The Origins of Alliances

Preface to the Paperback Edition

When I completed this book in 1986, the dramatic changes now sweeping the Soviet Union and Eastern Europe were largely unforeseen. Although I recognized that shifts in Soviet policy could weaken NATO’s cohesion significantly, I did not explore this possibility in any depth. The publication of this paperback edition provides an opportunity to do so now.

States form alliances primarily to balance against threats. Threats, in turn, are a function of power, geographic proximity, offensive capabilities, and perceived intentions. Throughout the Cold War, the Soviet Union posed a greater threat to the major powers of Eurasia than the United States did. As balance-of-threat theory predicts, these states balanced by allying with the United States, creating a global coalition that was both remarkably stable and significantly stronger than the Soviet alliance network.

The events of the past three years largely confirm balance-of-threat theory. Domestic reforms in the Soviet Union and the subsequent decision to permit independent regimes in Eastern Europe have sharply reduced the tendency for other states to balance against the Soviet Union. To note a few examples: the NATO countries are beginning major reductions in defense spending; the United States is planning to withdraw a substantial portion of its armed forces from Europe; movement for European unity has slowed; and U.S. reactions to the repression of pro-democracy forces in China suggest that China’s value as a counterweight to Soviet power has declined. As a more ominous instance, perceptions of U.S.-Japanese rivalry are growing, now that the Soviet threat no longer provides a powerful motive for cooperation. Significantly, these changes have occurred despite continued evidence of Soviet military might. Instead, the belief that Soviet aims are less dangerous and that Soviet power will be removed from Central Europe lies at the heart of these developments. Threats, not power alone, are crucial.

What does balance-of-threat theory reveal about international politics after the Cold War? By definition, the end of the Cold War implies a diffusion of threats. If current trends continue, we will see a reunified Germany bordering the newly independent states of Eastern Europe, a weakened and internally divided Soviet Union, and a stronger and more assertive Japan (China’s future course remains the most difficult to predict). It will be far less clear which states pose the most serious threats; as a result, international alignments will be more ambiguous and less durable. Perceptions of intent will be increasingly important, because the distribution of capabilities will be more equal and geography may not offer clear guidance. Thus the Eastern European states may lean toward the West should Soviet intentions appear more threatening, or tilt back toward Moscow if a reunified Germany poses the greater danger. Similarly, relations among Japan, China, and the Soviet Union are likely to be shaped less by power or geography than by each state’s assessment of the others’ intentions.

As for NATO itself, the optimistic rhetoric about maintaining the Atlantic Community should be viewed with some skepticism. Without a clear and present threat, neither European politicians nor U.S. taxpayers are likely to support a large U.S. military presence in Europe. Although NATO’s elaborate institutional structure will slow the pace of devolution, only a resurgence of the Soviet threat is likely to preserve NATO in anything like its present form. And shorn of U.S. protection and the unifying effects of the Soviet threat, the nations of Europe are likely to find cooperation more difficult to sustain, unless economic pressure from the United States and the Pacific Rim overrides familiar security concerns.

What do these trends mean for the United States? On the one hand, the United States’ external position appears even more favorable now than it did when this book was first published. (Its internal problems may be quite another matter, of course.) The traditional goal of U.S. grand strategy has been to prevent any single power from controlling the combined industrial resources of Eurasia; the decline of Soviet power means that goal will be even easier to achieve. But the end of the Cold War creates new problems as well. Because U.S. protection will be less important to its allies, U.S. influence over these states is virtually certain to decline. At the same time, the United States must avoid a return to isolationism as well as a renewed crusade to export democracy to the developing world via military intervention. History suggests that a major war is more likely when the United States withdraws from world affairs. History also suggests that U.S. military intervention in the Third World is more likely to lead to despotism and prolonged civil war than to stable and workable democracy. Paradoxically, therefore, U.S. leaders must show greater imagination, wisdom, and restraint than they did at the height of U.S.-Soviet rivalry, because the problems to be faced will be more ambiguous and the resources available to address them will be fewer.

In short, balance-of-threat theory reveals that the decline in the Soviet threat is creating a world in which the United States will enjoy greater security at less cost and risk. But it is also a world in which policy choices will be more difficult and U.S. influence will be smaller. And it is a world in which perceptions of intent will exert a powerful impact on how other states respond to U.S. actions. Although the United States retains enormous advantages when compared with other nations, it must still strike a balance between encouraging aggression through apparent indifference and provoking opposition through misplaced or unthinking belligerence.

