Chain gangs and passed bucks: predicting alliance patterns in multipolarity Thomas J. Christensen and Jack Snyder
Kenneth Waltz's rigorous recasting of traditional balance-of-power theory has provided the intellectual foundation for much of the most fruitful recent work in the fields of international politics and national security.1 But there is a tension between Waltz's theory and those who apply it in their practical research agendas. Waltz's is a theory of international politics; it addresses properties of the international system, such as the recurrence of war and the recurrent formation of balances of power.2 Those who have applied Waltz's ideas, however, have normally used them as a theory of foreign
This article combines the work of two unpublished papers. The theoretical sections are derived from Christensen's "Chained Gangs and Passed Bucks: Wallz and Crisis Management Before the Two World Wars," Columbia University, December 1987. The case study materia) is based on Snyder's "Offense, Defense and Deterrence in the Twentieth Century," a paper presented at the Conference on the Strategic Defense Initiative, University of Michigan, November 1986. We are grateful to Charles Glaser, Harold Jacobson, Robert Jervis, Stephen Krasner, Helen Milner, David Reppy, Cynthia Roberts, Randall Schweller, Stephen Van Evera, Stephen Walt, Deborah Yarsike, William Zimmerman, and an anonymous reviewer for comments on various earlier drafts. We also thank the Social Science Research Council and the MacArthur Foundation for Christensen's financial support and the Program in International Peace and Security Studies at the University of Michigan for sponsoring Snyder's original paper.
1. Kenneth Waltz, Theory of International Politics (Reading, Mass.: Addison-Wesley, 1979).
2. We feel no need to take a position on the epistemofogical debates surrounding Waltz's theory, spurred in particular by John Ruggie and Robert Cox. We are satisfied to accept Waltz's scheme as what Cox terms a "problem-solving theory." For current purposes, we hope to improve its problem-solving utility rather than to address its deeper epistemological adequacy. See Robert Keohane, ed., Neorealism and lis Critics (New York: Columbia University Press, 1986), especially pp. 208 and 214. See also David Dessler. "What's at Stake in the Agent-Structure Debate?" International Organization 43 (Summer 1989), pp. 441-74; and John S. Dryzek, Margaret L. Clark, and Garry McKenzie, "Subject and System in International Interaction," International Organization 43 (Summer 1989). pp. 475-504.
International Organization 44, 2, Spring 1990
© 1990 by the World Peace Foundation and the Massachusetts Institute of Technology
138
(
Inte).
(
■ aal Organization
policy to make predictions about or prescriptions for the strategic choices of states.3
This is a problem because for a particular state in particular circumstances, any foreign policy and its opposite can sometimes be deduced from Waltz's theory. In multipolarity, for example, states are said to be structurally prone to either of two opposite errors that destabilize the balancing system. On the one hand, they may chain themselves unconditionally to reckless allies whose survival is seen to be indispensable to the maintenance of the balance. This, Waltz argues, was the pattern of behavior that led to World War I. On the other hand, they may pass the buck, counting on third parties to bear the costs of stopping a rising hegemon. This was the pattern that preceded World War II.4
For Waltz, as a systemic theorist, this is not a crippling problem. He
_deduces logically that multipolarity is structurally prone to instabilities, and
the two major cases of this century illustrate his theory suitably. But for those who would use Waltz as a theorist of foreign policy, there is a problem. To explain, predict, or prescribe alliance strategy in particular circumstances, they need to specify which of the two opposite dangers—chain-ganging or buck-passing—is to be expected in those circumstances. An explanation that can account for any policy and its opposite is no explanation at all. Likewise, a prescription that warns simultaneously against doing too much and doing too little is of less use than one that specifies which of the two errors presents the more pressing danger in particular circumstances.
This does not mean that Waltz's insights about chain-ganging and buck-passing are of no use in a theory of foreign policy. Rather, it means that his ultraparsimonious theory must be cross-fertilized with other theories before it will make determinate predictions at the foreign policy level. Users of Walt2's theory already do this at various levels of explicitness, factoring in military technology, geography, and power variables that go beyond the mere counting of great power poles. In particular, they combine Waltz's insights with the variables stressed in Robert Jervis's version of the security dilemma theory.3 They also factor in biases affecting how policymakers and
3. See. for example, Stephen Walt, The Origins of Alliance (Ithaca, N.Y.: Cornell University Yess, 1987); and Barry Poser), The Sources of Military Doctrine (Ithaca, N.Y.: Cornell Uni-ersity Press, 1984). By "theory of foreign policy" we mean a theory whose dependent variable
the behavior of individual states rather than the properties of systems of stales. It does not :fer to a theory that explains all aspects of a state's foreign policy.
4. Waltz, Theory of International Politics, pp. 67 and 165-69.
5. See Robert Jervis. "Cooperation Under the Security Dilemma," World Politics 30 (January >78), pp. 167-214; Stephen Van Evera, "Causes of War," Ph.D. diss., University of California, :rkeley, 1984; Stephen Van Evera, "The Cult of the Offensive and the Origins of the First orld War," in Stephen E. Miller, ed., Military Strategy and the Origins of the First World
Alliance patterns 139
soldiers perceive the balance-of-power problem that faces them.6 By complicating the specification of the state's position in the international system— and in some cases by introducing the role of perception—determinate predictions can be made.7
Though a few scholars have de facto been working this way for some time, their method warrants more explicit specification. Toward this end, we will attempt to explain the opposite alliance choices of the European great powers before World Wars I and II, starting with Waltz's theory and adding a minimal number of variables from security dilemma theory and from perceptual theories that are necessary to derive a theoretically determinate and historically accurate account. In a nutshell, we argue that given Europe's multipolar checkerboard geography, the perception of offensive military advantages gave rise to alliance chain-ganging beforel9ITPwlvefeas the perception of defensive advantages gave rise to buck-passing before 1939. These perceptions of the international conditions constraining strategic choice were, however, misperceptions, rooted in patterns of civil-military relations and the engrained lessons of formative experiences. In the first two sections of the article, we review the theories needed to underpin this interpretation and show how they can be combined in a relatively parsimonious fashion. In subsequent sections, we present short case histories demonstrating the historical plausibility of the interpretation and offer comments on issues for further research.
This exercise should be of practical as well as theoretical and historical interest. Arguably, the world is again becoming more multipolar. Japan has caught up with the Soviet Union in terms of gross national product. Both the United States and the Soviet Union are playing a less dominating global
War (Princeton, N.J.: Princeton University Press, 1985), pp. 58-107; Stephen Van Evera. "Offense, Defense, and Strategy: When Is Offense Best?" paper presented at the annual . meeting of the American Political Science Association, Chicago, 1987. Fora work that preceded the publication of Waltz's and Jervis's theories but made many similar points, see George Quester, Offense and Defense in the International System (New York: Wiley, 1977), especially chap. 10 on alliance behavior in World War I.
6. In addition to the above-mentioned works by Van Evera, see Posen, Sources of Military Doctrine; and Jack Snyder, "Civil-Military Relations and the Cull of the Offensive, 1914 and 1984," in Miller, Military Strategy, pp. 139-^0. Levy points out that difficulties in measuring offensive and defensive advantage make such judgments problematic for social scientists as well as elusive for policymakers. See Jack S. Levy, "The Offensive/Defensive Balance of Military Technology," International Studies Quarterly 28 (June 1984), pp. 219-38.
7. By "determinate predictions" we mean thai if all other factors (such as checkerboard geography) are held constant, then knowing the polarity of the system and the perceived offense-defense balance will theoretically suffice to predict the alliance behavior of states. Of course, in the real world, other factors having some effect on alliance behavior may not be held constant, making our predictions probabilistic rather than strictly "determinate."
V
140   International Organization
role now than they were when Waltz began to write about the stability of the bipolar balance. As in the periods before World Wars I and II, Germany and Russia may once again be contending for markets and influence in an increasingly heterogeneous, independent, yet vulnerable belt of Eastern European states. Will multipolar alliance patterns make a reappearance? And if so, which pattern—chain-ganging or buck-passing? For which problem should scholars and policymakers begin devising antidotes?
As the new configuration of power emerges, we will need to know not only about its polarity but also about the key security dilemma and perceptual variables that interact with polarity in shaping international alignments. If the potentially unstable condition of multipolarity reemerges, we will need to know how its effects can be mitigated. Since the polarity of the system -is generally not subject to conscious manipulation by policymakers, our attention should be especially directed toward the variables that are somewhat more subject to conscious control, variables such as the offense-defense balance of technology and perceptions of it.
Chain gangs and passed bucks
Waltz argues that the structure of the international system determines what types of international behavior will be rewarded and punished (the process of selection) and, as a result, what types of foreign policy will seem prudent to actors in the system (the process of socialization). This structure comprises a constant element, anarchy, and a variable element, polarity. The fundamental, invariant structural feature, international anarchy, generally selects and socializes states to form balancing alignments in order to survive in the face of threats from aggressive competitors. However, a variable structural feature, polarity, affects the efficiency of the balancing process.
,   In multipolarity, two equal and opposite alliance dilemmas impede efficient \ balancing,8 The first js the chain-gang problem. Injmiltipolarity, the approximate equality of alliance partnersleac"
.. r_________to a high degree~15f-security
interdependence within an alliance. Given the anajxhjc_seiiing and this relative equality, each state feels its own Security is integrally intertwined with the security of its ajfiance j>artjjersTrAs a result, any nation that marches to war inexoraj3]y_j}^ with it. No state can restrain a
reckless ally by threatening tcTsiTouTTIie^conflict, since the demise of its reckless ally would decisively cripple its own security.9
J.:l?r.Z1^ arg"ml,!.ts ,hat use *e concepts of entrapment and abandonment, see Glenn
Alliance patterns 141
Waltz's entirely apt example of this dilemma is World War I: If Austria-Hungary marched, Germany had to follow: the dissolution of the Austro-Hungarian Empire would have left Germany alone in the middle of Europe. If France marched, Russia had to follow; a German victory over France would be a defeat for Russia. And so it was all around the vicious circle. Because the defeat or the defection of a major ally would have shaken the balance, each state was constrained to adjust its strategy and the use of its forces to the aims and fears of its partners.10
Vln short, as one member of the_chain gang stumbles off the precipice, the J/other musTfoTiow. Hyperactive balan^mgT5elTaTibT^h"reatenslhe stability of J the system by causing unrestrained warfare that threatens the survival of ' some of the great powers that form the system's poles.
