Social Foundations of Political Order in Interwar Europe Author(S): Gregory M

Trustees of Princeton University

Social Foundations of Political Order in Interwar Europe Author(s): Gregory M. Luebbert Source: World Politics, Vol. 39, No. 4 (Jul., 1987), pp. 449-478 Published by: Cambridge University Press Stable URL: http://www.jstor.org/stable/2010288 . Accessed: 21/07/2011 14:04

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

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

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

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

Cambridge University Press and Trustees of Princeton University are collaborating with JSTOR to digitize, preserve and extend access to World Politics.

http://www.jstor.org SOCIAL FOUNDATIONS OF POLITICAL ORDER IN INTER WAR EUROPE

By GREGORY M. LUEBBERT*

TOUR typesof regimesof historicimportance appeared in Europe between the two world wars: pluralistdemocracy, social or corporatist democracy,traditional dictatorship, and fascism.The vast body of liter- ature thathas grown up around themhas rarelycast thesepolitical orders as historicalalternatives to each other,however. When it has done so, it has normallycast pluralistdemocracy as the alternativeto fascism.Most commonly,this has taken the formof contrastsbetween Germany and Britain,and has been accompanied by the question, why was Germany not like Britain?' Yet, pluralistdemocracy such as appeared in Britain was actually the least relevantalternative between the wars, forthe pos- sibilityof stabilizingit where it did not alreadyexist had been foreclosed by World War I. Where liberal partieshad failedto establishresponsible parliamentaryinstitutions before the war, it would prove impossible to stabilize a pluralist democracy afterward. Henceforth, stabilization would require corporatism in eitherits fascistor social democraticvariant ratherthan pluralism. The appearance of fascism,however, was not predeterminedby the ir- relevance of pluralist democracy. Germany did not become fascistbecause it could not become a pluralistdemocracy such as Britain.It became fascistbecause it could become neithera corporatistdemocracy such as Norway nor a traditionaldictatorship such as appeared in much of east- ern Europe. It was stillpossible to establisha stable social or corporatist

* I would like to thankVinod Aggarwal,Gabriel Almond,Paul Buchanan,David Collier, Giuseppe Di Palma, JosephFizman, Ernst Haas, Andrew Janos,Daniel Verdier,and John Zysman fortheir thoughtful comments on the longer manuscriptfrom which this articleis taken. I am especiallyindebted to Daniel Verdierfor his researchassistance. Among the most influentialof theseworks are thoseby Ralf Dahrendorfand Karl Die- trichBracher, both of whom postulatea liberalor Anglo-Saxontheory of the stateand con- trastit with a German Idealist tradition.See Dahrendorf,Society and Democracyin Germany (New York: Norton, i967); Bracher,Die Aufidsungder WeimarerRepublik [The Dissolution of the Weimar Republic] (Stuttgart:Ring Verlag, I Two influentialbut more sociologi- callygrounded analyses that make use of thecontrast between Britain and Germanyare Bar- ringtonMoore, Social Originsof Dictatorship and Democracy(Boston: Beacon Press,i966), and AlexanderGerschenkron, Bread and Democracyin Germany(Berkeley: University of Califor- nia Press, 1943). 450 WORLD POLITICS democracy. This could be done if socialistscould forma coalition with the countryside.Even when socialistscould not do this,it was stillpossible to avoid fascism if interwar liberals could make such a coalition. Where they could do so, the result would be a traditionaldictatorship. Fascism, in otherwords, could only appear if pluralistdemocracy, social democracy,and traditionaldictatorship could not. An understandingof fascism,then, requires an understandingof whythese alternative regimes were not possible in some countries. To understandwhy theywere not,we must studythe determinantsof both stateand labor-marketinstitutions. The formthe interwarstate ul- timatelyassumed in the crises of the I920S and I930S cannot be under- stood apart fromthe manner in which workersbecame organized in the labor market.For both stateand labor marketinstitutions took shape si- multaneouslyas two halves of the same responseto working-classmobi- lization. Trade unions could be organized along corporatistor pluralistlines, or they could simply be repressedby the state. Corporatism between the wars denoted a systemin which economic interestsin a privateeconomy were becoming functionallyorganized in groups that were centralized, comprehensive,and authoritative.In its democraticvariant social democracy membershipin theseorganizations was voluntary;the organizations mediated the relationshipamong workers,capitalists, sectors of the economy, and the state. In social democracies, corporatistlabor unions have served to offsetthe otherwiseinherent advantages of employersin collectivebargaining and in the polity,and have played a critical role in the making of social policy.Corporatism has not been mainly a systemof restraintand controlof the working class, but of inclusion. Social democracies appeared between the wars in Norway, Sweden, Czechoslovakia, and to a lesserdegree in Denmark.2 In its authoritarianvariant fascism corporatistorganizations were mainly mechanisms throughwhich the state controlledand mobilized workers and capitalists. Here corporatism provided the institutional means to subordinatethe working class; inclusionwas mainly symbolic. Corporatisttrade unions served as the transmissionbelts of the material benefitsand coercionby which the controllingparty attempted to buy the support or force the acquiescence of labor. Between the wars, such re-

2 The interwarCzechoslovakian experienceand its corporatistaspects can be found in Harry Klepetar,Seit i9i8 ... Eine Geschichteder TschechoslowakischenRepublik [Since i9i8 ... A Historyof the Czechoslovak Republik](Moravska Ostrava: Verlag JuliusKittls Nach- folger,1937), esp. 283-375- POLITICAL ORDER IN INTERWAR EUROPE 451 gimesappeared in Nazi Germany,Fascist Italy and, to a lesserdegree, in Franco'sSpain. In pluralistlabor markets,trade unionswere legallyfree-within some basic rules-to organize,bargain collectively, and agitatepoliti- cally.In contrastto theircounterparts in socialdemocracies, unions re- mainedcomparatively weakly organized and lesseffective in bargaining withemployers and thestate. They did notexhibit the same comprehen- sivenessor centralization;leaders could not speak with authority for their memberson a broadrange of socialand economicquestions. There de- velopedno legal or culturalpresumption that union leaders had to be consultedand theirpolicy demands accommodated by the state.The UnitedStates, Britain, France, Ireland, and Switzerlandare examplesof pluralistdemocracies. The fourthvariety of politicalorder that appeared between the wars was thetraditional dictatorship. It combinedsevere curtailments of political competitionwith repressionin the labor market.Parliament was closedor existedonly as a rumpafter important parties had been outlawed. Trade unions were eithergreatly restricted or outlawedalto- gether,but theywere not,as in fascistregimes, replaced by state-spon- soredorganizations of control.Such regimesoffered no solutionto the problemof working-class integration; rather, they attempted to ignoreit. To thatextent, they were inherently unstable. Such traditionaldictator- shipsappeared in Finland,Austria, Hungary, and muchof easternEu- rope. The fourregimes are, to be sure,ideal types.No societyacquired in- stitutionsthat corresponded perfectly to theidealized models, and some institutionswere a mixture,especially of thepluralist and socialdemo- craticmodels. Belgium, the Netherlands, and Denmark,for instance, fell betweenthe pluralistand corporatistmodels, with the firsttwo more closelyapproximating the pluralistmodel, and Denmarkmore closely approximatingthe social democratic model. The intermediateposition of thesecountries reflects and can be explainedby themixed character of theirdevelopmental experience, which combined aspects of both the pluralistand corporatistexperiences. The fourregimes were the interwar byproducts of fourroutes from preindustrialpolitics to thecrises of the I920S and I930s. Each routewas delineatedby a historicallegacy and by the urban-ruralcoalition that formedin responseto theinterwar crises. The historicallegacy was providedby the dominance, or lackthereof, of liberal parties before the First WorldWar. Nowherebefore the war wereliberals in unambiguouspo- 452 WORLD POLITICS liticalcontrol. Rather, they exercised more or less influence,sometimes in an alliance with, and sometimes in competitionwith, rural interests. Whether theywere the senioror junior partner(or antagonist),however, was of decisive importance.Where they were dominant-as in Britain (where the Liberal Party and the liberalized Conservative Party alter- nated in power until the war), Switzerland (where the Radicals con- trolledevery government from i848 to i919), and France (where Repub- licans were dominantafter i877) workersentered politics before World War I; theydid so graduallyand in groups,rather than suddenlyand as a cohesive class. As a result,these societies,all of which acquired democratic polities,also acquired pluralistlabor markets. The situationwas dramaticallydifferent in countriesin which liberals were not dominant. Here, the war was a turningpoint in the transition to mass politics.Its trauma and the example of the Bolshevik Revolution served to radicalize workersand polarize politicsin all societies.Workers and peasants were now politicallyactive everywhere.The explosion of political participationwas vastlymore difficultto accommodate where workers had not been broughtinto politicsgradually under liberal tute- lage. Where theyhad not been, theyhad come to be organized in politics and in the labor marketin a more explicitlyclass-based manner. Yet, the politicaland economic citizenshipof thismore cohesivelyorganized class remained incomplete.In effect,modern politicalorders had yetto be cre- ated. Ultimately,corporatism in eitherits social democraticor fascistvar- iant,rather than pluralism,would be required to stabilizeclass politicsin these societies. The formof the politicalorder was determinedby the coalition of urban and rural intereststhat emerged in responseto the crises.In the making of these coalitions,the allegiance of the familypeasantry proved decisive: when it sided with an urban socialist party,the outcome was a social democraticregime; when it sided with the partyor partiesof traditional liberals,the outcome was invariablya traditionaldictatorship. In this essay,I will attemptto identifythe conditionsthat produced each of these coalitions. To set the stage, we must firstconsider the importance of liberal dominance in the creationof pluralistdemocracies, and why it was thatthe politicaleconomies establishedafter World War I could not become pluralistdemocracies.

