massive flight of investment capital. After all, where would capital flee? Certainly not to the D-mark, which, after the transition to EMU, would effectively cease to exist. And if a budgetary expansion in, say, France or Italy were to provoke a run on the euro, the economic costs resulting from such a crisis (e.g., the rise in the risk premium attached to euro-denominated assets) would be diffused across all members of the system; see Eichengreen 1994. It was for this very reason, argue David Cameron (1992a) and Wayne Sandholtz (1993), that the French Socialists and the Italians championed EMU during the late 1980s. And it was for the same reason—that is to say, the expectation that a full-blown monetary union would weaken the contractionary yoke of the existing EMS regime—that German central bankers were
246
E U R O P E A N
M O N E Y
Absolutely. In contrast with the situation facing the architects of the EMS, the architects of EMU encountered no major political opposition from would-be “losers” inside the regime’s pivotal Franco-German axis. Of critical importance here was the French government’s coming to terms with the preexisting EMS regime, a begrudging acceptance that began in 1983, as we have seen, and increased steadily thereafter.62 To be sure, the EMU initiative did meet with strong criticism from French politicians on both the left and the right. But with the exceptions of the country’s Communist party (which viewed the euro as an anti-inflationary sellout to the interests of global capitalism) and the far right (whose champion, Jean-Marie Le Pen, regarded the euro as an unacceptable surrender of national sovereignty), the target of France’s anti-EMU backlash was less the single-currency regime itself than the deficit-cutting measures that the French government was required to undertake in order to qualify for membership in the first wave of joiners.63 In Germany, meanwhile, the idea of moving to a single currency was strongly and consistently embraced by Chancellor Kohl. And though many German citizens viewed the idea with suspicion (if not outright hostility), party leaders across Germany’s political spectrum came out in favor of it anyway.64 Here, then, is a parsimonious explanation for EMU’s nondiscretionary institutional framework. Why was this framework acceptable to the French and German governments in the late 1990s but not the late 1970s? The answer is that this time around (in contrast with the earlier EMS experience) the Franco-German enactors encountered no internal opposition, or at least none with any realistic prospect of wielding political power. Consequently, initially opposed to EMU. On some accounts, Bundesbank officials raised the German discount rate immediately after the Maastricht summit (and then followed that raise with another increase six months later) as a way of sabotaging the EMS and thereby slowing progress toward a single European currency. See, for example, the account of Bundesbank behavior offered by Henning: “Whether some members of the Central Bank Council deliberately sought a realignment when taking the decision to increase interest rates in July 1992 cannot be known with certainty. The German central bank nonetheless appeared to be kicking the EMS and the Maastricht Treaty when they were most vulnerable.” Henning 1994, 237. 62 For further analysis of Mitterrand’s (belated) embrace of monetary union, see D. R. Cameron 1992a; Garrett 1998; Grieco 1996; Milner 1997, 215–18; Moravcsik 1998, 404–17. 63 Whitney 1997. As set forth in the Maastricht treaty, these requirements—the famous “convergence criteria”—included the stipulation that EMU participants limit their annual budget deficits to no more than 3 percent of gross domestic product. The fiscal discipline necessitated by this restriction (and there were others as well) prompted authorities in France, and indeed throughout Europe, to enact stringent austerity measures. Introduced by Chirac, these measures remained in effect even after the French parliamentary elections of 1997 gave the left a majority of seats in the National Assembly and a new period of “cohabitation” ensued. 64 R. Cohen 1997; Kaltenthaler 1998, chap. 5. There are a number of reasons for this. One is that if EMU were to become a reality, the German government would no longer have to be concerned, as it had been during the late 1970s (see chapter 8), about the deleterious effects of a strong D-mark on Germany’s export-driven economy. At least as important, the symbolism surrounding the passage and ratification of EMU reassured Germany’s European neighbors, as well as the United States, that “unification would not be the harbinger of a bellicose remilitarization.” Garrett 1994, 58; see also Grieco 1996; Joffe 1997.
R I G G I N G
T H E
S Y S T E M
247
the pivotal players behind the EMU initiative did not feel the same need to soften their new regime’s overarching institutional structure so as to discourage future French or German governments from withdrawing from it. The closing of the devaluation option aside, however, the EMU bandwagon of the late 1990s bears a striking resemblance to the EMS bandwagon of the 1970s and 1980s. Technically, the Maastricht treaty would have permitted the formation of a monetary union with as few as seven European Union member states. But while the possibility of a two-track or multispeed Europe was explicitly recognized by the agreement, most European governments were intent on adopting the euro at the earliest possible juncture. Prospects for the Euro: Can One Size Fit All? Most, but not all. In the UK, the current Labour government is in no particular hurry to abolish the pound, and has in any event promised to submit the matter to a national referendum. But Britain too will find it difficult— over time, increasingly so—to remain outside the continent’s new monetary union. The Labour party, which scored a resounding victory over the Tories on May 1, 1997, has seen what the UK’s earlier delay in joining the EMS did to British interest rates under the Conservative governments of the 1980s. For this reason, few observers view Britain’s relegation to the EMU “periphery” as a permanent condition. And yet, even if the country’s increasingly euro-friendly prime minister, Tony Blair, does go ahead and adopt the new European currency, the abolition of the British pound is not a change that most British citizens welcome. In Italy, the feeling is much the same—a sense of regret that France and Germany, rather than embracing the pre-EMS floating-rate status quo, have instead taken the opposite tack, embracing a full-blown common currency. But what if, in the wake of the 1992–93 currency crisis, France and Germany had in fact followed Italy and the UK into the floating-rate wilderness? Would Italian citizens really have been better off with a return to the “noncooperative” politics that characterized European monetary relations during the 1970s? There is ample reason to think so. While abolishing the lira made eminent sense given Italy’s limited range of options—its limited “choice set”—the adoption of the euro seems more likely to exacerbate Italy’s economic and political woes than it does to alleviate them. Even if the country’s business cycles are today somewhat more in sync with those of Europe’s “hard money” core than they were during the 1970s and 1980s, it is still the case that (1) Italy’s rates of labor mobility are lower than those of France and Germany, (2) its wage-setting system is more ossified, and (3) an EU-wide system of fiscal transfers remains a far-off prospect. Making matters worse, the institutional structure of the new monetary union is, as noted earlier, even less accommodating than that of the EMS, which permitted members experiencing temporary balance-of-payments difficulties to ask for (if not always
248
E U R O P E A N
M O N E Y
to obtain) fairly substantial exchange rate devaluations. Finally, whatever the long-term repercussions of Italy’s participation in a common currency may turn out to be, the stringent budgetary measures necessitated by Maastricht’s convergence criteria proved to be a hard sell politically. Prime Minister Romano Prodi’s 1997 budget, which called for a one-time levy (labeled a Eurotax) on middle- and upper-level incomes, was intensely debated, and the current conventional wisdom is that it has been highly damaging to his party’s left-wing constituencies. In Britain, too, the loss of an independent monetary policy is destined to produce a great deal of economic and political turmoil. As noted at the end of the previous chapter, business cycles in Germany are more out of line with those in the UK than with those in any other European country except Finland, Sweden, and Ireland.65 At the same time, Britain’s distinct strain of Euro-skepticism is as virulent as ever, as are fears that in “surrendering” its sovereignty Britain could end up losing its national identity. In this regard, it is worth noting that John Major’s policy toward monetary union—his preference for taking a wait-and-see approach rather than rejecting the euro outright—prompted two hundred Euro-skeptics within the Tory party to break ranks with Major during the last couple of weeks of the 1997 election campaign, a defection that certainly did not help the Tories’ cause. In truth, however, the Labour party is no more unified on the issue. “The lesson is clear,” writes the New York Times (Judt 1997): “Contrasting social and economic practices [that] are born of long-standing political and cultural differences . . . cannot be obliterated with the wave of a magic monetary wand.” If Britain ever does adopt the euro, prepare for a political backlash of major proportions. 65 Recall Table 8-5. This asymmetry was not lost on the British Treasury. In October 1997, the Treasury issued a report warning that because Britain’s economic recovery was “years more advanced” than those of its would-be EMU partners, “cutting British interest rates to continental rates—appropriate for countries where economic conditions are more depressed—could detonate a runaway boom, with inflationary pressure that could not be responded to by devaluation.” Joffe 1997, 26.
P A R T
V
Conclusions
C H A P T E R
1 0
Rethinking International Cooperation The definition of alternatives is the supreme instrument of power. —E. E. Schattschneider
Why are new regional and multilateral initiatives currently proliferating— and old ones expanding—at such a furious pace? Might these trends presage a full-scale transformation in the “anarchic” structure of today’s international system? If so, how will the interests of particular societal groups, governments, and nations be affected during the transition? What, in short, will the rise of supranational institutions mean for our collective future? As we make our way into the twenty-first century, few lines of inquiry in the social sciences seem more timely or, indeed, more pressing. To this point, however, the analytical apparatus that has framed scholarly discussion of these issues—and that informs current debates about the appropriate policy response1 —has left out an important part of the theoretical story.
Asking the Right Questions While there are many themes running throughout the recent literature on international cooperation, the dominant lines of argument all share one fundamental premise: all presuppose that international institutions, or international “regimes” more generally, are structures of mutual advantage. If these arrangements are difficult to establish—and they almost always are—that is not because anyone doubts their ultimate desirability. On the contrary, the arrangements that do make it into existence are assumed to facilitate a whole range of collectively desirable goals that their member states could never hope to accomplish (at least not as efficiently) on the basis of unstructured intergovernmental bargaining and negotiation. As for why international institutions are in greater demand today than in the past, that’s because we are moving into an era of ever-increasing “globalization,” in which the benefits of maintaining cooperative interstate relations are perceived to be higher— and the anticipated gains from unilateral action lower—than was the case in earlier years. This is the thrust of what theorists of international relations have had to say about recent changes in the institutional landscape of international politics. If one looks at what has actually been happening, however, one cannot 1
See, e.g., Haas, Keohane, and Levy 1993; Mearsheimer 1994/95.
252
C O N C L U S I O N S
help but be struck by the gulf separating the literature’s theoretical analysis from the emerging institutional reality. Is More Cooperation Necessarily Better? Take the recent “contagion” of regional trade institutions, for instance. Why are political leaders throughout Europe and the Americas actively seeking accession into the European Union and NAFTA—or, when their requests for admission are denied, forging ahead with similar (albeit smaller-scale) regional trade regimes of their own? The economic argument for eliminating protectionist trade barriers is predicated on the assumption that the domestic beneficiaries of economic liberalization compensate the domestic losers, so that in the end all individuals and groups in the free-trading society are made better off. In many countries, however, the prospect of higher taxes and lessened competitiveness discourages politicians from erecting the kinds of domestic economic and political institutions needed to ensure that the benefits of free trade and economic liberalization are in fact widely shared. And because governments are unwilling to “splurge” on social safety nets, the increase in cross-border flows of goods, services, and capital has been accompanied in many societies by rising income inequality.2 In turn, resentment and anger on the part of the individuals and groups whose livelihoods are threatened by rapid market integration (and who in a number of societies constitute a substantial proportion of the population) have created fertile soil for the growth of militant nationalist and fundamentalist movements.3 In sum, trade cooperation—the reciprocal elimination of protectionist trade barriers—may be conducive to economic liberalization, but economic liberalization is by no means a universal or unmitigated good. Nor, as the politically turbulent history of the European Monetary System makes clear, is the “good” obtained through regional monetary cooperation—namely, a stable exchange rate. To the contrary, the loss of devaluation as a tool of macroeconomic policy can lead to political turmoil in countries that lack alternative means of adjustment during periods of temporary downturn (chapter 8). Yet if international cooperation is not always Pareto-improving, how are we to explain the fact that virtually every East and Central European nation is presently clamoring to join the ranks of NATO and the European Union while, on the other side of the Atlantic, the countries of Latin American are pushing (without any prompting from the United States, Canada, or Mexico) for the creation of a “NAFTA-plus” hemispheric free trade regime? If the leaders of all of these countries—and many other examples could be cited as well—expected to be made worse off, why would they be lobbying so hard for admission? It is not as if their interest in participating in these arrange2 It is possible that we will one day see the emergence of compensating mechanisms at higher levels—regionally, if not globally—though we are of course a long way from that point today. 3 Cf. Rodrik 1997; see also Garrett 1998.
R E T H I N K I N G
C O O P E R A T I O N
253
ments—or (to take yet another example) the interest of China’s governing elites in gaining admission to the WTO—were inevitable, the natural outgrowth of some sort of technological imperative. When a country’s leaders request permission to join an existing regime, they do so by choice; had they wished, they could have chosen to remain on the outside. The same holds for the insiders as well; should a member government decide to opt out, it is free to go at any time. It would be one thing if each member government drew its support from domestic constituencies who cared only about their society’s long-term “national interests.” Most of the time, however, what seems to matter most to these constituencies are narrower, more immediate considerations. That this is the case makes the recent wave of government support for regimes promoting market integration, fixed exchange rates, and the like seem all the more perplexing. For in countries ranging from Canada (under the Liberals) to France (under the Socialists), much of this support has been coming from government officials who were elected to represent the interests of globalization’s short-term losers. By relinquishing control of their domestic economies and accelerating their countries’ integration into the global economy, these officials would appear to be increasing, not decreasing, their electoral vulnerability. And yet, throughout the 1980s and 1990s, international institutions have been spreading like wildfire. If these arrangements did not actually facilitate collective gains, why would governments throughout the world be lobbying so furiously to join them? This sort of thing should not be happening—or at least our current institutionalist theories would appear ill equipped to explain it. Shifting the Status Quo in International Relations Research But institutionalist theory is capable of explaining it. The problem is simply that, until now, the theory’s underlying rational-choice logic has been pursued in only one direction—toward that holy grail of institutionalist research known as the Pareto frontier. There has never been a very good reason for this. Although most scholars see international institutions as efficient, Pareto-improving responses to collective-action problems, this is not the only logical possibility. Nor, it appears, is the empirical basis for this point of view particularly compelling either. The European Monetary System and the North American Free Trade Agreement are two of the most important instances of international cooperation in recent memory. Yet, in each of these cases, the power-politics framework developed in the first two parts of this book provides at least as plausible an explanation for the relevant motivations and outcomes as do the mutual-gains interpretations so familiar to students of international relations. As we have seen, for example, the fear of being left behind was, for a number EMS and NAFTA entrants, a more important impetus for joining than the prospect of mutual gain. Indeed, rather than making things better, the
254
C O N C L U S I O N S
“goods” provided by these far-reaching regional initiatives would appear to have left some signatories considerably worse off. As for why these initiatives ended up taking the institutional forms they did, in both cases the evidence suggests that what mattered to the pivotal prime movers was less that “their” institutions be efficient, in the sense of mitigating coordination and free-rider problems, than that they be able to withstand the coming to power, at some point in the future, of an EMS or NAFTA loser whose participation in the regime would be necessary for its continuing survival. Note that the participation of each EMS and NAFTA member state remained strictly voluntary. In neither case did the government leaders who made up the “enacting coalition” force their counterparts in other countries to take part in the regime by subjecting them to coercive pressure. Nor was there really any need for the EMS or NAFTA beneficiaries to do this, for in each case the regime’s losers were confronted with a fait accompli. In addition to raising questions about the conventional understanding of power as involuntary coercion, this line of argument also underscores the limitations of thinking about international politics as a prisoners’ dilemmalike game of strategic interaction. As pervasive as it is, this way of looking at things may in fact be doubly misleading—first, because each player in the prisoners’ dilemma game predicates its own strategy on what it anticipates will be the strategy played by its partner and, second, because (in equilibrium) neither player in a prisoners’ dilemma ends up worse off than it was in the initial status quo. Whereas the players in a prisoners’ dilemma predicate their strategies on the anticipated responses of their partners, here the behavior of the winners is utterly nonstrategic. Though the effect of the winners’ cooperating is to remove the status quo from everyone else’s choice set, that is not why they cooperate. Again, the winners behave as they do for purely “internal” reasons: they behave strategically with respect to each other, but not with respect to anyone else. As for the second point, the big lesson to be learned from the single-shot prisoners’ dilemma game is that equilibria need not be collectively optimal. In the power-politics scenario explored here, however, the problem is not one of “suboptimality.” In equilibrium, no one really loses in a prisoners’ dilemma game; it’s just that none of the players is able to improve upon the initial (suboptimal) status quo. Here, by contrast, there really are losers—players whose utility was higher at the outset of the game (in the status quo) than it was after the game reached its equilibrium. To these players, the status quo is not suboptimal; it is the outcome they would actually prefer. Given that this status quo is not an option, these players decide they are better off cooperating with their partners—the winners—than not cooperating. But while the prospect of mutual gain is what drives their partners to cooperate, they—the losers—cooperate simply to avoid being left behind. This, of course, helps explain why the supranational initiatives we have been seeing of late often seem to impose high costs on some of their participants.
R E T H I N K I N G
C O O P E R A T I O N
255
There is a temptation to assume that, because this is so, the bandwagoning phenomenon of past years will be short-lived and isolated to a handful of states—that it is, in other words, self-limiting. But that is not how the logic of go-it-alone power works. As pressures on outsiders continue to mount, more and more countries may be “forced” to conform to rules of the game that members of the original enacting coalition have established for their own purposes—rules that may or may not be conducive to the economic interests of other states or the political interests of governing elites within them. The result would be greater cooperation. But along with more cooperation one may well find greater-than-necessary economic dislocations and political upheaval, as well as growing asymmetries between a pro-integration “core” and an anti-integration “periphery.” In time, of course, the beneficiaries of cooperation could lose their initial positions of power. Even then, however, it would be a mistake to assume that the bandwagon would immediately come grinding to a halt, for the institutional arrangements established by the original enacting coalition may have been specifically designed to become an enduring feature of the international landscape. Drawing upon the new economics of organization, most attempts to explain why international institutions look as they do emphasize the difficulty of writing “complete contracts” and the resulting—potentially unlimited—opportunities for reneging that this affords. In this explanation, states rely on supranational adjudication systems to deter would-be defectors from exploiting contractual ambiguities, that is, from trying to pass themselves off as good-faith cooperators when in fact they are blatantly violating the spirit (if not necessarily the letter) of the law. Others emphasize the role that supranational entities can play as equilibrium-selectors, choosing “focal points” around which states subsequently, and spontaneously, coordinate their behavior. Rather than undergoing a time-consuming process of direct bargaining and negotiation, states confronted with a multiplicity of Pareto-improving possibilities may be better off—so this argument runs— transferring control over the negotiating agenda to a supranational “agent,” empowering it to choose the equilibrium on their behalf. Note that, in both of these arguments, institutional choice is viewed a matter of finding the organizational arrangement that—by mitigating collectiveaction problems of one kind or another—most efficiently advances the collectivity’s interests. But what if that arrangement did not leave the whole collectivity better off? No doubt the would-be loser(s) would do everything in their power to keep such an arrangement from ever seeing the light of day. But what if “everything in their power” were not enough? Imagine, for example, that the beneficiaries of a blatantly inefficient, non-Pareto-improving organizational system had the capacity to “go it alone”—and their partners (the losers) knew it. Appreciating that some actors may be able to impose their preferred organizational structures on other actors in the system casts the issue of institutional choice in a rather different light. If it’s true that some cooperators
256
C O N C L U S I O N S
must accept whatever their partners dish out, the theoretical task becomes one of explaining why these partners, the enactors, would choose to do their bidding through supranational structures that afforded everyone in the collectivity, winners and losers alike, some role in organizational decision making. By offering the latter some (albeit limited and indirect) means of addressing their concerns—employing supranational dispute resolution systems, collective decision-making bodies, and the like to give the losers a greater voice than they would otherwise enjoy—the enactors would seem to be diminishing, not enhancing, their own gains from cooperation. That they would be willing to forgo these gains is all the more curious when one considers that, in the initial period at least, the other members of the collectivity—that is, those not part of the regime’s enacting coalition— are essentially boxed in. As long as opting out of the new regime would be (even) worse for them than participating in it, these other members will find it in their own interests to take part and, indeed, would decide to do so even if the enactors had purposefully engineered “their” regime so as to deny their partners any voice whatsoever. On the other hand, what is true for their initial cooperation partners would not necessarily be true for the enactors’ own domestic successors. What if the cooperative regime at issue were strongly opposed by the domestic challengers of one (or more) of the enacting governments? In that event, the members of the go-it-alone coalition would be in serious trouble, for now the ranks of regime “losers” would include an additional set of actors who could destroy their regime—as the enactors’ external opponents could not—simply by choosing not to participate in it. Herein lies what seems to me to be the most important yet, until now, little appreciated impetus for the establishment of supranational conflict resolution bodies and higher-level, EU-like governance structures. What makes these structures so attractive is not that they contribute to the overall efficiency of cooperation. It is that they provide the enacting coalition’s domestic successors a readily available means of moderating the high costs that cooperation imposes upon them, thereby making their continued participation in the regime somewhat less onerous than it might otherwise be. In this way, supranational structures discourage the beneficiaries of cooperation—but also, more importantly, the actors who do not benefit—from working to subvert the new (cooperative) status quo should they ever find themselves in a position to do so. Obstacles to Theoretical Progress Why have previous students of international relations failed to take up this line of analysis or, more generally, to incorporate the losers’ side of the story? Part of the answer goes back to the verification problems discussed at the end of chapter 3. As we have seen, the task of verifying my theoretical claims about power, cooperation, and institutions is confounded by a number of
R E T H I N K I N G
C O O P E R A T I O N
257
vexing methodological issues. In determining whether a particular party to an agreement is better or worse off, for example, the analyst is required to deduce what that party’s utility would have been if its neighbors had failed to enter into a cooperative partnership, thereby removing the previously existing, noncooperative status quo as a viable option. In addition to the counterfactual problem of determining how the ostensible “loser” would have behaved if the “winners” had been unable (or just unwilling) to go it alone, other difficulties of verification stem from the nontransparency of government utility functions. Even if a governing official publicly disparages a regional or multilateral arrangement in which her country has elected to participate, that does not mean that the official (or the constituents whom she represents) are necessarily worse off. In light of these considerations, it is difficult to prove that a winners-and-losers interpretation of a particular series of events is the correct one.4 In the end, however, this is not the main reason why the winners-and-losers side of international politics has gotten short shrift in the cooperation literature. If proponents of the dominant perspective have avoided the sort of counterfactual analysis pursued in this book, it is not because they think it cannot be done; it is because (in their view) it simply isn’t required. If ever a multilateral arrangement stopped producing collective benefits, the assumption is that all of the losers would promptly withdraw. And yet today we find that just the opposite is happening: Not only are such withdrawals exceedingly rare, but most international institutions are surrounded by more aspiring entrants than they know what to do with. But just because a country’s political leaders hope to join a particular institution does not mean they regard it as a structure of mutual advantage. Insofar as my argument holds—and, again, the empirical conditions that would need to be satisfied do not seem especially restrictive or demanding— one would actually expect new regional and global arrangements to exhibit a systematic bias in favor of some signatories (the A’s and B’s) and against others (the C’s and D’s, etc.). Some states may be cooperating, then, not because they regard the new cooperative regime as a good thing, but because they have concluded (correctly) that the alternative—allowing the A’s and B’s to go it alone—would be even worse. So far, however, previous analysts have failed to take a hard look at the winners-and-losers dimension to international cooperation. All too often (as in the literature’s treatment of the EMS and NAFTA) it is simply overlooked. To be sure, one can find researchers who do address these parts of the story. Typically, however, these analysts come from outside the discipline of political science. Most are either historians, such as Peter Ludlow, or economists, such as Sidney Weintraub, with expertise in a specific issue area or geo4 On the other hand, proponents of the more conventional, positive-sum models of cooperation and institution building cannot definitively say that their interpretations are correct either; methodological problems make it impossible for either side to score a knockout punch against the other. But that ought not to be our goal anyway (see below).
258
C O N C L U S I O N S
graphic region (or both): the politics of monetary policy in Western Europe, for example, or commercial relations between the United States and Mexico. That these individuals tell both sides of the empirical story is to their credit. On the other hand, the empirical knowledge that they bring to the table, as extensive as it is, does not automatically translate into a coherent theoretical whole—and yet it is precisely this overarching theory of international politics that is, or ought to be, our ultimate goal. To reach this goal, we will first have to move beyond our existing paradigms. While the world around them may be undergoing extraordinary change, however, the truth is that most scholars are quite content with the theoretical status quo. What is needed, in their view, is not a new collection of theories so much as a synthesis of the ones we already have.5 As a recent review of the EU literature concludes, “The efficiency considerations that are the economist’s bread and butter, the self-interested political behavior whose analysis comes naturally to the political scientist, and the institutional approach that has gained increasing favor under the banner of ‘the new institutionalism’ need to be blended to provide a balanced picture of the integration process.”6 This blending is already well underway. Indeed, for all the supposed acrimony between neoliberals and realists, proponents of these well-established traditions have spent the last several years laying the foundations for an elegant, higher-order synthesis, one that takes the diverse strands of a larger (primarily rational choice) literature on cooperation and institutions and fashions them into a single analytical framework. The result is what I earlier termed the “collective-action paradigm.” Whatever one chooses to call it, however, the point is that both sides have begun working together. And yet are we really any closer to providing the “balanced picture” everyone claims to want? The real problem with current research is not, I think, that our theories are too disparate. Nor is the problem that our empirical base is too limited. The real stumbling block is that these theories have been put to use in understanding only one side of the theoretical story—the side having to do with collective action, efficiency, and mutual gains. If we want to tell the other
5 Of course, not everyone is convinced that the mainstream theoretical literature is moving in the right direction, as evidenced by the recent outpouring of work by “constructivist” critics of this literature (see, e.g., Kratochwil and Ruggie 1986; Wendt 1987). Many of these critics explicitly reject the notion that theories of international relations must be rooted in microfoundational assumptions about state behavior. Inasmuch as the mainstream literature has advanced our understanding of international cooperation, however, it is precisely by offering a systematic way of thinking about where international institutions come from and why they take the particular forms they do. For this reason, the constructivist challenge is unlikely to result in the fundamental reorientation of future research that its proponents are seeking. The mainstream perspective may be too one-sided, but its microfoundational underpinnings are nonetheless a major asset. 6 Eichengreen, Frieden, and von Hagan 1995, 6.
