The Promise of Security Author(s): Charles A. Kupchan and Clifford A. Kupchan Source: International Security, Vol. 20, No. 1 (Summer, 1995), pp. 52-61 Published by: The MIT Press Stable URL: http://www.jstor.org/stable/2539215 . Accessed: 27/04/2011 06:55

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at . http://www.jstor.org/action/showPublisher?publisherCode=mitpress. .

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to International Security.

http://www.jstor.org The Promiseof CharlesA. Kupchan Collective Security and

I John J. Mearsheimer's critique of collective securitymisses its mark for three main reasons. First, Mearsheimer employs so narrow a definitionof collective securitythat he definesaway the issues most centralto evaluating the -causing effectsof institutionswithin the collectivesecurity family Second, he misrepresentshow collectivesecurity acts to promotestability, by portrayingit as based on mor- alistic principlesthat violate the logic of power balancing. But collectivesecu- rityis, if nothingelse, all about balancing and the aggregationof militaryforce against threatsto peace. Indeed, its main advantages over balancing under anarchyare thatit provides formore effectivebalancing against aggressorsand thatit promotesa more cooperativeinternational environment, thereby making inter-staterivalry and aggression less likely Third, Mearsheimer's general critique of institutionsstems froma theoreticalperspective-structural real- ism-that ignores the extent to which domestic politics,beliefs, and norms shape state behavior. By explaining and peace solely in termsof power balancing in an anarchicworld, Mearsheimermounts an attackthat is at once ahistoricaland internallycontradictory. We contend thata theoreticalperspec- tive that takes power seriously,but not to the exclusion of domestic and ideational variables,offers a richer,more accurate vision of internationalpoli- tics.It is withinthis vision thatcollective security has an importantrole to play in promotingpeace and cooperation.

DefiningCollective Security

The case forcollective security rests on the claim thatregulated, institutional- ized balancing predicated on the notion of all against one provides more stabilitythan unregulated,self-help balancing predicatedon the notionof each for his own. Under collectivesecurity, states agree to abide by certainnorms and rules to maintain stabilityand, when necessary,band togetherto stop

CharlesA. Kupchanis SeniorFellow for at theCouncil on ForeignRelations and a professorat GeorgetownUniversity. Clifford A. Kupchanis SeniorForeign Policy Adviserto CongressmanHarry Johnston.

The authorswould like to thankRichard Betts, George Downs, PeterKatzenstein, Robert Keohane, Lisa Martin,Jack Snyder, and AlexanderWendt for their comments on earlierdrafts of this article.

InternationalSecurity, Vol. 20, No. 1 (Summer 1995), pp. 52-61 ? 1995 by the Presidentand Fellows of Harvard College and the MassachusettsInstitute of Technology.

52 The Promiseof CollectiveSecurity | 53

aggression.Stability-the absence of major war-is the productof cooperation. In a world of balancing under ,states fend for themselves according to the dictates of a hostile internationalenvironment. Stability emerges from competition.The key question is whetherregulated balancing predicatedupon the notion of all against one, or unregulatedbalancing predicated upon the notionof each forhis own, is more likelyto preservepeace. Our task is to show only that collectivesecurity is preferableto balancing under anarchy,not that collectivesecurity is a panacea or the ultimateanswer to preventingwar. In his critique,Mearsheimer focuses only on ideal collective security-a variant in which states make automatic and legally binding commitmentsto respond to aggressionwherever and wheneverit occurs.He explicitlyexcludes fromconsideration other institutional formulations, such as concerts,that rely on looser and more informalregulation of balancing,arguing that they do not constitutecollective security. As a resultof this definitionalmaneuver, Mear- sheimerdirects his critiqueat a strawman and failsto engage the core concep- tual issue at stake:whether some formof regulated,institutionalized balancing is preferableto unregulatedbalancing under anarchy Of necessity,debate about the value of institutionsmust focus on generic formulations,not on the performanceof a specificinstitutional variant. Any institutionthat is predicatedupon the principlesof regulatedbalancing and all against one falls into the collective securityfamily Concerts do retain an undercurrentof competitive,self-help balancing. But they operate in a regu- lated, norm-governedenvironment and are predicated on the logic of all against one, not each forhis own. Accordingly,our originalterminology, which refersto a familyof collectivesecurity organizations ranging from ideal collec- tive securityto concerts,best captures the underlyingconceptual issues at stake.' Mearsheimer's formulationis, simply put, analyticallyunsustainable. He insists that concerts are "largely consistentwith realism" and logically "incompatible"with collectivesecurity, but writesthat concerts entail "coordi- nated balancing" among "greatpowers thathave no incentiveto challengeeach othermilitarily [and] agree on a set of rules to coordinatetheir actions" (p. 35). These featuresare fundamentalattributes of collective securityand stand in

