Patterns of Political Instability: Affiliation Patterns of

76 Studies in Comparative International Development Summer 2003

Marcus Kreuzer is assistant professor of political science at Villanova University His work focuses onhow electoral and legislative institutions shape the organizational and electioneering practices ofparties in interwar Europe and post-communist democracies He also is studying the origins of liberaldemocracy in nineteenth century Europe He is author of Institutions and Innovation ndash Voters Politi-cians and Interest Groups in the Consolidation of Democracy France and Germany 1870-1939(2001)

Vello Pettai is lecturer in political science at the University of Tartu Estonia He specializes in compara-tive ethnopolitics and party politics He has published previously in Nations and Nationalism Post-Soviet Affairs East European Politics and Society and Journal of Democracy

Studies in Comparative International Development Summer 2003 Vol 38 No 2 pp 76-98

Research Forum

Patterns of Political InstabilityAffiliation Patterns of Politiciansand Voters in Post-CommunistEstonia Latvia and Lithuania

Marcus Kreuzer and Vello Pettai

In contrast to established party systems the transformation of post-communist partysystems is not only shaped by shifts in electoral preferences but also by the chang-ing organizational loyalties of politicians Post-communist politicians pursue awide range of organizational strategies such as party fusions fissions start-ups andinterparty switching By focusing on the interaction between these organizationalstrategies and votersrsquo electoral preferences we argue that the seeming instability ofpost-communist party systems actually reveals distinct patterns of political changeThe article develops an analytical framework which incorporates politician-driveninterparty mobility and voter-induced electoral change It uses this framework toshow that the apparently inchoate party systems of Estonia Latvia and Lithuaniaactually follow definable modes of transformation

Among political scientists there is considerable agreement that party sys-tems are central if not paramount for the consolidation of democracies

Elster Offe and Preuss speak for a long line of scholars when they note thatldquothe structure and the interaction of political parties are the most significantvariables which contribute to the consolidation or failure of the political sys-tems of democratic politicsrdquo (1998 110-11) However there is far less agree-ment about what consolidates party systems in the first place One school ofthought inspired by Lipset and Rokkanrsquos work views party system institu-tionalization as a bottom-up process in which the gradual coagulation of voter

Kreuzer and Pettai 77

preferences produces stable socio-economic cleavages and reasonably durablevoter alignments Another school of thought points to partiesrsquo organizationalweakness volatile electorates and politiciansrsquo continuous interparty mobilityto conclude that party system formation is a top-down process shaped prima-rily by historical contingencies and strategic action

These two schools of thoughts have produced research that differs consid-erably in its theoretical sophistication and methodological rigor The first struc-tural line of inquiry borrows extensively from literatures on partisanidentification path dependence and political sociology and engages in exten-sive cross-national survey research (Kitschelt et al 1999 Miller et al 1998Reisinger et al 1995 Toka 1998) By contrast the second political-researchapproach has produced numerous empirically rich but theoretically and meth-odologically underdeveloped studies Commonly these studies stress the or-ganizational flux of parties along with exorbitant levels of volatility andhyper-fragmentation and conclude that party systems do not yet exist (Elsteret al 1998 Geddes 1995 Grofman et al 2000 349 Lewis 2000 Mair 1997175) Our goal is to build a bridge between these two schools by pointing toand partly remedying the analytical and theoretical shortcomings plaguing thesecond political line of inquiry

The key for developing a more sophisticated political analysis of party sys-tem development is to place affiliation patterns of both voters and politicianson a firmer analytical and conceptual footing Studies emphasizing politicalfactors are surprisingly unanimous in considering votersrsquo lack of partisan iden-tification and politiciansrsquo shifting organizational loyalties as the key micro-obstacles to the macro-institutionalization of party systems Yet despite thistacit agreement there has been no systematic attempt to differentiate thesedifferent forms of electoral and organizational disloyalty analyze their inter-actions compare them across countries and ultimately provide a theoreticallygrounded explanation1 We will do so first by elaborating a typology of politi-ciansrsquo organizational and votersrsquo electoral affiliations second by exploringhow the combination of these two levels of affiliation produces distinct pat-terns of party system institutionalization and third by using this analyticalframework to compare the development of party systems in Estonia Latviaand Lithuania In doing so we hope to bring the same level of analysis to thepolitical underpinnings of party systems that has been achieved in examiningtheir structural causes And in creating a more equitable playing field we alsohope that it will become easier to assess the respective causal importance ofthese two dimensions as well as investigate the possible interactions existingbetween them2

Analytical Framework Electoral and Organizational Affiliation Strategies

Given that stable Western party systems constitute the primary point of depar-ture for analyzing post-communist party systems the lack of stable electoralpartisan identification and loyal organizational affiliations among politiciansin post-communist democracies has widely been read as evidence for the ab-sence of any meaningful party systems Mair for example points out that ldquothe


very notion of a newly emerging party system [in Eastern Europe] may well bea contradiction in terms in that to speak of a system of parties is to ascribesome degree of stability and predictability to the interactions between the par-ties concernedrdquo (emphasis in original) (Mair 1997 175) Such Western-centricreasoning overlooks the fact that instability can have patterns just as distinct asthose of stability That is the formation of new party systems is subject toregularities just as much as the transformation of established party systemsHowever to identify such regularities we have to stop treating instability as anundifferentiated residual category and move towards a classificatory frame-work capable of capturing patterns of electoral and organizational non- orreaffiliation The key to detecting patterns of post-communist party systemtransformation is a closer analysis of how politicians shift their organizationalallegiances and how voters align themselves with parties

Organizational Affiliation Strategies

One important reason why scholars fail to recognize any patterns in the insta-bility of post-communist party systems is their neglect of politiciansrsquo organiza-tional affiliation choices (Kreuzer and Pettai 2002) Zielinski correctly notesthat the ldquoperiod of organizational confusion [in transitional democracies] isfrequently perceived as temporary phenomena lacking any wider theoreticalsignificancerdquo (2002 184-85) Yet Goldie Shabad and Kazimierz Slomczynskihave shown that the ldquopolitical tourismrdquo practiced by organizationally disloyalpost-communist politicians does follow distinct itineraries (2001 2) Buildingon this work we differentiate between five possible affiliation strategies avail-able to a politician in developing hisher political career We describe the strat-egies below a summary overview is provided in Appendix 1

i) Staying Put Staying put involves a politician remaining affiliated withhisher current party Staying put also covers a partyrsquos temporary membershipin an electoral alliance (apparentement) even when such an alliance adopts aname different from the politicianrsquos original party

ii) Party Switching Party switching occurs when a politician reaffiliateswith another existing party Both the exit and entry parties have to have beenin existence for at least one election prior to the switch otherwise we classifysuch reaffiliations as fission or start-up (see below) Party switching is alsodistinguished from fusion where the majority of a partyrsquos members reaffiliatein the form of a merger with another party In this respect it is important tonote that our definition of party switching is more restrictive than that em-ployed by others Shabad and Slomczynski as well as Desposato for exampleconsider any form of interparty mobilitymdashthat is both fusions and fissionsmdashas party switching

iii) Fusion Fusion is a collective reaffiliation strategy through which a ma-jority of members from two or more parties merge to form a totally new partyFusions can involve parties which have not participated in at least one electionprior to the merger Parties however are considered merged only if they inte-grate their organizations adopt a new name (even if only a hyphenated ver-sion of their old names) and agree on a single party leader Mere electoral


alliances are not considered mergers even if they involve a special name or ifthe parties continue the alliance in the subsequent legislature (ie sit as asingle party group)

iv) Fission Fission is a collective affiliation strategy in which a minoritygroup of politicians breaks away from an established party to form a newparty We also classify as fission cases where a politician breaks away from anestablished party to run as an independent3 or decides to join a fission partybefore the latter has participated in its first election

v) Starting Up Starting up differs from the other strategies since it involvespreviously unaffiliated individuals creating a brand-new party and does notinvolve mobility between an existing exit or entry party Start-up parties arehence distinguished from fusion and fission parties in that they lack any par-liamentary representation at the moment of their creation and have not previ-ously run in any parliamentary election They are often what Duverger (1951)labeled ldquoexternally createdrdquo parties that emerge from social movements orinterest groups and that differ from internally created largely memberless par-ties established by already elected politicians Although such parties may overtime attract individual incumbent deputies or other active politicians to theirranks such reaffiliations occur rarely as start-up parties are usually reluctantto compromise their novel status by welcoming incumbent politicians Suchre-affiliations are therefore considered start-up only if they follow the creationof a start-up party and are inconsequential to the formation of the party If suchreaffiliations are all from a single existing party then they are classified asfission