R E T H I N K I N G
C O O P E R A T I O N
259
side—the one concerning winners and losers, zero-sum conflict, the struggle to achieve (and maintain) power—we must first discard the analytical biases that have led international relations theorists to consistently overlook it. One source of bias has been their predilection for seeing “states” as the fundamental building blocks of international relations theory. A legacy of Waltzian realism,7 the assumption that states are (in the jargon) “unitary actors” greatly simplifies the explanatory task, allowing the analyst to trace changes in state behavior to corresponding changes in the larger international environment. That said, if our goal is to understand why international cooperation takes the form it does, abandoning the notion of states as internally undifferentiated entities is an absolute necessity. Rather than focusing on the preferences of states (i.e., their “national interests”), I think it makes more sense to talk about the preferences of the particular parties, groups, and individuals who govern them.8 Realists would counter that in a dangerous world, it does not really matter who holds office within the “units” that make up the international system, since all public officials are guided by the same objective: to defend the territories over which they preside against would-be foreign aggressors. The notion that domestic officeholders attach a great deal of importance to security considerations is certainly correct.9 Where realists go wrong is in assuming that because government officials who do not put security at the top of their list of priorities risk popular unrest and military defeat, the potential for significant domestic-level variation in state interests must therefore be trivial. For one thing, it may be possible for these officials to substantially reduce the security threats they encounter—by developing (or acquiring) a nuclear capability, for example.10 And even if they could not, the potential for governmentto-government variation in state interests could still be considerable, for only 7
See esp. Waltz 1979. Cf. Frieden and Rogowski 1996; Ikenberry, Lake, and Mastanduno 1988; Milner 1992 and 1997; Moravcsik 1998; Simmons 1994. 9 Just how single-mindedly public officials seek security is a subject of controversy, with some scholars arguing that “greed” and other motivations can be fully incorporated into realism’s explanatory edifice (see esp. Glaser 1994/95). But while security is surely not the only concern of government leaders, the preservation of political and territorial integrity often does take precedence over other goals. And for good reason: Government leaders who consistently ignore (or mishandle) security affairs do so at substantial political, and sometimes personal, risk. See Londregan and Poole 1990; Bueno de Mesquita, Siverson, and Woller 1992. 10 Realists have a stake in arguing that the minimal conditions for security are in fact quite demanding. The harder it is for states to secure their territorial and political integrity (e.g., because of uncertainty about the capabilities of potential enemies), the longer is realist theory’s explanatory reach. There are clear similarities here to theories of legislative behavior. One might think that incumbent legislators seeking reelection in “safe” districts would devote fewer resources to campaigning and enjoy greater latitude in position taking than incumbents running in districts where the margin of victory has historically been much closer. In fact, say congressional scholars, members of the U.S. Congress can never be assured of reelection; hence all legislators, even those who seem assured of victory, are perpetually “running scared.” Jacobson 1987; cf. Mayhew 1974. 8
260
C O N C L U S I O N S
some of the policies that government leaders pursue have immediate (or politically salient) security externalities.11 Even when it comes to formulating military and defense strategy, a “security-seeking” state may have ample room for discretion.12 For these reasons, it should come as little surprise that transfers in governmental authority can alter—sometimes in quite dramatic fashion—the policy orientations of the countries in which they occur.13 Changes in the polarity of the international system, the offense/defense balance, and other aspects of the international security environment tend to occur gradually, over decades or centuries. And yet “states” have been known to change their preferences quite rapidly—in some cases, overnight. What difference does this make? Quite a lot, actually. One goal of this book has been to try to understand why so many of today’s international treaties explicitly allow for ex post changes in the initial terms of cooperation. An important reason for this flexibility, I have argued, is that the governing officials who draft these treaties anticipate that their political successors might view them unfavorably and thus be inclined to destroy them. Affording their successors some means of moderating the net burden that cooperation imposes upon them is the prime movers’ way of insulating their new regime from the vicissitudes of domestic politics in their own countries. Note, however, that this dynamic is visible only insofar as the analyst allows that states—here, the nations that make 11 While virtually any policy can be shown, through its impact on national prosperity, to affect a country’s ability to defend itself (Grieco 1990; Gowa and Mansfield 1993), in the real world such security spillovers are often so small and attenuated as to be politically irrelevant. A possible exception is Germany’s stance on EMU after the implosion of the Soviet bloc in 1989. By expressing support for monetary union, public officials in Germany hoped to reassure their European counterparts that German reunification would not give rise to a new expansionary impulse (Garrett 1994, 58; see also D. R. Cameron 1992; Woolley 1994). As Garrett notes, however, the economic rationale for Germany’s embrace of EMU was “at least as significant as these security concerns.” Most important, Germany needed increased exports, “and further penetration of EC markets would be the best method of achieving this objective.” Garrett 1994, 57–59. 12 One reason is that systemic pressures do not always militate in favor of a single array of strategic doctrines. In a multipolar world containing four great powers, for example, there is more than one “optimal” balance-of-power arrangement. To say that balancing behavior is adaptive in a security environment where more than one equilibrium is viable is to beg the question: Balance with whom? And even if there were a single “first-best” strategy, there is no guarantee that “boundedly rational” policymakers would be able to identify it ( Jervis 1976). Nor, given their particularistic incentives, would they clearly wish to adopt it if they could. Even if a single, transparent, strategically optimal course of action were to exist, a country’s lawmakers might still choose to subordinate the “national interest” to their own parochial, partisan, or organizational interests (Van Evera 1984; J. Snyder 1991; Posen 1984). 13 That political reaction to the EMS varied widely at the domestic level was clearest in the case of France, where the regime was launched by the governing conservatives over the objections of the opposition Socialists. The Canadian response to free trade with the United States was also marked by internal division, with the Conservative party strongly in favor, the Liberal party strongly against.
R E T H I N K I N G
C O O P E R A T I O N
261
up the enacting coalition—may undergo profound changes in their underlying preferences, changes that may be wholly unrelated to developments taking place in the larger global security environment. Thinking of the actors in international politics as states, or even as particular kinds of states (distinguishing between democratic and nondemocratic states, for example), risks obscuring all of this, deflecting attention from the relationship between the uncertainties engendered by government turnover, on the one hand, and the types of supranational agreements and institutional structures we see emerging across today’s international landscape, on the other. There is, however, an even larger obstacle to theoretical progress. Not only has the literature failed to appreciate the scope for power politics within the countries that make up the enacting coalition; it has also failed to grasp how members of the enacting coalition induce support for their cooperative ventures from governments in other countries. How has this come about? Here, the answer can be traced to a second, even more important source of theoretical bias: the literature’s preoccupation with Mancur Olson’s positive-sum view of collective action, where cooperation—if it can be achieved—is better for everyone. Taking this idea as their starting point, students of international relations have tended to see collective action as something to be applauded and (insofar as possible) encouraged. In the end, however, this way of thinking provides an overly narrow conception of what international cooperation, institution building, and political integration are all about. Without cooperation, we would not necessarily be “lost,” as Keohane asserts in the passage quoted at the beginning of this book.14 To the contrary, some cooperators may be only too happy to return to the original (pre-cooperation) status quo. In fact, there has never been any inherent— that is to say, theoretical—reason for the literature’s positive-sum orientation. While my ideas about power, institutions, and cooperation may not be as familiar as the ones that have framed discussion of these issues in the past, both sets of arguments are firmly grounded in rational choice foundations, and both are equally deserving of our analytical attention.
Beyond Money and Trade: A Brief Look at Some Additional Applications Although the EMS and NAFTA have had a particularly strong impact upon the behavior of public officials—and though, when taken together, they afford considerable breadth in terms of regional and issue coverage—they are hardly the only cases of cooperation and institution building worth studying. My strategy up until now has been to focus on “big” cases where even hard-nosed realists would have to agree that cooperation mattered. These cases involve economic issues—trade, finance, money, and the like. But there 14
Keohane 1988, 393.
262
C O N C L U S I O N S
is no reason to think that the logic of power politics I have been developing applies only to the economic realm and not also to cooperation on matters concerning, say, national security or the protection of the global environment. Nor, by the same token, would it be surprising to find the same forces at work outside of North America and Western Europe—in Africa, for example, or in Asia. Not that the outcomes or motivations would necessarily be the same in every sphere of interstate interaction or across all regions of the world. When thinking about the broader empirical relevance of my arguments about bandwagoning and the dynamics of accession, for example, one would want to pay particular attention to the variables outlined in chapter 3. The cooperative “good” in question will not necessarily look bad to governments outside the enacting coalition. Nor would the “spillovers” produced by the good necessarily be harmful to outsiders—and even if they were, these negative externalities might nonetheless be too small to induce the outsiders to jump on the bandwagon. As with my theoretical claims about go-it-alone power, the propositions about “institutional” or “structural” power set forth in chapter 5 are also ripe for further empirical study. Take the case of Asia, a region where, for the moment at least, the transition from anarchy to organization would seem to be occurring at a decidedly slower pace than it is in other parts of the globe. Why is it that, as Grieco and others have noted, what cooperation there is in the region takes place through “strictly inter-governmental accords with little aspiration to significant forms of supranational authority”?15 The explanation given by Grieco is that Asian countries are uniquely sensitive to “relative gains.” I would offer a different explanation, one that flows directly out of the power-politics logic of institutional design I described in chapter 5. Insofar as Asian regimes have historically lacked the “governance” features that one finds in a NAFTA or an EMS, it is, I would suggest, because the political protection provided by these more flexible governance arrangements was simply not required. After all, many of these regimes were initiated by political elites who enjoyed political stability and encountered little or no significant (i.e., politically salient) domestic opposition at the time of their creation. The lesson here is that the insulation incentives discussed in chapter 5 may be stronger in some parts of the world than in others, in which case one would expect to find corresponding differences in each region’s preference for supranational delegation. In addition to varying by region, however, certain institutional features may also vary by issue or policy area. It is often assumed, for example, that left- and right-wing parties diverge on matters of security policy less than they diverge on questions of economic, social, or environmental policy. If this is correct (and I think it is), it suggests a parsimonious explanation for the greater “completeness” of most security arrange15
Grieco 1997, 169; see also Aggarwal 1995; Crone 1993; and Haggard 1995.
R E T H I N K I N G
C O O P E R A T I O N
263
ments. While other preferences display considerable partisan-induced variation, security preferences remain fairly constant as one moves across the ideological spectrum. As a result, the potential for significant government-togovernment variation in how the “national interest” is perceived tends to be lower in the security realm than in other areas of policy. And because domestic political uncertainty is lower, the attractions of institutional flexibility and delegated authority are correspondingly weaker.16 At first blush, then, the broad theoretical perspective I have set out would seem to capture something essential about international cooperation in a wide variety of settings—not just in Europe and the Americas or on matters concerning trade and monetary policy. To underscore the point, the remainder of this section takes a brief look at a few additional cases. Before getting started, however, I should note that in none of these cases will I be offering the kind of comprehensive analysis I sought to provide in my case studies of the EMS and NAFTA. Ideally, future research would move quickly to fill this gap, examining the applicability of my theoretical analysis to a much wider range of empirical developments than I have had the opportunity to look at here. Nation Building This book has offered a new way of thinking about why, particularly in contemporary Western Europe and the Americas, nations today are eagerly joining forces to establish new, higher-level structures of authority. Although no two cases are exactly alike, the world’s progression from anarchy to organization is best seen as part of a larger pattern or syndrome, the systematic study of which might be termed “political geography.” Just as economic geographers inquire into the spatial location of economic activity,17 the goal of political geographers would be to explain the spatial location of political activity: Why, historically, has political authority tended to migrate upward, with smaller “breakaway” units (the countries of the former Soviet Union, the Czech Republic, East Germany) relinquishing their autonomy to larger units (the Commonwealth of Independent States, the EU, etc.), rather than the larger units devolving into smaller ones? To what extent is this upward migration of authority a “natural” phenomenon, one that proceeds by its own logic, gaining momentum as it moves forward and stoppable only by concerted action on the part of the newly independent units? Are certain allocations of juridical authority more durable than others, and, if so, what accounts for their greater stability? Thus far, my discussion of these issues has been framed in terms of international relations. It is worth noting, however, that each of the nations that 16
This line of analysis suggests an interesting avenue for future research. For recent work on the institutional politics of NATO, see McCalla 1996. 17 Krugman 1991a.
264
C O N C L U S I O N S
make up the international system is itself a locus of political authority, and that the scope and allocation of its authority can strongly influence not only whether it thrives economically or militarily,18 but also whether it inspires the loyalty of a diverse citizenry and so functions effectively as a democracy.19 This is not to suggest that the forces propelling the rise of supranational institutions are exactly the same as those propelling the creation of national, or subnational, structures; the logic of political geography might well differ as one moves from an international to a purely domestic setting. Still, I have to believe that “state-building” is not really all that different from the process of (international) institution building explored in this book. In both cases, the fear of isolation may turn out to be an even more important factor in the integrating party’s decision-making calculus than any expectation of future economic, security, or other gains. I have already examined this possibility as it relates to European integration. Consider, then, the case of American integration: the forging of a new Federal constitution in the late 1780s. The basic facts of the case are not in dispute. Starting in the 1760s, the British parliament began raising taxes on the inhabitants of what were then known as the thirteen colonies. From Britain’s standpoint, these new levies—the Sugar Act (1764), the Stamp Act (1765), the Townshend Acts (1767), the Tea Act (1773)—were perfectly reasonable; the British were simply asking the colonists to help pay for the security umbrella that Britain had been providing them, at little or no cost, for years. The colonists, of course, took a different view. Two years after sending delegates to the First Continental Congress, the Americans, on July 4, 1776, formally declared their independence from the crown. By 1780, though the ensuing Revolutionary War was to drag on for several more years, written state constitutions had been drafted and ratified by each of the original thirteen colonies. The adoption of a national constitution took a bit longer, but in 1781, after a protracted four-year ratification campaign, the so-called Articles of Confederation went into effect, thus laying the foundations for a new central government. So much for background. The interesting question for present purposes is why, less than ten years after this new constitutional regime was established, the Americans uprooted it in favor of a much “deeper”—that is to say, more heavily integrated—system of governance. One theory is that the earlier system just did not work very well. Indeed, a mere two years after the Articles of Confederation went into effect, the Union promptly fell into a severe economic depression. Coinciding with the end of the Revolutionary War, this downturn might well have been temporary—or so, at any rate, James Madison, Alexander Hamilton, and other “nationalists” began arguing in the mid-1780s—but for the failure of Amer18 19
Casella 1992; Friedman 1977; McCallum 1995. Barro 1991; Sandel 1996; also see Alesina and Spolaore 1995 and 1996.
R E T H I N K I N G
C O O P E R A T I O N
265
ica’s new central government to solve what, in today’s lexicon, would be considered a classic collective action problem. There were, in fact, several such problems, each one pitting the (collective) interests of Congress against the (individual) interests of the thirteen newly independent states. There was, for example, the matter of paying off the national debt. Under a more centralized constitutional system, the problem could have been dealt with fairly easily: The central government could have imposed a national tax. As it was, however, the Articles of Confederation barred Congress (there was no executive) from levying taxes of any kind. The result was that the former colonies failed to pay off the financial obligations they had incurred during their just-concluded war with Britain—a failure that caused prospective lenders to lose faith in the American economy and, in so doing, contributed to the new nation’s postwar economic woes. A second problem stemmed from Congress’s inability to impose tariffs. In this case, the effect was to give Britain a decided bargaining advantage in its commercial relations with the United States. No longer part of the British empire, the Americans were nonetheless eager to maintain their preferential access to British markets. The problem was that with Congress unable to retaliate against British trade barriers, Britain was free to play one state off against the other, thereby assuring British exporters more favorable terms of entry into the U.S. market than American exporters enjoyed in Britain. As if that weren’t bad enough, the former colonies responded to the postwar flood of British imports by erecting tariff barriers, not against Britain, but against each other. And Congress, lacking the authority to regulate the flow of goods across state lines, could do nothing to stop it. Low external tariffs, high internal tariffs, a looming debt crisis—not exactly a recipe for economic recovery. Still, while the campaign for constitutional reform waged by Madison, Hamilton, and others may have been in the economic interests of the Union as a whole, the move to supplant the existing, highly decentralized governing apparatus with a new supra-state regime did not necessarily advance the individual interests of the Union’s thirteen member states. This may explain why only five of these states bothered sending delegates to the first constitutional convention organized by Madison and Hamilton, which met during September of 1786 in Annapolis, Maryland, and why their call for a second convention initially fell on deaf ears. Indeed, had it not been for the controversy ignited by Shays’ Rebellion in Western Massachusetts, that might well have been the end of the story.20 On February 21, 1787, however, Congress recommended that the states try again, and this time all but one state—Rhode Island—sent representatives. Meeting in Philadelphia’s town hall, the delegates to the second convention spent from May until September of 1787 devising a new constitution. What did this constitution look like? While Americans today are quick to 20
See, e.g., Wood 1969.
266
C O N C L U S I O N S
extol the virtues of their constitutional system of “checks and balances,” the truth is that the degree of centralization embodied in the document that came out of Philadelphia was far greater than that which Americans had endured—and, indeed, rebelled against—during the colonial era.21 Nor did this remarkable fact go unnoticed at the time. Even before the convention had ended, a number of delegates expressed concern that, if ratified, the new constitution would seriously encroach upon the “sovereignty” that each state enjoyed under the Articles of Confederation. A particular bone of contention was a constitutional provision granting Congress the right to regulate all foreign and inter-state (i.e., domestic) commerce. Among the principal opponents of this provision was New York, a state whose government was at that time being financed almost entirely out of revenues generated by a state tariff on goods shipped through its harbors. Were the Articles of Confederation to be superseded by the constitution drafted at Philadelphia, the new regulatory powers vested in the Congress were sure to be exploited by states such as Connecticut and New Jersey, whose “imports mostly passed through the port of New York”22 and, in the end, New York would almost certainly be forced to share its tariff revenues with the twelve other states in the Union.23 In the ensuing debate between Federalists, who supported ratification, and Anti-Federalists, who opposed it, the vast majority of New Yorkers sided with the Anti-Federalists. In the end, however, the state’s political leaders agreed to adopt the new constitution. “What makes the Federalist achievement in New York so impressive,” writes William Riker in his well-documented, book-length investigation of the ratification campaign, “is that they did not have electoral control either in the state government or in the convention.”24 To what, then, did the Federalists owe their New York victory?25 Riker’s own account gives credit to their “election-endowed control of the agenda.”26 To understand the logic behind his argument, one need simply envision a continuum of alternative possible constitutions, with the Articles of Confederation—that is, the status quo, SQ—located at the extreme left and unitary government, or U, at the extreme right (Figure 10-1). Riker assumes that the ideal outcome from the standpoint of the Union’s median “voter” would have been a constitution situated on this continuum 21
Riker 1996, 18. Riker 1996, 178. 23 This may account for the historical strength of New York’s opposition to the nationalist program embraced by such leaders as Madison and Hamilton. Five years before the “framers” convened in Philadelphia, New York played a key role in sinking an amendment to the Articles of Confederation that would have allowed Congress to impose a tax on import duties. 24 Riker 1996, 230. 25 At their convention in Philadelphia, the founders stipulated that ratification by nine of the thirteen states would be needed in order for the new constitution to take effect. On June 21, 1788, New Hampshire became the ninth state to do so. Five days later, Virginia became the tenth state to ratify, followed, on July 26, by New York. See below. 26 Riker 1996, 18. 22
R E T H I N K I N G
C O O P E R A T I O N
267
Figure 10-1. Riker’s Spatial Representation of Alternative Possible Constitutions
between SQ and U at the point labeled m. In other words, a majority-minusone of the new nation’s voting public would have preferred a constitution “located” between m and SQ, and another majority-minus-one would have preferred a constitution between m and U. That being the case, the framers could have been assured of winning a majority for constitution C1 —that is, a constitution positioned to the left of m (but to the right of SQ)—if they had adopted it. But because the framers themselves preferred a constitution located to the right of the median voter’s “ideal point,” they did not adopt C1. Instead, they adopted C2, a constitution situated to the right of m (and therefore closer to their own ideal point, U), but not so far to the right that the median voter would be better off with the status quo constitution, SQ. The result, argues Riker, is that each state was forced to choose between SQ (the Articles of Confederation) and C2 (the constitution drafted by the delegates at the Philadelphia convention). If these had in fact been the two choices, there is every reason to think that New York would have rejected the constitution drafted by the delegates at Philadelphia. That a state as avowedly Anti-Federalist as New York ultimately voted to accept the new constitution is due precisely to the fact that one of the two alternatives—namely, the status quo—was no longer in New York’s choice set when it came time for it to vote. Though the notion of a constantly “shifting” status quo does not fit comfortably within Riker’s spatial framework (Figure 10-1), this concept lurks just beneath the surface of the story he tells in the empirical chapters of his book. Like other students of the ratification campaign,27 Riker attaches enormous significance to the temporal nature of the process. Rather than settling the constitutional question on the basis of a single, nationwide referendum, the Framers stipulated that the issue be resolved sequentially—that is, state by state—in a campaign that ended up lasting nearly three years, from September 28, 1787, the day Congress initially submitted the Federal constitution to the states for approval, until May 29, 1790, when Rhode Island became the last of the holdouts to vote (by the slimmest of margins) in favor of ratification. To describe the ratification campaign as a state-by-state affair is not to say that all states counted equally. To the contrary, Riker discerns a clear “tipping point” in the campaign. In Massachusetts, he writes, the Federalist victory “made the difference between ratification and rejection for the entire 27
See esp. Wood 1969.
