Perverse institutionalism: NATO and the Greco-Turkish conflict.

Last year the U.S. Senate overwhelmingly approved the expansion of the Atlantic alliance. Whereas some advocates of enlarging NATO, particularly Eastern European leaders for whom the Soviets' iron grip is an all too recent memory, stress the extension of the alliance's traditional deterrent function, others acknowledge that Russia is in no position to reconquer its former empire. Rather, they argue that membership in NATO would stabilize the region by filling the power vacuum and eliminating the need for security competition. Traditionally a volatile area, East-Central Europe is rife with potential irredentist and ethnic conflicts, and NATO can help arbitrate and limit these disputes.(1) Critics have denounced the move as unnecessarily provocative to Russia, and they have also decried its hefty cost.(2) But they have not challenged the claim that alliances create zones of peace.(3)

Policymakers have swiftly embraced this intuitively appealing argument. In a 1996 campaign speech, President Bill Clinton trumpeted the virtues of alliance: "Through NATO, Western Europe became a source of stability instead of hostility. France and Germany moved from conflict to cooperation. . . . I came to office convinced that NATO can do for Europe's East what it did for Europe's West."(4) This logic is attractive, but it overstates the alliance's achievements. The triumph of Franco-German reconciliation must be balanced by the failure to suppress the persistent discord within NATO's southeastern flank. At least once in every decade since World War II, Greece and Turkey have nearly come to blows, and a host of more minor incidents and skirmishes have further marred their relations. Historians have maintained that the pair pursued their rivalry despite NATO's best efforts, yet I argue that, although the alliance did not create the divisive issues separating these longtime antagonists, it bears a significant degree of responsibility for the tensions between them.

Alliances can, under certain conditions, intensify conflict among their members. Unlike structural realists, who believe that institutions matter only at the margins of international relations, I contend that alliances are more than mere means of accumulating power. But unlike neoliberal institutionalists, who portray institutions as mitigating the dangers of anarchy and whose reasoning bolsters the Clinton administration's optimistic expectations, I doubt that their impact is always beneficial. In this article I seek to bridge the gap between these two approaches by taking the first steps toward a "realist institutionalism," marrying the belief that institutions matter with a skepticism as to their effects.(5)

I also seek to right a notable bias in the alliance literature, which has generally ignored the possibility of serious conflict among allies.(6) Over two decades ago Paul Schroeder argued that political scientists' emphasis on the operations of the balance of power had flattened the historical richness of alliance patterns and purposes and that the record demonstrated "that alliances in practice do not always serve to increase a nation's power and security, and that allies often clash with each other more than they unite in common cause." Alliances are "associative-antagonistic relationships": they are as often devices uniting rivals as means of linking friends.(7)

Schroeder's cry has gone largely unheeded over the last twenty years. Realist studies, focused on alliances as balance or bandwagon, have not concentrated on their internal politics.(8) Scholarship on intra-alliance conflict has explored the problem of burden sharing, not conflicts that strike at the core of the commitment.(9) Few institutionalists have ventured into security studies, and their insights have spoken more to the persistence of alliance institutions and policies or to how alliances foster cooperation.(10) Other scholars in traditions ranging from formal theory(11) to social psychology(12) have built their analyses on the assumption of mutuality of purpose. Even Glenn Snyder, who has shown unusual sensitivity to the political bargains at the heart of alliance, denies the possibility of armed conflict between allies, arguing that alliances generate patterns of amity and enmity, allies support growth in each other's power resources, and within alliance states develop shared interests.(13) Yet even as allies provide security, they can pose real threats, not just limits to a state's freedom of action. Alliance does not entirely eliminate conflicts of interest and concerns over relative position.

I examine the effects of membership in an institutionalized multilateral alliance on the relations of small powers and offer three core arguments in the realist institutionalist vein.(14) First, when accession to the alliance brings a security guarantee to small states, it eliminates the primary threat from their horizon and allows them to focus on secondary foreign policy objectives.(15) This shift becomes dangerous when it exposes potentially militarizable conflicts of interest with fellow allies, often long-standing regional rivals. States discontented with the status quo spot an opportunity to further their aggressive ends, while satisfied states, suspicious of their longtime enemies' ambitions, replace the alliance's target with their rivals as the major security threat. The states' entry into the alliance disrupts the bonds born of shared fear, permits the prosecution of a conflict that the common threat had moderated, if not entirely hidden, and revives competition among the allies.