PLURALIST DEMOCRACIES

The model of the pluralistpath, as exemplifiedby the Britishtransi- tion, has long fixed attentionon the essentialfeature that distinguished POLITICAL ORDER IN INTERWAR EUROPE 453 earlydemocratizers.3 In thesesocieties, the bourgeoisiegained a domi- nantposition, through the vehicle of liberalparties, before workers mo- bilizedautonomously. In societiesthat were becoming pluralist, working classunions and partiesgained access to thestate through collaboration with the liberalcommunity. Liberals gradually accommodated them- selvesto demandsfor an enlargedsuffrage, the secret ballot, the elimi- nationof unequalelectoral districts, the legalization of trade unions, collectivebargaining, the right to strike, and so forth.In largemeasure, these long-termreforms came aboutbecause they served the short-term elec- toralneeds of liberalparties. For example,because French Republicans needed the supportof workersin theirstruggle against Conservatives, theymet the threemost important demands of Frenchworker move- mentsafter i877: theestablishment of universal, free secular public edu- cation;the pardoning of theCommunards; and thelegalization of trade unions,collective bargaining, and theright to strike.4 Becauseworkers in thesesocieties gained the right to vote early on and because parliamentaryinstitutions had substantiallegitimacy among workers,liberal parties never found themselves confronted by the workingclass as a whole.Instead, they were confronted by scattered groups of graduallymobilized workers, usually organized on a local basis.In controlof the state,the liberals could definethe terms on whichworking- classparties were given access to politics;they controlled how voteswere counted,constituencies defined, parties registered, campaigns financed, and legislatorspaid. The liberalparties could extend participation with- out actuallychanging the balance of class power very much. Indeed, the Lib-Lab alliancesthat were so characteristicof thispattern of mobiliza- tionactually served to reinforcethe balance of class power. To theextent thatthe Lib-Lab strategyactually benefited workers (or merely appeared to benefitthem), the liberal ideology was made moreplausible, and the launchingof an autonomousworking-class movement was made more difficult. Liberalism withits emphasis on classharmony, individualism, com- petition,achievement, and merit-was an ideologythat legitimated the pluralistorder. Its extendedpre-eminence meant that it, rather than the

3 Early democratizersare definedas societiesthat acquired responsibleparliamentary institutions,manhood suffrage,reasonably equitable electorallaws, and trade union rightsbe- foreWorld War I. The countriesthat met all fourof the criteriawere Britain,Ireland (as a colony),France, Switzerland,and the United States.All othersfailed on at least one count. The classical statementof the pluralistpath is foundin T. H. Marshall,Class, Citizenship and Social Development(Garden City,NY: Doubleday, i965), chap. 4. 4 Val Lorwin, The FrenchLabor Movement(Cambridge: Harvard UniversityPress, 1954), i8. 454 WORLD POLITICS landedinterests or working-classmovements, defined the parameters of legitimatepolitical institutions and debate.Working-class parties thus foundthemselves in thedifficult position of arguingagainst an ideology and parliamentaryinstitutions that appeared legitimate to a largepart of the workingclass. In I920, FernandLoriot, a leaderof the new Com- munistParty, lamented the hold of the liberal ideology on Frenchwork- ers'attitudes: Whateverwe do, we cannotavoid the fact that in France,over 50 yearsof bourgeoisdemocracy has createdwithin the massesamongst whom we work ... an entirelydistinctive mentality. . .. People here are stillcon- vincedthat we havea patrimonyof libertiesto defend.The wholeof our mentalformation is basedon a revolutionary[i.e., Jacobin] tradition, skill- fullymanipulated by our bourgeois democracy.5 Ironically,one indicatorof the successof liberalsin integratingthe workingclass is thatin somerespects they actually had to do lessfor the materialinterests of workersthan did theleaders of authoritarian states. PeterFlora and JensAlber have shown that social welfare measures were most highly developed before I9I4 in those European societies that lacked parliamentaryinstitutions and political rightsfor workers.Flora and Alber argue thatauthoritarian regimes turned to thesemeasures as a substitutefor parliamentarismand the granting of political rights to workers.The greaterlegitimacy bestowed by a parliamentaryorder and liberal ideology actuallymilitated against the earlyintroduction of social welfareinitiatives that were of value to workers.6 Much the same processof accommodationunder liberaldominance occurred in the labor market. In France, the Republicans legislated the rightto formtrade unions, to bargain collectively,and to strikein i884. This law and the manner of its enforcementillustrate the impact of liberal dominance: unions were obliged to registertheir constitutions and bylaws and the names of theirofficers. Registered bourses du travailwere

5Quoted in Roger Magraw,France, i8I5-I9I4: The BourgeoisCentury (New York: Oxford University Press), 373. 6 Flora and Alber use a less restrictivedefinition of democraticregime (see fn. 3 for thc definitionemployed here). Their "constitutional-dualistmonarchies" are Sweden, Germany, Denmark, and Austria. Their parliamentaryregimes are Britain,Switzerland, France, Bel- gium, The Netherlands,Italy, and Norway. Under the more restricteddefinition employed here,The Netherlandsand Belgium almostfit the requirementsof earlydemocracies, falling shortonly on the extentof the suffrage.Norway (electorallaws) and Italy(franchise, electoral laws, trade union rights)clearly do not. Despite the definitionaldifferences, the association betweenliberalism and late social welfareinitiatives remains. Indeed, Flora and Alber's data show that it is strengthenedif Italy is properlyclassified as undemocratic.See "Moderniza- tion,Democratization, and the Developmentof WelfareStates in WesternEurope," in Peter Flora and Arnold Heidenheimer,eds., The Developmentof WelfareStates in Europeand Amer- ica (New Brunswick,NJ: TransactionBooks, i98i), 37-80. POLITICAL ORDER IN INTERWAR EUROPE 455 theneligible to receivestate subsidies and rent-freeoffice space. Registra- tionallowed municipalitiesto keep trackof radicals.The government used subsidiesto discourageradicalism and, fora time,the creationof labororganizations that stretched across different trades. The law of i884 meant that the state recognized the existence of unions.It did notmean that employers were required to recognizethem. Indeed, fortwo generationsmany French employers continued, often brutally,to resistrecognition. The reluctanceof many employersin Franceand theUnited States to recognizeunions until compelled to do so in the I930S was itselfa manifestationof the weightof the liberal ideology. The telling effectsof the i884 law, however, were to encourage decentralized unions, to enhance the legitimacyof the Republic, and to redefinesubstantially the battle as one between workersand employers ratherthan between workersand the state.7The infrequencyof political strikeswas indicativeof the success of French Republicans and theirliberal counterpartselsewhere. While political strikesby French, British, and American workers were exceptions,they were common in other countries.In Belgium, forinstance, there were fourgeneral strikesin the yearsbefore I9I4. Elsewhere, too, politicalstrikes were a regular feature of labor movements,even where as in Spain, Germany,and Italy the politicalrestraints on labor protestwere much more repressive.8 Because in liberal-dominatedsocieties workers were allowed to organize more or less as theysought to, trade unions became incorporatedinto the labor market as individual interestgroups. They were oftenorgan- ized locally and along craftlines, and developed traditionsof local or plant bargaining. By the time national working-classparties and trade- union movementsstarted to form,these locally organized unions had a generationor more of leaders and memberswhose vested interestsin the established organizations had to be overcome in order to create more comprehensivestructures.9 Among the pluralist democracies,France appears somewhat exceptional because of the eventual emergence of a radical working-class movementin the formof the CommunistParty. In the termsused above, the explanationseems straightforwardenough, however: liberalshad less

7The earlyhistory of the Frenchlabor movementis discussedin Lorwin (fn.4), chap. 2. 8 The absence of politicalstrikes by French workers-and the feeblenessof syndicalismin theFrench labor movementgenerally-has been carefullydocumented in PeterStearns, Rev- olutionarySyndicalism and FrenchLabor: A Cause WithoutRebels (New Brunswick,NJ: Rut- gers UniversityPress, 1971), esp. chap. 2. 9 For a discussionof the obstaclespresented by earlydecentralization for subsequent uni- ficationof French labor, see Lorwin (fn.4), chap. 2. For a paralleldiscussion of the Swiss experience,see Eduard Weckerle,The Trade Unionsin Switzerland(Bern: Swiss Federationof Trade Unions, 1947), i6-33. 456 WORLD POLITICS ofa needto solicitthe workers' support because of the combined electoral weightof the middle class and thepeasantry and becausethe centrality of clerical-anticlericalconflict left French workers with no alternativeto the RadicalRepublicans until the appearance of the unified Socialist Party in I905. The inconsistencyof thelate formationof theSocialist Party and theeventual appearance of the Communist Party points to anotherdeci- sivefeature of theFrench experience: France was distinguishednot only by theapparently lower level of workerintegration before the war, but byits uniquely traumatic experience in theFirst World War. As a result, the working-classmovement was, by thestandards of pluralistdemocracies,comparatively radicalized.