Electoral Affiliation Strategies

A second more extensively studied determinant of party system institutional-ization concerns change in the electoral affiliations of voters There are manyways to analyze these different patterns Studies on electoral volatility forexample measure the magnitude of electoral vote switching for an entire partysystem and thus provide a useful insight into the aggregate stability of voteraffiliations Election studies in turn further differentiate these patterns bymapping out the ideological trajectories of voters and their changing partypreferences (Kitschelt et al 1999 Tucker 2002 Whitefield 2002) Yet whiledetailed knowledge about votersrsquo partisan alignments is key for understandingparty system institutionalization it is frequently the case that the relevant poll-ing data is unavailable Given this problem we can classify votersrsquo electoralaffiliations according to seven basic choices available to them i) they can re-vote for the same party they supported in the previous election ii) they canswitch to another already existing party iii-iv-v) they can opt for either a fu-sion fission or start-up party vi) they can spoil their ballot or vii) they canabstain For our purposes it is important to note that the last two strategiesinvolve a form of political non-involvement since they signal either protest orpolitical alienation Although such non-involvement is clearly important forassessing democratic consolidation it does not directly shape the develop-ment of party systems As a result we exclude them from this analysis In-


stead we concentrate on the first five strategies that is the frequency of vot-ersrsquo affiliation changes and their preferences among for different party originsWe can thus observe the extent to which voters durably align themselves (bystaying with their current choice) realign themselves (by switching to anotherexisting party) or de-align themselves (by opting for a fusion fission or start-up party)

This interest in the continuity of votersrsquo electoral alignments raises the ques-tion of how best to differentiate between new and old parties Following theexample of established party systems new parties could for example be de-fined as having contested three or less elections However benchmarking lsquonewrdquoin post-communist democracies is more complicated since arguably every-thing is new For example defining ldquonewrdquo based on a comparison with inter-war parties is largely useless as such parties have played only a marginal rolesince the reemergence of electoral competition in the region Rather in theBaltic cases a much better baseline involves the reestablishment of indepen-dence in August 1991 for this served also as a clear watershed in the develop-ment of parties That is to say almost all of those parties which predateindependence in August 1991 were actually established at least seven monthsprior to that date In Estonia fifteen out of sixteen pre-independence (or estab-lished) parties were formed in 1990 or earlier in Latvia seven out of nine andin Lithuania eleven out of thirteen Moreover very few parties were formedbetween independence and the first post-independence election (none inLithuania one in Latvia and two in Estonia) In short there is a clear andsignificant temporal distance between established pre-independence partiesand new post-independence parties that justifies the selection of independenceas benchmark for classifying parties4 As a result we will classify all partiescreated before the reestablishment of Estonian Latvian and Lithuanian inde-pendence in August 1991 as established parties and all those established afterindependence as new parties Moreover as will be seen below the new post-independence parties closely overlap with parties that emerged through fu-sions fissions and start-ups Out of the three countriesrsquo forty-five fusion fissionand start-up parties only three parties were formed before independence Thusdiscussion of post-independence parties is tantamount to talking about fissionfusion or start-up parties

To make our typology broadly applicable it is worthwhile also taking aquick look at the usefulness of employing national independence as a bench-mark in other post-communist democracies This criterion becomes problem-atic outside the Soviet and Yugoslav successor states for two reasons First forCentral and East European countries the collapse of the Soviet Union did notproduce a clear-cut independence date but an expansion of national sover-eignty which oftentimes is difficult to date precisely Second the differenttypes of communist regimes in these countries often produced very differenttransition modes and durations (Kitschelt et al 1999) As a result it becomesdifficult to find transition points that are as clear-cut as in the Baltics We there-fore suggest using founding elections as an alternative benchmark for distin-guishing ldquoestablishedrdquo from ldquonewrdquo parties For example Reich defines foundingelections as the ldquofirst competitive multiparty elections occurring during a tran-


sition to democracy after (a) at least ten years of authoritarian rule and (b)following reforms that allow for the formation of multiple political parties in-dependent of the state and free from state repressionrdquo (2001 1239-40) Thiscriterion works well for national communist regimes (Poland Hungary) whichjust like the Baltics prompted resistance movements from which proto-partiesemerged prior to the founding election5 (Kitschelt et al 1999) In the case ofbureaucratic communism (eg East Germany and Czechoslovakia) and patri-monial communism (eg Rumania and Bulgaria) founding elections providea more ambiguous benchmark Bureaucratic and patrimonial communism neverfueled popular opposition movements rather they imploded virtually over-night As a result the lack of struggle against communism removed a keyimpetus for the formation of parties prior to founding elections (Kitschelt et al1999) These two types of communism thus come close to what Elster Offeand Preuss call ldquoactor-less transitionsrdquo or transitions in which parties existingprior to the founding election were so rudimentary that it does not make senseto consider them established parties (Elster Offe and Preuss 1998)

Patterns of Party System Transformation

The structure of a party system is therefore ultimately shaped by the interac-tion of both politiciansrsquo organizational affiliations and votersrsquo electoral choicesIn Table 1 we summarize these resulting patterns However rather than gener-ating a full five-by-five matrix we collapse voter options into a single ldquoprefer-ence changerdquo row in order to make the table more manageable6

A preliminary glance at Table 1 illustrates that of the ten possible transfor-mation patterns several can coexist at the same time For example during the1990s Italian politicians both switched from one existing party to another anddefected to new start-up parties while voters shifted their allegiances to otherexisting parties and to new start-up parties The multiple choices available tovoters and especially to politicians thus means that the dynamic of party sys-tem transformation can but does not have to be driven by multiple dynamicsThe prominence or even exclusivity of one pattern over another is thus contin-gent on the choices of politicians and voters Further on we present someempirical evidence about these choices For now however we will describethese transformation patterns and then discuss the ways in which they consoli-date or deconsolidate party systems The ten permutations can be grouped intothree general patternsmdashalignment realignment or de-alignmentmdashdependingon the consolidating effects of each

Alignment patterns Cells I a amp b represent the preeminent consolidatedparty system In such systems neither voters nor political actors change theirpreferences or if they do it is very slowly or only in small numbers Suchsystems have also been called ldquofrozenrdquo or ldquoover-institutionalizedrdquo This lackof change has been attributed to deeply rooted social cleavages tightly orga-nized mass mobilization or extensive patronage networks all of which im-pede changes in voter and politician affiliation (Lipset and Rokkan 1967Schedler 1995) In transitional democracies the standard for alignment is some-what less strict than in established democracies For these systems it makes


sense to define alignment in terms of a growing number of politicians stayingput along with an increasing number of voters developing stable partisan af-filiations