The secondhand opposite, pathology of multipolarity is buck-passing. In ju the face of a rising threat, balancing alignments fail to form in a timely fashion / (/ because some states try to ride free on other states' balancing efforts. They f may do this because they wish to avoid bearing unnecessary costs or because they expect their relative position to be strengthened by standing aloof from the mutual bloodletting of the other powers. Waltz illustrates with World War II:
French Foreign Minister Flandin told British Prime Minister Baldwin that Hitler's military occupation of the Rhineland in 1936 provided the occasion for Britain to take the lead in opposing Germany. As the German threat grew, some British and French leaders could hope that if their countries remained aloof, Russia and Germany would balance each other off or fight to the finish. Uncertainties about who threatens whom, about who will oppose whom, about who will gain or lose from the actions of other states accelerate as the number of states increases.'1
Barry Posen, in the same vein, shows that the defensive military postures adopted by both Britain and France in the face of German expansion were designed to pass the cost of fighting Germany to other allies.12 As a result, the balancing process operated inefficiently, giving the aggressor a chance to overturn the balance by eliminating the system's opposing poles through piecemeal aggression.
Waltz argues cogently that neither chain-ganging nor buck-passing dilemmas can arise in bipolarity. Bipolar superpowers do not need to chain themselves to small, reckless allies, since the superpowers are not dependent on allies for their survival. Superpowers also do not pass the buck, since smaller allies cannot possibly confront the opposing superpower alone.'3
10. Ibid., p. 167.
142   International Organization
Alliance patterns 143
Superficially, it might appear that Waltz's argument about bipolarity suffers from as much underdetermination as his argument about multipolarity does. That is, Wajtz. appears to associate bipolarity with two_£gujal and opposite stances S^rd_peripjieral allies. On th^onejiand, since the balance oi^werTfi^ipolarityTiingesorTtfie superpowers' internal efforts to generate power capabilities, the loss of peripheral allies is largely irrelevant.JThys, chain gangs need not occur, and tjhejnjgejpow^ non-invoTvfinientjn peripjhjrai_disputes. O^^e-otheMiand, each superpower understands that only it has the power to resist encroachments on third parties by the other. Consequently, the buck cannot be passed to others, so superpowers in bipolarity tend to "overreact" to threats in the periphery, Waltz says.14
— We believe that the tension in this part of Waltz's argument is not difficult to resolve. Since superpowers have no strong incentive to intervene in the periphery, the issue of buck-passing should be irrelevant. The structural logic of limited involvement should override the opposite logic leading to overreaction. Or put somewhat less categorically, bipolar superpowers should practice a policy of limited liability in intervening in defense of peripheral allies. That is, they should incur the costs of intervention only in proportion to the power assets that are at risk. In bipolarity, these assets will always, by definition, be of marginal importance, so superpower interventions in the periphery should be limited. Waltz's policy prescriptions suggest that this is his view.15
The behavior of Cold War policymakers has sometimes violated these prescriptions, but we believe that this had more to do with perceptual or domestic political factors than with the structural properties of bipolarity. The structural consequences of bipolarity, unlike those of multipolarity, do lead to a determinate prediction about alliance strategy, even though empirically the behavior of the superpowers sometimes falsifies that prediction. In short, bipolarity is an ameliorator rather than a panacea. It does not entirely rule out overreactions and underreactions caused, for example, by domestic politics or faulty ideas, but bipolarity mitigates the structuralxauses of such problems.
In creating a theory of international politics, Waltz is interested mainly in showing that a system of two is more stable than a system of many. He therefore evinces no interest in predicting which pathology of multipolarity will appear in particular circumstances. For his purposes, this may be acceptable. Even on Waltz's own terms, however, the failure to specify when chain-ganging and buck-passing will occur is at least mildly troubling. Waltz's
14. Ibid., pp. 169 and 171-72.
15. Ibid., chap. 9. This certainly is the vieW,of Waltz's students. See Stephen Walt, "The Case for Finite Containment: Analyzing U.S. Grand Strateev.-" Internntinnnl —
argument hinges on the notion that the structure of the system—that is, the number of poles—selects and socializes states to a particular form of behavior. But if chain-ganging and buck-passing, two starkly opposite forms of behavior, are equally selected under multipolar circumstances, how do states become socialized? Arguably, more information about the international setting must be provided in order for Waltz's crucial process of socialization to set states on a determinate path.
This indeterminacy is even more troublesome for Waltz's students, who attempt to adapt Waltz's ideas into a theory of foreign policy. For example, if Posen is to argue that the structural requirements of multipolarity led France to adopt a military strategy designed to pass the buck to Britain, then he must show that other strategies were not equally consistent with the logic of multipolarity. Posen understands this but leaves the solution to the problem only partially expressed.
Why, Posen asks, did states that passed the buck in the 1930s chain gang in the 1910s? In a single paragraph, Posen explains this as a consequence of the different effects of perceived offensive and defensive advantages on security calculations in multipolarity. Perceived offensive advantage before 1914 meant that war was considered cheap. Moreover, allies crucial to maintaining the balance of power were considered highly vulnerable to attack. Thus, states balanced aggressively and unconditionally. By contrast, in the 1930s, perceived defensive advantage led to buck-passing. According to Posen, "Each state had an interest in passing the costs of its own defense to its allies, because these costs [of defensive, attritional war] were high." He adds that "there was a widespread belief in a defensive advantage, so states did not believe that their allies might fold" and that "leaving one's ally a little bit in the lurch was not seen to represent a high risk to the ally's survival or one's own."16
Without spelling out the theoretical underpinning behind these arguments, Posen appears to dispense with evidence that falsifies his argument in an unsatisfying, ad hoc manner. By spelling out the underlying logic more explicitly and by combining balance-of-power theory with security dilemma theory, we hope to show that Posen's insight can be used to resolve Waltz's indeterminacy as a theorist of foreign policy. Far from being an ad hoc sleight-of-hand, this is a parsimonious, productive theoretical innovation that has general applicability for scholars working in the realist tradition.
Posen also notes that it was perceptions of offensive advantages, driven by the biases of "out-of-cdnTroTmilitary organizations,'* which shaped policy through shaping perceptions of systemic incentives before 1914. This introduces still more variables, but parsimony is still not utterly lost. Forces within the state affect alliance behavior and grand strategy, but they do so
144 Into
nal Organization
by affecting perceptions of the international environment. Thus, domestic and perceptual forces can be cleanly plugged into parsimonious international system theories. The next section lays out a framework for doing this.
Polarity, the security dilemma, and perception
To turn Waltz's ideas into a theory of foreign policy that accurately explains alliance behavior before World Wars I and II, two complications must be introduced. First, the variable elements of international structure must be broadened to include not only polarity but also the security dilemma variables: technology and geography. Second, perception of the strategic incentives inherent in the systemic structure must be introduced as a poten-tially autonomous factor.
Waltz approvingly cites Jervis's writings on the security dilemma as sup-f,] port for the notion that states in international anarchy are condemned to ^ehave_cojnp£litiyely^. Indeed, Waltz's and Jervis's theories are cut from the same "cloth, both stressing dilemmas that stem from the requirements of self-help in an anarchical political order. Both agree, moreover, that the intensity of the security dilemma is not constant but instead varies with the vulnerability of states. Waltz explores the stabilizing consequences of bi-polarity, which are due in part to the superpowers' greater self-sufficiency and consequently lesser vulnerability to the vicissitudes of international anarchy." Jervis explores the stabilizing consequences of defensive and deterrent military technologies, as well as geographical configurations that make conquest more difficult. Both see the same problem: vulnerability leads to self-help strategie^jMtJeaye^yäTorie lesssecureTßoth conceive of the international order similarly: as an anarchy. And botlrsee greater invulnerability as the source of greaterstabuftyin international anarchy. There is no reason that their two theories cannot be combined in order to explore interactions between their variables.
These interactions include the connection between offensive advantage and chain-ganging and, conversely, the connection between defensive advantage and buck-passing. In multipolarity, the greater the vulnerability of states (that is, the more propitious the technology or geography for the attacker), the greater is the propensity to align unconditionally and to fight all-out in defense of an ally from the first moment it is attacked. This happens because the expectation of rapid, easy conquest leads states to conclude
17. This is at least implicit in Waltz's arguments about interdependence in his Theory of International Politics, pp. 143-46, juxtaposed to his arguments about the relative invulnerability of the bipolar superpowers, p. 172. Note also Waltz's remarks about firms on p. 135: "More than any other factor, relative size determines the survival of firms. Firms that are large in comparison to most others in their field find many ways of taking care of themselves—of protecting themselves against other law firm* "l
Alliance patterns 145
lhat allies essential to maintaining the balance of power will be decisively defeated unless they are given immediate and effective assistance. Conversely, the less the vulnerability of states, the greater is the tendency to pass the buck. This is due both to the expectation that other states, even singly, will be able to stalemate the aggressor without assistance and to the expectation that the process of fighting will be debilitating even for a victorious aggressor. Such an aggressor will pose a reduced threat to buck-passing onlookers who remain at their full, pre-war strength. Thus, Jervis's variables provide the determinate predictions that Waltz's theory needs in order to become a theory of foreign policy.'8
On theoretical grounds alone, we could be entirely satisfied with this minor and parsimonious yet productive addendum to Waltz's theory. Unfortunately, for empirical reasons, still further adjustments are needed to explain alliance dynamics before World Wars I and II. This is because soldiers' and policymakers' perceptions of offensive and defensive advantages before the two wars were almost exactly wrong. Therefore, we need to add a perceptual dimension to explain why technological circumstances of defensive advantage were seen as encouraging offensives in 1914, whereas circumstances that were objectively much more favorable to the attacker in the late 1930s were seen as discouraging offensives.