STEPHEN M. WALT

Chicago, Illinois

June 1990

Preface

The ability to attract allies is a valuable asset in any competitive system. By contrast, those who cause others to align against them are at a significant disadvantage. Understanding the forces that shape international alliances should therefore be a vital concern for most statesmen. By elaborating and testing a number of hypotheses about alliance formation, this book attempts to contribute to that understanding.

Like most scholarly endeavors, this project began with several puzzles. While in graduate school, I was struck by the sharp discrepancy between what scholars wrote about alliances and what contemporary policy-makers apparently believed. Indeed, the two groups seemed almost 180 degrees apart. For example, the notion that alliances are formed to oppose strong states was then a truism among scholars. Yet both Soviet and U.S. statesmen repeatedly argued that the reverse was true, that any appearance of weakness would cause their allies to defect. Similarly, although most scholars viewed ideology as a weak cause of alignment, the Soviet fear of capitalist encirclement, Soviet support for countries in the socialist commonwealth, and U.S. fears about a Communist monolith suggested that the political leaders of both superpowers took ideology quite seriously. Disagreements also emerged over the importance of foreign aid and political penetration as causes of alignment. Because these competing beliefs underlie important policy disputes (including the importance of credibility, the need for overseas intervention, and the political effects of military spending and foreign aid), determining which views were most accurate seemed well worth the effort.

Along the way, as I examined the theory and practice of alliances and considered evidence from the Middle East, additional puzzles emerged. The most important involved balance of power theory. I was convinced that it was the most useful general theory available, but I was disturbed by several anomalies. According to Kenneth Waltz, whose Theory of International Politics contains the most elegant and rigorous presentation of this theory, the international system is characterized by a tendency to form balances of power. But if that is so, why did some alliances grow both larger and stronger over time (often dwarfing their opponents), as the Allies did in World War I and World War II? Furthermore, how might we explain the fact that for the past three decades the United States and its allies have controlled a combined gross national product roughly three times that of the Soviet Union and the Warsaw Pact while spending more each year on defense? Although balance of power theory can explain the Soviet response to this situation (i.e., the Soviet Union compensates for its weak allies by devoting a larger percentage of its GNP to military expenditures), it cannot tell us why the Soviet Union is so unpopular in the first place. Furthermore, if balance of power theory is a theory primarily about the behavior of great powers, then how do we explain the alliance preferences of lesser states? Do they balance as well? If so, against whom? Finally, could we incorporate the other factors that statesmen consider when deciding with whom to ally—factors that are not directly related to national power?

The solution presented in this book is a reformulation of balance of power theory which I call balance of threat theory. In anarchy, states form alliances to protect themselves. Their conduct is determined by the threats they perceive, and the power of others is merely one element in their calculations (albeit an important one). The power of other states can be either a liability or an asset, depending on where it is located, what it can do, and how it is used. By incorporating the other factors that create threats to national sovereignty, balance of threat theory provides a better explanation of alliance formation than does balance of power theory.

For reasons developed at length in chapter 1, most (but by no means all) of the evidence examined here is drawn from recent Middle East diplomacy. Unfortunately, there is no comprehensive and reliable diplomatic history of the Middle East. Patrick Seale’s excellent The Struggle for Syria ends in 1958, and the late Malcolm Kerr’s brilliant monograph The Arab Cold War is limited to inter-Arab politics from 1958 to 1970. The enormous literature on the Arab-Israeli conflict, by contrast, usually downplays inter-Arab relations. And the best scholarly studies—such as Nadav Safran’s Saudi Arabia: The Ceaseless Quest for Security, Steven L. Spiegel’s The Other Arab-Israeli Conflict, and Michael Brecher’s remarkable books on Israeli foreign policy (The Foreign Policy System of Israel, Decisions in Israel’s Foreign Policy, etc.)—focus primarily on one country. Having decided to test several hypotheses by examining Middle East alliances, I found that there was no reliable account of the full set of events I sought to explore.