1. For furtherdiscussion, see Richard Betts, "Systems for Peace or Causes of War? Collective Security,Arms Control,and the New Europe," InternationalSecurity, Vol. 17, No. 1 (Summer1992), pp. 5-43; George Downs and Keisuke lida, "Assessing the TheoreticalCase Against Collective Security,"in George Downs, ed., CollectiveSecurity Beyond the (Ann Arbor:University of Michigan Press, 1994), pp. 17-21; and Charles Lipson, "Is the Future of Collective SecurityLike the Past?" in ibid., pp. 105-131. InternationalSecurity 20:1 | 54

starkcontrast to the inescapable competitionand self-helpstrategies of realist balancing under anarchy2

The Advantagesof CollectiveSecurity

The advantages of collectivesecurity fall into two categories:it provides for more effectivebalancing against aggressors,and it promotes trustand coop- eration.

MORE EFFECTIVE BALANCING AGAINST AGGRESSORS Perhaps because of confusion over what collective securityis, Mearsheimer misunderstandsits underlyingcausal logic. Accordingto Mearsheimer,collec- tive securityrequires that states "ignore . . . balance-of-powerconsiderations" (p. 33). This characterizationis fundamentallymistaken. Collective security addresses head-on the centralconcern of realistswith the competitivenature of the internationalenvironment and its propensityto triggerspirals of hostil- ity Fully aware of the war-causingfeatures of the internationalsystem, collec- tive securityseeks to provide a more effectivemechanism for balancing against aggressors when they emerge, as well as to make aggression less likely by amelioratingthe competitivenature of internationalrelations. The challenge forproponents of collectivesecurity is not,as Mearsheimerwrites, to show that "institutionsare the key to managing power successfully"(p. 27). It is to show that thereis value added: that institutionsare betterthan no institutionsand offeran improvementupon the self-helpworld of balancing under anarchy Collective securityprovides formore effectivebalancing against aggressors than balancing under anarchybecause, when it works,it confrontsaggressors with preponderantas opposed to merely equal force.3Under anarchy,only those states directlythreatened by the aggressorand stateswith vital interests in the threatenedareas will band togetherto resistaggression. Under collective

2. See JohnJ. Mearsheimer, "The False Promiseof InternationalInstitutions," International Security, Vol. 19, No. 3 (Winter 1994/95), pp. 5-49. The conceptual muddle caused by Mearsheimer's restrictivedefinition of collectivesecurity is also apparentin his discussion of the empiricalrecord (pp. 33-34). Mearsheimerrefers to the and the as collective securityorganizations. Neither, however, comes close to fulfillingthe standardsof ideal collective security.The League Covenant and the UN Charterdo not entail automaticand binding commit- ments to respond to aggressionwith force.Both organizationscreated inner councils to enhance the influenceof the greatpowers. In theserespects, the League and the UN resembleconcerts more than they do ideal collectivesecurity organizations. 3. See Charles A. Kupchan and CliffordA. Kupchan, "Concerts,Collective Security, and the Future of Europe," InternationalSecurity, Vol. 16, No. 1 (Summer 1991), p. 117, n. 6. ThePromise of Collective Security | 55