In principle, any number of perceptual biases might affect perceptions of the structure of international incentives. In fact, however, two main hypotheses enjoy the greatest plausibility. The first is that soldiers' and policymakers' perceptions of international structural incentives, including the offense-defense balance, are shaped by their formative experiences, especially the last major war. Thus, since European wars before 1914 had often been short and decisive, most people expected offensives to succeed. But after the experience of 1914-18, most people expected defensives to succeed.19 The second hypothesis is that uncontrolled militaries favor offensive strategies, and since civilian control over the military was much greater in the 1930s than in the 1910s, the military-fueled "cult of the offensive" no longer dominated strategic perceptions. Instead, a civilian-based "cult of the defensive," aimed at finding strategic excuses for buck-passing, may have had an equal but opposite impact.20 It is not our main purpose here to
18. For related discussions, sec Posen, Sources of Military Doctrine, p. 232; Van Evera, "The Cult of the Offensive," pp. 96-101; Van Evera, "Why Cooperation Failed in 1914," in Kenneth Oye, ed., Cooperation Under Anarchy (Princeton, N.J.: Princeton University Press, 1986), especially pp. 83-84; and Walt, Origins of Alliance, especially pp. 24-25, fn 31, and pp. 32 and 165-67.
19. For the theory underlying this hypothesis, see Robert Jervis, Perception and Misper-ception in International Politics (Princeton, N.J.: Princeton University Press, 1976), especially chap. 6.
20. See Posen, Sources of Military Doctrine; Van Evera, "Causes of War"; Snyder, "Civil-Military Relations"; and lack Snyder, "International Leverage on Soviet Domestic Change," World Politics 41 (October 1989), pp. 1-30. On the cult of the defensive, see John Mearsheimer, Liddell Hart and the Weight of History (Ithaca, N. Y.; Cornell University Press, 1988), pp. 107.
r
146   International Organization
argue about the sources of such misperceptions. Rather, we are satisfied to note that either of the above hypotheses is parsimonious and can easily be joined with the Jervis-Waltz international system theory to improve the accuracy of its predictions.
The element of misperception is not as foreign to Waltz's theory as one might first imagine. Indeed, Waltz claims that the basic problem of multi-polarity is "miscalculation by some or all of the great powers."2' In the simpler world of bipolarity, a superpower's responsibilities and vulnerabilities are easier to gauge, and egregious strategic miscalculations are therefore less likely. Of course, Waltz is referring here to random errors of perception and calculation that are inherent in the structural complexity and uncertainty of multipolar conditions; he is not referring to systematic perceptual biases due to cognitive or organizational quirks. "But in explaining the differences between the two multipolar outcomes, Waltz goes much further. For example, he writes that "the keenness of competition between the two camps" led to the chain gang effect in World War I. The "perception of a common threat brought Russia and France together," he adds. "If competing blocs are seen to be closely balanced, and if competition turns on important matters, then to let one's side down risks one's own destruction."21 Waltz's use of the term "perception" here may have been accidental, but we think not. In purely structural terms, the fate of Austro-Hungarian power in 1914 was not more "important" for the European military balance than was the fate of Czechoslovak power in 1938.23 There was no structural reason for the competition over it to be less "keen." Consequently, it is entirely appropriate for Waltz to use perceptual language, rather than structural language, in discussing France's and Russia's sense of a common threat.
It is our purpose to make explicit the military and perceptual factors that made competition more keen, alliances tighter, and East European crises seemingly more important in 1914 than in 1938. By doing this, we can account for the differences in multipolar alliance balancing behavior before World Wars 1 and II and thus rescue Waltz's theory from its predictive indeterminacy. Our proposed theoretical framework is summarized in Figure 1 and discussed in detail below.
21. Waltz, Theory of International Politics, p. 172.
22. Ibid., pp. 165-67.
23. For a detailed description of Czechoslovakia's crucial role in the European balance, see Williamson Murray, The Change in the European Balance of Power, 1938-1939 (Princeton, N.J.: Princeton University Press, 1984).
Alliance r.^erns 147
The security dilemma
Perceived defensive advantage (arising from civilian control or defensive lessons of history)
Buck-passing
Neither buck-passing nor
chain-ganging
Perceived offensive advantage (arising from military autonomy or offensive lessons of history)
Chain-ganging
Neither buck-passing nor
chain-ganging
i. Polarity, the security dilemma, and resulting alliance
Alliance strategies before World Wars I and II
Proposed explanation for the differing alliance patterns
The (wo world wars starkly illustrate the consequences of differing assessments of the relative strength of the offensive and the defensive. The strategic situation in these two cases was, in most respects, quite similar: Germany threatened to overturn the balance among the same four leading European powers by establishing its hegemony over Eastern Europe. But because the prevailing perception of the relative strength of offense and defense differed in the two cases, the strategic behavior of the powers in 1938-39 was the opposite of their behavior in 1914.
In 1914, the continental states adhered to essentially unconditional alliances, committing themselves to immediate offensives in full strength to aid their ally with little regard to the circumstances giving rise to the hostilities. In 1938-39, in contrast, the powers tried to pass the buck, luring others to bear the burden of stopping the rise of German hegemony. Stalin said in 1939 that the Soviet Union would not pull others' chestnuts out of the fire, but that is precisely what Russia had done in August 1914 through its pre-
\
1
148  International Organization
mature, ill-fated offensive into East Prussia, an offensive designed to draw German fire away from France during the battle of the Marne.24
The aggressors' strategies were also opposite. The originators of the Schlieffen Plan sought to overturn the balance in a single bold stroke, whereas Hitler sought to overturn it through the piecemeal conquest of isolated targets. Finally, the causes of the two wars were essentially opposite. World War I was largely the result of a spiral process in which alliance dynamics magnified the consequences of local disputes, turning them into global issues. World War 11^ in contrast, has often been considered a deterrence failure in which buck-passing diplomacy by the status quo powers encouraged expansionist powers to risk piecemeal aggression.25
Behind these differences in strategic behavior were differing assumptions about the efficacy of strategic offense and defense. In 1914, quick victories that would decisively overturn the military balance were generally thought to be quite feasible. To uphold the balance and to have an effect on the outcome of the fighting, policymakers believed that they had to conclude binding alliances in advance and throw their full weight into the battle at the outset.26 In the late 1930s, in contrast, policymakers and strategists who had lived through the trench warfare stalemates of 1914-18 believed that conquest was difficult and slow. Consequently, they thought that they could safely stand aside at the outset of a conflict, waiting to intervene only if and when the initial belligerents showed signs of having exhausted themselves.
We contend that given the constant factors of the multipolar checkerboard configuration of power and Germany's aggressive aims, varying perceptions of the offense-defense balance constitute a sufficient explanation for the differing alliance patterns: chain-ganging before World War I and buck-passing before World War II. As we go through the evidence in support of this interpretation, readers may want to keep in mind the following alternative explanations and our reasons for rejecting them.
24. Stalin's statement of 10 March 1939, cited in Adam Ulam, Expansion and Coexistence, 2d ed. (New York: Praeger, 1974), p. 263.
25. This distinction works only as a rough first cut. There were deterrence failure aspects to the 1914 diplomacy. Conversely, even firm, early deterrent threats might not have deterred Hitler's aggression. For a recent corrective along these iines, see Sean M. Lynn-Jones, "Detente and Deterrence: Anglo-German Relations, 1911-1914," International Security II (Fall 1986), pp. 121-50. Recent correctives, however, do not negate the main point. Even followers of Fritz Fischer accept that Germany did not want a world war but that il stumbled into it as a result of misguided attempts to ensure German security. For a subtle discussion of these points and a commentary on Fritz Fischer's German Aims in the First World War (New York: Norton, 1967) and related works, see Jack S. Levy, "The Role of Crisis Management in the Outbreak of World War I," paper presented at the annual meeting of the International Studies Association, London, 1989, especially pp. 15-16! ,
26. This argument about World War I,' set in a theoretical perspective, is made by Quester in Offense and Defense, bv Jervis in "Cnnnentiinn 1 InrW the- Co/.,.-,.. r»:i».---■» — j ■--
(
Alliance paLterns 149
Alternative explanations for the differing alliance patterns Franco-Soviet ideological differences. It is occasionally argued that a balancing alliance failed to form in the 1930s owing to the deep ideological distrust between France and the Soviet Union. This ignores the fact that republican France and autocratic Russia managed to form a tight alliance before World War I, despite their deep ideological differences.27
The creation of independent states in Eastern Europe. It is sometimes argued that the creation of independent states in Eastern Europe, especially Poland, hindered Franco-Soviet security cooperation by depriving the Soviet Union of a common frontier with Germany. But after September 1939, Stalin did have a common frontier with Hitler, and he still passed the buck.
The lesson that tight alliances cause wars. It might be argued that tight alliances were shunned owing to the apparent lesson of 1914 that tight alliances cause wars. Even though today's scholars may argue that it was the offensive strategies of 1914 that caused the tight alliances, interwar observers may not have understood this underlying cause.28 Thus, they may have passed the buck not because perceived defensive advantages made it attractive but, rather, because they wanted to avoid what they thought were reckless alliance strategies. We have uncovered little evidence in favor of this interpretation, but it was not a major focus of our research.
Cost minimization. It might be argued that states passed the buck in the 1930s simply because the experience of 1914-18 had radically increased their perceptions of the cost of fighting. This explanation overlaps with our own, since one of the reasons that war was seen as too costly to fight was the expectation that defense dominance would create a slow-moving war of appalling attrition. It differs from our argument, however, in that we see policymakers making essentially strategic decisions driven by the security interests of their states and not by an absolute horror at inordinate bloodshed. Stalin passed the buck even though bloodshed obviously did not trouble him. Moreover, as we argue below, France and Britain passed the buck less than the cost-minimization explanation would lead us to expect.
Germany's greater relative power. It might be argued that France and Britain adopted defensive buck-passing strategies in the 1930s because they were weaker relative to Germany at that time than they had been in 1914. But defensive buck-passing is the preferred strategy of weak or declining powers only when defense is perceived to have the advantage and thus offers
27. Waltz makes this point. See Theory of International Politics, p. 125.
19 v,„ rv»ra. "The Cult of the Offensive," p. 97. We are grateful to Randall Schweller for
150  International Organization
compensation for weakness. When offense is perceived to have the advaiH tage, weak powers compensate through surprise attack, and declining powers! compensate through preventive aggression. Logically, Germany's greater^ power should have made no difference, independent of assessments of of-] fensive and defensive advantage.