As a result, I was forced to provide that account myself. This requirement was both a burden and an opportunity; chapters 3 and 4 provide the only complete account of postwar alliance diplomacy in the Middle East (including the diplomacy of the two superpowers) of which I am aware. Although this book is by no means the definitive diplomatic history of this region, I hope that both the theoretical and the historical material contained in it will be of use to students of international politics and the Middle East alike.

Writing this book taught me the importance of loyal allies. I was fortunate to have many. Kenneth N. Waltz encouraged me to ask an important question. Even more helpfully, he provided a model for the ways that important answers are reached. George Breslauer was unstinting in his criticism and generous with his praise. The friendship of both these scholars is much appreciated. Walter McDougall sharpened my use of diplomatic history and corrected several errors. I am also indebted to Robert Art, whose comments and encouragement made revising the manuscript much easier.

The following people offered valuable comments on all or part of the manuscript: L. Carl Brown, Lynn Eden, Charles Glaser, Lori Gronich, Fen Hampson, John Mearsheimer, Steven E. Miller, Laurie Mylroie, Kenneth Oye, Glenn Snyder, Jack Snyder, David Spiro, Marc Trachtenberg, John Waterbury, and Lynn Whittaker. I am grateful to all of them. I also thank Jonathan Shimshoni for his friendship and his many suggestions as I began the final draft.

I have profited from interviews with several scholars who helped me find my way through the tangled thicket of Middle East politics. William B. Quandt, Dennis Ross, and Nadav Safran provided advice in the early stages. Seth Tillman and Steven Rosen offered their own insights on the domestic politics of U.S. Middle East policy. In Egypt, discussions with Gehad Auda and Abdul-Monem Sayed of the Center for Strategic Studies at al-Ahram, Abdul-Monem al-Mashat of Cairo University, Ann Mosely Lesch of the American Universities Field Staff, and Mohammed Sid Aqmed helped sharpen my grasp of inter-Arab relations. In Israel, Ya’acov Bar-Siman-Tov, Galia Golan, Ya’acov Roi, and Dan Schueftan graciously interrupted their own work in order to discuss mine. I thank all of these scholars for sharing their ideas with me.

The opportunity to give a seminar at the Jaffee Center for Strategic Studies in Tel Aviv helped me refine my ideas further. I thank its director, Aharon Yariv, for the invitation and Shai Feldman for arranging my visit and for offering his own valuable insights.

Support from institutional allies also has been indispensable. The early drafts of this book were written while I was a research fellow at the Center for Science and International Affairs at Harvard University. CSIA was an ideal intellectual home, and I am grateful to its founder, Paul Doty, and the rest of the staff for their help. Some of the research was conducted during my various sojourns at the Center for Naval Analyses, and my associates there—especially Bradford Dismukes, James M. McConnell, Charles C. Peterson, and Robert G. Weinland—contributed greatly to my intellectual growth.

I have received financial support from the Institute for the Study of World Politics, the Center for Science and International Affairs, and the Center for International Studies at Princeton University. A grant from the University Committee on Research at Princeton made it possible for me to travel to the Middle East. In the Woodrow Wilson School, Agnes Pearson handled various administrative details with exceptional competence. Edna Lloyd, Gwen Hatcher, Lynn Caruso, Terry Barczak, and Malilia Clark provided efficient secretarial support.

Portions of chapters 2 and 8 originally appeared in the spring 1985 issue of International Security (vol. 9, no. 4, copyright © 1985 by the President and Fellows of Harvard College and the Massachusetts Institute of Technology). I thank the journal and its publishers for permission to use this material here. I also thank Cornell University Press for being tolerant of a young author, and especially Jo-Anne Naples for helping tidy up my prose.

Four others merit my deepest thanks. Stephen Van Evera has contributed ideas and encouragement throughout. He is, quite simply, everything a colleague should be. Helene Blair Madonick knew how much this book meant to me; I hope she knows that she means even more. Finally, I dedicate this book to my parents; the older I get, the more grateful I am.

STEPHEN M. WALT

Princeton, New Jersey

[1]

Introduction: Exploring Alliance Formation

This book is about the origins of alliances.¹ I seek answers to questions such as these: What causes states to support one another’s foreign policy or territorial integrity? How do statesmen choose among potential threats when seeking external support? How do the great powers choose which states to protect, and how do weaker states decide whose protection to accept? In short, how do states choose their friends?

