Schweller, R. (2004) - Unanswered Threats A Neoclassical Realist Theory of Underbalancing

Unanswered Threats: A Neoclassical Realist Theory of Underbalancing
Author(s): Randall L. Schweller

Source: International Security, Vol. 29, No. 2 (Fall, 2004), pp. 159-201
Published by: The MIT Press
Stable URL: http://www.jstor.org/stable/4137589 .
Accessed: 09/10/2014 22:09
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
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 support@jstor.org.
The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to International Security.
http://www.jstor.org
This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

All use subject to JSTOR Terms and Conditions
UnansweredThreats RandallL. Schweller
A Neoclassical Realist
Theory of Underbalancing
During thelate 1630s,

Charles I concentratedhis energieson the constructionof a new royal palace at
Whitehall. Designed in the classical style by JohnWebb, the new Whitehall
was to be the fulfillment of the king's lifelongdream to replace the sprawling
and obsolete palace thathe had inheritedfromthe Tudors with one thatwould
matchthe splendor and majestyof the Louvre or theEscorial. Charles I desired
nothingelse than thathis surroundingsshould reflectthe magnificenceof his
rule: "Here, at last,would be a seat of governmentappropriateto the systemof
'Personal Rule' Charles I had established since dispensing with Parliamentin
1629. At least until 1639,it was fromhere thatCharles could expect to govern
his realms,resplendentamid Webb's Baroque courtyardsand colonnades, dur-
ing the next decade and beyond."'
In making such ambitious plans, Charles I displayed supreme confidence
thathis regimewould not only survivebut thrivewell intothe future.Unfortu-
natelyforthe king,his reigndid not last out the 1630s. If the conventionalhis-
torical wisdom that "the collapse of Charles I's regime during the 1630s
appeared 'inevitable"' is correct,then Charles obviously sufferedfromself-
delusion-an unrealityall too characteristicof remoteand isolated rulers.2
Internationalpolitics, too, has seen many instances of this type of folly,
where threatenedcountrieshave failed to recognizea clear and presentdanger
or,more typically,have simplynot reacted to it or,more typicallystill,have re-
sponded in paltryand imprudentways. This behavior, which I call "under-
balancing," runs directlycontraryto the core predictionof structuralrealist
RandallL. Schwelleris AssociateProfessor
in theDepartmentofPoliticalScienceat TheOhio StateUniver-
sity.He is theauthorofDeadly Imbalances: Tripolarityand Hitler's StrategyofWorld Conquest. This
articleis drawnfroma recentlycompleted book-length
manuscript entitled"UnansweredThreats:Political
Constraintson theBalanceofPower."
The authoris gratefulto the followingindividuals fortheircommentson earlierdraftsof thisarti-

cle: Michael Desch, Colin Elman, Yoav Gortzak,Yoram Haftel,RichardHerrmann,Alan Lamborn,
JackLevy,Richard Rosecrance,David Schweller,Russell Schweller,JackSnyder,BrentStrathman,
and Kevin Sweeney.He would also like to thankthe seminarparticipantsof the Universityof Chi-
cago Programon InternationalPolitics,Economics, and Security,especially Vaidya Gundlupet for
his extensive writtencomments.
1. JohnAdamson, "England withoutCromwell: What If Charles I Had Avoided the Civil War?" in
Niall Ferguson,ed., VirtualHistory:Alternatives (New York:Basic Books, 1997),
and Counterfactuals
p. 92.
2. Ibid. This is not JohnAdamson's view, however.
International Vol.29,No. 2 (Fall2004),pp. 159-201

Security,
@ 2004by thePresident
and FellowsofIHarvard Collegeand theMassachusetts
Institute
ofTechnology.
159

International
Security29:2 ? 160
theory,namely,thatthreatenedstates will balance against dangerous accumu-

lations of power by formingalliances or building arms or both. Indeed, even
the most cursoryglance at the historicalrecordreveals many importantcases
of underbalancing.Consider,forinstance,thatnone of the greatpowers except
Britainconsistentlybalanced against Napoleonic France, and none emulated
its nation-in-armsinnovation.Later in the century,Britainwatched passively
in splendid isolation as the North defeated the South in the American Civil
War and as Prussia defeated Austria in 1866, and then France in 1871, estab-
lishing German hegemony over Europe. Bismarck then defied balance of
power logic by cleverlycreatingan extensive"hub-and-spoke" alliance system
that effectivelyisolated France and avoided a counterbalancing coalition
against Germany.The Franco-Russianalliance of 1893 emerged only afterBis-
marck's successor, Leo von Caprivi, refused to renew the 1887 Reinsurance
Treatywith Russia fordomestic political reasons and despite the czar's plead-
ings to do otherwise.Thus, more than twentyyears afterthe creation of the
new German state,a balancing coalition had finallybeen forgedby the dubi-
ous decision of the new German chancellorcombined withthe kaiser's soaring
ambitions and truculentdiplomacy.
Likewise, duringthe 1930s,none of the greatpowers (i.e., Britain,France,the
United States, the Soviet Union, Italy,and Japan) balanced with any sense of
urgencyagainst Nazi Germany.Instead, theybandwagoned, buck-passed, ap-
peased, or adopted ineffectivehalf measures in response to the growing Ger-
man threat.A similarreluctanceto check unbalanced power characterizesmost
interstaterelationssince 1945. Withthe exceptionof the U.S.-Soviet bipolar ri-
valry,a survey of state behavior during the Cold War yields few instances of
balancing behavior. As K.J.Holsti asserts: "Alliances, such a common feature
of the European diplomatic landscape since the seventeenthcentury,are nota-
ble by their absence in most areas of the Third World. So are balances of
power." Holsti furthernotes: "The search forcontinentalhegemony is rare in
the ThirdWorld,but was a common featureof European diplomacy under the
Habsburgs, Louis XIV, Napoleon, Wilhelmine Germany,Hitler, and Soviet
Union and, arguably,the United States."3In a continuationof this pattern,no
peer competitorhas yet emerged more than a decade afterthe end of U.S.-
Soviet bipolarityto balance against the United States. Contraryto realist pre-
3. K.J.Holsti, "InternationalRelationsTheoryand Domestic War in the ThirdWorld: The Limitsof

Relevance," in Stephanie G. Neuman, ed., International
RelationsTheoryand theThirdWorld(New
York: St. Martin's, 1998), p. 106.

UnansweredThreats 161
dictions, unipolarityhas not provoked global alarm to restorea balance of

power.4
If most states inside and outside of the Eurocentricdomain fromwhich bal-
ance of power theorywas derived and largelytested can resistits logic, then
the theoryis certainlynot akin to a law of nature,as it has been portrayed;at
the very least, it is underspecified.The question then becomes: what are the
necessary conditionsforthe proper operation of the balance of power? What
factorsconfound the logic and predictionsof the theory?The main problem is
realism's assumption of states as coherentactors. The closer the policymaking
process and actual state-societyrelationsapproximatea unitaryactor,the more
accurate realism's predictions.Conversely,when states are divided at the elite
and societal levels, theyare less likelyto behave in accordance with balance of
power predictions.The core of the argumentturnson the political and policy
risks associated with balancing behavior. Specifically,leaders of incoherent
states are less willing and able to undertake high political and policy risks to
balance than are leaders of coherentstates.5Given its focus on elite risk-taking
propensity,the logic of my argumentdoes not apply when balancing is per-
ceived as relativelycostless forelites and society.
The article unfolds as follows. First,I discuss the reasons why there have
been few studies of domestic politics and balance of power. Second, I situate
my domestic politics explanationof underbalancingwithinthe relevantlitera-
tures and the realistparadigm. Third, I offerprecise definitionsof balancing
and underbalancing and explain how the two concepts should be viewed
within the larger contextof balance of power theory.The fourthsection pre-
sents a neoclassical realisttheorythatpredictshow stateswithvarious domes-
tic situationswill respond to threats.Fifth,I offerreasons why social divisions
produce elite fragmentation, which in turncontributesto underbalancingand
policy paralysis. This analysis is followed by an examinationof interwarBrit-
ain and France,which serve as crucial cases against which to testthe theory.I
4. For various explanations of this phenomenon,see the essays in G. JohnIkenberry,ed., America

Unrivaled:The FutureoftheBalanceofPower(Ithaca, N.Y.: Cornell UniversityPress, 2002).
5. Yet,incoherentstatesmay be more likelythan coherentones to wage offensiveand diversionary
wars. These wars are motivated not by externalsecurityor balance of power logic but ratherby
greedy expansion on the cheap. As Edward D. Mansfield and JackSnyder argue, poorly institu-
tionalized democratizing states, which they call "incoherent" democracies, are especially war
prone because competingelites employ nationalistrhetoricto mobilize mass support and then get
locked in to the belligerentforeignpolicies thatare unleashed by this process. See Mansfield and
Snyder,"Democratization and the Danger of War," International Security,Vol. 20, No. 1 (Summer
1995), pp. 5-38; and Mansfield and Snyder,"Democratic Transitions,InstitutionalStrength,and
War,"International Organization,Vol. 56, No. 2 (Spring 2002), pp. 297-337.

Security29:2 162
International
conclude with a discussion of why the internationalrelations literaturehas

been biased in favor of theories about overreactionsto threatsand, conse-
quently,has overlooked the equally dangerous behavior of underresponding
to them.
BalanceofPoweras a StructuralLaw ofNature
Despite the frequencyof underbalancing,thereare surprisinglyfew studies on

thistypeof behavior and just as few on the role of domesticpoliticsin balance
of power theory.The reason for this theoreticallacunae is that balance of
power has been traditionallytreatedas a law of nature,whereinthe whole uni-
verse is pictured"as a giganticmechanism,a machine or a clockwork,created
and kept in motion by the divine watchmaker.""The origins of balance of
power theoryare importantin explaining why structureand natural laws,
ratherthan domestic politics,have dominated its discourse.
The idea of a balance of power in internationalpolitics arose during the Re-
naissance as a metaphoricalconceptborrowed fromotherfields,such as ethics,
the arts,philosophy,law, medicine,economics,and the sciences,where balanc-
ing and its relationto equipoise and counterweighthad already gained broad
popular acceptance. Whereverit was applied, balancing was conceived as a
law of nature underlyingconcepts viewed as generally appealing, desirable,
and sociallybeneficial(e.g., order,peace, justice,fairness,moderation,symme-
try,harmony,and beauty).' As Jean-JacquesRousseau observed about the
states of Europe: "The balance existingbetween the power of these diverse
members of the European society is more the work of nature than of art. It
maintainsitselfwithouteffort, in such a manner thatif it sinks on one side, it
reestablishesitselfvery soon on the other."
This Renaissance view of balancing behavior as a response driven by a law
6. Hans J.Morgenthau,PoliticsamongNations:The StruggleforPowerand Peace,4th ed. (New York:

AlfredA. Knopf, 1966), p. 197.
7. AlfredVagts,"The Balance of Power: Growthof an Idea," WorldPolitics,Vol. 1, No. 1 (October
1948), pp. 82-101. For comprehensivetreatmentsofbalance of power theory,see Michael Sheehan,
BalanceofPower:Historyand Theory(New York:Routledge, 1996); Evan Luard, TheBalanceofPower:
The SystemofInternational Relations,1648-1815 (London: Macmillan, 1992); Inis L. Claude Jr.,Power
and InternationalRelations(New York: Random House, 1962); Edward V. Gulick, Europe'sClassical
BalanceofPower(New York:W.W.Norton,1955); and JohnA. Vasquez and Colin Elman, eds., Real-
ismand theBalancingofPower:A New Debate (Upper Saddle River,N.J.:PrenticeHall, 2003).
de M. I'abbdde SaintPierre,as quoted in
8. Jean-JacquesRousseau, Extraitdu projetde paixperpetuelle
Ernst B. Haas, "The Balance of Power: Prescription,Concept, or Propaganda?" WorldPolitics,
Vol. 5, No. 4 (July1953), p. 453.

UnansweredThreats1163
of naturestillinfusesmost discussions of how the theoryoperates. Thus, Hans

Morgenthau wrote: "The aspiration forpower on the part of several nations,
each tryingeitherto maintainor overthrowthe statusquo, leads of necessityto
a configurationthatis called the balance of power and to policies thataim at
preservingit."' More recently,KennethWaltz has declared: "As natureabhors
a vacuum, so internationalpolitics abhors unbalanced power."'" Likewise,
ChristopherLayne has averred: "Great powers balance against each otherbe-
cause structuralconstraintsimpel them to do so."" Realists invoke the same
"law of nature" metaphorto explain opportunisticexpansion. In thisvein, Ar-
nold Wolfershas said of structuralincentivesforgains: "Since nations,like na-
ture,are said to abhor a vacuum, one could predictthat the powerful nation
would feel compelled to fillthe vacuum with its own power."'2 Using similar
logic, JohnMearsheimerhas claimed that"status quo powers are rarelyfound
in world politics,because the internationalsystemcreatespowerfulincentives
forstates to look foropportunitiesto gain power at the expense of rivals,and
to take advantage of those situationswhen thebenefitsoutweigh the costs."13
From the policymaker's perspective, however, balancing superior power
and fillingpower vacuums hardlyappear as laws of nature.Instead, these be-
haviors,which carryconsiderablepotentialpoliticalcosts and uncertainpolicy
risks,emergethroughthemedium of the politicalprocess; as such, theyare the
product of competitionand consensus building among elites with differing
ideas about the political-military world and divergentviews on the nation's
goals and challenges and the means that will best serve those purposes.14 As
Nicholas Spykman observed many years ago, "Political equilibriumis neither
a giftof the gods nor an inherentlystable condition.It resultsfromthe active
interventionof man, fromthe operationof politicalforces.States cannotafford
to wait passively forthe happy timewhen a miraculouslyachieved balance of
9. Morgenthau,PoliticsamongNations,p. 163.
10. KennethN. Waltz,"StructuralRealism afterthe Cold War,"International Vol. 25, No. 1
Security,
(Summer 2000), p. 28 n. 84.
11. ChristopherLayne, "From Preponderance to OffshoreBalancing: America's Future Grand
Strategy,"InternationalSecurity, Vol. 22, No. 1 (Summer 1997), p. 117.
12. Arnold Wolfers,Discordand Collaboration: Essayson International
Politics(Baltimore,Md.: Johns
Hopkins UniversityPress, 1962), p. 15.
13. JohnJ.Mearsheimer,TheTragedy ofGreatPowerPolitics(New York:W.W.Norton,2001), p. 21.
14. See, forexample,the descriptionof the policymakingprocess in WarnerR. Schilling,"The Poli-
tics of National Defense: Fiscal 1950," in Schilling,Paul Y. Hammond, and Glenn H. Snyder,Strat-
egy,Politics,and DefenseBudgets(New York:Columbia UniversityPress, 1962), pp. 5-27; and Roger
Hilsman, The PoliticsofPolicyMakingin Defenseand ForeignAffairs(New York: Rowman and Lit-
tlefield,1971).