SOCIAL DEMOCRACY AND FASCISM Societiesthat became social democracies, traditional dictatorships, or fascistdictatorships after World War I differedfrom pluralist democracies in thecomparative weakness of theirliberal parties before the war. There wereseveral reasons for this weakness. Most obviously, these so- cietiesbegan to industrializeat a laterdate in all casesafter i85o, and in manycases not until the beginning of the 20th century. The amountof timein whichthe subtending class, the bourgeoisie, could assert itself beforethe war, before the ensuing disruptions, and beforethe example of theBolshevik Revolution, was briefer.IoAlso, because late industrializa- tionsometimes brought about a greaterrole for the state, it often resulted in a correspondinglygreater dependence of thebourgeoisie on thestate and a lessplausible case forthe liberal ideology. Late democratizersentered the FirstWorld War withmore highly mobilizedworking classes than early democratizers. The war had up- rooted,alienated, disoriented and, therefore,made moresusceptible to radicalappeals upon their return home (from factories or thefront), mil- lionsof young men who had previouslybeen beyond the reach of radical organizers.The impactof the war was hardlymilder in neutralistcoun- tries,traumatized as theywere by the economic boom-and-bust cycle of war and peace. In short,a significantdifference between societies like theseand thosethat became pluralist democracies was notsimply that liberalismwas weaker,but thatworkers became class-consciousover a muchbriefer period of time.One indicationof this is providedby Table i, whichreports the percentage of the vote for the left in thelast election

Io An importantstudy of the impact of the war on politicalmobilization is provided in Charles S. Maier, RecastingBourgeois Europe (Princeton:Princeton University Press, 1975), esp. 3-15- POLITICAL ORDER IN INTERWAR EUROPE 457

TABLE I LEFT VOTING AND LEVEL OF INDUSTRIALIZATION

Prewar Postwar Labor in Left Votea Left Votea Non -primarySectorsb France 16.8 23.0 56.9 Switzerland 20.1 23.5 80.0 Britain 6.4 29.7 91.5

Netherlands 22.9 25.0 77.0 Denmark 29.6 28.7 63.3 Belgium 30.3 36.6 68.0

Norway 26.3 31.6 57.9 Sweden 30.1 38.9c 56.3 Finland 43.1 38.0 37.0 Germany 34.8 45.5 69.3 Austriad 23.0 40.8 68.1 Italy 22.8 34.3 43.8 a Last electionbefore the war and firstelection after the war. For both elections,the left vote is definedas the social democratic,labor, or socialistparty, and all partiesto the leftof it. Data forboth elections are fromPeter Flora, ed.,State, Economy and Societyin WesternEurope, 1815-1975.A Data Handbook. Vol. 1: The Growthof Mass Democracies(Chicago: St. James Press, 1983),89-153. b The percentageof the active labor forceemployed outside the primarysector in about 1920.Source: B. R. Mitchell,European Historical Statistics, 2d. rev.ed. (New York:Facts on File, 1980), 151-66. c This electiontook place in 1917. d Prewar figureis an average forseveral elections. Source: Walter Korpi, The Democratic ClassStruggle (London: Routledge& Kegan Paul, 1983),38. beforethe war and in the firstelection after the war, and the level of industrialization in twelve European countries.The table demonstrates thatthe societiesthat had not become democratichad had conspicuously higher levels of support for working-classparties even before the war. Moreover,in almostall cases the level of industrializationwas lower, and thereforethe size of the working class was smaller in these societies.In effect,working-class parties were having distinctlymore success under distinctlyless favorableconditions. The war raised the left vote everywhere,but did not close the gap. In none of the earlydemocratizers is the leftvote over 30 percentin the firstpostwar election. It was greaterthan 30 percentin all but two of the nine late democratizers. In some societies,foremost among them Italy,Norway, and Sweden, another generationof peace mighthave resultedin pluralistdemocratic regimes.In Italy,the Giolittiancoalition sought to bringtogether reform- 458 WORLD POLITICS ist parts of the bourgeoisieand what amounted to the Italian equivalent of the "aristocracyof labor" in a reformmovement that bore a remarka- ble resemblanceto the earlyLib-Lab coalitionin Britain.In Norway and Sweden, reformcoalitions between parts of the bourgeoisieand leading sectors of the still fairlysmall industrialworking-class movement had come into being during the years immediatelybefore the war. Even in Germany probablyamong the countriesleast likelyto become a stable pluralist democracy because of the high level of class-consciousness halting attemptsto create a reformistcoalition of liberals and labor had appeared in I9I2. Late industrializationand the war were themost proximate destabiliz- ing factors.It was of more fundamentalimportance that in societiesthat were to become social democracies,traditional dictatorships, or fascist, the liberal communitywas deeply divided. Norway and Czechoslovakia, with sociological fault lines that divided bourgeois political movements into linguistic,regional, cultural and religiousblocs, illustratethis well. In Norway, liberals were never able on theirown to sponsor successful reformmovements before World War I. As a result,rural interestsdom- inated the Storting,the dominantvalues of the partysystem were agrarian ratherthan industrial,and economic liberalizationwas long delayed. The shiftto parliamentarysovereignty itself was mainly broughtabout by rural representatives,and was aimed at controllingthe big spending of urban bureaucrats." In Czechoslovakia, liberal movementswere sim- ilarlyconstrained by theirdivisions and by subordinationto rural interests. Czech liberalismwas also constrainedby the peripheralposition of Bohemia and Moravia in the Habsburg Empire.12 These divisions within the liberal communitygenerally reflected the incompletenessof nation-stateformation before industrialization.The basic questions of territorialidentity, national language, and religionre- mained unresolved.These "preindustrialcleavages," to borrow Seymour Martin Lipset and Stein Rokkan's phrase,continued to divide the liberal community;as a result,the latterwas never able to obtain the pre-emi- nent positionthat its counterpartsacquired in countrieslike Switzerland and Britain.I3 Switzerland's experienceindicates that the decisive point is

II On the weakness of liberalismin Scandinavia,see FrancisCastles, The Social Democratic Image ofSociety (London: Routledge& Kegan Paul, 1978). On Czechoslovakia, see VictorMamatey, "The Establishmentof theRepublic," in Victor Mamateyand RadomfrLuza, eds.,A Historyof the Czechoslovak Republic, i918-I948 (Prince- ton: PrincetonUniversity Press, 1973), 3-39. 3 Sweden and Denmark are partialexceptions. The Swedish liberal communitywas divided into separate partiesby the high church-lowchurch conflict and by the temperance issue. This is in contrastto France, forinstance, where religiousconflict solidified rather than divided the liberalcommunity. The Danish liberalcommunity was divided by constitutional, POLITICAL ORDER IN INTERWAR EUROPE 459 notthat such cleavages exist in a partysystem, but that they prevent lib- eralsfrom acquiring a positionof dominance.Swiss Radicals and Cath- olic Conservativeswere constantly in conflictover clerical questions, but becausethe formeroutnumbered the latter by aboutfour to one in the legislature,the political dominance of the Radicals was neverchallenged. Takinga longerview of causality, then, we cansay that societies in which preindustrialcleavages were sufficientto underminethe dominanceof liberalsat theonset of workermobilization could onlybe stabilizedby corporatistinstitutions in theI920S and I930s, whetherdemocratic or fascist,because of the debilitatingeffect of these cleavages on the political claims of the liberal community. Liberal interestsin thesesocieties were attackedsimultaneously by traditional urban and rural oligarchs,peasants, and workers.Paradoxically, the same divisions that inhibitedliberal parties from establishingtheir pre-eminencealso preventedthem fromexpanding theirsupport outside the bourgeoisie.The vulnerabilityof the bourgeoisieapparently made it more difficultto persuade bourgeoisvoters and activiststo accept conces- sions to workers' interestsin exchange fortheir support. Thus, in countries like Italy, Sweden, and Germany, liberal parties appear never to have succeeded in gaining the percentageof the working-classvote that theircounterparts received in Britain,Switzerland, and France.'4 This inabilityto createsuccessful reformist alliances of the bourgeoisie and workersin the earlyperiod of workermobilization had momentous implications.Above all, it meant thatworkers were more likelyto be mobilized as a class than as a seriesof locally organized interestgroups and trade unions. In the face of limited opportunitiesto gain influence through alliances with bourgeois interests,arguments for class-based worker organizationswere made more compelling.Liberalism was em- piricallyrefuted by the political and labor-marketexperiences of workers. The outcome was a high level of class-consciousness,radicalization, and polarization. Under the circumstances,the conflictbetween the two urban classes became so intense that their representativescould not jointlyfound a stable politicaleconomy. It was not forlack of effort:the military,and foreignpolicy conflicts as well as bythe usual social reformquestions. The classic analysis of cleavage developmentis Seymour Martin Lipset and Stein Rokkan, "Cleavage Structures,Party Systems and Voter Alignments:An Introduction,"in Lipset and Rokkan, eds.,Party Systems and VoterAlignments (New York: Free Press,i967), I-64. 4 There is no directevidence that liberal parties in thesesocieties received fewer working- class votes,but-given theless frequentand less successfulappearance of Lib-Lab alliances- it seems impossibleto inferotherwise. Such alliancesare discussedin SeymourMartin Lipset, "Radicalism or Reformism:The Sources of Working-ClassPolitics," American Political Sci- enceReview 77 (JanuaryI983), 1-19. 460 WORLD POLITICS usual democraticcoalition of the I9205 was a coalitionof thereforming bourgeoisieand labor.Yet, in everycase thiscoalition either could not be formedat all at thegovernmental level (as in Italy,Sweden, Denmark, Finland,Norway, and Czechoslovakia)or soon failedunder the weight of its internalcontradictions (Weimar Germany and Spain). In all of thesecountries, the coalition was eventuallyformed between representativesof one ofthe urban classes and themiddle peasants.