The importance of this subject is manifest.² The forces that bring states together and drive them apart will affect the security of individual states by determining both how large a threat they face and how much help they can expect. At the same time, the factors that determine how states choose alliance partners will shape the evolution of the international system as a whole. The ability to establish durable empires, for example, will depend in large part on how potential victims respond. Will they work together to thwart these ambitions, or can a potential hegemon keep its opposition isolated and weak? Does aggression become easier with each new conquest, or does resistance increase at a faster rate?

Failure to understand the origins of alliances can be fatal. In the Franco-Prussian War, for example, France entered the war confident that Austria-Hungary would soon join it in battle against Prussia. When the Austrians chose to remain neutral (a decision Bismarck’s diplomacy had encouraged), a key element of French strategy collapsed.³ In the decades before World War I, Germany’s leaders ignored the possibility of a Franco-Russian alliance (1892) and an Anglo-Russian entente (1907), only to be surprised when their own actions helped create the very alignments they had believed were impossible.⁴ In much the same way, Japan’s leaders were convinced that their alliance with Nazi Germany and Fascist Italy would deter the United States from opposing their expansion in the Far East. They could not have been more wrong; the formation of the Axis encouraged the United States to resist Japanese expansion even more vigorously and to move closer to its wartime alliance with Great Britain and the Soviet Union.⁵

In each of these cases, the error lay in a faulty understanding of the causes of alliances. As a result, these states adopted grand strategies that were seriously flawed. In the simplest terms, a state’s grand strategy is a theory explaining how it can cause security for itself.⁶ Strategy is thus a set of hypotheses or predictions: if we do A, B, and C, the desired results X, Y, and Z will follow. Ideally, a statement of grand strategy should explain why these results are likely to obtain and provide appropriate evidence. Because the challenges one may face and the capabilities one can employ will be affected by the behavior of other states (e.g., will they help, remain neutral, or oppose?), the hypotheses that statesmen accept about the origins of alliances should play a major role in determining the strategies they select. The success of these policies will depend on whether the hypotheses they embrace are correct.

THE DEBATE

The United States offers no exception to these principles. More than anything else, the Cold War between the United States and the Soviet Union has been a competition for allies. As a result, many recurring debates over the conduct of U.S. foreign policy ultimately rest on disagreements about the causes of alliances. The policies that have emerged depend on which hypotheses of alliance formation were endorsed. The question is whether or not these hypotheses are correct. Answering this question is the principal goal of this book.

One central issue is how states respond to threats. Do states seek allies in order to balance a threatening power, or are they more likely to bandwagon with the most threatening state? This basic question lies at the heart of a host of policy issues. For example, should the United States increase its military spending and its commitment to NATO to prevent the growth of Soviet military power from causing the Finlandization of Europe? Alternatively, should the United States do less so its allies will do more? Similarly, will the fall of the Shah and the Soviet invasion of Afghanistan drive the Gulf states into a pro-Soviet position, or are they more likely to join forces with the United States and one another? The answer depends on whether states most often ally to oppose a threatening power or try to appease it.

Throughout the Cold War, U.S. statesmen have consistently embraced the latter view. As the Basic National Security Policy formulated by the National Security Council in 1953 stated: If our allies were uncertain about our ability or will to counter Soviet aggression, they would be strongly tempted to adopt a neutralist position.⁷ Pactomania was the logical result. In the same way, the lengthy involvement of the United States in Vietnam was justified by the widespread fear that U.S. allies would defect if the United States withdrew. As Secretary of State Dean Rusk put it: America’s alliances are at the heart of the maintenance of peace, and if it should be discovered that the pledge of America was worthless, the structure of peace would crumble and we would be well on our way to a terrible catastrophe.⁸ Because U.S. statesmen have believed that allies are attracted by displays of strength and will, they have sought to preserve an image of credibility and military superiority despite the obvious costs.⁹

A second issue is whether states with similar internal characteristics are more likely to ally than states whose domestic orders are different. The early debates over the implementation of containment, for example, were due in part to disagreements on this point. Where George F. Kennan saw the Communist bloc as prone to ideological rifts and internal divisions (and therefore vulnerable to U.S. blandishments), his opponents in the Truman administration saw the Communist world as a cohesive ideological alliance that had to be confronted militarily because it could not be dissolved through positive inducements. Different beliefs about what held the Soviet alliance system together thus gave rise to very different policy prescriptions.¹⁰ Since then, U.S. opposition to leftist and Marxist movements throughout the world has been based primarily on the belief that ideological solidarity will make these regimes loyal allies of the Soviet Union.¹¹ Here again, an unstated hypothesis about the causes of alignment has been a key element of contemporary U.S. foreign policy.