Case study: World War I
Germany: a strategy for decisive victory. The mainspring driving everyone's strategic calculations in 1914 was the Schlieffen Plan, Germany's strategy for a rapid knockout blow against France and a subsequent campaign against Russia. Whether German war aims were expansionism, self-defense, or "extended deterrence" of Russian pressure on Austria, the German General Staff argued that strategic circumstances dictated that any European war would have to be fought in this way.29
To say that this strategy was predicated on an erroneous belief in "offensive advantage" would be too simple. Schlieffen and his collaborators understood that increasing firepower enhanced the tactical advantage of the entrenched defender and that railroad mobility would help a country defend its own territory. However, he also argued that trenches could be outflanked, that railroads would allow a centrally positioned attacker to beat its opponents piecemeal, and that the slowness of Russian mobilization created a "window of opportunity" for implementing such a strategy. In this sense, Schlieffen saw an offensive advantage for Germany, which he generalized through the maxim that "if one is too weak to attack the whole, one should attack a section."30
German strategy was shaped even more strongly by fear of the offensive opportunities open to Germany's opponents. Schlieffen's mentor, the elder Moltke, had concluded that Germany could "extend deterrence" to Austria by mounting a limited attack on Russia in the East and maintaining a positional defense against France in the West. If France balked at attacking stout German defenses in the Saar, the war might be kept localized to Eastern Europe. Schlieffen, however, believed that France would rather easily overrun those defenses if Germany turned the bulk of its army eastward. Consequently, France would have to be disarmed before Germany could turn its attention to Russia.
29. For a discussion of the German strategy of 1914 in an alliance context, see Scott Sagan, ' 1914 Revisited: Allies, Offense, and Instability," International Security II (Fall 1986) pp 151-76. See also the dialogue between Scott Sagan and Jack Snyder in "Correspondence-'The Origins of Offense and the Consequences of Counterforce," International Security II (Winter 1986), pp. 187-98, On German strategy more generally, see Gerhard Ritter, The Schlieffen Plan (New York: Praeger, 1958); and Jack Snyder, Ideology of the Offensive: Military Decision--makmg and the Disasters of 1914 (Ithaca,lN.Y.: Cornell University Press, 1984), chaps. 4 and 5. ' e ;
(
Alliance patterns 151
In short, because Schlieffen and his successors greatly exaggerated France's offensive power and somewhat exaggerated their own, Germany adopted a war plan ensuring that a limited war in Eastern Europe would immediately escalate to a decisive showdown involving all of Europe's great powers. Moreover, the Schlieffen Plan increased Germany's strategic dependency on Austria by weakening German forces facing Russia early in a war. This meant that Germany had to run risks to keep Austria's strategic power intact, making the outbreak of an East European war all the more likely. In general, perceptions of offensive advantages and the adoption of offensive strategies led to unconditional alliances and aggressive balancing behavior.31
France: offensive advantages and support for Russia. When the defender enjoys a net strategic advantage, even a materially inferior power may feel secure. In the years before World War I, however, French authorities exaggerated the advantages of the attacker and thus concluded that a tight alliance with Russia was needed to offset the threat posed by Germany's larger population, army, and material base.
After the 1911 Moroccan crisis, in which Russia had offered only tepid support of France in its confrontation with Germany, the French resolved to tighten their alliance with Russia at all costs.32 Since active Russian help was seen as essential in parrying the danger from a German offensive, France concluded that the danger of being entrapped in a Russo-German dispute over the Balkans was less worrisome than the danger of being abandoned by Russia in some new Franco—German crisis.33 Indeed, some French officials concluded that it would be desirable for a war to arise over a Balkan issue, since that would ensure Russia's active participation. France was willing to balance aggressively in order to preclude Russian passivity.
Poincare\ the French President elected in the nationalist upsurge after the Moroccan crisis, was consequently more willing than his predecessors to support Russian efforts to form an alliance of the small Balkan powers against Austria. As it turned out, the Balkan states themselves were more interested in liberating European Turkey. To deter Serbia from excessive territorial aggrandizement as a consequence of the victory over Turkey, Austria mobilized part of its army. As a result, throughout November 1912, Russia and France confronted difficult decisions about what military measures to take ■n response to Austria's partial mobilization and what to do if Austria attacked Serbia. Though it would be an exaggeration to say that the French
31. For additional analysis, see Van Evera, "The Cult of the Offensive," especially pp.
32. For a perceptive analysis, see A. J. P. Taylor, The Struggle for Mastery in Europe (London: Oxford University Press, 1971), pp. 468 and 486.
33. For an analysis of the entrapment-abandonment trade-off in 1914 and in general, see
{
152   International Organization
actively sought war on this occasion, they seem to have been more keen forj the Russians to take military measures than the Russians were themselves.34 The Russians did take some precautionary steps, delaying the discharge of a year's cohort of draftees and mobilizing a light security force on the; Austro-Hungarian frontier.33 For the most part, however, the Russians did not think that the situation was especially dangerous and sought to avoid provoking a needless escalation. They believed that Germany was restraining Austria and that Austria's partial mobilization had made a full mobilization against Russia more complicated rather than easier.36 Russian caution was also based on an emergency review of Russia's material preparedness for war, which concluded that stocks were so low that Russia could not fight.37 According to A. J. P. Taylor, the Russians needed to find a scapegoat for
__their own timidity and "tried again and again to make Poincare say that he
would not support them if they went to war for the sake of Serbia, but Poincare refused to be caught,"38 telling the Russian ambassador that "if Russia goes to war, France will also. "39 Even more amazing was an interview in which French Defense Minister Alexandre Millerand took to task the Russian military attache in Paris for his government's weak response to Austrian military measures. At issue was "the hegemony of Austria throughout the entire Balkan peninsula," Millerand told him. If Russia fails to pick
34. Keiger puis this in perspective, arguing that Poincare was not bell/cose. See John Keiger, France and the Origins of the First World War (New York: St. Martin's Press, 1983).
35. For what is by far the clearest account of this misunderstood episode, see V. I. Bovykin, Jz istorii vozniknoveniia pervoi mirovoi voiny: Otnosheniia Rossii i Frantsii v 1912-1914 gg. (From the history of the origins of the First World War: Relations between Russia and France in 1912-1914) (Moscow: Moskovskii Universitet, 1961), pp. 151-53. See also E. C. Helmreich, The Diplomacy of the Balkan Wars (Cambridge, Mass.: Harvard University Press, 1938), p. 216; Louis Garros, "En marge de l'alliance franco-russe, 1902-1914" (A footnote to the Franco-Russian alliance, 1902-1914), Revue historique de i'armie, June 1950, p. 33; Frank M. Laney, "The Military Implementation of the Franco-Russian Alliance, 1890-1914," Ph.D. diss.. University of Virginia, 1954, p, 390; and Samuel Williamson, "Military Dimensions of Habsburg-Romanov Relations During the Era of the Balkan Wars," in Bela Kiraly and Dimitrije Djordjevic, eds., East Central European Society and the Balkan Wars (Boulder, Colo.: Social Science Monographs, 1986), pp. 317-37.
36. See Laney, "The Military Implementation of the Franco-Russian Alliance," p. 402; dispatch by General Marquis de Laguiche, the French military aitache in St. Petersburg, file 7NI478 (6/19 December 1912, 27 November/4 December 1912, and 30 November/13 December 1912) at the French military archive, Chateau de Vincennes; I. V. Bestuzhev, "Bor'ba v Rossii po voprosam vneshnei politiki nakanune pervoi mirovoi voiny, 1910-1914 gg." (The struggle in Russia on questions of foreign policy on the eve of the First World War, I910-I9I4), Istoricheskie zapiski, vol. 75, 1965, pp. 63 ff.; Garros, "En marge de l'alliance franco-russe," p. 36; Documents diplomatiquesfrancais {DDF), series 3, vol.V, no. 52, p. 65; and "Podgotovka pervoi mirovoi voiny" (Preparations for the First World War), Voenno-istoricheskii zhurnal, no. 3, 1939, pp. 132-33.
37. See Bovykin, h istorii vozniknoveniia pervoimirovoi voiny, p. 136; and A. A. Manikovskii, Boevoe snabzhenie russkoi armii, 1914-1918 gg. (Military supply in the Russian army, 1914-1918) (Moscow: Voennyi Redaktsionnyi Sovet, 1923). ,
38. This is Taylor's apt characterization of the situation in Struggle for Mastery in Europe,' p. 494. (
(
Alliance patterns 153
up the challenge, he said, "it is not our fault: we are ready."40 Similarly, the Russian ambassador reported that French generals saw great advantages in fighting a war in which Austria's strength would be dissipated in a Balkan campaign.41
In short, far from buck-passing in the crises of 1912 and 1914, France seems to have been at least as eager to stand up for Russian interests as were the Russians themselves. This contrasts sharply with the extremely tepid support that France offered Russia in the 1909 showdown over Bosnia-Herzegovina.42 The change in French calculations was primarily due to their belief after the 1911 Moroccan crisis that war between France and Germany was close to inevitable. Thus, abandonment by Russia became a greater risk than entrapment in Russia's quarrels.
Another factor promoting the tightening of the alliance was the rise of the doctrine of offensive a outrance, which was accompanied by the presumption that a decisive victory or defeat would be achieved with great speed on the Franco-German front.43 For this reason, the French felt more dependent on rapid aid from Russia at the earliest possible moment, and they pressed for a premature Russian offensive against East Prussia and offered to pay for the railroads needed to support this maneuver. Greater French faith in their own offensive prospects may also have increased the attractiveness of fighting Germany, especially under circumstances in which Austria's forces would be diverted to the Balkans.44 In this way, the belief in offensive advantage promoted aggressive balancing behavior.
Russia: short war expectations and a commitment to France. The growing belief that the clash of the French and German offensives would lead to an extremely rapid decision in the West also led to a tightening of Russia's commitment to France. This is an especially interesting case because it helps to refute an alternative explanation for balancing and buck-passing choices— namely, that states seek balancing alliances when they believe that they are the next target on the aggressor's list, but they try to pass the buck when they believe that others will be attacked first.
As late as. 1910, the Russian General Staff believed that Germany would
40. Millerand, cited in A. A. Ignat'ev, Piatdesiat let v strain (Fifty years of service), vol. 1 (Moscow: Khudozheslvennaia Literalura, 1959), p. 506.
41. Unpublished archival documents, cited in Bovykin, Jz istorii vozniknoveniia pervoi mirovoi voiny, pp. 137 and 146 ff.
