APSA-CP Winter 1997

APSA-CP Newsletter 1 Summer 1997

Dear Members:

When becoming President of the Section in Comparative Politics, I soughtcounsel from those who had organized it. Comparative Politics was rapidlychanging, they emphasized, and the Section should actively explore newdepartures and emerging perspectives. While striving to be inclusive, its leader-ship should be proactive. Before departing, I wish to express my thanks tothose who have provided leadership for the Section and preserved andstrengthened its sense of mission – the officers and members of the Executive,Nominating and Awards Committees, the Program Chairs, and the Editors ofthe Newsletter.

Over the last year, we have broadened the mandate of the Awards Committeeto include not only books but also articles, thereby giving recognition to contri-butions of a more technical nature. The Newsletter now works in concert withthe Program Chair, such that debates initiated in the pages of the Newsletternow form the focus of theme panels in the Annual Meeting. The Newsletter,under the leadership of Miriam Golden, remains provocative and accessible,and is regarded by the leadership of the Association as a model of what theSections can achieve.

What do I rue? The unhelpful intervention of The Chronicle of Higher Educa-tion dominates the list. Even that otherwise infuriating incident yielded benefits,however; it contributed to my education. In the future, I will be less dismissiveof ‘real’ Presidents when they rail against the fourth estate.

And now, David Collier, over to you. Enjoy!

Bob Bates

APSA-CPAPSA-CPAPSA-CPAPSA-CPAPSA-CPNewsletter of the APSA Organized Section in Comparative Politics

Volume 8, Number 2 Summer 1997

Letter from the PresidentFarewellRobert H. Bates

Harvard [email protected]

ContentsContentsContentsContentsContents

Letters to the EditorLetters to the EditorLetters to the EditorLetters to the EditorLetters to the Editor 2

Section PrizesSection PrizesSection PrizesSection PrizesSection Prizes 3

News and NotesNews and NotesNews and NotesNews and NotesNews and Notes 4

Notes from the AnnualNotes from the AnnualNotes from the AnnualNotes from the AnnualNotes from the AnnualMeetings/DebateMeetings/DebateMeetings/DebateMeetings/DebateMeetings/DebateCulture and Rational ChoiceCulture and Rational ChoiceCulture and Rational ChoiceCulture and Rational ChoiceCulture and Rational ChoiceNancy Bermeo, introduction 5Robert H. Bates 5James Johnson 6David D. Laitin 9Ian S. Lustick 11Ron Rogowski 14George Tsebelis 15Ashutosh Varshney 18

Continuing DebatesContinuing DebatesContinuing DebatesContinuing DebatesContinuing DebatesArea StudiesArea StudiesArea StudiesArea StudiesArea StudiesPeter A. Hall 22

Data Sets & ArchivesData Sets & ArchivesData Sets & ArchivesData Sets & ArchivesData Sets & ArchivesInternational FinancialInternational FinancialInternational FinancialInternational FinancialInternational FinancialRegulationRegulationRegulationRegulationRegulationDennis P. Quinn & Maria Toyoda 25

Book ReviewsBook ReviewsBook ReviewsBook ReviewsBook ReviewsOpen Economy POpen Economy POpen Economy POpen Economy POpen Economy Politicsoliticsoliticsoliticsolitics 27Remaking the Italian EconomyRemaking the Italian EconomyRemaking the Italian EconomyRemaking the Italian EconomyRemaking the Italian Economy 28Making and BreakingMaking and BreakingMaking and BreakingMaking and BreakingMaking and BreakingGovernmentsGovernmentsGovernmentsGovernmentsGovernments 29

The PThe PThe PThe PThe Political Economy of Demoolitical Economy of Demoolitical Economy of Demoolitical Economy of Demoolitical Economy of Demo-----cratic Tcratic Tcratic Tcratic Tcratic Transitionsransitionsransitionsransitionsransitions 31

Copyright 1997 American Political Science AssociationSubsidized by the College of Letters and Science, UCLA


Correction

MADAM – Bob Bates’s comments onarea studies in the Newsletter are need-lessly harmful and a disservice to com-parative politics, political science, andthe humanities because they unfairly dis-parage the work of many of his col-leagues, reflecting an inappropriatelyverkrampte, parochial perception ofeach of these fields. He fails to ac-knowledge (1) that abstract theory andgenuine knowledge in the social sci-ences, as in any science, must be vali-dated by empirical testing; (2) that validtheory in the field of comparative poli-tics must be empirically testable on amulti regional basis, not just in U.S. orWestern settings; (3) the extent towhich science and theory in compara-tive politics continue to rest implicitlydisproportionately upon U.S. and Eu-ropean experience—notably in the areaof democratic transitions; (4) theoreti-cal work in languages and the humani-ties and its potential contributions totheory in the social sciences; (5) thatwhile rational choice theory has muchto offer the social sciences, it has its ownlimitations, e.g. its tendency to reduc-tionism; and (6) the continuing need forhealthy competition among theoreticalapproaches as the path to the true ad-vancement of knowledge. He is outra-geously unfair in disparaging as non-rig-orous the contributions of a great manyof his colleagues in the comparativepolitics subdiscipline, political scienceas a whole, and in the languages and

humanities to theoretical approachesother than the one he prefers. He as-serts that specialists in non-Western ar-eas resist new theoretical approachesin the social sciences. In my experi-ence, however, those working in Ameri-can or western European settings oftenappear to deny even the potential forcontributions of research on non-West-ern experiences, particularly those ofsub-Saharan Africa, to the advance-ment of theory in the social sciences.

John W. HarbesonCity University of New [email protected]

Section OfficersSection OfficersSection OfficersSection OfficersSection Officers

PresidentPresidentPresidentPresidentPresidentRobert Bates

Harvard [email protected]

Vice-President and President-Vice-President and President-Vice-President and President-Vice-President and President-Vice-President and President-ElectElectElectElectElectDavid Collier

University of California, [email protected]

SecretarSecretarSecretarSecretarSecretar yyyyy-----TTTTTreasurerreasurerreasurerreasurerreasurerJames Caporaso

University of [email protected]

1997 APSA Program Coordinator1997 APSA Program Coordinator1997 APSA Program Coordinator1997 APSA Program Coordinator1997 APSA Program CoordinatorNancy Bermeo

Princeton [email protected]

At-Large Committee MembersAt-Large Committee MembersAt-Large Committee MembersAt-Large Committee MembersAt-Large Committee MembersRonald Herring

Cornell [email protected]

Edmond KellerUniversity of California, Los [email protected]

Ian LustickUniversity of [email protected]

Susan ShirkUniversity of California, San [email protected]

Gabor TokaCentral European [email protected]

The 1997 business meeting of the Organized Section in Comparative Politicswill be held on Thursday, August 28, 1997, at 5:30 pm in the Maryland C

room of the Sheraton Washington. All are welcome to attend.

Letters to the Editor

In the Winter 1997 [Volume 8, Issue1] issue of the Newsletter, Figure I ofGary King’s article should haveappeared as follows:

Figure I. Given these mar-ginal percentages from Loui-siana in 1990, what are thecell entries?

Voted Absent

Black ? ? 26.6%

White ? ? 73.4%

68.5% 31.5%


Newsletter Staff:Newsletter Staff:Newsletter Staff:Newsletter Staff:Newsletter Staff:

Edi torEdi torEdi torEdi torEdi torMiriam Golden

University of California, Los [email protected]

Assistant EditorAssistant EditorAssistant EditorAssistant EditorAssistant EditorDavid Yamanishi

University of California, Los [email protected]

Regional Editors-at-Large:Regional Editors-at-Large:Regional Editors-at-Large:Regional Editors-at-Large:Regional Editors-at-Large:

Soviet Successor StatesSoviet Successor StatesSoviet Successor StatesSoviet Successor StatesSoviet Successor StatesDaniel Treisman

University of California, Los Angeles

Middle EastMiddle EastMiddle EastMiddle EastMiddle EastLeonard Binder


Latin AmericaLatin AmericaLatin AmericaLatin AmericaLatin AmericaBarbara Geddes


AfricaAfricaAfricaAfricaAfricaEdmond Keller


WWWWWestern Europeestern Europeestern Europeestern Europeestern EuropeRon Rogowski


Eastern EuropeEastern EuropeEastern EuropeEastern EuropeEastern EuropeIvan Szelenyi


FFFFFormal Analysis and Methodologyormal Analysis and Methodologyormal Analysis and Methodologyormal Analysis and Methodologyormal Analysis and MethodologyJohn Londregan


The Newsletter is available on the Internet for viewing and downloading.

Visit us at http://shelley.sscnet.ucla.edu/apsacp/index.html.

Send technical inquiries to the Assistant Editor, David Yamanishi, at [email protected]..

Section prizes will be conferred dur-ing the Business Meeting. See page 2for details.

The Luebbert prize for best book incomparative politics has beenawarded to J. Mark Ramseyer andFrances Rosenbluth for The Politicsof Oligarchy: Institutional Choice inImperial Japan (Cambridge and NewYork: Cambridge University Press,1995).

The Politics of Oligarchy attemptsto explain the evolution of the Japanesepolitical system between 1868 and1932. This includes both broad pat-terns of institutional evolution, and twostriking episodes: the rise of electoralpolitics in the 1910s and the collapseof democracy in the 1930s. Ramseyerand Rosenbluth argue that both broadtrends and specific events can best beunderstood with the tools of carteltheory. The Meiji oligarchy’s openingto political competition stemmed fromits inability to control competition amongthe oligarchs themselves. Intra-oligar-chical conflict also contributed to weak-ness in electoral institutions, and to theunusual independence the oligarchy waswilling to grant the military. Both re-sults were central to the collapse ofJapanese democracy in the inter-warperiod.

Ramseyer and Rosenbluth’s pro-vocative case is presented cogently anddefended conscientiously. They are

careful to provide the logic of their theo-retical argument, and to show how itsempirical implications are borne out bythe evidence. The book is sure to becontroversial among Japan scholars,and comparativists more generally. Yetit stands as a model of clear argumen-tation, theoretical grounding, and em-pirical evaluation.

The runners-up for the Luebbertbook prize are:

• Miriam Golden for Heroic De-feats: The Politics of Job Loss(New York: Cambridge Univer-sity Press, 1997).

• John Huber for RationalizingParliament: Legislative Insti-tutions and Party Politics inFrance (New York: CambridgeUniversity Press, 1995).

• Stathis N. Kalyvas for The Riseof Christian Democracy inEurope (Ithaca: Cornell Univer-sity Press, 1996).

The Luebbert Prize for the best pa-per in comparative politics goes toJames D. Fearon and David D. Laitinfor their article “Explaining InterethnicCooperation” (American PoliticalScience Review 90 (December 1996):715-735).

Laitin and Fearon examine an issueof relevance to many sub-fields in com-parative politics: the nature of ethnic

Section Prizes


conflict. While interethnic tension ispervasive, they demonstrate empiricallythat interethnic relations are much moreoften characterized by cooperation thanconflict. Accordingly, they focus onhow ethnic cooperation is sustained (bydecentralized mechanisms of social con-trol rather than by state action) in theface of the inevitable misunderstandingsthat arise as individuals from variousgroups interact. In Fearon and Laitin’sview, interethnic tension often stemssimply from the lower density of socialnetworks across ethnic boundaries,which makes it difficult to evaluate andtrust members of another ethnic groupor to punish them when they transgressconventional norms of behavior. Theyidentify two different mechanisms bywhich one ethnic group might police thebehavior of another’s members, con-sider the conditions under which eachmight emerge, and trace out how eachhandles actual transgressions. Using aformal game-theoretic model, they showthat the two mechanisms they identifyhave very different implications whenthey fail – one more likely to allow pe-riodic outbreaks of ethnic conflict tospiral out of control, the other morelikely to contain such outbreaks. Byexamining which mechanisms can effec-tively sustain interethnic cooperation,Fearon and Laitin further an area of re-search with potentially important socialscientific and real-world payoffs. Theirsis a particularly effective use of formalmodelling, since it both makes explicitheretofore underanalyzed relationshipsin the empirical record and links theo-retical insights to evidence drawn froma wide range of cases.

The three runners-up for theLuebbert Prize for best paper are:

• Geoffrey Garrett and PeterLange. “Internationalization,Institutions and Political Change”(International Organization49:4 (Autumn, 1995): 627-55).

• Kathryn Firmin-Sellers. “ThePolitics of Property Rights”(American Political ScienceReview 89 (1995): 867-881).

• Lupia, Arthur, and Kaare Strom.“Coalition Termination and theStrategic Timing of ParliamentaryElections” (American PoliticalScience Review 89 (1995): 648-668).

The Sage Publications Prize for bestpaper at the annual conference will bedivided and awarded jointly to:

• Professor Duane Swank ofMarquette University for hispaper “Funding the WelfareState, Part 1,” an outstandingcomparative study ofglobalization influences ondomestic tax regimes. Swankemploys sophisticatedtechniques and brings rigor to thesubject of globalization.

• Professor Isabela Mares ofHarvard University for her paper:“Negotiated Risks: Employersand the Development ofUnemployment Insurance,” animaginative and originalcontribution to comparativestudies of welfare state formation.Mares’ study of BismarckGermany is a significantcontribution to historicalinstitutional analyses of thewelfare state.

International Industrial RelationsAssociation (IIRA) 11th WorldCongress in Bologna, Italy, 22-26September 1998

SPECIAL SEMINAR SERIES on“Change and Continuities inIndustrial Relations”(Coordinator: Marino Regini,University of Milan)

General theme: Globalization andNational Differences: the role ofmarkets and institutions in IR trends

Rationale. The six forums around whichthe IIRA Congress are organized arefocused on either specific issuesselected for their current relevance andcrucial presence in the internationaldebate, or new key sectoraldevelopments. While this choice hasmany merits, it cannot provide enoughroom for the more traditional themes ofindustrial relations, which a WorldCongress must be able to accomodate.In fact, these themes allow us to providemore balanced pictures of what isactually changing and new in industrialrelations versus what is part of long-term trends or of cyclical re-occurrences.

This “special seminar series” willconsist of four sessions. Each of themwill address a major traditional themeof industrial relations systems by askingtwo questions. First, to what extent doesrecent change represent a dramaticrupture with the past or lead instead tominor modifications? Second, to whatextent can such a change be capturedby outlining general trends or does itinstead point to different directionsdepending on pre-existing nationalinstitutions?

(continued on page 21)

Section Prizes News & Notes


Rational Choice andPolitical CultureRobert H. BatesHarvard [email protected]

When Weiner and Huntington reviewedthe field of political development, theynoted the renaissance of religious andethnic politics and viewed it as signalingthe inadequacy of rational choice ap-proaches to politics.1 Cultural conflictsuggests a lack of regard for calculatingdecisions, they argued, and recent his-tory has provided ample reason to agree.In the former Yugoslavia, people haveincurred massive costs for doubtfulgains. In committing acts of martyrdom,individuals sacrifice their own lives for‘greater’ causes. Wherever the politicsof community arises, hot-blooded emo-tion displaces reasoned discourse Themore prominent this kind of politics, thestronger the credibility of Weiner and

Huntington’s position.The arguments advanced by Weiner

and Huntington resonate with earlier in-tellectual traditions. On the one hand,they echo the discord that accompaniedpolitical collapse of the age of reason.Weiner and Huntington play EdmundBurke to, say, Peter Evan’s ThomasPaine. On the other hand, their argu-ments echo the legacy bequeathed bythe defeat of Weimar. The behavioral sci-ences were forged, after all, to explainthe rout of reason in Germany. The re-search accumulated by the behavioralmovement demonstrated the power ofthe need for social acceptance. It dem-onstrated the impact of social anchoringon human judgments. It emphasized therole of emotion and psychological forces.The tradition thus attacked the premiseof radical individualism and discountedthe significance of reason.

As the study of political culture arosewithin the behavioral tradition, it sharedmany the latter’s orientation toward the

study of politics. It too affirmed the pri-macy of the social and viewed the rootsof human conduct as lying as much inemotions and identities as in the facultyof reason.

In recent years, postmodernist formsof theorizing have joined in studying thepolitics of culture and in doubting thepossibility of its rational apprehension.While discordant in their ranks, post-modernists too challenge the premise ofrationality, viewing it as a conjuncturalproduct of a particular place and time;they too emphasize the role of commu-nity and the political power of symbols,rhetoric and the mass media. When ad-dressing the politics of identity, they, likethe ‘old style’ theorists of culture, reflex-ively discount the significance of ratio-nality in choice.

Rational choice theory and the studyof political culture thus appear to occupy“separate tables” in the study of politics.2

But in the study of politics, as shown inthis Newsletter, increasing numbers seek

Notes from the Annual Meetings:Culture and Rational Choice

Joining Tables?Nancy Bermeo

Princeton [email protected]

In 1988, in a widely read article in the APSR, Harry Eckstein suggested that the “single most important item on the agendaof political science” was “determining whether ‘culturalist’ or ‘rationalist’ modes of analysis [were] likely to give betterresults.” (APSR 82:3 (1988), p. 789) In the years since Eckstein made his assertion, hundreds of scholars have taken sideson the debate. We have willingly rushed for seats at what Gabriel Almond eloquently described as “separate tables” andoften limited bilateral contacts to heated discussions over hiring and tenure positions. In an effort to join some tables togetherand to examine some of the assumptions that brought us to different positions in the first place, I took advantage of myposition as head of the Comparative Politics section for the 1997 Annual Meetings to organize a Roundtable entitled , “Canthe Rational Choice Framework Cope with Culture?” Summaries of two of the roundtable presentations by Robert Batesand Ian Lustick appear below, along with additional contributions solicited by Miriam Golden, the Newsletter’s Editor. IanShapiro and Ronald Inglehart will join the APSA Roundtable in August in what promises to be a lively debate.