security,other states are likelyto join the opposing coalition,both because they have made eitherexplicit or implicitcommitments to do so and because they have interestsin protectingan internationalorder that they see as beneficialto theirindividual security4Furthermore, even when it does not work,collective securityat its worst (thatis, when all memberstates otherthan those directly threatenedrenege on theircommitment to resistaggression) is roughlyequiva- lent to balancing under anarchyat its best. Should non-threatenedstates opt out of collective action, the remainingcoalition would consist of the same directlythreatened states as the that would formthrough balancing under anarchy. The most powerful critique of the argumentthat collective securityat its worst is roughly equivalent to balancing under anarchy at its best is that collective securityencourages member states to count on the assistance of others,thereby leaving a directlythreatened coalition underpreparedfor war if the system unravels (p. 30). In a self-helpworld, the argumentruns, the opposing coalition would have known that it was on its own, and prepared accordingly Because thiscritique has been dealt withelsewhere, here we only summarize the main points of rebuttal.5First, it is the specter of a collective security organization unraveling on the eve of aggression that causes concern about directlythreatened states being left unprepared for war. Yet this scenario is highly improbable; the failure of collective securitymechanisms is likely to occur in stages, giving directlythreatened states adequate warning that the blockingcoalition will not containits fullcomplement of members.In addition, directlythreatened members of a collective securitysystem would be well aware that some of theirpartners might defect;prudence would dictate the maintenance of force levels greaterthan those needed should all members

4. See ibid., p. 126. Collective securityseeks to expand the realm of private interestso that even states whose securityis not immediatelythreatened have a stake in preventingaggression. It does not, as Mearsheimerwrites, require that states "not thinkin termsof narrow self-interest"(p. 29). Rather,it seeks to broaden how states define theirself-interest through two differentpathways. First,assuming thatinterests are fixedand confinedto realistnotions of rationalegoism, collective securityalters incentives so thatstates more oftenfind it in theirinterests to cooperate as opposed to compete. Second, collectivesecurity alters the characterof state intereststhemselves, not just the behavior thatstates adopt to attainthose interests.Through processes of learningand sociali- zation, states can come to definetheir interests in more collectiveterms. Through its participation in the EU and NATO, forexample, Germanyhas come to defineits interestsin European rather than in purely national terms.For furtheydiscussion, see pp. 57-59 below. 5. See Charles Kupchan, "The Case for Collective Security,"in Downs, CollectiveSecurity, pp. 59-63. InternationalSecurity 20:1 | 56

fulfillcommitments to collectiveaction. Mearsheimersuggests that,until out- rightwar breaks out, states in a collective securitysystem "must trusteach other" and eschew steps to balance against potentialaggressors (pp. 29-30). But it is absurd to suggest that collective security-even in its ideal form- requires its members to stand by idly as one among them arms itselfto its teeth.As we have argued, concertsare particularlywell suited to orchestrating pre-aggressiondeterrence and the early formationof a preponderantblocking coalition.6 Second, states do not set forcelevels simplyby assessing the capabilitiesof the enemyand determininghow much of theirown militarypower is required, given the strengthof coalitionpartners, to achieve preponderance.The level of militarycapability maintained by a given state is affectedby its general threat environment,but also by a complex mix of political and economic considera- tions. There is no one-to-one ratio between external threatand force level. Rather,as threatsincrease, governments and publics become generallymore willing to devote increased resources to the output of defense goods. Force levels rise with the political will to support the necessary expenditures,not only as militaryplanners calculate what it will take to defeatthe enemy When faced with an increasinglyhostile adversary of growing militarystrength, a directlythreatened state in a collective securitysystem would devote more resourcesto defense,just as it would in an alliance system.Indeed, it may well maintain forcelevels roughlyequivalent to the levels it would maintainas a memberof a defensivealliance. Third,although freeriding may contributeto the underproductionof mili- tary capability,there is no compelling deductive reason why the free-rider problemshould produce a weaker opposing coalitionunder collectivesecurity than under balancing under anarchy.7All coalitions,including defensivealli- ances, can fall prey to free riding. Indeed, the historicalexample that Mear- sheimer uses to illustrate the free-riderproblem is that of intra-alliance buck-passingamong Britain,France, and Russia during (pp. 31- 32). Again, the key question is not whethercollective security is flawless,but