42. DDF, series 2, vol. XII, nos. 51, 55, 74, 86 , 87, 90, 100, 113, and 266.
43. For example, in March 1910, Lt. Colonel Pelld, the French attache in Berlin, wrote to General Jean Jules Brun, the Minister of War, that "both on the German side and on the French, the bulk of the active forces of both countries are planned for deployment in first-line armies [near the frontiers]. The victory or defeat of these armies of the first line will very probably decide the outcome of the campaign" by the twentieth or thirtieth day after mobilization. See DDF, series 2, vol. XII, no. 453, p. 691.
44. For some pertinent comments on this matter, see Taylor, Struggle for Mastery in Europe,
154
^tional Organization
Allianci
.cms 155
direct its main offensive toward Russia if war broke out over a Balta dispute.45 Despite seeing themselves as the most immediately threatens power, defense-minded Russian staff officers resisted committing themselvei to a tighter French alliance. Though they thought Germany was more likel| to attack eastward, they thought that Germany might send the bulk of its' forces against France if war arose over some bilateral Franco-German disa pute. In that event, they believed that it would be unnecessary and ruinoiif to agree to a hasty Russian offensive into East Prussia. The Russian attache in Berlin told his French colleague that Russia's offensive could succeed only after its mobilization was completed and thus would have to lag two weeks behind that of France. Noting that "Napoleon's principle was to act with all his forces united," the Russian insisted that "we should not risk compromising this success by taking the offensive prematurely." The French
_attache rebutted that "in the case of two allied armies, such as ours, the
true application of the principle would consist not in waiting for the complete concentration of Russian forces, but rather in acting together at the moment when the French and Russian armies could produce simultaneously the maximum effect."46
Soon the Russians began to accept the logic of this argument. Even though the Russians came to believe that Germany would almost certainly attack first in the West, this did not reduce their dependency on the French alliance. On the contrary, they now desired a tighter alliance on French terms because they feared that France would be defeated without it. After about 1911, the Russians increasingly accepted the view that the collision of the offensive d outrance and the simultaneous German offensive would lead to a rapid decision, one way or the other. Russian planning documents now began to express fears of an immediate rout of the French, leading to a separate peace that would give Germany a free hand in the East.47 As one military official argued, Russia should mount an early offensive "to prevent Germany from finishing with France or weakening her in order to have the possibility of redeploying forces against us."48
In fact, in August 1914, Russia did invade East Prussia hastily, not waiting for the full mobilization of the Russian army or even of the supply trains of the attacking units. One result was that the Germans transferred two army corps to East Prussia from the second echelon of the offensive in the West,
45. See DDF, series 2, vol. XII, no. 399, p. 611; and Snyder, Ideology of the Offensive, chaps. 6 and 7.
46. Conversation between Colonel Mikhelsson and Lt. Colonel PelI6, reported by Pelle' to General Brun in March 1910 and cited in DDF, series 2, vol. XII, no. 453, p. 695, and no. 467, : p. 717.
47. See Valentin Alekseevich Emets, "O roli russkoi armii v pervyi period mirovoi voiny, 1914-1918 gg." (On the role of the Russian army in the first period of the World War, 1914-1918), Istoricheskie zapiski, vol. 77, 1965, p. 64. v'
48. General N. A. Kliuev, chief of staff of the Warsaw military district, cited by Emets in ibid., p. 64. *
perhaps marginally easing France's burden at the battle of the Marne. Another result was the encirclement and destruction of a Russian army of one hundred thousand men at the battle of Tannenberg.
This disaster has often been explained by the assertion that Russian strategy was held in thrall by French financial hegemony. Recent Soviet archival scholarship shows, however, that changes in Russian strategy were not extracted as the price for railroad loans. Rather, the headlong rush to Tannenberg was caused by Russian fears that the clash of offensive strategies would lead to a rapid decision in the West, forcing Russia to act hastily in order to have any chance of influencing the outcome.49
In short, Germany's decision to attack France first could have allowed Russia to pass the buck, waiting on developments in the West before committing forces to the fray. Indeed, some Russians advocated a strategy of initially standing aside in order to exploit German weakness after a slow, bloody, Pyrrhic victory over France. Instead, Russia chose to balance aggressively and unconditionally out of the fear that the German offensive might be quick, relatively bloodless, and decisive for the European balance of power.
Britain: defensive advantages. Though some British strategists shared Russia's fears of a lightning French defeat, Britain had the English Channel, the British fleet, and the resources of the British Empire to buffer it from the consequences of a shift in the continental military balance. Moreso than Russia, Britain could afford to wait on developments, seeking accommodation with Germany and Austria up to the very end and limiting Britain's initial liability to an expeditionary force of some four divisions.50 Moreover, it was not unreasonable for the British to believe that French and Russian power would suffice to contain German expansionism. Consequently, it made sense for Britain to limit its involvement in the attritional campaign in order to emerge from the war as the strongest, least damaged power.
During this period, as in the late 1930s, British policymakers sought to contribute "the smallest amount of money and the smallest number of men with which we may hope, some day, to win the war," through a blockade of the German economy and free riding on French casualties.51 Lloyd George, for example, anticipated in February 1912 that the German offensive in France would bog down in a stalemate and that modest British efforts would
49. Valentin Alekseevich Emets, Ocherki vneshnei polhiki rossii v period pervoi mirovoi voiny: Vzaimoolnosheniia rossii s soiuznikami po voprosam vedeniia voiny (Sketches of the foreign policy of Russia during the First World War: Relations of Russia with its allies on questions of the conduct of the war) (Moscow: Nauka, 1977), pp. 47-52.
50. Michael Howard, The Continental Commitment (London: Temple Smith, 1972).
51. This is Field Marshal William Robertson's apt characterization of the views of Walter Runciman and Reginald McKenna, cited by David French in British Strategy and War Aims, ]9H~10r* n^ndon: Allen & Unwin, 1986), p. 247.
(
156  International Organization
suffice to maintain French morale.52 Meanwhile, as David French has noted, \ Britain's own financial strength and expanded military forces would be hus-; banded "so that Britain would have the strongest army of all the belligerents^ when the time came to make peace."53 This approach changed decisively^ only in 1916, when French resources and morale came near exhaustion and; when it seemed likely that Russia would be knocked out of the war if Britain ? continued its strategy of limited liability.54
Thus, in 1914, Britain was the outlier, the country with the most invulnerable defensive position and the country with the most limited, conditional commitment to its allies. While Britain did not entirely pass the buck, it did take advantage of its protected position to pass costs and risks to France and Russia until their collapse seemed imminent. — For each of the major powers before 1914, there was a close connection between the perception of offensive advantage and the adoption of a strategy of aggressive, unconditional balancing. France and Russia tightened their alliance when French strategy became more offensive and when the expectation of a rapid and decisive victory, one way or the other, became more prevalent. Fearing a collapse of the western front, Russia accepted major self-sacrifices to bail France out, despite the temptation offered by the Schlieffen Plan to ride free on French efforts. Britain, in contrast, exploited its special defensive advantages to limit its liability until the strategic situation was clarified in the opening engagements.
Case study: World War II
Germany: Hitler's strategy of piecemeal expansion. Hitler's strategy in the late 1930s was the opposite of Schlieffen's earlier strategy. Instead of trying to overturn the European balance of power in one bold stroke, Hitler sought to accomplish this in a series of lightning campaigns against diplomatically isolated victims. Especially important in this strategy was the • capture of Czechoslovakia's thirty-four divisions and its heavy industrial, complex, the Skoda works. Through this piecemeal aggression, Hitler had by 1941 achieved an industrial and raw materials base that would allow him to prosecute a long war against the Soviet Union, despite the British blockade.55
A sufficient explanation for the German adoption of this piecemeal strategy of expansion is the buck-passing diplomacy of the other powers. Perhaps if Hitler had been tightly encircled by a Franco-Soviet alliance, he would have
52. French, ibid., p. 3; for related evidence, see also pp. xii, 106, 118, and 245-46.
53. This is French's characterization of Lord Kitchener's views, cited in ibid., pp. 200-201.^
54. French, ibid., pp. xii, 119, and 201., i
55. See Murray, Change in the European Balance of Power. On the tailoring of German % .:i:.„„, ~,~.,k;i;n, w chnrt ramMi.ru nnri rimlnmatie intimidation, see Posen. Sources ofM
Alliance
>ns 157
sought a Schlieffen-type strategic solution. But the buck-passing of his opponents meant that the easier, piecemeal route was available, so Hitler took it. What was important in this case was not so much German perceptions of the relative advantages of offense and defense but, rather, the perceptions of Germany's opponents on that dimension. Hitler himself was usually optimistic about offensive schemes, though even he expected General Heinz Guderian's blitz through the Ardennes in May 1940 to yield only a limited victory and not the utter collapse of France. Many German generals, steeped in the lessons of World War I, were even more pessimistic about the prospects for armored blitzkrieg breakthroughs.56 But because those same lessons led Germany's opponents to adopt strategies of passive buck-passing, the Germans never had to face the hard question of whether offense was easy enough to defeat all of Europe in a single campaign, the task that Schlieffen had confronted. Instead, Hitler had only to consider whether offense was feasible enough to lay low one enemy at a time.
The Soviet Union: Stalin's strategy of entrapment and buck-passing. Two key assumptions shaped Stalin's alliance diplomacy. The first was that France and Britain could hold out for a long time against German attacks, in part owing to the advantages of the defender, even if the Soviet Union offered them no assistance. Even if Germany did defeat France, a victory won through a grueling attritional campaign would be pyrrhic, leaving the free-riding Soviet Union in a strengthened position vis-a-vis the other powers. Khrushchev later reported that Stalin had been not only dismayed but also truly surprised by the collapse of France in 1940. "Couldn't they put up any resistance at all?" complained the stunned dictator to his Politburo colleagues.57
Stalin's dismay and surprise were due in part to his overrating of the strength of France and Britain. In his March 1939 speech warning that he would not pull others' chestnuts out of the fire for them, Stalin argued that 'the non-aggressive, democratic states are unquestionably stronger. . . both economically and militarily" than Germany and could therefore resist Germany on their own.58 In part, however, Stalin's reactions were also due to his overrating of the relative strength of the defense. General D. G. Pavlov, whom one historian ironically labels "Stalin's Guderian," returned from the Spanish Civil War and convinced Stalin that massed-armor blitzkrieg offen-
56. See John Mearsheimer, Conventional Deterrence (Ithaca, N.Y.: Cornell University Press, 1983), chap. 4.
57. Nikita Khrushchev, Khrushchev Remembers, vol. 1 (Boston: Little, Brown, 1970), p. ,34; see also p. 129. According to Deutscher, "the major premise of Stalin's policy and his Major blunder" were that "he expected Britain and France to hold their ground for a long «me." See Isaac Deutscher, Stalin (New York: Oxford University Press, 1949), p. 441.