Finally, can certain policy instruments cause other states to alter their alliance preferences? In particular, can the provision of economic or military aid create loyal allies? How easily and how reliably? Are foreign agents, advisers, and propaganda effective instruments of influence or control? The belief that these instruments will have a significant effect on alliance choices underlies U.S. concern for Soviet arms shipments to the Third World as well as the widespread conviction that states with a large Soviet or Cuban presence are reliable tools of the Kremlin.¹² Once again, an important element of U.S. national security policy rests on an untested assertion about the effectiveness of certain instruments on alignment.

These popular hypotheses paint a dramatic picture of U.S. insecurity. The belief that states are attracted to strength implies that any appearance of weakness or irresolution could damage the international position of the United States irreparably. The belief that ideology is a powerful cause of alignment implies that virtually all Marxist governments and leftist movements are reliable Soviet assets. And if foreign aid or foreign agents can create loyal satellites, then the United States is also threatened when non-Marxist countries receive material support from the Soviet Union. If these hypotheses are correct, in short, the United States faces an extraordinary challenge.

THE ARGUMENT

In the pages that follow, I argue that each of those beliefs is exaggerated. First, I demonstrate that balancing is far more common than bandwagoning. In contrast to traditional balance of power theorists, however, I suggest that states ally to balance against threats rather than against power alone. Although the distribution of power is an extremely important factor, the level of threat is also affected by geographic proximity, offensive capabilities, and perceived intentions. Thus I propose balance of threat theory as a better alternative than balance of power theory.

Second, the evidence shows that ideology is less powerful than balancing as a motive for alignment. Indeed, I argue that many apparently ideological alliances are in fact a form of balancing behavior. The record also shows that certain ideologies are extremely divisive. In other words, states sharing these ideologies are more likely to compete than to form durable alliances.

Third, I conclude that neither foreign aid nor political penetration is by itself a powerful cause of alignment. Even more important, neither is an effective way to gain leverage except under very unusual conditions.

Taken together, these results help explain why the international position of the United States is extremely favorable and likely to remain so. Because states balance against threats, (not against power alone), the United States has been able to create and sustain a global alliance whose capabilities exceed those of the Soviet Union and its own allies by a considerable margin. Ideological rifts (e.g., the Sino-Soviet split) reinforce Soviet isolation. Neither extensive foreign aid nor covert political penetration is likely to alter these tendencies.

Once we understand the origins of alliances, we can correctly judge the burden of preserving U.S. national security. It is relatively light. We can also see how recent U.S. foreign policy has been misguided, and we can identify how present errors can be corrected. Thus an enhanced theoretical understanding of the origins of alliances will yield important practical results as well.

THE ALLIANCE LITERATURE

Although the literature on alliances is enormous, much of it does not address the questions identified here.¹³ Most of the existing research on alliances has examined other issues, such as whether there is a relationship between alliance formation and the likelihood of war and whether the rate of alliance formation fits some specified mathematical model.¹⁴ Similarly, although the extensive collective goods literature on alliances implicitly assumes that alliances are created to provide security against threats, these models focus on explaining the distribution of burdens within existing alliances rather than on explaining why the alliances were formed in the first place.¹⁵

Nevertheless, a number of works do examine the origins of alliances. Among the traditional works on international politics are many accounts of individual alliances and several important theoretical treatments. Hans Morgenthau’s classic Politics among Nations, for example, contains a lengthy discussion of alliances supported by a variety of historical illustrations. Similar analyses are provided by George Liska and Robert L. Rothstein. Like Morgenthau, Liska relies on anecdotal evidence to support his points, whereas Rothstein bases his conclusions on case studies of Belgium and the Little Entente (Czechoslovakia, Rumania, and Yugoslavia) in the 1920s and 1930s.¹⁶