Second, membership in an institutionalized multilateral alliance often bestows greatly improved military capability on both the qualitative and quantitative dimensions, as more powerful alliance partners provide better equipment and training, more advanced weaponry, and larger numbers of weapons. Among states engaged in a limited dispute, these arms transfers exacerbate the security dilemma by creating suspicion that others harbor broader revisionist goals and by triggering a spiral of diplomatic tension.

Third, alliances, like other institutions, provide states with the means for issue linkage as well as with greater transparency, but such institutional functions need not yield cooperation and will contribute, under conditions specified later, to the deterioration of relations. Moreover, institutions are loci of power, and, rather than foster cooperation, the alliance can itself become the object of struggle as states seek to dictate its agenda, alter the distribution of benefits, and use its forums to generate political support. However, the central claim here - that alliance can deepen and intensify conflict among its members - does not deny that some features of alliance mitigate disputes nor does it gainsay that great powers can use the allied relationship to ease tensions and prevent the outbreak of war. Rather, I seek to highlight the multiple, contradictory effects of international institutions.

I first present neorealist and neoliberal institutionalist insights into the consequences of alliance for potential conflicts among members. I explore in greater detail the logic of the realist institutionalist propositions laid out briefly earlier. I then examine the case of Greco-Turkish relations as a probe of these hypothesized dynamics. I conclude by discussing the implications of this analysis for NATO enlargement.

International Relations Theory and the Consequences of Alliance

Structural Realism

When faced with either superior power or a great threat, states wishing to survive and maintain their autonomy must balance through either internal means (by mobilizing resources) or external means (by forming an alliance) or some combination of the two.(16) Alliances do not pattern state interactions but are essentially epiphenomenal intervening variables. In John Mearsheimer's stark formulation, "Realists maintain that institutions are basically a reflection of the distribution of power in the world. They are based on the self-interested calculations of the great powers, and they have no independent effect on state behavior." Alliances are short-term tools of foreign policy, easily dissolved and partners easily swapped, and allies can only occasionally constrain each other's behavior.(17) Structural realists acknowledge that the NATO allies generally paid little attention to unequal gains during the Cold War, but the alliance, qua alliance, did not alleviate the European security dilemma. The key was the allies' fear of the Soviet Union, which drove the formation of the alliance and reduced relative gains concerns. Alliances are simultaneously central and peripheral to these theories.

Neorealists would clearly be highly skeptical that alliances bring stability to troubled regions, but they would be equally skeptical that alliances can prove destabilizing. International institutions, including alliances, matter only at the margins and cannot alter states' calculations regarding matters of considerable national interest. Thus, NATO enlargement would not influence the course of potential conflicts in East-Central Europe, whether for good or ill. Although neorealists would reach these conclusions even when NATO appeared most effective, their arguments should seem more convincing in light of the disappearance of the Soviet threat. Agreeing with Kenneth Waltz that "NATO's days are not numbered, but its years are,"(18) they predict that in the long term the alliance will become at best an empty shell, preserved for its public relations and symbolic value, in which members no longer fulfill core treaty commitments.(19) Even if neorealists concede that heavily institutionalized alliances can influence state behavior, today's NATO, with its principal threat dissolved and its casus foederis unclear, would have only a minimal impact.

Neoliberal Institutionalism

A decade ago Robert Keohane insisted that "alliances are institutions."(20) Indeed, as Keohane briefly noted, the neorealist view does not describe the major postwar alliances, which possessed a relatively stable membership and acquired a formal bureaucratic organization with complex decision-making routines and differentiated functions. Were institutions epiphenomenal to the distribution of power, states would seem irrational for expending significant material resources and political capital on their formation and maintenance and on the subsequent battles for their control. Institutions, whether alliances or international economic regulatory agencies, shape the costs, information, even the preferences, and hence the actions of their constituent members.

Neoliberal institutionalists present several mechanisms linking institutional arrangements to the promotion of state welfare through international cooperation.(21) First, rule-governed interaction encourages an increased number of transactions among participants, thereby discouraging cheating among states that are reasonably sensitive to the "shadow of the future." Second, by providing a framework for further agreement, institutions reduce the costs associated with the bilateral negotiation, monitoring, and verification of individual accords, making cooperation more profitable and attractive.

Third, institutions link issue areas, that is, different zones of state interaction, creating greater opportunities for side payments and raising the price of defection. As Keohane explains, "Clustering of issues under a regime facilitates side payments among these issues: more potential quids are available for the quo. Without international regimes linking clusters of issues to one another, side payments and linkages would be difficult to arrange in world politics."(22) Such linkages also make cheating more costly because they provide the victim with multiple issue areas in which to exact revenge.