58. Stalin, cited in John Erickson, The Soviet High Command (New York: St. Martin's Press,
(
158   International Organization
sives were infeasible.59 Consequently, Stalin overrated not only the defens strength of France but also that of Poland.60 .% The second assumption behind Stalin's diplomacy was that Germany itiij get embroiled with the West first if the Soviet Union adopted a stance was militarily strong but diplomatically nonprovocative. This view was oi«. expressed in terms of the Leninist theory of interimperialist contradiction! which would arise from the uneven growth of the capitalist powers and tfL consequent need to fight for a redivision of the colonial spoils. As early % 1925, Stalin held the view that if war comes "we shall have to take actions but we shall be the last to do so in order to throw the decisive weight intolf the scales."6' \|ff Proceeding from the two assumptions of defensive advantage and inter-'g-imperialist contradictions, Stalin maneuvered to embroil Germany with the West and to pass to France the costs of checking German revisionism. IdJ| this, Stalin was greatly aided by the fact that France had a common border^ with Germany and alliance commitments to Czechoslovakia and Poland,^ whereas the Soviet Union did not. At the time of the Munich crisis, forM example, Soviet diplomacy tried to lure France to honor the Czech alliance^ by promising to help Czechoslovakia if France did too. Those who debatef whether Stalin's support for "collective security" was sincere in this in-:| stance miss the point. If France had agreed to these conditions, a German $ attack on Czechoslovakia would have triggered a major engagement of Frenchr and German forces at the Siegfried line. Meanwhile, even if Rumania allowed the Soviets to send some troops into Slovakia across Rumania's limited rail : connections, neutral Poland would have prevented German and Soviet forces .'■ from becoming fully engaged. In short, Stalin was pursuing a strategy of limited liability in 1938 as a means to lure France and Germany into an attritional campaign that would debilitate both of them.62 Of course, if Germany conquered Poland, Stalin would lose his buffer,
59. See John Erickson, The Road to Stalingrad, vol. 1 (New York: Harper & Row, 1975), pp. 8 and 26; and Erickson, Soviet High Command, p. 537.
60. See Deulscher, Stalin, p. 437. This genera! predisposition to underestimate the feasibility of blitzkrieg may even have lasted past May 1940 and contributed to the false hope that Hitler would not attack in June 1941. Politburo member Andrei Zhdanov believed in 1940 that "Germany is incapable of fighting on two fronts," and even after the fall of France, he considered that Germany was too "bogged down" by the war with England to attack the Soviet Union. Foreign Minister Vyacheslav Molotov said in June 1941 that "only a fool would attack us." See Gavriel Ra'anan, International Policy Formation in the USSR (Hamden, Conn.: Archon, 1983), p. 18. On some new revelations along the same lines, see Y. Perechenev, "Ten Volumes About the War," Moscow News. no. 38, 20 September 1987, p. 10, citing K. M. Simonov, "Zametkt k biografii G. K. Zhukova" (Notes for the biography of G. K- Zhukov), Voenno-isloricheskii zhurnal, no. 9, 1987, pp. 49-51. We are grateful to Cindy Roberts for this citation.
61. Stalin, cited in Louis Fischer, Stalin's Road from Peace to War (New York: Harper & Row. 1969), p. 304. ,
62. For insightful analyses of the situation, see Telford Taylor, Munich (Garden City, N.Y.:
(
Alliance ^„^erns 159
making a buck-passing strategy riskier and more difficult to arrange. France and Britain made Stalin's task easier, however, by guaranteeing their support for Poland after Hitler occupied Czechoslovakia. This greatly increased the likelihood that Hitler's next target after Poland would be France rather than the Soviet Union.
Thus, by the end of 1939, the Soviet Union was in a position strikingly analogous to that of Russia at the end of 1913. In the long run, a Russo-German war was likely to occur, but it would almost certainly be preceded by a Franco-German campaign. In both instances, Russia had an incentive to delay the confrontation for two or three years, when its military strength relative to Germany's would peak. Moreover, in both instances, the current military balance favored Germany over France only slightly.
Despite these similar circumstances, imperial Russia chose a strategy of aggressive balancing, whereas Stalin chose buck-passing. This was not due to the ideological antipathy between Soviet Russia and bourgeois France, which was only a little greater than that between reactionary Russia and bourgeois France. Rather, the available evidence suggests that it was due to Stalin's stronger faith in the power of the defense.63 If Stalin had understood that Germany could conquer France in a month, he probably would have acted just as Russia had in August 1914, mounting a simultaneous offensive regardless of the insufficiently prepared condition of his forces.
France: defensive advantages and buck-passing. French strategy in 1938-39 was powerfully influenced by the desire to pass the costs of France's defense to Britain and by the perception, based on French experiences in World War I, that offense was much more difficult than defense. However, the French inclination to pass the buck was not all-consuming. In 1939, France might have gambled on offering Hitler a free hand in the East, passing all the costs of French defense to Poland and the Soviet Union. Instead, France agreed 'o join in a guarantee of Poland, thereby passing only some of the costs of French defense to Britain. Likewise, French confidence in the holding power of the defense was not absolute. If it had been, France could have extended the Maginot line to the English Channel and remained indifferent to the alliance possibilities with Britain, Poland, and the Low Countries.
In fact, French strategy was more complex. The French believed that they would lose if they fought a long war alone against Germany, but they would win a defensive war fought with the assistance of a fully mobilized Britain. French strategy, including both its balancing and buck-passing aspects, was aimed at achieving this end.64
This perspective explains the most puzzling aspect of French strategic behavior: France's refusal to fight on extremely favorable terms in Septem-
63. For this evidence, see footnotes 57 and 59-62 above.
64. In Sources of Military Doctrine, chap. 4, Posen describes the French military strategy
----... „f fiohtino m ihe British.
(
160   International Organization
ber 1938 and its agreement to fight on extremely unfavorable terms a yearj later. At the time of the Munich crisis, the French potentially had strategic mastery in Europe. To overcome Czechoslovakia's thirty-four crack divisions and formidable frontier fortifications, Hitler planned to use—and would have had to use—the bulk of his army and air force. This would have left France with a seven-to-one advantage in the West. At this time, the Siegfried line (or Westwall) was only 5 percent complete, with recently poured con^l crete that had not yet set.
By September 1939, the Siegfried line consisted of 11,283 bunkers, in contrast to the 517 of a year before. The German army available for action in the West during the Polish campaign had thirty-five divisions, seven of which were first line, as opposed to only eight divisions in total in 1938. In - light of the deficiencies of the French army in offensive operations, these force balances suggest that even an all-out assault on the Siegfried line in September 1939 would not have saved Poland. The weak probes actually carried out by the French were probably the only offensives that were possible under the circumstances.65
The French seem to have made no gain, therefore, from declaring war as a result of Hitler's invasion of Poland. They succeeded only in ensuring that they, and not Russia, would be Hitler's next target. Thus, it is ironic that some historians have branded Foreign Minister Georges Bonnet a traitor for allegedly offering Hitler a free hand in the East.66 Under the circumstances, luring Hitler eastward would seem to have been a vastly superior course to guaranteeing Poland.
The paradoxical reversal in French behavior between September 1938 and September 1939 was due to the change in Britain's attitude. In 1938, Britain offered no help in a war to save Czechoslovakia. By 1939, however, Britain's guarantee to Poland and its decision to increase the size of the British army gave the French reason to expect that if they joined in the British guarantee, Britain would be prepared to deploy a significant "expeditionary force in France about six months after the outbreak of war. During this interval, France would be protected by the time needed to occupy Poland and by the winter. Thus, British aid plus defensive advantage would suffice to protect France from Germany at a tolerable price in French lives. In this sense, France's guarantee to Poland was part of a buck-passing strategy predicated on the expectation of defensive advantage.
French perceptions of a qualified defensive advantage played an important role at several stages of decision making. Often, however, assessments of defensive advantage seem to have been less a cause of buck-passing diplomacy than a manipulated rationalization of it. For example, during the Mu
65. Murray, Change in the European Balance of Power, p. 348.
66. For a review of the complexities o)f the evidence on this, see Anthony Adamthwaite,
(
Alliance patterns 161
'f»,/   iritn it___l_
nich crisis, when French Chief of Staff Maurice Gamelin briefed French politicians about the scenario of France attacking the Westwall, he portrayed a bloody campaign with no possibility of rapid results—in short, "a modernized Somme."67 However, when Gamelin briefed British Prime Minister Neville Chamberlain in September 1938, he argued that a joint Franco-British offensive would surely be successful, owing to the incompletion of German fortifications as well as their lack of trained reservists and raw materials. Alexander Cadogan, a British participant, astutely observed that what the French really had in mind was a "squib offensive (to bring us in) and then retirement on Maginot Line to wait (6 months) for our Kitchener armies."68 It also smacks of a manipulated double standard that in Gamelin's conversations with French politicians, he depicted the Westwall as a tough nut to crack while predicting a German walkover of the elaborate Czech fortification system.69
If the French selectively overrated German defenses, they did not greatly overrate Poland's ability to defend itself. Gamelin understood that Poland would be destroyed, with or without a Franco-British declaration of war. He argued, however, that it would buy France six months, which it did, during which British forces would start to arrive.70
Likewise, the French did not greatly overrate their ability to defend themselves behind the Maginot line, even if the line were extended to the sea. Prime Minister Daladier believed that "France could not make war alone against Germany," echoing the views of the French Chiefs of Staff that "France cannot long withstand effectives three times as numerous."7' Rather than complete the Maginot line, which might encourage Britain to ride free on French defense, French leaders thought it better to leave the invasion routes through Belgium open, thus luring Britain into a joint defense of the Low Countries.72 But once Britain was entrapped, the French seem to have been overconfident in the efficacy of their defenses.