6. See Kupchan and Kupchan, "Concerts,Collective Security, and the Futureof Europe," pp. 138- 144. Collectivesecurity institutions that do not make responsesto aggressionautomatic and legally binding also take care of Mearsheimer's charge that collective security"transforms every local conflictinto an internationalconflict" by mandating that all members respond to every act of aggression (p. 32). Concerts can play as importanta role in orchestratingmutual restraintas in coordinatingcollective action. 7. For furtherdiscussion, see Downs and lida, "Assessing the TheoreticalCase Against Collective Security,"pp. 26-29. ThePromise of Collective Security | 57

whetherit detersand blocks aggressorsmore effectivelythan balancing under anarchy8

PROMOTING TRUST AND COOPERATION Assessmentof collectivesecurity's historical performance poses particularem- pirical problems because it is when collective securityis functioningmost effectivelythat its benefitsmay be difficultto discern.9Collective securityis preferableto balancing under anarchy not only because it provides better balancing against aggressors,but also because it fostersan environmentin which aggressionis less likelyto take place. Indeed, its abilityto mitigatethe rivalryand hostilityof a self-helpworld is one of its key advantages. Mearsheimermisrepresents collective security's reliance on and promotion of trustamong statesas one of its chieflogical flaws. "Collectivesecurity is an incompletetheory," Mearsheimer writes, "because it does not provide a satis- factoryexplanation for how states overcometheir fears and learn to trustone another" (p. 30). We acknowledge that basic compatibilityamong the great powers in a system is foremostamong the conditionsnecessary for the suc- cessfuloperation of collectivesecurity And this compatibilityis a functionof the underlyinginterests and intentionsof states,not of theirparticipation in a collective security system. But collective security,through mechanisms we outline in "Concerts,Collective Security,and the Future of Europe" (pp. 130- 133) builds on this basic compatibilityand reinforcesconfidence in the inten-

8. We acknowledge thatit is conceivablethat collective security could produce a weaker opposing coalition than balancing under anarchy.At least hypothetically,aggression could take place as a bolt fromthe blue, or directlythreatened states could be dangerously overconfidentabout the willingnessof theircoalition partners to join the fray.But forthe reasons just enumerated,the risks of such an outcome are low. And these risksare well worthtaking in lightof collectivesecurity's considerableadvantages. 9. Mearsheimer incorrectlyclaims that the empirical record undermines the case for collective security.The preserved peace in Europe for fortyyears, not, as Mearsheimer asserts, for eight. The Concert's handling of the Belgian Crisis of 1830-32, the Unkiar-Skelessi question in 1833-34, and the Egyptian Crisis of 1839-41 provides evidence of its successful operation after1823. It ceased to functiononly afterthe revolutionsof 1848 destroyedthe condi- tions thatenabled it to operate.See Kupchan and Kupchan, "Concerts,Collective Security, and the Futureof Europe," pp. 142-143,note 81. The League of Nations enjoyed successes duringthe 1920s, as Mearsheimerenumerates (p. 33). Admittedly,it failed dramaticallyto counterJapanese and German aggressionduring the 1930s. But the existenceof the League had virtuallynothing to do with the status quo powers' underpreparationfor war and theirinitial inability to deter or stop Germanyand .The UN was never seriouslytested as a collectivesecurity institution because of the Cold War. Withthe Cold War only recentlyover, it is too soon to judge whetherthe UN's effectivenessis on the rise or to determinewhether some combinationof the Organization for Securityand Cooperationin Europe (OSCE, formerlyCSCE), NATO, and the Partnershipfor Peace will emerge as a functioningcollective security institution for Europe. InternationalSecurity 20:1 | 58