Realignment patterns Realignment patterns represent either consolidatedparty systems undergoing some restructuring or new proto-party systemsmoving towards consolidation In Table 1 patterns II a-e reflect dynamics inwhich voters change their partisan preferences and political actors switch partyaffiliations For example cell IIa encapsulates the classical realignment of South-ern Democrats in the United States to the Republican Party during the postwarera While this realignment was aided by the occasional reaffiliation of Demo-cratic congressmen it was overwhelmingly driven by voters changing theirpartisan affiliations Realignment patterns II b amp d meanwhile are relativelyuncommon since it is rare for incumbents to either switch or merge withoutsome voters following them at the ballot box Still factional switching withinthe Japanese Liberal Democratic Party could constitute an example for patternIIb while the fusion of small fringe parties demonstrates pattern IId PatternsIIc and IIe are more common The former can be found in transitional democ-racies (such as Brazil) in established democracies with undisciplined parties(such as French Third or Fourth Republics) or in established democraciesexperiencing a political crisis (such as Italy in the 1990s) The degree to whichsuch party switching consolidates a party system depends on whether switch-ers hop from smaller to larger parties and on whether they remain affiliated orcontinue to switch to whichever party has the best winning prospect (Desposato2000 Mainwaring 1999 Mershon and Heller 2001 Shabad and Slomczynski2001) Finally pattern IIe plays an important role in consolidating party sys-tems In postwar West Germany for instance the assimilation of various con-servative parties by the Christian Democratic Union and the votersrsquo endorsementof this move in subsequent elections together contributed centrally to the con-solidation of that countryrsquos party system

Table 1Patterns of Party System Transformation

Incumbentsrsquo Organizational Affiliations

Staying Put Switching Fusion Fission Start-up

Ia Alignmt IIb Legis IId Legis IIIa Legisl Ib Alignmt No Realignmt De-alignmt De-alignmt

Votersrsquo Preference (Frozen WElectoral Change EuropeChoices patronage (LDP faction

parties) switching) (Weimar)

IIa Elect IIc Elect amp IIe Elect amp IIIb Elect IIIc ElectRealignmt Legis Legis amp Legis amp Legis

Preference Realignmt De-alignmt De-alignmt De-alignmt Change

(Unfrozen (Fr III Rep (Germany (Japan (WeimarW Europe Brazil Italy 1950s 1990s CND amp ItalyLithuania) 1990s) Estonia) Baltics) 1990s Baltics)


De-alignment patterns The de-alignment patterns displayed in cells III a-crepresent the de-consolidation of established party systems or the continuedlack of consolidation of transitional party systems The instability of thesepatterns results from the fact that voters andor politicians change their prefer-ences but they do so by opting for new rather than existing parties Fissionsconstitute one source of new parties In established democracies these are ad-mittedly rare occurring once every three elections (ie 9 to 12 years) theyalso rarely prove electorally successful (Mair 1990 187) In transitional de-mocracies they occur more frequently but as far as we can tell from the Balticexamples their electoral appeal is still limited The most common source ofde-alignment comes from start-up parties Parties like Silvio Berlusconirsquos ForzaItalia Hitlerrsquos NSDAP or the Bloc Queacutebecois and the Reform Party in Canadaare all examples of groups that have fundamentally transformed theircountryrsquos political landscape Moreover it is important to underscore thatthis transformation often owes much more to the appeal or organizationalskills of new political entrants than to the reaffiliation of incumbents7 Follow-ing our earlier definition we will therefore count as start-up only those politi-cians who reaffiliate after a start-up party has been founded and who do so onan individual basis Generally speaking start-up reaffiliations are not verycommon since most start-up parties run against the political establishment andthus have more to lose by inviting incumbents into their ranks (Kreuzer 2001133-65)

Empirical Application

The typology outlined above provides a firm basis for analyzing the develop-ment of party systems in post-communist and other transitional democraciesBy combining the organizational affiliations of politicians with the electoralchoices of voters we are less biased towards detecting patterns of stability andhence better able to capture the wide array of transformation patterns found inmost transitional democracies From Table 2 moreover it becomes apparentwhy relying on conventional indicators of party system institutionalizationmdashie the effective number of electoral parties to indicate fragmentation or gen-eral electoral volatility to demonstrate the durability of voter affiliationsmdashareof limited use in the Baltics 8 By these two measures the three Baltic partysystems appear so atomized and in flux that they barely resemble anythinglike established democracies Their effective number of electoral parties is twoto three times higher than the average 33 parties found in advanced industri-alized democracies (Lijphart 1999 76-77) Their volatility levels in turn arebetween four and nine times higher than the average 84 percent reported forpost-war European democracies and two to three times higher than other EastEuropean democracies (Bartolini and Mair 1990 Mair 1997 182) These vola-tility and fragmentation figures underscore a degree of voter preference changethat is dramatic even by post-communist standards Overall then these con-ventional indicators leave us with the impression of a totally inchoate partysystem without a clear developmental pattern By now this should not surpriseus since these indicators (with the implicit assumption of parties as stable uni-


tary actors) only capture three (Ia IIa and IIId) of the ten possible transforma-tion patterns outlined in Table 1

The next two sections therefore inquire whether our alternative analyticalframework allows us to actually detect some patterns in this apparent fluxFirst we look at the frequency with which active politicians select among thedifferent available organizational affiliation strategies Second we look at thefrequency with which voters choose among the different types of politicalparties As indicated earlier we classify the different party types by their originrather than ideological orientation Thus the remainder of this article analyzesthe extent to which politicians coordinate their organizational choices and vot-ers coordinate their electoral choices9

Organizational Affiliation Patterns of Politicians

Many commentators of East European politics capture the lack of organiza-tional continuity among post-communist politicians by referring to them sim-ply as ldquopolitical touristsrdquo or ldquoinstitutional nomadsrdquo (Shabad and Slomczynski2001 2) To analyze such organizational affiliation patterns in a more sys-tematic matter however we base our analysis on the five organizationalstrategies listed in the top row of Table 1 Second we define our set ofldquopoliticiansrdquo as anyone who ran as a candidate in any two consecutiveparliamentary elections in the Baltic states during the period 1992-200210

Thus we compiled complete candidate lists for the ten elections that tookplace during this period11 Next we compared each pair of consecutive elec-tions and for each pair generated a list of repeat candidates12 Finally we com-pared the party affiliations of each repeat candidate for each election pair Anyshifts were classified according to our five affiliation strategies The results arereported in Table 3

Unquestionably the most distinct feature of Table 3 is the considerable or-ganizational disloyalty evident among Baltic politicians The latter would ap-pear to have very different affi liation calculi than their counterparts in

Table 2Baltic Party Systems

Fragmentation Volatility1

First Second Third Average By 2nd By 3rd

Election Election Election Election Election

Estonia 95 106 70 90 611 404

Latvia 62 101 76 80 577 742

Lithuania 40 73 56 56 409 729

1The possibility and extensive use of apparentements (ie electoral alliances) complicates the compu-tation of volatility in the Baltics We therefore disaggregated the vote shares of apparentements usingpost-election seat apportionment of apparentement members If an apparentement did not win seatsthen we used pre-election seat distribution


established democracies The fact that an average of only 571 percent of re-peat candidates stayed with the same political party between any pair of elec-tions suggests that the payoffs of continuous party affiliation are smaller inpost-communist than in established democracies Partiesrsquo lack of a brand nametheir limited logistical infrastructure and the generally high political price thatthe transition to a market economy has imposed on politicians are only someof the reasons why staying put with the same party is not necessarily the mostadvantageous affiliation choice13 Moreover the average of politicians stay-ing put has only marginally risen over time from 544 to 57 percent