268
C O N C L U S I O N S
country.”28
What Riker never satisfactorily explains, however, is why the Massachusetts victory proved to be the campaign’s defining moment. Let me offer a simple explanation: it shifted the status quo. Indeed, after Massachusetts voted to ratify on February 7, 1788, most observers viewed the Articles of Confederation as a dead letter. It is true that, as mentioned earlier, the new constitution could not begin operating until nine states had agreed to adopt it, and Massachusetts was but the sixth state to do so. After the Massachusetts vote, however, the contest between the Federalists and Anti-Federalists was effectively over, for even if all of the remaining states had gone on to reject the new constitution, the initial group of ratifiers was fully able, by virtue of its combined wealth and resources, to go it alone. For that reason, the status quo that New York and other states outside this group had to consider when deciding whether or not to support C2 was not SQ — a union governed by the Articles of Confederation—but a new and, in the eyes of the recalcitrant states, greatly inferior alternative—a “dismembered” union in which the holdouts went their separate ways while (at least) six states, including ones as large and economically important as Pennsylvania, New Jersey, Connecticut, and Massachusetts, adopted a common, overarching system of governance.29 Riker’s account of the dilemma facing New York State’s Anti-Federalist coalition is worth quoting at some length.30 It is easy to understand why the Anti-Federalists split. The thirty persons in the majority for ratification [the final vote was 30-27] consisted of eighteen Federalists . . . and twelve Anti-Federalists. The Federalists were from the counties in and around New York City. . . . This was the area that would lose international trade if New York separated. It would also lose the national capital, which was thought to bring in about one hundred thousand pounds per year in hard money. And, most of all, it was the area of potential civil war. The twelve Anti-Federalists were from counties in a wider circle around New York City. . . . In the very first speech of the convention Chancellor Livingston [a Federalist] emphasized the inability of New York State to defend Long Island and Staten Island from the potentially unfriendly powers, Connecticut and New Jersey. He also pointed out that the surrounding states had no interest in allying themselves with New York, which was a broad 28
Riker 1996, 183. The term “dismemberment” was used by Alexander Hamilton in Federalist Paper No. 13 to describe the outcome that would result should the new constitution fail to be ratified by all thirteen states: “The ideas of men who speculate upon the dismemberment of the empire seem generally turned toward three confederacies—one consisting of the four Northern, another of the four Middle, and a third of the five Southern States. . . . According to this distribution, each confederacy would comprise an extent of territory larger than that of the kingdom of Great Britain. No well-informed man will suppose that the affairs of such a confederacy can be properly regulated by a government less comprehensive in its organs or institutions than that which has been proposed by the [Philadelphia] convention.” Madison, Hamilton, and Jay [1788] 1987, 139. 30 See also Kaminski 1985 and 1988. 29
R E T H I N K I N G
C O O P E R A T I O N
269
hint that the southern counties would have considerable interest in secession and alliance with the neighboring states. So throughout the convention these Anti-Federalists from the south had to face the question of their future in an isolated New York State, a future that might include secession and civil war, realistic concerns considering that New York had, only a few years earlier, lost Vermont.31
Under the circumstances, then, it is not at all surprising that New York, bastion of Anti-Federalist sentiment though it was, voted in favor of the new constitution. New Yorkers did not like the new constitution to which they agreed to adhere. Quite the contrary, the state’s government would have been perfectly happy to continue charging other states a fee for the “privilege” of using New York ports. Once it was clear that New York was in the position of our actor D, however, and that Massachusetts (actor C) was fully prepared to join Pennsylvania, New Jersey, Connecticut, and Georgia (the A’s and B’s) in forging a new nation, it would have been foolish for New Yorkers not to jump on board the Federalist bandwagon. True, it took some time for the nature of their predicament—the fact that the status quo had been removed their choice set—to sink in.32 In the end, however, the state of New York followed Virginia in (narrowly) approving the Federalist project on July 26, 1788. By May of 1790, the only remaining holdouts—North Carolina and Rhode Island—had also come aboard, making the ratification unanimous. Security Cooperation I want to begin the discussion of my next case by reiterating my earlier point about the conditionality of go-it-alone power. Even if most real-world cases of “zero-sum” cooperation involve the use of go-it-alone power, this is not the only logical possibility. In some cases, the winners may be forced to exert their will over the losers more directly—through good, old-fashioned coercion. One such case, the Chemical Weapons Convention of 1997, has the added attraction—methodologically speaking—of not being a case of economic cooperation. This is desirable in that it affords yet another opportunity for addressing the generalizability of my theoretical argument: How, if at all, might the power-politics logic of economic cooperation differ from that of security cooperation—and of global disarmament in particular? Signed by 130 nations at a Paris summit on January 13, 1993, the Chemical Weapons Convention (CWC) is a multilateral treaty aimed at prohibiting the development, production, acquisition, stockpiling, transfer, and use of chemical weapons by the year 2007.33 The United States was among the earliest and staunchest supporters of the CWC, which officially entered into 31
Riker 1996, 239. Cf. Riker 1996, 235. 33 In addition to calling for permanent elimination of an entire class of weapons of mass destruction, the CWC also established a highly complex and intrusive verification regime. For a detailed analysis of the treaty, see Morel and Olson 1993. 32
270
C O N C L U S I O N S
force in 1997, for the treaty promised not only to diminish the threat of chemical warfare but to reduce international terrorism as well. There was, however, one big problem: these potential benefits could be realized only if chemical-weapons-producing “rogue nations” such as Iran and China agreed to participate in the new regime. The CWC thus presents a case of multilateral cooperation in which the United States and other prospective beneficiaries could not as easily threaten to go it alone. While proceeding unilaterally remained an option for this group, the cooperative gains accruing to the “winners” were contingent on many, if not all, of the CWC “losers” faithfully upholding the terms of a weapons ban whose very existence they would normally have opposed. Unable to rely upon their go-it-alone potential, members of the CWC’s enacting coalition took a different tack, directly threatening the losers with economic sanctions. It is in this light that the convention’s express prohibition against all trade in “military” chemicals between member and nonmember countries should be understood. Included in the list of banned substances were “dual use” chemicals ranging from plastics and pesticides to ceramics and pharmaceuticals—even common, everyday products like ink. In denying nonmember countries access to such a wide range of chemical substances, it was as if the sponsors of the CWC had established permanent economic sanctions against these countries.34 This, argues one commentator, was the CWC’s innovative genius: Because of the flexibility of the chemical technology, the treaty’s punishment by denial of “military” chemicals amounts to broad and immediately painful sanctions against the civilian economy. And these sanctions are a good reason not just to stay in compliance, but to sign the treaty in the first place. . . . [The CWC] would make the manufacture of chemical weapons an endeavor with a significant risk of unmasking, and unmasking would bring painful penalties—penalties that no Security Council member would have the chance to veto.”35
All of this goes a long way toward explaining the impressive scope, and resulting success, of the CWC regime. Indeed, of all those countries that had, or were suspected of having, chemical weapons programs at the time the convention was being negotiated in the early 1990s—the United States, Russia, Iraq, and about twenty others, including North Korea, India, Pakistan, Libya, Iran, China, and Israel—only Iraq and Libya remained wholly outside the convention’s institutional framework when the CWC officially entered into force on April 29, 1997.36 Ultimately, one should not be surprised if all nations, even ones as intransigent as Iraq and Libya, decide to come 34 Wishing these sanctions to take effect as soon as possible, the architects of the CWC set a low ratification threshold, stipulating that only 65 ratifiers would be necessary for the treaty to take effect. 35 Wright 1995, 22. 36 Though all were signatories to the original treaty, the United States, Iran, China, and Russia had not yet ratified the convention when it entered into force.
R E T H I N K I N G
C O O P E R A T I O N
271
aboard. Not only would holdouts who refused to sign the agreement risk incurring the wrath of the international community, accepting the social stigma associated with their “outlaw” status, but thanks to the CWC’s in-built sanctioning device—a provision necessitated by the absence of the standard goit-alone means of inducing “support” for the regime—nonparticipants would also have to bear a more concrete burden: the prospect of continued exclusion from many of the world’s most important international markets. Environmental Diplomacy To what extent can international institutions influence the underlying preferences and values of their member governments? Over the last several years, the preference-altering potential of international agreements has emerged as a topic of considerable interest among political scientists and legal scholars alike. Thus far, however, virtually all of this work has focused on just one type of agreement: those aimed at protecting the global environment.37 This is no accident. As an empirical matter, the preference-transforming effects of international trade, monetary, and security regimes are just not as large as those produced by environmental regimes. As for why the “transformational” properties of economic and arms control agreements do not play as fundamental a role, let me propose a simple answer: the beneficiaries of these agreements—the winners—do not need them to.38 As a vehicle for making this point, the remainder of this section explores the institutional politics of one particular agreement: the Montreal Protocol on Substances that Deplete the Ozone Layer. Signed in 1987, this landmark environmental treaty limits the production and sale of ozone-depleting chlorofluorocarbons (CFCs) by all member states.39 Detailed accounts of the negotiations leading up to the passage of the protocol make it clear that the United States, Canada, Sweden, Norway, and (belatedly) West Germany were the prime movers behind the Montreal initiative. The 1985 discovery of a hole in the ozone layer energized large 37 See, e.g., Choucri 1993; P. Haas 1990; Mitchell 1994; Ostrom 1990; Sebenius 1991; O. Young 1994; and Young and Osherenko 1993. 38 My claim here is that compared with economic and security institutions, environmental institutions have a stronger and ultimately more consequential impact on the underlying values of the government actors who participate in them. That said, economic and security institutions almost certainly have some value-altering “transformational” impact. Ikenberry and Kupchan (1990) and Katzenstein (1996) discuss the transformational properties of a number of different collective security and arms control arrangements. For an inquiry into the preference-transforming properties of an economic arrangement—the European Community—see Sandholtz 1993. Building upon the “constructivist” analyses of Ruggie (1983), Wendt (1992), and others, Sandholtz tries to show that Community decisions reflect state interests, but that “those interests are shaped in part by membership in the EC.” Sandholtz 1993, 3. 39 By 1992, some eighty-seven governments—plus the EC—had formally agreed to the terms of the treaty. For more on the protocol and its origins, see Benedick 1991; Bryk 1991; P. Haas 1992; Morrisett 1989; and Parson 1993.
272
C O N C L U S I O N S
portions of the voting public in all of these countries, intensifying pressures on their respective governments to impose an outright ban on CFCs, ozonedepleting substances which were at that time widely employed as refrigerants, solvents, and cleaning agents. Identifying the losers in the environmental regime established by the protocol is not easy. Environmental treaties are not the sort of things government leaders like to complain about—at least not openly—and in the long run, of course, such treaties promise “utility gains” for the world’s entire population. It is also true that technological advances introduced since the Montreal agreement went into effect have greatly reduced the overall burden of compliance. Attention has now turned from the regulation of CFCs, which were completely phased out in 1996, to that of the ozone-depleting pesticide methyl bromide.40 All that said, the fact remains that governing officials in the United Kingdom and France, as well as in a number of developing countries including China and India, went to extraordinary lengths to maintain the ozoneunfriendly, pre-Montreal status quo.41 Notwithstanding the treaty’s longterm environmental benefits, government officials in these countries did not relish the prospect of losing their right to decide which substances (if any) needed to be controlled or how those controls would be administered. True, in countries with strong grass-roots environmental organizations, public opinion was often highly sensitive to environmental issues. In other countries, however, environmental issues were regarded as more of a nuisance; they were obstacles to the pursuit of other, more pressing objectives. This was certainly the case in France and Britain, two countries whose domestic CFC manufacturers—Atochem in France, Imperial Chemical Industries (ICI) in the UK—enjoyed unusually close ties to government decision makers.42 Both companies were major suppliers to the rapidly growing markets of the developing world, and by the mid-1980s Atochem and ICI were exporting nearly half of their total output. In contrast, DuPont and other U.S.-based manufacturers exported hardly any of the CFCs they produced. As for public opinion, relatively few citizens in either France or Britain viewed ozone depletion as an urgent matter requiring immediate action. Instead, attention focused on more immediate, closer-to-home environmental problems such as acid rain, chemical spills, and the 1986 Chernobyl nuclear power plant accident.43 Indeed, even after the ozone hole’s discovery, doubts remained within Europe’s—and especially Britain’s—scientific community as to whether the ozone layer was in fact deteriorating and, if so, whether CFCs were the culprit.44 This skepticism was shared throughout much of the developing world as 40
Stevens 1993. See esp. Benedick 1991; Jachtenfuchs 1980. 42 Benedick 1991, 39; Jachtenfuchs 1980. 43 Benedick 1991, 28. 44 Benedick 1991, 39; also see P. Haas 1992, 210; Roan 1990, 102– 3. 41
R E T H I N K I N G
C O O P E R A T I O N
273
well. If anything, the developing countries were even more hostile to Montreal’s anti-CFC agenda. After all, air conditioning, food preservatives, and other products containing CFCs had played an integral role in the economic development of the “North.” Why should the “South” have to forgo these necessities? Nor was the alternative—asking developing countries to purchase CFC substitutes—any less unacceptable, since the effect (as the Chinese and Indian delegates to a meeting held shortly after the Montreal protocol’s entry into force were quick to point out) would be to “enrich the very chemical industries that had created the ozone problem in the first place.”45 In short, many developing countries regarded access to CFCs less as a luxury than as a basic right or entitlement. At stake, in their view, were fundamental moral issues—issues concerning equity, justice, and fair play—rather than simply the economic and environmental issues on which the protocol’s champions had tended to focus. To be sure, the U.S.-led coalition of countries in favor of banning CFCs was perfectly capable of proceeding unilaterally. Like supporters of the Chemical Weapons Convention, however, supporters of the Montreal Protocol had to contend with a “holdup” problem (see chapter 3): As much as they stood to benefit from a substantial reduction in the emission of harmful, ozone-depleting substances, the would-be beneficiaries lacked the ability to supply this “good” by themselves. If, for example, the UK and France—or, even worse, these two countries along with a large, populous Third World country such as China or India—had refused to coordinate their policies in accordance with the protocol (or had agreed to do so but then later reneged), the protocol’s costly restrictions on CFC emissions would have had little effect in combating the ozone depletion problem. Thwarted by a counter-coalition of governments opposed to environmental cooperation, members of the activist coalition would have failed to achieve the goals they set out to accomplish. Analytically, this holdup problem is what most clearly differentiates the Montreal Protocol and Chemical Weapons Convention from the other cases we have been discussing. In cases where the losers are in a position to block the provision of a particular “good” (because the enacting coalition cannot supply it unilaterally), members of the enacting coalition will naturally look for ways of compensating for their limited go-it-alone potential. Coercion is the alternative that comes most readily to mind; and in some cases—the chemical weapons treaty, for example—the threat of sanctions may well be sufficient to get the losers to cooperate (however reluctantly) with the winners. But what if, confronted by a pivotal coalition of holdouts, the winners are unable to exert either go-it-alone or coercive pressure? The Montreal case bears directly on the issue—first, because the utility gains derived by the treaty’s environmentally progressive sponsors were contingent upon their being joined by other nations (France, the UK, etc.); sec45
Benedick 1991, 124.
274
C O N C L U S I O N S
ond, because the governments who stood to benefit from the ozone treaty could not credibly threaten to exclude their opponents from the new regime; and third, because they could not force their opponents to comply either. Though the architects of the protocol could prohibit signatories from importing CFCs from companies located in nonparticipating countries, the effects on ICI (Britain’s leading CFC manufacturer) and Atochem (France’s producer) were unlikely to be very great, since both companies enjoyed large domestic markets. Nor was there much hope of using trade restrictions to deter startup CFC producers in developing countries such as China and India, considering the low cost of CFC technology and (again) the size of these countries’ potential domestic markets.46 Limited in their go-it-alone potential and yet unable to threaten nonparticipating governments coercively, the prime movers behind the ozone treaty bent over backward to accommodate the concerns of the losers. Thus, for instance, they declined to impose an immediate ban on CFCs, instead requiring only a 20 percent reduction in the 1986 consumption and production rates by 1994, and another 30 percent by 1998.47 The protocol’s sponsors further stipulated that a permanent supranational governing body would meet at least once every four years to review, assess, and—pending evidence that the economic costs of compliance were proving too demanding for some signatories—selectively relax the control measures outlined in the original treaty. But that was only the beginning. In addition to accommodating the concerns of pivotal holdout countries, the enacting coalition also sought to alter these concerns—by, for example, sensitizing their populations to the dangers of CFCs and encouraging the development of affordable CFC substitutes (thereby weakening the political clout of CFC-producing industries). It is true that the prime movers behind the EMS, NAFTA, and CWC initiatives also sought to manipulate the underlying preferences of their opponents. These efforts at consciousness raising were not nearly as intense or systematic, however, since the beneficiaries of these other arrangements had the ability to shift the status quo unilaterally, a luxury not enjoyed by supporters of the Montreal Protocol. This difference can account for several unusual features of the Montreal story. One has to do with the extraordinary intensity of the media campaign that the coalition of anti-CFC countries waged throughout the negotiations. The United States government, in particular, “undertook major efforts to reach out to foreign public opinion.”48 In addition to “senior U.S. officials and scientists [giving] speeches, press conferences, and radio and television interviews in numerous foreign capitals,” the head of the U.S. Environmen46
Benedick 1991, 151. Actually, these reduction amounts were meant to apply only to industrialized countries. Article 5 of the protocol entitled developing-country members some 10 to 15 percent leeway in these figures, as well as a ten-year extension, from the year 2000 until 2010, for implementation of the total phaseout. 48 Benedick 1991, 56. 47
R E T H I N K I N G
C O O P E R A T I O N
275
tal Protection Agency, the Secretary of State, and on numerous occasions (e.g., the G-7 summit of 1987) the president himself all used “high-level personal diplomacy” to convince their counterparts in other countries that ozone depletion was indeed a serious cause for concern.49 The fact that Montreal’s anti-CFC coalition had a special stake in “transforming” the preferences of their opponents is also reflected in the treaty itself. For example, the protocol explicitly provides for regular get-togethers of the world’s foremost environmental scientists. In addition to guaranteeing worldwide media exposure, these international conferences have lent the ozone-depletion problem a legitimacy it would not otherwise have enjoyed. And as citizens throughout the world have been sensitized to the issue, so too, predictably, have the domestic political authorities who represent them. As for countering the offsetting political influence of such well-entrenched CFC-producing firms as Britain’s ICI and France’s Atochem, the sponsors of the protocol went out of their way to promote the development of inexpensive CFC substitutes. Thus, “key international groups of technical experts [were brought together] to develop new non-CFC solutions in specific applications,” an effort that culminated in a proliferation of “new technical solutions and new business opportunities.”50 In sum, the cooperative arrangement established by the Montreal Protocol on the Protection of the Ozone Layer offers yet another example of a major international regime whose original membership encompassed both winners and losers. At the same time, however, the particular power-politics dynamic underlying this case differs in subtle, and interesting, ways from the power-politics model set forth in chapter 3. Recognizing their weak position vis-à-vis potential holdout countries, the prime movers behind the protocol took special efforts to alter the fundamental preferences of their (initially) ozone-insensitive counterparts in Britain, France, and the developing world. Again, this emphasis on consciousness raising and persuasion was a matter of necessity. If the champions of the protocol had been able either to go it alone (as in the EMS, NAFTA, and Federalist cases) or, failing that, to apply coercive pressure directly against nonparticipating governments (as was the case in the Chemical Weapons Convention), opponents of the Montreal regime would have been likely to accede to it despite their antienvironmentalist leanings.
A New Face of Power? The relative weakness of the go-it-alone dynamic in the arms-control and environmental cases just discussed represents an important contrast to the nation-building example looked at earlier, a contrast that is certainly worth 49 50
Benedick 1991, 56– 57; see also Parson 1993, 19. Parson 1993, 68; cf. Sprinz and Vaahtoranta 1994.
276
C O N C L U S I O N S
exploring. But are the multilateral chemical weapons and CFC agreements really so different from the constitutional agreement negotiated between the thirteen newly independent American states? From the standpoint of international relations theory, what is most striking about the chemical weapons and CFC treaties is not that the winners employed a different means of eliciting the support and compliance of the losers. It is that—as with the EMS, NAFTA, and Federalist examples—only some of the actors involved in these agreements regarded them as good things. There were winners in all of these cases. But there were also losers: governments, states, or nations whose leaders would never have “cooperated” had their neighbors not presented them with a fait accompli. Again, I do not mean to imply that all cases of international collaboration and institution building look like this; I am not suggesting that cooperation is always and everywhere about power and domination. Nor, insofar as power is involved, must it always operate in the hidden ways I have suggested. Sometimes—as in the chemical weapons and CFC cases—the winners’ power over the losers will be overt. Here, once again, there is a parallel to the debate I referred to earlier between proponents of pluralist theory such as Robert Dahl, on the one hand, and critics such as E. E. Schattschneider, Peter Bachrach, and Morton Baratz, on the other. The critics never claimed that Dahlian (i.e., overt) power could not exist or was somehow irrelevant. Their point was simply that there were other, less transparent—but not necessarily any less effective—ways of exercising power, and hence that power was potentially even more pervasive than Dahl and his followers, having embraced an overly narrow theoretical conception, had been led to believe. To be sure, there are differences between the understanding of power presented in this book and the one advanced by Dahl’s critics. For one thing, Dahl’s critics tended to assume that “creating or reinforcing social and political values” was the primary means by which the powerful actor exerted its will.51 As we have seen, however, changing the underlying preferences of weaker actors may not actually be necessary for the stronger one to get its way. Indeed, in neither the EMS nor the NAFTA case study did we find evidence of an emerging ideational consensus, and yet the Franco-German and Canadian-American “enacting coalitions” were, in each case, still able to act as de facto agenda setters, removing the status quo ante from the range of choices available to other actors. There are other differences, as well. For example, the agenda setters in Bachrach and Baratz’s model use their agendacontrol power to keep new legislative initiatives from coming up for a vote.52 Their objective, in other words, is to preserve the initial status quo and not (as it is for the agenda setters in my model) to remove it from the set of feasible alternatives. 51 52
Bachrach and Baratz 1962, 948; cf. Lukes 1974. Cf. Shepsle and Weingast 1981.
R E T H I N K I N G
C O O P E R A T I O N
277
These differences aside, however, the basic point that Bachrach and Baratz were making—that power can influence political behavior (in fact, all behavior) even when actions appear uncoerced and voluntary—is certainly correct. And though applications of Bachrach and Baratz’s insights have until now concentrated on decision making within domestic political settings, these same insights apply with a vengeance to the world of contemporary international relations—a world where, on the surface at least, the use of intimidation and bullying is no longer as prevalent as it once was. What I have been calling go-it-alone power is really just another form of agenda control, that is, where one actor or group exerts its will by circumscribing the choice sets of other actors or groups. Casting the issue in this way accentuates the essential similarities between the thesis of this book and a broader set of arguments concerning the strategic—or, in some cases, nonstrategic (i.e., inadvertent)—manipulation of alternatives. In closing, then, it is important to be clear about what I am not arguing. Even if I am right, it does not mean that the collective-action paradigm developed by Olson, Keohane, Krasner, Grieco, and other scholars is therefore useless or invalid. I am certainly not advocating that we discard this paradigm. The power-politics perspective elaborated in these pages and the more familiar “Olsonian” paradigm are, in fact, two sides of the same coin. It’s just that the importance of the first side—and its integral connection to the second—have not always been fully appreciated, or sometimes even acknowledged, by contributors to the mainstream literature. As a result, commentators on international politics have spent much of the past fifteen years concentrating on aspects of international cooperation and institution building that may not in fact be the most pervasive or important. Fortunately, redressing this imbalance does not mean starting from scratch. As I have noted at various points throughout this book, the analytical foundations of the approach I have sketched out here are exactly the same as those upon which the mainstream theory has been constructed, their differing logics derived from the same (rational-choice) premises. But while its theoretical underpinnings may be similar, the analytical orientation elaborated in this book does offer a strikingly different way of thinking about international cooperation, and of understanding international politics more generally. Most importantly, it directs our attention to the politics of winning and losing, a subject that has somehow managed to escape the careful scrutiny of previous scholars, neoliberal optimists and realist skeptics alike. This book calls, then, not merely for synthesis but rather for a shifting of the theoretical center of gravity—a reorientation that will require: • inquiring more deeply into the nature and dynamics of what I have termed go-it-alone power and, more generally, into the possibility that voluntary cooperation—whether it takes place in an international or purely domestic setting—has as much to do with power, co-optaton, and outright domination as it does with bargaining or strategic interaction;
278
C O N C L U S I O N S
• reassessing the welfare consequences of NAFTA and the EMS, and grappling with the broader question of whether certain regional or multilateral institutions may actually be wreaking havoc on the domestic economic and political systems of their member states, even, perhaps, those governed by “enthusiastic” supporters of these institutions; and • exploring how the underlying logic of all of this might play out differently as one moves beyond NAFTA and the EMS to other issue areas, geographic regions, and historical periods. There is, of course, a great deal more work to be done on each of these fronts. In the end, however, the payoff is likely to be substantial, for the issues raised by this sort of analysis are of more than purely theoretical interest. Indeed, as the dust left behind by a collapsing Soviet empire settles and world leaders set about the task of fashioning a new global economic and security order, the work of analyzing the recent trends toward delegated authority and political integration has taken on special significance. Are we now witnessing the early stages of a full-scale transformation from international anarchy to supranational organization? If so, how will this transformation affect the “utilities” of individual citizens, governments, and nations in the coming years and decades? Are today’s integrationists jumping on the multilateral bandwagon because they expect to be made better off? Or is it simply because they know their neighbors are in a position to proceed without them and so, under the circumstances, they have no (better) choice? My hope is that in laying the analytic foundations for a broader (albeit less sanguine) theoretical perspective, this book will encourage others to take a fresh look at these important questions. Should this happen, I am convinced it will only be a matter of time before we arrive at a truly “grand theory” of international cooperation: a theoretical edifice large enough to incorporate and explain all aspects of interstate collaboration—the good, the bad, and the downright ugly.
B I B L I O G R A P H Y
Abbott, Kenneth W., and Duncan Snidal. 1998. “Why States Act through Formal International Organizations.” Journal of Conflict Resolution 42: 3–32. Aggarwal, Vinod K. 1995. “Comparing Regional Cooperation Efforts in the AsiaPacific and North America.” Pp. 40– 65 in Andrew Mack and John Ravenhill, eds., Pacific Cooperation: Building Economic and Security Regimes in the Asia-Pacific Region. Boulder, Colo.: Westview Press. Aguilar Zinser, Adolfo. 1993. “Authoritarianism and North American Free Trade.” In Ricardo Grinspun and Maxwell A. Cameron, eds., The Political Economy of North American Free Trade. New York: St. Martin’s Press. Alesina, Alberto. 1987. “Macroeconomic Policy in a Two-Party System as a Repeated Game.” Quarterly Journal of Economics 102: 651–78. ———. 1989. “Politics and Business Cycles in Industrial Democracies.” Economic Policy 8: 57– 98. Alesina, Alberto, and Howard Rosenthal. 1995. Partisan Politics, Divided Government, and the Economy. Cambridge: Cambridge University Press. Alesina, Alberto, and Enrico Spolaore. 1995. “On the Number and Size of Nations.” Working Paper no. 5050. Cambridge, Mass.: National Bureau of Economic Research. Alho, Kari. 1995. “The Economic Consequences of EU Enlargement for the Entrants: The Case of Finland.” Pp. 240 – 60 in Richard Baldwin, Pertti Haaparanta, Jaakko Kiander, eds., Expanding Membership of the European Union. Cambridge: Cambridge University Press. Alt, James. 1985. “Political Parties, World Democracy, and Unemployment.” American Political Science Review 79: 1016 – 40. Alter, Alison. 1993. “The Politics of Structural Funds in Italy.” Stanford University. Manuscript. Alvarez, R. Michael, Geoffrey Garrett, and Peter Lange. 1991. “Government Partisanship, Labor Organization, and Macroeconomic Performance.” American Political Science Review 85: 539 – 56. Anderson, Kym, and Rod Tyers. 1995. “Implications of EU Expansion for European Agricultural Policies, Trade and Welfare.” Chap. 8 in Richard Baldwin, Pertti Haaparanta, Jaakko Kiander, eds., Expanding Membership of the European Union. Cambridge: Cambridge University Press. Andrews, David M. 1994. “Capital Mobility and Monetary Adjustment in Europe, 1973 –1991.” Policy Sciences 27: 425 – 45. Antola, Esko. 1991. “Finland.” Chap. 11 in Helen Wallace, ed., The Wider Western Europe: Reshaping the EC/EFTA Relationship. London: Pinter Publishers. Artis, Michael J. 1987. “The European Monetary System: An Evaluation.” Journal of Policy Modeling 9: 175 – 98. Artis, Michael J., and Mark P. Taylor. 1988. “Exchange Rates, Interest Rates, Capital Controls, and the European Monetary System: Assessing the Record.” Pp. 185 – 210 in Francesco Giavazzi, Stefano Micossi, and Marcus Miller, eds., The European Monetary System. Cambridge: Cambridge University Press. Axelrod, Robert. 1984. The Evolution of Cooperation. New York: Basic Books.