tions of other states, therebydeepening cooperation.10It promotes a more benigninternational environment in which statescan devote less attentionand fewer resourcesto ensuringtheir survival and more to improvingtheir wel- fare-unless and until an aggressoremerges.11 There are profoundadvantages to institutionalizinga securitysystem that promisesto deepen accord among statesrather than lettinga self-helpsystem take its course and simplyhoping thatgreat power conflictdoes not reemerge. Collectivesecurity ameliorates the securitydilemma, thereby enhancing stabil- ity and reducing the likelihood of unintendedspirals of hostility12Collective securitywould also enable states to focus more on absolute as opposed to relativegains, a conditionthat Mearsheimeradmits would facilitatecoopera- tion (pp. 19-24). A statewill focusmore on absolute gains when it believes that the relativegains of otherswill not come back to haunt it. This beliefis in turn based on deep-seated assessments of the intentionsof those states enjoying relativegains. By building confidenceamong memberstates about each others' intentions,collective security thus mitigatesthe constraintsimposed on coop- erationby relative-gainsconsiderations.13 Collective securitywould not allow its members to focus exclusivelyon absolute gains, but states would be less concernedabout relativegains than in a self-helpworld. Finally,collective security institutions would help statesdefine their national interestsin ways that contributeto internationalstability Especially in post- Cold War Europe, where the strategiclandscape is ill-definedand major

10. Germany's participationin NATO provides an illustrationof this institutionalevolution. The Federal Republic became part of NATO because of the strategicobjectives it shared with other members.But the currentcloseness of Germany'srelations with its West European neighborsand with the is a functionnot just of shared interestsbut also of its steady participation in the web of Westerninstitutions. It is hard to imagine that Germany's relations with other establisheddemocracies would be as close as theyare today had these statesbeen interactingwith each otheronly as like-mindedpowers in an internationalenvironment without institutions. 11. In this sense, it is wrong to argue that collectivesecurity works only when it is not needed. On the contrary,it is self-reinforcing;as a collectivesecurity organization functions, it promotes the conditionsthat make it even more effective.The idea is not,as Mearsheimerinsists, that states must trusteach otherand be confidentthat status quo powers "will not change theirminds at a later date" (pp. 29-30). Rather,collective security affordsstates the opportunityto be more confidentabout the intentionsof othersuntil a given state's behavior proves otherwise. 12. See Kupchan and Kupchan, "Concerts,Collective Security, and the Futureof Europe," pp. 133- 137. 13. For example, the United States today would be relativelyunconcerned should Britainacquire a new offensiveweapons systemor enjoy a relativegain in a trade deal, not because the United States could best Britainif war broke out or readily find allies to forma blocking coalition,but because it is virtuallyinconceivable that Britainand the United States would findthemselves on opposing sides of a conflict.This confidencein Britain'sintentions is a productof decades of close, institutionalizedcooperation. The Promiseof Collective Security | 59

powers are in the midst of reformulatingtheir identities and interests,institu- tions will shape, and not just be shaped by,the distributionof power. NATO was a response to, not the cause of,the division of Europe into two competing blocs. But the institutionhas taken on a life of its own despite the collapse of the balance-of-powerconsiderations that led to its formation.NATO continues to enable and encourage Germany to define its interestsin European, not nationalterms. It provides a justificationand a vehicle forAmerica's continued militaryengagement in Europe. Its integratedmilitary structure encourages national militaryestablishments to formulateobjectives and strategiesthat are multinational,not national,in characterand outlook. NATO's futurewill affectnot only how its currentmembers interactwith each other,but also how the statesof theformer Soviet bloc definetheir security needs. If NATO expands into Central Europe as a defensivemilitary alliance and then stops, it will effectivelydraw a new dividing line between Europe's east and west. It would be the lines and resultantpower blocs created by institutions,not by other political or ideological cleavages, that would help definefor Russia what its new sphere of influenceis, whetherit is a European or a Eurasian power, and whetherits relationswith NATO will be cooperative or competitive.Instead, Russia should be gradually drawn into a European collective securitysystem, increasing the chances that Russians will come to define themselves as members of a European communityof nations, not as outsiders. For reasons of its own, Russia may well veer from the path of democraticreform and pursue foreignpolicies incompatiblewith its participa- tion in a collectivesecurity system. But takingcautious, prudentsteps toward its inclusion unless and untilRussia demonstratesmalign intentionsoffers far more promise of preservingpeace in Europe than exposing a fragileRussia to the vagaries and insecuritiesof a self-helpworld.