International
Security29:2 164
power will bringpeace and security.If theywish to survive,theymust be will-

ing to go to war to preservea balance against the growinghegemonicpower of
the period."15
In an era of mass politics,the decision to check unbalanced power by means
of arms and allies-and to go to war if these deterrentmeasures fail-is very
much a politicalact made by politicalactors.War mobilizationand fightingare
distinctlycollectiveundertakings.As such, political elites carefullyweigh the
likelydomesticcosts ofbalancing behavior against the alternativemeans avail-
able to them (e.g., inaction,appeasement, buck-passing,bandwagoning, etc.)
and the expected externalbenefitsof a restoredbalance of power. Structural
imperativesrarely,if ever, compel leaders to adopt one policy over another;
decisionmakersare not sleepwalkers buffetedabout by inexorable forcesbe-
yond theircontrol.This is not to say,however,thattheyare oblivious to struc-
tural incentives.Rather,statesrespond (or not) to threatsand opportunitiesin
ways determinedby both internaland externalconsiderationsof policy elites,
who must reach consensus within an often decentralized and competitive
political process.
The Role ofDomesticPoliticsin RealistTheory
The themeof thisarticlefitssquarely withinthe new wave of neoclassical real-

ist research,which emerged in the early 1990s and posits that systemicpres-
sures are filteredthroughinterveningdomestic variables to produce foreign
policy behaviors. Works by Thomas Christensen,Aaron Friedberg,Randall
Schweller,JackSnyder,William Wohlforth,and Fareed Zakaria all show that
statesassess and adapt to changes in theirexternalenvironmentpartlyas a re-
sult of theirpeculiar domestic structuresand political situations. More spe-
cifically,complex domestic political processes act as transmissionbelts that
channel, mediate, and (re)directpolicy outputs in response to externalforces
(primarilychanges in relativepower). Hence, states oftenreact differently to
similarsystemicpressuresand opportunities,and theirresponses may be less
motivated by systemic-levelfactorsthan domestic ones.1"
15. Nicholas JohnSpykman,America'sStrategyin WorldPolitics:The UnitedStatesand theBalanceof

Power(New York: Harcourt,Brace, 1942), p. 25.
16. For a review and analysis of the various works by these neoclassical realists,see Gideon Rose,
"Neoclassical Realism and Theoriesof ForeignPolicy," WorldPolitics,Vol. 51, No. 1 (October 1998),
pp. 144-172; and Randall L. Schweller,"The Progressivenessof Neoclassical Realism," in Colin
Elman and Miriam Fendius Elman, eds., Progressin International RelationsTheory:Appraisingthe
Field (Cambridge, Mass.: MIT Press, 2003), chap. 9.

Curiously, neoclassical realists have yet to examine the subject of under-

balancing in response to externalthreat(s).Instead, theyhave offeredexplana-
tions for normal expansion (Zakaria), reckless overexpansion (Snyder),
adaptation to relativedecline (Friedberg),why leaders inflateexternalthreats
to sell costlyinternalmobilizationcampaigns (Christensen),how perceptions
of the balance of power affectstatebehavior (Wohlforth),and alignmentdeci-
sions based on state motivations(Schweller).'7
There have been several challenges to the conventionalrealistwisdom that
balancing is more prevalentthanbandwagoning, a formof underbalancingbe-
havior.Paul Schroeder's broad historicalsurvey of internationalpoliticsshows
thatstateshave bandwagoned with or hidden fromthreatsfarmore oftenthan
theyhave balanced against them.Similarly,I have claimed thatbandwagoning
behavior is prevalentamong revisioniststates,whose behavior modernrealists
have ignored,because theiralliance choices are driven more by the search for
profitthan security.' More recently, RobertPowell treatsstatesas rationaluni-
tary actors within a simple strategicsettingcomposed of commitmentissues,
informationalproblems,and thetechnologyof coercionand findsthat"balanc-
ing is relativelyrare in the model. Balances of power sometimes form,but
thereis no general tendencytoward thisoutcome. Nor do statesgenerallybal-
ance against threats .... States frequentlywait, bandwagon, or,much less of-
ten, balance.""' Yet Powell notes that a rational-unitary-actor assumption
"does not mean thatdomesticpoliticsis unimportant.""'None of these studies
have offereda domestic politics explanation forbandwagoning or a theoryof
17. See Fareed Zakaria, FromWealthto Power:The Unusual OriginsofAmerica'sWorldRole (Prince-

ton,N.J.:PrincetonUniversityPress,1998); JackL. Snyder,MythsofEmpire:DomesticPoliticsand In-
ternational Ambition(Ithaca, N.Y.: Cornell UniversityPress, 1991); Aaron L. Friedberg,The Weary
Titan:Britainand theExperienceofRelativeDecline,1895-1905 (Princeton,N.J.:PrincetonUniversity
Press, 1988); Thomas J.Christensen,UsefulAdversaries:Grand Strategy,DomesticMobilization,and
Sino-American 1947-1958 (Princeton,N.J.:PrincetonUniversityPress, 1996); William Curti
Conflict,
Wohlforth, TheElusiveBalance:Powerand Perceptions duringtheCold War(Ithaca, N.Y.: Cornell Uni-
versityPress, 1993); and Randall L. Schweller,DeadlyImbalances:Tripolarity and Hitler'sStrategyof
WorldConquest(New York: Columbia UniversityPress, 1998).
18. Paul Schroeder,"Historical Realityvs. Neo-realistTheory,"International Vol. 19, No. 1
Security,
(Summer 1994),pp. 72-107; and Randall L. Schweller,"Bandwagoning forProfit:Bringingthe Re-
visionistState Back In," ibid., pp. 108-148. See also RobertJervisand JackSnyder,eds., Dominoes
and Bandwagons:StrategicBeliefsand GreatPowerCompetition in theEurasianRimland(New York:Ox-
fordUniversityPress, 1991); and Kevin Sweeney and Paul Fritz,"Jumpingon the Bandwagon: An
Interest-basedExplanation forGreatPower Alliances," JournalofPolitics,Vol. 66, No. 2 (May 2004),
pp. 428-449.
19. Robert Powell, In theShadowof Power:Statesand Strategiesin International Politics(Princeton,
N.J.:PrincetonUniversityPress, 1999), p. 196.
20. Ibid., p. 26.

International
Security29:2 166
the broader phenomenon of underbalancing behavior, which includes buck-

passing, distancing, hiding, waiting, appeasement, bandwagoning, and in-
effectivehalf measures.
Another strand of realist literaturefocuses on buck-passing, a form of
underreactionto threatsby which states attemptto freeride on the balancing
effortsof others. There are two basic structural-systemic explanations for
buck-passing. Christensen and Snyder claim that great powers under multi-
polaritywill pass the balancing buck when they perceive a defensive advan-
tage; on the other hand, Mearsheimer argues that buck-passing occurs
primarilyin balanced multipolarsystems,especially among greatpowers that
are geographically insulated from the aggressor.21I argue, instead, that
whetherstatesbalance against threatsis not primarilydeterminedby systemic
factorsbut rather,like all national securitydecisions, by the domestic political
process.22
WHAT IS MEANT BY BALANCING BEHAVIOR?
Although arguably the most frequentlyused termin the fieldof international

relations,balancing remains an ambiguous concept. Balancing means the cre-
ation or aggregation of militarypower throughinternalmobilization or the
forgingof alliances to preventor deter the territorialoccupation or political
and militarydominationof the stateby a foreignpower or coalition.Balancing
exists only when the stakes concern some form of political subjugation or,
more directly,the seizure of territory,eitherone's homeland or vital interests
abroad (e.g., sea-lanes, allies, colonies, etc.). Thus, balancing requires that
statestargettheirmilitaryhardwareat each otherin preparationfora potential
war. If two statesare merelybuilding arms forthe purpose of independent ac-
tion against thirdparties,theyare not balancing. Indeed, stateA may be build-
ing up its militarypower and even targetinganother state B and still not be
balancing against B, thatis, tryingto matchB's overall capabilitiesto preventB
frominvading A or its vital interests.Instead, the purpose of A's actions may
be coercive diplomacy: to gain bargainingleverage with B. Finally,the timing
21. Thomas J.Christensenand JackSnyder,"Chain Gangs and Passed Bucks: PredictingAlliance

Patternsin Multipolarity,"International
Organization,Vol. 44, No. 2 (Spring 1990), pp. 137-168; and
Mearsheimer,The TragedyofGreatPowerPolitics,pp. 271-273.
22. See, for example, JackS. Levy and Michael M. Barnett,"Domestic Sources of Alliances and
Alignments:The Case of Egypt, 1962-1973," InternationalOrganization,Vol. 45, No. 3 (Summer
1991), pp. 369-395; and Levy and Barnett,"Alliance Formation,Domestic Political Economy,and
Third World Security,"Jerusalem Relations,Vol. 14, No. 4 (December 1992),
JournalofInternational
pp. 19-40.

of a state's response to threatdetermineswhetherit is balancing or not. Bal-

ancing does not mean simply fightingback ratherthan surrenderingafteran
attack.As I have writtenelsewhere,balancing involves a situationin which "a
state is not directlymenaced by a predatory state but decides to balance
against it anyway to protectits long-termsecurityinterests";or, as JackLevy
puts it,"it would not be balancing ifwar is forcedon the potentialbalancer by
a directmilitaryattackby the aggressor."23
Balancing and underbalancingcan be brokendown intofourdistinctcatego-
ries. The firstis simplyappropriatebalancing,which occurs when the targetis
a trulydangerous aggressor that cannot or should not be appeased and the
state's military capabilities are indispensable to counterbalance the rising
state's power. The second is overbalancing(or inappropriatebalancing),which
unnecessarilytriggersa costlyand dangerous arms spiral because the targetis
misperceivedas an aggressorbut is, instead,a defensivelyminded state seek-
ing only to enhance its security.24The thirdcategoryis nonbalancing,which
may take the form of buck-passing, bandwagoning, appeasement, engage-
ment,distancing,or hiding. These policies (which may also be underbalancing
behaviors when adopted under the circumstancesspecified below) are pru-
dent and rationalwhen the stateis therebyable to avoid the costs of war either
by satisfyingthe legitimategrievances of the revisioniststate,or by allowing
othersto do so, or by lettingothersdefeattheaggressorwhile safelyremaining
on the sidelines. In addition,ifthe state also seeks revision,thenit may wisely
choose to bandwagon with the potential aggressor in the hope of profiting
fromits success in overturningthe established order.25In still othercases, one
stateis so overwhelminglypowerfulthata harmonyof interestsexistsbetween
the hegemon (or unipole) and the restof the greatpowers-those thatone day
could eitherbecome peer competitorsor join togetherto balance against the
predominantpower. The other states do not balance against the hegemon be-
cause they are too weak (individually and collectively)and, more important,
because they perceive theirwell-being as inextricablytied up with the well-
being of the hegemon. Here, potential "balancers" bandwagon with the he-
23. I am gratefulto Thomas Christensenfor his insightsregardingthis definitionof balancing.

Schweller,"Bandwagoning forProfit,"p. 135; and JackLevy, "Balance and Balancing: Concepts,
Propositions,and Research Design," in Vasquez and Elman, Realismand the Balancingof Power,
p. 135.
24. This view of appropriateand inappropriatebalancing follows RobertJervis'sspiral and deter-
rence models. See Jervis,Perceptionand Misperceptionin InternationalPolitics (Princeton,N.J.:
PrincetonUniversityPress, 1976), chap. 3.
25. See Schweller,"Bandwagoning forProfit."

Security29:2 168
International
gemon not because they seek to overthrowthe established order (the motive
for revisionistbandwagoning) but because they perceive themselves to be
benefitingfromthe status quo and, therefore,seek to preserve it.21
The fourthcategoryis underbalancing,which occurs when the state does
not balance or does so inefficientlyin response to a dangerous and unappeas-
able aggressor,and the state's effortsare essential to deter or defeat it. In this
case, the underbalancing state brings about a war that could have been
avoided or makes thewar more costlythanit otherwisewould have been. This
articleconcernsunderbalancing.What I offer, therefore, is essentiallya "theory
of mistakes,"so to speak, provided thatone were to consider the policy choice
solely in terms of the internationalstrategicsetting.In other words, under-
balancing is the opposite mistake of overbalancing. When the state under-
balances, it either misperceives the intentionsof the rising power as more
benign than they in fact are or, if it correctlyperceives the threat,does not
adopt prudent policies to protectitselfforreasons of domestic politics.
A NeoclassicalRealistExplanation
Variationin the way states respond to similar changes in theirexternalenvi-

ronmentturnson the preferencesof relevantpoliticaland social actorsand the
unique structuralcharacteristicsof society and governmentthat constitute
constraintsand opportunitiesforthese actors,all of which leads to one or an-
other political outcome. An explanation for why some states and not others
underreactto structural-systemic incentives,therefore, cannot ignorethe strat-
egies of eitherthose who are more or less interested preservingthe state or
in
those who, placing othervalues higher,are unwilling to defend the state and
may even seek to overthrowor otherwise destroyit.
Because the probabilitythat a state will balance is a functionof the prefer-
ences of politicalelitesand social groups,underreactionsto dangerous shiftsin
relative power may arise for one of two reasons: actors' preferences,which
may be more influencedby domesticthan internationalconcerns,do not create
incentives to adopt a balancing policy (even when these same power shifts
would cause most otheractors to adopt a prudentbalancing strategy),or the
26. See E.H. Carr,TheTwentyYears'Crisis,1919-1939: An Introduction totheStudyofInternationalRe-

lations(New York:Harper and Row, 1964 [19391),pp. 80-82. See also William C. Wohlforth,"The
Stabilityof a Unipolar World," International
Security,Vol. 24, No. 1 (Summer 1999), pp. 5-41.

potentialdomesticpoliticalrisksand costs ofbalancing are too high.27The first

reason concernsthe "willingness"of actorsto balance; the second, the "ability"
of actorsto balance given the political and materialhurdles thatmust be over-
come to do so.
This kind of analytic framework-one whose architectureis built on elite
calculationsof cost and risk-emphasizes the point thatstatecraft is not simply
a functionof the particular geostrategicrisks and opportunitiespresented
by a given systemic environment,that is, of objective material factors at
the structural-systemic level of analysis. Statecraftis also a consequence of
(1) elites' preferencesand perceptionsof the externalenvironment,(2) which
elites' preferencesand perceptions "matter" in the policymaking process,
(3) the domestic political risks associated with certainforeignpolicy choices,
and (4) the variable risk-takingpropensities of national elites. Once these
unit-levelfactorshave been established, they can then be treated as inputs
(state strategiesand preferences)at the structural-systemic level to explain
how unit- and structural-levelcauses interact to produce systemic out-
comes.
To be sure, thereare many factorsthatmightincrease the domestic political
risks of balancing behavior and raise obstacles to resource extraction.Never-
theless, I posit four unit-level variables that are comprehensive enough to
explain variationacross space and time in state responses to threats:elite con-
sensus, governmentor regime vulnerability,social cohesion, and elite cohe-
sion. Elite consensus and cohesion primarilyaffectthe state's willingness to
balance, whereas government/regimevulnerabilityand social cohesion in-
fluencethe state's abilityto extractresources forthis task. The combinationof
these fourvariables determinesthe degree of state coherence.
Unlike standard balance of power theoryas articulatedby Waltz and other
structuralrealists,in which states respond in a timelyand systematicway to
dangerous changes in relative power, the theory proposed here presents a
more elaborate causal chain of how policy adjustmentsto changes in relative
power occurs:
Changes in relative power elite consensus about the nature of the threat
-- 4 mobilization hurdles as a functionof re-
and the degree of elite cohesion
gime vulnerabilityand social cohesion -4 continuityor change in foreignpol-
icy (i.e., balancing, bandwagoning, appeasement, half measures, etc.)
27. I am gratefulto Alan Lamborn forpointingthis out.