THE COALITION WITH THE COUNTRYSIDE The coalitionhad to be with middle peasantsbecause, within the agrarianpopulation, only middle peasants combined a massbase, a fairly highand stablerate of politicalparticipation, a distinct political agenda, and theability to destabilizethe polity if this agenda was notrealized. By contrast,landed elites almost everywhere were a spentpolitical force in the I920S and I930s; theywere no longerin controlof themiddle peas- antryand therural proletariat. Although the rural proletariat had itsown agenda and, in some countries,the potentialto providea mass base, it was,by itself, the postwar equivalent of Marx's "sack of potatoes."Inca- pable of leadingitself or creatingits own organizations,it could be or- ganizedas a durableforce only by groups that were not actually of it: socialists,reform-minded liberals, and Christiandemocrats. Hence, it could be ignoredif the sources of external leadership were neutralized. Middlepeasants, by contrast, usually had autonomousinterest groups or a powerfulvoice in largeragrarian groups. In Norway,Sweden, Den- mark,and Finland,they had theirown agrarianparties. In CatholicGer- many,Austria, Czechoslovakia, Spain, and partsof northernItaly, they had a decisiveposition in less homogeneousparties. In ProtestantGer- manyand otherparts of northernItaly, they had theability to abandon swiftlytraditional representatives whom they judged to have served them poorly. A socialistmovement that sought to mobilizethe rural proletariat (landlesslaborers and smallholdersdependent on thelabor market) con- tainedthe seeds of its own failure,because it would become entrapped in ruralclass conflict and therebyalienate middle peasants. A coalitionwith middlepeasants required that socialists address three kinds of conflict: ruralclass conflict, consumer-producer conflict, and, sometimes, clerical- anticlericaland regionalconflict. My contentionis thatdefusing the second setof conflicts was essentialto thecreation of a worker-peasantcoa- lition,but thatthese conflicts could notbe defusedif thesocialist party founditself in confrontationwith middle peasants on ruralclass-struc- POLITICAL ORDER IN INTERWAR EUROPE 461 tureissues. If the consumer-producerissues could be defused,the less im- portantclerical-anticlerical and regionalissues could also be defused. Class conflictwas as centralto the politicallives of peasantsas it was to the political lives of urban workersand the bourgeoisie.Where socialist partiessought to mobilize the ruralproletariat, they became enmeshed in thisconflict and threatenedmiddle peasantsin severalways. The greatest threatarose fromcampaigns forland reform.Redistribution was a cen- tral issue of rural politicsin Spain, Italy,Finland, Denmark, and Czech- oslovakia. The fact that redistributionwould have occurred at the ex- pense of large landowners seems not to have reduced the threatfelt by the middle peasants, especially in areas such as northernItaly, where there was a substantialrural proletariatand littlesurplus land to redis- tribute.When the assertionthat their land was secure came fromparties born of a collectivistambition, it appeared especiallyincredible. For peasants,the securityof the asset around which theirentire existence revolved had to be beyond question. Hence, the Spanish Socialists' campaign for land reform in the south of Spain antagonized peasants even in the north.'5 Socialist mobilization of agrarian workersalso threatenedthe middle peasants' productiveprocess. Middle peasants,who, by definition,did not rely heavily on hired labor, nonethelesshad an employer'sperspective. For, to the extentthat theydid hire a worker or two, socialistefforts to unionize agrarian workers, to establish centralized hiring halls, to fix wages, and to extend social welfarebenefits to agrarian workersall im- pinged on the peasants' costsof productionand theirfreedom to manage theirland. The cumulativeeffect was to threatenmiddle peasants in a thirdway. At stake, ultimately,was theirsocial status.In this context,it is easy to understandwhy socialists'protestations that they were committedto the interestsof both agrarian workers and middle peasants sounded as im- plausible to middle peasants as the reformistliberals' claims that they were committedto the interestsof both workersand the bourgeoisiedid to urban workers. And just as the socialists'efforts on behalf of urban workers aroused intenseresentment among the petitebourgeoisie, even when it was not directlyand materiallythreatened, their efforts on behalf of agrarian workers were bound to generateintense resentment among middle peasants.Indeed, resentmentand a sense of neglectwere the common featuresof all antisocialistpeasant movements. When Socialist partiesdid succeed in foundingregimes, it was not be-

I5See Edward Malefakis,Agrarian Reform and PeasantRevolution in Spain: Originsof the Civil War (New Haven: Yale UniversityPress, 1970), 219 and passim. 462 WORLD POLITICS cause theyhad a superiorgrasp of the strategicrequirements of the mo- ment,but because theydid not attemptto organize the rural proletariat. The level of industrializationor the class structureof the countrysidehad littlebearing on this,as can be seen by examiningfirst the size of the primarylabor forcein comparisonto the totallabor force,and then the size of the proletarianlabor force in comparison to the total agrarian labor force.These data, forthe early I920S, appear in Table 2. Of the threeleast agrarian countries,one became fascist,one became a traditionaldictatorship (Austria,under the Dollfuss-Schuschniggregime), and one became a social democracy.Of the threemost agrarian countries, two became fascist,and the third(Finland) became semi-authoritarian.

TABLE 2 RURAL POPULATION AND REGIME OUTCOME

PrimaryLabor Force AgrarianProletariat (as a percentage (as a percentage oftotal labor oftotal primary Regime force) a laborforce) a Outcome Germany 30.7 26 Fascism Austria 31.9 26 Dictatorship Denmark 37.7 43 Socialdemocracy

Czechoslovakia 40.3 35 Socialdemocracy Norway 42.1 15 Socialdemocracy Sweden 43.7 26 Socialdemocracy

Spain 56.0 35-40 Fascism Italy 56.2 35-40 Fascism Finland 63.0 30 Dictatorship a Data in the firsttwo columnsare fromabout 1920.The agrarianproletariat is definedas landless workersor smallholdersdependent on the agrarian labor marketfor some part of theirincome. Sources: For the firstcolumn, Mitchell (Table 1, fn.b), 151-66.For the secondcolumn, the figuresare estimatesbased on data reportedin Frieda Wunderlich,Farm Labor in Germany, 1810-1945(Princeton: Princeton University Press, 1961), 15 (Germany);Statistisches Handbuch far denBundesstaat Osterreich, vol. 26 (Vienna: StatisticalOffice, 1936), 43 (Austria);Danmarks Statistik.Statistiske Meddelelser. Landbrugets Arbejdeskraft. 1904 og 1935,series 4, vol. 102,no. 3 (Copenhagen: StatisticalDepartment, 1937), 15 (Denmark); Manuel Statistique,1931, vol. 3 (Prague: StatisticalBureau of the Czechoslovak Republic, 1931),300 (Czechoslovakia); Kare Lunden, "The Growth of Co-operativesAmong the Norwegian Dairy FarmersDuring the Period 1856-1905,"Cahiers Internationaux d'Histoire Economique et Sociale, vol. 6 (Geneva: Li- brairieDroz, 1973),-342(Norway); StatistiskArsbokfdr Sverige: 1932 (Stockholm:Statistical Bureau, 1932),24-25 (Sweden); Statistisk4rsbokfdr Finland: 1931 new series(Helsinki: Cen- tral Bureau of Statistics,1931), 50-52 (Finland); Edward Malefakis,"Peasants, Politics and Civil War in Spain, 1931-1939,"in RobertBezucha, ed.,ModernEuropean Social History(Lex- ington,MA: Heath, 1972), 196,passim (Spain); J.C. Adams, "Italy,"in WalterGalenson, ed., ComparativeLabor Movements(New York: Russell & Russell, 1968),413-14 (Italy). POLITICAL ORDER IN INTERWAR EUROPE 463 The size of the agrarianproletariat is also unrelatedto the regimeout- come. Spain and Italy,both of which became fascist,had huge agrarian proletariats.Germany and Austria,one of which became fascistand one of which became a traditional dictatorship,had comparativelysmall agrarianproletariats. We have fascistoutcomes, then, with both large and small agricultural proletariats.The lack of a pattern continues: Den- mark, with a large proletariat,produced a democraticoutcome. Sweden, with an agrarian proletariatsimilar to thatof Germanyand Austria,produced a democraticoutcome. Norway, with a verysmall proletariat,and Czechoslovakia, with a ratherlarge agrarian proletariat,also produced democraticoutcomes. The outcomes were determinedby politicsrather than by rural class structure.Socialists succeeded in making a coalitionwith middle peasants whereverthe agrarianproletariat had been mobilized by othersbefore so- cialistshad an opportunityto do so. So pre-empted,the lattercould freely negotiatethe issues that divided rural producersand urban workers.By contrast,where the rural proletariatwas available as in Spain, Italy, Finland, and Germany it presentedsocialists with a reservoirof poten- tial supporttoo appealing to ignore.In thesecases, the logic of democratic competitionand the lure of immediatepower underminedthe socialists' abilityto acquire long-termpower because the acquisition of a constituency among the rural proletariatinvariably required commitmentsthat alienated the middle peasants. Where socialistparties were able to avoid thisalienation, it was possible for them to find a way out of the regime crisesof the early I93os by making a bargain with middle peasants.The typicalagreement involved some combination of price supports,interest-rate reductions, debt re- structuring,protective tariffs, and tax relief in exchange for increased public works, deficitspending, and expanded unemploymentbenefits. Agreements of this sort were negotiated in Denmark, Sweden, and Czechoslovakia in I933, and in Norway in I935.i6 By enteringa coalition that satisfiedthe interestsof middle peasants and made the latterdependent on the social democraticstate, social democrats attained several objectives.Above all, theycreated a stable power base.I7 On this theycould build the corporatistinstitutions that consoli-

i6 On Czechoslovakia,see Klepetair(fn. 2), 283-375;on Scandinavia,see Sven Nilsson et al., eds., Kriseroch krispolitiki Norden under mellan rigstiden[Crises and crisispolitics in the in- terwarNordic countries](Uppsala: Nordic HistoryAssociation, I974). 17 For fascism,too, the coalitionwith rural interests provided the social foundationsfor the stabilizationof a new order.Once in power,fascists could obviouslyrely on coercion(rather than inducement)to a vastlygreater degree than could social democrats.Yet, fascistsby no means entirelyignored the materialneeds of theirconstituents; a plausible case can be made 464 WORLD POLITICS dated theirhold on the stateand the working class.'8From this position of dominance,they could, in turn,induce thecollaboration of the business community.'9And theycould carrythrough the social reformsthat the interestsof their working-classconstituents required. In Denmark, for example, in the years immediatelyfollowing the Kanslergade Agree- ment,the Social Democratic governmentsucceeded in gaining passage of legislation establishing or improving workers' compensation, invalid, unemployment,and health insurance,a pension plan, and public assist- ance.20