This selective and partial overrating of the efficacy of defense strongly implies that the desire for buck-passing was driving the estimates of the relative strength of offense and defense, rather than the reverse. Alleged offensive advantages, such as the ease with which the Germans could bomb Paris, were also invoked whenever they served to justify taking no action without British assistance. This raises the question of whether the fear of the high costs of fighting might have been the ultimate force shaping French strategy and not perceptions of defensive advantage per se. This simplication fails, however, to explain the Polish guarantee. If France had been single-
67. Gamelin, cited in ibid., p. 232.
*8. Cadogan, cited in ibid., p. 232. Kitchener had organized the expansion of the British ^y for deployment in France in World War I.
69. Gamelin, cited in ibid., pp. 232-34.
70. Gamelin's opinion of 23 August 1939, cited in ibid.j. 340. See also ibid., p. 311.
71. Daladier, ciled in ibid., pp. 226 and 230.
162   International Organization
mindedly bent on minimizing combat casualties, the best strategy wouldl have been to offer no guarantee to Poland, hoping that Hitler's ultimate aiin*: was the Ukraine and not France. This might have been risky, jeopardizing Britain's continental commitment if Hitler were to strike France first, but was not an unreasonable gamble, since France's own estimates were tha Hitler's main goals lay in the East.73
Instead of gambling on passing the whole costs of the war to the Soviets!! France took what it thought was the safer but more costly course of passing^ part of the costs of fighting to Britain. It was because France overrated the g chances of a successful defense with Britain's help that this policy looked | superior. Though perceptions of defensive advantage were manipulated in the service of a buck-passing diplomacy, there was also at bottom a real '■■ perception that France and Britain together could stalemate Germany, as they had in 1914-18, aided by the inherent advantages of the defender.74
Arguably, this left France with the worst of all possible strategies. If France had had more faith in the holding power of the defense, the Maginot;;. line might have been extended to the Channel and the Polish guarantee would ^; have been shunned, even at the loss of British support. That is, France would have tried to pass the buck entirely to Russia, rather than partially $ to Britain, while preparing to fight successfully on its own if that plan misfired. If, on the other hand, France had had more confidence in offensive v operations, supporting Czechoslovakia in 1938 might have looked more attractive. As it was, the British expeditionary force amounted to only four divisions by May 1940, a measure of the illusory success of French buck-passing.75
Britain: a strategy of limited liability. Like France, Britain did not count on riding scot-free on inherent defensive advantages and the balancing efforts of other powers. Nonetheless, Britain did count heavily on such advantages to allow it to contribute a minimum to upholding the balance of power as well as the luxury of waiting until the last minute to see what that minimum would be. In short, Britain pursued a strategy of limited liability, based on the defensive advantage provided by the English Channel and on the expectation that a new European war would be a slow-moving rerun of the last one. Chamberlain, both before and after September 1939, thought that French defenses were so strong that Hitler might not even attack them, that Germany would be worn down by a long blockade, and that Hitler's only
73. Adamthwaite, France and the Coming of the Second World War, pp. 252 and 274 ;^°r ev,dence    support of this interpretation, see Eleanor M. Gates, End of ■
ifomlfteiuS^IK'sT-S.^"^V JWSM9(Berilelcy: University of Cal,
.J.5: Brian Bond. British Military Policy Between the Two World Wars (Oxford; Clarendon.
(
Alliance patterns 163
offensive option would be to try to grab Rumania's oil to help him endure the Sitzkrieg.''6
The Munich crisis is easily explainable in these terms. The specter of a defensive attritional land war, coupled with the fear of a costly air war, gave Britain a strong incentive to make sure that war was absolutely unavoidable before deciding to fight. At the same time, confidence in the Maginot line and the extra cushion provided by Britain's off-shore position gave the British the luxury of waiting until the evidence of Hitler's intentions was all in. As Cadogan remarked after the Munich crisis, "I know that it is said that Mitteleuropa will turn round and rend us. But many things may happen before that."77
The puzzling Polish guarantee also seems more sensible when viewed in the light of a strategy of limited liability, anchored on France's apparently formidable defensive military power. Like the French, the British Chiefs of Staff had few illusions that they could take any action to prevent the destruction of Poland. Only the Soviets could stop Hitler in the East, they believed, and then only if Germany attacked the Soviet Union, which was devoid of offensive power.78 They did believe, however, that Poland might take months to conquer, exacting attrition on German forces and buying time for preparing a defense of the Low Countries.75 Of course, Polish efforts might have achieved this result without a Franco-British guarantee. Nevertheless, as Brian Bond indicates, "Halifax and Chamberlain feared that the Poles were about to do a deal with Germany which would demolish the hope of a second front in the east."80
Another important aim of the guarantee was its effect on France. The British tended to rate the Maginot line and, more generally, France's defensive posture quite highly.81 The secretary of state for war, Leslie Hore-Belisha, even argued that Britain should announce irrevocably that no expeditionary force would go to the continent under any circumstances, for then France would extend the Maginot line to the sea and give up the game
76. Chamberlain's letters to family members, cited in Maurice Cowling, The Impact of Hitler (London: Cambridge University Press, 1975), pp. 355-57. See also Bond, British Military Power, P- 253; and N. H. Gibbs. Grand Strategy, vol. 1, Rearmament Policy (London: Her Majesty's Stationery Office, 1976), pp. 637-38. Bond and Gibbs also note, however, that the notions of defensive advantage held by Chamberlain and Leslie Hore-Belisha were not universally held within British official circles. We are grateful to Randall Schweller for help on this point.
77. David Dilks, ed., The Diaries of Sir Alexander Cadogan, 1938-1945 (London: Casscll, 1971), p. 119.
78. Murray, Change in the European Balance of Power, p, 298.
79. Simon Newman, March 1939: The British Guarantee to Poland (Oxford: Clarendon, 1976), pp. 155-56.
80. Bond, British Military Policy, p. 306.
81. See, for example, the views of Sir John Simon, Chancellor of the Exchequer, cited in Murray, Change in the European Balance of Power, p. 274; and the views of the Chiefs of Staff, cited in Newman, March 1939, p. 139. See also Adamthwaite, France and the Coming
164   Internauonal Organization
if
of luring Britain to help defend the invasion corridor across Belgium.82 This would make both Britain and France more secure—and at France's expense."' The more typical view, however, was that it would harm British security: if France mounted a defense on the French borders. Instead, Britain had to "J induce France to mount a forward defense of the Low Countries to prevent Germany from using them as a base for air attacks on Britain.83 To make sure that France was willing and able to mount such a forward defense, Britain considered it worthwhile to agree to a limited British commitment to the continent.
Beyond this, after the collapse of Czechoslovak power, there was even a fear that France proper might fold under German pressure. The British Chiefs of Staff, for example, worried that "France might give up the unequal struggle unless supported with the assurance that we should assist them to the utmost."84 Nonetheless, Britain still held to the assumption of defensive advantage, which implied that a small British force with a primarily moral impact would probably suffice to stiffen French resistance. Defensive advantage would permit Britain the luxury of limiting its initial liability, awaiting further developments to see whether a greater contribution was needed.
Robert Vansittart captured the essence of British thinking:
We are proceeding on two assumptions both of which I am sure will be falsified: first that France can hold out on two or perhaps three frontiers [German, Italian, and Spanish] with no expeditionary force from us. . . . Secondly we are assuming that the war, if it comes, will be a long one and we must therefore lay great stress on conserving our [financial] staying power.85
Though Vansittart offered this characterization in early 1938, it still captures Britain's basic thinking even after the Polish guarantee. The strategy was still one of limited liability based on the exploitation of defensive advantage and the balancingefforts of others. After April 1939, however, there was a mild upward adjustment in the estimate of the minimal British liability needed to ensure that those balancing forces would operate successfully. In this way, Britain hoped to strike an optimal trade-off between the benefits of riding free and the benefits of balancing aggressively, guaranteeing British security at a minimal cost. If in retrospect the trade-off appears less than optimal to some, that is because the expectation of defensive advantage was
82. Adamthwaite, France and the Coming of the Second World War, p. 71.
83. See Murray, Change in the European Balance of Power, pp. 276-77; Dilks, Diaries of Sir Alexander Ccdogan, p. 139; and Bond, British Military Policy, p. 297.
84. British Chiefs of Staff, cited in Adamthwaite, France and the Coming of the Second World War, p. 253. See also Dilks, Diaries of Sir Alexander Cadogan, p. 166. In Change in the European Balance of Power, p. 71, Murray argues, mostly by inference, that the British change on the continental commitment in 1939 was due to the loss of Czechoslovakia's thirty-four divisions from the European military (equation. In British Military Policy, p. 296, Bond notes that Britain's military attache' in Pari? took this view.
Alliance litems 165
too sanguine and not because British deductions from that assumption were faulty.
Finally, the air power element in British liability calculations merits additional attention because it is especially relevant to the choice between strategic deterrence and strategic defense. One of the reasons that Chamberlain appeased Hitler at Munich was his exaggerated estimate of German strategic bombing capabilities and his fear that Britain's own retaliatory capability would not deter attacks on British cities. After the Munich crisis, Chamberlain pushed for a reorientation of British air power expenditures from bombers to fighters. Believing these efforts to be successful, he concluded by mid-1939 that a German air attack on Britain would probably fail. This allowed him to guarantee Poland with less fear of the immediate casualties that this might produce.86 By analogy, ballistic missile defenses, if they were believed to be highly effective, might encourage future policymakers to be more assertive in their balancing behavior.
Chain-ganging and buck-passing in World Wars I and 11
To sum up the findings from the two world wars, in every case perceptions of offensive advantage were associated with chain-ganging—that is, with unconditional balancing behavior. Conversely, perceptions of defensive advantage were associated with buck-passing—that is, with strategies of limited liability. Given a choice, states preferred to pass the costs of balancing to other states or to await developments before making irrevocable commitments. But when offensive advantages were believed to make states extremely vulnerable and wars short, buck-passing strategies were deemed too risky.
This hypothesis is more successful than some obvious competitors. For example, it is not true that states balance when they believe they are an aggressor's next target but pass the buck when they believe they are farther down on the list. If first-line states are seen as vulnerable but witling and able to balance if assisted, second-line states tend to accept the buck.