280
B I B L I O G R A P H Y
Axelrod, Robert, and Robert O. Keohane. 1985. “Achieving Cooperation under Anarchy: Strategies and Institutions.” World Politics 38: 226–54. Bachrach, Peter, and Morton S. Baratz. 1962. “Two Faces of Power.” American Political Science Review 56: 947– 52. Baer, M. Delal. 1991. “North American Free Trade.” Foreign Affairs 70, no. 4 (Fall): 132– 49. Baer, M. Delal, and Sidney Weintraub. 1994. “The Pressures for Political Reform in Mexico.” Chap. 7 in Baer and Weintraub, eds., The NAFTA Debate: Grappling with Unconventional Trade Issues. Boulder, Colo.: Lynne Rienner Publishers. Baldwin, David A. 1985. Economic Statecraft. Princeton: Princeton University Press. ———, ed. 1993. Neorealism and Neoliberalism: The Contemporary Debate. New York: Columbia University Press. Baldwin, Richard E. 1995. “A Domino Theory of Regionalism.” Chap. 2 in Richard Baldwin, Pertti Haaparanta, Jaakko Kiander, eds., Expanding Membership of the European Union. Cambridge: Cambridge University Press. Baldwin, Robert E. 1976. “The Political Economy of Postwar U.S. Trade Policy.” The Bulletin 4. New York: New York University, Center for the Study of Financial Institutions. ———. 1985. The Political Economy of U.S. Import Policy. Cambridge: MIT Press. Barlow, Maude. 1990. Parcel of Rogues: How Free Trade Is Failing Canada. Toronto: Key Porter Books. Barnard, Chester I. 1938. The Functions of the Executive. Cambridge: Harvard University Press. Barro, Robert J. 1991. “Small Is Beautiful.” Wall Street Journal, October 11. Barro, Robert J., and David Gordon. 1983. “Rules, Discretion, and Reputation in a Model of Monetary Policy.” Journal of Monetary Economics 12: 101–21. Bates, Robert. 1981. Markets and States in Tropical Africa. Berkeley: University of California Press. Bates, Robert H., Avner Greif, Margaret Levi, Jean-Laurent Rosenthal, and Barry R. Weingast. 1998. Analytic Narratives. Princeton: Princeton University Press. Baumgartner, Tom, Walter Buckley, and Tom R. Burns. 1975. “Meta-Power and Relational Control in Social Life.” Social Science Information 14: 49 –78. Baumgartner, Tom, and Tom R. Burns. 1975. “The Structuring of International Economic Relations.” International Studies Quarterly 19: 126–59. Bayard, Thomas O., Joseph Pelzman, and Jorge Perez Lopez. 1983. “Stakes and Risks in Economic Sanctions.” World Economy 6: 73–88. Bayoumi, Tamin, and Barry Eichengreen. 1993. “Shocking Aspects of European Monetary Union.” Pp. 193–229 in Francisco Torres and Francesco Giavazzi, eds., Adjustment and Growth in the European Monetary Union. Cambridge: Cambridge University Press. Bendor, Jonathan. 1987. “In Good Times and Bad: Reciprocity in an Uncertain World.” American Journal of Political Science 31: 531–58. Benedick, Richard Elliot. 1991. Ozone Diplomacy: New Directions in Safeguarding the Planet. Cambridge: Harvard University Press. Bhagwati, Jagdish. 1991. The World Trading System at Risk. Princeton: Princeton University Press. Bhagwati, Jagdish, and Arvind Panagariya. 1996a. “Preferential Trading Areas and Multilateralism: Strangers, Friends, or Foes?” Pp. 1–78 in Jagdish Bhagwati and Arvind Panagariya, eds., The Economics of Preferential Trading Agreements. Washington, D.C.: American Enterprise Institute Press.
B I B L I O G R A P H Y
281
———. 1996b. “The Theory of Preferential Trade Agreements: Historical Evolution and Current Trends.” American Economic Review 86: 82–87. Bhagwati, Jagdish, and Hugh T. Patrick, eds. 1990. Aggressive Unilateralism: America’s 301 Trade Policy and the World Trading System. Ann Arbor: University of Michigan Press. Biersteker, Thomas J. 1993. “Evolving Perspectives on International Political Economy: Twentieth-Century Contexts and Discontinuities.” International Political Science Review 14: 7– 33. Blackburn, Keith, and Michael Christensen. 1989. “Monetary Policy and Policy Credibility: Theories and Evidence.” Journal of Economic Literature 27: 1– 45. Blanchard, Oliver Jean, and Pierre Alain Muet. 1993. “Competitiveness through Disinflation: An Assessment of the French Macroeconomic Strategy.” Economic Policy 16: 11– 56. Blanchard, Oliver, Rudiger Dornbusch, Stanley Fischer, Paul Krugman, Franco Modigliani, Paul Samuelson, and Robert Solow. 1993. “No Reason to Mourn.” Financial Times, December 23. Block, Fred L. 1977. The Origins of International Economic Disorder: A Study of United States International Monetary Policy from World War II to the Present. Berkeley: University of California Press. Boddez, Thomas M., and Michael J. Trebilcock. 1993. Unfinished Business: Reforming Trade Remedy Laws in North America. Policy Study 17. Toronto: C. D. Howe Institute. Bosworth, Barry P., Robert Z. Lawrence, and Nora Lustig. 1992. “Introduction.” Pp. 1–25 in Nora Lustig, Barry P. Bosworth, and Robert Z. Lawrence, eds., North American Free Trade Agreement: Assessing the Impact. Washington, D.C.: The Brookings Institution. Bouzas, Roberto, and Jaime Ros, eds. 1994a. Economic Integration in the Western Hemisphere. Notre Dame, Ind.: University of Notre Dame Press. ———. 1994b. “The North-South Variety of Economic Integration: Issues and Prospects for Latin America.” Pp. 1– 45 in Roberto Bouzas and Jaime Ros, eds., Economic Integration in the Western Hemisphere. Notre Dame, Ind.: University of Notre Dame Press. Bowker, Marjorie Montgomery. 1988. On Guard for Thee. Quebec: Voyageur Publishing. Bradley, John, and Karl Whelan. 1992. “The Irish Experience of Monetary Linkages with the United Kingdom and Developments since Joining the EMS.” Pp. 121–43 in Ray Barrell, ed., Economic Convergence and Monetary Union in Europe. London: Sage Publications. Bradsher, Keith. 1995. “The World Shifted, but Not Mexico.” New York Times, January 2, C1. Branson, William H. 1994. “German Reunification, the Breakdown of the EMS, and the Path to Stage Three.” Pp. 16–28 in David Cobham, ed., European Monetary Upheavals. Manchester: Manchester University Press. Bruno, Michael, and Jeffrey Sachs. 1985. The Economics of Worldwide Stagflation. Cambridge: Harvard University Press. Bryk, Dale S. 1991. “The Montreal Protocol and Recent Developments to Protect the Ozone Layer.” Harvard Environmental Law Review 15: 275–98. Buchanan, J. M. 1965. “An Economic Theory of Clubs.” Economica 32: 1–14. Bueno, Gerardo M. 1988. “A Mexican View.” Chap. 5 in William Diebold, Jr., ed.,
282
B I B L I O G R A P H Y
Bilateralism, Multilateralism, and Canada in U.S. Trade Policy. Cambridge, Mass.: Ballinger Publishing Company. Bueno de Mesquita, Bruce, Randolph M. Siverson, and Gary Woller. 1992. “War and the Fate of Regimes: A Comparative Analysis.” American Political Science Review 86: 638 – 46. Burley, Anne-Marie, and Walter Mattli. 1993. “Europe before the Court: A Political Theory of Legal Integration.” International Organization 47: 41–76. Cameron, David R. 1988. “The Canada-United States Free Trade Agreement.” Paper prepared for Workshop II of the Ford Foundation Project on Domestic-International Interaction, Center for Advanced Study in the Behavioral Sciences, Stanford, Calif., December. ———. 1989. “Socialism in an Open Economy: Constraint and Choice in French Economic Policy.” Yale University. Manuscript. ———. 1990. “Sovereign States in a Single Market: Integration and Intergovernmentalism in the European Community.” Paper presented at the Conference and European Political Institutions and Policymaking after 1992, The Brookings Institution, March 29 – 30. ———. 1992a. “The Maastricht Agreement on Economic and Monetary Union: Initiation, Negotiations, Implications.” Paper prepared for the Workshop of the Consortium on 1992 and the Emerging European Polity, Center for Advanced Study in the Behavioral Sciences, Stanford, Calif. ———. 1992b. “The 1992 Initiative: Causes and Consequences.” Chap. 2 in Alberta M. Sbragia, ed., Euro-Politics: Institutions and Policymaking in the “New” European Community. Washington, D.C.: The Brookings Institution. ———. 1998. “Creating Supranational Authority in Monetary and Exchange-Rate Policy: The Sources and Effects of EMU.” Chap. 7 in Wayne Sandholtz and Alec Stone Sweet, eds., European Integration and Supranational Governance. Oxford: Oxford University Press. Cameron, Duncan, ed. 1986. The Free Trade Papers. Toronto: James Lorimer. ———, ed. 1988. The Free Trade Deal. Toronto: James Lorimer. Campbell, Robert M., and Leslie A. Pal. 1989. The Real Worlds of Canadian Politics. Peterborough, Ont.: Broadview Press. Canada. 1985. Royal Commission on the Economic Union and Development Prospects for Canada [Macdonald Commission]. Report, 3 v. Ottawa: Supply and Services Canada. Canzoneri, Matthew B. 1985. “Monetary Policy Games and the Role of Private Information.” American Economic Review 75: 1056–70. Carr, E. H. 1946. The Twenty Years’ Crisis. London: St. Martin’s. Casella, Alessandra. 1992. “On Markets and Clubs: Economic and Political Integration of Regions with Unequal Productivity.” American Economic Review 82: 115 –21. Castañeda, Jorge G. 1990. “Salinas’ International Relations Gamble.” Journal of International Affairs 43: 407–22. ———. 1993a. “Can NAFTA Change Mexico?” Foreign Affairs 72, no. 4 (September/October): 66 – 80. ———. 1993b. “The Clouding Political Horizon.” Current History 92: 59 –66. Centeno, Miguel Ángel. 1994. Democracy within Reason: Technocratic Revolution in Mexico. San Diego: University of California, San Diego. Chapman, Anthony. 1988. Canada-U.S. Free Trade: The Road to Implementation. Ottawa: Library of Parliament, Research Branch, Economics Division.
B I B L I O G R A P H Y
283
Chayes, Abram, and Antonia Handler Chayes. 1993. “On Compliance.” International Organization 47: 175 –206. Choucri, Nazli, ed. 1993. Global Accord: Environmental Challenges and International Responses. Cambridge: MIT Press. Christensen, Thomas J., and Jack Snyder. 1990. “Chain Gangs and Passed Bucks: Predicting Alliance Patterns in Multipolarity.” International Organization 44: 137– 68. Clarke, Harold D., and Allan Kornberg. 1992. “Support for the Canadian Federal Progressive Conservative Party since 1988: The Impact of Economic Evaluations and Economic Issues.” Canadian Journal of Political Science 25: 29 –53. Coase, Ronald. 1937. “The Nature of the Firm.” Economica 4: 386 –405. Cohen, G. A. 1979. “Capitalism, Freedom, and the Proletariat.” In Alan Ryan, ed., The Idea of Freedom. New York: Oxford University Press. Cohen, Roger. 1997. “Economics Aside, Europeans Like Their Marks and Francs.” New York Times, September 18, A1. Collins, Susan. 1988. “Inflation and the European Monetary System.” Pp. 112–39 in Francesco Giavazzi, Stefano Micossi, and Marcus Miller, eds., The European Monetary System. Cambridge: Cambridge University Press. Conroy, Michael E., and Amy K. Glasmeier. 1992/93. “Unprecedented Disparities, Unparalleled Adjustment Needs: Winners and Losers on the NAFTA ‘Fast Track.’” Journal of Interamerican Studies and World Affairs 34: 1–37. Conybeare, John A. C. 1984. “Public Goods, Prisoners’ Dilemmas, and the International Political Economy.” International Studies Quarterly 28: 5–22. Cook, Maria Lorena, Kevin J. Middlebrook, and Juan Molinar Horcasitas. 1994a. “The Politics of Economic Restructuring in Mexico: Actors, Sequencing, and Coalition Change.” Chap. 1 in Cook, Middlebrook, and Molinar Horcasitas, eds., The Politics of Economic Restructuring: State-Society Relations and Regime Change in Mexico. San Diego: University of California, San Diego. ———, eds. 1994b. The Politics of Economic Restructuring: State-Society Relations and Regime Change in Mexico. San Diego: University of California, San Diego. Cooper, Richard N. 1980. The Economics of Interdependence: Economic Policy in the Atlantic Community. 2d ed. New York: Columbia University Press. ———. 1986. Economic Policy in an Interdependent World. Cambridge: MIT Press. Corden, W. Max. 1972. “Economies of Scale and Customs Union Theory.” Journal of Political Economy 80: 465 –75. Cornelius, Wayne A. 1996. Mexican Politics in Transition: The Breakdown of a OneParty-Dominant Regime. San Diego: University of California, San Diego. Cowhey, Peter F. 1993. “Domestic Institutions and the Credibility of International Commitments: Japan and the United States.” International Organization 47: 299– 326. Crone, Donald. 1993. “Does Hegemony Matter? The Reorganization of the Pacific Political Economy.” World Politics 45: 501–25. Dahl, Robert A. 1957. “The Concept of Power.” Behavioral Science 2: 201–5. ———. 1961. Who Governs? Democracy and Power in an American City. New Haven: Yale University Press. de Boissieu, Christian, and Jean Pisani-Ferry. 1998. “The Political Economy of French Economic Policy in the Perspective of EMU.” Chap. 3 in Barry Eichengreen and Jeffry Frieden, eds., Forging an Integrated Europe. Ann Arbor: University of Michigan Press. De Cecco, Marcello. 1989. “The European Monetary System and National Inter-
284
B I B L I O G R A P H Y
ests.” In Paolo Guerrieri and Pier Carlo Padoan, eds., The Political Economy of European Integration: States, Markets, and Institutions. Savage, Md.: Barnes and Noble. De Grauwe, Paul. 1990. “The Cost of Disinflation and the European Monetary System.” Open Economies Review 1: 147–73. ———. 1992. The Economics of Monetary Integration. Oxford: Oxford University Press. de la Torre, Augusto, and Margaret R. Kelly. 1992. Regional Trading Arrangements. IMF Occasional Paper no. 93 (March). Washington, D.C.: International Monetary Fund. de Melo, Jaime, and Arvind Panagariya. 1993. “Introduction.” Chap. 1 in de Melo and Panagariya, eds., New Dimensions in Regional Integration. Cambridge: Cambridge University Press. Dennis, Geoffrey, and Joseph Nellis. 1986. “The European Monetary System and UK Membership.” In Robert E. Baldwin and J. David Richardson, International Trade and Finance: Readings. Boston: Little, Brown and Company. Destler, I. M. 1992. American Trade Politics, 2d ed. Washington, D.C.: Institute for International Economics. Destler, I. M., and C. Randall Henning. 1989. Dollar Politics: Exchange Rate Policymaking in the United States. Washington, D.C.: Institute for International Economics. Destler, I. M., and John S. Odell. 1987. Anti-Protection: Changing Forces in United States Trade Politics. Washington, D.C.: Institute for International Economics. Diebold, William, Jr. 1988. “The New Bilateralism?” Chap. 6 in William Diebold, Jr., ed., Bilateralism, Multilateralism, and Canada in U.S. Trade Policy. Cambridge, Mass.: Ballinger Publishing Company. Dixit, Avinash K. 1996. The Making of Economic Policy: A Transaction Cost Politics Perspective. Cambridge: MIT Press. Doern, G. Bruce, and Brian W. Tomlin. 1991. Faith and Fear: The Free Trade Story. Toronto: Stoddart. Dornbusch, Rudiger. 1989. “Credibility, Debt, and Unemployment: Ireland’s Failed Stabilization.” Economic Policy 4: 173 –209. ———. 1993a. Comments on Barry Eichengreen and Charles Wyplosz, “The Unstable EMS.” Brookings Papers on Economic Activity 1: 130–36. ———. 1993b. Comments on Jeffrey Frankel, Steven Phillips, and Menzie Chinn, “Financial and Currency Integration in the European Monetary System: The Statistical Record.” Pp. 306 –11 in Francisco Torres and Francesco Giavazzi, eds., Adjustment and Growth in the European Monetary Union. Cambridge: Cambridge University Press. Downs, George W., and David M. Rocke. 1995. Optimal Imperfection? Domestic Uncertainty and Institutions in International Relations. Princeton: Princeton University Press. Durch, William J. 1996. UN Peacekeeping, American Policy, and the Uncivil Wars of the 1990s. New York: St. Martin’s Press. Edwards, Sebastian. 1995. “Comments.” In Stephan Haggard, Developing Nations and the Politics of Global Integration. Washington, D.C.: The Brookings Institution. Eichengreen, Barry. 1989. “Hegemonic Stability Theories of the International Monetary System.” Pp. 255 – 98 in Richard N. Cooper, Barry Eichengreen, Gerald
B I B L I O G R A P H Y
285
Hotham, Robert D. Putnam, and C. Randall Henning, eds., Can Nations Agree? Issues in International Economic Cooperation, Washington, D.C.: The Brookings Institution. ———. 1992. “Should the Maastricht Treaty Be Saved?” Working Paper 1.10 of the Political Economy of European Integration Research Group, University of California Center for German and European Studies. ———. 1994. “Fiscal Policy and EMU.” Chap. 9 in Barry Eichengreen and Jeffry Frieden, eds., The Political Economy of European Monetary Unification. Boulder, Colo.: Westview Press. Eichengreen, Barry, Jeffry Frieden, and Jürgen von Hagen. 1995. “The Political Economy of European Integration: Introduction.” Pp. 1–10 in Eichengreen, Frieden, and von Hagen, eds., Politics and Institutions in an Integrated Europe. Berlin: Springer. Eichengreen, Barry, and Charles Wyplosz. 1993. “The Unstable EMS.” Brookings Papers on Economic Activity 1: 51–143. Elster, Jon. 1983. “Exploitation, Freedom, and Justice.” In J. Roland Pennock and John Chapman, eds., Nomos XXVI: Marxism. New York: New York University Press. Evans, Peter B., Harold K. Jacobson, and Robert D. Putnam, eds. 1993. DoubleEdged Diplomacy: International Bargaining and Domestic Politics. Berkeley: University of California Press. Farnsworth, Clyde H. 1992. “U.S.-Canada Rifts Grow over Trade.” New York Times, February 18, A1. Fearon, James D. 1991. “Counterfactuals and Hypothesis Testing in Political Science.” World Politics 43: 169 – 95. ———. 1994. “Domestic Political Audiences and the Escalation of International Disputes.” American Political Science Review 88: 577–92. ———. 1998. “Bargaining, Enforcement, and International Cooperation.” International Organization 52: 269 – 305. Feinberg, Richard E. 1997. Summitry in the Americas. Washington, D.C.: Institute for International Economics. Feldstein, Martin. 1997. “The Political Economy of the European Economic and Monetary Union: Political Sources of an Economic Liability.” Journal of Economic Perspectives 11: 23 – 42. Fieleke, Norman. 1992. “One Trade World, or Many: The Issue of Regional Trading Blocs.” New England Economic Review (May/June): 3–20. Finger, J. Michael. 1991. “The GATT as an International Discipline over Trade Restrictions: A Public Choice Approach.” Pp. 125– 41 in Roland Vaubel and Thomas D. Willett, eds., The Political Economy of International Organizations: A Public Choice Approach. Boulder, Colo.: Westview Press. Finnemore, Martha. 1996. National Interests and International Society. Ithaca: Cornell University Press. Flanagan, Robert J., David W. Soskice, and Lloyd Ulman. 1983. Unionism, Economic Stabilization, and Incomes Polices: European Experience. Washington, D.C.: The Brookings Institution. Frankel, Jeffrey A. 1997. Regional Trading Blocs in the World Economic System. Washington, D.C.: Institute for International Economics. Frankel, Jeffrey, and Shang-Jin Wei. 1995. “European Integration and the Regionalization of World Trade and Currencies: The Economics and the Politics.” In Barry
286
B I B L I O G R A P H Y
Eichengreen, Jeffry Frieden, and Jürgen von Hagen, eds., Monetary and Fiscal Policy in an Integrated Europe. Berlin: Springer. Fratianni, Michele, and Jürgen von Hagen. 1992. The European Monetary System and European Monetary Union. Boulder, Colo.: Westview Press. Frieden, Jeffry A. 1991. “Invested Interests: The Politics of National Economic Policies in a World of Global Finance.” International Organization 45: 425–51. ———. 1994. “Making Commitments: France and Italy in the European Monetary System, 1979 –1985.” Chap. 2 in Barry Eichengreen and Jeffry A. Frieden, eds., The Political Economy of European Monetary Unification. Boulder, Colo.: Westview Press. ———. 1998. “The Euro: Who Wins? Who Loses?” Foreign Policy 112 (Fall): 25– 40. Frieden, Jeffry A., and Ronald Rogowski. 1996. “The Impact of the International Economy on National Policies: An Analytical Overview.” Chap. 2 in Robert O. Keohane and Helen Milner, eds., Internationalization and Domestic Politics. New York: Cambridge University Press. Friedman, David. 1977. “A Theory of the Size and Shape of Nations.” Journal of Political Economy 85: 59 –77. Friman, H. Richard. 1993. “Side-Payments versus Security Cards: Domestic Bargaining Tactics in International Economic Negotiations.” International Organization 47: 387– 410. Fudenberg, Drew, and Eric Maskin. 1986. “The Folk Theorem with Discounting or with Incomplete Information.” Econometrica 54: 533–54. Funabashi, Yoichi. 1995. Asia Pacific Fusion: Japan’s Role in APEC. Washington, D.C.: Institute for International Economics. Gaines, Brian J. 1994. “Introduction to Canada and the Politics of Free Trade.” Working paper. Graduate School of Business, Stanford University. Gandhi, Prem. 1990. “Review Essay: Canada-United States Economic Relations Today: Living with the Free Trade Agreement.” American Review of Canadian Studies 20: 423– 46. Garrett, Geoffrey. 1992. “International Cooperation and Institutional Choice: The European Community’s Internal Market.” International Organization 46: 533– 60. ———. 1993. “The Politics of Structural Change: Swedish Social Democracy and Thatcherism in Comparative Perspective.” Comparative Political Studies 25: 521– 47. ———. 1994. “The Politics of Maastricht.” Chap. 3 in Barry Eichengreen and Jeffry A. Frieden, eds., The Political Economy of European Monetary Unification. Boulder, Colo.: Westview Press. ———. 1995. “The Politics of Legal Integration in the European Union.” International Organization 49: 171– 81. ———. 1998. “The Transition to Economic and Monetary Union.” Chap. 2 in Barry Eichengreen and Jeffry Frieden, eds., Forging an Integrated Europe. Ann Arbor: University of Michigan Press. Garrett, Geoffrey, R. Daniel Kelemen, and Heiner Schulz. 1998. “The European Court of Justice, National Governments, and Legal Integration in the European Union.” International Organization 52: 149 –76. Garrett, Geoffrey, and Peter Lange. 1991. “Political Responses to Interdependence: What’s ‘Left’ for the Left?” International Organization 45: 539–64.
B I B L I O G R A P H Y
287
———. 1995. “Internationalization, Institutions and Political Change.” International Organization 48: 627– 55. Garrett, Geoffrey, and Barry R. Weingast. 1993. “Ideas, Interests and Institutions: Constructing the EC’s Internal Market.” Chap. 7 in Judith Goldstein and Robert O. Keohane, eds., Ideas and Foreign Policy: Beliefs, Institutions, and Political Change. Ithaca: Cornell University Press. Gaston, Noel, and Daniel Trefler. 1994. “The Role of International Trade and Trade Policy in the Labour Markets of Canada and the United States.” World Economy 17: 45 – 62. ———. 1995. “The Labour Market Consequences of the Canada-U.S. Free Trade Agreement.” Seminar paper 95– 09, Centre for International Economic Studies, University of Adelaide, Australia. Gaventa, John. 1980. Power and Powerlessness: Quiescence and Rebellion in an Appalachian Valley. Urbana: University of Illinois Press. General Agreement on Tariffs and Trade. 1994. The Results of the Uruguay Round of Multilateral Trade Negotiations: The Legal Texts. Geneva: GATT Secretariat. Giavazzi, Francesco, and Alberto Giovannini. 1988. “The Role of the ExchangeRate Regime in a Disinflation: Empirical Evidence on the European Monetary System.” Pp. 85 –111 in Francesco Giavazzi, Stefano Micossi, and Marcus Miller, eds., The European Monetary System. Cambridge: Cambridge University Press. ———. 1989. Limiting Exchange Rate Flexibility: The European Monetary System. Cambridge: MIT Press. Giavazzi, Francesco, and Marco Pagano. 1988. “The Advantage of Tying One’s Hands: EMS Discipline and Central Bank Credibility.” European Economic Review 32: 1055 –75. Giavazzi, Francesco, and Luigi Spaventa. 1989. “Italy: The Real Effects of Inflation and Disinflation.” Economic Policy 4: 133 –72. Gilpin, Robert. 1975. U.S. Power and the Multinational Corporation. New York: Basic Books. ———. 1981. War and Change in World Politics. Cambridge: Cambridge University Press. ———. 1984. “Structural Constraints on Economic Leverage: Market-Type Systems.” Chap. 6 in Gordon H. McCormick and Richard E. Bissell, eds., Strategic Dimensions of Economic Behavior. New York: Praeger. Glaser, Charles L. 1990. Analyzing Strategic Nuclear Policy. Princeton: Princeton University Press. ———. 1994/95. “Realists as Optimists: Cooperation as Self-Help.” International Security 19: 50 – 90. Globerman, Steven, and Michael Walker, eds. 1993. Assessing NAFTA: A Trinational Analysis. Vancouver: Fraser Institute. Goldstein, Judith. 1988. “Ideas, Institutions, and American Trade Policy.” International Organization 42: 179 –217. ———. 1993. Ideas, Interests, and American Trade Policy. Ithaca: Cornell University Press. ———. 1996. “International Law and Domestic Institutions: Reconciling North American ‘Unfair’ Trade Laws.” International Organization 50: 541–64. Goldstein, Judith, and Robert O. Keohane, eds. 1993. Ideas and Foreign Policy: Beliefs, Institutions, and Political Change. Ithaca: Cornell University Press.