The traditional literature almost always falls within the broad compass of balance of power theory, although other hypotheses appear as well. Thus Liska writes that alliances are against, and only derivatively for, someone or something, and Morgenthau refers to alliances as a necessary function of the balance of power operating in a multiple state system.¹⁷ At the same time, however, Liska suggests that alignment…may [also] express ideological or ethnic affinities, and he states that opportunistic alignments may occur when a state believes that the effort to balance power will fail.¹⁸ To complicate matters further, Paul Schroeder has argued that alliances are formed either to (1) oppose a threat, (2) accommodate a threat through a pact of restraint, or (3) provide the great powers with a tool of management over weaker states¹⁹ In short, although most of the traditional literature relies heavily on balance of power concepts, doubts remain regarding the universal applicability of this hypothesis.

A limitation of the traditional approach is that its proponents rarely offer systematic tests of general hypotheses. Although Liska provides many apt examples, he does not attempt to assess the relative validity of his many interesting propositions. Case studies on individual alliances can provide more reliable evidence but may not tell us much about how different states would behave in different circumstances. Schroeder fares better on this score, because he supports his arguments with a survey of the major European alliances from 1815 to 1945. But even he does not examine which of the possible motives for alignment is most common or identify the factors that might affect the strength of each.²⁰ The question of which hypotheses provide the best guide for policy is left unanswered.

The belief that states ally to oppose powerful or threatening states has been challenged by several quantitative studies as well. Using sophisticated indices of national capabilities and a cooperation versus conflict scale created by coding diplomatic events, Brian Healy and Arthur Stein argue that European great power alliances from 1870 to 1881 result from bandwagoning (which they term the ingratiation effect) rather than from a desire to balance power.²¹ Because Germany’s ascendance after 1870 was associated with increased cooperation from most other states and because an anti-German alliance did not form in this period, they reject the hypothesis that states act to oppose any state [or coalition] which tends to assume a position of predominance.²²

Despite the sophistication and originality of this work, there are several problems. First, the methodology is suspect, because it assumes that coders can estimate the true meaning (i.e., the level of conflict or cooperation) of a discrete diplomatic event divorced from its historical context.²³ Second, the results are based on events that were atypical. Germany’s favorable position (including the free hand it enjoyed during the Franco-Prussian War) was due primarily to the great effort Bismarck devoted to convincing others that Germany was not an aggressive state.²⁴ Far from refuting the tendency of states to oppose predominant powers, Healy and Stein’s work reveals that power can be less important than other factors, such as perceived intentions.²⁵ In particular, Germany failed to provoke a countervailing coalition because Bismarck’s adroit diplomacy made friendship with Germany seem both possible and preferable. The ingratiation effect may thus be largely a testimonial to Bismarck’s diplomatic artistry.

Another challenge to balance of power theory has emerged from the ranks of expected utility theorists. Claiming to offer a formal theory of how states choose alliance partners, these authors suggest that states form alliances to increase their utility, measured in terms of security, risk, or welfare.²⁶ After correlating observed alliance dyads with various measures of utility, Michael Altfeld concludes that alliances do not appear to be random;…potential alliances which fail to increase both partners’ security almost never form.²⁷ In a similar work, David Newman claims to disprove balance of power theory by showing that states whose power is increasing are more likely to form alliances, because they are more attractive partners.²⁸

There are serious problems here as well. Expected utility theory does not identify who will ally with whom; it can only predict when states may seek alignment with someone. Furthermore, the fact that states whose capabilities are increasing tend to form alliances does not refute balance of power theory. After all, a state whose security position is threatened will probably attempt to increase its relative power (e.g., by spending more on defense) while simultaneously seeking an alliance with another state. Thus what Newman claims is a causal relationship (increases in power encourage alignment) may well be spurious. Even were this not the case, two states whose capabilities are increasing might well form an alliance against a third state that is growing still faster or that appears especially aggressive. This type of response would still be an example of balancing behavior.

Finally, some authors have used game theory to analyze alliance behavior. William Riker’s seminal work on political coalitions examines the optimal size of n-person alliances, and Glenn Snyder used two-person game theory to illuminate the trade-offs that states face in seeking to maintain allied support while avoiding the risk that their allies will entrap them in unwanted wars.²⁹ Both authors reach conclusions that are consistent with balance of power theory (i.e., Riker’s prediction that the players will seek a minimum winning coalition implies that states will join the weaker side).