NORWAY, SWEDEN, DENMARK, AND CZECHOSLOVAKIA

The criticaldistinction between cases of socialistsuccess and failureis to be found in the marginal number of agrarian workers in successful movements.Because thereare no standardizeddata thatallow us to make easy crossnationalcomparisons on this score, I have relied on data on trade-union enrollments,collective agreements, intergenerational voter surveys,and election and census reports.Taken together,these figures are revealingin spiteof theirlack of strictcomparability. We will begin with Norway, Sweden, and Denmark, threecountries in which socialistssuccessfully aligned themselveswith middle peasants. Norway provides the simplestcase. Because of earlier land reformsand the exodus of much of the rural proletariatto North America, the rural social structurehad become remarkablyegalitarian by the interwaryears. As early as I900, 97 percentof farmersowned theirland and employed littleor no hired labor.2,According to an inter-generationalsurvey made by the late Stein Rokkan, agrarianvoters (middle peasants,smallholders, and day laborers) did not account formore than I3 to I5 percentof the total Labor Party vote in the I920S and I93OS.22 Even if we assume that thattheir regimes, too, acquired widespreadlegitimacy. See A. JamesGregor, Italian Fascism and DevelopmentalDictatorship (Princeton: Princeton University Press, I979), chaps. 5-9. i8 The close associationbetween social democraticpower, democratic corporatism, and the worker-peasantalliance in Scandinavia has been extensivelydiscussed. See G0sta Esping-An- dersen,Politics Against Markets (Princeton: Princeton University Press, i985), chap. 5; Walter Korpi, The DemocraticClass Struggle(London: Routledge& Kegan Paul, i983), chap. i; Phi- lippe C. Schmitter,"Modes of InterestIntermediation and Models of Societal Change in WesternEurope," ComparativePolitical Studies io (October I978), I-20. 19 See Korpi (fn. i8), 48 and passim. 20 HarryHaue et al.,Det nyDanmark, i890-I978, Udviklingslinierog tendens[The new Den- mark, i890-I978, lines of developmentand tendencies](Copenhagen: Munksgaard, i980), i68. 21 Kare Lunden, "The Growth of Co-operativesAmong the Norwegian Dairy Farmers During the Period i856-I905," CahiersInternationaux d'Histoire Economique et Sociale 6 (Ge- neva: LibrarieDroz, I973), 342. 22 Stein Rokkan, "Geography,Religion and Social Class: CrosscuttingCleavages in Nor- wegian Politics,"in Lipset and Rokkan (fn. I3), Table 26, 429. POLITICAL ORDER IN INTERWAR EUROPE 465 two-thirdsof these voterswere at least partlydependent on the agrarian labor marketfor some income,we findthat at most io percentof the party'sconstituency could have consistedof agrarianworkers. Finally, it appears thatwithin the Norwegian trade union movement,agrarian workers accounted forno more than 5 percentof the membership.23 Sweden, with a considerablyless equal distributionof land, providesa moredemanding test of thehypothesis. In the early I930S, 26 percentof the agrarian work force about 250,000 men and women-were day laborers (see Table 2). Since almost all of these were adult males, we may assume that,including wives and otheradult dependents,at least 400,000 votes were contained in thislabor force.To put thisin perspective,recall thatthe Socialist electorateaveraged about 750,000, and the total electorate about 2,200,000, during theseyears.24 Yet, the Socialist Partyand the unions almost entirelyignored thispo- tentialconstituency. As Sven Anders Soderpalm has noted: The Swedishagricultural workers' movement was weak and dividedfor a longtime. ... In themiddle of the twenties, not even 5 percentof a totalof 250,000 agriculturalworkers were organized, and onlyin thelarge estate areason Skaone,Ostergdtland and theMilar provinceswas thereany great supportfor the movement. The declinein partresulted from the advantage theagricultural crisis gave to employers,but it was also connectedto the lack of interestwhich for a long timehad distinguishedthe labor movement'sattitude toward the rural proletarians' efforts.25 The Social Democrats did not seriouslydebate appealing to agrarian workers until I930; by then,these workershad already been mobilized by the Liberal and Agrarian Parties. Liberals and Agrarianshad in fact, succeeded in mobilizing these workersagainst the Social Democrats. "In most of the election districts,"Soderpalm reports,"the majorityof farm workersvoted forthe bourgeoisparties."26 Although we cannot estimateaccurately the actual amount of support the Social Democratic Partyreceived from agrarian laborers and theirde- pendents,we can estimaterather precisely the maximum amount of support it could have received under the most favorablecircumstances. As-

23 Two sources indicatethat within the Norwegian trade union movement(LO) agricul- turaland forestryworkers together constituted only io% of themembership in I93i, and that forestryworkers were a majority.See Hans Fredrik Dahl, Norgemelloi krigene.Det norske samfunneti kriseog konflikt[Norway betweenthe wars. The Norwegian societyin crisisand conflict](Oslo: Pax Forlag,1971), 91, and BergeFurre, Norsk historic, I905-I40 [Norwegian history,I905-I940] (Oslo: Det NorskeSamlaget, I971), 200. 24 Election data forSweden are fromStig Hadenius et al., Sverigeefter 90oo[Sweden after i900] (Stockholm:Bonniers, I978), Table I.C.,306-7. 25 Soderpalm,"The CrisisAgreement and theSwedish Social DemocraticRoad to Power," in Steven Koblik, ed., Sweden'sDevelopmentfrom Poverty to Affluence, I750-I970 (Minneapo- lis: Universityof MinnesotaPress, I975), 263-64. 26 Ibid., 264. 466 WORLD POLITICS suming that the rate of participation was slightlyless for agrarian workersthan forthe population as a whole (50 percentas against 6o percent),we have an electorateof about 200,000.27 We know that threepar- ties competed forthese votes: the Social Democrats, the Liberals, and the Agrarians. If we assume that in the early I930S one-halfof these votes went to the Social Democrats an assumptionfavorable to the latter this constituencycould not have representedmore than I3 or I4 percent of the total Social Democratic vote. Denmark, with the largestday-labor population of the three Scandi- navian countries,provides an even more demanding test. Of the total agrarian adult labor forceof 48i,000 in I934, 2I0,000 (43 percent)were hired laborers (see Table 2). Because, again, theywere overwhelmingly male, we may suppose that,including wives and otheradult dependents (parents, grandparents),they representa voting population of about 400,000. To putthis in perspective,we shouldnote that in I932 theSocial Democratic electoratewas just over 66o,ooo and the total electoratewas just over i.6 million. In principle,this day-labor population should have presentedthe Social Democrats with an extraordinarilytempting target. In fact,it did not because thispopulation had alreadybeen heavilymo- bilized by anotherparty. The Radical Liberals had come into existencein i9o6 after breaking away from the Liberal Party. The Liberals were quintessentiallythe partyof the middle peasants,and the Radicals came into existence to mobilize agrarian workersagainst the Liberals. From 19o6 untilthe introductionof proportionalrepresentation in i9i8, the So- cial Democrats agreed not to competeagainst the Radicals (a partyled by bourgeois reformers)in rural Jutland.In exchange, the Radicals prom- ised not to challenge the Social Democrats in a numberof urban constituencies. This gave the Radicals an open field in the districtsin which agrarianlaborers were concentrated.As a result,by the timeproportional representationwas introduced,the Radicals had captured this constituency. If we assume that the rate of participationfor this potentialelectorate of 400,000 was slightlyless than for the entirepopulation (6o percentas against70 percentin the earlyI930s), we have an actual electorateof 240,000.28 We will furtherassume thatthe partythat had historicallyor- ganized thiselectorate received half of itsvote-or I20,000 votes which

27 The rate of participationequals the share of the valid vote cast as a percentageof the populationover twentyyears of age. See PeterFlora, ed.,State, Economy and Societyin Western Europe, i8I5-I975. A Data Handbook.Vol. i: The Growthof Mass Democracies(Chicago: St. JamesPress, i983), I4I-44. 28 The rate of participationequals the share of the valid vote cast as a percentageof the populationover 20 yearsof age. Ibid., I05-6. POLITICAL ORDER IN INTERWAR EUROPE 467 in factapproximates what the Radicals receivedin rural districtsin I935. We will also assumed thatthe remainderwas distributedamong the So- cial Democrats and all other partiesat a ratio of four to one. Under the foregoinggenerous assumptions,90,ooo of these votes went to the Social Democrats.29With a total Social Democratic vote of 760,000 in I935, this representsabout I3 percentof the Social Democratic electorate.Finally, we should note that the participationof agrarian workers in Socialist trade unions during theseyears was virtuallynil. In sum, in Norway and especiallyin Sweden and Denmark, the mobilization of agrarian workers by other partieswas decisive. The objec- tion can be made thatthe real distinctionbetween these countries and the countriesthat became fascist-especially Italy and Spain-is that,in the former,agrarian workers were mobilized by other parties because the transitionto mass politicswas more gradual. With a more sudden transition,this population would not have been captured,and socialistswould have stepped in. In thisconnection, the case of Czechoslovakia, in which the transitionwas extremelyabrupt, becomes significant. In Czechoslovakia, the working-classmovement was divided ethni- cally, regionally,and ideologically.The German Social Democrats, because of theirethnic base, organized almost exclusivelyindustrial workers. The [National] Socialists,limited to Bohemia and Moravia, were also exclusivelyurban. Bohemian and Moravian agriculturalworkers-there were not many of them-were in fact initiallyorganized by the Social Democratic Partyduring the firstfew years of the Republic. But, because these were heavily industrializedareas, agriculturalworkers were in a less dependent position.They routinelymoved back and forthbetween industrialand agriculturalemployment. In I92I, only 28.4 percentof the labor forceof Bohemia was in agriculture.The percentagesfor the other provinceswere: Moravia and Silesia, 33.8; Slovakia,58.7; and Ruthenia, 62.4.30 The great majorityof agrarian workerswere thus located in Slo- vakia and Ruthenia. A conservativeestimate would be that,in Slovakia, 40 percentof the agrarianlabor forceconsisted of proletarians;in Ruthe- nia it was probablyclose to half. In these provincesthere was almost no industry,and large surplus rural populations were dependent on the agrarian labor market. The Social-Democratic Party never had an opportunityto organize theseworkers. In the schismbetween social democracyand communism, the Communists capturedthe partyorganization in Slovakia and Ruthe-