Despite this low overall level of organizational loyalty however there arestill important cross-national variations suggesting that country-specific fac-tors interact with more general post-communist environments in shaping partyaffiliation These variations differ by degree but they will become more pro-nounced in the next section where we look at votersrsquo electoral choices Fornow we see that Lithuania demonstrated the least organizational disloyalty asfully 69 percent of its repeat candidates chose on average to stay put betweenelections When Lithuanian politicians did switch their affiliation they gener-ally adhered to a risk-adverse realignment dynamic as 535 percent of allreaffiliating politicians favored party switching to the other reaffiliation alter-natives The shake-up of existing parties through fusions or fissions played afar smaller role in Lithuania than in the other Baltic countries accounting foronly 328 percent of reaffiliations Finally defections to new start-up partiesplayed a modest role in Lithuania involving just 58 percent of repeat candi-dates Overall then the affiliation choices of Lithuanian politicians were byno means random but contributed to an elite-driven realignment of the partysystem The greater organizational loyalty of Lithuanian politicians also meant

Table 3Organizational Affiliation Patterns of Baltic Politicians

Staying Put Switching Fusion Fission Start-Up

Estoniabw 1st-2nd election 252 109 398 212 29bw 2nd-3rd election 570 191 203 26 09 Average 411 150 300 119 19

Latviabw 1st-2nd election 672 157 80 42 49bw 2nd-3rd election 470 143 250 07 130bw 3rd-4th election 696 88 58 55 102 Average 613 129 129 34 94

Lithuaniabw 1st-2nd election 707 193 31 19 81bw 2nd-3rd election 673 139 25 128 35 Average 689 166 27 74 58

Baltic average 571 149 153 76 57


that established parties were relatively stable and provided the most preferrednew home for defecting politicians Lithuanian politicians thus realigned theparty system by most frequently alternating their allegiances among existingparties Organizational choices like fission and start-up which would havecontributed to party system de-alignment were of secondary importance al-though not entirely irrelevant

Estoniarsquos party system also experienced a significant realignment eventhough it was more indirect (via fusions) and hence more disruptive thanLithuaniarsquos realignment Established parties in Estonia clearly had great diffi-culty securing the organizational loyalty of their electoral candidates The lowaverage of Estonian politicians who stayed putmdashjust 411 percentmdash indicateseither a lack of party discipline or the presence of unpopular established politi-cal parties Yet despite this large number of defectors the Estonian party sys-tem did not disintegrate since 255 percent of these defectors switched toother established parties and 509 percent were part of new fusion partiesEstoniarsquos party system was thus reshaped through switching-induced directrealignment and through fusion-driven party mergers Interparty mobility flowedfrom established parties to other established parties as well as from the organi-zational reconfiguration of established parties De-alignment played a rela-tively modest role with a paltry 19 percent joining start-up parties and 119percent breaking away to set up their own party14

Latviarsquos party system arguably is characterized by the most complex dy-namic which combines elements from its two neighbors Just as in Lithuaniaa high percentage of politicians (613 percent) stayed put thus providing par-ties with considerable organizational stability The reaffiliation choices of de-fectors however were more evenly distributed than in Estonia or LithuaniaLatviarsquos party system underwent a clear realignment in that 129 percent ofdeputies switched between existing parties and another 129 percent were in-volved in the reorganization of established parties Likewise there was a no-ticeable de-alignment streak in that an average of 94 percent of all repeatcandidates opted for start-up parties thus displaying a greater willingness totake political risks

Overall this look at interparty mobility clearly demonstrates that the no-madic drive of Baltic politicians follows distinct itineraries rather than beingaimless and random Realignment clearly outweighs de-alignment even thoughthe levels of realignment and its types vary somewhat across the three coun-tries Latvia and especially Estonia stand out for the importance of fusion inrealigning party systems while party switching occurs at similar levels in allthree countries With respect to de-alignment the consistently high rate ofstartups in Latvia is the only striking feature Baltic politicians thus select con-solidating re-affiliation strategies (ie fusion and switching) far more frequentlythan deconsolidating ones (ie fission and startups) These transformationpatterns resulting from intraparty mobility could be refined if we were to dis-aggregate affiliation choices by incumbency ideological trajectories or elec-toral payoffs but current space constraints make this impossible We thereforeturn our attention to votersrsquo electoral choices which constitute the second andmore commonly studied determinant of party system transformation


Electoral Affiliation Patterns of Voters

While politiciansrsquo organizational affiliations structure the choices available tovoters it is ultimately the voters and their electoral decisions which structureand hence transform party systems Consequently we examine now how vot-ers responded during 1992-2002 to politiciansrsquo organizational affiliation strat-egies We do so by clustering the parties and their vote totals according to theirorigins and then tracking their electoral performance Our party sample is basedon detailed political histories of thirty-four Estonian parties thirty-one Latvianparties and twenty-four Lithuanian parties We gathered information on eachparty including its founding date the circumstances of its creation and itsparticipation in electoral alliances mergers fissions and dissolutions Assources we consulted political histories local newspapers and individual coun-try specialists15 In selecting our eighty-nine parties we used parliamentaryrepresentation as the key criterion for inclusion in our dataset All eight-nineparties had to control at one point or another at least one parliamentary seatUnder this criterion three types of parties qualified

i) parties having won at least one seat through electionsii) parties emerging from breakaway legislative factions andiii) parties whose successor or predecessor (ie through a fusion or a fission) won at

least one seat

Analyzing the electoral performance of parties according to their organiza-tional origins not only echoes but also accentuates some of the transforma-tional patterns evident among politiciansrsquo affiliation patterns Below we providea brief synopsis of these patterns before using individual country profiles toelaborate them in more detail

The realignment dynamic among Lithuanian politiciansrsquo seen in the previ-ous section is closely mirrored in the electoral arena During the countryrsquos firstthree post-independence elections voters opted most frequently for establishedparties thus rewarding the staying put and party switching reaffiliation strate-gies of political actors Voters however also offset their preference for statusquo parties by massively shifting their support among them Their reticence tosupport new parties was much more pronounced for fission and start-up par-ties than it was for fusion parties Lithuaniarsquos party system thus functionedmost closely to that of an established democracy in that realignment took placeamong established albeit still very numerous parties In Estonia electoralchoices mirrored those of Lithuania in that the realignment dynamic was cuedby politiciansrsquo organizational reaffiliations For example Estonian voters simi-larly shifted their preferences among existing parties but preferred fusion par-ties just as frequently as established parties They were more hesitant in optingfor newer fission or start-up parties which ultimately limited the de-alignmentof the party system Lastly in Latvia the balance between realignment and de-alignment dynamics that we observed among its politicians clearly tipped inthe electoral arena in favor of the latter Latvian voters cast their ballots mostfrequently for new parties thus providing far fewer rewards for staying putand switching than in Lithuania Among the new parties they clearly preferred


start-up parties over fission and fusion parties Latvian voters thus contributedto consistent de-alignment as start-up parties quickly replaced establishedpolitical parties and even some earlier start-up parties

The following three tables and descriptive accounts elaborate on these elec-toral transformation patterns more fully They also add a classificatory nuancethat is not covered by Table 1 by breaking down the established pre-indepen-dence parties into three additional categoriesmdashldquointerwarrdquo ldquocommunist suc-cessorrdquo and ldquofoundingrdquo parties This allows us to further differentiate howvoters affiliated themselves among incumbents who stayed put or switchedamong existing parties These three additional party types also reflect whatElster Offe and Preuss have called Eastern Europersquos ldquotriple pastrdquomdashthe inter-war communist and independence struggle eras (1998 69 Ishiyama 2001a)For example interwar parties usually readopted the name of their interwarprecursor and in some cases had some of the original leaders or their offspringplay an important leadership role Communist successor parties emerged fromthe more nationalist andor moderate factions within the old Communist Partyof the Soviet Union They were the ldquogoverning party in the communist regimewhich inherited the preponderance of the former ruling partyrsquos resources andpersonnelrdquo (Ishiyama 2001a 3) Finally founding parties already had a dis-tinct organizational existence during the pre-independence struggle even thoughmost of them participated in umbrella movements like the Baltic popular fronts