Security29:2 1170
International
For incoherentstates,the causal scheme thatproduces underbalancingmay

be the same as the one above, but the most logical sequence is:
Rise of an externalthreat - social fragmentation- governmentor regime

vulnerability - elite fragmentation- elite disagreement about how to re-
spond to the threator elite consensus not to balance - underbalancing
A discussion of each of the variables serves to clarifyhow these causal

chains work and why.
ELITE CONSENSUS
Elite consensus/disagreementis the most proximatecause of a state's response
or nonresponseto externalthreats.To say thisis to acknowledge thatstatesdo
not make policy; governmentsthroughtheirleaders do. Thus, elite consensus
is the dependent variable: when there is a consensus among policymaking
elites to balance, the statewill do so. The only questions are (1) against whom
will it balance, and (2) will mobilization hurdles created by social fragmenta-
tion and regimevulnerabilitylimitthe state's abilityto meet the threat?In con-
trast,when a consensus to balance is absent, the state will pursue some other
policy-one thatmay or may not be coherent,one thatmay arise throughcare-
ful deliberationand political bargainingor simply by default.
At its core, elite consensus concerns the degree of shared perceptionabout
some factsin the world as being problems (vs. not) of a particularnature (vs.
some othernature)requiringcertainremedies (vs. others).28It is a conceptthat
takes into account the intuitivelyunderstood but too often ignored idea
thatthe process of problem construction(or representation)is a subjectiveone
thatis only partlydeterminedby objectivefacts.29Specifically,elite consensus
is a measure of the similarityof elites' preferencesover outcomes and theirbe-
liefs about the preferencesand anticipated actions of others.It is also a func-
tion of the strategicsetting,which includes the perceived actions available to
the actors and the informationstructureof the environment,that is, what the
28. For a similardiscussion on collectiveaction problems and institutionaldesign, see Alexander

Wendt,"Driving with the Rearview Mirror:On the Rational Science of InstitutionalDesign," Inter-
nationalOrganization,Vol. 55, No. 4 (Autumn 2001), pp. 1023-1024.
29. See Donald A. Sylvan and JamesE Voss, eds., ProblemRepresentation in ForeignPolicyDecision
Making(New York:Cambridge UniversityPress, 1998). Fromthe perspectiveof criticalhistory,the
classic statementof this kind of explanation and understandingis R.G. Collingwood, The Idea of
History(New York: Oxford UniversityPress, 1956), especially p. 283.
30. David A. Lake and RobertPowell, eds., StrategicChoiceand International Relations(Princeton,
N.J.:PrincetonUniversityPress, 1999).

actors can know forcertainand what theymust inferfromothers' behavior.30

Obviously,the more ambiguous the informationavailable to the actors about
theirenvironment,the less one would expect thereto be elite consensus about
policy preferencesand strategies.Under such circumstances,pronounced pol-
icy splits will likelyemerge withinthe regime (and society) thatplace intense
political constraintson decisive and costly governmentactions. Conversely,
the less ambiguous the strategicenvironment,themore likelythatelite consen-
sus will emerge around bold policy initiatives.
The key questions withrespectto eliteconsensus and balancing behavior are
(1) do policy elites agree that there is an external threat?(2) do they agree
about the natureand extentof the threat?(3) do theyagree about which policy
remedywill be most effectiveand appropriateto deal with a threatand protect
the state's strategicinterests?31 and (4) do theyagree on the domestic political
risksand costs associated with the range of policy options to balance a threat?
Variationin elite consensus on these centralissues is a functionof the "mix of
internationaland domestic incentives attached to differentoptions, actors'
risk-takingpreferences,theirtime horizons,and how they discount costs and
benefits."32
Balancingbehavior requiresthe existenceof a strongconsensus among elites
that an externalthreatexists and must be checked by eitherarms or allies or
both. As theproximatecausal variable in themodel, elite consensus is the most
necessary of necessary causes of balancing behavior. Thus, when thereis no
elite consensus, the prediction is either underbalancing or some other non-
balancing policy option.Developing such a consensus is difficult, however,be-
cause balancing, unlike expansion, is not a behavior motivated by the search
forgains and profit.It is instead a strategythatentails significantcosts in hu-
man and materialresourcesthat could be directedtoward domestic programs
and investmentratherthan national defense. In addition, when alliances are
formed,the state must sacrificesome measure of its autonomy in foreignand
militarypolicy to its allies. In the absence of a clear majorityof elites in favorof
a balancing strategy,therefore,an alternativepolicy,and not necessarilya co-
herent one, will prevail. This is because a weak grand strategycan be sup-
ported for many differentreasons (e.g., pacifism, isolationism, pro-enemy
31. For adjustmentfailuresthattake the formof eitheroverly cooperative or overly competitive

behaviors in response to rapidly changing strategicenvironments,see Charles A. Kupchan, The
VulnerabilityofEmpire(Ithaca, N.Y.: Cornell UniversityPress, 1994).
32. Alan C. Lamborn, "Theory and the Politics in World Politics,"InternationalStudiesQuarterly,
Vol. 41, No. 2 (June1997), p. 212.

International
Security29:2 172
sympathies,collectivesecurity,a belief in conciliation,etc.). Consequently,ap-

peasement and otherformsof underbalancingwill tend to triumphin the ab-
sence of a determinedand broad political consensus to balance simplybecause
these policies representthe path of least domestic resistanceand can appeal to
a broad range of interestsalong the political spectrum.Thus, underreactingto
threats,unlike an effectivebalancing strategy,does not requireoverwhelming,
united, and coherentsupport fromelites and masses; it is a default strategy.
As the foregoingsuggests,democracies will be particularlyslow to balance
against threats.This is because (1) nonbalancingbehavior is (or should be) the
status quo policy prior to the emergence of a dangerous threat,and (2) there
are typicallymany "veto players" in a democraticpolicymakingprocess, that
is, "individual or collectiveactors whose agreementis necessaryfora change
in the status quo.""' As George Tsebelis explains, "Significantdeparturesfrom
the status quo are impossible when the winset is small-that is, when veto
players are many-when they have significantideological distances among
them,and when theyare internallycohesive.""34 It follows,therefore,that,for
democracies to move away fromthe nonbalancing status quo to a costlybal-
ancing strategy,it oftentakes a dramaticevent to expand the "winset" forpol-
icy change. Thus, as Condoleezza Rice recentlytestifiedbeforethe September
11 Commission:
Historically,democraticsocieties have been slow to reactto gatheringthreats,

tendinginstead to wait to confrontthreatsuntil theyare too dangerous to ig-
nore or untilit is too late. Despite the sinkingof the Lusitaniain 1915 and con-
tinued German harassmentof American shipping, the United States did not
enter the FirstWorld War until two years later. Despite Nazi Germany's re-
peated violations of the Versailles Treaty and its string of provocations
throughoutthe mid-1930s,the Westerndemocracies did not take action until
1939. The U.S. governmentdid not act against the growingthreatfromImpe-
rial Japan until the threatbecame all too evident at Pearl Harbor. And, tragi-
cally, for all the language of war spoken before September 11, this country
simply was not on a war footing.35
33. George Tsebelis,VetoPlayers:How PoliticalinstitutionsWork(Princeton,N.J.:PrincetonUniver-

sity Press, 2002), p. 19.
34. Ibid., p. 2.
35. National Security Adviser Condoleezza Rice, testimonybefore the Independent National
Commission on TerroristAttacksupon the United States, April 8, 2004. Rice mighthave added
that the Korean War was the dramaticevent that triggeredthe United States to ratchetup its de-
fensespending at the outsetof theCold War.See RobertJervis,"The Impact of the Korean War on
the Cold War,"JournalofConflictResolution,Vol. 34, No. 4 (December 1980), pp. 563-592.

In summary,balancing behavior requires elite consensus thatthe state con-

frontsa serious threatthatmust be checked by internalor externalmeans; it is
a necessary condition forbalancing behavior.-"All other scenarios regarding
elite consensus will result in eitherunderbalancing or the adoption of other
nonbalancing strategies.States are not expected to balance against threats
when thereis (1) significantelite disagreementin termsof threatperception;
(2) elite consensus that a threatexists but disagreementover the appropriate
remedy(e.g., whetherto appease or stand firm);or (3) elite consensus to adopt
other policy options such as appeasement, bandwagoning, buck-passing, or
bilateral or multilateralbinding strategies.
GOVERNMENT/REGIME VULNERABILITY
In its most basic sense, the concept of governmentor regime vulnerability
"asks what is the likelihood thatthe currentleadership will be removed from
political office.""' Specifically,do the governingelites face a serious challenge
fromthe military,opposing politicalparties,or otherpowerfulpoliticalgroups
in society? Are such groups threateningto prematurelyremove the current
leaders fromoffice?Have they done so in the recentpast?
In a related but more general sense, the concept of governmentor regime
vulnerabilityseeks to capturethe relationshipbetween rulersand ruled at any
given moment. Hence, the following questions related to elite-masslinkages
are also relevant:(1) is the government'sauthoritybased primarilyon coercion
or is it self-legitimatingin the eyes of the public? (2) is thegovernmentmeeting
36. It is not a sufficientconditionforbalancing,however. Even in countrieswhere thereis a mod-

erate to high level of elite consensus that(1) a particularstate poses the most dangerous threatto
the state's survival and (2) its militarypower must be checked by building arms and formingalli-
ances, the policy may stillbe thwartedbecause of a disjuncturebetween elites' policy preferences
and those of the society.A good example is the Hoare-Laval debacle in response to Italy's attackon
Ethiopia in 1935. The problemwas thatBritishand Frenchelites wanted to build and bolstertheir
alliance with Italy against Germany,but theirmore liberal societies demanded punitiveLeague of
Nations actions in response to Italy's unprovoked attack against a weak Africancountry.In the
end, the Anglo-Frenchpublics got theirway, as the League of Nations imposed sanctionson Italy.
French and Britishparticipationin these collective securitymeasures quickly drove the Italian
leader, Benito Mussolini, into Adolf Hitler's orbitand, in the process, dashed a key element (i.e.,
maintainingthe Italian check against German expansion in Austria and the Balkans) in the grand
strategiesof Britainand France.
37. Joe D. Hagan, "Regimes, Political Oppositions, and the Comparative Analysis of Foreign Pol-
icy,"in Charles F. Hermann, Charles W. Kegley Jr.,and JamesN. Rosenau, eds., New Directionsin
theStudyofForeignPolicy(Boston:Allen and Unwin, 1987), p. 346. 1 define"regime" in the conven-
tional way as the permanentinstitutionswithin the state, whereas "government"refersto the
group of people occupying the importantpositions in the regime.

International
Security29:2 174
the expectations of the people? (3) does it enjoy broad support from the
masses? and (4) can it minimizedomesticinterference in its policy decisions?
These questions go to the heartof a government'seffectivenessand political
authorityand the trade-offbetween external securityand internal stability.
Leaders, especially vulnerable ones, cannot simply choose securitypolicies
based on theirlikelihood of neutralizingthe externalthreator satisfyingna-
tional ambitionsforgreaterpower and influence.They must also consider the
domestic costs attached to the policy options. Vulnerableleaders will typically
be more constrainedthan popular ones, and theywill be less effectivein mobi-
lizing resources fromsociety.As James Morrow observes: "Leaders and do-
mestic groups often disagree about the appropriate response to a threat.
Leaders choose policies fortheirabilityto countera threatand to provide do-
mesticsupport.Withoutthe latter,securitypolicies will failto do the former."38
In otherwords, the strengthof elite-masslinkages plays a major role in de-
terminingthe range and viability of policy choices available to decision-
makers. Weak governmentshave less policy capacity than do legitimateones;
thatis, theyare less able to detectand assess threats;to control,mobilize, and
allocate national resources;to articulateand choose policies; and to implement
those policies. On this point, Edward Azar and Chung-in Moon write:
Weak legitimacyexhausts domestic capabilities and turns the overall policy
capacity rigid and ineffectual.As a regime attempts to ensure its survival
throughthe use of forceor by co-optationand appeasement of the opposition,
it begins rapidly to deplete the nation's capabilities and scarce resources....
Among the serious side-effects of the erosion of domestic policy capacity are: a
sharp decrease in the people's loyaltyand conformity to, and compliance with,
governmentpolicies and decisions; a decrease in the government'sabilityto
mobilize materialand human resourcesin the event of [a] national securitycri-
sis; diminutionin the accountabilityor the virtual terminationof the steering
functionof the government;and a serious reductionin the co-ordinationand
implementationof policies.39
Moreover, vulnerable rulers will be especially wary of fomentinghyper-
nationalism and mobilizing a mass army because theyfear,forgood reasons,
thatweapons put in the hands of a newly energized,nationalistpublic are just
38. JamesD. Morrow,"Arms versus Allies: Trade-Offsin the Search forSecurity,"InternationalOr-

ganization,Vol. 47, No. 2 (Spring 1993), p. 216.
39. Edward E. Azar and Chung-inMoon, "Legitimacy,Integration,and Policy Capacity: The 'Soft-
ware' Side of ThirdWorldNational Security,"in Azar and Moon, eds., NationalSecurityin theThird
World:TheManagementofInternaland ExternalThreats(College Park, Md.: CenterforInternational
Development and ConflictManagement,Universityof Maryland, 1988), p. 84.