29 For electiondata forDenmark, see ibid. 3?Vaclav L. Beneg,"Czechoslovak Democracyand Its Problems,i9i8-I920," in Mamatey and Luza (fn. I2), Table 2, p. 4I. 468 WORLD POLITICS nia, while Social Democrats prevailed in the industrialareas of Bohemia and Moravia. The result was that the Social Democratic Party became mainly a Bohemian and Moravian party and mainly one of industrial workers.Thus, in the firstpost-schism election, in 1925, the Social Dem- ocraticParty was wiped out in Slovakia and Ruthenia,receiving 4.3 per- centand less than 2 percentof the vote,respectively. In contrast,the Com- munistParty received I3 percentof the Slovakian vote; with 30.8 percent of the Ruthenian vote,it became the largestparty in Ruthenia.3' The Communist Partycaptured the social democraticorganization in these regions, but it did not succeed in mobilizing agrarian workers.32 These workerswere mainlyorganized by the regionalCatholic Populist partiesand the secular Agrarianparties. The agrarians,who were mainly parties of middle peasants, were especially successfulin capturing this constituencybecause of theirleading role in the land reformprogram. They could unite middle peasants and day laborersbecause land reform, which consolidated mainlythe positionof the former,was aimed specif- ically at Magyar and German landowners; it consequentlyexploited ethnic cleavages in the countryside.33The net effectof the schism and the land reformwas thatthe Social Democrats neverdeveloped a significant involvementin rural class conflict.The democraticworking-class movement (Social Democrats, [National] Socialists,and German Social Dem- ocrats) remained urban and industrialin its orientation,and freeto col- laborate with the representativesof the middle peasants (the secular Agrarian partiesand the Catholic Populist parties). The case of Czechoslovakia also demonstratesthat the many overlap- ping cleavages were not sufficientto preventa stable democratic coalition; neither was the absence of a coherentpeasant party. In this, the Czech example stands as a useful correctiveto the propositiondrawn fromthe Spanish, Italian, and German cases: thatthe absence of coherent peasant representationand the multiplicityof cleavages in thosecountries enfeebleddemocracy. It was still possible for the Socialists to make a coalition with middle peasants to overcome theirconsumer-producer, religious, and regional conflicts-even when peasants were themselvesdivided by these cleavages. The socialistswere divided among themselvesby ethnicityand region (autonomist German Social Democrats versus centralizingBohe-

31Victor S. Mamatey, "The Development of Czechoslovak Democracy, I920-I938," in Mamateyand Luza (fn. I2), I29. 32 The ruralcommunist vote came mainlyfrom ethnic minorities. See Richard V. Burkes, The Dynamicsof Communismin EasternEurope (Westport,CT: Greenwood Press, I976), 42 and passim. 33Czech land reformis discussedin Klepetair(fn. 2). POLITICAL ORDER IN INTERWAR EUROPE 469 mian-Moravian Social Democrats versus centralizingSlovakian Social Democrats). The peasant partieswere divided fromthe Social Democrats by region (centralizing Social Democrats versus autonomist Slovakian Populists) and by religion(secular Social Democrats versusCatholic Bo- hemian-Moravian Populists versus Catholic Slovakian Populists). And the peasant partieswere themselvesdivided by religionand region (secular, centralizingBohemian-Moravian Agrarians versus Catholic, cen- tralizingBohemian-Moravian Populists versus secular,centralizing Slo- vakian Agrariansversus Catholic autonomistSlovakian Populists).These cleavages provoked conflictsbitter enough to be definingfeatures of the partysystem; but so long as the coalitionserved the otherinterests of the participants,they were willing to avoid a showdown on theseissues.

SPAIN, ITALY, AND GERMANY

The positionof agrarianlaborers in the socialistmovement looks quite differentwhen we turn our attentionto Spain, Italy,and Germany. Al- though it is difficultto estimatethe votingpreferences of agrarian workers, data on the composition of trade unions make it clear that these workersexercised enormous leverage in the socialistmovements of Italy and Spain; in Germany theirinfluence was less, but still substantial.In these countries agrarian labor had remained largely unmobilized by other political groups; therefore,it presentedsocialist parties and trade unions with a vast reservoirof potentialsupport. The decisive class conflictin the Italian countrysideoccurred in the northand center;it pittedday laborers(braccianti) against a mixed group of sharecroppers,tenants, and small and large owners. The two most im- portantpoints of contentionwere the local monopolyover hiring,which the sharecroppers,as periodic employers,resented no less than tenants and peasant owners, and the ownership of harvestingequipment, the sharecropperscaring little for the socialistvision of collectiveownership.34 The socialistcause was pressed in the countrysideby the Syndicalists (USI) and Federterra.The latterwas the agrariansection of the normally reformist,socialist Confederazione Generale del Lavoro (CGL); it provided 33 percentof the CGL's membershipin 1920.35 Federterrasought to bridge the conflictbetween laborers and sharecroppers,but its massive postwar mobilization of laborersand its collectivistambitions made this impossible. As its membershiprose fromabout 93,000 in i9i8 to over

34 C. Seton-Watson,Italy from Liberalism to Fascism(London: Methuen,i967), 303-4. 35 Daniel L. Horowitz, The Italian Labor Movement(Cambridge: Harvard University Press, i963), I24. 470 WORLD POLITICS

300,000 in i9I9, Federterraleaders lost controlof theirmembers and felt compelled to act more militantlythan theirpast experiencewould other- wise have recommended.36 The strategyof jointlyorganizing sharecroppersand laborersbecame untenable as well, because the end of the war saw a dramaticincrease in the size of the middle peasantryas laborers became sharecroppersand landowners: the percentageof owners rose from i8.3 of the males engaged in agriculturein I9II to 32.4 in I921.37 Federterraleaders feared that theirconstituency would shiftits politicalloyalty as its class position changed. As a solution to this, Federterra advocated collective ownership of equipment among sharecroppers.This was an attemptto bind sharecroppers to the Federterra organization. Instead, this new emphasis drove sharecroppersinto the arms of local Catholic unions (CIL), thus trans- formingFederterra: by I920, agrarian laborersprovided virtuallyall of Federterra'smembership, but theyaccounted foronly io percentof the agrarian membershipof the CIL. The remainderconsisted of sharecroppers,tenants (78.4 percent),and small owners (io percent).38 Attentionto the Federterraalone providesan incompletepicture. The CGL, or at least itsurban, industrial wing, had always followeda reformist course, and this did not change substantiallyafter the war. But the CGL was not the only importantsocialist organizer of day laborers. In I920, the three syndicalistunions (USI, UIL and the independent rail- waymen s union) organized about one-thirdof the socialistlabor force.39 The agrariancomponent-overwhelmingly laborers, mainly in Emilia probably accounted for about half of the syndicalistmembership.4o The effectof the syndicalistunions was to furtherincrease the dependence of the socialistmovement on day laborers:once theseworkers had been po- liticallymobilized by a part of the left,it became imperativethat the So- cialist Partysolicit their support. The role of the day laborers,however, did not stop even there,because of the special place theyoccupied in bolsteringthe positionof the intran- sigentsin the Socialist Party. Intransigentswere dominant in the party

36 Maurice Neufeld, Italy:Schoolfor Awakening Countries: The Italian Labor Movementin Its Political,Social and EconomicSetting from i8oo to 1960 (Ithaca, NY: Cornell University School forIndustrial and Labor Relations,i961), 367. 37Maier (fn. I 0), 3 I I . 38 Horowitz (fn.35), I24- 39 See Neufeld (fn. 36), 347. Also see Michael F. Hembree, "The Politicsof Intransigence: CostantinoLazzari and the Italian SocialistLeft, i882-I9I9," Ph.D. diss. (Florida State Uni- versity,I98I), 232. 4?The agrarianmembership is an estimate.For the USI, it was assumed to be equal to the membershiptaken away fromthe CGL: 8o,ooo.See Horowitz (fn.35), 75. POLITICAL ORDER IN INTERWAR EUROPE 471 leadershipand in the syndicalistunions. Accommodationistswere dominant in the parliamentaryparty and in the CGL. Since the intransigents could not rely on the supportof the CGL, theywere inclined to accen- tuate the claims of the syndicalistunions and, therefore,of the agrarians. So much so, it turnedout, thateven when the fascistthreat had convinced the leaders of the CGL of the need to reach an accommodationwith the Catholic unions in organizing the sharecroppersand small landowners, the Socialist Partyleadership resisted.4' The lines of class conflictin the Spanish countrysidediffered from those in Italy. In Spain, class conflictpitted day laborers,tenants, and marginal owners (all of whom were dependent on the labor market) against middle and large owners; some of the latter were directlyin- volved in cultivationand some were membersof the urban bourgeoisie. Two unions had traditionallymobilized farmlabor: the anarchists(in the CNT, Confederacion Nacional del Trabajo) and the socialists (in the UGT, Union General de Trabajadores). Until Primo de Rivera's dictatorship,the CNT had been by farthe more importantorganizer of agrarian workers. By the I930s, however, the CNT was a spent force in the countryside,mainly because of the heavy hand of the dictatorship.The UGT, by contrast,had been favored by the dictatorshipand, with its well-establishedpresence among industrialvoters, was ready to move forcefullyin the agriculturalareas. The Federacion Nacional de Traba- jadores de la Tierra (FNTT), the agrarian wing of the UGT, increased its membershipfrom a mere 36,639 in 1930 to 451,337 in 1933, thereby accounting for about 40 percent of the total UGT membership.42The proportionof the FNTT membershipthat was dependentmainly on the labor market forits income was veryhigh, probably in excess of 8o percent. One indicationof thisis, accordingto Edward Malefakis,the close re- lationshipbetween the Socialist Party(PSOE) vote and the size of the la- boringpopulation. Malefakis observed the disparityin the patternof land tenure between the north (small familyowners), the centerand littoral (small and intermediateproperties), and the southwest (large estates worked by day laborers). He demonstratedthat this distributiontrans- lated into a steady voting pattern:rightist deputies were elected almost entirelyby the small peasant proprietorsof Old Castile and Navarre; the Socialists were the largestparty in southwesternSpain; and the political centerwas elected in the geographiccenter.43 The PSOE leftno doubt about the kind of reformit wanted; land had