Lithuania The realignment pattern we observed in the interparty mobilityof politicians clearly repeats itself in the electoral arena (See Table 4)Lithuanian voters fostered party system realignment by switching their loyal-ties among existing pre-independence parties and largely ignoring newer par-ties formed after independence The countryrsquos realignment dynamic is mostclearly evident in the overwhelming vote shares controlled by pre-indepen-dence parties The latter won 887 percent of the votes in 1992 783 percent in1996 and 689 percent in 2000 Indeed these first two elections characterizebest Lithuaniarsquos realignment dynamic since in 1992 the original democraticopposition movement Sajudis (later renamed the Homeland Union-LithuanianConservatives) was ousted by the ex-communist Democratic Labor Party in astunning electoral comeback In 1996 the reverse happened What is morethe pivotal role played by these two parties can be seen by the fact that in 1996they accounted for 729 percent of the overall electoral volatility

By the third election in 2000 the Lithuanian party system showed somesigns of de-alignment even though the share of pre-independence parties re-mained at 689 percent An internal power struggle and resulting defectionsfrom the Homeland Union were one contributing factor Meanwhile the twomajor leftist parties (the Social Democratic Party and the Democratic LaborParty) along with two other marginal parties shored up their electoral appealby forming in 2000 an apparentement called the ldquoSocial Democratic AlliancerdquoThe appeal of newer post-independence parties remained however modestOnly one start-up party the New Union scored in the top five and fissionparties won no more than 6 percent This astonishing continuity of Lithuaniarsquospre-independence parties is also illustrated by the fact that over the course ofthese three elections they occupied thirteen out of fifteen possible top five


vote-winning spots Lithuanian voters thus seem to have been far more risk-averse than their Latvian and Estonian counterparts Whatever disenchant-ment they had during an election they translated into defections to otherpre-independence parties rather than opting for new untried parties The ma-jor benefactors (and victims) of these electoral realignments were the Home-land Union and the Democratic Labor Party Others like the Social DemocraticParty the Christian Democratic Party the Liberal Union and the Peasant Partyalso saw their electoral fortunes fluctuate

The overall prominence of Lithuaniarsquos pre-independence parties is reflectedin the strength of each pre-independence party type On a general levelLithuaniarsquos founding parties mirrored those of Estonia and Latvia in that fourout of seven succeeded in placing among the electoral top five during thecountryrsquos three post-independence elections Where Lithuania differed how-ever was in the prominence of communist successor and historical partiesLithuaniarsquos revived interwar partiesmdashthe Christian Democratic Party the Na-tionalist PartymdashldquoYoung Lithuaniardquo and the Social Democratic Partymdashall hadconsiderable success winning 142 percent and 220 percent of the votes inthe first two elections before dropping to 5 percent in the 2000 election TheSocial Democratic Party was the most successful left-wing revival party in the

Boxed party names indicate ranking among top five vote-getting parties Lithuania has a dual electoralsystem (Pettai and Kreuzer 1999) The percentages consequently report the average of the single- andmulti-member district election results

Only SMD results Formed an apparentement for MMD individual party vote shares not available

Table 4Lithuanian Party System


Baltics while the Democratic Labor Party was by far the most successful com-munist successor party

The flip side of Lithuaniarsquos realignment pattern was the marginality of itspost-independence parties Only on two occasions did such parties rank amongthe top five finishers Fissions were entirely inconsequential as no breakawayparty ever won more than three seats Only one start-up party the New Unionmanaged to attract a significant vote share Fusions were also inconsequentialand confined to small fringe parties such as the two competing Christian Demo-cratic parties and various right-wing splinter groups

Overall then Lithuaniarsquos fragmentation and volatility reflected the risingand falling fortunes of parties that had established themselves early and thatdrew on historical legacies going as far back as the interwar period or datingfrom as recently as the independence struggle These parties experienced anorganizational continuity unmatched in the Baltics even as their electoral sup-port proved highly volatile Lithuanian parties thus differed from their Esto-nian and Latvian counterparts who instead experienced both organizationaldiscontinuity and electoral volatility

Latvia Latvian voters behaved in a much more straightforward manner thantheir politicians Whereas the politicians had shown only a slight trend towardsde-alignment the votersrsquo primary transformation pattern was de-alignment asestablished pre-independence parties lost votes to newer post-independenceparties Indeed what makes this de-alignment process particularly striking isthe fact that it was driven by start-up parties whose electoral success led to thereplacement of older pre-independence parties in the electoral market place

The rapidly declining vote share of pre-independence parties as demon-strated in Table 5 provides the most compelling evidence for Latviarsquos rapidde-alignment These parties won 498 percent of the vote in 1993 366 per-cent in 1995 and just 16 percent in 1998 Thus in only five years the partiesthat had roots in either the interwar communist or independence periods lostnearly all their electoral support Indeed in 2002 they fell even further downto just 134 percent Moreover Table 5 illustrates another aspect of de-align-ment It lists the top five vote-getting parties for each election and thus allowsus to count how many pre-independence parties ended up holding such posi-tions In Latvia only six pre-independence parties held one out of the twentypossible positions this is far lower than in Estonia and Lithuania Moreoverthe electoral fortunes of post-independence parties seem as volatile and short-lived as that of the pre-independence parties Table 5 underscores this fact byshowing that ten different post-independence parties or coalitions rotated inand out of the sixteen top five spots that they occupied Only one party LatviarsquosWay scored among the top five in at least three elections but by 2002 it toofell below the electoral threshold In addition a number of Latviarsquos post-inde-pendence parties proved short-lived organizationally At least six parties (in-cluding some that garnered as much as 15 percent of the vote) participated inonly one or two elections before fading into oblivion This rapid de-alignmentprocess thus precipitated a lack of continuity in partiesrsquo electoral strength andorganizational existence which together accounts for much of Latviarsquos highfragmentation and volatility


The relative insignificance of pre-independence parties is the flip side ofLatviarsquos de-alignment Of the ten pre-independence parties listed in Table 5only fivemdashthe Farmers Union For Fatherland and Freedom the National In-dependence Movement the Social Democratic Workersrsquo Party and the Social-ist Partymdashever held a top five spot Their limited appeal was further underscoredby the fact that most of these parties ranked in the top five only once More-

Table 5Latvian Party System

Boxed party names indicate ranking among top five vote-getting partiesPart of apparentement led by Harmony Party individual party results not available Part of apparentement led by SDWP individual party results not availablePart of apparentement ldquoFHRULrdquo led by Harmony Party individual party results not availablePart of apparentement led by Farmers Union individual party results not available


over in contrast to Estonia and Lithuania founding parties in Latvia wereespecially weak Among the countryrsquos five founding parties only For Father-land and Freedom and the National Independence Movement scored in the topfive and by the third election only the Greens survived as independent politi-cal organization All the rest fused into new parties some of which in turndisappeared entirely Likewise historical parties played a modest role in Latviawith only the Farmers Union the Social Democratic Workerrsquos Party and theDemocratic Center Party winning a high of 17 percent in 1993 and a low of 3percent in 1995 Latviarsquos two communist successor partiesmdashthe DemocraticLabor Party and the Socialist Partymdashwere far less successful than LithuaniarsquosDemocratic Labor Party but more substantial than Estoniarsquos namesake partyStill their continued existence relied heavily on apparentements with otherstronger parties