as likely,ifnot more so, to be firedat themas theyare to be used against the ex-
ternalthreat."4Illegitimaterulerswill also be less able than legitimateones to
persuade the mass public of the existenceof national securitythreatsand of
the need to make sacrificesto resistthem.All ofthese factorslead one to expect
that vulnerable regimes will be more likely to search for allies than to build
arms to counterexternalthreats,because the formeris a quick and relatively
low-cost method to counterbalancea dangerous threat.
SOCIAL COHESION
Social cohesion and its opposite, social fragmentation,describe the relative
strengthof ties thatbind individuals and groups to the core of a given society.
Social cohesion does not mean politicalunanimityor the absence of deep polit-
ical disagreementswithinsociety.All societies exhibitnormal conflictsarising
fromvarious sources and cleavages, including divergentclass interests,eco-
nomic inequalities,competingpolitical goals, ethnicanimosities,and so-called
normativeconflicts(e.g., differencesover the definitionofnational identity,the
relationshipbetween religionand the state,culturewars, etc.).4'
The key to social cohesion is that all members of society accept the same
rules of the game; thatis, theysupport the society's institutionsas legitimate
and appropriate mechanisms to settle disputes among them no matterhow
profoundtheirdisagreementsor grievances (i.e., loyal opposition). Dangerous
political divisions exist when groups within society do not conferlegitimacy
on the institutionsthat structureit and, even more so, when a significant
segment of the population intends to overthrow the state (i.e., disloyal
opposition).
In its widest sense, the concept of social cohesion encompasses more than
mere institutionallegitimacy;it is about a psychological feelingof solidarity
40. This argumentis widely used to explain the absence of mass armies in the thirdworld. For the
claim thatweak states led by illegitimateregimescannot affordthe internalrisks posed by mass
armies, see Alexander Wendt and Michael N. Barnett,"Dependent State Formation and Third
WorldMilitarization,"ReviewofInternational Studies,Vol. 19 (1993), pp. 321-347; and Steven R. Da-
vid, "The Primacyof InternalWar," in Neuman, International Relationsand theThirdWorld,p. 88.
For a related argumentwith respectto China, see Erika StreckerDowns and Phillip C. Saunders,
"Legitimacyand the Limits of Nationalism: China and the Diaoyu Islands," International Security,
Vol. 23, No. 3 (Winter1998/99), pp. 114-146.
41. See VolkerThen, "Introduction,"in Peter L. Berger,ed., The Limitsof Social Cohesion:Conflict
and Mediationin PluralistSocieties,a ReportoftheBertelsmann Foundationto theClub ofRome(Boulder,
Colo.: Westview,1998), p. xvi. Peter Berger writes: "Essentially the notion of 'normative order'
means the way in which any human group triesto answer two fundamentalquestions: Whoare
we? And, How arewe tolivetogether?" Berger,"Conclusion," in ibid., p. 355 (emphasis in original).

International
Security29:2 176
within a society.When all members of society "feel interconnectedand inte-

grated into the vast series of networksthatmake up society,good social cohe-
sion is likely.As soon as one group feels excluded fromthe society in which
that group is nonethelessformallypresent,social division is likely."42
A common themein internationalrelationsis thatthe emergenceof a serious
externalthreatcauses an increase in social cohesion.43This is the logic that
drives theoriesof diversionarywars, also known as "scapegoat wars," and the
notionthatwartimeleaders enjoy a "rallyaround the flageffect."44These theo-
ries assume that,in the presence of an outside threat,divided or dissatisfied
citizenswill put aside theirconflictswith one anotheror theirfrustration with
the rulingregimeand uniteto repel the invader or,when invasion is not the is-
sue, defeat the externalthreatto theircore values or othernational interests.
Thus, ErnstHaas and Allen Whitingassert: "In timesof extremedomestic ten-
sion among elites,a policy of unitinga badly divided nation against some real
or alleged outside threatfrequentlyseems useful to a ruling group."'4
All otherthingsbeing equal, the emergenceof an outside threatshould in-
crease social cohesion. How much of an increase in national unity depends,
interalia, on the perceived strategicobjectives of the aggressorand the likeli-
hood thatthe pooled effortsof the nation will lead to victory.The more wide-
spread the belief among the target'spopulation thatthe invader will establish
a brutallyrepressiveoccupation,forinstance,the more likelytherewill be high
levels of civilian resistanceand social cohesion in response to the threat.Fur-
thermore,the higherthe probabilityof success in war, the more likelythe tar-
42. Jacques Semelin, UnarmedagainstHitler:CivilianResistancein Europe,1939-1943, trans.Suzan

Husserl-Kapit (Westport,Conn.: Praeger,1993), pp. 64-65.
43. The classic statementsof the group-bindingfunctionsof conflictare Georg Simmel, Conflict,
trans. Kurt H. Wolff(Glencoe, Ill.: Free Press, 1955); and Lewis A. Coser, The Functionsof Social
Conflict(New York: Free Press, 1956).
44. For comprehensivereviews of works on this subject,see ArthurA. Stein,"Conflictand Cohe-
sion: A Review of the Literature,"Journalof ConflictResolution,Vol. 20, No. 1 (March 1976),
pp. 143-172; and Jack S. Levy, "The Diversionary Theory of War: A Critique," in Manus I.
Midlarsky,ed., TheHandbookofWarStudies(Boston: Unwin Hyman, 1989),pp. 259-288. For studies
thatembrace the propositionthatexternalthreatincreases internalcohesion, see Daniel S. Geller,
DomesticFactorsin ForeignPolicy(Cambridge, Mass.: Schenkman,1985); and Bruce Russett,"Eco-
nomic Decline, ElectoralPressure,and the Initiationof InterstateConflict,"in Charles S. Gochman
and Alan Ned Sabrosky,eds., Prisonersof War?Nation-Statesin theModernEra (Lexington,Mass.:
LexingtonBooks, 1990), chap. 9.
45. Ernst Haas and Allen Whiting,DynamicsofInternational Relations(New York: McGraw-Hill,
1956), p. 62. See also Levy,"The DiversionaryTheoryof War"; Richard Rosecrance,Actionand Re-
actionin WorldPolitics(Boston:Little,Brown,1963), p. 294; and Michael Stohl, Warand DomesticPo-
liticalViolence:The AmericanCapacityforRepressionand Reaction(BeverlyHills, Calif.: Sage, 1976),
pp. 19-20, chap. 3.

get's population will cohere and voluntarilymobilize to put down the threat.
Nevertheless,historyrecords many instances of externalthreatscausing the
opposite effect,namely,increased in-group conflictand disintegration.What
might account for this variation? The theories of Georg Simmel and Lewis
Coser on the effectsof externalconflicton in-groupcohesion provide answers
to this question.
According to Simmel, the primaryfunctionof social conflictis to establish
and maintain the identityand boundary lines of societies and groups. He
furthersuggests thatexternalconflictstrengthensinternalgroup cohesion and
increases its centralization:"The group in a state of peace can permitantago-
nisticmemberswithinit to live with one anotherin an undecided situationbe-
cause each of them can go his own way and can avoid collisions. A state of
conflict,however,pulls the members so tightlytogetherand subjects them to
such uniformimpulse thattheyeithermust get completelyalong with,or com-
pletelyrepel,one another.This is the reason why war with the outside is some-
times the last chance for a state ridden with inner antagonisms to overcome
these antagonisms,or else to break up definitely.""4
Likewise, Coser asserts: "Internal cohesion is likely to be increased in the
group which engages in outside conflict."47 The presence or creationof a hos-
tile out-group,by supplyinga targetof aggression,enhances in-groupintegra-
tionor cohesion (and sometimesboth) by reinforcing common values, interests
(the most basic of which is each member's interestin self-preservation and the
survival of the group on which thatdepends), and anythingelse thatserves to
bind the group together.In this way, intergroupconflictimposes a degree of
harmonyupon otherwiseconflictingelements within the group.
Because members need only perceive or be made to perceive an external
threatto pull themselves together,Simmel suggests that the search for ene-
mies, even theirfabricationwhen none exist,is a rational strategyto maintain
or enhance group cohesion. This logic leads Simmel to observe: "A group's
completevictoryover its enemies is thus not always fortunate.... Victorylow-
ers the energy which guarantees the unity of the group; and the dissolving
forces,which are always at work,gain hold.""' This insightnicelysupports the
age-old proposition in internationalpolitics that,once the common enemy is
defeated, the alliance breaks apart.
46. Simmel, Conflict,

pp. 87-88.
47. Coser, The Functionsof Social Conflict,
p. 92.
48. Simmel, Conflict,
pp. 97-98.

International
Security29:2 178
Unlike Simmel, however, Coser emphasizes that the group-bindingfunc-

tions of conflictare conditionedby the precrisisdegree of group cohesion and
the uniformity of threatperceptionacross groups withinsociety:"The relation
between outer conflictand inner cohesion does not hold true where internal
cohesion beforethe outbreakof conflictis so low thatthe group membershave
ceased to regard preservationof the group as worthwhile,or actually see the
outside threatto concern'them' ratherthan 'us.' In such cases disintegrationof
the group, ratherthan an increase in cohesion, will be the result of outside
conflict.'"49
Coser also argues that externalthreatswill produce increased group cohe-
sion only when theyare seen as a menace to the entiregroup. Thus, according
to Coser, the degree of precrisisgroup cohesion and the general perception
among group members regardingthe scope of the threatare key intervening
variables that determine whether a threatenedgroup will unite or disinte-
grate."' There are many intuitivereasons why this should be so. Wars to repel
determinedinvaders,even when successful,typicallyrequirelarge sacrificesin
blood and treasurefromthe population. If one defends only what one holds
dear and is willing to sufferhigh costs and take greatrisksonly forthingsthat
one values, then groups thatperceive themselvesto be excluded fromsociety
priorto the crisiswill not sense thattheyhave much to lose by siding withthe
49. Coser, TheFunctionsofSocial Conflict, p. 93. Similarly,Coser writes:"The degree of group con-
sensus prior to the outbreakof the conflictseems to be the most importantfactoraffectingcohe-
sion. If a group is lacking in basic consensus, outside threatleads not to increased cohesion,but to
general apathy,and the group is consequently threatenedwith disintegration."Ibid., pp. 92-93.
See also Levy,"The DiversionaryTheoryofWar,"p. 272. Unfortunately, as ArthurSteinpointsout,
the "outgroup-conflict-increases-internal-cohesion" idea "has been called a 'ubiquitous principle'
and 'Coser's strongesttheme'. . . . None of Coser's qualificationsand necessary conditions are
mentioned, nor is the counter-hypothesiselucidated. As [R.W. Mack] points out, 'Coser's The
FunctionsofSocialConflictruns the riskof being accorded thatpeculiar formof academic obeisance
in which a work is cited by everyone and heeded by no one."' Stein, "Conflictand Cohesion,"
p. 145.
50. For exceptional studies thatcorrectly employ Coser's hypothesis,see ArthurA. Stein,TheNa-
tionat War(Baltimore,Md.: JohnsHopkins UniversityPress, 1978); Christensen,UsefulAdversaries;
and Michael C. Desch, CivilianControloftheMilitary:TheChangingSecurityEnvironment (Baltimore,
Md.: Johns Hopkins UniversityPress, 1999). For a critique of the "external-conflict-produces-
internal-cohesion"logic thatinformsthe scapegoat theoryof war, see GeoffreyBlainey,TheCauses
ofWar(New York:Free Press, 1973),chap. 5. It is undeniable thatembattledleaders have oftenbe-
lieved thatan externalthreatwould increase internalcohesion, and so theycreated threatswhere
none existed in the hope of gaining domestic unity and support for their regime. Yet, as Stein
wisely pointsout,thismisguided and riskystrategyusually backfires:"Politicalleaders who count
on foreignadventuresto unifytheircountryand cementtheirpositions should thinkagain. Manu-
facturingcrises may enable leaders to expand the powers of the state but, in one manifestationor
another,cohesion will surelydecrease. Only if thereis a perceived threatwill cohesion increase."
Stein, The Nationat War,p. 87.

aggressor.Indeed, disenfranchisedgroups (e.g., disloyal and semiloyalopposi-

tion) may even perceive thattheyhave much to gain fromcollaborationwith
the enemy.A couple of propositionsarise fromthis line of thought.First,the
deeper the divisions withinsocietypriorto the threat,the more likelya part of
the communitywill eitheractivelycollaborate with the enemy or remain pas-
sive ratherthan resist the aggressor.Second, the deeper the social divisions
withinthe state,the greaterthe resistanceto militarymobilizationagainst the
threat.
Problems of loyaltyand social cohesion typicallyarise within "oversized"
states-those thatdefinethemselveson an ethnicratherthan a civic basis (i.e.,
theychoose the option of nationbuilding over statebuilding),but whose terri-
torial boundaries do not correspond closely to an ethnicallydefined nation.
Oversized states are, by definition,ethnicallystratifiedsocieties,wherein cer-
tain minoritygroups are highlydiscriminatedagainst by the state and conse-
quentlyview themselvesas captive populations. Because the national identity
does not apply to all of the citizens of the multinationalstate,it does not fur-
nish a sufficientfoundationforstate loyalty.Instead, social fragmentationand
in-grouprigidityweaken the political authorityof the state,oftento the point
where it will be unable or unwilling to balance against externalthreats.In ex-
tremecases, ethnonationalistgroups may advocate thatthe state relinquishits
sovereign independence in favor of bandwagoning with a larger coethnic
neighbor.51When subjugated minoritygroups oppose the state in principle
and assume the role of disloyal opposition, the state and the general public
confronta Trojan horse dilemma: will ethnic soldiers with kin in an enemy
countryturn theirguns against theirfellow soldiers and citizens or fightfor
theircountry?52
The precrisisdegree of social cohesion withinthe threatenedstate may also
explain its choice of arms or allies to deal with the threat-that is, whetherthe
targetwill be more likely to respond with an internalor externalbalancing
strategy.Internalbalancing entails greaterand more immediate sacrificesfrom
the general population than the alternativeof externalbalancing,by which the
state gives up a measure of foreignpolicy autonomyto shifta part or all of the
51. Mark R. Thompson, "Building Nations and CraftingDemocracies-Competing Legitimacies

in InterwarEastern Europe," in Dirk Berg-Schlosserand JeremyMitchell,eds., Authoritarianism
and Democracyin Europe,1919-39: ComparativeAnalyses(New York: Palgrave Macmillan, 2002),
pp. 20-38.
52. Alon Peled, A QuestionofLoyalty:MilitaryManpowerPolicyin MultiethnicStates(Ithaca, N.Y.:
Cornell UniversityPress, 1998).