41Seton-Watson (fn 34),599. 42 Malefakis(fn. I5), 292. 43Ibid., 206. 472 WORLD POLITICS tobe redistributedto thepoorer rural workers, especially the landless day laborers.The partybore primaryresponsibility for the I932 legislation thatthreatened the land titlesnot only of large landowners in thesouth- west,but also of smallowners who leasedout theirland in Old Castile and Navarre.In I93I, thePSOE and theother San Sebastianparties en- acteda seriesof emergencydecrees that forbade expulsion of smallten- ants;forbade owners from withdrawing their land fromcultivation; es- tablishedrules that favored collectives in therenting of large properties; establishedan eight-hourday and a defacto wage increase;established whatamounted to a closedshop in thelabor market, thereby vastly in- creasingthe power of the FNTT; and establishedarbitration boards that generallyreinforced the leading position of the FNTT. Emboldenedor driven by its growing support among the agrarian proletariat,the FNTT was thesocialist organization that initiated the turn towardradicalization. This radicalizationalienated the PSOE and UGT fromthe Catholic middle peasants of thenorth and center.Indeed, this radicalizationbegan with an agriculturalissue. In JuneI934, theFNTT calleda nationwidestrike in responseto therepeal of the I93I Municipal Labor Act,the act thathad createdthe arbitration boards on whichso muchof the FNTT's influencedepended. The strikereceived only luke- warmsupport from the urban wing of theUGT and eventuallyfailed, causingthe loss of tens of thousands of union members. In itsaftermath, and especiallyin thewake ofthe failed October Revolution, the gulf betweenthe two wingsof thesocialist movement, and betweenthe move- mentand the middlepeasants, became even greater.The polarization came to a head in February I936, when middle peasantsin the centerand northvoted overwhelmingly for the Catholic Confederacion de Derechas Autonomas(CEDA) and agrarianlaborers in thecenter and southwest votedoverwhelmingly for the socialists.44This electoraloutcome mir- roredthe alignments that formed on behalfof fascismand theRepublic a fewmonths later.45 AlthoughGermany presents a morecomplicated picture, the conclu- sionis thesame. Notwithstanding their conventional images as quintes- sentiallyurban and working-class,the Social DemocraticParty and the freetrade unions (ADGB) wereboth dependent on theagrarian proletariat.Let us note,to beginwith, that in I928 purelyrural districts gave no less than 25 percentof theirvote to the socialistparties (SPD and KPD) and 2I.5 percentof theirvote to theSPD alone.46These data be-

44 Ibid., 214. 45 Ibid., 2I9. 46 Derek Urwin and Frank Aarebrot,"Socio-geographic Correlates of LeftVoting in Wei- mar Germany,I924-I932," in Per Torsvik,ed., Mobilization,Center-Periphery Structures and POLITICAL ORDER IN INTERWAR EUROPE 473 come more tellingwhen we considerthe strengthof the ADGB-affiliated agrarian workers' union, Deutscher Landarbeiterverband(DLV). Until i9i8, Germany'sagrarian workers about 3,000,000 in all had remained politicallyunorganized. Agrariantrade unions were illegal until the collapse of the empire. This segmentof the proletariatmust have seemed an especially temptingtarget for the Social Democrats and the free trade unions. Indeed, in the SPD's vision of the political and economic developmentof the country,these workerswere assigned a vital role. Politically,it was necessaryto weaken the position of the Junkers. Land reformwas not a plausible solutionbecause in early i919, when the SPD mighthave had thepolitical power to introduceit, there were severe food shortages.Moreover, it was unsatisfactorybecause it conflictedwith the SPD's axiom that the "economic development of the countryde- pended on large-scale production in both industryand agriculture."47 Land reformwould have increased the number of allegedly inefficient small farmsas well as the size of a middle peasantrythat "showed no in- clinationtoward a socialismthat considered them fated to lose theirhold- ings."48Like their Spanish and Italian counterparts,the German Social Democrats faced the dilemma of mobilizing agrarianworkers while not raisingthem out of the proletariat. The German solutionwas to apply an industrialmodel to agriculture. Agriculturalworkers would be collectivelyorganized and would share in all of the labor market and social policy rightsthat applied to industrial workers.This solution addressed all of the concernsof the Social Dem- ocrats.Increases in wages and taxes would weaken the Junkers,preserve large units of production,prevent food shortages,and would not create more middle peasants.At the same time,workers, "at least a considerable number of them, could be won for the Social Democratic Party and unionized....549 At the end of i9i8, there were only 20,000 unionized agricultural workersin all of Germany.By I920, theirnumber had risento about one million.Membership then declined and stabilizedat about 200,000 forthe remainingperiod of the Republic. This is in contrastto about 95 million industrial workers in the ADGB in I923, and a stabilized industrial membershipof about 4.9 millionthereafter. It would appear thatsocialist unions in the countrysidehad no more than a residual role. At the peak of unionization, about one in ten ADGB members was an agrarian

Nation-Building (Oslo: Universitetsforlaget,198I), 259-60. 47 Frieda Wunderlich,Farm Labor in Germany,i8io-I945 (Princeton:Princeton University Press, 196I), 35. 48 Ibid. 49Ibid., 35-36. 474 WORLD POLITICS worker; during the last half of the Republic, only about one in twenty- fivewas.50 These figuresseverely understate the impact of the agrarian unions, however. This becomes clear when we examine data on the industrial and agrarian workers who were covered by collective contracts.The number of industrialworkers covered by collectiveagreements union members or not was I3,I35,000 in I924, I2,276,000 in I 929, and II,950,000 in I93I.5I The equivalent figures for agrarian workers were 2,37I,7I9 in I924, 2,749,398 in I926, and 2,I23,IIO in I928.52 Since the total population of agricultural laborers was only about 3,000,000, this means thattwo-thirds of themor more were coveredby collectiveagreements.Twenty percentof the total labor forcecovered by ADGB agree- mentswere in agriculture.We now begin to see the potentialimportance of thisconstituency to the SPD and the ADGB. It seems likelythat a large part of the 25 percentof the rural vote that went to the SPD and KPD came fromthis constituency. We can see even more clearlythe mannerin which thispowerful union presence-and the largerpolitical and economic strategyof which it was a part-was bound to antagonize the middle peasants.For the corollary of that strategywas that the SPD and the ADGB took a stronginterest in the living standardsof agrarian laborers.Indeed, in the firstyears of the Republic, the SPD succeeded in winning passage of legislationthat extended all of the social policy rightsof industrialworkers to agrarian workers,guaranteed them the rightto unionize, and extended to agriculture the same systemof compulsoryarbitration that was used in in- dustrialwage disputes.All agrarianworkers became covered by compul- sory health, pension, unemployment,and disabilityinsurance schemes. The costs of these programswere covered in the same manner as those for industrialworkers: by joint employee and employercontributions.53 German agriculturalworkers were surelythe mostcomprehensively pro- tectedin Europe in the I920S. These benefitswere being extended to agrarian workersduring years when the middle peasants' income was collapsing. The net share of national income going to agriculturefell from I3.0 percentin I9I3 to 7.2 percentin I929.54 Several studiesprovide plausible evidence that the in-

5?Gerhard Bry, Wages in Germany,i87I-I945 (Princeton: Princeton UniversityPress, i960), 32. 51Ibid., 42. 52 Wunderlich(fn. 47), 84, n. I4. 53Ibid.,I 26-59. 54Paul Jostock,"The Long-Term Growth of National Income in Germany,"in Simon Kuznets, ed., Incomeand Wealth,Vol. V (London: Bowes and Bowes, I955), I09. POLITICAL ORDER IN INTERWAR EUROPE 475 come of a peasant actuallyfell below thatof a worker in the late I 920S.55 In thiscontext, the peasants' almosthysterical attacks on Sozialpolitik,and theirantipathy to theSPD, becomecomprehensible. On balance,then, it appearsthat the GermanSocial Democrats- throughtheir efforts on behalfof thesocial rights of agrarianworkers, legislationthat improved the market position of theseworkers, and the generalizationof collectiveagreements also becameentangled in rural classconflict.