The debate below is the second in our continuing series linking the Newsletter and Section panels at the APSAmeetings in a more explicit fashion. Tentatively scheduled on Friday, August 29, at 10:45 am, the panel “Can theRational Choice Framework Cope with Culture?” will further explore the issues addressed in these articles.


to span the divide.Ironically, perhaps, many who seek to

extend game theoretic tools to the studyof culture have been inspired by the ar-guments of post-modernists. Few are assensitive as they to the relationship be-tween culture and power or as aware ofthe manner in which ‘universal’ valueshave been appropriated for private pur-poses. Emotional, communal politics isalso guileful, they contend. Joining an-thropologists and social historians, con-temporary scholars explore the creationof tradition3 and the construction of com-munity.4 The politicization of identity, theyrecognize, is a strategy that can be cho-sen or, perhaps with greater difficulty,abandoned.

References1 Myron Weiner and Samuel P. Hunting-

ton, Understanding Political Devel-opment (Boston: Little, Brown, andCo., 1987).

2 Gabriel Almond, A Discipline Di-vided (Newbury Park, CA: Sage,1990).

3 Eric Hobsbawn and Terence Ranger,eds. The Invention of Tradition(Cambridge: Cambridge UniversityPress, 1993).

4 Elizabeth Colson, Tradition and Con-tract (Chicago: Aldine, 1974).

Symbol and Strategyin ComparativePolitical Analysis1

James JohnsonUniversity of [email protected]

IntroductionCulture and politics are related in impor-tant, pervasive, intimate ways that, whileintuitively obvious, resist systematicanalysis. That said, the subject of thissymposium may well strike many politi-cal scientists as quite improbable. Whatcan rational choice theorists conceivablysay about the relations between cultureand politics? This judgement reflects atheoretical impasse in which advocates

and critics of rational choice theory de-fend mutually reinforcing but largely mis-taken positions. I first sketch this impasseand then propose one way beyond it.

Rational choice theorists typicallyadopt one of three attitudes toward therole of cultural factors in political expla-nations. Many are dismissive. GeorgeTsebelis, for example, identifies “culture”as among the very many factors thatsimply “do not enter directly into anyrational choice explanation.” Othersadopt a reductionist posture. RobertBates, for instance, speculates that “con-temporary game theory” affords the ap-paratus necessary “to provide formalstructure for kinds of symbolic displays”that occupy the intersection of politicsand culture. On this view cultural fac-tors can be reduced to signals in gamesof incomplete information. Finally, someare accommodationist. For example,while he suggests that rational choicetheory and cultural analysis are “comple-mentary,” John Ferejohn argues they areengaged in dissimilar endeavors, and so,can be pursued along parallel, largelyindependent tacks. On his view the twosorts of analysis may, but need not, in-form one another. In what follows I ar-gue that each of these positions is mis-guided. I show that cultural consider-ations enter directly, if usually tacitly, intogame theoretic models of strategic in-teraction. I suggest that such consider-ations currently are not reducible to theformal apparatus that game theoristshave developed. I conclude that com-parative analysis necessarily must attendto both symbolic and strategic factors.

Critics of rational choice theory rangefrom the more or less dyspeptic(Chalmers Johnson, Harry Eckstein) tothe reasonably measured (Gabriel Al-mond, Ronald Inglehart). Such critics,however different their tone, agree onone point: rational choice theory contrib-utes little or nothing to analyses of therelation between culture and politics.Some claim that rational choice theories(for some typically mysterious reason)cannot contribute anything to such analy-ses. Others complain by turns that ratio-nal choice theories “ignore cultural fac-

tors” and that they treat such factors “asconstant.” These criticisms seem war-ranted precisely to the extent that ratio-nal choice theorists invite them in waysthat I just mentioned. Nevertheless thesecriticisms are either wrong or insuffi-ciently nuanced. They surely warrantnothing like the wholesale repudiation ofrational choice theory that the critics pro-pose.

Although rational choice theoristsmight like to circumscribe or even elimi-nate cultural considerations in politicalanalysis they cannot, and presently donot, do so. Critics of rational choicetheory may think this claim vindicatestheir views. Yet I actually offer them littlesupport. For even if we concede thatrational choice theories currently do notoffer much explicit aid in analyzing theinteraction of culture and politics, criticsof rational choice have not done muchbetter. David Laitin noted a decade ago(in a view endorsed by anthropologistDavid Kertzer) that “the systematicstudy of culture and politics is moribund.”The situation (recent revivals ofDurkheim and Parsons, or enthusiasmsfor ‘social capital’ notwithstanding) hasnot changed markedly. Despite both theirown misconceptions and the skepticismof their critics, rational choice theoristsoffer theoretical ingredients – namely awell-developed theory of strategic inter-action – crucial to any remedy to thispredicament.

A Theoretical ProblemAt this point political scientists must re-sist their initial impulse – which is to treateverything as an empirical problem. Thedifficulties that plague efforts to analyzethe nexus of culture and politics are, atbottom, theoretical ones. This point is notnew. Nor is it simply the carping of animpertinent political theorist. A quartercentury ago anthropologist CliffordGeertz offered a remark that still reso-nates as both description and diagnosis.

“Culture . . . is . . . the structures ofmeaning through which men give shapeto their experience; and politics is . . .one of the principle arenas in which suchstructures publicly unfold. The two be-


ing thus reframed, determining the con-nection between them becomes a prac-ticable enterprise, though hardly a mod-est one.

“The reason the enterprise is immod-est, or anyway especially venturesome,is that there is almost no theoreticalapparatus with which to conduct it;the whole field . . . is wedded to an ethicof imprecision. Most attempts to findgeneral cultural conceptions displayed inparticular social contexts are content tobe merely evocative . . . Explicit argu-ment is rare because there are, as muchby design as neglect, hardly any termsin which to cast it, and one is left with aseries of anecdotes connected by insinu-ation, and with a feeling that thoughmuch has been touched little has beengrasped.”

As description Geertz captures the in-timate relation of culture and politics. Asdiagnosis he rightly insists that what welack are the proper theoretical tools withwhich to analyze that relation.

Geertz provided one crucial compo-nent of the “theoretical apparatus” thatwe need to analyze the relation of cul-ture and politics in a systematic way. Heelaborated a revisionist conception ofculture as consisting in publicly sharedsymbols and the cultural practices (i.e.,tradition, ritual, and so on) that social andpolitical actors wield in their ongoing ef-forts to impose conceptual order on oth-erwise indeterminate experience. Politi-cal scientists, by and large, ignored thisconceptual advance. This partly was dueto Geertz’s notorious pronouncements oninterpretive method. His claims on thatscore have been misconstrued and ex-aggerated and are, in any case, only con-tingently related to his conception of cul-ture. Nothing in my argument here rideson received assessments of Geertz’smethodological views. The major reasonthat political scientists ignored Geertz’stheoretical views, however, is that theybecame infatuated with survey methodsand conveniently defined ‘political cul-ture’ in ways that purportedly make itsusceptible to analysis through survey re-search. Ironically, just at the point thatanthropologists reconceptualized culture

in ways that rendered it public and ob-servable, political scientists defined ‘po-litical culture’ in subjective, psychologi-cal terms that rendered it unobservable.All of that is, for present purposes,largely beside the point, so I leave it toone side.

Geertz’s conceptual revision involved“cutting the culture concept down tosize.” Instead of focusing on cultures writlarge, we should, on his account, focuson symbols and the varieties of symbolicpractices through which social and po-litical agents deploy them. The problem,as his critics point out, is that neitherGeertz nor his intellectual progeny everdeveloped much in the way of a theoryof what he calls “symbolic action.” I ar-gue that an adequate conception of sym-bolic action has a prominent and unavoid-able strategic dimension. Indeed, insteadof worrying about culture and politics weshould, I suspect, focus on the inescap-able intersection of symbol and strategyin politics. I first show how models ofstrategic interaction tacitly trade uponsymbolic considerations. I then arguethat symbols have an unavoidable stra-tegic component. In this way I both iden-tify the limits of current analyses of stra-tegic interaction and indicate how thedomain of symbolic interaction invites anextension of strategic analysis.

Symbolic Action Among the RationalChoice TheoristsSamuel Popkin supplies an unlikely ex-ample here. Many treat his book TheRational Peasant as an exemplary em-pirical study by a rational choice theo-rist. Early on in the book Popkin pro-claims that he will treat cultural consid-erations as “givens” in order to advancean exclusively rational choice analysis ofpeasant politics. Toward the end of thebook, however, Popkin reflects briefly onthe decisive role leaders played in facili-tating collective action among Vietnam-ese peasants. He suggests that the suc-cess or failure of these leaders dependedcrucially on their competence and cred-ibility. But, somewhat surprisingly, Popkininsists that those leaders establishedcredibility on “cultural bases.” He at-

tributes the variable success of religiousand political leaders over time to theirdifferential ability “to utilize culturalthemes” to orchestrate “terms and sym-bols” that resonate with their potentialconstituents. In short, on Popkin’s ac-count, political leaders create and sus-tain credibility through symbolic action.They seek to coordinate and mobilize rel-evant constituencies by more or less skill-fully deploying symbolic forms that haveforce over them. Here, Tsebelis notwith-standing, symbolic action emerges un-comfortably but centrally in a purport-edly austere rational choice explanation.

Symbolic action also emerges centrally,if still implicitly, much closer to the‘core’of the rational choice research tra-dition. Specifically, non-cooperativegame theory tacitly incorporates ‘sym-bolic action’ at at least two junctures.First, and most obviously, there is the un-resolved problem of indeterminacy inthose very common instances where agame generates multiple equilibria.Thomas Schelling long ago suggestedthat this problem requires that gametheorists attend to two seemingly irrel-evant issues. The first is what game theo-rists consider “incidental detail” such assymbols and traditions. The second is theway that such incidental detail providesactors with a resource upon which to relyas they make “strategic moves” aimedat defining or redefining the context oftheir interactions. Despite misgivingsabout the elusiveness of these observa-tions, game theorists have yet to improveupon the suggestion that social and po-litical actors engage in symbolic action– that they exploit what Schelling callsthe “symbolic contents of the game” –in their efforts to resolve the pervasiveequilibrium selection problems that gametheoretic models highlight. Conversely,this symbolic action has an inescapablestrategic dimension because, insofar assuch coordination problems typically areasymmetrical, success at endowing oneor another option with salience will havedistributive consequences.

Second, and less obviously, game theo-rists tacitly incorporate symbolic actioninto their standard technique for trans-


and identities.

Symbolic ForceCulture, according to Geertz, affordssocial and political actors with a sort of“symbolic strategy for encompassingsituations,” for imposing conceptual or-der on otherwise indeterminate pro-cesses of interaction. A symbol, in turn,is “any object, act, event, quality, or re-lation which serves as a vehicle for aconception – the conception being thesymbol’s ‘meaning.’” To see how sym-bols work we must, as anthropologistSherry Ortner rightly notes, identify a“comprehensible mechanism” that canaccount for the ways that they influencesocial and political interaction. Thismechanism, on Ortner’s account, mustallow for “a kind of elastic distance”between symbols and relevant popula-tions in the sense that it enables us tosee how symbolic forms affect socialand political interaction without program-ming it.

Geertz differentiates for analytical pur-poses between the “scope” of symbolicforms and their “force.” The scope of asymbol or a cultural practice consists in“the range of social contexts” to whichrelevant actors consider it to have “moreor less direct relevance.” Its force re-fers to its “centrality or marginality” inthe lives of relevant actors, to the “psy-chological grip” it exercises over them.The efficacy of a set of symbolic formsclearly depends on the scope that actorsattribute to it. Yet force takes analyticalpriority over scope. Claims about thescope of particular symbols or practicespresuppose that they have force. If sym-bols lacked such force they would notbe relevant to any social context.

The force of symbolic forms is at bot-tom cognitive or conceptual. Symbolicforms, deployed in cultural practices ofvarious sorts, structure the ways thatactors understand social and political life.They provide actors with, in Geertz’swords, “extrinsic sources of information”not in the sense that they convey detailedmessages, but in a broader sense of im-parting a view of how the world actuallyis and how it operates - the sorts of en-

tities it contains, how those entities canbe expected to behave, and so on. Sym-bolic forms establish the focal catego-ries of social and political interaction and,thereby, establish parameters on beliefformation. They do not (contrary to theviews of political scientists from Almondand Verba through Wildavsky) directlydetermine beliefs or values. Instead,symbolic forms help delineate the rangeof things over which actors might havebeliefs (or, for that matter, which theycan invest with emotion or value). In thisway they constitute the realm of socialand political possibility for relevant ac-tors in at least two analytically separableif empirically related ways.

In the first place, symbolic forms op-erate indicatively to focus the attentionof actors, directing it toward certainranges of alternatives and away fromothers. They allow relevant actors toimpose order, relation, and predictabilityin the face of indeterminacy. They fore-close possibilities. This process is notnaive. Symbolic force discriminates. Bycalling attention to certain identities andoptions, thereby defining them as viableor feasible, it forecloses others. It con-stitutes social and political interactionson particular, partial terms. If symbolicforms operated only indicatively theywould have an unvaryingly constraining,conservative impact on social and politi-cal interaction. But symbols also oper-ate subjunctively to disclose possibilitiesoften not immediately discernable inmundane existence. They thus open op-tions and identities that might go other-wise unconsidered. Orchestrated in moreor less complex cultural practices suchas ritual, for example, symbolic formsgive palpable existence to as yet unreal-ized possibility. They nourish the imagi-nation of social and political actors. Thus,this process is not naive either. By imagi-natively disclosing and exploring possi-bilities actors can, within limits, redefinetheir options and identities.

Symbolic forms, then, exert force oversocial and political actors by command-ing their attention and capturing theirimagination. They govern the mentalcapacities with which actors delimit the

forming games of incomplete informa-tion into equivalent, technically more trac-table, games of complete but imperfectinformation. This procedure is complex.For analytical purposes it reduces allforms of uncertainty to mutual uncer-tainty about the payoff functions char-acteristic of players in the game. Play-ers then construct conditional probabili-ties over the ‘types’ of player that theymight encounter based on commonknowledge of an initial objective distri-bution over possible types. Game theo-rists regularly attribute this initial distri-bution to ‘nature.’ Geertz, by contrast,notes that “the everyday world in whichmembers of any community move, theirtaken for granted field of social action,is populated not by anybodies, facelessmen without qualities, but by somebod-ies, concrete classes of determinant per-sons positively characterized and appro-priately labeled. And the symbol systemswhich define these classes are not givenin the nature of things — they are his-torically constructed, socially maintainedand individually applied.” As anthropolo-gists recognize but game theorists typi-cally do not, it is not ‘nature’ but socialand political agents engaged in symbolicaction who construct the range of pos-sible ‘types’ in any population. Here wesee that symbolic considerations are ana-lytically prior to, and indeed are a pre-condition for, the sort of signalling pro-cesses that Bates analyzes.

For the actors who populate starkgame theoretic models, then, both therange of strategic options they confrontand the range of identities available torelevant players are constituted and con-strained symbolically. Absent some suchsymbolic constraint the strategic inter-actions that game theorists seek to cap-ture in their models remain highly inde-terminate. Yet because symbols con-strain indeterminacy in partial, contestedways, because, that is, they render someranges of options and identities availableat the expense of others, political actorshave a powerful incentive to contestthem for strategic advantage. This, as Iwill now argue, is not a contingent fea-ture of symbolically constituted options


litical entrepreneurs who sought to co-ordinate the collective activities of theVietnamese peasantry encountered thedaunting prior problem of projecting com-prehensible, credible “visions of the fu-ture.” This required that they recast aworld populated by “rational peasants”into one in which other “types,” specifi-cally credible, committed entrepreneurs,were genuinely possible. And it requiredthem to articulate this possibility in “termsand symbols” that had force for relevantconstituencies. Popkin argues that Com-munist organizers did not enjoy successcomparable to that of religious leadersuntil they learned to articulate their vi-sion of the future in indigenous culturalidioms. Prior to that point the “peasantsdid not understand why organizerswere offering to help them and werereluctant to join with them for even smalllocal projects.” Their world was sym-bolically constituted in such a way that itcould not accommodate the entrepre-neurial “type” as a genuine possibility.

Scott’s peasants and Popkin’s compet-ing organizers, like many social and po-litical actors, are parties to what Geertzcalls the “struggle for the real.” Thoseengaged in this struggle seek, with dif-ferential proficiency and success, to ex-ploit symbolic force in the hope of defin-ing the context of their ongoing interac-tions. Their objective is to establish asauthoritative a particular and partial con-ception of the world and the possibilitiesit contains. This struggle is not a contin-gent aspect of culture. Cultures affordrelevant actors both ample opportunityand strong incentives to engage in sym-bolic contests. First, because symbolscan accommodate multiple meaningsthey obviously invite discordant interpre-tations. Second, “cultures” are not seam-less. They inevitably contain intersticesthat provide openings for strategic im-provisation. Finally, although the force ofsymbolic forms does not directly instillindividual values, beliefs or preferencesit does circumscribe the range of possi-bilities over which actors might expresspreferences, values or beliefs. More-over, it does so in a discriminatory man-ner. Symbolic force sustains particular

conceptions of political possibility at theexpense of others and so constitutes apotent strategic resource.

My argument here is obviously incom-plete and preliminary. It nevertheless hasseveral virtues. It outlines a way to ana-lyze the relation of culture and politicsthat allows, indeed requires, that we at-tend to both symbolic and strategic con-siderations. It identifies a “comprehen-sible mechanism” that can generate ex-planations of observable events. Itgrounds “cultural” accounts of politicsin a systematic theory of action. It al-lows us to see conflict and change (andtheir bases) as central to “cultural” ex-planation. It provides a theoretical ratio-nale for generalizing beyond particularcases. (All of these things are notoriousshortcomings of survey based studies of“political culture.”)

These virtues are all theoretical. Theymay not appeal either to resolute defend-ers or devoted critics of rational choicetheory. Yet by indicating how we bothcan and must extend existing rationalchoice models in profitable ways, myargument challenges defenders and crit-ics alike to reconsider rather than sim-ply reiterate their views.

1 Due to considerations of format andlength I provide no citations or notesto substantiate the claims that I makein this essay. I gladly will supply rel-evant references to any reader uponrequest.