The principle reason behind Latviarsquos de-alignment stemmed from a highnumber of start-up parties and to a lesser extent fusion parties Latviarsquos eightstart-up parties won on average 12 percent of the votes and replaced in rapidsuccession most of the pre-independence parties and eventually also somepost-independence ones Latviarsquos Way was the most successful start-up partywinning an average of 217 percent of the votes for its first three electionsbefore slipping to just 49 percent in 2002 During 1998 and 2002 a numberof single-leader start-up parties emerged such as the Peoplersquos Party (led by theformer prime minister Andris Skele) and New Era (founded by the formercentral bank president Einars Repse) Both scored over 20 percent of the voteduring their debut election and the Peoplersquos Party continued strong in 2002Among other start-ups success was more ephemeral but it still demonstratedwhat one might even call an eagerness among Latvian voters to choose new-comers to the party system The Unity Party the New Party and the First Partyall had respectable debuts on the electoral scene in 1995 1998 and 2002respectively Compared to Lithuania and Estonia the considerable success ofLatvian start-up parties suggests that Latvian voters were greater risk-takersand that the Latvian political market had fewer entry barriers Fusion partiesalso had some success in Latvia attracting an average of 102 percent of thevotes Three such partiesmdashthe Democratic Party-Saimnieks For Fatherlandand FreedomLNNK and the First Partymdashwon enough seats to score in topfive positions Only one fission party For Latvia played a significant albeitshort-lived role

In sum Latviarsquos high fragmentation and volatility did not reflect the risingand falling fortunes of established parties (as it did in Lithuania) but rather itfollowed Estoniarsquos example of a wholesale party system transformation Latviahowever differed from Estonia because the transformation was driven far morecentrally by the entry of start-up parties rather than by the organizationalreconfiguration of existing parties through fissions and especially fusions

Estonia The electoral dynamic of Estoniarsquos party system closely followedthat of its politiciansrsquo interparty mobility Established parties lost support butnot nearly as rapidly as in Latvia Merger parties fared better in Estonia than inany other Baltic country thus reinforcing the organizational realignment of itspoliticians Voters however opted for fusion parties although not with the


same high frequency as politicians chose to form them Instead they were justas likely to vote for fission or start-up parties The development of the Esto-nian party system was thus characterized by a hybrid dynamic combining theelements of the Latvian and Lithuanian ones As Table 6 demonstrates Estoniarsquospre-independence parties did not assume the same electoral strength or elec-toral swings as in Lithuania and their vote share also did not drop as rapidly asin Latvia Estoniarsquos pre-independence parties won 484 percent of the seats in1992 37 percent in 1995 and 309 percent in 1999 and they occupied sevenout of the fifteen possible top five spots Pre-independence parties in Estoniawere thus able to maintain their electoral strength longer than their Latviancounterparts and there was also far less volatility than in Lithuania The sup-port of the two most prominent pre-independence parties the Center and Coa-lition Parties remained steady enough in all three elections to land them in thetop five vote-getting spots (although by 2002 the Coalition Party had beendissolved) Meanwhile the average volatility for pre-independence parties was156 percent in Estonia compared to 39 percent in Lithuania Estonia thus didnot experience the same swing-like realignments among pre-independenceparties observable in Lithuania Furthermore Estoniarsquos established parties weredominated by founding parties which won 404 percent of the vote in 1992284 percent in 1995 and 309 percent in 1999 As Table 6 shows six found-ing parties managed to win votes in 1992 and by 1995 their electoral strengthhad largely concentrated on the Center and Coalition parties The prominenceof founding parties in Estonia rivals that in Lithuania while the other types ofpre-independence parties played an insignificant role Estoniarsquos only interwarparty the Rural Union was an important partner in apparentements with theCoalition Party in 1992 and 1995 only to merge however in 1999 with theCountry Peoplersquos Party to form the Peoplersquos Union Meanwhile the successorto the Estonian Communist Party the Democratic Labor Party remained mar-ginal failing to win any seats in the first two elections before finally winningtwo spots in 1999 via inclusion in the list of the United Peoplersquos Party

The most important transformation of Estoniarsquos party system occurred throughthe sort of indirect organizational realignments we observed among the countryrsquospoliticians Whereas Lithuanian voters swung their allegiances back and forthbetween established parties Estonian voters followed politicians into six merg-ers which subsumed a total of seventeen parties and won on average 256percent of the vote These mergers took place in two waves The first occurredin the run-up to the 1992 election with the Moderates bringing together theRural Center Party and the Social Democrats while Pro Patria subsumed fiveproto-parties (the Christian Democratic Party the Christian Democratic Unionthe Conservative Peoplersquos Party the Liberal Democratic Party and the Repub-lican Coalition Party) A second wave of mergers took place after the 1999election when on the one hand the Moderates merged with the small PeoplesrsquoParty and on the other hand the Country Peoplesrsquo Party joined with the RuralUnion and Pensioners and Families Party to form the Peoplersquos Union Thereplacement of established parties by new start-up parties thus did not play animportant role in Estonia As Table 6 illustrates start-up parties assumed a farless prominent place in Estonia than in Latvia They won an average of only


12 percent of the votes and only one such party the Country Peoplersquos Partyever scored a top five spot The four fission parties in turn controlled onaverage 84 percent of the votes The Reform Party (after its creation fromsections of Pro Patria and the Moderates in 1994) was the only Estonian fissionparty to gain any significant strength winning 162 percent and 159 percentof the votes in the 1995 and 1999 elections

Overall Estoniarsquos high fragmentation and volatility reflected a complex partysystem dynamic It was neither an already set party system trying toreequilibrate itself nor was it a party system turning itself inside out Thestrength of its founding parties provided it with some continuity and vot-ers endorsed the various party mergers thus contributing most of all to a re-alignment pattern De-alignment still played a role but it was not nearly asprominent as in Latvia

Table 6Estonian Party System

Boxed party names and vote percentages indicate ranking among top five vote-gettersThe entire RU-CP apparentement won 136The entire RU-CP-CPP apparentement won 322The entire UPP-RP apparentement won 59The entire UPP-DLP apparentement won 63


Conclusion

This comparison of top-down party system transformation via politician-ledinterparty mobility and bottom-up party system change via voter-induced elec-toral shifts has demonstrated that distinct patterns of party system formationexist in post-communist countries despite conditions of highly unstable partyidentification as well as organizational disloyalty among politicians The analy-sis also showed a considerable degree of congruence between shifts in thesetwo levels of party system evolution Political instability is thus characterizedby regularities just as is political stability If we go beyond looking at post-communist party systems merely through concepts like fragmentation or vola-tility (which are only useful for detecting the absence of stability without tellingus what shape instability takes) we can indeed detect patterns of instability Inthis article we have tried to show that the key for detecting such patterns ofchange and transformation is to incorporate the organizational preferencechanges of politicians These constitute the central political dimensions towhich observers of post-communist party systems have frequently pointedbut which so far have not been systematically investigated Ultimately detect-ing these patterns of change is only the first and easiest step towards explain-ing the formation of party systems since the interactions between politicalactorsrsquo organizational choices and votersrsquo electoral choices constitute onlyproximate causes or underlying processes of party system transformation Theydo not provide actual explanations for why one pattern occurs rather thananother Such explanations require analysis of various distal causes (such asinstitutions historical legacies transition modes) to see how these might haveconstrained the choices of both politicians and individual voters which in turnhave produced different aggregate transformation patterns This represents thenext step in understanding the institutionalization of post-communist partysystems