288
B I B L I O G R A P H Y
Goodman, John B. 1992. Monetary Sovereignty: The Politics of Central Banking in Western Europe. Ithaca: Cornell University Press. Gowa, Joanne. 1986. “Anarchy, Egoism, and Third Images: The Evolution of Cooperation and International Relations.” International Organization 40: 167–86. ———. 1994. Allies, Adversaries, and International Trade. Princeton: Princeton University Press. Gowa, Joanne, and Edward D. Mansfield. 1993. “Power Politics and International Trade.” American Political Science Review 87: 408–20. Greenhouse, Steven. 1994. “Clinton Outlines Plan to Bolster Europe-Wide Security.” New York Times, November 16. Grieco, Joseph M. 1988. “Anarchy and the Limits of Cooperation: A Realist Critique of the Newest Liberal Institutionalism.” International Organization 42: 485–507. ———. 1990. Cooperation among Nations: Europe, America, and Nontariff Barriers to Trade. Ithaca: Cornell University Press. ———. 1993. “Understanding the Problem of International Cooperation: The Limits of Neoliberal Institutionalism, and the Future of Realist Theory.” Chap. 12 in David A. Baldwin, ed., Neorealism and Neoliberalism: The Contemporary Debate. New York: Columbia University Press. ———. 1996. “State Interests and Institutional Rule Trajectories: A Neorealist Interpretation of the Maastricht Treaty and European Economic and Monetary Union.” Security Studies 5: 176 –222. ———. 1997. “Systemic Sources of Variation in Regional Institutionalization in Western Europe, East Asia, and the Americas.” Chap. 7 in Edward D. Mansfield and Helen Milner, eds., The Political Economy of Regionalism. New York: Columbia University Press. Grilli, Vittorio. 1994. Comments on “The European Monetary System and European Integration: An Evaluation.” Pp. 74–78 in Paul J. J. Welfens, ed., European Monetary Integration: EMS Developments and International Post-Maastricht Perspectives, 2d ed. Berlin: Springer-Verlag. Grinspun, Ricardo, and Maxwell A. Cameron, eds. 1993. The Political Economy of North American Free Trade. New York: St. Martin’s Press. Gros, Daniel, and Niels Thygesen. 1992. European Monetary Integration: From the European Monetary System to European Monetary Union. London: Longman. Grossman, Sanford J., and Oliver D. Hart. 1986. “The Costs and Benefits of Ownership: A Theory of Vertical and Lateral Integration.” Journal of Political Economy 94: 691–719. Grunwald, Joseph. 1993. “Hemispheric Economic Integration? Some Reflections.” Annals of the American Academy of Political and Social Science 526: 135–50. Guerrieri, Paolo, and Pier Carlo Padoan, eds. 1989. The Political Economy of European Integration: States, Markets, and Institutions. Savage, Md.: Barnes and Noble. Haas, Ernst. 1964. Beyond the Nation State: Functionalism and International Organization. Stanford, Calif.: Stanford University Press. Haas, Peter M. 1990. Saving the Mediterranean: The Politics of International Environmental Protection. New York: Columbia University Press. ———. 1992. “Banning Chlorofluorocarbons: Epistemic Community Efforts to Protect Stratospheric Ozone.” International Organization 46: 187–224. Haas, Peter M., Robert O. Keohane, and Marc A. Levy, eds. 1993. Institutions for the Earth: Sources of Effective Environmental Protection. Cambridge: MIT Press.
B I B L I O G R A P H Y
289
Haggard, Stephan. 1988. “The Institutional Foundations of Hegemony: Explaining the Reciprocal Trade Agreements Act of 1934.” International Organization 42: 91–119. ———. 1990. Pathways from the Periphery: The Politics of Growth in the Newly Industrializing Countries. Ithaca: Cornell University Press. ———. 1995. Developing Nations and the Politics of Global Integration. Washington, D.C.: The Brookings Institution. ———. 1997. “Regionalism in Asia and the Americas.” Chap. 2 in Edward D. Mansfield and Helen V. Milner, eds., The Political Economy of Regionalism. New York: Columbia University Press. Haggard, Stephan, and Beth A. Simmons. 1987. “Theories of International Regimes.” International Organization 41: 491– 517. Hall, Peter A. 1986. Governing the Economy: The Politics of State Intervention in Britain and France. Oxford: Oxford University Press. Hankel, Wilhelm. 1980. “Germany: Economic Nationalism in the International Economy.” Pp. 22– 30 in Wilfrid Kohl and Giorgio Basevi, eds., West Germany: A European and Global Power. Lexington, Mass.: Lexington Books. Hardin, Russell. 1982. Collective Action. Baltimore: Johns Hopkins University Press. Harmon, Mark D. 1994. “‘If I Can’t Change the Rules, Then I Won’t Play Your Game’: Britain in and out of the Exchange Rate Mechanism of the European Monetary System.” Working Paper 2.24 of the Political Economy of European Integration Research Group. University of California Center for German and European Studies. June. Harsanyi, John C. 1965. “Measurement of Social Power, Opportunity Costs, and the Theory of Two-Person Bargaining Games.” Pp. 378 –85 in J. David Singer, ed., Human Behavior and International Politics. Chicago: Rand McNally. Hart, Michael. 1990. A North American Free Trade Agreement: The Strategic Implications for Canada. Ottawa: Centre for Trade Policy and Law and Institute for Research on Public Policy. ———. 1994. Decision at Midnight: Inside the Canada-U.S. Free-Trade Negotiations. Vancouver: University of British Columbia Press. Heisenberg, Dorothee. 1999. The Mark of the Bundesbank: Germany’s Role in European Monetary Cooperation. Boulder, Colo.: Lynne Rienner Publishers. Helleiner, Gerald K. 1993. “Considering U.S.-Mexico Free Trade.” Chap. 3 in Ricardo Grinspun and Maxwell A. Cameron, eds., The Political Economy of North American Free Trade. New York: St. Martin’s Press. Hellman, Judith Adler. 1993. “Mexican Perceptions of Free Trade: Support and Opposition to NAFTA.” In Ricardo Grinspun and Maxwell A. Cameron, eds., The Political Economy of North American Free Trade. New York: St. Martin’s Press. Henderson, Michael D. 1987. The Future on the Table: Canada and the Free Trade Issue. Toronto: Masterpress. Henning, C. Randall. 1994. Currencies and Politics in the United States, Germany, and Japan. Washington, D.C.: Institute for International Economics. Heraclides, Alexis. 1993. Helsinki II and Its Aftermath: The Making of the CSCE into an International Institution. London: Pinter Publishers. Heredia, Carlos. 1994. “NAFTA and Democratization in Mexico.” Journal of International Affairs 48: 13 – 38. Hibbs, Douglas. 1977. “Political Parties and Macroeconomic Policy.” American Political Science Review 71: 1467– 87.
290
B I B L I O G R A P H Y
———. 1987. The Political Economy of Industrial Democracies. Cambridge: Harvard University Press. Hirschman, Albert. 1945. National Power and the Structure of Foreign Trade. Berkeley: University of California Press. ———. 1970. Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Cambridge: Harvard University Press. Horlick, Gary N., and F. Amanda DeBusk. 1992. “The Functioning of FTA Dispute Resolution Panels.” Chap. 1 in Leonard Waverman, ed., Negotiating and Implementing a North American Free Trade Agreement. Vancouver: Fraser Institute. ———. 1993. “Dispute Resolution under NAFTA: Building on the U.S.-Canada FTA, GATT, and ICSID.” Journal of International Arbitration 10: 51–71. Horlick, Gary N., Geoffrey D. Oliver, and Debra P. Steger. 1988. “Dispute Resolution Mechanisms.” Chap. 3 in Jeffrey J. Schott and Murray G. Smith, eds., The Canada-United States Free Trade Agreement: The Global Impact. Washington, D.C.: Institute for International Economics. Howe, Geoffrey. 1990. Resignation Speech to the House of Commons. Hansard, November 13, cols. 461– 65. Hufbauer, Gary Clyde, and Jeffrey J. Schott. 1992. North American Free Trade: Issues and Recommendations. Washington, D.C.: Institute for International Economics. ———. 1993. NAFTA: An Assessment. Rev. ed. Washington, D.C.: Institute for International Economics. ———. 1994. Western Hemisphere Economic Integration. Washington, D.C.: Institute for International Economics. Hufbauer, Gary Clyde, Jeffrey J. Schott, and Kimberly Ann Elliott. 1990. Economic Sanctions Reconsidered. 2d ed., rev., 2 vols. Washington, D.C.: Institute for International Economics. Huntington, David S. 1993. “Settling Disputes under the North American Free Trade Agreement.” Harvard International Law Journal 34, no. 2 (Spring): 407–43. Hurrell, Andrew. 1995. “Regionalism in the Americas.” Chap. 9 in Louise Fawcett and Andrew Hurrell, eds., Regionalism in World Politics. Oxford: Oxford University Press. Hurtig, Mel. 1991. The Betrayal of Canada. Toronto: Stoddart. Ikenberry, G. John, and Charles Kupchan. 1990. “Socialization and Hegemonic Power.” International Organization 44: 283 – 316. Ikenberry, G. John, David A. Lake, and Michael Mastanduno. 1988. “Introduction: Approaches to Explaining American Foreign Economic Policy.” International Organization 42: 1–14. IMF. 1984a. Direction of Trade Statistics Yearbook. Washington, D.C.: International Monetary Fund. ———. 1984b. International Financial Statistics Yearbook. Washington, D.C.: International Monetary Fund. ———. 1992. Direction of Trade Statistics Yearbook. Washington, D.C.: International Monetary Fund. Ishiyama, Yoshihide. 1975. “The Theory of Optimum Currency Areas: A Survey.” IMF Staff Papers 22: 344 – 83. Jachtenfuchs, Markus. 1980. “The European Community and the Protection of the Ozone Layer.” Journal of Common Market Studies 28: 261–77.
B I B L I O G R A P H Y
291
Jackson, John H. 1989. The World Trading System: Law and Policy of International Economic Relations. Cambridge: MIT Press. ———. 1994. “Dispute Settlement Procedures.” Pp. 117–23 in The New World Trading System: Readings. OECD Documents. Paris: OECD. Jacobson, Gary C. 1987. “Running Scared: Elections and Congressional Politics in the 1980s.” Chap. 3 in Matthew D. McCubbins and Terry Sullivan, eds., Congress: Structure and Policy. Cambridge: Cambridge University Press. James, Scott C., and David A. Lake. 1989. “The Second Face of Hegemony: Britain’s Repeal of the Corn Laws and the American Walker Tariff of 1846.” International Organization 43: 1–29. Jenkins, Roy. 1989. European Diary. London: Collins. Jervis, Robert. 1976. Perception and Misperception in International Politics. Princeton: Princeton University Press. ———. 1978. “Cooperation under the Security Dilemma.” World Politics 30: 167– 214. Joffe, Josef. 1997. “The Euro: The Engine That Couldn’t.” New York Review of Books 44, no. 19 (December 4): 26 – 31. Johnson, Jon R. 1993. “NAFTA and the Trade in Automotive Goods.” Pp. 87–129 in Steven Globerman and Michael Walker, eds., Assessing NAFTA: A Trinational Analysis. Vancouver: Fraser Institute. Johnston, Richard, André Blais, Henry E. Brady, and Jean Crête. 1992. Letting the People Decide: Dynamics of a Canadian Election. Stanford, Calif.: Stanford University Press. ———. 1993. “Free Trade in Canadian Elections: Issue Evolution in the Long and the Short Run.” Pp. 231– 54 in William H. Riker, ed., Agenda Formation. Ann Arbor: University of Michigan Press. Joskow, Paul L. 1985. “Vertical Integration and Long Term Contracts: The Case of Coal Burning Electric Generating Plants.” Journal of Law, Economics, and Organization 1: 33 – 80. Josling, Tim. 1992. “NAFTA and Agriculture: A Review of the Economic Impacts.” Pp. 144 –75 in Nora Lustig, Barry P. Bosworth, and Robert Z. Lawrence, eds., North American Free Trade Agreement: Assessing the Impact. Washington, D.C.: The Brookings Institution. Josling, Tim, and Rick Barichello. 1993. Agriculture in the NAFTA. Toronto: C. D. Howe Institute. Judt, Tony. 1997. “Continental Rift.” New York Times, June 5, A39. Kaempfer, William H., and Anton D. Lowenberg. 1992. International Economic Sanctions. Boulder, Colo.: Westview Press. Kahler, Miles. 1992. “External Influence, Conditionality, and the Politics of Adjustment.” Chap. 2 in Stephan Haggard and Robert R. Kaufman, The Politics of Economic Adjustment. Princeton: Princeton University Press. ———. 1995. International Institutions and the Political Economy of Integration. Washington, D.C.: The Brookings Institution. Kaltenthaler, Karl. 1998. Germany and the Politics of Europe’s Money. Durham, N.C.: Duke University Press. Kaminski, John P. 1985. “New York: The Reluctant Pillar.” Pp. 48–117 in Stephen L. Schechter, ed., The Reluctant Pillar: New York and the Adoption of the Federal Constitution. Troy, N.Y.: Russell Sage College.
292
B I B L I O G R A P H Y
———. 1988. “Adjusting to Circumstances: New York’s Relationship with the Federal Government, 1776–1788.” Pp. 225 – 49 in Patrick T. Conley and John P. Kaminski, eds., The Constitution and the States: The Role of the Original Thirteen in the Framing and Adoption of the Federal Constitution. Madison, Wisc.: Madison House. Katzenstein, Peter J., ed. 1996. The Culture of National Security: Norms and Identity in World Politics. New York: Columbia University Press. Kaufman, Hugo M. 1982. “The European Monetary System and National Policy Constraints.” Comparative Social Research 5: 107–28. Kaufman, Robert R., Carlos Bazdresch, and Blanca Heredia. 1994. “Mexico: Radical Reform in a Dominant Party System.” Pp. 360 – 410 in Stephan Haggard and Steven B. Webb, eds., Voting for Reform: Democracy, Political Liberalization, and Economic Adjustment. Oxford: Oxford University Press. Kees, Andreas. 1987. “The Monetary Committee of the European Community.” Kredit und Kapital 2: 258. Kehoe, Timothy J. 1994. “Assessing the Economic Impact of North American Free Trade.” Pp. 3 – 33 in M. Delal Baer and Sidney Weintraub, eds., The NAFTA Debate: Grappling with Unconventional Trade Issues. Boulder, Colo.: Lynne Rienner Publishers. Keohane, Robert O. 1980. “The Theory of Hegemonic Stability and Changes in International Economic Regimes, 1967–1977.” Pp. 131–62 in Ole R. Holsti, Randolph M. Siverson, and Alexander L. George, eds., Change in the International System. Boulder, Colo.: Westview. ———. 1984. After Hegemony: Cooperation and Discord in the World Political Economy. Princeton: Princeton University Press. ———. 1988. “International Institutions: Two Approaches.” International Studies Quarterly 32: 379 – 96. Keohane, Robert O., and Stanley Hoffmann. 1991. The New European Community: Decisionmaking and Institutional Change. Boulder, Colo.: Westview Press. Kesselman, Mark. 1980. “The Economic Analysis and Program of the French Communist Party.” Pp. 177– 90 in Philip Cerny and Martin Schain, eds., French Politics and Public Policy. New York: St. Martin’s Press. Kiewiet, D. Roderick, and Douglas Rivers. 1984. “A Retrospective on Retrospective Voting.” Political Behavior 6: 369 – 93. Kindleberger, Charles P. 1973. The World in Depression, 1929–1939. Berkeley: University of California Press. King, Gary, Robert O. Keohane, and Sidney Verba. 1994. Designing Social Inquiry: Scientific Inference in Qualitative Research. Princeton: Princeton University Press. Kitschelt, Herbert P. 1999. Partisan Competition and Welfare State Retrenchment: When Do Politicians Choose Unpopular Policies? Duke University. Manuscript. Klein, Benjamin, Robert G. Crawford, and Armen A. Alchian. 1978. “Vertical Integration, Appropriable Rents, and the Competitive Contracting Process.” Journal of Law and Economics 21: 297– 326. Klesner, Joseph L. 1994. “Realignment or Dealignment? Consequences of Economic Crisis and Restructuring for the Mexican Party System.” Chap. 7 in Maria Lorena Cook, Kevin J. Middlebrook, and Juan Molinar Horcasitas, eds., The Politics of Economic Restructuring: State-Society Relations and Regime Change in Mexico. San Diego: University of California, San Diego.
B I B L I O G R A P H Y
293
———. 1997. “Democratic Transition? The 1997 Mexican Elections.” PS: Political Science and Politics 30: 703 –11. Knight, Jack. 1992. Institutions and Social Conflict. Cambridge: Cambridge University Press. Koechlin, Timothy, and Mehrene Larudee. 1992. “The High Cost of NAFTA.” Challenge (September/October): 19 –26. Koremenos, Barbara, Charles Lipson, and Duncan Snidal, eds. 1999. “Rational International Institutions.” University of Chicago. Manuscript. Kramer, Gerald H. 1971. “Short-Term Fluctuations in U.S. Voting Behavior, 1896 – 1964.” American Political Science Review 65: 131–43. Krasner, Stephen D. 1976. “State Power and the Structure of International Trade.” World Politics 28: 317– 45. ———, ed. 1983. International Regimes. Ithaca: Cornell University Press. ———. 1985. Structural Conflict: The Third World Against Global Liberalism. Berkeley: University of California Press. ———. 1991. “Global Communications and National Power: Life on the Pareto Frontier.” World Politics 43: 336 – 66. ———. 1992. “Power Asymmetries and Relations between Mexico and the United States.” Stanford University. Manuscript. Kratochwil, Friedrich, and John Ruggie. 1986. “International Organization: A State of the Art on an Art of the State.” International Organization 40: 753–75. Kreps, David. 1990. “Corporate Culture and Economic Theory.” Chap. 4 in James E. Alt and Kenneth A. Shepsle, eds., Perspectives on Positive Political Economy. Cambridge: Harvard University Press. Kreps, David, Paul Milgrom, John Roberts, and Robert Wilson. 1982. “Rational Cooperation in the Finitely Repeated Prisoners’ Dilemma.” Journal of Economic Theory 27: 245 – 52. Kreps, David, and Robert Wilson. 1982. “Reputation and Imperfect Information.” Journal of Economic Theory 27: 253 –76. Krueger, Anne O. 1993. Economic Policies at Cross-Purposes: The United States and Developing Countries. Washington, D.C.: The Brookings Institution. Krugman, Paul. 1988. “EFTA and 1992.” Occcasional Paper no. 23. Geneva: Economic Free Trade Association Secretariat. ———. 1991a. Geography and Trade. Cambridge: MIT Press. ———. 1991b. “Is Bilateralism Bad?” Pp. 9 –23 in Elhanan Helpman and Assaf Razin, eds., International Trade and Trade Policy. Cambridge: MIT Press. ———. 1993a. “Lessons of Massachusetts for EMU.” Chap. 8 in Francisco Torres and Francesco Giavazzi, eds., Adjustment and Growth in the European Monetary Union. Cambridge: Cambridge University Press. ———. 1993b. “The Uncomfortable Truth about NAFTA.” Foreign Affairs 72, no. 5 (November/December): 13 –19. ———. 1994. Peddling Prosperity: Economic Sense and Nonsense in the Age of Diminished Expectations. New York: W. W. Norton and Company. ———. 1995. “Dutch Tulips and Emerging Markets.” Foreign Affairs 74, no. 4 (July/August): 28 – 44. Krugman, Paul R., and Maurice Obstfelt. 1994. International Economics: Theory and Policy. 3d ed. New York: HarperCollins. Kruse, Donald C. 1980. Monetary Integration in Western Europe: EMU, EMS, and Beyond. London: Butterworths.
294
B I B L I O G R A P H Y
Kupchan, Charles A. 1998. “After Pax Americana: Benign Power, Regional Integration, and the Sources of a Stable Multipolarity.” International Security 23: 40–79. Kupchan, Charles A., and Clifford A. Kupchan. 1991. “Concerts, Collective Security, and the Future of Europe.” International Security 16: 114–61. Kydland, Finn E., and Edward C. Prescott. 1977. “Rules Rather Than Discretion: The Inconsistency of Optimal Plans.” Journal of Political Economy 85: 473–91. Lake, David A. 1993. “Leadership, Hegemony, and the International Economy: Naked Emperor or Tattered Monarch with Potential?” International Studies Quarterly 37: 459 – 89. Lange, Peter. 1993. “Maastricht and the Social Protocol: Why Did They Do It?” Politics and Society 21: 5 – 36. Lange, Peter, and Geoffrey Garrett. 1985. “The Politics of Growth: Strategic Interaction and Economic Performance in the Advanced Industrial Democracies, 1974– 1980.” Journal of Politics 47: 792– 827. Lauber, Volkmar. 1983. The Political Economy of France: From Pompidou to Mitterrand. New York: Praeger. Lawrence, Robert Z. 1991. “Emerging Regional Arrangements: Building Blocks or Stumbling Blocks?” Pp. 22– 35 in Richard O’Brien, ed., Finance and the International Economy, vol. 5 of the AMEX Bank Review Prize Essays. New York: Oxford University Press. ———. 1996. Regionalism, Multilateralism, and Deeper Integration. Washington, D.C.: The Brookings Institution. Lawson, Nigel. 1990. Speech in the House of Commons Following Sterling’s Entry into the ERM. Hansard, October 23, cols. 214 –16. ———. 1992. The View from Number 11: Memoirs of a Tory Radical. London: Bantam. Lebow, Richard Ned. 1994. “The Long Peace, the End of the Cold War, and the Failure of Realism.” International Organization 48: 249–77. LeDuc, Lawrence. 1991. “Voting for Free Trade? The Canadian Voter and the 1988 Federal Election.” Pp. 350– 66 in Paul W. Fox and Graham White, eds., Politics: Canada. Toronto: McGraw-Ryerson. Levy, Jack S. 1990. “The Offense/Defense Balance of Military Technology: A Theoretical and Historical Analysis.” International Studies Quarterly 28: 222–30. Linz, Juan. 1975. “Totalitarian and Authoritarian Regimes.” In Fred Greenstein and Nelson Polsby, eds., Handbook of Political Science, vol. 3. Reading, Mass.: Addison-Wesley. Lipsey, Richard G. 1990. Canada at the U.S.-Mexico Free Trade Dance: Wallflower or Partner? Toronto: C. D. Howe Institute. ———. 1991. “The Case for Trilateralism.” Pp. 89–124 in Steven Globerman, ed., Continental Accord: North American Economic Integration. Vancouver: The Fraser Institute. Lipsey, Richard G., Daniel Schwanen, and Ronald J. Wonnacott. 1994. The NAFTA: What’s In, What’s Out, What’s Next? Toronto: C. D. Howe Institute. Lipsey, Richard, and Robert York. 1988. Evaluating the Free Trade Deal. Toronto: C. D. Howe Institute. Lipson, Charles. 1984. “International Cooperation in Economic and Security Affairs.” World Politics 37: 1–23. ———. 1985. Standing Guard: Protecting Foreign Capital in the Nineteenth and Twentieth Centuries. Berkeley: University of California Press.