Game Theory andCultureDavid D. LaitinUniversity of [email protected]

There is an elective affinity between“culture” and “game theory.” The studyof the former through the techniques ofthe latter should do far better than thegodforsaken marriages in Goethe’s sa-tiric novel Elective Affinities.

Culture has two faces. Its first facereveals the “points of concern” (for ex-ample, individual versus group interests

possibilities embodied in their extant situ-ation and envision those that lie beyondit. Two points are important here. First,because symbolic force does not directlyinstill beliefs, values, attitudes, orienta-tions or whatever, it incorporates pre-cisely the sort of distance that Ortnerdemands. Second, anthropologists regu-larly complain that if social and politicalactors hope to use them for strategicpurposes, symbols must have indepen-dent force. They typically neglect to no-tice that just insofar as symbols do haveforce and just insofar as that force dis-criminates in the ways that I suggest,they offer a nearly irresistible strategicresource. Here we see whyaccommodationist approaches are un-satisfactory - symbol and strategy aremutually implicated in our explanatoryaccounts. And while existing game theo-retic analyses do not fully capture sym-bolic considerations, they, potentially atleast, do offer a systematic way of think-ing about the strategic dimensions ofsymbolic action.

ConclusionConsider two hopefully familiar ex-amples. First, the peasants in the Ma-laysian village where James Scott didfieldwork engaged in “a struggle over theappropriation of symbols . . . over howthe past and present should be under-stood and labeled.” They transformedtwo villagers, Razak and Haji Broom, into“symbols,” “social banners” that theydeploy in an ongoing contest over thedefinition of just and decent social rela-tions. Razak and Haji Broom (poor andrich respectively) each in his own wayupsets prevailing expectations and there-fore embodies notions of behavior thatis unacceptable because it transgress the“symbolic order” of the village. Each isa central character in the cultural prac-tices, the cycles of gossip, rumors, andtales that, “suitably embroidered, elabo-rated and retold,” are the media throughwhich more and less prosperous villag-ers continually seek to define or rede-fine the bounds of that symbolic order totheir relative advantage.

Second, according to Popkin, the po-


in American society) that people whoshare a culture understand as the corequestions that are worth debating. It isthrough debates on these points of con-cern that people form preferences andbeliefs about their world. Culture’s sec-ond face reveals the symbolic resourcesinherent in culture (language, religion,rituals) that are available to political en-trepreneurs who can exploit those sym-bols in order to enhance group cohesionand power. To those who gaze at culture’sfirst face culture looks “primordial;” tothose who gaze at culture’s second faceculture looks “instrumental.” Theoriesthat appreciate only one face, but ignorethe other, will fail to account for culture’srichness and its ambiguities.1

Because of culture’s symbolic contentand its Janus-faced ambiguities, anthro-pologists such as Geertz have insistedthat culture must be “interpreted” ratherthan put into some kind of mechanicalcausal framework.2 I agree with thisadmonition, and wish to argue – givingJames Johnson’s insight a new spin3 –that game theory is a powerful interpre-tive tool well-positioned to reconcileculture’s two faces. To support this ratherunconventional view, I shall here takethree separate lines of attack.

Stability and ChangeConsider the situation of a culturally dis-tinct population that faces assimilatorypressures as the region in which itspeople live gets progressively incorpo-rated into a nationalizing state, whosedominant cultural group is distinct fromthe culture of the regional population.From the point of view of the culturalvirtuosi in the region (Gellner calls it“Ruritania”) it would be criminal for theirmembers to assimilate into the cultureof the state (Gellner’s “Megalomania”).4

Yet from the point of view of a Ruritanianpeasant, migration to a greatMegalomanian city, getting a workingclass job there, matriculating his child intoa Megalomanian school, and watchinghis grandchildren becomeMegalomanians is a bright prospect. Ifa critical mass of peasants is similarlyinduced, Ruritanian culture could practi-

cally disappear. In this case, cultural shiftbased on economic and status incentiveslooks simple. Yet, if a Ruritanian culturalmovement were able to organize politi-cally, raise the costs of migration, andprovide government jobs to speakers ofRuritanian, mass assimilation ofRuritanians into Megalomanian culturewould seem impossible, like bridging aHuntingtonian civilizational divide.

These scenarios of linguistic and cul-tural assimilation can be modeled as aSchelling “tipping game.”5 In it, there aretwo equilibria: nearly all Ruritanians be-coming Megalomanians, and virtually noRuritanians becoming Megalomanians.The beauty of the tipping model is that itcan account for both the instrumental orconstructed nature (to those who studyit and those who wish to change it) andthe primordiality or naturalness (to thosewho live it) of cultural identities. The tip-ping game, by showing both the sourcesof stability (after all, that is what an equi-librium points to) and of change (the mi-cro-incentives of individuals to seek as-similation), gets us beyond the unproduc-tive instrumental and primordial divide.

BeliefsAn important aspect of culture’s firstface is that people who share a culturehave a tacit understanding of what fel-low members of their cultural commu-nity would do in new situations, and theywould thus find it easy to coordinate withmembers of their own culture even with-out specifying precisely the operantnorms. When we are invited to celebrateoccasions for the first time (maybe ananniversary of a divorce), we are oftensurprised at how well we coordinate ourdress with other invitees. Each of us wastrying to assess what each other wouldwear; that entails assessing what eachof the others thought we would be wear-ing; that entails assessing what each ofthe others thought that we thought theywould be wearing; and so forth. Becauseour beliefs are largely correct about eachothers’ beliefs within a cultural commu-nity, this type of coordination is possible.

Game theorists, whether they analyze“focal points” (Schelling) ,“beliefs about

off the equilibrium path behavior” (Greif),or the creation of “common knowledge”(Chwe) are intensely interested in know-ing how these beliefs get formed andhow they are sustained. It is no surprise,then, that a few pioneering game theo-rists have gone to the literature in an-thropology – with field methods uniquelyable to discern people’s beliefs abouttheir neighbor’s beliefs – in order to ex-amine the dynamics of beliefs and howthey impact upon equilibrium selection.6

The promise of this approach is that itprovides a new realm for the isolation ofculture. If by definition “off the equilib-rium path” situations never (or rarely)occur, beliefs about what others will doin such situations can never (or hardlyever) be updated. Therefore, culturalbeliefs about off the equilibrium pathbehavior – although they may drive equi-librium selection – can be derived fromsymbolic imagining, or perhaps from bi-zarre events (“man bites dog”) that oc-cur within a society and become part ofcultural memory, but they cannot be de-rived from empirical generalization. Suchbeliefs, because of their imagined or me-morialized source, help outline culture’sfirst face. Since ethnographic descrip-tion is a powerful way to get at thesebeliefs, this point powerfully illustratesthe affinity between game theory andanthropological interpretation.

MechanismsMy third line of attack is that the macrostudy of culture has largely run its course.It will not likely make new advances untila new generation works out the micro-mechanisms that underlie the macro pat-terns. The real divide in the discipline,therefore, is not “area studies” vs. “so-cial science”7 – as the best area studiesscholars were always engaged in macrotheory – but rather between identifyingcausal patterns among variables (themacro approach) and specifying the un-derlying mechanisms that drive behav-ior.8

The study of political culture, goingback for a century to Weber, has beendominated by macro theorizing. Re-search has sought to identify the broad


historical and ideological forces that in-fluenced individuals living within a soci-ety. Whether the independent variableis “the Protestant ethic” or “Americanliberalism,” the methodological challengehas been to trace the causal impact of acultural variable on social, political oreconomic behavior. In his recent diatribeagainst rational choice, ChalmersJohnson has identified a generation ofclassic works in comparative politics thatfind patterns between macro historicalvariables and political, social and eco-nomic outcomes.9

While the macro tradition has had aglorious era in comparative politics, it hasbeen egregiously weak in working outthe micro mechanisms that translate val-ues on the independent variable to theircorresponding values on the dependentvariable. It is in the specifying of thesemechanisms that game theory has a com-parative advantage. We all know thepower of sacred symbols to mobilizepopulations for enormous acts of cour-age (for example, the Shi’ite symbologyused by the Ayatollah Khomeini); yet weare dimly aware of such symbols flyinglike lead balloons (images of the GreatMotherland War used to hold togetherthe Soviet Union in 1991). Macro theorycan tell us how powerful those symbolscan be; a new generation of micro theo-rists, relying on signalling theory, seek tospecify the conditions when sacred sym-bols can successfully coordinate behav-ior.10 If our goal is to specify mecha-nisms, the model of causal inference mayhave to give way to a new model of thickdescription – far closer to anthropologythan to demography.

ConclusionGame theory holds promise for the fu-ture study of culture. Its models allowus to get beyond the primordial/instru-mental divide of culture’s two faces. Ithas an elective affinity with anthropo-logical interpretation, in that both arehungry for information on beliefs, andbeliefs about others’ beliefs. And itpromises to complement a generation ofmacro theory with a focus on micromechanisms

Footnotes1.This is the principal argument of my

Hegemony and Culture (Chicago:University of Chicago Press, 1977). Agloss of that argument is available inmy “Political Culture and PoliticalPreferences” (American PoliticalScience Review 82:2 (June, 1988), pp.589-93).

2.See “Thick Description” in Interpre-tation of Cultures (New York: BasicBooks, 1973).

3.“Rational Choice as a ReconstructiveTheory,” in K. Monroe, ed., The Eco-nomic Approach to Politics (NewYork: Harper Collins, 1991), pp. 113-42.

4.Ernest Gellner, Nations and Nation-alism (Ithaca: Cornell UniversityPress, 1983).

5.Thomas Schelling, Micromotives andMacrobehavior (New York: Norton,1978). This argument is developed inmy Identity in Formation (Ithaca:Cornell University Press, 1998).

6.Thomas Schelling, Strategy of Con-flict (Cambridge: Harvard UniversityPress, 1960). Avner Greif, “CulturalBeliefs and the Organization of Soci-ety” (Journal of Political Economy102:5 (October, 1994)). Michael Chwe,“Culture, Circles, and Commercials:Publicity, Common Knowledge, andCollective Action” (Rationality andSociety, forthcoming).

7.This is the way Robert Bates sets theterms of debate in “Area Studies andThe Discipline” (PS 30:2 (June 1997)).

8.The best statement of causes andmechanisms in Jon Elster, Alchemiesof the Mind (forthcoming).

9.“Perception vs. Observation” (PS 30:2(June 1997), pp. 170-74).

10.This is one goal in Robert Bates andBarry Weingast, “Rationality and In-terpretation: The Politics of Transition”(paper presented at the 1995 Ameri-can Political Science Association An-nual Meetings, San Francisco).

Culture and the Wagerof Rational ChoiceIan S. LustickUniversity of [email protected]

Concepts, hypotheses, theories,paradigms...what are these? Most fun-damentally they are tools available fordoing what scientists, including socialscientists, do, namely learning about theworld. Because they are good for somethings and not for others, tools are toolsand not magical devices. They may wellhave uses beyond the purposes for whichthey were designed, but they will alwaysbe better for some things than for oth-ers. For many, if not most, tasks theywill actually get in the way. Try using asaw to hammer a nail.

This is all very well and leads easilyinto homilies on the need to match theconceptual and theoretical tools one em-ploys to the problems one wishes tosolve. One problem, however, is that atthe level of paradigms science operatesby endowing one tool with an appear-ance to those who use it that it is not, infact, a tool, but the template of realityitself. This is what happens when a re-search program, with a fecund positiveheuristic, achieves hegemonic statuswithin a scholarly or scientific commu-nity – a community in which some par-ticipants may lose all consciousness ofthe program’s negative heuristic.

For what may be termed “hard core”rational choice theorists, that is for thosewho experience rational choice theoryas “theory” rather than as a form oftheory, the question of what rationalchoice can contribute to the study ofculture takes on a special aspect. Eitherrational choice can contribute to the studyof culture or it cannot. If it can, then it iscapable of uncovering everything we, aspolitical scientists at any rate, would wishto know about the subject. If it cannot,then whatever “culture” might be, itwould not, for these hard core rationalchoicers, be relevant to the world of poli-tics.

But we may leave this approach aside.


Adopting the view that no particular re-search program is hegemonic, or shouldbe, at this stage of the discipline’s devel-opment, we may turn to the homileticmode and ask about the comparativeadvantages of the tools provided by ra-tional choice practioners for the studyof political behavior associated with“culture.”

This requires a definition of culture.Only by defining “culture” can we pro-ceed to ask what about the character ofrational choice theory would make thistool useful or not for studying culture inrelation to politics. I define culture as thearray of symbols, shared expectations,and interactive patterns that limit andstabilize the boundaries of variation ob-servable within groups as individualswithin those groups perform life func-tions. With this conception in mind thereare three aspects of rational choicetheory that suggest its disutility for thestudy of culture.

First, because of the focus on choice,the approach discourages attention toalgorithmic processes – sets or se-quences of behaviors performed with-out those so engaged making any deci-sions or choices. To be sure, rationalchoice might consider such processes asroutines or recipes chosen at some pre-vious point over either non-routinizedbehavior or other routines and recipes.The claim of culture, however, is thathowever the routines and recipes arose,they can operate over time, across gen-erations, and across individuals, withoutany conscious relationship (and perhapswithout any interesting relationship at all)to rationales that could putatively havebeen involved in some discrete and origi-nal choice.

Second, because of its focus on theformalization of rationality, or even theformalization of bounded rationality, therational choice approach has very littleif anything to do with claims about theactual cognitive processes experiencedby, reported, and/or observed in humanbeings when they make choices, or aboutthe actual discursive and deliberativepractices observable within organizationsengaged in choice behavior. Accord-

ingly, rational choice scholars are oftendrawn to models of individual behaviorthat are not only very wrong, but knownto be very wrong, as depictions of whatpolitical subjects actually do, think, andfeel.

For example, the well-known Ferejohnand Fiorina model of voting is based onnever-observed individual calculationsabout minimization of maximum regret.Defending their argument, the authorsreaffirmed their commitment to “joustwith our critics about the descriptiveaccuracy of various decision-theoreticmodels” and conclude that “the availabledata provide no empirical basis for re-jecting the minimax regret model as adescriptive model of the turnout deci-sion.” But the data they treat as “avail-able” are patterns of correlations amongsurvey respondent estimates of howclose a coming election is likely to be,how prospective voters perceive the im-portance of issues in the election, andturnout. An elaborate array of inferencesand imputations then connect these cor-relations to claims about the motivationof individuals. But the real disinterest ofthe authors in the descriptive accuracyof their rationality model at the individuallevel (compared to their real interest inthe explanation of patterns in aggregatedata) is apparent from the fact that noevidence to test the argument is soughtor given from cognitive psychology, sur-vey research about subjectively experi-enced rationales for voting, discussionswith informants, or reporting of the au-thors’ own experience of voting.

The analytic wager represented by thisand most formal theories of rationality isthat despite their inability to illuminatebehavior at the individual level, analystsusing such theories can account for ag-gregate behavior by pretending theirmodels have versimilitude at the indi-vidual level and by relying on marketmechanisms and the laws of large num-bers to produce distributions of outcomesconforming to predictions. By contrast,cultural explanations are intrinsically at-tempts to connect group-level social andpolitical phenomena to motivational andpsychological mechanisms which pro-

duce individual behavior. Therefore onlyif a cultural or psychological analysisyielded evidence that, in a particular set-ting formalist-rational calculations weredriving actual choices by individuals,would rational choice theory be posi-tioned to make substantive contributionsto the ‘content’ of that culture.

Finally, it can be noted that no matterhow rational choice theorists imagine theaggregate outcomes which are their de-pendent variables (whether as a uniform,optimizing behavior throughout the popu-lation or as a normal distribution of varia-tion surrounding the optimal strategy), theapproach directs little or no attention tothe character of the variation within theaggregate and to the possible significanceof the ‘non-dominant’ strategies whichcomprise that variation. This is a severehindrance for any study of culture andcultural change based on processes ofinteraction, selection, and evolutionwithin a cultural repertoire, that is amongsimilar but not identical codes, symbols,or routines. This is not to say that for-mal, deductivist theorizing about culturalchange is impossible, only that formalapproaches of the rational choice vari-ety are likely to be severely handicappedwhen used for this purpose. For example,formal theory and computer simulationsshow great promise for the analysis ofprocesses of cultural change.

These techniques, however, reflect a‘bottom-up’ evolutionary approach (asin the work of Robert Axelrod, andJoshua M. Epstein and Robert Axtell),in which cultural traits are emergentproperties of complex adaptive systems.They stand in sharp contrast to rationalchoice’s ‘top-down’ approach, focusedon intended consequences of deliberate,‘rational’ decisions or retrospective iden-tifications of optimal strategies.

These three characteristics of ratio-nal choice theory limit its usefulness forthe study of culture. There are a varietyof ways, however, in which the specifictools of rational choice can help pose andanswer important questions about poli-tics and culture. Consider the designerof an institution who understands that inaddition to competition within the insti-


tution, actors will also be involved in co-ordination games. The task of the insti-tutional architect is to identify norms (con-stituent elements in the ‘culture’ of theinstitution) that can be fashioned andpromoted to encourage strategies lead-ing to relatively efficient outcomes inthese games. Rational choice theory isa useful way to anticipate these games,probe their dynamics, contribute to thespecification of norms that would leadto relatively efficient solutions, and ac-count for the consequences of failuresto solve coordination problems via ap-propriate codes and standards.