Notes

We would like to thank for Artis Pabriks and Darius Zeruolis for sharing their knowledge ofLatvian and Lithuanian party politics as well as John T Ishiyama Scott Desposato and twoanonymous SCID reviewers for commenting on an earlier draft Funding for this research camefrom an Estonian Science Foundation grant nr 4904 We gratefully acknowledge their support

1 The one exception we came across is mentioned in Shabad and Slomczynski (2001)2 For a first attempt in this direction see Bielasiak (1997)3 These cases pertain only to Estonia and Lithuania where such independents are permitted4 In a similar vein Bielasiak (1997 33-37) reports that in East Central Europe party politics during

the closing days of the communist regimes was quite distinct from party politics in the opening ofdays of democracy

5 For Poland we would suggest the 1991 election as the first genuine post-Soviet era election Whilethe 1989 election was free the organization of parties was still quite circumscribed making itcomparable to the Balticsrsquo 1990 Supreme Soviet elections In both of these cases communistincumbents competed with single broad opposition movements rather than with individual parties

6 This also acknowledges the interdependent definition of politiciansrsquo and votersrsquo affiliation choicesSince realigning voters switch by definition among existing parties their electoral choices do notinteract with fusion fission and start-up parties Conversely since de-aligning voters opt bydefinition for new parties their choices do not interact with politicians who stay put or switch


7 Table 1 poorly reflects this fact Its last column defines a start-up induced de-a1ignment as beingprimarily the result of incumbentsrsquo reaffiliation and changing voter preferences However this lastcolumn should also be read as including de-alignment resulting from the entry of new politicalactors forming a start-up party

8 For the computational details of these two measures see Taagepera and Shugart (1989) andBartolini and Mair (1990) respectively

9 For now we do not address the more complicated issue of a possible interaction between politiciansrsquoand votersrsquo affiliation choices For example to what extent do voters and politicians coordinate theirchoices with each other Do voters follow the organizational choices of politicians or do politiciansrespond to changing electoral circumstances

10 In this respect we include not just incumbent candidates (politicians par excellence) but also non-incumbents (politicians aspiring to office)

11 In Estonia 1992 1995 1999 in Latvia 1993 1995 1998 2002 in Lithuania 1992 1996 2000The total number of candidates across these elections was over 9000

12 Each of the resulting seven lists yielded an average sample of around 300 repeat candidates It isinteresting to note that for any pair of elections this figure represented around 32 of all candidateswho ran in the first election of the pair meaning that after any election roughly a third of candidatestried their luck again

13 The political calculus underlying these affiliation choices is explored further in Kreuzer and Pettai(2002)

14 Moreover preliminary analysis of the candidate lists for Estoniarsquos 2003 parliamentary electionsshows even further stabilization of politiciansrsquo organizational affiliations Fully 80 of the 402repeat candidates in 2003 remained loyal to their 1999 party Another 15 switched merely to otherestablished parties

15 Our principle source was the Baltic News Service together with the following country-specificsources For Estonia (Arter 1996 Grofman et al 2000 Ishiyama 1993 Pettai and Kreuzer 1999)For Latvia (Bungs 1993a 1993b 1993c 1993d Ishiyama 1993 Gobins and Kerner 1997 Pettaiand Kreuzer 1999 Plakans 1997) For Lithuania (Girnius 1992a 1992b 1993 Pettai and Kreuzer1999 Krickus 1997 Krupavicius 1998 Veser 1995 Clark 1995)

References

Arter David 1996 Parties and Democracy in the Post-Soviet Republics The Case of Estonia BrookfieldAldershot

Bartolini Stefano and Peter Mair 1990 Identity Competition and Electoral Availability CambridgeCambridge University Press

Bielasiak Jack 1997 ldquoSubstance and Process in the Development of Party Systems in East CentralEuroperdquo Communist and Post-Communist Studies 30 1 23-44

Bungs Dzintra 1993 ldquoThe Shifting Political Landscape in Latviardquo RFE-RL Research Report 2 12 28mdashmdashmdash 1993 ldquoTwenty-three Groups Vie for Seats in the Latvian Parliamentrdquo RFE-RL Research Report

2 23 44mdashmdashmdash 1993 ldquoModerates win parliamentary elections in Latviardquo RFE-RL Research Report 2 28 1-7mdashmdashmdash 1993 ldquoElections and Restoring Democracy in the Baltic Statesrdquo RFE-RL Research Report 2

38 12-17Clark Terry 1995 ldquoThe Lithuanian Party Systemrdquo East European Politics amp Societies 9 41-62Desposato Scott 2000 ldquoInstitutional vs Societal Explanations of Party Development A Comparative

Analysis of Brazilrsquos State Legislaturesrdquo Paper presented at the Midwestern Political ScienceAssociation Meeting Chicago (April)

Duverger Maurice 1951 Les Partis Politiques Paris Armand ColinElster Jon Claus Offe and Ulrich Klaus Preuss 1998 Institutional Design in Post-Communist Soci-

eties Rebuilding the Ship at Sea Theories of Institutional Design Cambridge Cambridge Uni-versity Press

Geddes Barbara 1995 ldquoA Comparative Perspective on the Leninist Legacy in Eastern Europerdquo Com-parative Political Studies 28 2 239-274

Girnius Saulius 1992a ldquoLithuaniarsquos Sajudis Declines but Parties Remain Weakrdquo RFE-RL ResearchReport 1 15 8


mdashmdashmdash 1992b ldquoThe Parliamentary Elections in Lithuaniardquo RFE-RL Research Report 1 48 6mdashmdashmdash 1993 ldquoLithuanian politics seven months after the electionsrdquo RFE-RL Research Report 16Gobins Marcis and Manfred Kerner 1997 ldquoPolitische Parteien in Lettland Personlichkeiten Programme

Perspektivenrdquo Osteuropa 47 2 139-149Grofman Bernard Evald Mikkel and Rein Taagepera 2000 ldquoFission and Fusion of Parties in Estonia

1987-1999rdquo Journal of Baltic Studies 31 4 329-357Ishiyama John 1993 ldquoFounding Elections and the Development of Transitional Parties The Cases of

Estonia and Latvia 1990-1992rdquo Communist and Post-Communist Studies 26 3 277-99________ 2001 ldquoSickles into Roses Successor Parties and Democratic Consolidation in Post-commu-

nist Politicsrdquo Pp 32-54 in Party Development and Democratic Change in Postcommunist Eu-rope The First Decade ed P G Lewis London Frank Cass

Kitschelt Herbert Zdenka Mansfeldova Radoslaw Markowski and Gabor Toka 1999 Post-commu-nist Party Systems Competition Representation and Inter-Party Cooperation Cambridge Cam-bridge University Press

Kreuzer Marcus 2001 Institutions and Innovation Voters Parties and Interest Groups in the Con-solidation of Democracy France and Germany 1870-1939 Ann Arbor University of MichiganPress

Kreuzer Marcus and Vello Pettai 2002 ldquoThe Calculus of Party Affiliation in Post-communist Democ-racies Party Switching Fusions Fissions and the Institutionalization of Party Systemsrdquo Paperpresented at the Annual Meeting of the American Political Science Association August 29-Sep-tember 1 Boston

Krickus Richard 1997 ldquoDemocratization in Lithuaniardquo Pp 290-334 in Consolidation of Democracy inEast-Central Europe eds K Dawisha and B Parrot Cambridge Cambridge University Press

Krupavicius Algis 1998 ldquoThe Post-Communist Transition and Institutionalization of Lithuaniarsquos Par-tiesrdquo Political Studies Annual 46 3 465-91

Lewis Paul 2000 Political Parties in Post-Communist Eastern Europe London RoutledgeLijphart Arend 1999 Patterns of Democracy Government Forms and Performance in Thirty-Six