B I B L I O G R A P H Y
295
———. 1991. “Why Are Some International Agreements Informal?” International Organization 45: 495 – 538. Lohmann, Susanne. 1992. “Optimal Commitment in Monetary Policy: Credibility versus Flexibility.” American Economic Review 82: 273–86. Lomax, D. F. 1987. “The United Kingdom Case.” In Paul Van den Bempt, ed., The European Monetary System: Towards More Convergence and Closer Integration. Leuven, Belgium: Academic Publishing Company. Londregan, John, and Keith Poole. 1990. “Poverty, the Coup Trap, and the Seizure of Executive Power.” World Politics 42: 151– 83. Loriaux, Michael. 1991. France after Hegemony: International Change and Financial Reform. Ithaca: Cornell University Press. Lowenfeld, Andreas F. 1994. “Remedies Along with Rights: Institutional Reform in the New GATT.” American Journal of International Law 88 (July): 477–88. Luce, R. D., and Howard Raiffa. 1957. Games and Decisions. New York: Wiley. Ludlow, Peter. 1982. The Making of the European Monetary System: A Case Study of the Politics of the European Community. London: Butterworth Scientific. Luif, Paul. 1991. “Austria.” Chap. 10 in Helen Wallace, ed., The Wider Western Europe: Reshaping the EC/EFTA Relationship. London: Pinter Publishers. Lukes, Steven. 1974. Power: A Radical View. London: Macmillan Press. Lustig, Nora. 1992. Mexico: The Remaking of an Economy. Washington, D.C.: The Brookings Institution. ———. 1994. “NAFTA: Potential Impact on Mexico’s Economy and Beyond.” Pp. 46 – 80 in Roberto Bouzas and Jaime Ros, eds., Economic Integration in the Western Hemisphere. Notre Dame, Ind.: University of Notre Dame Press. Lustig, Nora, Bosworth, Barry P., and Lawrence, Robert Z., eds. 1992. North American Free Trade Agreement: Assessing the Impact. Washington, D.C.: The Brookings Institution. Mace, Gordon, and Gérard Hervouet. 1989. “Canada’s Third Option: A Complete Failure?” Canadian Public Policy 15: 387– 404. Mackenzie, Hugh. 1988. “Free Trade and the Auto Industry.” In Duncan Cameron, ed., The Free Trade Deal. Toronto: James Lorimer. Madison, James, Alexander Hamilton, and John Jay. [1788] 1987. The Federalist Papers. London: Penguin Books. Magee, Stephen P., William A. Brock and Leslie Young. 1989. Black Hole Tariffs and Endogenous Policy Theory: Political Economy in General Equilibrium. New York: Cambridge University Press. Mansfield, Edward D. 1994. Power, Trade, and War. Princeton: Princeton University Press. Manzetti, Luigi. 1993/94. “The Political Economy of Mercosur.” Journal of Interamerican Studies and World Affairs 35: 101– 41. March, James G., and Johan P. Olsen. 1990. Rediscovering Institutions: The Organizational Basis of Politics. New York: Free Press. Marks, Gary. 1992. “Structural Policy in the European Community.” Chap. 6 in Alberta M. Sbragia, ed., Euro-Politics: Institutions and Policymaking in the “New” European Community. Washington, D.C.: The Brookings Institution. Marsh, David. 1992. The Most Powerful Bank: Inside Germany’s Bundesbank. London: Heinemann. Marston, Richard C. 1993. “Interest Differentials under Bretton Woods and the PostBretton Woods Float: The Effects of Capital Controls and Exchange Risk.” Pp.
296
B I B L I O G R A P H Y
515 – 38 in Michael D. Bordo and Barry Eichengreen, eds., A Retrospective on the Bretton Woods System: Lessons for International Monetary Reform. Chicago: University of Chicago Press. Martin, Lisa L. 1992a. Coercive Cooperation: Explaining Multilateral Economic Sanctions. Princeton: Princeton University Press. ———. 1992b. “Interests, Power, and Multilateralism.” International Organization 46: 765 – 92. ———. 1993. “Credibility, Costs, and Institutions: Cooperation on Economic Sanctions.” World Politics 45: 406 – 32. ———. 1994. “International and Domestic Institutions in the EMU Process.” Chap. 5 in Barry Eichengreen and Jeffry A. Frieden, eds., The Political Economy of European Monetary Unification. Boulder, Colo.: Westview Press. Martin, Lisa L., and Beth A. Simmons. 1998. “Theories and Empirical Studies of International Institutions.” International Organization 52: 729–57. Mastanduno, Michael. 1991. “Do Relative Gains Matter? America’s Response to Japanese Industrial Policy.” International Security 16: 73–113. Mattli, Walter, and Anne-Marie Slaughter. 1995. “Law and Politics in the European Union: A Reply to Garrett.” International Organization 49: 183–90. Mayhew, David R. 1974. Congress: The Electoral Connection. New Haven: Yale University Press. McCalla, Robert. 1996. “NATO’s Persistence after the Cold War.” International Organization 50: 445 –75. McCallum, John. 1995. “National Borders Matter: U.S.-Canada Regional Trade Barriers.” American Economic Review 85: 615 –23. McCarthy, Patrick. 1990. “France Faces Reality: Rigueur and the Germans.” Pp. 25– 78 in David P. Calleo and Claudia Morgenstern, eds., Recasting Europe’s Economies: National Strategies in the 1980s. Lanham, Md.: University Press of America. McDonald, Frank, and George Zis. 1988. “The European Monetary System: Towards 1992 and Beyond.” Journal of Common Market Studies 27: 183–202. McKinnon, Ronald. 1963. “Optimum Currency Areas.” American Economic Review 53: 717–25. McNamara, Kathleen R. 1998. The Currency of Ideas: Monetary Politics in the European Union. Ithaca: Cornell University Press. McQuaid, Linda. 1991. The Quick and the Dead: Brian Mulroney, Big Business, and the Seduction of Canada. Toronto: Viking. McQueen, Rod. 1985. Leap of Faith: The Macdonald Report. Toronto: Cowan & Company. Mearsheimer, John J. 1990. “Back to the Future: Instability in Europe after the Cold War.” International Security 15: 5 – 56. ———. 1994/95. “The False Promise of International Institutions.” International Security 19: 5 – 49. Melitz, Jacques. 1988. “Monetary Discipline and Cooperation in the European Monetary System: A Synthesis.” Pp. 51– 84 in Francesco Giavazzi, Stefano Micossi, and Marcus Miller, eds., The European Monetary System. Cambridge: Cambridge University Press. Michalski, Anna, and Helen Wallace. 1993. The European Community: The Challenge of Enlargement. Washington, D.C.: The Brookings Institution. Milgrom, Paul R., Douglass C. North, and Barry R. Weingast. 1990. “The Role of
B I B L I O G R A P H Y
297
Institutions in the Revival of Trade: The Medieval Law Merchant, Private Judges, and the Champagne Fairs.” Economics and Politics 2: 1–23. Milgrom, Paul, and John Roberts. 1990. “Bargaining Costs, Influence Costs, and the Organization of Economic Activity.” Chap. 3 in James E. Alt and Kenneth A. Shepsle, eds., Perspectives on Positive Political Economy. Cambridge: Harvard University Press. ———. 1992. Economics, Organization, and Management. Englewood Cliffs, N.J.: Prentice-Hall. Miller, Marcus, and Alan Sutherland. 1992. “Britain’s Return to Gold and Entry into the EMS: Joining Conditions and Credibility.” In Paul Krugman and Marcus Miller, eds., Exchange Rate Targets and Currency Bands. Cambridge: Cambridge University Press. Milner, Helen. 1991. “The Assumption of Anarchy in International Relations Theory.” Review of International Studies 17: 67– 85. ———. 1992. “International Theories of Cooperation among Nations: Strengths and Weaknesses.” World Politics 44: 466 – 96. ———. 1997. Interests, Institutions, and Information: Domestic Politics and International Relations. Princeton: Princeton University Press. Mitchell, Ronald B. 1994. “Regime Design Matters: International Oil Pollution and Treaty Compliance.” International Organization 48: 425–58. Mitrany, David. 1966. A Working Peace System. Chicago: Quadrangle. Moe, Terry M. 1984. “The New Economics of Organization.” American Journal of Political Science 28: 739–77. ———. 1990a. “Political Institutions: The Neglected Side of the Story.” Journal of Law, Economics, and Organization 6: 213 – 53. ———. 1990b. “The Politics of Structural Choice: Toward a Theory of Public Bureaucracy.” In Oliver E. Williamson, ed., Organization Theory: From Chester Barnard to the Present and Beyond. New York: Oxford University Press. Moe, Terry M., and Michael Caldwell. 1994. “The Institutional Foundations of Democratic Government: A Comparison of Presidential and Parliamentary Systems.” Journal of Institutional and Theoretical Economics 50: 171–95. Monnet, Jean. 1978. Memoirs. Translated by Richard Mayne. Garden City, N.Y.: Doubleday. Moravcsik, Andrew. 1991. “Negotiating the Single Act: National Interests and Conventional Statecraft in the European Community.” International Organization 45: 19 – 56. ———. 1998. The Choice for Europe: Social Purpose and State Power from Messina to Maastricht. Ithaca: Cornell University Press. Morel, Benoit, and Kyle Olson, eds. 1993. Shadows and Substance: The Chemical Weapons Convention. Boulder, Colo.: Westview Press. Morgenthau, Hans J. 1958. Politics among Nations: The Struggle for Power and Peace. 6th ed. New York: Knopf. Morici, Peter, ed. 1990. Making Free Trade Work: The Canada-U.S. Agreement. New York: Council on Foreign Relations. Morris, Stephen D. 1992. “Political Reformism in Mexico: Salinas at the Brink.” Journal of Interamerican Studies and World Affairs 34: 27–58. Morrisett, Peter M. 1989. “The Evolution of Policy Responses to Stratospheric Ozone Depletion.” Natural Resources Journal 29: 793–820.
298
B I B L I O G R A P H Y
Morrow, James D. 1994. “Modeling the Forms of International Cooperation: Distribution versus Information.” International Organization 48: 387–423. Moyer, Homer E., Jr. 1993. “Chapter 19 of the NAFTA: Binational Panels as the Trade Courts of Last Resort.” International Lawyer 27: 707–26. Mundell, Robert. 1961. “A Theory of Optimum Currency Areas.” American Economic Review 51: 657– 65. North, Douglass C., and Barry R. Weingast. 1989. “Constitution and Commitment: The Evaluation of Institutions Governing Public Choice in Seventeenth-Century England.” Journal of Economic History 49: 803–32. Nye, Joseph S., Jr. 1971. “Comparing Common Markets: A Revised Neo-functional Model.” In Leon N. Lindberg and Stuart A. Scheingold, eds., Regional Integration: Theory and Research. Cambridge: Harvard University Press. Oatley, Thomas H. 1997. Monetary Politics: Exchange Rate Cooperation in the European Union. Ann Arbor: University of Michigan Press. Oatley, Thomas, and Robert Nabors. 1998. “Redistributive Cooperation: Market Failure, Wealth Transfers, and the Basle Accord.” International Organization 52: 35 – 54. OECD. 1992. Economic Outlook—Historical Statistics, 1960–1990. Paris: Organization for Economic Cooperation and Development. O’Halloran, Sharyn. 1994. Politics, Process, and American Trade Policy. Ann Arbor: University of Michigan Press. Olson, Mancur. 1965. The Logic of Collective Action. Cambridge: Harvard University Press. ———. 1982. The Rise and Decline of Nations: Economic Growth, Stagflation, and Social Rigidities. New Haven: Yale University Press. Ostrom, Elinor. 1990. Governing the Commons: The Evolution of Institutions for Collective Action. New York: Cambridge University Press. Oudiz, Gilles, and Henri Sterdyniak. 1985. “Inflation, Employment, and External Constraints.” Pp. 9 – 50 in Jacques Melitz and Charles Wyplosz, eds., The French Economy: Theory and Policy. Boulder, Colo.: Westview Press. Oye, Kenneth A., ed. 1986a. Cooperation under Anarchy. Princeton: Princeton University Press. ———. 1986b. “Explaining Cooperation under Anarchy: Hypotheses and Strategies.” Chap. 1 in Kenneth A. Oye, ed., Cooperation under Anarchy. Princeton: Princeton University Press. ———. 1992. Economic Discrimination and Political Exchange: World Political Economy in the 1930s and 1980s. Princeton: Princeton University Press. Padoa-Schioppa, Tommaso. 1986. “Lessons from the EMS.” Lecture at the European University Institute, February 20. Reprinted in Banca d’Italia, Economic Bulletin 2 (February). “Pan-American Trade: The Mirage That Won’t Go Away.” 1997. Economist, May 10, p. 42. Pape, Robert A. 1997. “Why Economic Sanctions Do Not Work.” International Security 22: 90 –136. Parisi, Arturo, and Gianfranco Pasquino. 1980. “Changes in Italian Electoral Behaviour: The Relationships between Parties and Voters.” Pp. 6–30 in Peter Lange and Sidney Tarrow, eds., Italy in Transition: Conflict and Consensus. London: Frank Cass and Company. Parson, Edward A. 1993. “Protecting the Ozone Layer.” Chap. 2 in Peter M. Haas,
B I B L I O G R A P H Y
299
Robert O. Keohane, and Marc A. Levy, eds., Institutions for the Earth: Sources of Effective Environmental Protection. Cambridge: MIT Press. Pastor, Manuel, and Carol Wise. 1994. “The Origins and Sustainability of Mexico’s Free Trade Policy.” International Organization 48: 459–89. ———. 1997. “State Policy, Distribution, and Neoliberal Reform in Mexico.” Journal of Latin American Studies 29: 419– 56. Pastor, Robert A. 1980. Congress and the Politics of U.S. Foreign Economic Policy, 1929 –1979. Berkeley: University of California Press. ———. 1990. “Salinas Takes a Gamble: Mexico’s Postrevolutionary Opening.” New Republic, September 10 and 17, pp. 27– 32. Peres Núñez, Wilson. 1990. “Foreign Direct Investment and Industrial Development in Mexico.” Paris: OECD. Perrow, Charles. 1986. Complex Organizations: A Critical Essay. 3d ed. New York: Random House. Pescatore, Pierre. 1993. “The GATT Dispute Settlement Mechanism: Its Present Situation and Its Prospects.” Journal of World Trade 27: 5 –20. Pierson, Paul. 1995. Dismantling the Welfare State? Reagan, Thatcher, and the Politics of Retrenchment. Cambridge: Cambridge University Press. Poitras, Guy, and Raymond Robinson. 1994. “The Politics of NAFTA in Mexico.” Journal of Interamerican Studies and World Affairs 36: 1–35. Pollack, Mark A. 1997. “Delegation, Agency, and Agenda Setting in the European Community.” International Organization 51: 99 –134. Popkin, Samuel L. 1979. The Rational Peasant: The Political Economy of Rural Society in Vietnam. Berkeley: University of California Press. ———. 1994. The Reasoning Voter: Communication and Persuasion in Presidential Campaigns. 2d ed. Chicago: University of Chicago Press. Posen, Barry R. 1984. The Sources of Military Doctrine: France, Britain, and Germany between the World Wars. Ithaca: Cornell University Press. Posner, Alan R. 1978. “Italy: Dependence and Political Fragmentation.” Pp. 225–54 in Peter J. Katzenstein, ed., Between Power and Plenty: Foreign Economic Policies of Advanced Industrial States. Madison: Wisconsin University Press. Powell, Robert. 1992. “Absolute and Relative Gains in International Relations Theory.” American Political Science Review 85: 1303–20. ———. 1993. “The Relative-Gains Problem for International Relations.” American Political Science Review 87: 735 – 37. Przeworski, Adam. 1985. Capitalism and Social Democracy. Cambridge: Cambridge University Press. Przeworski, Adam, and John Sprague. 1986. Paper Stones. Chicago: University of Chicago Press. Puchala, Donald J., and Raymond F. Hopkins. 1983. “International Regimes: Lessons from Inductive Analysis.” Pp. 61– 91 in Stephen D. Krasner, ed., International Regimes. Ithaca: Cornell University Press. Putnam, Robert D. 1988. “Diplomacy and Domestic Politics: The Logic of TwoLevel Games.” International Organization 42: 427–60. Putnam, Robert D., and Nicholas Bayne. 1986. Hanging Together: The Seven-Power Summits. Cambridge: Harvard University Press. Putnam, Robert D., and C. Randall Henning. 1989. “The Bonn Summit of 1978: A Case Study in Coordination.” Pp. 12–140 in Richard N. Cooper, Barry Eichengreen, Gerald Hotham, Robert D. Putnam, and C. Randall Henning, eds., Can Na-
300
B I B L I O G R A P H Y
tions Agree? Issues in International Economic Cooperation. Washington, D.C.: The Brookings Institution. Raiffa, Howard. 1982. The Art and Science of Negotiation. Cambridge: Harvard University Press. Ramirez de la O, Rogelio. 1993. “The North American Free Trade Agreement from a Mexican Perspective.” Pp. 60– 86 in Steven Globerman and Michael Walker, eds., Assessing NAFTA: A Trinational Analysis. Vancouver: Fraser Institute. Randall, Stephen J., ed. 1992. North America without Borders? Integrating Canada, the United States, and Mexico. Calgary: University of Calgary Press. Rapoport, Anatol. 1960. Fights, Games and Debates. Ann Arbor: University of Michigan Press. Reding, Andrew. 1991. “Mexico: The Crumbling of the ‘Perfect Dictatorship.’” World Policy Journal 8: 255 – 84. Riker, William H. 1980. “Implications from the Disequilibrium of Majority Rule for the Study of Institutions.” American Political Science Review 74: 432–46. ———. 1996. The Strategy of Rhetoric: Campaigning for the American Constitution. New Haven: Yale University Press. Risse-Kappen, Thomas. 1996. “Exploring the Nature of the Beast: International Relations Theory and Comparative Policy Analysis Meets the European Union.” Journal of Common Market Studies 34: 53 – 80. Roan, Sharon L. 1990. Ozone Crisis: The 15-Year Evolution of a Sudden Global Emergency. New York: Wiley. Robinson, Ian. 1993. North American Trade As If Democracy Mattered: What’s Wrong with NAFTA and What Are the Alternatives? Ottawa: Canadian Centre for Policy Alternatives. Rodrik, Dani. 1997. Has Globalization Gone Too Far? Washington, D.C.: Institute for International Economics. Rohter, Larry. 1997. “Backlash from NAFTA Batters Economies of the Caribbean.” New York Times, January 30, A1. Ros, Jaime. 1992. “Free Trade Area or Common Capital Market? Notes on MexicoU.S. Economic Integration and Current NAFTA Negotiations.” Journal of Interamerican Studies and World Affairs 34: 53 – 91. Rubinstein, Ariel. 1982. “Perfect Equilibrium in a Bargaining Model.” Econometrica 50: 97–109. ———. 1985. “A Bargaining Model with Incomplete Information about Time Preferences.” Econometrica 53: 1151–72. Ruggie, John G. 1983. “International Regimes, Transactions, and Change: Embedded Liberalism in the Postwar Economic Order.” Pp. 193–231 in Stephen D. Krasner, ed., International Regimes. Ithaca: Cornell University Press. ———. 1992. “Multilateralism: The Anatomy of an Institution.” International Organization 46: 561– 98. ———. 1993. “Territoriality and Beyond: Problematizing Modernity in International Relations.” International Organization 47: 139–74. Rugman, Alan M., and Andrew D. M. Anderson. 1987. Administered Protection in America. London: Croom Helm. Russell, Peter H. 1992. Constitutional Odyssey: Can Canadians Become a Sovereign People? Toronto: University of Toronto Press. Saborio, Sylvia. 1992. “Overview: The Long and Winding Road from Anchorage to Patagonia.” Pp. 3 – 31 in Sylvia Saborio, ed., The Premise and the Promise: Free Trade in the Americas. New Brunswick, N.J.: Transaction Publishers.
B I B L I O G R A P H Y
301
Sachs, Jeffrey, and Charles Wyplosz. 1986. “France under Mitterrand.” Economic Policy 2: 261– 321. Salinas de Gortari, Carlos. 1990. “Cinco premisas sobre las relaciones comerciales con el exterior.” Comercio exterior 40, no. 6 ( June): 524–25. Sandel, Michael J. 1996. Democracy’s Discontent: America in Search of a Public Philosophy. Cambridge: Harvard University Press. Sandholtz, Wayne. 1993. “Choosing Union: Monetary Politics and Maastricht.” International Organization 47: 1– 39. Sandholtz, Wayne, and John Zysman. 1989. “1992: Recasting the European Bargain.” World Politics 42: 95 –128. Sapir, André, and Lars Lundberg. 1984. “The U.S. Generalized System of Preferences and Its Impacts.” Chap. 6 in Robert E. Baldwin and Anne O. Krueger, eds., The Structure and Evolution of Recent U.S. Trade Policy, Chicago: University of Chicago Press. Sargent, Thomas J., and Neil Wallace. 1981. “Some Unpleasant Monetarist Arithmetic.” Federal Reserve Bank of Minneapolis Quarterly Review 9, no. 1 (Winter): 15 – 31. Sassoon, Donald. 1981. The Strategy of the Italian Communist Party: From the Resistance to the Historic Compromise. New York: St. Martin’s Press. Saxonhouse, Gary R. 1993. “Trading Blocs and East Asia.” Chap. 12 in Jaime de Melo and Arvind Panagariya, eds., New Dimensions in Regional Integration. Cambridge: Cambridge University Press. Scharpf, Fritz W. 1987 (trans. 1991). Crisis and Choice in European Social Democracy. Ithaca: Cornell University Press. Schattschneider, E. E. 1935. Politics, Pressure, and the Tariff: A Study of Free Private Enterprise in Pressure Politics, as Shown in the 1929 –1930 Revision of the Tariff. Hamden, Conn.: Archon Books. ———. 1960. The Semisovereign People: A Realist’s View of Democracy in America. New York: Holt, Rinehart, and Winston. Schelling, Thomas. 1960. The Strategy of Conflict. Cambridge: Harvard University Press. Schott, Jeffrey J. 1988. United States-Canada Free Trade: An Evaluation of the Agreement. Washington, D.C.: Institute for International Studies. ———. 1989. “More Free Trade Areas?” Chap. 1 in Jeffrey J. Schott, ed., Free Trade Areas and U.S. Trade Policy. Washington, D.C.: Institute for International Economics. Schott, Jeffrey J., and Murray G. Smith, eds. 1988. The Canada–United States Free Trade Agreement: The Global Impact. Washington, D.C.: Institute for International Economics. Schwanen, Daniel. 1993. A Growing Success: Canada’s Trade Performance under Free Trade. Toronto: C. D. Howe Institute. Scott, James C. 1976. The Moral Economy of the Peasant: Rebellion and Subsistence in Southeast Asia. New Haven: Yale University Press. Sebenius, James K. 1983. “Negotiation Arithmetic: Adding and Subtracting Issues and Parties.” International Organization 37: 281–316. ———. 1991. “Designing Negotiations toward a New Regime: The Case of Global Warming.” International Security 15: 110 – 48. Serra Puche, Jaime. 1990. “Lineamientos para una estrategia comercial.” Comercio exterior 40, no. 6 ( June): 526– 30. Shepsle, Kenneth A. 1986. “Institutional Equilibrium and Equilibrium Institutions.”
302
B I B L I O G R A P H Y
In Herbert F. Weisberg, ed., Political Science: The Science of Politics. New York: Agathon Press. Shepsle, Kenneth A., and Barry R. Weingast. 1981. “Structure-Induced Equilibrium and Legislative Choice.” Public Choice 37: 503–19. Shugart, Matthew, and John M. Carey. 1992. Presidents and Assemblies: Constitutional Design and Electoral Dynamics. New York: Cambridge University Press. Shulman, Jennifer B., and Harold K. Jacobson. 1998. “The Proliferation of International Organizations: Building a New International Order?” Paper prepared for the annual meetings of the American Political Science Association, Boston, September. Silva-Herzog, Jesus. 1991. “Why the Rush to Free Trade?” New Perspectives Quarterly 8, no. 1 (Winter): 29 – 32. Simmons, Beth A. 1994. Who Adjusts? Domestic Sources of Foreign Economic Policy during the Interwar Years. Princeton: Princeton University Press. Simon, Bernard. 1992. “Free Trade Is No Longer Such a Hot Potato.” Financial Times, December 9. ———. 1993a. “Canada Hails Victory on NAFTA Talks.” Financial Times, August 16. ———. 1993b. “Canada Liberals Pledge to Renegotiate NAFTA.” Financial Times, September 16. Simon, Herbert A. 1947. Administrative Behavior. New York: Macmillan. Simpson, Jeffrey. 1988. Spoils of Power: The Politics of Patronage. Toronto: HarperCollins. Smith, Adam. [1776] 1965. The Wealth of Nations. New York: Random House. Smith, David. 1993. From Boom to Bust: Trial and Error in British Economic Policy. London: Penguin. Smith, Jay. 1994. “Dispute Settlement in NAFTA.” Stanford University. Manuscript. Snidal, Duncan. 1985a. “Coordination versus Prisoners’ Dilemma: Implications for International Cooperation and Regimes.” American Political Science Review 79: 923 – 42. ———. 1985b. “The Limits of Hegemonic Stability Theory.” International Organization 39: 579– 614. ———. 1991. “Relative Gains and the Pattern of International Cooperation.” American Political Science Review 85: 701–26. ———. 1996. “Political Economy and International Institutions.” International Review of Law and Economics 16: 121– 37. Snyder, Glenn H. 1991. “Alliances, Balance, and Stability.” International Organization 45: 121– 42. Snyder, Jack. 1991. Myths of Empire: Domestic Politics and International Ambition. Ithaca: Cornell University Press. Spaventa, Luigi. 1980. “Italy Joins the EMS—A Political History.” Occasional paper no. 32, Johns Hopkins University, Bologna Center. Spotts, Frederic, and Theodor Wieser. 1986. Italy: A Difficult Democracy. Cambridge: Cambridge University Press. Sprinz, Detlef, and Tapani Vaahtoranta. 1994. “The Interest-Based Explanation of International Environmental Policy.” International Organization 48: 77–105. Stålvant, Carl-Einar, and Carl Hamilton. 1991. “Sweden.” Chap. 14 in Helen Wallace, ed., The Wider Western Europe: Reshaping the EC/EFTA Relationship. London: Pinter Publishers.