Within a broadly functionalist perspec-tive, rational choice theory can also helpexplain traits, dispositions, and norms asresponses of a population that have been“selected” via market mechanisms, dy-namic patterns of increasing returns toscale, or other evolutionary processes,as effective survival strategies within aparticular mix of environmental con-straints and opportunities (i.e. within aparticular incentive structure). As notedabove, the key here is to understand therole of rational choice as a non-descrip-tive heuristic strategy for seeing patternsin cultural variation – patterns identifi-able as such only if emergent propertiesreflecting the interaction of countlesssmall decisions by individuals are imputedto have been drawn from sets of avail-able ‘strategies’ for non-existent ‘ratio-nal’ entities at higher levels of abstrac-tion, e.g. nation, tribe, gender, or MotherNature. Thus changing standards forsexual behavior in postwar America, forexample, including changing norms re-garding divorce, can be seen as rationalresponses to the changing role of womenin the economy, even though no ‘choice’was ever made to change these stan-dards and norms qua attributes ofAmerican culture. That is, no one imag-ines the operation of a ‘gender high com-mand’ charged with deciding how to re-fashion gender roles in the new socio-economic context. From this same heu-ristic perspective, rational choice theorycan also be used to highlight the effectsof culture by identifying gaps betweenobserved outcomes, or the pace of

change in culture, and that which mightbe predicted as rational from aneconomistic or otherwise narrowly con-ceived formula of self-interest.

When rational choice theorists do claimecological validity for their models theyspecify lists of conditions under which itcan be expected that the calculationsimputed to individuals are actually madeand are actually determinative of behav-ior. One list is provided by Stanley Kelly:“uncomplicated goals for agents, widelyavailable knowledge about ways andmeans to achieve these goals, choicesthat continually repeat themselves,agents who care a great deal about theirgoals, and situations that reward (appre-ciably) choices of efficient means andpunish (severely) choices of inefficientones.” For the most part students of cul-ture will seldom find all or even most ofthese conditions fulfilled, especially inview of the habitualized nature of cul-turally conforming practices. But manyof these conditions are fulfilled when con-sidering the choices which entrepeneursof culture make as they devise gambitsand full-fledged strategies in the kind ofwar of position that is also referred to as‘Kulturkampf.’ David Laitin’s gametheoretic work on the language politicsstrategies available to state-builders andthose who would resist their efforts isan excellent example of how rationalchoice techniques can be used, in thisparticular kind of domain, to model andexplain the actual behaviors and strate-gic commitments of political agents us-ing cultural resources for political pur-poses.

In an earlier issue of this Newsletter, Iindeed argued that rational choice theo-rists themselves were involved in such awar of position over the culture of politi-cal science as a discipline. I identifiedspecific discursive practices as elementsin an ambitious strategy to make ratio-nal choice hegemonically institutionalizedas ‘political science’ and rational choicetheory naturalized as ‘theory.’ In this wayI suggested, and here again suggest, thatrational choice makes yet another con-tribution to the study of culture by pre-senting itself, and its career, as an in-

stance of how politics problematicallyproduces and is reproduced by culture.Indeed applying rational choice to the ex-planation of its own culture highlights itsparticular capabilities, even as the exer-cise illuminates rational choice’s com-parative disadvantages measured againstother available theories.

Rational choice thus does have muchto offer the study of culture, once it isunderstood that what it has to offer fo-cuses attention on very specific parts ofthe cultural ‘elephant.’ For much of whatwe would want to know about cultureand politics is not directly accessible viarational choice. After all rational choicemust, at some level, take preferenceorderings as a priori. Culture, on theother hand, pertains directly to the pro-duction of preference orderings withoutthe presumption that change in thoseorderings can themselves be reduced tohigher level, inherited, or otherwise an-tecedent orderings. But this very pointalso means that understandings of cul-ture can make a vital contribution to therational choice approach. If culturalanalysis can help explain the stability ofa structure of incentives for a commu-nity and also help explain patterns in thechange of such a structure, then rationalchoice theory can be applied in full forcefor the duration of that stability, withinwhat can quite comfortably be under-stood as the boundaries of the boundedrationality that make any choice possible.

References1.John A. Ferejohn and Morris P.

Fiorina, “Closeness Counts Only inHoreshoes and Dancing” (AmericanPolitical Science Review 69 (Septem-ber 1975), pp. 921 and 925) (empha-sis in original). The original article wasJohn A. Ferejohn and Morris P.Fiorina, “The Paradox of Not Voting:A Decision-Theoretic Analysis”(American Political Science Review68 (June, 1974), pp. 525-536).

2.Robert Axelrod, “The Disseminationof Culture: A Model with Local Con-vergence and Global Polarization”(Journal of Conflict Resolution(1997), forthcoming) and Joshua M.


Epstein and Robert Axtell, GrowingArtificial Societies: Social Sciencefrom the Bottom Up (Cambridge: TheMIT Press, 1996).

3.Stanley Kelley, Jr., “Rational Choice:Its Promises and Limitations” (Criti-cal Review 9 (Winter-Spring 1995), p.101).

4.David D. Laitin, “Language Games”(Comparative Politics 20 (April1988), pp. 289-302) and David D.Laitin, Language Repertoires andState Construction in Africa (Cam-bridge: Cambridge University Press,1992).

5.“Rational Choice as a HegemonicProject, and the Capture of Compara-tive Politics” (APSA-CP Newsletter5:2 (Summer 1994), pp. 7, 31) and“Coping with Paradigm Competition:Studying the Culture of RationalChoice as a Case in Point” (PS: Po-litical Science & Politics 30 (June1997).

Rational Choice as aWeberian View ofCultureRon RogowskiUniversity of California, Los [email protected]

Rightly, rationalists have rejectedculturalist explanations’ frequent tautol-ogy and untestability – their propensityto “explain” instances of Germanauthoritarianism or Japanese insularity byinvoking alleged underlying propensitiestoward authoritarianism or insularity inthose respective cultures remains on apar with Moliere’s physician, who proudlyattributed opium’s narcotic effects to thedrug’s “dormitive power.” To restate anattribute is to explain nothing.

Wrongly, rationalists long dismissedalso in culturalist accounts (a) the ob-servation that different groups of peoplesometimes respond quite differently, yetpredictably, to identical incentives and (b)the serious efforts of the best studentsof culture to provide ‘covering law’ ex-planations of how such differences arise,

persist, and (sometimes) are overthrown.Inevitably rationalists have come up

against these very issues in their ownwork. Trying to explain why much ofsub-Saharan Africa has failed to developeconomically, or even to accumulate thehuman capital that would permit devel-opment, analysts as diverse as RobertBarro and Adrian Wood have foundthemselves invoking unknown, but pos-sibly cultural, peculiarities of the region.So have some students and participantsin the effort to revive the post-Sovieteconomies. More dangerously, the per-sistence of sociopathologies among dis-advantaged groups within states (e.g.,growing rejection of education and ac-ceptance of early single motherhoodamong inner-city underclasses in manyadvanced economies) is increasingly at-tributed, even among otherwise devot-edly rationalist scholars and policy ana-lysts, to irreducible cultural differences– so much so that some, as Joe Kleinnotes in a perceptive recent New Yorkerarticle, now accept religious conversionas the most effective remedy.

But of course the best students of cul-ture addressed precisely these issues.Max Weber, the great German sociolo-gist who argued explicitly that one shouldalways seek a “goal-rational” explana-tion first, was nonetheless inspired tosome of his greatest work by the obser-vation that German Catholics around1900 exhibited a lesser propensity to in-vest in human and physical capital thantheir identically-situated Protestant coun-terparts. (Students of management atHeidelberg, he initially noticed, wereoverwhelmingly Protestant, students ofart history – then as now an unremu-nerative specialization – disproportion-ately Catholic.) Yet closer study con-vinced Weber that it was the EuropeanProtestants who were actually history’sodd ducks, and whose intense propen-sity to invest required explanation.

But modern culturalists too often for-get that Weber’s Protestant Ethic andthe Spirit of Capitalism emphaticallydoes not argue that people of differentcultures differ irredeemably, does notsuppose that socialization is destiny, and

does investigate above all how culturechanges. Taking culture not as individualpredispositions but as a socially sharedand logically interrelated set of symbols,codes, and norms – something that suchearly students as Parsons and Shilsrightly analogized to a language – We-ber noted how powerfully medievalCatholic culture anathematized invest-ment (which it normally characterized as‘avarice’), privileged specific kinds ofconsumption (lavish altarpieces, monas-tic endowments), and encouraged trans-fers from the industrious to the idle(almsgiving). A normative structure sostrong and so multiply reinforced, We-ber contended, would have interdictedthe development of modern capitalism.Economic transformation would have hadas its necessary precondition a religiousrevolution like the Reformation – or,more specifically, the Calvinist theologythat bizarrely regarded extreme propen-sity to work and invest as proof of sal-vation.

It seems no great stretch to interpretWeber as portraying a situation of dualcultural equilibria, which modern-dayrationalists could readily translate, with-out great loss of meaning and probablywith some gain in precision and deduc-tive fertility, into different notation and amore formal model. An example thatcomes very close, and that to me is themost impressive demonstration to dateof how rationalists can and should dealwith culture, is Avner Greif’s “CulturalBeliefs and the Organization of Society,”(Journal of Political Economy 102(1994): pp. 912-50). Considering the cul-tural requisites for successful long-rangetrading (including use of physically re-mote agents) in the Mediterranean of theearly Renaissance, Greif models a situ-ation in which identical actors wouldcoordinate on one of two equilibria: one(roughly typified, he believes, by Jewishmerchants resident in the Muslim worldduring the early Renaissance) that de-pended on personal knowledge, mutualtrust, and a “grim trigger” strategy of ex-communicating any agent who betrayedany principal’s trust; or an alternative (ex-emplified, perhaps, by the nascent


Genoese non-Jewish merchant class)that employed one-shot contracts, wasmore accepting of occasional betrayal,and demanded less extensive informa-tion about potential partners. Among themany rich (and, to me, Weber-like) im-plications of Greif’s argument is this: thatthe first, more ‘traditional,’ culture willhave been more stratified and will haveoffered less opportunity for social mo-bility.

Whether one finds Greif’s particularapproach congenial or not, it illustratesthe three things that any social-scientifictreatment of culture must do:

1. regard all actors as governed bythe same fundamental laws of be-havior (and not invent one socialscience for Americans, anotherfor Russians, yet a third for Ira-nian fundamentalists, etc.);

2. understand culture not as a setof individual propensities but asa coordinator of strategies andexpectations among independentbut mutually reliant actors, i.e. asa social institution akin to lan-guage; and

3. offer a coherent explanation ofwhat sustains, and hence also ofwhat can change, culture evenamong fully socialized adults.

Again, some of the earlier students ofculture tried to do exactly this. GabrielAlmond’s pioneering 1956 article on po-litical culture in the Journal of Politics,or David Laitin’s Gramscian attempt ofthe mid-1980s to explain the absence ofreligious conflict among the Yoruba ofNiger (Hegemony and Culture (Uni-versity of Chicago Press, 1986)), soughttrue covering-law explanations of culturalvariety, survival, and change. It is a threadthat too much of mainstream culturalisttheory has subsequently ignored and thatnow must be re-addressed. Currently, themost promising way of doing so is as partof a multiple equilibrium story among ra-tional actors.

Rational Choice andCultureGeorge TsebelisUniversity of California, Los [email protected]

“Why did the dinosaurs disappear?” Iasked my three year old son. He did notunderstand that it was a rhetorical ques-tion and answered with conviction, “Be-cause they died.” There are lots of ar-guments in political science stated withequal seriousness and with similar tau-tological qualities.

Consider the answers to the questionof why people vote. Several decadesago, this was not a question in the mindsof political scientists. At that time, em-pirical studies concluded that people withhigher income or more education weremore likely to vote than people of lowersocioeconomic status, but all of thesestudies took voting (as well as non-vot-ing) for granted. Only after the seminalwork of Olson did political scientists (par-ticularly those subscribing to the rationalchoice research program) ask the ques-tion, “Why do people vote?” There is noshortage of rational choice ‘explanations’of voting. I focus on one in particular,which claims that people vote becausethey derive satisfaction from the act ofvoting. There are several variants of this‘psychic income’ approach: ‘consump-tion value,’ ‘D term’ and other versions.One thing is certain: having heard suchexplanations, the enquiring mind knowsnothing new. While it may be true thatpeople vote because they like to, this doesnot constitute an explanation (rationalchoice or any other kind) of voting. Thereason is that the added value introducedby the statement “people vote becausethey like to” is nil, or very little.

Added Value As An Evaluative Crite-rionI submit that the major criterion by whichwe should evaluate scientific work is notwhether it is theoretical, empirically valid,consistent with what we know alreadyor iconoclastic. These are important cri-teria, but they are subordinate to whether

and how much a piece of work affectsour prior beliefs. Whether our beliefs arefalsified (as with the paradoxes of so-cial choice, starting with Arrow’s Theo-rem) or corroborated (as with evidencethat the composition of committees inCongress affects policy outcomes), im-portant work significantly affects our pri-ors. It makes us understand somethingthat we did not understand before, in-forms us of something that we did notknow, changes our minds about how theworld works or reinforces beliefs thatare otherwise diffuse and/or unjustified.

This is the yardstick that I will apply inthe remainder of my argument, so thereader should try to evaluate the crite-rion at this point. Does the work that youlike provide significant added value, andthe work that you dislike little or none?When we hear fascinating presentations,can we not readily summarize the mainpoints while with trivial work we leavethe room saying, “So what?” If this isthe case, then added value is the mostimportant evaluative criterion, and we areon safe ground when we apply it to anyfield of work, including the variants ofrational choice analysis that deal with po-litical culture.

The purpose of any ‘analysis’ includ-ing rational choice is not to say whathappens, but to explain why knownevents or empirical regularities happen.For example, we know that plurality elec-toral systems are associated with twoparty systems (Riker has traced state-ments of this association back some 150years). An analysis (known as‘Duverger’s law’) explains that this as-sociation is not accidental, but due to twoeffects: the mechanical (that pluralityelectoral systems favor big parties) andthe psychological (that voters who un-derstand the mechanical effect will avoid‘wasting’ their votes on small parties).Duverger did not use rational choice ter-minology, but the essence of his argu-ment is that voters perform expectedutility calculations and don’t vote for par-ties which have a low probability of win-ning. Duverger’s account has significantadded value, because he persuasivelyexplains the mechanism underlying a long


recognized but little understood associa-tion. He organizes our beliefs about theworld such that we expect plurality elec-toral systems to lead to two party sys-tems. This new prior is so strong thatwhen we find countries where the asso-ciation breaks down (like Canada or In-dia), we need to explain why these vio-lations of Duverger’s law occur.

Culture, Rational Choice and AddedValueWhat is the role of culture in rationalchoice analysis? There are three signifi-cantly different ways that culture ap-pears in rational choice arguments. Thefirst (and most frequent) is to use cul-ture as a constraint along the equilib-rium path, the second is to use cultureas information for equilibrium selection(in both these cases, culture is used asan independent variable) and the third isto use culture as a dependent variable.I will argue that there is an hierarchy ofadded value among these three variants.The first approach – even in the bestcase – provides little added value, thesecond adds significantly in our under-standing of the world and the third bestcombines rational choice and culture.1. Culture as independent variableRational choice analysis assumes that in-dividuals are goal-oriented and try tomaximize the achievement of their goals,given existing constraints. The basic con-cept for rational choice analysis is ‘equi-librium.’ Equilibrium is a situation fromwhich no rational actor has an incentiveto deviate. If a rational actor had an in-centive to deviate, then she would notselect that option and we would not ob-serve that outcome as an equilibrium.

From the above discussion, it is obvi-ous that the selected actions depend onthe existing constraints. What is the na-ture of these constraints? Some of themmay be imposed by existing institutions.For example, in parliamentary systemsmost of the bills considered by the par-liament are introduced by the govern-ment. Similar rules may define what kindof amendments (if any) are permitted,who is recognized from the floor,whether discussion will be made on the

basis of the government proposal or ofthe corresponding committee report, andso on. Institutional approaches to poli-tics (rational choice or not) focus on in-stitutions as the independent variablesthat explain human action.

Other constraints may be imposed bythe choices of other actors. For ex-ample, the government may not admit aparliamentary amendment, or a witnessmay be treated as hostile (different ruleswill apply to her) or you may be late foran appointment because of a traffic jam.

Finally, constraints may be imposed bysome person’s beliefs, ideology or cul-ture. I may be not be allowed to beararms by my religion. I may believe (asChristians did long ago) that asking forinterest on a loan is immoral because itis equivalent to charging for time, whichis a gift from God. Such restrictions (forthe people who believe in them) are noless real than those in the previous cat-egories. I may not violate the law be-cause the police are present, or I mayobey because I believe that compliancehas an inherent moral or transcendentalvalue. My beliefs predict my behavioras well as institutional constraints do; insome cases, they may provide more ac-curate predictions.

While cultural accounts of human ac-tion may be true, there is a significantdifference in their explanatory value. Ex-planations by culture or ideology may betrivial. Under what conditions will a cul-tural explanation be trivial, as opposedto non-obvious? The crucial differenceis whether culture is used to define anactor’s choice directly (as a constrainton an actor’s behavior) or to define anactor’s response to the constraints ofother actors (in which case culture isused as information for equilibrium se-lection).a. Culture as a constraint along theequilibrium pathSuppose that a model predicts some(maximizing) set of actions, but thatsome actors do not follow the prescribedbehavior. The cultural explanations thatthey did not think of it, did not have thecognitive capacities, were prohibited bytheir ideology or culture from acting in

this way or ‘reacted spontaneously’ aread hoc. Even if true, these assertionsdon’t explain anything, and worse yetthey often seem invented just to save themodel. We do not account for behaviorby identifying the categories of actors(whether our categories are ethnicgroups, genders, races or even ideolo-gies). Arguing that ‘Italians’ have ‘sub-ject’ culture, as Almond and Verba do,even if accurate, does not constitute anexplanation of their behavior, merely arelabelling.

I am afraid that this is too often thepattern in scholarly articles, not onlythose that belong to the rational choicetradition. For example, explanations ofvoting on the basis of ‘party identifica-tion’ have an unpleasant tautological ringto them (people with Democratic identi-fication vote Democratic). Similarly, ge-netic explanations (Germans behave dif-ferently than Italians because they havealways done so) certainly don’t push thelimits of imagination. So tautologies arenot exclusive to rational choice analysis.What is particular to poorly wrought ra-tional choice articles is that they dressup tautological arguments with a ratio-nal choice vocabulary. But familiar vo-cabulary does not mean that we under-stand the phenomenon better – as theinitial example of voting indicates.