Fourth, institutions increase the level of transparency among members, raising the expected cost of cheating within an institutional context, as the likelihood of detection and retaliation increases. Moreover, the expected gain decreases, since transparency permits prospective victims to take protective measures. Fifth, institutions prescribe regular interaction among individual decision makers and policymaking groups from member states, generating agreement on policy questions among participants, who then coordinate their respective states' policies.(23) Even institutionalists outside the epistemic-community research program at times abandon their theoretical commitment to exogenous, fixed interests and aver that interaction can alter preferences over the long run.

Despite Keohane's plea, relatively little work has examined alliances from an institutionalist perspective,(24) yet this approach yields expectations for intra-alliance relations. If in a neorealist world states are reluctant to remain in alliances longer than they must because they do not trust others, neoliberals expect that formalized, well-integrated alliances help their members overcome incentives to internal mobilization, attenuate the fear of unequal gains, and foster numerous cooperative ventures. Not all alliances have as deep an institutional form as NATO, however, and neoliberals would expect the degree of cooperation to correlate with the degree of institutionalization. For an alliance akin to NATO, evidence consistent with the institutionalist approach would reflect at least some of the following processes:

1. States use the issue-linkage opportunities provided by the alliance to offer side payments, eventually reaching a mutually satisfactory compromise.

2. Participation in alliance functions generates increasing knowledge of and trust in others' benign motives, decreasing concerns of relative gains and perceptions of intra-alliance threat. Further, if more powerful states provide military assistance, the state's absolute level of security rises.

3. Joint force structures and planning reshape each member's military posture to reflect alliance interests, reducing the fear of attack and obviating security dilemma dynamics. Force specialization constrains members' abilities to use their forces for exclusively national objectives.

4. Military officers that interact regularly with their allied colleagues are socialized into a shared alliance culture, transcending their national interest.(25) These officials endeavor to shape their states' preferences accordingly.

Common Flaws

Thanks to simplifying assumptions regarding the interests of allies, neither structural realism nor neoliberal institutionalism is well suited to elucidating the politics within alliance. Neorealists typically portray international institutions as epiphenomenal to national power and alliances as fleeting agglomerations of capabilities: they are "only temporary marriages of convenience, where today's alliance partner might be tomorrow's enemy, and today's enemy might be tomorrow's alliance partner."(26) However, at times neorealists have implied that once states conclude that the international threat environment demands the formation of an alliance, their behavior exhibits a near identity of interest and residual clashes do not arise. Although Waltz famously noted that states "are compelled to ask not 'Will both of us gain?' but 'Who will gain more?'" he also explained European integration as a consequence of the Soviet menace. Bound together by their common fear, Germany, France, and the other former great powers of Europe overcame centuries of hostility and strife and even welcomed their fellows' growth and development. Among these consumers of security, "not all impediments to cooperation were removed, but one important one was - the fear that the greater advantage of one would be translated into military force to be used against the others."(27)

Yet, contra Waltz, states do not put aside their colliding interests when they enter into alliance. The decision to join forces reflects their common estimate of a particular threat, but they will continue to disagree on other matters and may pursue those goals with great vigor. For example, as a recent study of the early Cold War in Asia notes, "Although the United States and its allies agreed on the broad outlines of containment strategy, . . . historians should be sensitive to the regional priorities of America's allies and to the subtle competition for spheres of influence that often underlined their diplomacy towards the Third World." As much as France supported U.S. efforts to counter the Soviet Union in Europe, it also perceived U.S. diplomacy in Vietnam as a threat to its overseas empire.(28) Insofar as states continue to seek aims beyond the casus foederis of the alliance, they remain competitors on the international scene, blurting the distinction between allies and adversaries. Both of these neorealist strains of thought on intra-alliance relations fail to capture the complexity of states' motives and the resulting interactions.

On occasion neoliberal institutionalists have displayed greater sensitivity to this problem. In the context of a related debate, Keohane observed that "the concept of relative gains becomes fundamentally ambiguous as the number of actors becomes greater than two. Relative gains for state B in a dyadic relationship with state A may help state A in a contest with state C, if A and B are allies and B and C are adversaries. Which 'relationship' counts?" However, rather than explore the interaction of absolute and relative gains within "cooperative clusters," such as alliances, neoliberal institutionalists fall back on the claim that such "defensive cooperation" allows groups of states to achieve relative gains collectively versus others and ignore the distributional consequences among themselves.(29)

Neoliberals commit a still more serious error by treating all institutions, regardless of the sphere in which they operate, as …