Unfortunately, because game theory models are based solely on the distribution of power and the structure of possible payoffs, they do not take into account the impact of perceptions, ideology, and geography. Among other things, this limitation helps explain why Riker’s attempt to apply his insights to international politics is only partly successful.³⁰ And as Snyder admits, two-person game theory tells us more about the behavior within coalitions than it does about the players’ choice of partners: game theory does not predict who will align with whom.³¹ As with most of the other literature on alliances, in short, game theory has provided interesting answers to a different set of questions.

The existing scholarship on alliances is useful as a source of hypotheses. It does not, however, tell us which hypotheses are valid. As one student of this subject has observed, We have little if any reliable information about the relative potency of the various reasons why nations…join alliances.³² Thus, despite the enormous impact that debates about the origins of alliances exert on the conduct of U.S. foreign policy, the same disputes persist. What is needed is a strategy for resolving them.

RESEARCH METHODS AND PROCEDURES

The method I have employed is straightforward. The principal historical evidence I have used is the diplomatic history of the Middle East between 1955 and 1979.³³ Through a survey of these events, I have identified thirty-six separate bilateral or multilateral alliance commitments, involving eighty-six national decisions. After identifying the motives that led each state to select certain allies at different times, I have compared these results with the predictions of each hypothesis. Two questions are central: (1) Which hypothesis explains the greatest number of alliances? and (2) Are there identifiable conditions that affect which type of behavior is to be expected (i.e., which hypothesis is likely to apply)?

This approach is designed to overcome some of the limitations found in the works described earlier. Although historical case studies provide the most detailed evidence regarding the causes of a particular alliance, attempting to test several general hypotheses through a single case study is obviously problematic. Comparative case studies are more promising, but a large number of cases would be needed to establish valid conclusions. This difficulty is especially troublesome when the hypotheses under consideration are not mutually exclusive (a problem faced throughout this book).

Reliance on historical anecdotes or a large statistical sample is equally troublesome. Anecdotal evidence cannot tell which causes are most powerful or widespread, and statistical manipulations cannot provide direct evidence about the perceptions and motivations that inspired a particular alliance decision. Nor can they readily take into account the novel contextual features of a given case.

In order to overcome these limitations, I have employed a methodology that combines the features of a focused comparison and a statistical-correlative analysis.³⁴ By examining a large number of alignments (thirty-six), I have enhanced the external validity of my results and increased their robustness. Even if the ambiguity of the evidence leaves my interpretation of a few cases open to debate, my conclusions are likely to be valid as long as most of the analysis is sound. Finally, because my evaluation of each alliance is based on a careful reading of available historical accounts, the results can be informed by evidence on perceptions and motivations as well as the impact of unique contextual factors. Given the objectives of the study, this compromise between generality and specificity seems necessary and appropriate.

Definitions

I use the terms alliance and alignment interchangeably throughout the book. For my purposes, an alliance is a formal or informal arrangement for security cooperation between two or more sovereign states. Employing this rather broad definition makes sense for several reasons. First, many contemporary states are reluctant to sign formal treaties with their allies. To limit my analysis to formal alliances would omit a large number of important cases. Second, precise distinctions—for example, between formal and informal alliances—would probably distort more than they would reveal. There has never been a formal treaty of alliance between the United States and Israel, but no one would question the level of commitment between these two states. Changes in that commitment, moreover, have been revealed primarily by changes in behavior or by verbal statements, not by the rewriting of a document. Similarly, the Soviet Union and Egypt did not sign a formal treaty until 1971 but were obviously close allies long before then. And the 1971 Soviet-Egyptian Treaty of Friendship and Cooperation was actually a sign of growing tension between the two countries, not a symbol of enhanced commitment. Finally, any effort to take formal statements of inter-Arab solidarity at face value would be fraught with peril, as any student of Middle East politics knows. Thus an attempt to employ a strict typology of alliance commitments could easily be misleading, because the true meaning of either formal or informal arrangements is likely to vary from case to case.³⁵

Even more important, establishing a strict typology of commitments is simply not necessary for my purposes. I am interested in identifying the broad forces that lead states to support one another in international affairs, but I do not seek to explain the precise arrangements the parties ultimately choose. The specific commitments that allies accept will reflect a host of idiosyncratic features that are unlikely to be easily generalized. Thus I make no claim to be able to predict exactly how states will choose to implement their mutual commitments, but I do seek to explain why they choose to do so in the first place and to identify which of several potential partners they are likely to prefer. Resolving these questions will be challenge enough.