I have tried to distance myself fromthis use of culture in rational choiceanalysis in Nested Games. This kind ofanalysis gives rational choice approachesa bad name among scholars who studyculture. They justifiably believe, afterreading a tautological ‘explanation,’ thatthey have learned nothing new. Fortu-nately, this is not the only intersection ofculture and rational choice.b. Culture as an equilibrium selectionmechanismAndre Malraux was General de Gaulle’sMinister of Culture. He contributed notonly to the content of the General’sspeeches but also to the selection of thetime and place that the General deliv-ered his speeches. He made his selec-tions to maximize the cultural impact ofeach speech. Today, media consultantsadvise candidates to package themselves


in order to have maximum impact, giventhe preferences, biases, stereotypes,beliefs, ideologies and cultures of the pub-lic. The scholarly works of Tarrow andPopkin describe how revolutionary lead-ers study and incorporate the culture oftheir followers into their strategies. In allof these examples, beliefs about the cul-ture of other people affect the optimalstrategies of some actors. Culture helpsanswer the question of why particularactors select particular courses of ac-tion as optimal.

I emphasize that cultural reasons mayexplain not only the selection of strat-egies, but also their avoidance. One ofthe procedures used by the EuropeanUnion for legislative decisions is the co-operation procedure. The important fea-ture of this procedure is (I simplify herefor the sake of the argument) that a pro-posal emanating from the Commissionand the Parliament can be accepted bythe Council with a qualified majority, whileit can be modified only with unanimity.

Consider the figure below. The statusquo (SQ) is outside the area defined bythe ideal points of the members of theCouncil (the numbers), while the Parlia-ment (EP) and the Commission (C) thatmake the proposal fall on the other sideof the Council. The required qualifiedmajority in the Council is 5 of 7.

What proposal will the Parliament andthe Commission (hereafter denoted P+C)make, knowing that the Council cannotmodify their proposal except by unanim-ity? P+C will make the proposal X thatmakes the pivotal member of the Coun-cil (3) almost indifferent between X andthe status quo. Indeed, in this case 3 willprefer X to the status quo, and so will allthe members to his right (4, 5, 6 and 7).Note that in this example we did not usethe power of the Council to modify theproposal.

Suppose now that we learn from theempirical literature on the EuropeanUnion that the Council is a consensus-oriented body which tries to reach deci-sions by unanimity whenever possible,and that most of the time there are noformal votes. (This information is readilyavailable in the EU literature.) Does this

information about the culture of theCouncil affect the proposal of P+C? Yes,because unless they make a proposalthat makes every member of a qualifiedmajority better off than any possibleunanimous decision, some potential mem-ber of the qualified majority will success-fully make that unanimously favored pro-posal.

Follow my reasoning in the figure. TheCouncil can unanimously approve any-thing in the area [SQ, SQ’] where SQ’is symmetric to the status quo with re-spect to the ideal point of the voter piv-otal for unanimity (1). The new P+C pro-posal Y makes the qualified majority piv-otal member of the Council (3) just shyof indifferent between it and SQ’ (whichhe can get by unanimity). So the restric-tion of the off-equilibrium beliefs of P+Cleads us to the selection of equilibriumY rather than X.

Note that had the empirical literatureon the Council found that one countrywas always in the minority and that theothers never tried to incorporate it intotheir bargains (same institutional rules butdifferent culture), X would emerge againas the predicted equilibrium.

The reason that this use of culture pro-

duces added value is that the selectionof strategy is not intuitively obvious. Theactor did not select this course of actionbecause of some constraint on his ownbeliefs or capacities, but because of theinformation he possessed about the otheractors’ cultures and, therefore, theirlikely courses of action.2. Culture as a dependent variableAn even more interesting way of look-ing at culture is as a dependent variable.In general, the assumption of rationalityand the use of game theory does not re-strict the number of equilibria very much.Indeed, under conditions of incompleteinformation (the norm in politics) or re-peated play (also quite frequent), equi-libria are infinite and the real question ishow to select among them. For example,while the outcome of a confrontationwith complete information may never bea war (a point raised as an argumentagainst rational choice analyses by un-informed critics), war becomes a pos-sible equilibrium with incomplete infor-mation.

One way of understanding ‘cultures’is as such manifold equilibria. In thisconceptualization, different equilibriacome from different antecedent condi-

S Q S Q ’ Y X C E P

1 2 3 4 5 6 7

S Q 1 1S Q ’ S Q ’3 3Y S Q 3 3 X = , > , >

R e str ic ting o ff-equ ilib r ium be lie fs to a con sensu s c ounc il leads to the se lec tion o f Y as the equ ilib rium . R e str ic t ing o ff - equ ilib r ium be lie fs to a d iv ided c ounc il leads to the se lec tion o f X a s the equ ilib r ium . S ee the tex t fo r d eta ils .

F igu re . E qu il ib r ium S elec tion on the B a sis o f D iffe re n t O ff-E qu il ib riu m B elie fs

.


tions. If one wants to explain why cer-tain rural cultures practice female infan-ticide, she may assert that parents con-sider their children to be assets or liabili-ties; if physical strength leads to survival,parents will keep male babies.

Another example will make my pointmore clearly. Suppose that two peopleare to divide a dollar. Any division of thedollar that leaves no residual is an equi-librium. If the amount to be divided issignificant, disputes among individualscan last forever.

Up to the 1970s, bargaining was abranch of cooperative game theory inwhich criteria of ‘fairness,’ symmetryand mathematical elegance produceddifferent solutions. Ariel Rubinstein bril-liantly produced a game which simulatedreal bargaining: Player One makes anoffer for a division of the dollar to PlayerTwo. If Player Two accepts, the gameends; if not, he makes a counter-offer toplayer one. If Player One accepts thecounter-offer, the game ends; if not, thegame goes continues until the two play-ers agree. To bring the game to an end,Rubinstein endowed his players with ‘im-patience’ – that is, a preference for thegame to end sooner rather than later. Hethereby calculated a unique perfect equi-librium as a function of who makes thefirst offer and the levels of impatienceof the players. If we call the level ofimpatience (the time discount factor) ofeach player d, the final division of thedollar gives the first player x=(1-d

1)/(1-

d1d

2).

What is interesting in this approach isthat if both players are infinitely patient(d

1 and d

2 tend to 1), the final outcome

is x=1/2. So the familiar Western habitof splitting the difference evenly can bederived as the equilibrium outcome of agame if both players are infinitely pa-tient. The same outcome results if theplayers are not infinitely patient, butequally patient and equally likely to movefirst. If in a different society men madethe first move, the split of the dollar wouldnot be symmetric but would favor men.I suppose (although I do not know it fora fact) that in some cultures men andwomen do not split dollars (or other cur-

rencies) equally.This is an example where ‘culture’ is

the equilibrium corresponding to a seriesof exogenous conditions (sequence ofmoves, impatience). Rubinstein selectsthe unique perfect equilibrium from theinfinite possible equilibria, and this is whatgives power to his result. It may, how-ever, be the case that the set of perfectequilibria is infinite, in which case ana-lysts will look for some additional refine-ment that further restricts the predictedoutcome.

ConclusionsCultural studies produce a wealth of in-formation about how different people –from Africa to Capitol Hill – think andbehave. If these reports yield beliefs,behaviors and rituals that we did not pre-viously recognize, then they produceadded value. Their existence does andshould alter the way we analyze thesesocieties. Rational choice does not haveanything to offer to such studies, butmuch to learn from them. Repetition ofthese studies with a rational choice vo-cabulary helps neither tradition. Ratio-nal choice contributes by incorporatingthese cultural findings into the rationalcalculations of actors. Even better, itenables researchers to understand thereasons why particular cultural patternsemerged as equilibria from the wide va-riety of possible behaviors.

Cultures and Modesof RationalityAshutosh VarshneyHarvard [email protected]

Rational choice theory has made re-markable contributions to two subfieldsof comparative politics. It has deepenedour understanding of political economyissues – especially the politics of eco-nomic growth and distribution. And itsability to explain behavior in highly insti-tutionalized settings – as in the rule-gov-erned universe of a Western bureau-cracy, legislature and executive – hasbeen strikingly impressive. Politics, how-

ever, is not just political economy, nor areall forms of politics highly institutional-ized – especially in the developing world.

Consider how different ethnic conflictis from legislative or bureaucratic battlesover economic policy. Ethnic conflicts area form of mass politics marked by highlyrisky or costly forms of behavior in whichethnic partisans not only kill but are will-ing to die. Just as it is hard to explain –given rational calculations of cost andbenefit, why people vote – it is also hardto understand – with tools of rationalchoice – why so many people in theworld demonstrate ethnic fervor or em-brace nationalism. From an individual per-spective, the instrumental benefits of par-ticipating in nationalist mobilization areobvious only under two strict conditions:(a) when nationalists are already closeto capturing power and much can begained, or anticipated losses cut, by join-ing the bandwagon; or (b) when law andorder have broken down, ethnic animosi-ties have soured group relations, andeven neighbors of longstanding belong-ing to a different ethnic group can’t betrusted, creating a “security dilemma” forindividuals (Posen, 1993) and makingpreemptive violence against neighbors ofa different ethnic group an exercise inpersonal security (Hardin, 1995).

These extreme conditions constitute arather small proportion of the universeof ethnic conflict. The former Yugosla-via, Rwanda and Burundi are not typi-cal; they are simply the most dramaticand gruesome cases of ethnic conflict.Violence may be common in ethnic con-flicts, but a complete disintegration of thestate is not. The latter breakdown hasseldom marked ethnic conflicts in Asia,Europe and North America. This doesnot mean that there are no risks or costsassociated with participation in ethnicmobilizations in societies where the statehas not collapsed. Risks of incarcera-tion, injury and death remain, but in theabsence of state disintegration, ethnicconflicts don’t produce security dilem-mas. By and large, the situation from anindividual perspective can be summa-rized as follows: the benefits of partici-pation – a better job, a political office –


have moved from behavior under cer-tainty to that under risk, uncertainty andincomplete information, especially withthe use of subjective probabilities underBayesian decision rules.

Cognitive psychology heavily critiquesthe economic concept of rationality. Onthe basis of experimental data, this cri-tique suggests that rationality, as speci-fied in economic models, is impossible.In making decisions, human beings re-act excessively to current information(ignoring prior information, thereby mak-ing Bayesian probabilities irrelevant), areinsensitive to sample size (thereby mak-ing reliability of information irrelevant todecisions), and respond to how thechoice-set is framed rather than whatthe choice-set is. Thus, economic ratio-nality is a normative, not a descriptive,notion. The leading proponents of thisview are Amos Tversky and DanielKahneman (1990). Some of the seminalrational choice scholars in economicshave become quite favorable to thesepsychological theories (Arrow, 1982).

The concept of rationality in philoso-phy is still different. In philosophical dis-cussions, rationality refers to “reasonedassessment as the basis of action” (Sen,1982). Such an assessment can be basedon self-interest, but also on larger val-ues. Self can be broadly defined – interms of group goals, religious values,aesthetic considerations, etc. This is notsimply a philosophical abstraction. Phi-losophers claim also that many in reallife are driven by such considerations.

These three concepts of rationalityhave come to acquire different labels.The economic view has become synony-mous with the term ‘instrumental ratio-nality,’ the psychological view with‘bounded rationality,’ and the philosophi-cal view – or the view in that branch ofeconomics which remains aligned withphilosophy and is today most commonlyassociated with Amartya Sen – is sim-ply called ‘rationality’ with no prefixesattached. Instead, in philosophical treat-ments, the various forms of rationalityare, more often than not, freely admit-ted. This larger view would also includewhat Max Weber called “value rational-

ity”. In Economy and Society, Weberhad categorized social action into fourtypes: instrumental-rational, value-ratio-nal, norm-oriented (based on conventionsand traditions without critical delibera-tion) and affective or impulsive (express-ing anger, envy, love, etc.).

The alternatives to instrumentally ra-tional behavior are, thus, not simply emo-tional or irrational behavior. Of the fourWeberian categories of human action, thefirst two are goal-directed; only one isinstrumental-rational. Instrumental ratio-nality entails a strict cost-benefit calcu-lus with respect to goals, necessitatingthe abandonment or adjustment of goalsif the costs of realizing them are too high.Value-rational behavior is produced bya conscious “ethical, aesthetic, religiousor other” belief, and is more or less cost-inelastic. Behavior, when driven by suchvalues, can entail great personal sacri-fices. Some spheres of life – value-ra-tional individuals would argue – are notup for sale or compromise.

Value-rationality does not mean thatthe values expressed by such behaviorare necessarily laudable. Indeed, thevalues in question may range from his-torical prejudice vis-a-vis some groupsor belief-systems to goals such as dig-nity, self-respect and commitment to agroup or a set of ideals. Likewise, value-rational acts can range from long-runsacrifices to achieve distant goals on theone hand to violent expressions of preju-dice on the other.

Which of these categories of behav-ior is represented by the term ‘rationalchoice?’ Almost without exception, it isinstrumental rationality with which ratio-nal choice theorists identify. They eitherdo not speak of goals, concentrating in-stead on the means; or they assume thatself-interest is the goal of human action.Some other standard positions also markrational choice. Proponents of rationalchoice theories believe that universaltheories of human behavior – includingpolitical behavior – can be formulatedwithout consideration of cultural con-texts. Moreover, considerable resistanceremains to the idea that different moti-vations can underlie behavior in differ-

may accrue far into the future or not atall, but the likely costs – incarcerationand injury, if not death – are often fartoo obvious. Still, a large number ofpeople participate in ethnic mobilizations.Moreover, martyrdom remains a widelynoted phenomenon in such conflicts. Astrictly rational choice explanation can’texplain why, given the risks of participa-tion on the one hand and the distanceand uncertainty of benefits on the other,such movements or mobilizations takeoff and gather momentum. Once theyhave gathered momentum, it is easier toexplain, in a rational choice framework,why people join them.

Ethnic partisanship is just one exampleof culturally driven behavior. Less dra-matic forms of politics – withdrawal frommainstream politics by some groups, ordemand for a certain conception ofschool education – can also be rooted inculture. Can rational choice make a con-tribution to the study of the less dramaticforms of cultural behavior? If so, in whatways?

To answer these questions, we firstneed to ask what rationality is. Are theterms ‘rational choice’ and ‘rationality’interchangeable? We need to inquirewhether rational choice theories, as op-posed to rationality, can explain why cul-tures exist, and how they might deter-mine human behavior.

It is not often realized that the threedisciplines that have dwelt most on thenature of rationality – economics, psy-chology and philosophy – perceive it verydifferently. In economics, rationality hastwo meanings. First, it means consistencyof choice: if I prefer A over B and Bover C, then I must prefer A over C.The second meaning is identical with self-interest. Action is rational if it is aimedat realizing self-interest. If costs of anaction outweigh benefits, self-interestwould not be served; hence a cost-ben-efit calculus accompanies analysis basedon self-interest. Following the economicconcept of rationality, we not only havetheories of individual rational behavior(utility theories) and models of rationalbehavior of two or more interacting in-dividuals (game theory), but theorists


ent spheres of life: that it may be per-fectly rational for human beings to beinstrumentally rational when buying a car,but value-rational when examining ques-tions of national liberation or of genderbalance, affirmative action, andmulticulturalism in the universities. Fi-nally, rational choice also remains highlyskeptical of the notion that individualaction can be rooted in group values orinterests rather than in self-interest.

Can rationality conceptualized as in-strumental rationality explain the role ofculture (or religion) in human life? Canit explain why and how culture mightshape behavior?

As already stated, instrumental ratio-nality is used in two ways: either it isdeployed as a conception of the means,not of the ends, while the ends remainunspecified; or self-interest is assumedto be the end of human life. In eithercase, rational choice cannot explain someof the fundamental puzzles of human lifewith which cultures deal. Can societieslive without notions of right and wrong?Can human beings live without ideas thatcan guide them as to how to relate tothe family, the community and lovedones? Students of culture would claimthat these are some of the central ques-tions in their field. Many also claim thatdominant cultural practices concerningthe family and the community, and, some-what less so, the dominant notions ofright and wrong tend ultimately to berooted in religious traditions. Secularhomes and societies do have cultures;even secularized cultures owe a histori-cal debt to their religious foundations.Religion and culture are not interchange-able terms, but they have had a deep in-terrelationship historically.

Very few religious traditions of theworld elevate self-interest and worldlymatters into the highest moral obligationof human beings. Sikhism and the Puri-tan sects of Protestantism come readilyto mind. In such traditions, self-interestbegins to acquire a moral status. In othertraditions, self-interest can at best givehuman beings their immediate or inter-mediate ends, not their ultimate ends orvalues. In these traditions, self-interest

may be seen as a necessity in severalspheres of life, but not in all, nor do thesetraditions view self-interest as a higherend or value.

Instrumental rationality, in short, is notabout values. Moreover, there may bespheres of life where most human be-ings can’t do without such values. Thisidea has been very effectively expressedby some of the greatest rationalists ofthe century. Albert Einstein, for example,has written insightfully about the rela-tionship between rationality on the onehand and religion and cultural traditionson the other. To illustrate what is at is-sue here, let me quote from Einstein atlength:

“Knowledge of what is does notopen the door directly to whatshould be... One can have theclearest and most complete knowl-edge of what is, and yet not be ableto deduct from that what should bethe goal of human aspirations. Ob-jective knowledge provides us withpowerful instruments for theachievement of certain ends, but theultimate goal itself and the longingto reach it must come from anothersource. And it is hardly necessaryto argue for the view that our exist-ence and our activity acquire mean-ing only by the setting up of such agoal and of corresponding values...Here we face, therefore, the limitsof the purely rational conception ofour existence...”

“To make clear these fundamen-tal ends and valuations, and to setthem fast in the emotional life of theindividual, seems to me precisely themost important function which reli-gion has to perform in the social lifeof man... And if one asks whencederives the authority of such funda-mental ends, since they cannot bestated and justified merely by rea-son, one can only answer: they ex-ist in a healthy society as powerfultraditions, which act upon the con-duct and aspirations and judgmentsof the individuals; they are there, thatis, as something living, without it

being necessary to find justificationfor their existence...”