International
Security29:2 180
burden ofbalancing the externalthreaton to anotherstate.Thus, it follows that

the more social divisions withina state,the more likelyit will be forcedto rely
on externalmeans (alliances) as opposed to internalmeans (the mobilizationof
arms and troops) to balance against the threat.Conversely,the greaterthe de-
gree of social cohesion in the precrisisperiod, the more likelythe state will be
able to resistan attackon its own, or,at the veryleast, the more likelyinternal
balancing will be a viable option forthe state.
In summary,Coser claims that the relationshipbetween outer conflictand
inner cohesion holds only when there is some minimal consensus that the
group is a "going concern,"whose preservationas an entityis worthwhile,
when the outside threatis seen as a danger to the group as a whole, not just
a part of it, and when thereis a reasonable chance that,if the group unites,
the enemycan be defeated.Thus, Coser's logic yields two general predictions:
(1) stateswith high levels of political and social integrationwill be most likely
to balance against externalthreatsand to experiencean increasein the public's
support and compliance with the regime; and, conversely,(2) fragmented
states will underreactto dangerous threats(i.e., responding with inefficient
balancing,bandwagoning,buck-passing,distancing,incoherenthalfmeasures,
or ineffectivepolicies defined by the lowest common denominator) because
the threatwill not be perceived as a universal one by the masses, and leaders
will view the politicalrisksof an arms buildup and internalmobilizationcam-
paign as intolerablyhigh.Moreover,severelyfragmentedstatesin the precrisis
period can be expected to furtherdisintegrateand even crack under external
pressure(i.e., surrenderto the enemyor undergo a regimechange,civil war, or
revolution).
ELITE COHESION
Elite cohesion concerns the degree to which a central government'spolitical
leadership is fragmentedby persistentinternal divisions. Elite polarization
may arise over ideological, cultural,or religious divisions; bureaucraticinter-
ests; party factions;regional and sectoral interests;or ethnicgroup and class
The concept of elite cohesion is a continuous variable. At one ex-
loyalties."53
treme,political elites are divided into two armed camps, with hypernational-
ists on one side and disloyal collaboratorswith the enemy on the other.It is a
situationdevoid of politics,forthereis no room forbargainingamong the po-
liticalfactionsto reach compromisesettlements.At the otherextreme,all polit-
53. Hagan, "Regimes, Political Oppositions, and the Comparative Analysis of Foreign Policy,"
p. 344.

ical elites and groups belong to a dominant party"and theyuniformlyprofess

its ideology, religious belief, or ethnonationalistcreed-an 'ideocratic' con-
figurationthat is primarilycoerced."54 In practice, the structureof political
elites within most states falls somewhere between these two ideal types.
There are five relevant questions regarding elite cohesion and balancing
behavior. First,is therea struggleamong elites for domestic political power?
Second, ifso, are thereopportunisticelites withinthe threatenedstatewho are
willing to collaboratewith the enemy to advance theirown personal power or
to gain office?Third,if thereare multiplethreats,do elites agree on theirrank-
ings of externalthreatsfrommost to least dangerous to the state's survival
and vital interests?Fourth,are theredeep disagreementsamong elites regard-
ing the question: with whom should the state align? Fifth,are elites divided
over the issue of whetherto devote scarce resourcesto defend interestsin the
peripheryor the core?
When elites are fragmented,it is highlyunlikelythatthe statewill be able to
constructa coherentand effectivebalancing strategy.Some elites will want to
balance against one threat,whereas others will want to balance against an-
other; some will want to invest resources to defend interestsin the core, yet
others will want to protectinterestsin the periphery.Moreover,the political
costs and policy risksof balancing increasewhen elites are fragmented.On the
one hand, any firmdecision will be publicly criticizedby opposition elites as
too costlyand misguided. On theotherhand, bargainingefforts to appease op-
position groups and therebygain policy consensus will typicallyresultin inco-
herenthalf measures, in which some parts of the state's balancing strategy
contradictotherpartsof it,such thatthe riskof policy failureincreasesdramat-
ically.Hence, effectivebalancing behavior is most likelywhen elites agree on
the targetofbalancing (i.e., which statepresentsthe greatestthreat)and where
best to devote scarce militaryresources.55
EliteFragmentation
and Overexpansion
versusUnderbalancing
The argumentpresented here-that elite fragmentationcontributesto under-

balancing-appears to contradictJackSnyder's claim that elite fragmentation
54. Mattei Dogan and JohnHigley, "Elites, Crises, and Regimes in Comparative Analysis," in
Dogan and Higley,eds., Elites,Crises,and theOriginsof Regimes(Lanham, Md.: Rowman and Lit-
tlefield,1998), p. 18. See also JaroslawPiekalkiewicz and AlfredWayne Penn, PoliticsofIdeocracy
(Albany: State Universityof New York Press, 1995).
55. Among democracies,thereshould be a strongrelationshipbetween the degree of social cohe-
sion and that of elite cohesion; the lattermirroringthe former.

International
Security29:2 182
leads to overexpansionistgrand strategies.In his view, governmentsthatrep-

resenta narrow set of societal interestshave incentivesto cater to the military
(by spending more on defense) and provoke rally effects(by overinflatingin-
ternationalthreats).When elites are disunited,therefore, the resultis not paral-
ysis but ratherexpansionistlogrolls.5"
Two points are essential here. First,overexpansion and overbalancing are
differenttypes of behaviors, and so it is logical that elite fragmentationcan
cause both overexpansion and underbalancing. Expansion is motivated by
greed and profit;it is a power-maximizing,predatorypolicy to revise the exist-
ing order. Balancing behavior,in contrast,is motivated by the desire to keep
what one has; it is a security-maximizing policy to avoid losses and maintain
the status quo."7 Expansion is a relativelyeasy sell because it promises to in-
crease the size of the pie and make everyone betteroff.Balancing is a costly
policy thatwill, at best,maintainthe existingpie and, more likely,dramatically
shrink it for most citizens, because it requires ratchetingup taxation and
switching from butter to guns. Thus, unless the target is unambiguously
known to be an insatiablepredator,many elites and the public will see balanc-
ing as a self-defeating policy thatrisksunnecessaryarms racingand possibly a
bloody war.
Second, Snyder's theory is underspecified and pertains only to constitu-
tional oligarchies, not emerging democracies. A primary determinantfor
whetherelite fragmentationleads to overexpansion or underbalancingis the
patternof democratictransition,thatis: (1) whetherthe transitionfroma non-
democraticregime(eithertotalitarianor authoritarian)is to a constitutional(or
liberal) oligarchyor to a restrictedor fulldemocracy;and (2) whetherthe tran-
sition to democracy drasticallychanges the underlyingconstellationof con-
tending political and social forces.
Overexpansion throughelite logrollingcan occur only when a majorityof
elite groups favor expansion in differentgeographic areas. These are excep-
tional cases, however. Typically,a bargainingprocess among competing elite
groups will generate blocking coalitions, resulting in policy inertia or the
adoption of unfocused half measures or both."8In general, such policies offer
elites the benefitof low domestic costs (e.g., no increased taxes) but carrythe
risk of policy failure(namely,underreactingto threatsand opportunities).
In the transitionfroman authoritarianor a totalitarianregimeto a liberal or
56. See Snyder,MythsofEmpire,pp. 6-7.

57. See Schweller,Deadly Imbalances,pp. 19-26.
58. See Tsebelis, VetoPlayers.

constitutionaloligarchy,the dominantclasses and the old rulingelitesstillcon-

trol the state apparatus, but they must now compete for mass allies against
newly mobilized groups. In thishighlypoliticized atmosphereof mass politics,
elites will be tempted, in varying degrees, to mobilize the masses through
nationalistpropaganda designed to whip up support forpredatoryexpansion.
These jingoisticappeals to militarismand nationalism,when combined with
a weak central authority,an entrenched military,unstable coalitions, and
logrolling-typebargaining among elites, unleash an impassioned hyper-
nationalism among the mass public, which then becomes an uncontrollable
forceforrecklesslyexpansionistforeignpolicies with a high risk of war.
This is the theoryput forthby Edward Mansfieldand JackSnyderto explain
theirstatisticalfindingthat democratizingstates are more likelyto fightwars
than are states thathave not undergone a regime change."sThe problem with
theirdiscussion of historicalcases, however,is thatit focuses on Palmerston's
Britain,Napoleon III's Second Empire,WilhelmineGermany,Imperial Japan,
and present-daySerbia and Croatia. These are not democratizing states but
ratherliberal or constitutionaloligarchies.As DietrichRueschemeyer,Evelyne
Stephens,and JohnStephens point out: "If participationis limitedto a few (as
in mid-nineteenth-century Britain),the regime may be liberal because issues
are openly discussed or because state action is limited by solid individual
rights,but it remains an oligarchicregime;one cannot speak of democracy.If
the state apparatus is not made responsible [to the elected parliament](as in
the Germany of Bismarckand Wilhelm II), the most inclusive system of suf-
to createa 'rule of
frageand the best protectionof civil rightsare not sufficient
the people' in any meaningfulsense."6"
Logrollingamong eliteswithinliberaloligarchiestypicallyengendersoverly
expansionistpolicies because the participantsat thebargainingtable are exclu-
sively or overwhelminglymembers of the old ruling classes (e.g., the landed
elite,the military,the bureaucracies,imperial interestgroups, and other con-
servative elites fromthe dominant classes). This should come as no surprise
because liberal oligarchiesare, by definition,regimesin which the state appa-
ratusis not responsibleto the parliament.Hence, the interestsand views of the
subordinateclasses have littleeffecton the decisionmakingprocess,regardless
of whetherthe regime allows mass participation.6'
59. Mansfield and Snyder,"Democratization and the Danger of War."

60. DietrichRueschemeyer,Evelyne Stephens, and JohnStephens, CapitalistDevelopment and De-
mocracy(Chicago: Universityof Chicago Press, 1992), pp. 43-44.
61. For example, logrollingin WilhelmineGermany,an exemplarycase in Mansfieldand Snyder's
theoreticaldiscussion of how democratizationleads to war, was done by old economic and politi-

Security29:2 184
International
Returningto the cases of WilhelmineGermany,Second Empire France, Im-

perial Japan,and VictorianEngland, the key question is: who was and was not
invitedto thebargainingtable? Political agents of thesubordinateclasses were
conspicuously absent in Germany and France because the liberal movement
could not overcomepreindustrialcleavages withinthe middle classes; as such,
liberal hegemony and the inclusion of the working classes within a competi-
tive political process had yet to be established in these countries.In Imperial
Japan,a council of militaryelder statesmendecided who would govern.As for
Britain,it cannot even be called a democratizingstateuntilthe Second Reform
Act passed in 1867, giving the vote to one in threemales. Indeed, only after
WorldWar I did all male citizenstwenty-oneyears or older and femalesthirty
and over who met residencyrequirementshave the rightto vote. In all these
cases, the power brokerswere the same economic and political elites thathad
representedthe dominantclasses priorto the move toward democracy.Still in
power but threatenedby social change and mass participation,the old elites
triedwith varyingsuccess to create a nationalistfrenzyforexpansion.
Whetherthe masses actuallybought into the elite-drivenmythsof empire is
debatable. History shows that subordinate classes, especially the working
classes, are not so easily duped. That liberal oligarchies have indeed overex-
panded is not evidence thatthe masses actuallysupported any of thebrands of
rash imperialismthatwere being peddled by the rulingelites. Afterall, liberal
oligarchies,by definition,can do any fool thing they choose to without the
public's support (which is not to say thattheywould be wise to do so or have
typicallyacted as iftheywere completelyunconstrainedby theirpublics). Nor
is the argumentsupported, much less proven, by the observation that,once
war came, the masses volunteeredto fight.Presenteda faitaccompli to either
fightor surrender,most citizenswill naturallyrally around the flag and fight;
thisis not proof,however,thatthe median voterwas brainwashed by elitepro-
paganda or that he or she mindlessly supported reckless expansion.
For all these reasons, the claim that democratizingstates are the most war
prone is misleading. What Mansfield and Snyder show, instead, is thatliberal
oligarchies,not emergingdemocracies, exhibitthe peculiar domestic politics
that are most conducive to overexpansion. Their cases are of states that had
adopted some of the formaltrappingsof democracy;the new democraticinsti-
cal elites representingthe established interestsof the Junkers,Industrialists,Navy League, and

Pan-Germangroups. The economic and class interestsof the newly enfranchisedmasses fromthe
workingand middle classes were nowhere to be found.

tutions,however, did not rest on the power of formerlyexcluded classes but

ratheron the dominant influenceof agrarian elites, the military,the govern-
mentbureaucracy,or a hegemonicparty.What, then,are the characteristicdo-
mesticdynamics and externalbehaviors of democratizingstates?
If the dominantelitesremainheavily representedafterthe initialinstallation
of a restrictedor full democracy,the new democratic elites and their allies
among the previously excluded classes will be especially wary of strengthen-
ing the military(e.g., France during the period 1877 to 1898). In these situa-
tions, there will be high levels of elite dissensus, elite fragmentation,and
governmentor even regime instability.
Assuming thatthe primarygoal of most elites is to gain or maintainpolitical
office,intenseinterpartycompetitionwill generateuniversal appeals to subor-
dinate classes based on materialimprovements;that is, butterover guns will
be the most politicallyexpedient choice forboth old and new elites. It can be
furtherexpected that an energized, if not militant,labor movementwill find
naturalallies among new elitesin the liberal party.This labor-liberalunion will
work against defense spending,urgingthe state,instead,to take more respon-
sibilityfor guaranteeing full employment,a more equitable distributionof
goods and services,more equal opportunity,and greatersecurityagainst the
hazards of illness and old age. In cases, such as Latin America,where the mid-
dle class needed the workingclass to succeed in the transitionto democracy,
thereis an even morepowerfulbias against militaryspending and expansionist
foreignpolicies.12
When the middle class is highly fragmentedor hyperfractionalized,how-
ever,democratizationunleashes unmanageable societal and political polariza-
tion.The inabilityof the middle class to consolidate itselfsubvertsany hope of
a liberal stabilization by means of Lib-Labism-the mobilization of labor
throughan alliance with the liberal establishment."6Indeed, the presence or
absence of Lib-Labism is what distinguishesliberal and aliberal democratiza-
tion processes. In democratizingaliberal societies,it is because labor was shut
out of the politicalprocess priorto reformthattrade unions were forcedto cre-
62. In those rare cases where the power of the once-dominantclasses has been so marginalized
thatthe old rulersno longerpose a serious threatto the new regime,pro-democraticelites may be
able to support the militarywithoutfearof a breakdown in the democratizingprocess. Balance of
power theoryshould make fairlyaccurate predictions about the behavior of these states in re-
sponse to externalthreats(e.g., France, 1899-1913).
63. GregoryM. Luebbert,Liberalism, Fascism,or Social Democracy:Social Classesand thePoliticalOri-
gins ofRegimesin InterwarEurope(New York: Oxford UniversityPress, 1991).