THE TRADITIONAL DICTATORSHIPS One questionremains: Why, in thecountries discussed so far,did lib- eralsnot enter into a coalitionwith middle peasants? The factis thattra- ditionalliberals were involvedin such coalitionsin Finland,Austria, Hungary,and muchof easternEurope. In thesecountries, the outcome was a rathertraditional authoritarian regime that was distinguished, aboveall, byits lack ofinstitutional innovation. Finlandexperienced a semi-competitiveparliamentary order that was premisedon theharassment and outlawingof the Communist Party, the disablingof trade unions, and theenfeeblement of the Social Democratic Party.With participation and dissentthus constrained, it was possibleto preservea circumscribedparliamentarism. The Dollfuss-Schuschniggre- gimein Austriawas a morefully formed dictatorship; it lastedfrom the briefcivil war in I934 untilthe Anschlussin I938. Under the dictatorship, the parliamentwas closed, the Communistand Social Democratic parties and trade unions were outlawed, and democraticfreedoms were circum- scribed.Even so, the interestsincluded in the rulingcoalition were given rather broad latitude for dissent,and the working-classorganizations thatwere formallysuppressed were allowed to carryon underground.In Hungary, political competitionwas also severelycircumscribed, and the Social Democratic Partyand trade unions were able to operate only under veryrestricted conditions. These regimeswere thus typicalof traditional authoritarianismin two ways: in theirtolerance of limiteddissent and in theirdisinterest in creatingsubstitutes for the working-classinsti- tutionsthey suppressed.56

55 Adolph Mtinzinger,"Eine bduerlicheBetriebserhebung in Wiirttemberg"[A surveyof peasant production in Wiirttemberg];Der Arbeitsertragder bduerlichenFamilienwirtschaft, Vol. II (Berlin I929), 873; Constantinvon Dietze, "Die Lage der deutschenLandwirtschaft" [The position of German agriculture],Jahrbficher fir Nationalakonomieund Statistik(Jena: I929), Vol. I30, p. 3; Vol- 75, pp. 659-60. 56 See Risto Alapuro and Erik Allardt,"The Lapua Movement:The Threat of a Rightist Takeover in Finland, 1930-32," and Walter Simon, "Democracy in the Shadow of Imposed 476 WORLD POLITICS They were certainlynot fascist.That much is made clear not only by the absence of fascistinstitutions, but by the factthat in Finland and Aus- tria the regimescame about in part as a reactionto growing fascistmovements-the Lapua Movement in Finland and the Nazi Partyin Austria. Indeed, the Austrian governmentput down an attemptedNazi coup in I934 and, then,in the face of the impendingAnschluss in I938, called a referendumon Austrian independence and invitedSocial Democrats to participatein a cabinetof national unity. What distinguishedthese traditionalliberals and allowed them to in- stall these dictatorshipswith the aid of agrarians?The answer appears to be that in these countriesliberals were more cohesive. In Finland, the bourgeoisie was divided into a Swedish Liberal Party,a Finnish Liberal Party,and a ConservativeParty. These divisions remained in the party system,but were largelyovershadowed by the experienceof the civil war, which forcedthe bourgeoisieto uniteor perish,and by the constantthreat posed by the existenceof a Bolshevik regime in neighboringRussia. In Austria,the bourgeoisiewas less divided. A substantialpart of it was already organized by the clericalChristian Social Party,which itselfwas a coalition of bourgeois and middle peasant interests.Along with some anticlerical German nationals, the Christian Social Party provided the core of the authoritarianVaterland Front. What is strikingabout Fin- land, Austria, and Hungary is that all threeexperienced civil wars that were foughton explicitlyclass lines and thatrequired the bourgeoisieto overcome its historicaldivisions in order to survive.57 Finland and Austria were also distinguishedby the exceptional co- hesivenessof the middle peasantry.In each of these countries,the peasants were organized in one party the Agrarian in Finland and Chris- tian Social in Austria. Because these were the largestnonsocialist parties in each countryand because the bourgeoisiewas less cohesive,it was possible for them to dominate any coalition. In Hungary the peasants were not as cohesivelyorganized, but theynonetheless provided a mainstayof the Horthy regime. Elsewhere, traditionalliberals failed; the liberal communitywas di-

Sovereignty:The First Republic of Austria,"both in JuanLinz and AlfredStepan, eds., The Breakdownof Democratic Regimes: Europe (Baltimore:The JohnsHopkins UniversityPress, I978). On Hungary, see Andrew Janos,The Politicsof Backwardnessin Hungary,i825-I945 (Princeton:Princeton University Press, i982), chaps. 4-6. 57 Lithuania,Latvia, and Estonia providetextbook cases of coalitionsbetween liberals and middle peasants leading to traditionaldictatorships between the wars. In each case, a brief class-basedcivil war followedimmediately on World War I, and in each case, socialistswere active in organizingagrarian labor in the period beforethe instaurationof the dictatorship. See Georg von Rauch, Die Geschichteder baltischenStaaten [The historyof the Baltic states] (Stuttgart:W. Kohlhammer,1970). POLITICAL ORDER IN INTERWAR EUROPE 477 vided intoa multiplicityof parties,none of whichwas largeenough to claimplausibly to speakfor the bourgeoisie. The result,on theone hand, was thatnone of them could offer middle peasants a valuableally and, on theother hand, that none of them,individually, had an incentiveto be- comesuch an ally.In thesecases, a bargainwas struckwith middle peas- antsby either socialists (in Norway,Sweden, Denmark, Czechoslovakia) or by entirelynew fascistmovements (in Germany,Italy, Spain), which sufferedneither the historicalcleavages that divided and paralyzedlib- eral partiesnor the involvement in ruralclass struggle that handicapped socialistparties. In short,the preconditions of fascistsuccess were a di- videdliberal community and a working-classmovement engaged in a de- fenseof the rural proletariat.

CONCLUSIONS In developingthis proposition, we haveseen that explanations of interwar politicscouched in termsof a multiplicityof overlappingcleavages or theabsence of a singlecoherent peasant party fail to accountfor the regime outcomes.58The foregoingdiscussion has also placed in a new lightseveral common interpretations offascism and democraticinstabil- itybetween the wars.Barrington Moore and AlexanderGerschenkron haveargued that a fascistoutcome was causedby the existence of a landed elitewhich added the supportor acquiescenceof a subordinatedrural mass to a coalitionwith the bourgeoisie.59In thisview, what distin- guishedsocieties was thepresence of a landedelite in controlof the rural masses:England and France becameliberal democracies because they lackeda politicallydecisive landed elite; Germany became fascist because itpreserved one. When appliedto theinterwar years, Moore's and Gerschenkron'sar- gumentis questionablefor several reasons. First, as we have observed, thereis no correlationbetween the rural social structure (in particularthe size ofthe dependent labor force), and theregime outcome. Second, even ifa landedelite exercised economic control of a ruralmass, it did notnec- essarilycontrol the latter'spolitical behavior. In Spain, for example, southernday laborersvoted overwhelminglyfor the SocialistParty throughoutthe life of the Republic. Third, an authoritarianoutcome did notrequire that a landedelite have political control of a ruralmass. When an independentpeasantry sided withthe bourgeoisie,an authoritarian outcomeensued even in theabsence of an importantlanded elite.Fin- 58 See the discussionof the Czechoslovak case. 59 Moore (fn. i); Gerschenkron(fn. i). 478 WORLD POLITICS land's smalland mediumpeasants, never noted for their subordination, alignedwith the bourgeoisie, first to win thecivil war and thento main- tainan authoritarianregime. In Spain,the critical support for a fascist solutioncame not from southern landed elites, who had alwaysbeen au- thoritarianbut unable to controltheir workers, but from middle peasants in thenorth and center. A finalexplanation for the weakness of interwardemocracy points to a lack of cooperationamong democrats, engendered by acutepolariza- tion.60However, polarization did notso muchcause specific outcomes as reflectthe inabilityto find any stable solution.Some countriesthat adoptedauthoritarian regimes-Spain and Austriain particular-were in fact conspicuously depolarized in 1931-1933 and 1919-1923 respectively.6'Some countriesthat became democratic-particularly Norway and Sweden were in factamong the most polarized in the 1920S and had someof themost radicalized labor parties in Europe.62Where dem- ocratsfailed, it was not becausetheir societies had becomepolarized; rather,the societies became polarized because democrats had failed. In fact,it mattered little in theend whethersocialist leaders were committedto radicalismor reformism.63The unqualifiedcommitment to the reformist,electoral road to power,which was pursuedby the German and,initially, the Spanish socialists, could be as self-defeatingas the more radicalposture taken by theirItalian counterparts. Indeed, the Spanish experiencesuggests that such a strategywas thecause of radicalization. Moderationand democraticstability required that socialists ignore a substantialpart of the working class, and no socialistleaders consciously de- cided to do so. Wherethey did, it was becausethey had no choice.And wherethey had no choice-whetherthey were initially radical after the war,as in Norwayand Sweden,or reformist,as in Denmarkand Czech- oslovakia-the outcomewas stabledemocracy.

60 This is the centraltheme of the essaysin the volume editedby Linz and Stepan (fn.56). 6, See Juan Linz, "From Great Hopes to Civil War: The Breakdown of Democracy in Spain," and Walter Simon, "Democracy in the Shadow of Imposed Sovereignty:The First Republic of Austria,"in Linz and Stepan (fn.56). 62 The levels of Norway's and Sweden's strikeactivities were among the highestin Europe duringthe I920S. See WalterKorpi and Michael Shalev,"Strikes, Power and Politicsin West- ern Nations, I900-I976," PoliticalPower and Social Theoryi (i980), 30I-34; Douglas Hibbs, "On the PoliticalEconomy of Long-Run Trends in StrikeActivity," British Journal of Political Science2 (April I978), 26-43. 63 Among the mostimportant efforts to accountfor the origins of radicalismand reformism in labor movementsis Lipset (fn. I4), esp. pp. I4-i6. Lipset contendsthat radicalism and re- formismwere one among severalimportant variables in accountingfor regime outcomes betweenthe wars.