“[A] conflict arises when a reli-gious community insists on the ab-solute truthfulness of all statementsrecorded in the Bible. This meansan intervention on the part of reli-gion into the sphere of science; thisis where the struggle of the Churchagainst the doctrines of Galileo andDarwin belongs. On the other hand,representatives of science have of-ten made an attempt to arrive at fun-damental judgments with respect tovalues and ends on the basis of sci-entific method.. . These conflictshave all sprung from fatal errors.”(Einstein, 1954. pp. 42-5)

Seen this way, rationality and religionbelong to two different realms of humanexperience – the former having little todo with the ends of life. For those unin-spired by religion and some of its ex-cesses, however, culture – a set of insti-tutions and normative practices that welive by – has been a source of such val-ues. Culture replaces religion in the ag-nostic or unbelieving homes.

A rational choice theorist may say thatindividuals create culture (or religion).What appears as an inheritance todaywas created by individual acts in the past,making it possible for a methodologicalindividualist to explain the existence ofculture instrumentally. In a fundamentalsense, this view cannot be correct. Cul-ture may indeed have been created byindividuals, but each individual engagedin such acts of creation also acted in re-lation to an inherited set of practices. Inorder for an individual to create, affirm,deny or innovate a set of cultural prac-tices – and a good deal of that happensin everyday life – there has to be a pre-existing set of normative practices in theframework of which the creation, affir-mation, denial or innovation acquiremeaning. As philosophers of languageare fond of saying, a sentence or wordhas no meaning until a language exists.The acts of creation, innovation or de-nial draw their rationale, negative or posi-tive – from an existing set of values. Cul-


ture, in this sense, is embedded in ourlife; it exists as a framework of mean-ing within which human deliberation andrationality operate. That is why it is nota privately underprovided public good, aswe should expect if we are true to ratio-nal choice. Rather, to borrow a phrasefrom Charles Taylor (1995), it is an “ir-reducibly social good.”

To conclude, cultural choice or behav-ior is different from buying a car or ahouse on the one hand and forming po-litical strategies to defeat adversaries forpolitical office on the other. Rationalchoice theories may be more applicableto marginal decisions – or to decisionsabout political strategies in legislaturesor elections, and less so to decisionsabout how people choose fundamentalvalues. And for those spheres of lifewhere these values guide us – in manybut not all families, in many but not allcommunities, and in many widely prac-ticed religions of the world – we need torework our view of rationality. Behaviorthat appears to be highly principled orrisky may be value-rational – i.e., ratio-nal with reference to these values — butirrational by rational choice canons ofjudgment. Finally, whether or not cultur-ally driven behavior is rational, such be-havior exists in plenty.

ReferencesKenneth Arrow, “Risk Perception in

Psychology and Economics” (Eco-nomic Inquiry 20:1 (January 1982)).

Albert Einstein, “Religion and Science:Irreconcilable?” in Ideas and Opin-ions (New York: Crown Publishers,1954).

Russell Hardin, One for All (Princeton:Princeton University Press, 1995).

Barry Posen, “Ethnic Conflict and Se-curity Dilemma” in Michael Brown,Ethnic Conflict and InternationalSecurity (Princeton, Princeton Uni-versity Press, 1993).

Amartya Sen, “Rational Fools” inChoice, Welfare and Measurement(Cambridge: MIT Press, 1982).

Charles Taylor, “Irreducibly SocialGoods,” in Philosophical Arguments(Cambridge: Harvard University

Press, 1995).Amos Tversky and Daniel Kahneman,

“Judgment Under Uncertainty” in PaulMoser, Rationality in Action (NewYork: Cambridge University Press,1990).

Max Weber, Economy and Society(Vol. 1) (Berkeley: University of Cali-fornia Press, 1978).

(continued from page 4)

Session titles and invited participants

1. Trade unions and employers’associations: towards greaterdependence on the market orinstitutionally-based variations inpower?Chair and introduction: Jelle Visser(University of Amsterdam)Invited papers and discussants:Miriam Golden (UCLA), JanineGoetschy (Université de Paris X), ColinCrouch (European University Institute),Torben Iversen (Harvard University),Jesper Due (University of Copenhagen)

2. New and old sources of work forcesegmentation (by gender, ethnicity,occupational structure, stability ofemployment) and the fate of solidarity.Chair and introduction: DavidMarsden (London School ofEconomics)Invited papers and discussants: GöstaEsping-Andersen (University ofTrento), Fausto Miguélez (AutonomousUniversity of Barcelona), RichardLocke (MIT), Rainer Zoll (Universityof Bremen), Martin Gannon and StanleyNollen (University of Maryland andGeorgetown University)

3. Industrial relations in the workplace:collective representation versus directemployee participation, conflict versuscooperation.Chair and introduction: WolfgangStreeck (Max-Planck-Institut, Köln)Invited papers and discussants: KeithSisson (University of Warwick), IdaRegalia (University of Turin), KathleenThelen (Northwestern University), PaulMarginson (Leeds University), AlainChouraqui (LEST, Aix-en-Provence),P. Gunnigle (University of Limerick)

4. Industrial relations and the politicaleconomy: decline versus re-emergenceof tripartite concertation.Chair and introduction: Marino Regini(University of Milan)Invited papers and discussants:Philippe Schmitter (European UniversityInstitute), Franz Traxler (University ofVienna), Peter Lange (Duke University),Michael Shalev (Hebrew University ofJerusalem), Anton Hemerijck(University of Rotterdam), DieterSadowski (University of Trier)

(continued on page 24)

News & Notes

Use the Newsletter in the class-room. The APSA has authorizeduniversity teachers to reproducearticles from the Newsletter for

use in the classroom at no charge.Take advantage of this policy,and introduce your graduate

students to the latest research,issues and debates in compara-

tive politics.


Peter Hall responds to the presiden-tial column of Robert Bates in volume7, issue 1 of the Newsletter. The origi-nal column is available on theNewsletter’s web site. See page threefor details.

Comparative Politicsand Area StudiesPeter A. HallHarvard [email protected]

The active discussion in the APSA-CPNewsletter and elsewhere about theappropriate relationship between areastudies and the study of comparativepolitics is an important one with seriousinstitutional implications. Many depart-ments are rethinking how they hire schol-ars in comparative politics. Foundationsare reconsidering their support for areastudies, and interdisciplinary centers forarea studies face intellectual and finan-cial challenges in many universities. Atstake is the future both of comparativepolitics and area studies, and on this out-come will depend America’s knowledge-base about other nations in an increas-ingly interdependent world. For that rea-son, lively debate should be welcomed.

However, if “truth proceeds morereadily from error than from confusion,”as Francis Bacon asserted long ago,there is a risk that this debate willfounder, since it is marked by pervasiveambiguities about both the nature of areastudies as practiced today and the alter-native forms that the study of compara-tive politics might take. Whether areastudies deserves our support dependsvery much on how we define it, and theoutlines of a comparative politics divorcedfrom area studies remains as yet unclear.

The concept of ‘area studies’ has threequite distinctive connotations. First, it issometimes used to refer to studies thatprovide detailed description of a nationalor regional case but remain uninterested

in generalizing beyond the case at hand,i.e. to studies that do not seek to provideor address general propositions thatmight illuminate the politics of other na-tions or regions. Second, it can refer tothe acquisition of a relatively deep andcontext-rich knowledge of the politics ofa specific nation or region that is thenused as the basis for developing propo-sitions of more general applicabilityacross nations or regions. Third, the termis sometimes used to refer to the organi-zation of teaching or research via inter-disciplinary clusters of scholars groupedtogether, more or less loosely, in a pro-gram or center oriented towards the poli-tics, history and social complexion of aparticular region.

The value of ‘area studies’ to com-parative politics, and to the social sci-ences more generally, depends heavilyon which sense of the term one is using.In my view, even area studies in the firstsense of the term, namely, the study of aparticular nation or region for its ownsake, has value. There is much to belearned from studying a culture on itsown terms; and those who fail to do soare fated to misunderstand other nationsand peoples, a potentially serious prob-lem in an era when the prosperity andsecurity of all nations depend heavily oninterchange with other peoples. As sev-eral contributors to this debate havepointed out, even those social scientistsmost committed to the development of‘portable truths’ must rely on such workif they are to generate accurate gener-alizations.

However, it is important to recognizethat, in political science, economics andsociology at least, area studies in this firstsense of the term ceased for the mostpart to be practiced many years ago. Anattack on area studies of this sort wasmounted and largely won in the 1950sand 1960s on the grounds that researchin social science had to be oriented to-ward generalizations that would be use-ful for understanding politics or society

in many national contexts. As a result,the kind of ‘area studies’ that predomi-nates in political science today is stronglyoriented toward producing general propo-sitions capable of informing our under-standing of politics across nations. Insome cases, that involves the develop-ment of causal propositions meant to holdacross societies; in others, the genera-tion of concepts and categories of po-tentially wide applicability. Some excep-tions can still be found, but it would be aprofound disservice to the vast majorityof scholars in comparative politics not torecognize that what they practice is anarea studies in the sense of the seconddefinition given above.

Defined in this sense of the term, areastudies has great value for comparativepolitics and the social sciences moregenerally. To those who have done workin this vein we owe some of the mostfruitful propositions to emerge from thediscipline of political science as a wholeand much of our understanding of na-tions other than the United States. In-deed, it is the study of American politicsthat seems to have been relatively lim-ited, at least from time to time, by a fail-ure to ask whether and how the propo-sitions it generates might apply acrossother nations.

The value of area studies defined inthe third sense – with regard to areacenters – is a slightly different issue be-cause it speaks to the fruitfulness of al-ternative ways of organizing teaching, re-search and intellectual interchange in thesocial sciences. If discipline-based aca-demic departments did not exist, I wouldargue for their invention because, al-though it is a contestable proposition, Ibelieve that undergraduate and gradu-ate teaching should reflect the ‘disci-pline’ that such disciplinary organizationprovides. However, in most institutions,such departments do exist and the issue,then, is whether it is useful to have areacenters in addition to them.

This can appropriately be the subject

Continuing Debates


of debate since area centers are notcostless in either financial or intellectualterms. Their central purpose is to en-hance interdisciplinary exchange amongscholars interested in a specific regionof the world. To the degree that it suc-ceeds, it diminishes the intensity of in-terchange within the separate disciplines.However, a number of considerationsargue strongly for the value of area cen-ters. First, much of the cutting-edgework in the social sciences has long beengenerated by scholars working on theborder of one discipline and another, in-formed by precisely the kind of interdis-ciplinary exchange that area studies fa-cilitate. Second, given the strong hold thatdisciplines exercise over virtually all theother modalities of scholarly organiza-tion, the presence of area centers seemsa small incursion beyond such bound-aries; and it is my experience that theinsights colleagues bring from such in-terdisciplinary forums tend to enrich in-terchange within departments more thanthey detract from it. Third, the visitingscholars and speakers that area centersattract from the regions that they studyprovide students and faculty membersalike with direct insights and informationfrom the region that might otherwise bevery difficult to acquire.

The other set of issues implicated inthis debate about the relationship be-tween area studies and comparative poli-tics centers on the conception of com-parative politics invoked, explicitly orimplicitly, as an alternative to what ispracticed today. By and large, those whocriticize area studies do so in order toadvance another conception of how com-parative politics should be studied. But,again, it is useful to distinguish amongthe different images of this alternativethat have become jumbled together in thecurrent debate.

First, there is the image of a compara-tive politics oriented toward the genera-tion of fruitful cross-national generaliza-tions. Those who associate area studieswith a purely descriptive or idiographicendeavor frequently argue for such analternative, but, as noted above, this is abattle that was fought and largely won

twenty years ago. To invoke it againseems to be little more than a rhetoricalexercise. Most area specialists in com-parative politics already value and pur-sue research based on cross-nationalgeneralization.

Second, one might construe the alter-native as a comparative politics basedpredominantly on the framework of en-quiry associated with rational choiceanalysis. Many of those who are cur-rently critical of area studies seem tohave this alternative in mind. However,there is a danger here of confusing twoseparate issues: the demand for gener-alization on the one hand, and the de-mand that those generalizations becouched in terms of rational choiceanalysis on the other. Rational choiceanalysis is by no means the only frame-work capable of generating general hy-potheses or propositions in political sci-ence, and its advantages or disadvan-tages as a framework for enquiry shouldbe assessed independently of the debateabout area studies. In my judgment atleast, these advantages are by no meansfirmly established enough for one to as-sume that the only good kind of analysisin comparative politics is an analysisfounded on rational-choice precepts.

Third, some suggest that comparativepolitics should ideally be defined as in-volving comparison across regions. Here,the issue seems to be how hard onewants to press this particular demand.Cross-regional studies are clearly valu-able and we have too few of them, butthis is not a coincidence. Apart fromlarge-scale statistical studies, which ex-ist in some number but are appropriateonly for some problems, cross-regionalcomparisons are inordinately expensiveand difficult to do with any degree ofaccuracy – expensive because of thetime and funding required to gather pri-mary data in multiple sites and difficultbecause of the effort required to acquireaccurate information, including relevantcontextual information and the associ-ated language skills, across regions. Pre-cisely for these reasons, we should con-sider such studies valuable and worthyof support but perhaps not press them

on everyone interested in doing com-parative politics.Whether cross-regionalresearch should be privileged overcross-national research or even overnational case studies (often comprehend-ing sub-national cases) involves the as-sessment of some complex heuristic is-sues that are beyond the scope of thisbrief note. However, it is important toobserve that, when it comes to the se-lection of cases in a context of scarceresearch resources, more cases andmore regions are not automatically su-perior to fewer. There are many topicsfor which a comparison of Latin Ameri-can nations may be more suitable than aresearch design that compares a LatinAmerican nation to an African nation.In instances in which the results of theanalysis depend heavily on the intensivegathering of data, a single national casemay even produce superior results, pro-vided the case selection and frameworkof analysis is defined in suitably com-parative terms. The issues that suchchoices pose have been with the fieldfor some time and are addressed by along and distinguished literature.

In sum, the debate about the role ofarea studies in comparative politics islikely to make progress only if it remainsclear about the multiple issues that areinvolved and about the labels attachedto them. In particular, while area cen-ters undoubtedly vary one from another,it would be wrong to associate them ex-clusively with the kind of old area stud-ies that concentrates on purely idio-graphic enquiry. Large numbers of themcontain and nurture scholars who areactively contributing to disciplinary agen-das and to the formulation and testing ofgeneral propositions of the sort that havelong been central to advance in the so-cial sciences.

Similarly, while it is important that doc-toral students whose work focuses onareas of the world other than the U.S.acquire the theoretical knowledge andmethodological tools that are central tothe discipline, it is also important that theyhave intellectual and logistical support forsecuring an adequate knowledge of thecountries about which they write; and


area studies centers have long been cru-cial to the provision of such support. Intheir absence, we risk graduating stu-dents who know a great deal about thearcane details of our writings but verylittle about the part of the world that issupposed to feature in their own.

As for the future of comparative poli-tics, I suggest that we should distinguishmore clearly between three separatequestions. The first asks: how muchknowledge of the region one studies doesgood work require? The second asks: isinterdisciplinary knowledge of the sortpromoted by area studies centers usefulfor social science enquiry? And the thirdasks: what kind of theoretical frame-works offer the most promise for ad-vancing our knowledge of the politicalworld? The answer to the first questionwill depend upon the issues and rangeof one’s research but, in general, com-parative politics will be impoverished ifthose who study other nations of theworld do not have a good first-handknowledge of their politics and society.The answer to the second may be amatter of taste but I remain of the viewthat interdisciplinary interchange can beof great value to political science. Al-though many political scientists currentlyprefer to engage in such interchange witheconomics, there are also insights to befound on the borders of history, sociol-ogy and anthropology, not to mentionseveral other disciplines. As to the lastquestion about the choice of theoreticalframeworks, all I will say here is thatwe would be well advised to rememberthat it is a perennial one, which shouldconcern us all, but not one that is likelyto be advanced by neglecting or abjur-ing detailed knowledge about the poli-tics of other nations.

single nation, but more than one partymay be discussed in that chapter – andthere may be more than one chapter pernation. The organization and theme ofeach book is up to its editors to de-cide.

Indigenous parties scholars in theMiddle East, Asia and Africa who areinterested in exploring the possibility oftaking part in this project are invited towrite to the Workgroup Organizer, KayLawson, Department of Political Sci-ence, San Francisco State University,San Francisco, California 94132 (ex-cept January through May:Departement de Science Politique dela Sorbonne, 17, rue de la Sorbonne,75005 Paris France), or to the News-letter Editor, Andrea Rommele,Mannheimer Zentrum für EuropaischeSozialforschung, Universitat Mannheim,Steubenstrasse 46, D-68131Mannheim, Germany. Please enclose avita and a letter saying which party orparties are of interest to you. We wouldalso be pleased to hear from partyscholars from other regions who havenames (and addresses, please!) to rec-ommend.

(continued from page 21)

Recruiting Possible Authors forChapters in Forthcoming Volumeson the Parties of Asia, Africa, andthe Middle East

The Workgroup on Parties and Elec-tions is a sub-group of the Committeefor Political Sociology which is affiliatedwith both the International Political Sci-ence Association and the InternationalSociology Association. Nearly 300 ofthe world’s parties scholars are enrolled.The Workgroup is presently sponsor-ing the publication of a series of vol-umes on contemporary political parties,with an emphasis on the internal life ofparties. The first volume, How Politi-cal Parties Work (edited by KayLawson), was published by Praeger in1994, and includes studies of partiesfrom around the world. Subsequent vol-umes are devoted to the parties of asingle region, including The Organiza-tion of Political Parties in SouthernEurope (edited by Piero Ignazi andColette Ysmal) and Political Cleav-ages and Parties in Eastern and Cen-tral Europe (edited by Kay Lawson,Andrea Rommele, and GeorgiKarasimeonov), which are under con-tract to Praeger and are expected toappear in late 1997. Planning for afourth volume on parties in the MiddleEast is underway, as is that for a fifth onAsian parties. A sixth, on the new po-litical party systems in Africa, is plannedfor the more distant future.