Why the Middle East?

I have chosen to investigate alliance formation in the Middle East for several reasons. First, the Middle East has been and remains an area of considerable strategic importance. Its importance is revealed by the efforts that both superpowers have devoted to acquiring and supporting allies in the region, efforts that have led to serious superpower confrontations on several occasions.³⁶ Second, alliance commitments in the Middle East have shifted frequently throughout the postwar period, as these states adjusted to changing internal and external circumstances. As a result, the diplomacy of the Middle East provides a large number of cases for consideration and is likely to reveal more about the factors that determine alliance choices than would examination of a less turbulent region.

Most important of all, the Middle East provides a strong test of many familiar hypotheses. Because most propositions about alliance formation (or international relations theory in general, for that matter) have been derived from the history of the European great power system, it is especially appropriate to examine their utility in predicting the behavior of states that are neither European nor great powers. Moreover, many of these regimes are relatively young and lack the diplomatic experience and traditions of the European great powers. Thus, if familiar hypotheses apply to this region as well, that is strong testimony to their explanatory power. Furthermore, the Middle East has been swept by intense ideological rivalries, major shifts in relative power, and significant superpower involvement throughout the period in question. As chapter 2 will show, these factors lie at the heart of the most popular hypotheses about alliance formation. Accordingly, the Middle East is an especially appropriate arena within which to assess them.

Methodological Barriers

Despite these strengths, I have faced a number of potential methodological problems in writing this book. Although several can be only partially alleviated, none presents an insurmountable barrier.

As I already noted, the concept of alliance is difficult to define and measure with precision. As will become clear in the next chapter, the same is true for such independent variables as level of threat and ideological solidarity. Moreover, many of the alliances examined in this study are overdetermined: they result from a number of separate causes. In such circumstances, measuring the importance of each different cause precisely (in order to distinguish between different hypotheses) is extremely difficult. To deal with this problem, I consider a large number of separate alignments while remaining alert for crucial cases that support one hypothesis while excluding others. I also include direct evidence (e.g., elite testimony) that identifies which causal factors were the most important in a particular instance.

A second potential difficulty arises from my focus on the Middle East. It might be argued that this region is sui generis, that any results derived from examining alliances in the Middle East cannot be applied to other areas or to different time periods. It might also be suggested that adequate understanding of Middle Eastern diplomacy requires specialized training and a knowledge of unique cultural factors that I cannot claim.³⁷

Although these considerations are not without merit, they do not present an overwhelming barrier. The argument that the Middle East is sui generis applies with equal force to any other region. Yet international relations scholars have long relied on historical cases and quantitative data drawn from European diplomatic history without being accused of a narrow geographic, temporal, or cultural focus. Nonetheless, I have addressed this problem in two ways. First, I have drawn on European history in elaborating the different hypotheses. Second, after testing these propositions on the Middle East, I have applied them to the current array of global alliance commitments. By drawing on evidence from several different contexts, I have significantly reduced the limitations of relying primarily on evidence from the Middle East.

Of course, it may be true that alliances in the Middle East exhibit unique patterns of behavior. If so, that situation is not so much a barrier to theoretical work as a challenge to the theorist. The task is to explain how unique regional characteristics produce the observed behavior. Indeed, that is precisely what I attempt in later chapters, when I examine the impact of pan-Arabism on alliances in the Arab world.

Finally, because I am relying primarily on secondary sources in investigating these alliances, my assessment of contemporary Middle East diplomacy rests on the scholarship that area specialists have provided. Unfortunately, the historiography on recent Middle East politics is uneven—because of the difficulty of archival research and the obvious biases with which many accounts are written. Even primary sources and elite testimony must be treated with caution, given the instrumental motives that most participants in Middle East diplomacy have in offering their accounts. To compensate for these problems, I have tried to document events and arguments as extensively as possible, relying on multiple sources and the most widely accepted historical accounts.

I will proceed as follows. In chapter 2, I will develop the concepts and hypotheses that will guide the remainder of the study. In Chapters 3 and 4, I will describe the evolution of alliance commitments in the Middle East,