Countries New Haven Yale University PressLipset Seymour M and Stein Rokkan 1967 ldquoCleavage Structures Party Systems and Voter Align-

ment An Introductionrdquo Pp 1-64 in Party Systems and Voter Alignments eds S M Lipset and SRokkan New York Free Press

Mainwaring Scott 1999 Rethinking Party Systems in the Third Wave of Democratization The Case ofBrazil Stanford Stanford University Press

Mair Peter 1990 ldquoThe Electoral Payoffs of Fission and Fusionrdquo British Journal of Political Science 201 131-141

mdashmdashmdash 1997 Party System Change Approaches and Interpretations Oxford Clarendon PressMershon Carol and William B Heller 2001 ldquoParty Fluidity and Legislatorsrsquo Vote Choices The Italian

Chamber of Deputies 1996-2000rdquo Paper presented at Annual Meeting of the American PoliticalScience Association August 30-September 2 San Francisco

Miller Arthur H Klobucar Thomas F Reisinger William M and Hesli Vicki L 1998 ldquoSocialIdentities in Russia Ukraine and Lithuaniardquo Post-Soviet Affairs 14 3 248-87

Pettai Vello and Marcus Kreuzer 1999 ldquoParty Politics in the Baltic States Social Bases and InstitutionalContextrdquo East European Politics and Societies 13 1 148-90

Plakans Andrejs 1997 ldquoDemocratization and Political Participation in Post-communist Societies TheCase of Latviardquo Pp 245-89 in The Consolidation of Democracy in East Central Europe eds KDawisha and B Parrot Cambridge Cambridge University Press

Reich GM 2001 ldquoCoordinating Party Choice in Founding Elections Why Timing Mattersrdquo Com-parative Political Studies 34 10 1237-1263

Reisinger William M Arthur H Miller and Vickie L Hesli 1995 ldquoPublic Behavior and PoliticalChange in Post-Soviet Statesrdquo The Journal of Politics 57 4 941

Schedler Andreas 1995 Under- and Overinstitutionalization Some Ideal Typical Propositions Con-cerning New and Old Party Systems Working Paper 213 Kellogg Institute University of NotreDame

Shabad Goldie and Kazimierz Slomczynski 2001 ldquoInterparty Mobility among Political Elites in Post-Communist East Central Europerdquo Paper presented at the Annual Meeting of the American PoliticalScience Association August 28-September 2 San Francisco


Taagepera Rein and Matthew Shugart 1989 Seats and Votes New Haven Yale University PressToka Gabor 1998 ldquoParty Appeals and Voter Loyalty in New Democraciesrdquo Political Studies Annual 46

3 589-610Tucker Joshua A 2002 ldquoThe First Decade of Post-Communist Elections and Voting What Have We

Studied and How Have We Studied It rdquo Annual Review of Political Science 5 271ndash304Veser Reinhard 1995 ldquoPolitische Parteien in Litauenrdquo Osteuropa 45 10 936-945Whitefield Stephen 2002 ldquoPolitical Cleavages and Post-Communist Politicsrdquo Annual Review of Politi-

cal Science 5 181-200Zielinski Jakub 2002 ldquoTranslating Social Cleavages into Party Systems The Significance of New

Democraciesrdquo World Politics 54 2 184-211

Appendix 1Coding for Organizational Affiliation Patterns of Baltic Politicians

Categories Sub-Categories Explanation

Staying put Loyal Candidate remained loyal to a partythat remained free-standing

Loyal following Candidate remained loyal to one ofapparentement the parties which separated after anbreak-up apparentement from the first election

broke down by the secondLoyal as part of Candidate remained loyal to partyapparentement after it formed an apparentement in

the second election

Switching Switched to Candidate left original party andestablished party switched to an established free-

standing partySwitched to post- Candidate left original party andapparentement party switched to an established party

that had previously been a memberof an apparentement

Switched to Candidate left original party andapparentement switched to an apparentement made

up of parties different from originalone

Fusion Loyal as part of merger Candidate remained loyal as originalparty merged with another

Switched to merger Candidate left original party andswitched to a merged party

Fission Fission party Candidate was part of a fission partySwitched to fission Candidate left original party and

switched to a fission partySwitched to self Candidate left original party and

became a self-nominated candidate(Estonia and Lithuania only)

Start-up Start-up Candidate left original party andjoined a start-up party


preferences produces stable socio-economic cleavages and reasonably durablevoter alignments Another school of thought points to partiesrsquo organizationalweakness volatile electorates and politiciansrsquo continuous interparty mobilityto conclude that party system formation is a top-down process shaped prima-rily by historical contingencies and strategic action

These two schools of thoughts have produced research that differs consid-erably in its theoretical sophistication and methodological rigor The first struc-tural line of inquiry borrows extensively from literatures on partisanidentification path dependence and political sociology and engages in exten-sive cross-national survey research (Kitschelt et al 1999 Miller et al 1998Reisinger et al 1995 Toka 1998) By contrast the second political-researchapproach has produced numerous empirically rich but theoretically and meth-odologically underdeveloped studies Commonly these studies stress the or-ganizational flux of parties along with exorbitant levels of volatility andhyper-fragmentation and conclude that party systems do not yet exist (Elsteret al 1998 Geddes 1995 Grofman et al 2000 349 Lewis 2000 Mair 1997175) Our goal is to build a bridge between these two schools by pointing toand partly remedying the analytical and theoretical shortcomings plaguing thesecond political line of inquiry

The key for developing a more sophisticated political analysis of party sys-tem development is to place affiliation patterns of both voters and politicianson a firmer analytical and conceptual footing Studies emphasizing politicalfactors are surprisingly unanimous in considering votersrsquo lack of partisan iden-tification and politiciansrsquo shifting organizational loyalties as the key micro-obstacles to the macro-institutionalization of party systems Yet despite thistacit agreement there has been no systematic attempt to differentiate thesedifferent forms of electoral and organizational disloyalty analyze their inter-actions compare them across countries and ultimately provide a theoreticallygrounded explanation1 We will do so first by elaborating a typology of politi-ciansrsquo organizational and votersrsquo electoral affiliations second by exploringhow the combination of these two levels of affiliation produces distinct pat-terns of party system institutionalization and third by using this analyticalframework to compare the development of party systems in Estonia Latviaand Lithuania In doing so we hope to bring the same level of analysis to thepolitical underpinnings of party systems that has been achieved in examiningtheir structural causes And in creating a more equitable playing field we alsohope that it will become easier to assess the respective causal importance ofthese two dimensions as well as investigate the possible interactions existingbetween them2

Analytical Framework Electoral and Organizational Affiliation Strategies

Given that stable Western party systems constitute the primary point of depar-ture for analyzing post-communist party systems the lack of stable electoralpartisan identification and loyal organizational affiliations among politiciansin post-communist democracies has widely been read as evidence for the ab-sence of any meaningful party systems Mair for example points out that ldquothe



















































Estonia 95 106 70 90 611 404

Latvia 62 101 76 80 577 742

Lithuania 40 73 56 56 409 729




















least one seat






































Conclusion


Notes
















References
















































the second election






























































Estonia 95 106 70 90 611 404

Latvia 62 101 76 80 577 742

Lithuania 40 73 56 56 409 729




















least one seat






































Conclusion


Notes
















References
















































the second election






























least one seat






































Conclusion


Notes
















References
















































the second election















































Conclusion


Notes
















References
















































the second election






















References
















































the second election
















































the second election












Patterns of Political Instability: Affiliation Patterns of

Documents

Transcript of Patterns of Political Instability: Affiliation Patterns of