A key feature of all the Workgroup vol-umes devoted to particular regions isthat the scholars who write the indi-vidual chapters are themselves citizensand residents in the nations whose par-ties they cover. Each chapter covers a

News & Notes

The Newsletterappears twice

annually.

The deadline forthe winter issue is

December 15.

The deadline forthe summer issue

is June 15.


International Finan-cial Regulation forTwenty-one Coun-tries, 1950-1994Dennis P. Quinn and A. MariaToyodaGeorgetown [email protected]@guvax.georgetown.edu

National economic policies have changedremarkably over the past fifty years.Some of the most dramatic changes haveoccurred in the regulation of internationalfinance. Scholars have been hampered,however, in studying changes in interna-tional financial regulation by the absenceof adequate data measuring regulationacross time and space.

We offer a catalogue of data measur-ing both the forms and degree of restric-tiveness of domestic laws that regulateinternational financial transactions. Thecatalogue allows us to compare acrosscountries over time the restrictivenessof government regulations on inward andoutward current and capital accountflows. Data are available for 21 mem-ber countries of the Organization forEconomic Cooperation and Developmentfor the years 1950-1990. The cataloguedescribes which nations regulated inter-national finance, when and by how much.The source for the data is the text of theInternational Monetary Fund’s AnnualReport on Exchange Arrangementsand Exchange Restrictions.

Our systematic description of interna-tional financial regulation offers twomajor benefits to interested scholars.First, the data describe precisely overtime and space variation in magnitudeand form of regulation. Second, the dataseries have enough observations so thatit may be used in either cross-sectionalor individual country regression analy-ses, which will assist political economistsin sorting through some widely varying

hypotheses about the origins and effectsof financial liberalization.

The Data SourceSince 1950, the International MonetaryFund (IMF) has issued its Annual Re-port on Exchange Restrictions (laterre-titled, Annual Report on ExchangeArrangements and Exchange Restric-tions), which reviews the restrictionsgoverning international financial transac-tions found in the laws and internationalagreements.

The laws reviewed in the annual re-port are those that regulate either the pay-ments to or receipts from non-residentsby residents. Some reported laws gov-ern the proceeds of transactions (e.g.,exchange surrender requirements); oth-ers govern the underlying transactionsthemselves (e.g., license requirementsfor direct foreign investment). Trade re-strictions are covered by the text only“insofar as they are closely integratedwith exchange restrictions.” (IMF 1950,p. 1) Both current and capital accountrestrictions are reviewed in ExchangeRestrictions.

In the initial 1950 volume, the IMF dis-tinguished between “exchange pay-ments” and “exchange receipts.” Under“exchange payments,” the IMF sepa-rately reported restrictions on paymentsfor imported goods, restrictions on pay-ments for “invisibles” (i.e., internation-ally traded services such as insuranceor legal services) and restrictions on pay-ments or transfers of capital abroad.Under “exchange receipts,” the IMFseparately reported restrictions on theproceeds from exports, the proceedsfrom services (invisibles) transactions,and the proceeds from inward transfersof capital (e.g., foreign direct invest-ment).

The treatment of the laws regulatinginternational financial transactions is con-sistent across time and space: ExchangeRestrictions has used almost all thesame categories regarding legal regula-

tion from 1950 to the present. This con-sistency and continuity allowed us tosystematically convert the qualitative textto quantitative indicators of each nation’slevel of financial openness in each year.

CodingWe developed coding rules based on theregulatory taxonomy described by Ar-row (1973) and IMF (1950). Arrow sug-gests that, in general, regulations andadministrative rules that forbid or limitbehavior are inherently more restrictivethan is taxation of the same behavior. Intheir initial assessment of governmentlaws regulating international finance, theauthors of Exchange Restrictions con-clude similarly, noting in particular thatthe quantitative or administrative regu-lation of international transactions aremore restrictive than are multiple cur-rency practices (i.e., taxation) (IMF1950, pp. 4, 13).

In coding and assigning scores to thelaws reported in Exchange Restrictions,we follow the six category format de-vised by the IMF in assessing financialrestrictions: three describe restrictions onexchange payments (imports, invisibles,capital) and three describe restrictionson exchange receipts (exports, invisibles,capital). For each of the six categories,we score on a scale of zero to two withhalf-integer intervals (i.e., 0, .5, 1, 1.5,2). Zero represents the most severe fi-nancial restriction and 2 represents theabsence of restrictions, with half-integerincrements representing intermediatecases.

We present separate scores for capi-tal account transaction restrictions(CAPITAL) and current account trans-action restrictions (CURRENT). CAPI-TAL is scored on a 0-4 scale (the sum ofthe scores for capital exchange pay-ments and capital exchange receipts),and CURRENT is scored on a 0-8 scale(the sum of the scores for imports, ex-ports, payments for invisibles, and re-ceipts of invisibles).

Data Sets & Archives


We also include a measure of interna-tional laws and agreements that constraina nation’s ability to restrict exchange andcapital flows (on a 0 to 2 scale). Themeasure AGREE reflects, for example,agreements among member nations ofthe OECD and the European Union(EU), and acceptance of other bindingrules (e.g., acceptance of the IMF’sArticle 14 status). We include this mea-sure because, when nations agree to jointhe OECD, IMF, or EU, they also agreeto liberalize some aspects of their finan-cial markets. AGREE may therefore beinterpreted by international investors asimplying commitments that they find tobe more credible than changes in domes-tic laws in isolation. (See Pindyck 1991.)

Adding AGREE, CAPITAL, and CUR-RENT produces a 0-14 measure of in-ternational financial regulation, with 0representing a closed economy, and 14representing a fully open economy. Thefull measure is called OPENNESS.

Other Indicators of International Fi-nancial RegulationThe IMF provides a rough measure ofthe presence or absence of financial re-strictions, using a “-,.” indicator in a table,“Summary Features of Exchange andTrade Systems in Member Countries,”which appears at the back of ExchangeRestrictions from 1967. A dash, “-”,shows the absence of restrictions on ei-ther current payments or capital pay-ments; a vertically centered period, “·”,shows the presence of a restriction ofsome type on either current or capitalpayments. Many scholars have con-verted the IMF’s “-,·” indicators into a“0,1” measure, and used it in regressionanalysis.

The 0,1 measure has serious deficien-cies as a measure of international finan-cial regulation, a point readily acknowl-edged by most scholars using it. Wemention a few here. (See Quinn andToyoda 1997 for a more comprehensivereview of these limitations.) First, theindicator of the presence of regulationcontains no information about the mag-nitude of a nation’s financial restrictionson current or capital payments. This pre-

sents a particularly severe limitation be-cause the majority of reported cases arethose where government laws imposesome restrictions on international finan-cial transactions, but where the economyis neither fully closed nor fully open.Second, the table does not contain infor-mation about important aspects of finan-cial openness. The most important omis-sion is that restrictions on inward capitalor current flows are not reported. A thirdproblem is that, in a few cases, the lawsreported in the text of the document donot appear to match what is recorded inthe table. A fourth, particularly severe,problem involves the use of the categori-cal 0,1 measure as a dependent variablein regression analysis, which is almostalways done using techniques like logitor probit. The data are characterized byinertia, which implies the presence ofserial correlation in regression analysis.The estimation of standard errors in logitand probit analysis is biased downwardin the presence of serial correlation.Strategies for correcting for serial cor-relation are less developed in logit andprobit analysis, however, than in ordinaryleast squares analysis. The measure re-ported here has enough information toallow for use of least squares proce-dures.

Haggard and Maxfield 1996 deviseda measure of international financial regu-lation for four countries, 1970-1990, fromthe same source. Their measure is notdirectly comparable, however, as it con-tains information about domestic bank-ing laws.

Data AvailabilityThe catalogue, a complete description ofthe coding rules, and related papers areavailable from the authors. Address in-quires to either Dennis Quinn or A.Maria Toyoda.

References and Related PapersArrow, Kenneth J., “Economic Effi-

ciency and Social Responsibility”(Public Policy 21 (Summer 1973), pp.303-13).

Haggard, Stephan, and Sylvia Maxfield,“The Political Economy of Financial

Internationalization in the DevelopingWorld” (International Organization50 (Winter 1996): pp. 35-68).

International Monetary Fund, AnnualReport on Exchange Arrangementsand Exchange Restrictions (previ-ously Annual Report on ExchangeRestrictions, 1950-1978). (Washing-ton: IMF, 1950-1990).

Pindyck, Robert S., “Irreversibility, Un-certainty, and Investment” (Journalof Economic Literature 29 (Decem-ber 1991), pp. 1110-1148).

Quinn, Dennis P., “International CapitalFlows: A Twenty-One Country Studyof Financial Liberalization 1950-1988”(Paper presented at the 1992 AnnualMeeting of the American Political Sci-ence Association, Chicago, Ill.)

Quinn, Dennis P., “The Correlates of In-ternational Financial Regulation”(American Political Science Review91 (forthcoming)).

Quinn, Dennis P. and Carla Inclán, “TheOrigins of Financial Openness: AStudy of Current and Capital AccountLiberalization.” (American Journalof Political Science 41 (July 1997):pp. 771-813).

Quinn, Dennis P. and A. Maria Toyoda,“Measuring International FinancialRegulation.” (Manuscript.)

Book Reviewers Welcome

Doctoral students at any in-stitution are welcome to sub-mit proposals for book re-views. To do so, contact theeditor and/or assistant editorwith the name of the recentlypublished book that you wishto review and a short (ap-proximately 200 word) dis-cussion of the book’s impor-tance and why you wish toreview it. Reviewers are re-sponsible for procuring theirown copies of books.


Open Economy Politics:The Political Economy ofthe World Coffee TradeRobert H. BatesPrinceton University PressPrinceton, 1997

Reviewed by Matthew KocherUniversity of [email protected]

As with many good books in the socialsciences, the excellence of RobertBates’s Open-Economy Politics: ThePolitical Economy of the World Cof-fee Trade is driven by the brilliant iden-tification of empirical puzzles. Batestakes a handful of geographical and tem-poral slices from the recent history ofthe international coffee trade, and showsthat, in crucial instances over the courseof the twentieth century, the behavior ofstates involved in the coffee trade hasdeviated substantially from what sys-temic theories of the internationaleconomy, on their own, would predict.Bates makes a compelling case that, forBrazil, Colombia and the United States,domestic political institutions, and thestructure of incentives they created,were at least as significant as determi-nants of behavior in the internationalcoffee market as systemic causes andconstraints.

Brazil’s response to the conditions ofthe international coffee market in theearly part of this century represents acritical case for systemic theories of in-ternational political economy. Followingthe leaf rust epidemic which devastatedAsian coffee production during the latenineteenth century, Brazil became a veri-table market hegemon, controlling be-tween 70% and 90% of world exportsduring the interval 1900-1920. Systemictheories predict that states occupying amonopoly position in international mar-kets should employ their market domi-nance to manipulate prices. For a sig-

nificant portion of the time that it enjoyeda monopoly on coffee production, Brazilfailed to exploit its position to fullest ad-vantage.

Bates shows that Brazilian political in-stitutions and the structure of domesticinterests limited the ability of coffee pro-ducers and their political allies to forge adomestic coalition responsive to the con-ditions of the international market. Bra-zil under the Old Republic was highlyfederalized, with individual states con-trolling their delegations to the nationallegislature. Coffee production was con-centrated in a small number of stateswhich, while politically powerful, couldnot by themselves muster a legislativemajority. Thus, coffee interests werecompelled to tie market regulation toeconomic reforms benefiting other do-mestic interests in order to secure a win-ning coalition.

In particular, the appreciation of thenational currency which would accom-pany an increase in the international priceof coffee threatened the interests of foodproducers by making imports more com-petitive. By tying export restrictions tocurrency valorization, Brazilian coffeeproducers were able to secure the legis-lative coalition necessary to carry outtheir economic program. On the otherhand, artificially maintaining the value ofthe national currency entailed substan-tial risk of inflation, which was damag-ing to the interests of other crucial politi-cal interests, notably the military. In theevent, Brazilian coffee interests wereunable to act on the strategic opportu-nity presented by their hegemonic posi-tion in the international coffee marketuntil 1906, and they periodically failed tosustain a program of export restrictionuntil the rationalization of coffee policyunder Vargas.

To the extent that Brazilian coffee in-terests did succeed in artificially main-taining the international price of coffee,it provided a public good to the free-ridingcompetitive fringe of coffee-producing

nations. Colombia, in particular, aggres-sively exploited its advantage by under-cutting the price of Brazilian coffee andrapidly expanding both production andmarket share. Although this outcome isconsistent with the predictions of sys-temic theories, Bates shows thatColombia’s competitive market rational-ity was also the product of a domesticpolitical game, one which raises impor-tant empirical puzzles as well.

Unlike the relatively concentrated pro-duction of Brazil, Colombia’s coffeegrowers were mostly small peasant pro-ducers. How did they overcome thecosts of organization and avoid govern-ment predation to secure a favorableprogram of export promotion? Batesshows that coffee producers occupied apivotal median voter niche in the elec-toral competition of republican Colom-bia. Thus, the structure of political insti-tutions created incentives for politiciansnot only to pursue the votes of peasantproducers, but also to pay the costs oforganizing coffee interests into a nationalfederation which was capable of pursu-ing an aggressively competitive tradepolicy.

At times, Bates may press his caseagainst systemic theories too strongly.For instance, he argues that U.S. par-ticipation in the International Coffee Or-ganization (ICO) was inconsistent withsystemic rationality because it fostereda cartel which ultimately raised the priceof coffee for U.S. consumers. U.S. se-curity concerns over Castroism dur-ing the late 1950s and early 1960s ex-plains why the US would agree to helpthe coffee-producing nations restrict pro-duction: the International Coffee Agree-ment offered the U.S. a way to subsi-dize the economies of Latin Americanstates, which were perceived to be un-der siege by Communism. However,security does not explain the timing ofU.S. action, which was postponed by arecalcitrant Congress through some ofthe worst periods of U.S./Soviet secu-

Book Reviews


rity competition. On these grounds,Bates argues that systemic theories“don’t work.”

Nevertheless, on Bates’s own account,it appears that two systemically deter-mined national interests explain almosteverything about U.S. participation in theICO: both the impetus for a coffee-con-suming nation to commit itself to produc-tion limits and the initial resistance of theU.S. Senate to such an agreement. Po-litical pressure from U.S. coffee roast-ers may have helped to tip the balancein favor of the ICO (as Bates argues),but it is equally clear that, absent the per-ceived Soviet threat, the ICO wouldnever have come into being. A moresympathetic interpreter of this data mightargue that systemic theories of interna-tional politics and political economy farewell here, though they fail to exhaus-tively explain some of the details of U.S.participation. Indeed, the U.S. case ischaracteristic of the findings of this book:for each case examined, it is not so muchthat systemic causes were not operative,but rather that political actors on the in-ternational stage were simultaneouslyconstrained by their need to play domesticpolitical games.

In the concluding chapter of Open-Economy Politics, Bates reveals hismethodological anxiety over a studywhich seeks to make generalizable in-ferences from the study of a single com-modity and a single institution. He wor-ries that a study of this sort may be use-ful only as a set of disconfirmatorycounterexamples to existing theory(which, indeed, is its greatest virtue),rather than being a contribution in its ownright to the theory of the internationalpolitical economy. This concern promptsBates to provide three supplementaryempirical vignettes intended to improvethe reader’s impression of the robust-ness of the book’s findings by increas-ing the sample size. In general this con-cluding material is not of the same highquality as the balance of the book; muchof it could have been folded into earlierchapters.

Furthermore, this concluding move be-lies Bates’ commitment to an ‘analytic

narrative’ methodology, in which histori-cal events of intrinsic interest are micro-modeled in order to provide a rich set ofpredictions concerning the individual‘case.’ In effect, analytic narrative is it-self a form of empirical disaggregationwhich increases the number of obser-vations which can be drawn from a singletemporal or geographical unit. The mod-els developed to explain particular casesmay well not apply cross-sectionally be-cause they formalize peculiar domesticinstitutions. Instead, Bates’s analyticnarratives suggest a rich set of toolswhich students of political economy willfind useful to employ in the investigationof other cases and topics in IPE.

Remaking the ItalianEconomyRichard LockeCornell University PressIthaca, 1995

Reviewed by Angelo Del PrioreNorthwestern [email protected]

Despite prolonged periods of rampantinflation, high unemployment and politi-cal instability, the Italian economy hasperformed exceptionally well in the post-war period, outperforming many otherOECD member countries on numerousmacro-economic indicators. Disparitiesbetween the North and the South andeven within the regions, however, haveundermined traditional cross-nationalcomparative studies. Few studies thatoffer national level explanations can ac-commodate the Italian case within theirframework. On the other hand, Italianarea specialists have long recognized theimportance of regional differences, andhave sub-divided Italy into two, three ormore regions in order to develop morenuanced understandings. While com-parative studies have generally failed tocapture many of the varied details thatoccupy Italy specialists, the latter grouphave typically found their work of lim-ited applicability to other states. Into thisgap steps Richard Locke’s Remakingthe Italian Economy, which proposes

an alternative framework for the studyof political economy.

Rather than view Italy as an excep-tional case, Locke proposes that itmerely exhibits more clearly the hetero-geneity that exists within all countries.Therefore, instead of viewing nations ashomogenous units, we should study thedifferences within them since they areat least as significant as those amongthem. Locke’s sub-national, yet com-parative, study of the local social andpolitical factors that affect economicreform and industrial transformation pro-vides a convincing portrayal of Italianeconomic and political developments inthe ‘80s, and retains a level ofgeneralizability lacking in many single-country case studies. Although Locke’sbook focuses on one country, its rel-evance extends to scholars whose inter-ests include industrial transformation,industrial relations, and the socialembeddedness of economic actors in theadvanced industrialized countries.