International
Security29:2 186
ate coherentorganizationswith centralizedleadership; and it was this organi-

zational strength that enabled trade unions to penetrate the lives and
consciousness of the workingclasses. In contrast,where a consolidated middle
class had already established the hegemonyof liberalismand Lib-Labism,the
workingclasses remained politicallyweak and disorganized. Thus, when de-
mocratization finally eliminated the traditional barriers to working-class
power in aliberal societies,the new balance of class power derailed the liberal
democratic experiment.A powerful labor movement and disunited middle
class prohibited the building of alliances across the middle and working
classes and led, instead, to a corporatistsolution, eithersocial democracy or
fascism.
This patternis well illustratedin interwarEurope. According to Gregory
Luebbert'sstructuralist theoryof regimechange, only the relativelymaturelib-
eral democracies (Britain,France,and Switzerland) were capable of surviving
not just the Great War but also the subsequent interwarcrises."4The triumph
of marketsover politics,the co-optationof labor, and right-center dominance
enabled liberal democracies to muddle throughthese crises despite extremely
fragmentedelites and polarized societies. In contrast,the democratizingre-
gimes of theiraliberal neighborswere too fragileto withstand these shocks,
and theywere replaced by eitherthe hegemonyof the workingclass (social de-
mocracy)or its total destructionunder brutallycentralizedregimesthatput a
premium on the extractivecapacity of the state (fascism). Social democracy
emerged where urban socialism attractedthe support of the familypeasantry
(Czechoslovakia, Denmark,Norway,and Sweden), which solidifiedthe politi-
cal dominance of the workingclass and the subordinationof marketsto poli-
tics; these states demonstrablyunderbalanced. Where this process did not
occur (Germany,Italy,and Spain), fascismarose, crushingthe labor movement
and harnessingthe workingclass to the state; these countriesoverexpanded.
In sum, when thereis a highlyfragmentedmiddle class withina competitive
political process thatincludes labor,elite fragmentationwill push in the direc-
tion of underbalancingbehavior and policy paralysis."6 Thus, although regime
survival in liberal democracies during the interwarperiod remained strong,
theircapacity to forgecoherentand effectivebalancing strategiesin response
to dangerous externalthreatsproved dangerously weak. This point is elabo-
rated upon in the case studies of interwarBritainand France.
64. Ibid.
65. Moreover,as discussed above, elite fragmentationwill lead to underbalancingin cases of de-
mocratizingstates where the old elites retainenough power to derail the transition.

Case Studies:The Underbalancing

ofInterwarBritainand France
This section presents two case studies: Britishand French grand strategies
priorto WorldWar II. These cases may be seen as "crucial" ones forthe theory
because they"have come to define,or at least to exemplify,a concept or theo-
retical outcome,""''namely,underbalancing behavior. For two reasons, how-
ever,theyare not "easy" or "extreme"cases forthe theory.First,the values of
two of the independent/interveningvariables-regime vulnerabilityand so-
cial cohesion-are relativelylow when compared with "most likely" cases,
such as late Qing China, late Tokugawa Japan, Moghul India, and Austria-
Hungary.~'Second, theyare "most likely" or "easy" cases forrealistbalance of
power theoryin thattheyinvolve two greatpowers directlythreatenedby an-
other unmistakablyaggressive great power-one that they had just fought
against in a bloody life-and-deathstruggle a mere twenty years prior. If
HitleriteGermanyis not a threatto be balanced against,what would qualifyas
one? Thus, by consideringonly greatpowers under conditionsof clear threats
to theirsurvival,the cases challenge balance of power in its own bailiwick,so
to speak.
BRITAIN: A DOMESTIC CLIMATE FOR APPEASEMENT

InterwarBritishdefense policies provide an exemplary case of poor strategic
adjustment in the formof overly cooperative behavior in response to an in-
creasingly threateningexternal environment.Indeed, as Andrew Roberts
notes,between "1920 and 1938, Britishdefense spending was consistentlyless
than 5 per cent of national income per annum-less than at any timebeforeor
since; and this at a time when Britain's imperial commitmentshad almost
reached theirmaximum historicextent.""'To borrow Arnold Wolfers'smeta-
phor, the house was unmistakablyon fire,and yet the Britishdid not rush to
exit as if compelled by an irresistibleforce."''Britain's policy of appeasement
(accompanied by its commitmentto "peaceful change," "limited liability,"di-
plomacy, disarmament,and unilateral arms limitationunder the aegis of the
Ten Years' Rule of 1919) priorto World War II is all the more puzzling because
it was so distinctlydifferentfromthe balance of power doctrine Britainhad
66. JohnGerring,Social ScienceMethodology:A CriticalFramework

(Cambridge: Cambridge Univer-
sityPress, 2001), p. 219.
67. I am gratefulto one of the anonymous reviewersforpointingthis out.
68. Andrew Roberts,"Hitler's England: What If GermanyHad Invaded Britainin May 1940?" in
Ferguson, VirtualHistory,p. 284.
69. Wolfers,Discordand Collaboration,
pp. 13-16.

International
Security29:2 188
adhered to prior to 1914. Why did interwarBritishelites so consciously reject

balance of power as a mode of behavior when, objectivelyspeaking, Anglo-
German antagonism was much more marked in the 1930s than beforeWorld
War I?
The key to understandingBritishforeignpolicy during the interwarperiod
is to recognize the trade-offbetween internaland externalstability.More pre-
cisely,Britishelites refusedto put at risk theirabilityto stabilize the domestic
political systemin exchange forenhanced externalsecurity.Accordingly,Brit-
ish policy toward the outside world was circumscribedby the imperative-
overwhelminglysupported by policy elites of all parties,captains of industry,
and public opinion-that foreignand securitypolicies must not interferewith
the primary goal of safeguarding the viability and efficiencyof Britain's
sociopolitical system; this required Britain's complete freedom of action to
develop internallyas its elites saw fit.
In practice,this meant a rejectionof the "old" balance of power strategy,
which would have put Britishdomestic policy at the mercyof alliance politics.
Specifically,Britishelites were not prepared to give France (or Belgium or
Soviet Russia) a say in British domestic disputes by including the French
(or any other ally) in joint militarystafftalks or coordinated rearmament
measures. Determinedto maintainits isolation fromthe continentand its non-
aligned strategy,the Britishgovernmentformulateda two-track,appeasement-
cum-rearmamentstrategy.Limited rearmament,which focused solely on
strengtheningthe Royal Air Force and air defenses against a bomber attack
fromGermany,would support the primarystrategyof appeasement, which
centeredon disarmamenttalks and negotiations(mostlybilateral) forpeaceful
change. As Gustav Schmidtargues, "Britishpolicy in the fieldsof security,ar-
maments and foreignaffairsin the 1930s may be seen ... as motivated by a
fear of destabilizationof the domestic status quo, which led . .. to a greater
willingnessto accept changes in the internationalstatusquo .... [It]was linked
to the expectationthat the relative autonomy of decision-makingin all ques-
tions relevantto the development of Britishsocietyand politics had to be up-
held."'7 The articulationof this guiding principlecorrespondedwith,and was
in response to, the social unrestand protestthatarose during Britain'stransi-
tion froma wartimeto a peacetime economy in the 1920s, when the govern-
ment was also forcedto set prioritiesto check inflationcaused by the cost of
70. Gustav Schmidt,The Politicsand EconomicsofAppeasement:BritishForeignPolicy in the1930s

(Hamburg, Germany:Berg, 1986), pp. 25-26, 28.

war. Thus, Stanley Baldwin maintained in 1936: "In the postwar years we had
to choose between . . . a policy of disarmament,social reformand latterly
financialrehabilitationand ... a heavy expenditureon armaments.Under a
powerful impulse for development every governmentof every party elected
forthe former."7'
The primacy of sociopolitical stabilization over rearmamentand external
securitypersisted despite mountingevidence of the "German peril" and even
after the annulment of the Ten Years' Rule-a document prepared by the
Finance Committee of the British cabinet in August 1919 and made self-
perpetuating in 1928. The Ten Years' Rule "stipulated that British defense
forces should formulatetheirarmament programs based on the assumption
thatBritainwould not be engaged in a major war forthe next decade"; it was
annulled in 1932.72 Indeed, the Defense Requirements Committee's 1935
guideline, which pronounced the DRC's assumption that the urgentand im-
mediate need for Britain to achieve greater security took precedence over
financialconsiderations,was never implemented.Britishpolicy did not adjust
in response to a dramaticallychanged strategicenvironmentbecause elites
were paralyzed by the threatof social unrestand fearsthatthe domestic politi-
cal systemcould not stand the strainsof rearmament.Once again, Schmidtex-
plains: "'Security policy' was still subject to the directivethat Britain could
ratheraffordto take risksover the defence issue than run the risk of dividing
the countryalong class lines. This maxim put into practice the crucial lesson
drawn frompolitical experience since the turn of the century;namely, that
trade unions had learnthow to exertpressure throughindustrialaction in key
industries (mining, railways, shipping). Governmentsin any case preferred
'willing co-operation'to 'industrialconscription'but hesitatedin enforcingac-
celeratedrearmamentbecause theyfearedgrantingconcessions to labour in re-
lated-issue areas." Throughoutthe 1930s, trade unions remained profoundly
suspicious of national rearmamentbecause of the danger of industrial con-
scription.Thus, as Schmittfurthernotes: "Against this backdrop, the rearma-
ment campaign, justifiedon foreignpolicy grounds, seemed riskyeven when
71. Stanley Baldwin, 11 March 1936, cab. 28-83, p. 295, as quoted in ibid., p. 228.
72. Michael L. Roi, Alternativeto Appeasement:Sir RobertVansittartand AllianceDiplomacy,1934-
1937 (Westport,Conn.: Praeger,1997), p. 5. As Gustav Schmittnotes about the annulmentin 1934:
"By consentingto the chiefof staff'sdemand forthe abrogationof theTen Years' Rule only on con-
hopes of disarmamentnegotiationsin Geneva would not therebybe torpedoed
dition that,firstly,
and that,secondly,the economic crisis would be taken into consideration,the cabinet reinforced
the tenet that avoidance of domestic political risks must take precedence over security risks
abroad." Schmitt,The Politicsand EconomicsofAppeasement, p. 229.

Security29:2 190
International
individual, influentialunion leaders ([Ernest] Bevin and [WalterMcLennan]

Citrine) made it known that,in theiropinion, 'collective security'demanded
considerable effortsfor rearmamentin Britain,regardless of which govern-
ment was in power."73
Labor's preferenceforsocial legislationwas the primaryobstacle to military
spending, but not the only one. The economic dislocations caused by World
War I and the GreatDepression gave rise to popular demands among the mid-
dle class for expenditureson social programs,making militaryspending ex-
tremely unpopular among the general public and, therefore,politically
unfeasible.Arguablythe most famous example of the popularityof pacifismin
Britainoccurred in February 1933, when the celebrated Oxford Union, after
heated debate, sustained the motionby a vote of 275 to 153: "That thisHouse
will in no circumstancesfightforits King and Country."74Driven by these do-
mesticconstraintson Britishsecuritypolicy,totalexpendituresforall threeser-
vices accounted forroughly10 percentof the national budget in 1933-a paltry
sum when compared with the 30 percentfigureprior to World War I.
The British case may be conveniently,if somewhat arbitrarily,divided
into two periods: January1933-February1936 and March 1936-March 1939.
Throughoutboth periods, Britishdomestic politics were fragmentedamong
the three major parties (Labour, Conservative, and Liberal) and competing
elite factionswithinthose parties. The level of Britishfragmentationwas not
uncharacteristicof democraticsystemsin "normal" times. Prior to Nazi Ger-
many's remilitarizationof the Rhineland on March 7, 1936, however, British
elite fragmentationwas accompanied by an overwhelmingconsensus among
elites and the public forappeasement. Isolationistson the Right,pacifistson
the Left,and pro-League enthusiastsat the Center all supported, forvarious
and not entirelycompatible reasons, conciliation of the revisionistdictators
through the machinery of collective security as against a more proactive
balancing policy of rearmamentand defensive alliances (see Table 1).
Conservatives, who dominated the government, supported limited re-
armament,focusingon bombers and air defense,to ensure Britishsecurity-
defensivemeasures thatwould not be useful as contributionsto European and
Frenchsecurity.Their preferredgrand strategywas one of FortressBritainas
against Britain'straditional"balancer" role vis-Ti-visthe continent.In contrast,
73. Schmitt,The Politicsand EconomicsofAppeasement, p. 254.

74. Quoted in R.J.Q.Adams, BritishPoliticsand ForeignPolicyin theAge ofAppeasement,
1935-39
(Stanford,Calif.: StanfordUniversityPress, 1993), p. 9.