The Povertyof StructuralRealism

Underlyingthis debate about the value of internationalinstitutions is a funda- mental differenceof opinion about the causes of war and peace. In the end, our assessment of the promise of collectivesecurity stems froma theoretical perspectivethat is incompatiblewith Mearsheimer's structuralrealism. It is thereforeappropriate to end this replyby making explicitthe precise areas of disagreement. In Mearsheimer'sworldview, all greatpowers are createdequal. When they see the opportunityto do so, greatpowers will take advantage of one another, InternationalSecurity 20:1 | 60

fearfulof being exploited later if theydo not. World War I, World War II, and the Cold War were nothingmore and nothingless than greatpowers actingas they must, given the exigencies of an anarchic,self-help world. From within this worldview,collective security, and internationalinstitutions more gener- ally,matter at the margins,if at all. Sooner or later,balance-of-power consid- erationswill overridethe rules and normsof institutionalstructures. Collective securityorganizations may be not only irrelevant,but also dangerous. States that place illusory faithin collective securitywill find themselvesworse off than had theyacted as if in a self-help,anarchic setting. In our worldview, all great powers are not created equal. Although the behavior of major states is heavily influencedby balance-of-powerconsidera- tions, domestic politics, beliefs, and norms mattertoo, and not just at the margins.World War I, World War II, and the Cold War came about not from the warp and woof of internationalcompetition, but as a result of the emer- gence of aggressor states-states that for reasons of ideology and domestic politics became predatoryand sought power, not security WilhelmineGer- many,,and interwarJapan were malign great powers infected with virulentdomestic pathologies,not garden-varietygreat powers dealing with legitimate security concerns. Each commenced an ambitious buildup and embarked down the path of aggression during peaceful periods in which they faced no imminentsecurity threats. Domestic politics and na- tionalism,not just the rivalryof a self-helpworld, were at play. Similarly,the United States and the were not equally to blame forthe Cold War. The United States sought its share of wealth and power but,with some notable exceptions,conducted itselfas a benign great power. Soviet Russia was the principal aggressorstate in the Cold War,driven in part by vulnerabilityand the search forsecurity, but also by domestic and ideological pathologies.14 Our contentionthat it is not only power politicsbut also the nature of both domestic and internationalsocieties that affectsgreat-power behavior is the basis forour optimismabout the promiseof collectivesecurity It is conceivable that Russia will emerge as a benign, democraticgreat power and that all of Europe's major stateswill share similarvalues and interests,the underpinnings forthe successfulfunctioning of a collectivesecurity system. Even Mearsheimer admits thatideational variables can play a role in shaping relationshipsamong

14. For discussion of aggressorstates and theircauses, see Charles Kupchan, The Vulnerabilityof Empire(Ithaca, N.Y: Cornell UniversityPress, 1994); and Stephen Van Evera, "Primed forPeace: Europe Afterthe Cold War,"International Security, Vol. 15, No. 3 (Winter1990/91), pp. 7-57. The Promiseof Collective Security | 61

states. What, afterall, does he mean when he writes that "some states are especiallyfriendly for historical or ideological reasons" (p. 31)? We submitthat he is scratchingthe surfaceof the povertyof his own theoreticalframework, forcedto resortto variables other than the balance of power to explain why states sometimescooperate to the extentthey do. The case for collective securityrests not on woolly-headed moralism or naivete about the demands imposed on states by power politics.It restson a more nuanced understandingof internationalpolitics than that offeredby structuralrealism. The post-Cold War era offersan excellent laboratoryin which to pit these competingtheoretical perspectives against each other.If, one or two decades fromnow, Russia is a fullmember of a pan-Europeancollective securitybody, Mearsheimer will have to recant.If, on the otherhand, collective securityis given a trybut Europe's great powers again fall prey to national rivalriesand its multilateralinstitutions founder, we will have to reconsider not just collectivesecurity, but the theoreticalsuppositions thatundergird our confidencein it. Unless collectivesecurity is given a chance, however,oppor- tunitiesto preserve peace in Europe will be missed and unresolved debates between structuralrealists and institutionalistsof various stripeswill continue to fillthe pages of InternationalSecurity.