B I B L I O G R A P H Y
303
Steger, Debra A. 1988. A Concise Guide to the Canada–United States Free Trade Agreement. Vol. 11 in the Research Studies of the Royal Commission on the Economic Union and Development Prospects for Canada. Toronto: University of Toronto. Stein, Arthur. 1983. “Coordination and Collaboration: Regimes in an Anarchic World.” Pp. 115– 40 in Stephen D. Krasner, ed., International Regimes. Ithaca: Cornell University Press. ———. 1990. Why Nations Cooperate: Circumstance and Choice in International Relations. Ithaca: Cornell University Press. Stephens, Philip. 1990. “Philip Stephens Chronicles the Drama behind the Government’s Decision.” Financial Times, October 6/7. Stevens, William K. 1993. “Scientists Report an Easing in Ozone-Killing Chemicals.” New York Times, August 26. Stokes, Susan. 1999. “Mandates, Neoliberalism, and Democracy.” University of Chicago. Manuscript. Strange, Susan. 1983. “Cave! hic dragones: A Critique of Regime Analysis.” Pp. 337– 54 in Stephen D. Krasner, ed., International Regimes. Ithaca: Cornell University Press. ———. 1996. The Retreat of the State: The Diffusion of Power in the World Economy. Cambridge: Cambridge University Press. Streeck, Wolfgang, and Philippe Schmitter. 1991. “From National Corporatism to Transnational Pluralism.” Politics and Society 19: 133–64. Sutton, John. 1986. “Non-cooperative Bargaining Theory: An Introduction.” Review of Economic Studies 53: 709 –24. Tabellini, Guido. 1988. “Monetary and Fiscal Policy Coordination with a High Public Debt.” Pp. 90–134 in Francesco Giavazzi and Luigi Spaventa, eds., High Public Debt: The Italian Experience. Cambridge: Cambridge University Press. Taylor, Michael. 1987. The Possibility of Cooperation. New York: Cambridge University Press. Taylor, Paul. 1980. “Interdependence and Autonomy in the European Communities: The Case of the European Monetary System.” Journal of Common Market Studies 18: 370– 87. ———. 1982. “Intergovernmentalism in the European Communities in the 1970s: Patterns and Perspectives.” International Organization 36: 741–66. ———. 1983. The Limits of European Integration. New York: Columbia University Press. Thatcher, Margaret. 1993. The Downing Street Years. New York: HarperCollins. Thygesen, Niels. 1984. “Exchange-Rate Policies and Monetary Targets in the EMS Countries.” In Rainer S. Masera and Robert Triffin, eds., Europe’s Money: Problems of European Monetary Co-ordination and Integration. New York: Oxford University Press. Tollison, Robert D., and Thomas D. Willett. 1979. “An Economic Theory of Mutually Advantageous Issue Linkages in International Negotiations.” International Organization 33: 425 – 49. Tornell, Aaron, and Gerardo Esquivel. 1995. “The Political Economy of Mexico’s Entry to NAFTA.” National Bureau of Economic Research Working Paper no. 5322, October. Trebilcock, Michael J., and Robert C. York, eds. 1990. Fair Exchange: Reforming Trade Remedy Laws. Toronto: C. D. Howe Institute.
304
B I B L I O G R A P H Y
Triffin, Robert. 1979. “The American Response to the European Monetary System.” In Philip H. Trezise, ed., The European Monetary System: Its Promise and Prospects. Washington, D.C.: The Brookings Institution. Truman, David B. 1951. The Governmantal Process. New York: Alfred Knopf. Tsebelis, George. 1995. “Decision Making in Political Systems: Veto Players in Presidentialism, Parliamentarism, Multicameralism, and Multipartyism.” British Journal of Political Science 25: 289– 325. Tsoukalis, Loukas. 1989. “The Political Economy of the European Monetary System.” Pp. 58 – 84 in Paolo Guerrieri and Pier Carlo Padoan, eds., The Political Economy of European Integration: States, Markets, and Institutions. Savage, Md.: Barnes and Noble. Tufte, Edward R. 1978. Political Control of the Economy. Princeton: Princeton University Press. Ungerer, Horst, Owen Evans, Thomas Mayer, and Philip Young. 1986. The European Monetary System: Recent Developments. IMF Occasional Paper no. 48. United Nations. 1989. International Trade Statistics Yearbook, 1987. New York: United Nations. Van Evera, Stephen W. 1984. “The Cult of the Offensive and the Origins of the First World War.” International Security 9: 58 –107. ———. 1990/91. “Primed for Peace: Europe after the Cold War.” International Security 15: 33 – 40. Van Ypersele, Jacques, with Jean-Claude Koeune. 1985. The European Monetary System: Origins, Operation, and Outlook. Brussels: Commission of the European Communities. Verdier, Daniel. 1994. Democracy and International Trade: Britain, France, and the United States, 1860 –1990. Princeton: Princeton University Press. Viner, Jacob. 1950. The Customs Issue. New York: Carnegie Endowment for International Peace. Vogel, David. 1995. Trading Up: Consumer and Environmental Regulation in a Global Economy. Cambridge: Harvard University Press. Wagner, R. Harrison. 1988. “Economic Interdependence, Bargaining Power, and Political Influence.” International Organization 42: 461–83. Walt, Stephen M. 1987. The Origins of Alliances. Ithaca: Cornell University Press. Waltz, Kenneth N. 1959. Man, the State, and War. New York: Columbia University Press. ———. 1979. Theory of International Politics. Reading, Mass.: Addison-Wesley. ———. 1986. “Reflections on Theory of International Politics: A Response to My Critics.” Chap. 11 in Robert O. Keohane, ed., Neorealism and Its Critics. New York: Columbia University Press. ———. 1990. “Nuclear Myths and Political Realities.” American Political Science Review 84: 731– 45. Watkins, G. C. 1993. “NAFTA and Energy: A Bridge Not Far Enough?” Pp. 193– 225 in Steven Globerman and Michael Walker, eds., Assessing NAFTA: A Trinational Analysis. Vancouver: Fraser Institute. Weaver, R. Kent, and Bert A. Rockman, eds. 1993. Do Institutions Matter? Government Capabilities in the United States and Abroad. Washington, D.C.: The Brookings Institution. Weber, Axel A. 1991. “Reputation and Credibility in the European Monetary System.” Economic Policy 12: 57–102.
B I B L I O G R A P H Y
305
Weiler, J. H. H. 1991. “The Transformation of Europe.” Yale Law Journal 100: 2403 – 83. Weingast, Barry R., and William Marshall. 1988. “The Industrial Organization of Congress.” Journal of Political Economy 96: 132–63. Weintraub, Sidney. 1990a. “The Impact of the Agreement on Mexico.” Chapter 5 in Peter Morici, ed., Making Free Trade Work: The Canada–U.S. Agreement. New York: Council on Foreign Relations. ———. 1990b. A Marriage of Convenience: Relations between Mexico and the United States. New York: Oxford University Press. ———. 1990c. “The North American Free Trade Debate.” Washington Quarterly (Autumn): 119 – 30. ———. 1993. “The Economy on the Eve of Free Trade.” Current History 92 (February): 67–77. ———. 1994. NAFTA: What Comes Next? Washington, D.C.: Center for Strategic and International Studies. Weintraub, Sidney, and M. Delal Baer. 1992. “The Interplay between Economic and Political Opening: The Sequence in Mexico.” Washington Quarterly (Spring): 187–201. Wendt, Alexander E. 1987. “The Agent-Structure Problem in International Relations Theory.” International Organization 41: 335 –70. ———. 1992. “Anarchy Is What States Make of It: The Social Construction of Power Politics.” International Organization 46: 391– 425. ———. 1994. “Collective Identity Formation and the International State.” American Political Science Review 88: 384 – 96. Wertheimer, Alan. 1987. Coercion. Princeton: Princeton University Press. Whalley, John. 1992. “CUSTA and NAFTA: Can WHFTA Be Far Behind?” Journal of Common Market Studies 30: 125 – 41. ———. 1993. “Regional Trade Arrangements in North America: CUSTA and NAFTA.” Chap. 11 in Jaime de Melo and Arvind Panagariya, eds., New Dimensions in Regional Integration. Cambridge: Cambridge University Press. Whalley, John, and Colleen Hamilton. 1996. The Trading System after the Uruguay Round. Washington, D.C.: Institute for International Economics. White, Randall. 1988. Fur Trade to Free Trade: Putting the Canada-U.S. Trade Agreement in Historical Perspective. Toronto: Dundurn Press. Whitney, Craig R. 1997. “On French Campaign Trail, Few Support New Currency.” New York Times, April 28, A8. Wiarda, Howard J. 1994. “The U.S. Domestic Politics of the U.S.-Mexico Free Trade Agreement.” Chap. 5 in M. Delal Baer and Sidney Weintraub, eds., The NAFTA Debate: Grappling with Unconventional Trade Issues. Boulder, Colo.: Lynne Rienner Publishers. Wijkman, Per Magnus. 1994. “EFTA Countries.” Chap. 4 in C. Randall Henning, Eduard Hochreiter, and Gary Clyde Hufbauer, eds., Reviving the European Union. Washington, D.C.: Institute for International Economics. Williamson, Oliver E. 1985. The Economic Institutions of Capitalism. New York: Free Press. Winham, Gilbert R. 1993. “Dispute Settlement in NAFTA and the FTA.” Pp. 251– 70 in Steven Globerman and Michael Walker, eds., Assessing NAFTA: A Trinational Analysis. Vancouver: Fraser Institute. Wolf, Martin. 1996. “The Dog That Failed to Bark: The Climate for Trade Policy in
306
B I B L I O G R A P H Y
the European Union.” Chap. 7 in Jeffrey J. Schott, ed., The World Trading System: Challenges Ahead. Washington, D.C.: Institute for International Economics. Wonnacott, Paul, and Mark Lutz. 1989. “Is There a Case for Free Trade Areas?” Chap. 2 in Jeffrey J. Schott, ed., Free Trade Areas and U.S. Trade Policy. Washington, D.C.: Institute for International Economics. Wonnacott, Ronald J. 1991a. “Canada’s Role in the U.S.-Mexico Free Trade Negotiations.” World Economy 14: 79 – 86. ———. 1991b. The Economics of Overlapping Free Trade Areas and the Mexican Challenge. Toronto and Washington, D.C.: C. D. Howe Institute and National Planning Association. ———. 1991c. “Mexico and the Canada-U.S. Auto Pact.” World Economy 14: 103– 11. ———. 1996. “Free-Trade Agreements: For Better or Worse?” American Economic Association Papers and Proceedings 87: 62– 66. Wood, Gordon S. 1969. The Creation of the American Republic, 1776–1787. New York: W.W. Norton. Woolley, John T. 1992. “Policy Credibility and European Monetary Institutions.” Chap. 5 in Alberta M. Sbragia, ed., Euro-Politics: Institutions and Policymaking in the “New” European Community. Washington, D.C.: The Brookings Institution. ———. 1994. “Linking Political and Monetary Union: The Maastricht Agenda and German Domestic Politics.” Chap. 4 in Barry Eichengreen and Jeffry A. Frieden, eds., The Political Economy of European Monetary Unification. Boulder, Colo.: Westview Press. World Bank. 1997. World Development Indicators. CD-ROM. Wright, Robert. 1995. “Be Very Afraid: Nukes, Nerve Gas and Anthrax Spores.” New Republic, May 1: 19 –27. Yarbrough, Beth V., and Robert M. Yarbrough. 1990. “International Institutions and the New Economics of Organization.” International Organization 44: 235–59. ———. 1992. Cooperation and Governance in International Trade: The Strategic Organizational Approach. Princeton: Princeton University Press. Young, Alwyn. 1994. “The Tyranny of Numbers: Confronting the Statistical Realities of the East Asian Growth Experience.” National Bureau of Economic Research, Working Paper no. 4680. Young, Oran R. 1983. “Regime Dynamics: The Rise and Fall of International Regimes.” Pp. 93 –113 in Stephen D. Krasner, ed., International Regimes. Ithaca: Cornell University Press. ———. 1986. “International Regimes: Toward a New Theory of Institutions.” World Politics 39: 118–19. ———. 1994. International Governance: Protecting the Environment in a Stateless Society. Ithaca: Cornell University Press. Young, Oran R., and Gail Osherenko, eds. 1993. Polar Politics: Creating International Environmental Regimes. Ithaca: Cornell University Press.
I N D E X
Abbott, Kenneth W., 21n, 29n, 62n, 75n, 83n agenda-setting power, 12n, 40 – 43, 79, 276 – 77. See also go-it-alone power Aggarwal, Vinod K., 262n Aguilar Zinser, Adolfo, 151n Alchian, Armen A., 73n Alesina, Alberto, 56n, 224, 264n Alho, Kari, 164n Alt, James, 208n Alter, Alison, 197n Alvarez, R. Michael, 193n Andean Pact, 161–62 Anderson, Andrew D. M., 100n Anderson, Kym, 164n Andreotti, Giulio, 180 – 81, 183, 186, 192– 94, 197, 218 Andrews, David M., 42n antitrust law, xiii, 112n Antola, Esko, 164n APEC (Asia-Pacific Economic Cooperation), 165–67, 262 Articles of Confederation, 264 – 68 Artis, Michael J., 173n, 204 Austria, accession to EU of, 163 – 65 Axelrod, Robert, 65– 66, 78n, 106n Bachrach, Peter, 41–42, 276 –77 Baer, M. Delal, 124n, 126n, 127n Baldwin, David A., 19n, 38n Baldwin, Richard E., 164 – 65 Baldwin, Robert E., 100n, 101n Balladur, Edouard, 240 bandwagoning: in Asia, 165 – 66; in developing world, 160–63, 165 – 67; and EMS, 171, 178, 180–91, 200 – 4, 212, 233; and EMU, 247–48; in Europe, 163 – 65, 252; and GATT, 166–67; and go-it-alone power, 47–50, 255; and NAFTA, 97, 123, 127n, 129–38, 160 – 63, 252. See also goit-alone power Baratz, Morton S., 41– 42, 276 –77 Barca, Luciano, 181 bargaining power, 7, 33 – 35, 45n, 51– 54. See also power Barichello, Rick, 152n Barlow, Maude, 119n Barnard, Chester I., 74n Barre Plan, 177–78, 222, 225 Barre, Raymond, 178, 221, 239
Barro, Robert J., 217n, 264n Bates, Robert, 57n, 90n “battle of the sexes” game, 77– 80 Baumgartner, Tom, 41n, 83n Bayard, Thomas O., 38n Bayne, Nicholas, 55n Bayoumi, Tamin, 201n, 243n Bazdresch, Carlos, 148n Bendor, Jonathan, 67n Benedick, Richard Elliot, 271n, 272n, 273n, 274n, 275n Bhagwati, Jagdish, 101n, 142n Biersteker, Thomas J., 42n Blackburn, Keith, 217n Blair, Tony, 247 Blanchard, Oliver Jean, 192n, 237n Block, Fred L., 83n blocking coalition, 46, 273 Boddez, Thomas M., 120n Bosworth, Barry P., 124n, 139n, 141n, 146n Bouzas, Roberto, 129n, 142, 160n, 161n, 162 Bowker, Marjorie Montgomery, 16n Bradley, John, 206n Brandt, Willy, 175n Branson, William H., 242n Brazil, 161– 62 Bretton Woods, 34 – 35, 171, 183n, 228 Britain, and Single European Act, 52– 53, 207n. See also EMS; EMU Brock, William A., 106n Bruno, Michael, 216n Bryk, Dale S., 271n Buchanan, J. M., 46n Buckley, Walter, 41n Bueno de Mesquita, Bruce, 259n Bueno, Gerardo M., 126n, 129n Bundesbank. See EMS, and Bundesbank Burley, Anne-Marie, 87n, 96n Burns, Tom R., 41n, 83n Bush, George, 26, 122, 127, 139, 161 Caldwell, Michael, 85 Callaghan, James, 184, 185n, 187, 207, 218 Camdessus, Michel, 232n Cameron, David R., 79n, 98n, 108n, 171n, 207n, 236, 245n, 246n, 260n Cameron, Duncan, 104n, 116n, 155n Cameron, Maxwell A., 142n Campbell, Robert M., 99, 115n, 155n
308 Canada. See FTA; NAFTA Canada-U.S. Free Trade Agreement. See FTA Canzoneri, Matthew B., 178n Cárdenas, Cuauhtémoc, 126 Carey, John M., 85n Carli, Guido, 184 Carr, E. H., 15n cartels, 28 Casella, Alessandra, 264n Castañeda, Jorge G., 125n, 128n, 142n, 148n, 149n, 151n Centeno, Miguel Ángel, 128n CERES group, 236, 238 CFCs (chlorofluorocarbons), 271–76 Chapman, Anthony, 99n Chayes, Abram, 22n, 73n Chayes, Antonia Handler, 22n, 73n Chemical Weapons Convention (CWC), 269–71, 273, 275 –76 Chile, 127n, 141n, 161– 62, 165 – 66 China, 165–66, 253, 270, 273 –74 Chirac, Jacques, 231, 239 – 40, 245, 246n Choucri, Nazli, 271n Chrétien, Jean, 120, 123, 139, 154, 158 – 60 Christensen, Michael, 217n Christensen, Thomas J., 47n Clarke, Harold D., 118n, 119n Clinton, Bill, 121n, 122 club goods, 46 Coase, Ronald, 28n, 73n, 74n coercion, 3, 7, 28, 34 – 38, 138 – 39, 184, 254; credibility of, 37– 38; vs. go-it-alone power, 7, 38 –40, 42– 43, 269 –71, 273, 277. See also power Cohen, G. A., 42n collective action. See cooperation “collective-action paradigm,” 29 – 31, 153, 258, 277 Collins, Susan, 194n collusion, 28 Commonwealth of Independent States, 263 Conference on Security Cooperation in Europe, 70 constructivist theories of international relations, 258n. See also critical theory Conybeare, John A. C., 65n, 106n Cook, Maria Lorena, 124n, 125n, 126n, 151n, 152n Cooper, Richard N., 17n cooperation: and coordination, 76 –79; definition of, 17; and distributional conflict, 3, 22–27, 29–31, 33 – 38, 45n, 51– 53; dynamics of, 47–50, 255; enforcement of, 8, 17–18, 20 –23, 63 – 68, 70n, 72–78, 90;
I N D E X
and externalities, 17, 46, 51, 262; incentives for, xiii–xv, 3 –12, 16 –18, 23, 27– 32, 34 – 40, 43 – 45, 47, 53, 63 – 68, 76 – 78, 86, 90, 251– 55, 261– 62, 276 –78; and institutions, xiii, 3 –10, 15 –18, 27– 33, 61– 80, 255 – 56; recent trends in, xiii, 3, 19, 251– 53, 257; utility consequences of, xiii, 3 –10, 16 –18, 23 –24, 27– 32, 34, 36n, 38 – 39, 43 – 45, 47– 57, 61, 63, 251– 55, 257, 261, 276. See also environmental cooperation; institutions; monetary cooperation; security cooperation; trade cooperation “cooperation game,” 43 – 47; stage two of, 47–50 co-optation. See power politics of institutional design coordination dilemma, 76 – 80. See also cooperation Corden, W. Max, 141n Cornelius, Wayne A., 125n, 151n counterfactual problems, 4, 11, 50, 53, 90, 178 – 80, 186, 205 – 6 Couzen, Ken, 207 Cowhey, Peter F., 85n Crawford, Robert G., 73n Craxi, Bettino, 200 credibility, 37– 38, 40, 54n, 66n, 101, 108, 177–78, 181, 187, 190, 222–23, 237, 274 critical theory, 18n. See also constructivist theories of international relations Crone, Donald, 166n, 262n Dahl, Robert A., 40 – 42, 276 D’Alema, Massimo, 199n de Boissieu, Christian, 221n, 237n, 238n DeBusk, F. Amanda, 120n, 121n, 145n De Cecco, Marcello, 173n, 176n, 182n, 200 –201 De Grauwe, Paul, 175n, 194n, 201n de la Madrid, Miguel, 124 –28, 139, 146 de la Torre, Augusto, 95n Delors, Jacques, 232n, 235 Delors Plan, 232n de Melo, Jaime, 166n Dennis, Geoffrey, 186n, 207n Destler, I. M., 86n, 101n Diebold, William, Jr., 126n divergence indicator, 24n, 228 Dixit, Avinash K., 72n Doern, G. Bruce, 99, 154n, 158n dominant strategy, definition of, 39 Dornbusch, Rudiger, 192n, 194n, 201n, 206n, 244n
I N D E X
Downs, George W., 67n Durch, William J., 70n economic geography, 142, 263 Edwards, Sebastian, 127n EFTA (European Free Trade Association), 164 –65 Eichengreen, Barry, 173n, 175n, 192n, 201n, 203n, 204n, 222n, 242n, 243n, 258n Elliott, Kimberly Ann, 38n Elster, Jon, 42n empires, voluntary, 163 EMS (European Monetary System), xiv–xv, 3, 11–12, 25, 42n, 61, 74, 90, 171, 252– 54, 274 –76, 278; Britain’s motivations for joining, 171, 178, 181, 186 – 91; and Bundesbank, 108n, 171, 175 –77, 201, 208, 214, 217, 221–22, 224, 226 –28, 242– 44, 245n; and coercion, 184, 232– 33; collapse of, 172–73, 242– 44; and cooptation of French Socialists, 114n, 213 – 14, 222– 41; costs of exclusion from, 182– 91; disciplinary rationale for, 178, 181, 186n, 192– 94, 237– 38; divergence indicator of, 24n, 228; efficiency account of, 215 –22, 240; exchange rate mechanism (ERM) of, 172; and Franco-German enacting coalition, 173 –78, 184, 213 –14, 220 –26, 228 – 30, 241; and governance role of European Monetary Committee, 201, 214 –15, 219 –20, 230, 234; impact on state behavior of, 91, 261; incomplete contracting account of, 215 –22, 240; Italy’s motivations for joining, 171, 178, 180 – 91, 200 – 4, 212, 233; negotiation and ratification of, 171–72, 183, 196 – 97, 220, 229, 240 – 41; as optimum currency area, 201– 4, 208 – 9, 211, 243n; as political scapegoat, 54 – 56, 191– 92; sidepayments in, 196 – 97; vs. Snake regime, 182, 215n, 225 –26, 230, 232– 34, 236; terms and institutions of, 172–73, 184, 186, 213 –16, 218 –20, 222–24, 227–28; and utility consequences for Britain, 171, 178 – 80, 191, 206 –12; and utility consequences for French Socialists, 234, 236 – 41; and utility consequences for Italy, 171, 178 – 80, 191–206, 211–12. See also EMU EMU (European Monetary Union), 3, 95, 173; and Britain, 247– 48; and France, 245–47; and Germany, 245 – 47, 260n; and Italy, 245n, 247– 48; negotiation and ratification of, 214, 245 – 48; terms and in-