Table 1. BritishEliteViews on GrandStrategy,January1933-February1936

EliteGroups GrandStrategy
Mass of Conservatives(Stanley Passive appeasement; limitedliability;armed
Baldwin,AnthonyEden, Neville forcesfornationaldefenseto promoteBritish
Chamberlain) isolationismand neutrality
WinstonChurchillConservatives Armsand allies against Germany
Labour Disarmamentand international conciliation;"yes"

to limitedarmamentsforthe League of Nations
and collectivesecurity;"no" to armed forcesfor
nationaldefense
Liberals Stronglypro-Leagueof Nations;opposed to
rearmamentand Britishneutralityin Southern
and EasternEurope; stronglyopposed to
economic nationalism
LloydGeorge Liberals Activeappeasement; determinedpursuitof

settlementwithGermany
PolicyOutcome:EliteConsensus forAppeasement
left-wingpartiessupported a more internationalist, multilateralgrand strategy

centered on the "rule of law" and collective security.Yet, they consistently
denied the governmentthe necessary means forBritainto play a leading role
in a systemof collectivesecuritybased on "peace throughstrength."
These aspects of the Britishcase underscore two points: (1) elite consensus
regardingpolicy preferencesis not necessarilyaccompanied by a similar con-
sensus on grand strategyor elite cohesion, and (2) elite consensus is not a
sufficientcondition for appropriate balancing behavior. Consensus among
elites and the masses may formin support of an alternativepolicy to balancing
(e.g., appeasement and conciliationof the threat).In this case, an overly coop-
erative strategywith respectto Germany arose because it was a common de-
nominatoramong elites-one that promised the least domestic political costs
but high policy risks of failure.
Afterthe Rhineland crisis,policy consensus forappeasement among British
elites was replaced by an intense but unevenly matched political struggle.A
new consensus arose among elites regardingthe external environment:Ger-
many was seen as the primarythreatto Britishsecurityand, in response, a
more proactive policy had to be undertaken.Britishelites, however, divided
into two camps: (1) those who supported the government'spolicy of appease-

Security29:2 192
International
ment,but who now favoredactive over passive appeasement and limitedrear-

mament for Britishdefense,75and (2) those who favoredWinston Churchill's
Grand Alliance proposal for an offensive-defensivealliance among France,
Britain,and the Soviet Union, supported by staffarrangementsand the moral
backing of the League of Nations (see Table 2).
Churchill's supporters,referredto elsewhere as the "anti-appeasers,"7"'be-
lieved thatGerman aggressioncould be deterredonly by an Anglo-Sovietmili-
taryalliance. Thus, RobertBoothby,Churchill's close friend,confided to Lloyd
George in late 1938: "There is no doubt whatever that [Foreign Minister
Joachimvon] Ribbentroptold Hitler that the Britishgovernmentwas much
more concerned about class intereststhan national interests;and that unless
and until direct political and military contacts were established between
the Britishand the Russians, he could restassured thatthe Britishgovernment
did not mean business."77Although theywere badly outnumbered,the "anti-
appeasers," whose members included Leopold Amery, Clement Atlee,
Boothby, Churchill, Hugh Dalton, Anthony Eden, Lloyd George, Harold
Macmillan, Harold Nicolson, Archibald Sinclair,and Maj. Gen. Sir Edward
Spears, mighthave been predictedto carrythe day because theywere farmore
talentedand eloquent politiciansthan Neville Chamberlain's supporters.They
did not triumphover the appeasers, however,partlybecause theywere hope-
lessly fragmented politically and could not-or, more accurately, would
not-form a united frontin opposition to Chamberlain,whose supportersin-
cluded the mass of Conservatives,old isolationistson the Right,Liberal and
Labour pacifists,and a few Nazi sympathizers.Churchill's problem was that
his supportersfromthe extremerightwing of the Toriesrefusedto become po-
litically intimatewith Labour. Likewise, Labour's leadership worried about
upsetting its own supportersby mingling with Tories. Moreover, within the
75. William R. Rock definesthe difference between these two formsof appeasement in the follow-
ing terms:"[Passive appeasement] involves the impromptuand piecemeal sanctioning of an ad-
versary's advance--by means of armament amassing treatyviolation, or the development of
economic and ideological spheres of interest-to a more powerful position in an area where the
appeasing power has vital interests.... [Active appeasement] begins with a sympathetichearing
forthe concretegrievancesand demands laid down by a potentiallyor activelyaggressive adver-
sary,perhaps even with an invitationto formulateand state grievancesor demands more clearly,
and proceeds throughnegotiationtowards the satisfactionof the demands." Rock, BritishAppease-
mentin the1930s (New York: W.W. Norton, 1977), p. 25.
76. Neville Thompson, TheAnti-appeasers: ConservativeOppositionto Appeasementin the1930s (Ox-
ford:OxfordUniversityPress, 1971); and Louise Grace Shaw, TheBritishPoliticalEliteand theSoviet
Union,1937-1939 (London: Frank Cass, 2003), chap. 4.
77. Quoted in Shaw, The BritishPoliticalEliteand theSovietUnion,p. 82.

Table 2. Two CompetingBritishGrandStrategies,March 1936-March1939

EliteGroups:The Appeasers GrandStrategy:ChamberlainAppeasement
Mass of Conservatives Activeappeasement; limitedliabilityon the
ToryIsolationists continent;rearmamentfornationaldefense of
Nazi sympathizers the homelandand the empire
Labour and Liberalpacifists
EliteGroups:The Anti-appeasers GrandStrategy:Churchillian

Deterrence
Mass of Labour (ClementAttlee, Armedcoercive League of Nationsto resistthe
HaroldMacmillan) aggressors and impose disarmamenton
Right-wingTories Germany;GrandAlliancewithRussia and
LloydGeorge Liberals Franceto encircleGermany;commitment to
defendEasternand SouthernEurope
PolicyOutcome: Underbalancing
Conservative Party,Eden "had no intentionof voluntarilysubordinatinghim-

selfto Churchill... and, withouthim, the prospect of a substantialConserva-
tive secession disappeared."'"
Thus, appeasement and "the spiritof Munich," in which Adolf Hitler was
portrayedas a mere pan-Germanwho could be satisfiedby negotiated settle-
ments that put Germans living in other countriesback in Germany,shaped
Britishgrand strategyup untilHitler's seizure of the restof Czechoslovakia in
March 1939. In the years 1937-39, the treasury,Chamberlain, and his col-
leagues (Sir JohnSimon and Sir Thomas Inskip) continuedto hope beyond rea-
son thatan internationalsettlementcould be reached. Defyingthe warnings of
the BritishForeign Officeand military,they succeeded in delaying broad and
prudent rearmamentin favor of the so-called Inskip Doctrine, which advo-
cated stayingthe "appeasement" course throughlimitedrearmamentcoupled
with more active effortsat peaceful preventivediplomacy.79
Yet,the question arises: ifChurchilland opposition partiesshiftedtheirposi-
tion to favor balancing, what prevented the governingpartyfromdoing so?
The answer is thatChamberlain and the rightwing of the Conservative Party
had compelling political and social reasons to remain committedto appease-
78. R.A.C. Parker,"Alternativesto Appeasement," in PatrickFinney,ed., The OriginsoftheSecond

WorldWar (London: Arnold, 1997), p. 219.
79. For details on the Inskip Reportof March 1938, see RobertPaul Shay,BritishRearmament in the
Thirties:Politicsand Profits(Princeton,N.J.:PrincetonUniversityPress, 1977), pp. 190-191.

Security29:2 194
International
ment. A peaceful solution to Europe's problems would strengthenChamber-

lain's positionwith the partyat large,checkingrivals such as Eden and further
isolatingChurchill.In addition,with the electionyear of 1940 drawing nearer,
the avoidance of war would not only sustain but enhance the Conservative
Party's appeal in the countryat large. More importantstill,Chamberlain and
the governingpartyfeared that anotherwar of attrition,like that of 1914-18,
would resultin a dramaticshiftin the domestic social ordertoward the work-
ing class. As Paul Kennedy points out: "Already the Labour Partyspokesmen
and trade unions were making it clear that,if war came, the price for their
full-heartedsupport would be the 'conscriptionof wealth' as well as man-
power, and the nationalizationof certainkey industries.Churchill,Eden and
theirfriendsmighthave been willingto pay thisprice; but therewas littlesign
that Chamberlain,or many industrialists,or most of the Conservative Party,
were. As Oliver Stanley,the Presidentof the Board of Trade, put it to an ac-
quaintance in September1938,' . . . whetherwe win or lose [a war], it will be
the end of everythingwe stand for."'1
FRANCE: A DOMESTIC CLIMATE FOR HALF MEASURES AND MUDDLING THROUGH

Like Britain,France during the interwarperiod representsa prototypicalcase
of underreactionto an extremelydangerous externalenvironment.In a mere
twenty-twoyears, France went fromhegemonic grandeur to humiliatingde-
feat.The primarycause of Frenchdecline was that,while Germany rearmed,
France was wholly distracted by internal troubles. French society was ex-
tremelyfragmentedby deep and wide ideological and class divisions, and
these conflictswere played out in a weak political systemthatencouraged in-
decisive and muddled leadership. As David Thompson noted about French
democracy during the Third Republic: "If the functionof democraticgovern-
ment is to be a mirroror a photographicnegative-reflectingor reproducing
accurately the conflictsof social and political forces-the Republic worked
well enough. If the functionof democraticgovernmentis somethingmore pos-
itive than this . . . it worked less well."81
The Frenchgovernmentexhibitedextremeregime instability,witnessingan
astounding thirty-five changes in prime ministersbetween 1918 and 1940,and
80. Paul M. Kennedy,TheRealitiesbehindDiplomacy:Background

Influenceson BritishExternalPolicy,
1865-1980 (London: Fontana, 1981), p. 298.
81. David Thompson, Democracyin France:The Thirdand FourthRepublics,2d ed. (London: Oxford
UniversityPress, 1952), p. 171.

twenty-four changes of ministrybetween 1930 and 1940. Withthe exceptionof

the Communists,France also lacked strong,disciplined parties,though there
was no shortage of them. As the Socialist leader Leon Blum once lamented
prior to becoming France's prime ministeron June4, 1936, "Ah, if only there
were political parties in France and if these parties had an organization and
doctrine."82The resultwas a political systemdominated by personalitieswith
only loose and shiftingparty allegiances.
In addition to France's vulnerable and unstable ruling regimes and highly
fragmentedsociety,French elites made differentand contradictoryassess-
ments about the degree of threatin the external environment.One faction,
dubbed the "optimists"by the historianRene Girault,saw Hitler as a buffoon
with a penchantforswaggeringbut who would, nonetheless,act rationallyas
long as the Britishand Frenchinsisted on normal negotiations.A second fac-
tion, dubbed the "realists," maintained that Hitler had expansive ambitions
thatwould be tamed only by the threatof war froman overwhelminglypow-
erfulcountercoalitionof states aligned against Germany.A thirdfaction,the
"pessimists,"agreed with the realiststhatthe optimistswere underestimating
the extent of Hitler's revisionist goals and his willingness to risk war to
achieve them. But they,unlike the realists, supported Hitler's expansionist
aims in Centraland EasternEurope at the expense of Poland, Czechoslovakia,
and the Soviet Union, believing that these gains would satisfyHitler and
therebypreserve a separate peace in the West.83
In the absence of any elite consensus on thenatureof the externalthreat,it is
not surprisingthat Frenchleaders could not agree on a policy to confrontit.
Given Germany's overwhelming population and industrial advantages,
France would not be able to defeatGermany on its own afterthe remilitariza-
tion of the Rhineland in 1936, which broughtto a close the era of Versailles.84
Yet,elites were hopelessly divided along ideological lines over the question of
with whom France should ally. The Communists demanded that France be
ready to fightin an alliance with the Soviet Union against Germany.Socialists
and most othersin the noncommunistLeft,however,insisted thatwar should
be out of the question and held firmin theirbelief that therewere peacefully
82. Quoted in AnthonyAdamthwaite,Franceand theCoiningof theSecond WorldWar,1936-1939

(London: Frank Cass, 1977), p. 10.
83. Ernest R. May, StrangeVictory:Hitler'sConquestof France(New York: Hill and Wang, 2000),
p. 150.
84. For the details and ramifications,see JamesThomas Emmerson,The RhinelandCrisis,7 March
1936: A Studyin MultilateralDiplomacy(Ames: Iowa State UniversityPress, 1977).

International
Security29:2 196
negotiable solutionsto all foreignpolicy problems.Attributingthe outbreakof

World War I to tightalliances and balance of power politics,they supported
the League of Nations and collectivesecurityas the only legitimateinstrument
to maintaininternationalorder and preventanotherdisastrous war. The Cen-
ter in France supported traditionalmilitaryalliances with Italy,Poland, and
the LittleEntenteagainst Germany,but it would not honor its commitmentsto
these allies withoutBritishsupport.The British,however,consistentlyclaimed
thatthey had no interestsin Eastern Europe and were perfectlyhappy to see
the entirearea fallunder German,ratherthan Soviet,control.This led to a per-
verse chain reaction,wherebyPoland, forinstance,would not balance against
GermanywithoutFrenchsupport,and the Frenchwould not support Poland
withoutBritishsupport.8"
Meanwhile, the Right,whose slogans were "betterHitler than Blum" and
"betterHitlerthanStalin,"wanted to balance with Germanyagainst the Soviet
Union and its "communistconspiracyof InternationalJewry,"which theysaw
as the most dangerous threatto Frenchvalues and way of life.Finally,"radi-
cals and othersin the centerdivided threeways. Some sided with the pacifist
socialists. Some sided with the right.Some, however,said thatFrance had to
uphold the alliance treatiesand, if necessary,go to war.""8Paralyzed by elite
fragmentation, France proved politicallyincapable of choosing sides and forg-
ing a reliable and internallyconsistentalliance system.Thus, when war came,
France,having sold out its allies in the East and clingingto a reluctantally in
Britain,found itselffightingessentiallyalone against Germany.
In summary,shaped by the inherentweaknesses of Frenchsociety and the
fragilityand fragmentationof its political system,the Frenchresponse to the
German challenge was an incoherentseries of half measures and indecisive
muddling through.Frenchgrand strategy,if it can be called that,rested on a
combination of contradictorypolicies that included elements of balancing,
buck-passing, bandwagoning, and appeasement-a grand strategybest de-
scribedby the foolhardymaxim that"half a Maginot line is betterthan none."
And so, by 1938,France stillhad no greatpower allies and was devoting a pal-
try8.2 percentof national income to militaryexpenditures."7Incredibly,"from
85. See Anna M. Cienciala, Polandand theWesternPowers,1938-1939:A Studyin theInterdependence

ofEasternand Western Europe(London: Routledge and Kegan Paul, 1968); and Roman Debicki, For-
eignPolicyofPoland,1919-1939:FromtheRebirthofthePolishRepublicto WorldWarII (London: Pall
Mall, 1962).
86. May, StrangeVictory,p. 154.
87. Accordingto the CorrelatesofWar data set,France's percentageshare of militaryexpenditures

1935 through1938,the Italians spent more on theirarmed forcesin actual dol-

lars than did the French."" The dismal failureof the Frenchhigh command
and the ministryof armamentsto train and equip the Frencharmy has been
well documented elsewhere and need not be repeated here.89What was argu-
ably more responsible forthe French defeat but has been less emphasized in
the literaturewas the lack of patriotismand fightingspiritamong the French
population. As an astute observer-one who had regular and confidential
meetings with the leading French statesmen and militaryelites prior to and
during the war-wrote in 1942:
Something more serious than a shortage of indispensable arms afflictedthe
FrenchArmy.Somethingwas amiss. At a juncturemost criticalforthe fate of
France the armywas no longerthe source of inspirationand strengthit had al-
ways been to the nation.This timeit did not rejuvenatepatriotism."If only the
civilians hold firm."That is what [the political cartoonistJean-Louis]Forain
quoted Verdun's defendersas saying in the last war. The phrase would have
been meaninglessin 1939 and 1940. The Armywas the epitome of a nation cut
to its depth by politicaland social quarrels. Of course therehad been reconcili-
ation as the war drew nearerand nearer.But the passionate elementin individ-
uals often seemed to be kept in reserve for tasks other than fightingthe
enemy.9"
In the end, as a U.S. diplomat keenlyremarked,Frenchinterwarpolicy was
not thatof a "statesmanbut the policy of an undertaker.""9
This insightis con-
among the European greatpowers was a mere5.6 in 1938,and 3.7 in 1939. By contrast,France held
a 19.3 percentshare of militaryexpendituresamong European great powers in 1910 and an 18.7
percentshare in 1914. See Dale C. Copeland, TheOriginsofMajor War(Ithaca, N.Y.: Cornell Univer-
sityPress, 2000), pp. 250-252.
88. Williamson Murray,"Armored Warfare:The British,French,and German Experiences," in
Murrayand Alan R. Millett,eds., MilitaryInnovationin theInterwarPeriod(New York:Cambridge
UniversityPress, 1996), p. 13. See also MacGregor Knox, Mussolini Unleashed,1939-1941 (New
York: Cambridge UniversityPress, 1982), pp. 294-295.
89. For a briefbut extraordinarilydetailed account of the inadequacies of Frenchmilitaryprepara-
tions priorto 1940, see Pertinax,TheGraveDiggersofFrance:Gamelin,Daladier,Reynaud,Pdtain,and
Laval (New York: Howard Fertig,1968 [19421),pp. 20-33. For more recentaccounts, see Murray
and Millet,MilitaryInnovationin theInterwarPeriod;and May, StrangeVictory. For the most charita-
ble account of Frenchmilitarypreparations,see R.H.S. Stolfi,"Equipment forVictoryin France in
1940," History,Vol. 52, No. 183 (February1970), pp. 1-20.
90. Pertinax,TheGraveDiggersofFrance,p. 33. Lord Wolseleymade a similarobservationabout the
Britishpublic during the Victorianera: "The people of England lack the public spiritthatwould
cause themto supporttheburden of an armyand a navy of a strengthcommensuratewithour du-
ties and responsibilities."Quoted in JohnR. Ferris,"'The GreatestPower on Earth': Great Britain
in the 1920s," InternationalHistoryReview,Vol. 13, No. 4 (November 1991), p. 732.
91. Quoted in Anthony Adamthwaite, Grandeurand Misery: France's Bid for Power in Europe,
1914-1940 (London: Arnold, 1995), p. 228.