After citing various economic mea-sures that demonstrate the Italianeconomy’s laudable performance, Lockesets about to prove his argument thatnations should be treated as heteroge-neous rather than homogenous units. Hedoes this by studying several failed at-tempts by the Italian central governmentto impose unifying economic reforms andthe incapacity of various sociopoliticalactors to transform the industrial rela-tions system in an orderly and predict-able fashion. These efforts not only failedto import and institutionalize systemsexisting in Italy’s seemingly more suc-cessful neighbors, but actually exacer-bated intranational differences in someinstances. The diversity that character-izes Italy not only continues to exist, butis crucial to understand why companiesadjusted differently to comparable eco-nomic challenges despite being similarlysituated. Following others (Putnam,Making Democracy Work, 1993),Locke finds that these initiatives unin-tentionally fostered regional differences,but that deregulation has not necessarilyled to free markets’ unfettered rule(Regini, Uncertain Boundaries, 1995).


Locke moves beyond these and otherscholars, however, by treating adminis-tratively demarcated regions as complexand internally differentiated units that donot conveniently correspond to culturalboundaries, and by bringing to the forethe political factors that influence par-ticular firms and industries’ adjustmentstrategies. Through a detailed, compel-ling analysis of specific instances of ac-tors dealing with industrial transforma-tion, Locke clearly illuminates the causalmechanism that links micro-political fac-tors to labor and management’s strate-gies, and shows how local sociopoliticalnetworks can outweigh macro-variablessuch as national wage bargaining ac-cords.

In the latter part of the book, Lockepresents detailed case studies of the au-tomobile manufacturing and textile indus-tries efforts to restructure in the 1980s,which further demonstrates the influenceof local sociopolitical networks. To dem-onstrate that the same technology canbe harnessed in different ways, provid-ing varying amounts of benefits for capi-talists and labor, Locke examines the his-torical experiences and worldviews, aswell as the material and institutional set-tings, of different groups. The type ofsociopolitical network, which can be oneof three ideal types – policentric, hierar-chical, and polarized, affects actors’ con-straints and opportunities. While FIATand Alfa Romeo share the same own-ership, utilize similar technologies andinteract with the same unions, the localcontext shaped management and labor’schoices and led to radically differentstrategies and outcomes. Whereas dis-agreements in the FIAT plant, located inTurin, escalated into pitched battles inwhich neither side was willing to backdown and the unions were eventuallycrushed, Alfa Romeo and its employeesbenefited from Milan’s more conciliatoryenvironment. In Milan sociopolitical net-works are organized in a policentric fash-ion, in which power and information aremore evenly distributed; this fosters con-sensual agreements. In Turin, however,the predominance of polarized relationsled to a zero-sum approach between the

policy areas – and not just industrial policy,Locke’s subnational comparative ap-proach should find wide applicability. Evenif the European Union fails to reach itsmost ardent backers’ goals, the generalideological movement towards laissezfaire ensures that the existing social en-vironment will play a comparatively sig-nificant role as central governments de-volve power to the market and local in-stitutions and increasingly restrict indus-trial policies.

While Locke makes a convincing ar-gument and provides a firm footing forfurther study, the Italian case is one ofthe easiest ones for him to make his ar-gument with since Italy has historicallybeen disunited and lacks strong unifyingnational-level institutions. The next stepwould seem to be a similar study of acountry with a less heterogeneous back-ground and stronger national-level insti-tutions, such as France or Sweden, or asubnational comparative analysis acrossdifferent countries. Nonetheless, Remak-ing the Italian Economy deserves to bewidely read as it achieves the benefits ofthe comparative approach without sacri-ficing the attention to detail that attractsarea specialists, and it provides a theo-retical perspective that could usefully beapplied elsewhere.

contending parties. While differences inlabor organization matter, it is the densesocial networks that connect differentactors that play the key role. While theoutcomes in neither case were prede-termined and sociopolitical networks arenot immutable (as Locke explicitly dem-onstrates in a chapter on the Biella tex-tile industry), local sociopolitical net-works do impose limits on actors’choices. These networks and limits typi-cally change only slowly and in waysthat cannot be guided or foreseen bygovernment programs.

Besides the prodigeous amount ofresearch and lucid synthesis of compli-cated material, Locke also expertlynavigates methodological issues andfully exploits the economic, political andsocial diversity within Italy that has dis-mayed many scholars who employ na-tional level explanations. By conduct-ing a subnational comparative approach,Locke holds a greater proportion of fac-tors constant, which increases the reli-ability of his findings. He demonstratesfurther methodological adroitness bysystematically employing, within thesame industry, matched pairs of com-panies that display radical differencesin the way that they handled industrialtransformation (Locke’s dependent vari-able). In addition, comparing two diverseindustries that differ radically from oneanother and share little in common en-ables Locke to exclude a number of al-ternative hypotheses, such as techno-logical requirements, since each indus-try employs substantially different meth-ods. Despite and because of the tre-mendous differences between thesetwo industries, Locke’s claim that “thestrategic choices of company manag-ers and local union leaders were shapedby the local economic order in whichthey were situated” (Locke, 175) is fullysustainable. The type of sociopoliticalnetworks that exist for a company af-fects unions and companies’ strategiesby influencing information flow, strate-gies, and goals, and favoring distinctpatterns of behavior.

Particularly in Europe, where the re-gions’ roles are expanding in many

Book Reviewers Welcome

Doctoral students at any in-stitution are welcome to sub-mit proposals for book re-views. To do so, contact theeditor and/or assistant editorwith the name of the recentlypublished book that you wishto review and a short (ap-proximately 200 word) dis-cussion of the book’s impor-tance and why you wish toreview it. Reviewers are re-sponsible for procuring theirown copies of books.


Making and Breaking Gov-ernments: Cabinets and Leg-islatures in ParliamentaryDemocraciesMichael Laver and Kenneth A.ShepsleCambridge University PressCambridge, 1996

Reviewed by M. Shawn ReichertRice [email protected]

In Making and Breaking Govern-ments, Laver and Shepsle offer a theoryof government formation and durationthat is distinct from previous coalitiontheories. Whereas much research hasviewed government formation as agame among legislators deciding on howto divide the spoils of office, Laver andShepsle adopt a portfolio allocation ap-proach in which the political parties arethe key actors and specific cabinet posi-tions and the policy authority they entailare the parties’ goals. The authors basethe portfolio allocation approach on anassumption of policy motivated parties.While the stance taken toward the issueof policy versus electoral motivations hasdistinguished earlier coalition theories, theauthors argue that policy may influencethe strategies of political parties as ei-ther an instrument in pursuit of electoralgain or as an end in itself. Thus, politi-cians will attempt to implement thosepolicies with which they are associatedif the opportunity presents itself.

Given that political parties adopt policypositions which differentiate the partiesin the electoral arena, and given the de-gree of party discipline that is usuallyfound in parliamentary regimes, it is pos-sible for parties to rationally forecast thepolicy implications of different cabinetportfolio allocations. The policy positionsadopted by the political parties on issuedimensions – dimensions which are pre-sumed to correspond with cabinet min-istry jurisdictions – provide a “lattice ofpolicy positions,” the intersections ofwhich provide the “feasible govern-ments” that can potentially form. The

authors assume that the dimensions areseparable and that the control over aparticular portfolio by a political partydoes not influence the control over otherportfolios by other parties in the govern-ment. Both assumptions could be con-sidered somewhat controversial to manybecause of the convention of ‘collectiveresponsibility’ that stipulates that the par-ties in government adopt a common pub-lic stance regardless of differences be-tween the parties over policies in cabi-net deliberations. However, collective re-sponsibility is not the same as collectivedecision-making. Laver and Shepsle sug-gest that the complexity of ministerialresponsibilities is sufficient to warrant adivision of labor in which each ministersits at the apex of his or her ministry andis able to draw upon the careerists’knowledge and expertise of issues thatfall within the domain of the ministry.This ministerial expertise leads to minis-terial discretion, or control, over that partof the policy agenda that falls within thedomain of a particular ministry held by aspecific political party. Such ministerialdiscretion substantially reduces the fea-sible policy positions that may beadopted by the cabinet.

Since political parties can forecast thepolicy positions of other parties on theissue dimensions, given the presumptionof ministerial discretion, Laver andShepsle are able to derive distinct pre-dictions of which governments, with par-ticular policy outputs, are likely to form.This is perhaps the greatest contributionof the Laver and Shepsle model.Whereas cooperative game spatial theo-ries ‘predict’ governments and policieswithin vague, continuous policy space,Laver and Shepsle can make ‘point’ pre-dictions that can more readily be testedby empirical data such as that producedby the Comparative Manifestoes Project.The predictions are based on Black’smedian voter theorem as expanded ton-space by Kadane. Political parties thatare potential partners in a governmentrealize the need to attain majority sup-port and look to the median party on eachdimension. The dimension-by-dimensionmedian is the most likely to produce a

“cabinet equilibrium” in which no partywith the ability to bring down the gov-ernment has an incentive to do so.

Unlike some earlier theories, non-empty policy win sets do not indicate gov-ernment instability in the Laver andShepsle model. Rather, it is when an in-tersection of party policy positions, a“feasible government,” is present withinthe cabinet win set that stability mightbreak down. However, instability mightbe curbed by the presence of a “strongparty,” a party that is included in everyfeasible government preferred by a ma-jority. Since it is assumed that eachpolitical party possesses veto power overany cabinet in which it is a potentialmember, strong parties are those withveto power over all feasible governmentspreferred by a majority. Such a strongparty could then veto all governmentsother than its ideal point (in which it re-ceives all portfolios) and a standoff be-tween parties would ensue. A “verystrong party” on the other hand is onewhose ideal point is the dimension-by-dimension median with no alternativegovernment majority-preferred to it.Strong parties are then key actors in thegovernment formation process.

In computer simulations based on theirWINSET program, the authors find thatstrong parties are more likely in partysystems with a smaller number of par-ties, fewer policy dimensions, and in situ-ations in which one party is substantiallylarger than the other parties. One po-tential criticism of the simulation is thatthe number of policy dimensions variedonly between two and four while in thereal world the number of cabinet posi-tions is much larger. The finding that in-creasing the number of policy dimensionssubstantially reduces the probability ofan equilibrium cabinet to near zero wouldseem to suggest that the Laver andShepsle model could not pass the test ofprediction. However, Laver and Shepslerelax the assumption of separable dimen-sions and show that higher correlationsbetween a particular party’s positions ondifferent jurisdictional dimensions in-creases the probability of equilibriumcabinets forming. Thus, by focusing on


a few key portfolios equilibrium predic-tions are possible. The authors also sub-ject their model to more orthodox em-pirical tests, qualitative and quantitative,and find tentative support for their theory.

Curiously, Laver and Shepsle all butneglect the role of prime ministers. Whilethe authors note that prime ministersnominate cabinet ministers, dissolve leg-islatures, and call new elections, it is onlyin passing that they mention the deci-sional power of prime ministers. In theirearlier edited volume, Cabinet Minis-ters and Parliamentary Government(1994), Laver and Shepsle concludedthat prime ministers are central actorsin cabinet deliberations and that primeministers possess effective veto powerover many cabinet decisions. It is notexactly clear how the latter view of primeministers, with veto power, would fit intheir portfolio allocation model. Primeministers, after all, do not hold controlover a specific ministry but rather exer-cise jurisdictional discretion over thecabinet as a whole.

Making and Breaking Governmentsoffers several novel changes to the ex-isting coalition literature and is sure tobe an important book for future research.The book is written in clean, easy-to-fol-low prose without the heavy reliance onGreek notation that is often found in for-mal work. Such progression in coalitiontheory in tandem with the recently col-lected data of the Comparative Mani-festoes Project will provide significantgains in our understanding of governmentformation and termination in parliamen-tary democracies.

The Political Economy ofDemocratic TransitionsStephan Haggard and Robert R.KaufmanPrinceton University PressPrinceton, 1995

Reviewed by Carolina WielandUniversity of California, Los [email protected]

Challenging ‘choice-based’ approachesto the analysis of political transitions,

Stephan Haggard and Robert Kaufmanemploy an extensive case-study exami-nation of political and economic reformsin twelve middle-income developingcountries. In doing so, they stress the im-portance of both economic conditions andpolitical institutions in determining thenature of regime transitions, as well asin shaping the economic performanceand future prospects of new democraticregimes.

Focusing on transitions from authori-tarian rule in the first part of their book,Haggard and Kaufman point out that ei-ther economic crisis or economic suc-cess can contribute to the demise ofauthoritarianism, albeit with differentconsequences for the subsequent politi-cal transition. Economic crisis can causethe bargain between state elites and sup-portive interest groups to become “un-glued,” while economic success willeventually lead to the rise of new middleincome groups which will seek to checkthe power of the state. But authoritarianwithdrawals during crises differ signifi-cantly from “non-crisis” cases, as stateelites departing in crisis situations are lessable to delay and structure democratictransitions to their benefit. Therefore,democratic transitions in “crisis” casesare more likely to lead to open politicalcompetition over a short period and en-courage greater political polarization.Moreover, such crises tend to empowerstrong leftist or populist parties, whichtend to threaten not only subsequentdemocratic consolidation, but desperatelyneeded economic reforms.

Whether new developing-country de-mocracies can make the sacrifices re-quired for successful economic reformconstitutes the topic of the book’s sec-ond section. In addressing this question,Haggard and Kaufman again make aclear distinction between crisis and non-crisis cases. They argue that since non-crisis cases must only sustain previouslyinstituted policies, and they often enjoythe support of private-sector groups withvested interests in neo-liberal reforms,they are better able to foster the devel-opment of a market-based economy. Butin both crisis and non-crisis cases, Hag-

gard and Kaufman find that strong ex-ecutive authority and a centralized partysystem aid the initiation and maintenanceof neo-liberal reforms. Yet, the impactof such institutional factors is alsoshaped by economic circumstances. In-deed, in many crises, reforms were de-layed following initial transitions towarddemocracy and were instituted only af-ter further economic decline. As theauthors say, it was the dire economicconditions of many developing countriesduring the late 1980s which allowednewly elected presidents to finally takethe initiative to impose harsh economicreforms, while simultaneously erodingthe social base of interests opposed tothese reforms.

In their third section, Haggard andKaufman conclude that economic cir-cumstances are also crucial for the suc-cessful consolidation of democracy.Though they admit that many new de-veloping-country democracies are atleast “surviving,” they posit that economiccrisis will tend to undermine democrati-zation, albeit by an “indirect” route. Moreprecisely, they point out that consider-able tensions between concurrent eco-nomic and political reforms do exist.Strong executive authority, though con-ducive to economic adjustment, tends toimpede institutionalized consultation withelected legislators and social groups overthe long-run. Moreover, a centralizedparty system may not be sufficientlyaccountable to previously neglected so-cietal groups, and may thereby increasethe appeal of extremist groups, whichmay destabilize tentative steps towarddemocracy.

Seeking solutions to these tensions be-tween economic and political reform,Haggard and Kaufman then proposefour party-system options for effectivelymanaging developing economies underdemocratic rule. All four systems – two-party systems, consociationalism, andmultiparty systems dominated by a cen-ter-left or a center-right party – presentpotential problems of their own, depend-ing on the historical legacy and socio-economic circumstances of each coun-try. In conclusion, however, the authors

tentatively contend that multiple viablepaths to the successful maintenance ofdemocracy and a stable market economyin developing countries appear possible.

While this book constitutes an ambi-tious and praiseworthy effort at exam-ining political transitions, its analysis lackssufficient rigor. The case studies them-selves, though detailed, are not particu-larly well focused. Much of the infor-mation tends to be anecdotal and doesnot develop any falsifiable hypotheses.What constitutes a fragmented partysystem or a centralized executive in onecase, for instance, is not systematicallycompared across case studies. In addi-tion, Haggard and Kaufman often fail todefine their central terms. At what point,for example, does a democracy become“consolidated?” It remains unclearwhether any of Haggard and Kaufman’scases have successfully completed theirtransition to democracy, or seem poisedto do so.

Finally, Haggard and Kaufman’s analy-sis falls short because it remains depen-dent on the very choice-based frame-work whose inadequacy it seeks to dem-onstrate. Though few would disputeHaggard and Kaufman’s central conten-tion that economic conditions and politi-cal institutions shape the nature of bothpolitical transitions and economicpolicymaking, this does not run counterto choice-based analysis. Indeed, sev-eral prominent choice-based analysts ofdeveloping-country politics have increas-ingly incorporated the role of economicand institutional factors into their work.The authors themselves, in fact, describeall political regimes as a “bargain be-tween political leaders and key supportgroups” (p. 7). Though the terms of thisbargain are undoubtedly subject to botheconomic and institutional constraints, itstill represents an explicit choice of self-interested political actors.

How to Subscribe

Subscriptions to the APSA-CP Newslet-ter are a benefit to members of the Or-ganized Section in Comparative Politicsof the American Political Science As-sociation. To join the Section, check theappropriate box when joining the APSAor renewing your Association member-ship. Sections dues currently run $7 an-nually, with a $2 surcharge for foreignaddresses. The printing and mailing ofthis Newsletter are paid for out of mem-bers’ dues. To join the APSA, contact:

American PoliticalScience Association

1527 New Hampshire Ave., NWWashington, DC 20036

Telephone: (202) 483-2512Facsimile: (202) 483-2657

Email: [email protected]

APSA-CP NewsletterMiriam Golden, EditorUniversity of California, Los AngelesDepartment of Political Science405 Hilgard AvenueLos Angeles, California [email protected]

APSA-CP Newsletter 32 Winter 1997

Please file changes of address with the APSA. Your Newsletter address will change automatically.

Use the Newsletter in the classroom. The APSA has authorized university teachers to reproduce articlesfrom the Newsletter for use in the classroom at no charge. Take advantage of this policy, and introduce your

graduate students to the latest research, issues and debates in comparative politics.

APSA-CP Winter 1997

Documents

Transcript of APSA-CP Winter 1997