309 stitutions of, 214 –15, 244 – 47. See also EMS enacting coalition, incentives of, 6 –10, 81– 90, 255 – 56, 260 – 63, 271 Enterprise for the Americas Initiative, 161 environmental cooperation, xiii, 73n, 142, 271–76 Esquivel, Gerardo, 139n euro, 95, 173, 214 –15, 245 – 48. See also EMU European Central Bank (ECB), 214 –15, 245. See also EMU European Community (EC). See European Union European Regional Development Fund (ERDF), 196, 203 – 4 European Union (EU), 19, 33n, 66n, 72, 74, 86n, 95 – 96, 171n, 271n; enlargement of, 163 – 65; and GATT, 25, 166 – 67; and Single European Act, 52– 53, 79, 164 – 65, 184n; and Structural Funds, 148. See also EMS; EMU Evans, Peter B., 54n exchange rate mechanism (ERM), 172. See also EMS externalities, 17, 46, 51, 262; security, 24 – 25, 260. See also cooperation, go-it-alone power Fabius, Laurent, 232n Fearon, James D., 3n, 7, 33n, 50n, 54n, 77n Federal constitution, U.S. ratification of, 264 – 69, 275 –76 Feinberg, Richard E., 161n Feldstein, Martin, 192n, 245n Fieleke, Norman, 95n Finger, J. Michael, 75 Finland, accession to EU of, 163 – 65 Finnemore, Martha, 18n fiscal federalism, 203 – 4, 247 Flanagan, Robert J., 194n, 195n focal points, 6, 78 –79 foreign aid, xiii, 36 France. See EMS; EMU Frankel, Jeffrey A., 91n, 129n, 140n, 162n, 163n, 175n Fratianni, Michele, 173n, 174, 178n, 194n, 219n, 222n, 234n free trade areas, vs. custom unions, 95. See also APEC; EFTA; FTA; FTAA; NAFTA Frieden, Jeffry A., 8n, 37n, 42n, 56n, 73n, 173n, 180, 182n, 194n, 215n, 233n, 258n, 259n Friedman, David, 264n
310 Friman, H. Richard, 33n FTA (Canada-U.S. Free Trade Agreement): and Canadian national election of 1988, 154–56; and Canadian national election of 1993, 117–19; and domestic politics in Canada, 97, 102– 5, 115 –17, 123, 154 – 60; and domestic politics in U.S., 99 –102, 106, 108–9; incomplete contracting account of, 106–12; negotiation and ratification of, 97–98, 108 – 9; and power politics of institutional design, 96 – 97, 112– 14, 119–21; previous literature on, 98 – 99; terms and institutions of, 105 – 8, 109–14, 120–21. See also NAFTA FTAA (Free Trade Agreement of the Americas), 161–63, 252 Fudenberg, Drew, 78n Funabashi, Yoichi, 166n Gaines, Brian J., 103n Gandhi, Prem, 100n Garrett, Geoffrey, 3n, 6n, 8n, 33n, 56n, 72n, 77n, 79n, 87n, 96n, 98n, 189n, 193n, 207n, 234n, 246n, 252n, 260n Gaston, Noel, 119n, 156n GATT (General Agreement on Tariffs and Trade), 5, 19, 25, 35, 69 –70, 75, 86, 114n, 124n, 161n, 163, 166 – 67. See also WTO Gaventa, John, 41n Germany, 88n, 163; reunification of, 148 – 49, 242–43, 246n. See also EMS; EMU Giavazzi, Francesco, 173n, 174n, 178n, 181n, 192n, 200n, 204, 226n, 228n, 237n, 242n Gilpin, Robert, 20n, 28n, 35n, 37, 38n, 74n Giovannini, Alberto, 173n, 174n, 178n, 204, 226n, 228n, 242n Giscard d’Estaing, Valéry, 173 –75, 177–78, 180–81, 183, 196, 213 –14, 216 –18, 220–31, 233, 237, 240 – 41 Glaser, Charles L., 25n, 26n, 259n globalization, 18, 251, 253 Globerman, Steven, 141n go-it-alone power, xiv, 7–12, 38 – 52; and agenda-setting power, 12n, 40 – 43, 276 – 77; and bargaining power, 7, 33 – 35, 45n, 51–54; and coercion, 7, 38 – 40, 42– 43, 269–71, 273, 277; conditionality of, 11, 40n, 45–47, 262– 63, 269, 276 –77; dynamics of, 47–50, 255; and institutional design, 8–11, 62, 81– 90, 255 – 56, 260 – 63, 271; and issue linkage, 8, 36 – 37, 66n, 87, 184; and Marxist theories of exploita-
I N D E X
tion, 42– 43; as three-player game, 43 – 47. See also bandwagoning Goldstein, Judith, 16n, 85n, 99, 101n, 120n, 121n Goodman, John B., 173n, 175n, 203n Gordon, David, 217n governance structures, xiii. See also institutions Gowa, Joanne, 23n, 25n, 65n, 260n Grieco, Joseph M., 3, 4, 8n, 16, 19n, 21n, 22–28, 76 –77, 197n, 246n, 260n, 262, 277. See also relative gains Grilli, Vittorio, 194n Grinspun, Ricardo, 142n Gros, Daniel, 176n, 186n Grossman, Sanford J., 72n Group of Seven (G-7), 55, 217n Grunwald, Joseph, 149n Guerrieri, Paolo, 173n Haas, Ernst, 18n, 86n Haas, Peter M., 251n, 271n, 272n Haggard, Stephan, 36n, 69n, 95n, 100n, 138, 139n, 142n, 165n, 262n Hall, Peter A., 232n, 235n Hamilton, Carl, 164n Hamilton, Colleen, 69n, 70n Hankel, Wilhelm, 177n Hardin, Russell, 17n Harmon, Mark D., 186n, 187n, 190n Harsanyi, John C., 37 Hart, Michael, 99, 102n Hart, Oliver D., 72n hegemonic stability theory, 28n, 36n, 37 Heisenberg, Dorothee, 173n, 177n, 226n Helleiner, Gerald K., 141n Hellman, Judith Adler, 151n Henderson, Michael D., 116n Henning, C. Randall, 217n, 241n, 242n, 243n, 246n Heraclides, Alexis, 70n Heredia, Blanca, 148n Heredia, Carlos, 151n Hervouet, Gérard, 117n Hibbs, Douglas, 56n, 224 Hirschman, Albert, 9n, 33, 45n, 73n, 109n Hoffmann, Stanley, 18n holdup problem, 46, 73n, 273 Hopkins, Raymond F., 28n Horlick, Gary N., 106n, 120n, 121n, 145n Howe, Geoffrey, 210 –11 Hufbauer, Gary Clyde, 20n, 38n, 96n, 141n, 142n, 146n, 147n, 149n, 162n human rights, xiii, 21, 70
311
I N D E X
Huntington, David S., 109n, 145n Hurrell, Andrew, 139n Hurtig, Mel, 99n, 119n
Johnston, Richard, 104n, 155n, 156n Joskow, Paul L., 73n Josling, Tim, 125n, 152n
Ikenberry, G. John, 56n, 259n, 271n institutions: early literature on, 15; as equilibrium-selectors, 6, 78 – 80, 255; and focal points, 6, 78 –79; incomplete contracting rationale for, 71–75, 80, 255; and information, 65 – 69, 74 –75; insulation and longevity of, 84 – 90, 262; as political scapegoats, 54 – 56; preferencetransforming impact of, 6, 16, 87n, 271, 274–76; and prospect of mutual gain, xiii, 3–10, 16–18, 23 –24, 27– 34, 38, 47– 57, 61, 63, 68, 78–79, 251– 55, 257, 261, 276; rapid expansion of, xiv, 47– 50; relevance of, 15, 19, 21–22, 27, 33 – 34; as supranational governance structures, xiii– xiv, 9–10, 61–63, 68 – 80, 251; supranational vs. intergovernmental, xiv, 79, 262; variation in rules and structures of, 8–11, 29–31, 61–62, 70, 95, 262– 63, 278; as vehicles for international cooperation, 5– 6, 17–19, 27–28, 33, 61– 80; voluntary vs. imposed, xiii, 7, 28, 38, 253 – 54. See also cooperation; nation building integration, political. See cooperation; institutions; nation building interdependence, 17. See also cooperation; externalities International Court of Justice (ICJ), 21 International Energy Agency (IEA), 21 International Labor Organization (ILO), 21 International Monetary Fund (IMF), 35, 55, 228n Ireland, 172, 196–97, 202, 206n, 233 Ishiyama, Yoshihide, 201n issue linkage, 8, 36–37, 66n, 87, 184 Italy. See EMS; EMU
Kaempfer, William H., 38n Kahler, Miles, 5n, 62n, 69, 130n, 139n, 145n, 167n Kaltenthaler, Karl, 226n, 246n Kaminski, John P., 268n Katzenstein, Peter J., 16n, 271n Kaufman, Robert R., 148n Kees, Andreas, 228n Kehoe, Timothy J., 141n Kelemen, R. Daniel, 87n Kelly, Margaret R., 95n Keohane, Robert O., 3 – 6, 11–12, 15 –21, 23, 28 –29, 63, 66n, 68, 71, 72, 85n, 87n, 88n, 106n, 113n, 187n, 251n, 261, 277. See also neoliberalism Kesselman, Mark, 238n Kiewiet, D. Roderick, 56n Kindleberger, Charles P., 74n King, Gary, 187n Kitschelt, Herbert P., 83n Klein, Benjamin, 73n Klesner, Joseph L., 126n Knight, Jack, 3n Koechlin, Timothy, 142n Koeune, Jean-Claude, 175n Kohl, Helmut, 79, 241, 242, 246 Koremenos, Barbara, 29n, 62n Kornberg, Allan, 118n, 119n Kramer, Gerald H., 56n Krasner, Stephen D., 3 – 4, 7, 16n, 21n, 28n, 29n, 29, 31, 33n, 34, 35n, 36n, 37, 45n, 74n, 76 –77, 78n, 83n, 85n, 145n, 148n, 277 Kratochwil, Friedrich, 18n, 258n Kreps, David, 66n, 72n Krueger, Anne O., 143n, 166n Krugman, Paul, 54 – 55, 139n, 140n, 141n, 142n, 150n, 164n, 192n, 201n, 263n Kruse, Donald C., 173n, 183n Kupchan, Charles, 16n, 162n, 271n Kupchan, Clifford A., 16n Kydland, Finn E., 217n
Jachtenfuchs, Markus, 272n Jackson, John H., 19n, 62n, 69n, 70n, 114n Jacobson, Gary C., 259n Jacobson, Harold K., 54n, 91n Jamaica, 160–61 James, Scott C., 43 Japan, 140–41, 157n, 163, 177n; and relations with U.S., 26–27; and United Nations, 88n Jenkins, Roy, 172–73, 175, 177, 207n Jervis, Robert, 25n, 260n Joffe, Josef, 246n, 248n Johnson, Jon R., 100n, 116n
Lake, David A., 36n, 37n, 43, 56n, 259n Lamont, Norman, 190, 244 Lange, Peter, 56n, 164n, 193n, 234n Larudee, Mehrene, 142n Lauber, Volkmar, 178n Lawrence, Robert Z., 61, 91n, 124n, 139n, 141n, 142n, 144, 146n, 166
312 Lawson, Nigel, 187, 189 – 90, 207, 209 –11 Lebow, Richard Ned, 63n LeDuc, Lawrence, 104n, 154n, 155n Le Pen, Jean-Marie, 239, 246 Levitt, Kari, 116n Levy, Jack S., 26n Levy, Marc A., 251n Linz, Juan, 57n Lipsey, Richard G., 109n, 116n, 146n, 148n, 157n, 159n Lipson, Charles, 16n, 25, 29n, 62n, 65n, 67n, 72n Lohmann, Susanne, 219n Lomax, D. F., 186n Londregan, John, 259n Loriaux, Michael, 174n, 175n Lowenberg, Anton D., 38n Lowenfeld, Andreas F., 69n Luce, R. D., 38n Ludlow, Peter, 173n, 180, 181n, 182n, 183n, 185n, 194n, 196n, 197n, 225n, 226n, 229n, 257 Luif, Paul, 164n Lukes, Steven, 40n, 41n, 276n Lundberg, Lars, 143n Lustig, Nora, 124n, 128n, 139n, 141n, 146n Lutz, Mark, 140n Lynch, Jack, 197 Maastricht treaty, 214, 246n, 247– 48. See also EMU Macdonald Commission, 101– 3 Mace, Gordon, 117n Mackenzie, Hugh, 116n Magee, Stephen P., 106n Major, John, 190, 210n, 211–12, 248 Mansfield, Edward D., 25n, 260n Manzetti, Luigi, 162n maquiladora program, 143n, 158 March, James G., 63n market failures, 28, 63. See also cooperation; externalities Marks, Gary, 204n Marsh, David, 175n, 226n Marshall, William, 71n, 72n Marston, Richard C., 186n Martin, Lisa L., 7, 8n, 15n, 16n, 28n, 29n, 34, 37, 47n, 54n, 56n, 77n, 78n, 87n Maskin, Eric, 78n Mastanduno, Michael, 26 –27, 56n, 259n Mattli, Walter, 87n, 96n Mauroy, Pierre, 235 Mayhew, David R., 259n McCalla, Robert, 263n
I N D E X
McCallum, John, 264n McCarthy, Patrick, 231n, 239n McDonald, Frank, 173n McKinnon, Ronald, 201n McNamara, Kathleen R., 173n, 174n, 215n, 234n McQuaid, Linda, 103n Mearsheimer, John J., 15n, 16n, 19n, 21n, 23n, 25n, 26n, 68n, 251n Melitz, Jacques, 216n Mercosur (Common Market of the South), 122n, 161– 63 Mexico. See NAFTA Michalski, Anna, 164n Middlebrook, Kevin J., 124n, 125n, 126n, 151n, 152n Milgrom, Paul, 68n, 72n Miller, Marcus, 186n Milner, Helen, 16n, 17, 22n, 28n, 33n, 54n, 56n, 108n, 246n, 259n Mitchell, Ronald B., 271n Mitrany, David, 86n Mitterrand, François, 184n, 213, 224–25, 226n, 231– 41, 246n Moe, Terry M., 62, 71n, 72n, 82n, 85 Molinar Horcasitas, Juan, 124n, 125n, 126n, 151n, 152n monetary cooperation, xiii, 91, 95, 252, 261– 62. See also EMS; EMU Monnet, Jean, 86n Montreal Protocol, 271–76 Moravcsik, Andrew, 6n, 16n, 33n, 52– 53, 56n, 72n, 86n, 98n, 171n, 175n, 177n, 186n, 207n, 246n, 259n Morel, Benoit, 269n Morgenthau, Hans J., 15n, 20n Morici, Peter, 99 Morris, Stephen D., 126n, 148n Morrisett, Peter M., 271n Morrow, James D., 3n, 33n, 77n Moyer, Homer E., Jr., 113n, 121n Muet, Pierre Alain, 237n Mulroney, Brian, 98, 103 – 4, 107– 8, 112– 13, 115, 119, 121, 127, 154 – 56 multilateralism. See cooperation; institutions multiple equilibria problem, 29 – 31, 76 –79, 255 Mundell, Robert, 201n Murphy, Peter, 98, 109 Nabors, Robert, 40n NAFTA (North American Free Trade Agreement), xiii–xv, 3 – 4, 11–12, 61, 90, 253 – 54, 274 –76, 278; APEC as response to,
I N D E X
165 – 66; Canada’s motivations for joining, 139, 154 – 60; and coercion, 130 – 31, 138 – 39, 159; and Chiapas uprising, 151– 53; compared to GATT, 124n, 143, 145; as empirical puzzle, 95 – 97, 123 –27, 153 – 54; EU response to, 163 – 65; FTAA as response to, 160 – 63; and hub-andspokes issue, 123, 159 – 60; impact of FTA on Mexican support for, 97, 123, 127n, 129 – 38; impact on state behavior of, 91, 261; Mexico’s motivations for proposing, 123 –24, 127– 39; negotiation and ratification of, 122; previous literature on, 122–23, 124n, 129, 257; as stimulus to GATT, 161n, 166 – 67; terms and institutions of, 95 – 96, 142, 145 – 49, 157– 59; U.S. interests in, 96, 139, 145, 161n; and utility consequences for Canada, 96, 123, 154 – 60; and utility consequences for Mexico, 96 – 97, 123, 140 – 53. See also FTA National Solidarity Program (PRONASOL), 148 nation building, 263– 64, 275 –76; and ratification of U.S. Federal constitution, 264 – 69, 275–76 NATO (North Atlantic Treaty Organization), 71n, 252, 263n natural experiment, 187 Nellis, Joseph, 186n, 207n neofunctionalism, 18n, 86n neoliberal institutionalism. See neoliberalism neoliberalism, 3, 15–20, 27– 31, 33, 52– 53, 63–69, 71–73, 76, 81, 258, 277. See also new institutionalism new economics of organization, 71–75, 81, 255. See also new institutionalism new institutionalism, 15, 63, 215. See also new economics of organization nontariff barriers (NTBs), 25, 69, 75, 138, 146, 152 North, Douglass C., 68n, 72n Norway, 164n, 165 Nye, Joseph S., Jr., 18n, 86n Oatley, Thomas H., 40n, 171n, 173n, 207n, 215n Obstfelt, Maurice, 140n Odell, John S., 86n O’Halloran, Sharyn, 100n, 109n, 139n Oliver, Geoffrey D., 106n Olsen, Johan P., 63n Olson, Kyle, 269n Olson, Mancur, 4, 17n, 57n, 277
313 Organization for Security and Cooperation in Europe (OSCE), 70 Organization of Petroleum Exporting Countries (OPEC), 28, 177, 200, 207 Osherenko, Gail, 91n, 271n Ostrom, Elinor, 271n Oudiz, Gilles, 178n Oye, Kenneth A., 3n, 7, 16n, 33n, 46, 67n, 86n Padoan, Pier Carlo, 173n Padoa-Schioppa, Tommaso, 220n, 222n, 226n Pagano, Marco, 178n, 181n, 192n, 237n Pal, Leslie A., 99, 115n, 155n PAN (Partido de Acción Nacional), 151 Panagariya, Arvind, 142n, 166n Pandolfi Plan, 181 Pape, Robert A., 38n Pareto efficiency, xiii, 23, 34, 44, 63. See also cooperation, utility consequences of Parisi, Arturo, 199n Parson, Edward A., 271n, 275n Pasquino, Gianfranco, 199n Pastor, Manuel, 125n, 127n, 128, 144n, 150n, 151n Pastor, Robert A., 100n path dependence, 47– 50 Patrick, Hugh T., 101n PCI (Partito communista italiano), 181– 83, 193 – 96, 197–200 Pelzman, Joseph, 38n Peres Núñez, Wilson, 128n, 147n, 151n Perez Lopez, Jorge, 38n Perrow, Charles, 42n Pescatore, Pierre, 70n Pierson, Paul, 207n Pisani-Ferry, Jean, 221n, 237n, 238n pluralist theory: concept of power in, 40 – 42, 276 –77; omission of institutions from, 15 Poitras, Guy, 128n, 143n, 151n political geography, 263 – 64. See also institutions; nation building Pollack, Mark A., 79n Poole, Keith, 259n Popkin, Samuel L., 55n, 148n Posen, Barry R., 260n Posner, Alan R., 194n Powell, Robert, 22n, 24n power, 3, 7, 10 –12, 29 – 52. See also agendasetting power; bargaining power; coercion; go-it-alone power power politics of institutional design, 8 –11,
314 power politics of institutional design (cont.) 62, 81–90, 255 – 56, 260 – 63, 271. See also institutions PRD (Partido de la Revolución Democrática), 151 preferences, instability of, 18n, 41n, 82– 84, 89, 260–63. See also power politics of institutional design preferential trading arrangements. See APEC; EFTA; FTA; FTAA; NAFTA Prescott, Edward C., 217n PRI (Partido Revolucionario Institucional), 123–26, 128, 144, 151, 153. See also NAFTA prisoners’ dilemma, 64–65, 76, 78n, 217, 254 Prodi, Romano, 248 Przeworski, Adam, 56n, 89n public goods, 46. See also prisoners’ dilemma Puchala, Donald J., 28n Putnam, Robert D., 33n, 54n, 55n, 108n, 217n, 226n Raiffa, Howard, 34n, 39n Ramirez de la O, Rogelio, 125n Randall, Stephen J., 142n Rapoport, Anatol, 37 Reagan, Ronald, 99, 107– 8, 113, 121, 139 realism, 3, 15 –16; anarchy assumption of, 20–22, 68, 71n, 277; critique of neoliberal theory, 18–27; and domestic politics, 6n, 10, 259–61; and institutions, 15 –16, 19, 33, 63n; and neoliberal consensus, 27– 29, 258. See also cooperation, enforcement of; relative gains reciprocity, 17. See also cooperation Reding, Andrew, 126n regimes, international. See institutions regime theory, 15–16. See also neoliberalism Reisman, Simon, 98, 109 relative gains, 19, 22–27, 76 –77, 262; vs. absolute gains, 4, 7, 23 –24; and East-West conflict during Cold War, 25. See also realism Riker, William H., 89n, 266 – 69 Risse-Kappen, Thomas, 6n Rivers, Douglas, 56n Roan, Sharon L., 272n Roberts, John, 72n Robinson, Ian, 142n Robinson, Raymond, 128n, 143n, 151n Rocard, Michel, 238 Rocke, David M., 67n Rockman, Bert A., 85n
I N D E X
Rodrik, Dani, 252n Rogowski, Ronald, 56n, 259n Ros, Jaime, 128n, 129n, 141n, 142, 144n, 147n, 151n, 160n, 161n, 162 Rosenthal, Howard, 56n RPR (Rassemblement pour la République), 178, 224, 239 – 40 Rubinstein, Ariel, 34n Ruggie, John G., 18n, 19n, 258n, 271n Ruggiero, Renato, 183 Rugman, Alan M., 100n Russell, Peter H., 103n Saborio, Sylvia, 161n Sachs, Jeffrey, 216n, 232n, 233n, 236n Salinas de Gortari, Carlos, 122, 126 –29, 138 – 39, 152– 53, 158 sanctions, 34, 36n, 37, 38n. See also coercion Sandel, Michael J., 264n Sandholtz, Wayne, 6n, 66n, 98n, 245n, 271n Sapir, André, 143n Sargent, Thomas J., 201n Sassoon, Donald, 196n Saxonhouse, Gary R., 166n Scharpf, Fritz W., 177n Schattschneider, E. E. , 12n, 41n, 42, 100n, 276 Schelling, Thomas, 37, 54n, 78 Schmidt, Helmut, 173 –75, 180 – 83, 213 – 14, 216 –18, 220 – 30, 240 – 41 Schmitter, Philippe, 86n Schott, Jeffrey J., 20n, 38n, 96n, 98n, 99, 116n, 126, 141n, 142n, 146n, 147n, 149n, 162n Schulz, Heiner, 87n Schwanen, Daniel, 109n, 116n, 146n, 148n, 156n, 157n, 159n Scott, James C., 148n Sebenius, James K., 8n, 271n security cooperation, xiii, 70, 262– 63, 269 – 71, 275 –76 Serra Puche, Jaime, 128 –29, 131 Shepsle, Kenneth A., 41n, 71n, 276n Shugart, Matthew, 85n Shulman, Jennifer B., 91n side-payments, 8, 24, 36n, 44, 252n. See also issue linkage Silva-Herzog, Jesus, 129 Simmons, Beth A., 15n, 29n, 36n, 56n, 259n Simon, Herbert A., 74n Simpson, Jeffrey, 115n Siverson, Randolph M., 259n Slaughter, Anne-Marie, 87n
I N D E X
Smith, Adam, 143n Smith, David, 186n, 190n, 207n Smith, Jay, 146n Smith, Murray G., 99 Snake regime, 182, 213, 215n, 221, 225 –26, 230, 232–34, 236, 241 Snidal, Duncan, 16n, 21n, 22n, 24n, 28n, 29n, 37n, 62n, 75n, 76n, 83n Snyder, Glenn H., 47n Snyder, Jack, 47n, 260n Soskice, David W., 194n, 195n sovereignty, 21, 61–63, 68 – 80, 95 – 97, 214–15, 219–20. See also institutions Spaventa, Luigi, 173n, 182n, 196n, 200n spillover effects, 46, 262. See also externalities Spolaore, Enrico, 264n Sprague, John, 56n Sprinz, Detlef, 275n Stålvant, Carl-Einar, 164n Steger, Debra P., 98n, 106n Stein, Arthur, 23n, 76n, 85n, 113n Stephens, Philip, 190n Sterdyniak, Henri, 178n Stokes, Susan, 83n Strange, Susan, 84, 95n Streeck, Wolfgang, 86n supranational institutions. See institutions Sutherland, Alan, 186n Sutton, John, 34n, 44 Sweden, 202; accession to EU of, 163 – 65 Tabellini, Guido, 203n Taylor, Mark P., 173n Taylor, Michael, 65n, 67n Taylor, Paul, 18n, 33n, 62n, 197n Thatcher, Margaret, 52, 79, 186 – 91 threat point, 34 Thygesen, Niels, 176n, 186n, 215, 226n, 228n Tollison, Robert D., 66n Tomlin, Brian W., 99, 154n, 158n Tornell, Aaron, 139n trade cooperation, xiii, 34 – 36, 65, 74 –75, 87, 91, 252, 261–62. See also APEC; FTA; FTAA; GATT; NAFTA; WTO trade creation vs. trade diversion, 140 – 41 transaction costs, 28, 87. See also cooperation Trebilcock, Michael J., 112n, 120n Trefler, Daniel, 119n, 156n Triffin, Robert, 177n Truman, David B., 15 Tsebelis, George, 85n
315 Tsoukalis, Loukas, 186n, 220n, 228n, 240n Tufte, Edward R., 56n Turner, John, 98, 103, 120n, 153 – 55 two-level games, 54n, 108n Tyers, Rod, 164n UDF (Union pour la Démocratie Française), 178, 224, 239 – 40 Ulman, Lloyd, 194n, 195n Ungerer, Horst, 194n, 234n unilateral power. See go-it-alone power United Nations, 15, 19, 21, 70n, 71n, 83, 88 Vaahtoranta, Tapani, 295n Van Evera, Stephen W., 26n, 260n van Ypersele, Jacques, 175n Verba, Sidney, 187n Verdier, Daniel, 101n Viner, Jacob, 140 Vogel, David, 43n von Hagen, Jürgen, 173n, 174, 178n, 194n, 219n, 222n, 234n, 258n Wagner, R. Harrison, 33n Waigel, Theo, 245 Walker, Michael, 141n Wallace, Helen, 164n Wallace, Neil, 201n Walt, Stephen M., 25n, 47n Waltz, Kenneth N., 10n, 15n, 16, 19, 23n, 25n, 47n, 73n, 259n. See also realism Watkins, G. C., 148n Weaver, R. Kent, 85n Weber, Axel A., 173n, 194n Wei, Shang-Jin, 175n Weiler, J. H. H., 96n Weingast, Barry R., 6n, 41n, 68n, 71n, 72n, 77n, 276n Weintraub, Sidney, 96n, 124n, 125, 126n, 129n, 130n, 141n, 146n, 257 Wendt, Alexander E., 16n, 18n, 258n, 271n Wertheimer, Alan, 37n Whalley, John, 3n, 69n, 70n, 147n, 157n, 161n Whelan, Karl, 206n White, Randall, 115n Wiarda, Howard J., 139n Wijkman, Per Magnus, 164n Willett, Thomas D., 66n Williamson, Oliver E., xiii, 28n, 72n, 73n, 85 Wilson, Robert, 66n Winham, Gilbert R., 96n, 145n
316 Wise, Carol, 125n, 127n, 128, 144n, 150n, 151n Wolf, Martin, 166n Woller, Gary, 259n Wonnacott, Paul, 140n Wonnacott, Ronald J., 100n, 109n, 116n, 117n, 141n, 146n, 148n, 157n, 158 – 59 Wood, Gordon S., 265n, 267n Woolley, John T., 173n, 234n, 260n World Bank, 35, 83 Wright, Robert, 270n WTO (World Trade Organization), 3, 5, 61, 69–70, 74 –75, 95, 253. See also GATT Wyplosz, Charles, 173n, 192n, 232n, 233n, 236n
I N D E X
Yarbrough, Beth V., 16n, 62n, 72n, 73n, 74n, 75n, 87n, 99n, 127n, 145n Yarbrough, Robert M., 16n, 62n, 72n, 73n, 74n, 75n, 87n, 99n, 127n, 145n York, Robert C., 112n, 116n Young, Alwyn, 165n Young, Leslie, 106n Young, Oran R., 7, 23n, 28, 78n, 91n, 271n Zapatista Army of National Liberation (EZLN), 151– 52 Zedillo, Ernesto, 153 Zis, George, 173n Zysman, John, 66n, 98n