Likewise, it is not true that buck-passing has been driven strictly by a craven desire to minimize the costs of fighting, regardless of strategic consequences. The French decision to join in guaranteeing Poland was a strategic attempt to ensure the resources needed to stalemate Germany in a costly attritional campaign. If the French had been concerned only with minimizing casualties, but at a greater strategic risk, they would have shunned the British guarantee to Poland and tried instead to embroil Hitler and Stalin. Similarly,
86. For information on Chamberlain's views, see William R. Rock, Neville Chamberlain (New York: Twayne, 1969), p. 180; Cowling, Impact of Hitler, p. 395; and Ian Colvin, The Chamberlain Cabinet (London: Gollancz, 1971), p. 174. For background on air power policy and perceptions, see Posen, Sources of Military Doctrinerchap. 5; and Murray, Change in the
F----- „ ,-----fD....,.,, ™  lft« and ?<M_5V
166
Intel
.ia! Organization
Allian
Stalin's buck-passing was aimed not at saving the lives of Soviet soldiers; per se but, rather, at conserving Soviet power until the decisive moment] when the other powers would be exhausted by the first round of fighting. :
Finally, the evidence cited above belies the commonly expressed viewj that appeasers in Britain and France did not calculate strategically at all butj were simply reacting to public opinion or inchoate emotion. ''Chamberlain! was not primarily, if at all, motivated by strategic factors," states a recent] historian;87 likewise, "muddle not machination" is said to have been at the| bottom of French policy.88 Domestic political pressures and other sources! of perceptual bias undoubtedly influenced strategic calculations. But this is f not the same as saying that no calculations were made. V.-|
This was the case in 1914 as well as in 1938-39.89 In 1914, the military' had been highly successful in propagating what General Joseph Joffre himself _later called a "cult of the offensive," which served military organizational . interests. Conversely, in 1938, there existed a civilian "cult of the defensive," headed by B. H. Liddell Hart and others who sought to use any : strategic rationalization to avoid a British commitment to fighting a large land war on the European continent. Critics of Liddell Hart have clearly established that the strategy of limited liability came first for him and that ] only later did he develop his ideas of defensive advantage in armored warfare ' in order to explain how France could stalemate Germany without the help of a large British expeditionary force.90
The point is that strategic calculations were in fact made, if only to sell a policy as plausible, given a certain view of the offense-defense balance. Indeed, policy tended to dovetail with the logic of those arguments. Though these arguments may have sometimes been ex post facto rationalizations rather than root causes, assessments of offensive and defensive advantage were directly tied to grand strategic choices.
These choices had effects on the stability of the system. Strategies of aggressive balancing, based on perceptions of offensive advantage, and passive buck-passing, based on perceptions of defensive advantage, were both destabilizing. These instabilities were triggered by the fact that the underlying strategic assumptions were incorrect. Thus, the European confrontation of July 1914 escalated because of the expectation that states were vulnerable to conquest, but it was prolonged by the fact that they were not. Conversely, Hitler's opponents failed to appreciate that blitzkrieg operations against
87. Bond, British Military Policy, p. 282.
88. Adamthwaite, France and the Coming of the Second World War, p. 320. ,
89. For a discussion of the events in 1914, see Snyder, ideology of the Offensive. For similar points about 1938-39, see Posen, Sources of Military Doctrine. In "Causes of War," Van Evera offers a general theory of this type.
90. See Brian Bond and Martin Alexander, "Liddell Hart and De Gaulle: The Doctrine of Limited Liability and Mobile Defense," in Peter Paret, ed.. Makers of Modern Strategy, ed. (Princeton, N.J.: Princeton University Press, 1986), p. 612; and Mcarshetmer, LiddellHw and the Weight of History, pp. 107, 111-12. and 128.
rns 167
isolated targets would allow Germany to seize the assets needed to mount a serious bid for European hegemony. By 1941, for example, 40 percent of German steel production came from outside the Reich's 1937 borders.91
However, perceptions of defensive advantage need not always lead to this result, even when the aggressor occupies the center of the alliance checkerboard. In the late 1880s, Germany was dissuaded from attacking anyone because each of its opponents looked individually impregnable. However, in the late 1930s, perceptions of defensive advantage were destabilizing because the status quo states saw stronger defensive advantages than did the aggressors. In checkerboard conditions, therefore, the aggressor was not dissuaded from attacking isolated opponents, whereas the status quo states were dissuaded from aiding their allies by attacking the aggressor's
rear.
Conclusions and issues for research
Contemporary balance-of-power theory has become too parsimonious to yield determinate predictions about state alliance strategies in multipolarity. Waltz's theory predicts only that multipolarity predisposes states to either of two opposite errors, which we call chain-ganging and buck-passing. To predict which of these two policies will prevail, it is necessary to complicate Waltz's theory by adding a variable from Jervis's theory of the security dilemma; the variable of whether offense or defense is perceived to have the advantage. At least under the checkerboard geographical conditions in Europe before World Wars I and II, perceived offensive advantage bred unconditional alliances, whereas perceived defensive advantage bred free riding on the balancing efforts of others.
The marriage that we propose between Waltz's theory'and Jervis's suggests a number of issues for further research as well as a number of applications to current policy analysis. One question of considerable theoretical a"d policy interest is the source of stability in multipolar periods that lacked major wars. For example, the diplomacy of Bismarck's era managed to avoid [he pitfalls of both chain-ganging and buck-passing, despite its multipolar setting. Above, we briefly suggested that this may have been the result of the increasing perception in the 1880s that each of the European powers was individually too well defended to conquer. But Bismarck's limited aims and diplomatic skills may also have been factors. In any event, given the likelihood that the world will become increasingly multipolar, it would be useful 'o ask what role the offense-defense balance has played in cases in which multipolarity has been managed successfully.
The interaction of polarity and the offense-defense balance might also
91. Murray, Change in the European Balance of Power-, -p. 13.
(I
168   International Organization
(
yield interesting intepretations of regional conflict dynamics outside Europe. The June 1967 Arab-Israeli War might be interpreted as multipolar chain-ganging stemming from perceptions of offensive advantage, whereas the subsequent war of attrition can be seen as Syrian buck-passing stemming from perceptions of a defensive stalemate.92 Such regional multipolar processes are likely to become a more and more important feature of interna-' tional politics as the superpowers increasingly withdraw from their overextended positions in the Third World and even in Eastern Europe.
Analytically more difficult are multipolar settings that lack the familiar checkerboard geography which makes one's neighbor an enemy and makes the enemy's neighbor one's friend. Checkerboard balancing hypotheses become decreasingly helpful when sea and air power supplant land power as the dominant factor in the military equation. Insofar as the multipolarity of the twenty-first century is likely to feature the rise of Japan as a major sea and air power, heuristic historical cases of noncheckerboard alliance politics should focus on multipolar naval competition in the Eastern Mediterranean (the nineteenth-century's "Eastern question") or in East Asia.93
Nuclear weapons will also have to be factored in to any assessment of multipolar balancing in the future, both because their global reach undermines traditional checkerboard balancing logic and because the nuclear deterrent stalemate is likely to benefit the defender of the status quo.94 Insofar as nuclear weapons are likely to make each pole individually invulnerable to conquest, a nuclear-armed multipolarity may resemble the stable 1880s more than it will the chain-ganging 1910s or buck-passing 1930s. It cannot be excluded, however, that states with small, vulnerable nuclear arsenals will have to form alliances with larger nuclear powers or with each other to mount a credible deterrent. In that case, the dynamics of chain-ganging and buck-passing may still apply in future nuclear showdowns.
We make no claim to be able to foretell the balancing dynamics of the corning decades. We do claim, however, that realist scholars will have to prepare for this analytic challenge by developing a theory that combines the insights of Waltz's balance-of-power theory and Jervis's security dilemma theory. This is the most parsimonious international system theory that has any hope of explaining and prescribing great power alliance strategies.
92. We thank Stephen Walt for suggesting this possibility. Walt's Origins of Alliances applies a variant of balance-of-power theory to Middle Eastern case studies. : .
93. For background, see C. J. Bartlett, Great Britain and Sea Power. 1815-1853 (Oxfor*J| Clarendon, 1963). ,
94. For this argument as applied to the present bipolar setting, see Robert Jervis, The Meaning of the Nuclear Revolution (Ithaca, N.Y.: Cornell University Press, 1989).
The international telecommunications regime: the political roots of regimes for high technology Peter F. Cowhey
The international telecommunications regime provided a multilateral framework that reinforced domestic monopolies and bilateral agreements in the global market and thereby created one of the most lucrative and technologically significant international cartels in history. For over a century, it defied the laws of economics that caused other cartels to crash. Through multilateral agreements negotiated primarily by the International Telecommunications Union (ITU) and the International Telecommunications Satellite Organization (Intelsat), the regime helped create what is now estimated to be a $550 billion market in voice, visual, and data communications services and equipment.1 The regime was in fact a political invention so successful that it eventually disappeared from sight.
Today, there are significant efforts to restructure the telecommunications regime by introducing competition and, in particular, by granting some jurisdiction over telecommunications to trade institutions that serve new political constituencies. Although traditionally the General Agreement on Tariffs and Trade (GATT) has covered goods and not services, the current GATT round will most likely give GATT authority over services in order to break traditional regulatory arrangements both in domestic and in international markets.2
An earlier draft of this article was presented at a Social Science Research Council conference ai the University of Chicago. I thank Steve Krasner, David Laitin, John Ruggie, and Mark ^cher for their thoughtful comments. Steve Saideman provided valuable research assistance.
Equipment constitutes about $100 billion of this total. About $150 billion of the services market might be open to international competition by 1992. Market estimates are from Jonathan
; Aronson and Peter F. Cowhey, When Countries Talk: International Trade in Telecommunications Servicer (Cambridge, Mass.: Ballinger, 1988), p. 7. Note thai the study presented in n|s article excludes the market for television and radio programming.
2. The telecommunications regime reflects a central feature of the modern slate: a rough 'vision in the economic arrangements for goods and services. Services were at the core of the state's command over the economic infrastructure, which includes money, communications, asportation, the administration of laws, and the provision of health and education. Govern-
'Wernational Organization 44, 2, Spring 1990 '990 by the World Peace Foundation and the Massachusetts Institute of Technology