Security29:2 1198
International
sistentwith the argumentspresented in this article: confrontedby a serious

threat,incoherentstates that sufferfromdeep social and political divisions
have a greaterpropensityto adopt "undertaker"grand strategiesthan to act in
accordance with balance of power theory.
Conclusion
Despite the historicalfrequencyof underbalancing,littlehas been writtenon

the subject.Indeed, GeoffreyBlainey's memorable observationthatfor"every
thousand pages published on the causes of wars thereis less than one page
directlyon the causes of peace" could have been made with equal veracity
about overreactionsto threatsas opposed to underreactionsto them.92Library
shelves are filled with books on the causes and dangers of exaggerating
threats,ranging fromstudies of domestic politics to bureaucraticpolitics,to
political psychology,to organization theory.By comparison, therehave been
few studies at any level of analysis or fromany theoreticalperspectivethatdi-
rectlyexplain why states have with some, if not equal, regularityunderesti-
mated dangers to theirsurvival.
There may be some cognitiveor normativebias at work here. Consider,for
instance,thatthereis a commonlyused word, paranoia, forthe unwarranted
fear that people are, in some way, "out to get you" or are planning to do
one harm. I suspect that just as many people are afflictedwith the opposite
psychosis:the delusion thateveryoneloves you when, in fact,theydo not even
like you. Yet,we do not have a familiarword forthis phenomenon. Indeed, I
am unaware of any word that describes this pathology (hubris and over-
confidencecome close, but they plainly define somethingother than what I
have described).
That noted, internationalrelations theory does have a frequentlyused
phrase forthe pathologyof states' underestimationof threatsto theirsurvival,
the so-called Munich analogy.The termis used, however,in a disparagingway
by theoriststo ridicule those who employ it. The central claim is that the
naivet6 associated with Munich and the outbreakof World War II has become
an overused and inappropriateanalogy because few leaders are as evil and un-
appeasable as Adolf Hitler.Thus, the analogy eithermistakenlycauses leaders
92. GeoffreyBlainey,The Causes of War,3d ed. (New York: Free Press, 1988), p. 3.

to adopt hawkish and overly competitivepolicies or is deliberatelyused by

leaders to justifysuch policies and mislead the public.
A more compellingexplanationforthe paucity of studies on underreactions
to threats,however,is the tendencyof theoriesto reflectcontemporaryissues
as well as the desire of theoristsand journals to provide society with policy-
relevanttheoriesthat may help resolve or manage urgentsecurityproblems.
Thus, born in the atomic age with its new balance of terrorand an ongoing
Cold War,the fieldof securitystudies has naturallyproduced theoriesof and
prescriptionsfornational securitythathave had littleto say about-and are, in
fact,heavily biased against warnings of-the dangers of underreactingto or
underestimatingthreats.Afterall, the nuclear revolutionwas not about over-
killbut,as Thomas Schellingpointed out, speed ofkill and mutual kill."3Given
the apocalyptic consequences of miscalculation,accidents, or inadvertentnu-
clear war, small wonder that theoristswere more concerned about overreact-
ing to threatsthanunderrespondingto them.At a timewhen all ofhumankind
could be wiped out in less than twenty-five minutes,theoristsmay be excused
forstressingthe benefitsof caution under conditionsof uncertaintyand erring
on the side of inferringfromambiguous actions overly benign assessments of
the opponent's intentions.The overwhelmingfearwas thata crisis "mightun-
leash forcesof an essentiallymilitarynature thatoverwhelm the political pro-
cess and bringon a war thatnobody wants. Many importantconclusionsabout
the riskof nuclear war, and thus about the political meaning of nuclear forces,
reston this fundamentalidea."94
Now that the Cold War is over,we can begin to redress these biases in the
literature.In that spirit,I have offereda domestic politics model to explain
why threatenedstates oftenfail to adjust in a prudent and coherentway to
dangerous changes in theirstrategicenvironment.The model fitsnicely with
recentrealiststudies on imperialunder-and overstretch.Specifically,it is con-
sistentwith Fareed Zakaria's analysis of U.S. foreignpolicy from1865 to 1889,
when, he claims,the United States had the national power and opportunityto
expand but failed to do so because it lacked sufficientstate power (i.e., the
statewas weak relativeto society).95Zakaria claims thatthe United States did
93. Thomas C. Schelling, Armsand Influence(New Haven, Conn.: Yale UniversityPress, 1967),
chap. 1.
94. Marc Trachtenberg,"The Meaning of Mobilizationin 1914,"International Vol. 15,No. 3
Security,
(Winter1990/91), p. 120.
95. Zakaria, FromWealthto Power,pp. 44-89.

International
Security29:2 1200
not take advantage of opportunitiesin its environmentto expand because it

lacked the institutionalstate strengthto harness resources fromsociety that
were needed to do so. I am making a similar argumentwith respectto balanc-
ing ratherthan expansion: incoherent,fragmentedstatesare unwillingand un-
able to balance against potentiallydangerous threatsbecause elites view the
domestic risks as too high, and they are unable to mobilize the required re-
sources froma divided society.
The argumentspresentedhere also suggest thatelite fragmentationand dis-
agreementwithina competitivepolitical process,which JackSnydercites as an
explanation for overexpansionistpolicies, are more likely to produce under-
balancing than overbalancingbehavior among threatenedincoherentstates.9'
This is because a balancing strategycarriescertainpoliticalcosts and riskswith
few,if any,compensatingshort-term political gains, and because the strategic
environmentis always somewhat uncertain.Consequently,logrollingamong
fragmentedelites within threatenedstates is more likely to generate overly
cautious responses to threats than overreactions to them. This dynamic
captures the underreactionof democratic states to the rise of Nazi Germany
during the interwarperiod."' In addition to elite fragmentation,I have sug-
gested some basic domestic-levelvariables that regularlyinterveneto thwart
balance of power predictions.
The main problemwithbalance of power theoryis thatit assumes "constant
mobilizationcapacity,"thatis, thatall stateshave a similarabilityto extractre-
sources from their society,such that aggregate national resources may be
equated with actual state power and global influence."'This simplifyingas-
96. Snyder,MythsofEmpire.
97. Relatedly,Deborah W. Larson offersan institutionalargumentto explain why small states in
Eastern and Central Europe cracked under Hitler's pressure ratherthan balancing against it. See
Larson, "Bandwagoning Images in American Foreign Policy: Myth or Reality?" in Jervisand
Snyder,Dominoesand Bandwagons,chap. 4. For historicaldetails of the various cases, see Joseph
Rothschild,East CentralEuropebetweenthe Wars (Seattle: Universityof WashingtonPress, 1974).
98. For a discussion of the assumptionof constantmobilizationcapabilityand systemictheory,see
Andrew Moravcsik, "Introduction:IntegratingInternationaland Domestic Theories of Interna-
tional Bargaining," in Peter B. Evans, Harold K. Jacobson, and Robert D. Putnam, eds.,
Double-EdgedDiplomacy:International Bargainingand DomesticPolitics(Berkeley:Universityof Cali-
forniaPress, 1993), pp. 10-11. For correctives,see Azar and Moon, "Legitimacy,Integration,and
Policy Capacity"; Michael Barnett,"High Politics Is Low Politics: The Domestic and Systemic
Sources of Israeli SecurityPolicy,1967-1977," WorldPolitics,Vol. 43, No. 4 (July1990), p. 561; Ste-
ven R. David, "Explaining Third World Alignment," WorldPolitics,Vol. 43, No. 2 (January1991),
pp. 223-256; Klaus Knorr,The WarPotentialofNations(Princeton,N.J.:PrincetonUniversityPress,
1956),p. 198; Alan C. Lamborn,ThePriceofPowoer: Riskand ForeignPolicyin Britain,France,and Ger-

sumption ignoresthe trade-offbetween internaland externalstability.Conse-

quently,incoherentstates,whethergreat powers or small ones, oftencannot
balance against threatsas the theorypredictsbecause theyare constrainedby
domestic political considerations.
many(Boston: Unwin Hyman, 1991); Michael Mastanduno, David A. Lake, and G. JohnIkenberry,
"Toward a Realist Theoryof State Action,"International Vol. 33, No. 4 (December
StudiesQuarterly,
1989), pp. 457-474; JohnMueller,Quiet Cataclysm:Reflectionson theRecentTransformation of World
Politics(New York:HarperCollins,1995),chap. 2; A.EK. Organski and JacekKugler,The WarLedger
(Chicago: Universityof Chicago Press, 1980); and Zakaria, From Wealthto Power.


Schweller, R. (2004) - Unanswered Threats A Neoclassical Realist Theory of Underbalancing

Загружено:

Сведения о документе

Оригинальное название

Авторское право

Доступные форматы

Поделиться этим документом

Поделиться или встроить документ

Параметры публикации

Этот документ был вам полезен?

Это неприемлемый материал?

Авторское право:

Доступные форматы

Schweller, R. (2004) - Unanswered Threats A Neoclassical Realist Theory of Underbalancing

Загружено:

Авторское право:

Доступные форматы

Unanswered Threats: A Neoclassical Realist Theory of Underbalancing

Author(s): Randall L. Schweller

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

During thelate 1630s,

The authoris gratefulto the followingindividuals fortheircommentson earlierdraftsof thisarti-

International Vol.29,No. 2 (Fall2004),pp. 159-201

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

theory,namely,thatthreatenedstates will balance against dangerous accumu-

3. K.J.Holsti, "InternationalRelationsTheoryand Domestic War in the ThirdWorld: The Limitsof

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

dictions, unipolarityhas not provoked global alarm to restorea balance of

4. For various explanations of this phenomenon,see the essays in G. JohnIkenberry,ed., America

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

conclude with a discussion of why the internationalrelations literaturehas

BalanceofPoweras a StructuralLaw ofNature

Despite the frequencyof underbalancing,thereare surprisinglyfew studies on

6. Hans J.Morgenthau,PoliticsamongNations:The StruggleforPowerand Peace,4th ed. (New York:

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

of naturestillinfusesmost discussions of how the theoryoperates. Thus, Hans

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

power will bringpeace and security.If theywish to survive,theymust be will-

The Role ofDomesticPoliticsin RealistTheory

The themeof thisarticlefitssquarely withinthe new wave of neoclassical real-

15. Nicholas JohnSpykman,America'sStrategyin WorldPolitics:The UnitedStatesand theBalanceof

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

Curiously, neoclassical realists have yet to examine the subject of under-

17. See Fareed Zakaria, FromWealthto Power:The Unusual OriginsofAmerica'sWorldRole (Prince-

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

the broader phenomenon of underbalancing behavior, which includes buck-

WHAT IS MEANT BY BALANCING BEHAVIOR?

Although arguably the most frequentlyused termin the fieldof international

21. Thomas J.Christensenand JackSnyder,"Chain Gangs and Passed Bucks: PredictingAlliance

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

of a state's response to threatdetermineswhetherit is balancing or not. Bal-

23. I am gratefulto Thomas Christensenfor his insightsregardingthis definitionof balancing.

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

Variationin the way states respond to similar changes in theirexternalenvi-

26. See E.H. Carr,TheTwentyYears'Crisis,1919-1939: An Introduction totheStudyofInternationalRe-

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

potentialdomesticpoliticalrisksand costs ofbalancing are too high.27The first

27. I am gratefulto Alan Lamborn forpointingthis out.

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

For incoherentstates,the causal scheme thatproduces underbalancingmay

Rise of an externalthreat - social fragmentation- governmentor regime

A discussion of each of the variables serves to clarifyhow these causal

28. For a similardiscussion on collectiveaction problems and institutionaldesign, see Alexander

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

actors can know forcertainand what theymust inferfromothers' behavior.30

31. For adjustmentfailuresthattake the formof eitheroverly cooperative or overly competitive

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

sympathies,collectivesecurity,a belief in conciliation,etc.). Consequently,ap-

Historically,democraticsocieties have been slow to reactto gatheringthreats,

33. George Tsebelis,VetoPlayers:How PoliticalinstitutionsWork(Princeton,N.J.:PrincetonUniver-

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

In summary,balancing behavior requires elite consensus thatthe state con-

36. It is not a sufficientconditionforbalancing,however. Even in countrieswhere thereis a mod-

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

38. JamesD. Morrow,"Arms versus Allies: Trade-Offsin the Search forSecurity,"InternationalOr-

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM

within a society.When all members of society "feel interconnectedand inte-

42. Jacques Semelin, UnarmedagainstHitler:CivilianResistancein Europe,1939-1943, trans.Suzan

This content downloaded from 190.3.14.67 on Thu, 9 Oct 2014 22:09:00 PM