Throwing Out the Bums Protest Voting and Unorthodox Parties after Communism By Grigore Pop-Eleches* FOR a brief period in the second part of the 1990s, following the victories of broad democratic coalitions in Romania (1996), Bulgaria (1997), Moldova (1998), Slovakia (1998), and Croatia (2000), it looked as though postcommunist East Europe was moving decisively toward party system and democratic consolidation. The results seemed to indicate that East European voters had finally rejected their former rulers (tainted by communist pasts, nationalism, and corruption) in favor of politicians dedicated to pursuing faster economic and political reforms in line with the long-term goal of Western integration.1 However, the crushing defeat of the center and the alarmingly strong showing of the extremist Greater Romania Party (prm) in the Romanian elections of November 2000 marked the onset of a reverse wave in the electoral choices of postcommunist voters who now endorsed unorthodox parties (uops) of various stripes at the expense of their mainstream competitors. This wave swept through not only the more fragile democracies of southeastern Europe (Romania in 2000 and Bulgaria and Moldova in 2001), but also affected relatively more consolidated democracies such as the Czech Republic, Poland, Slovakia, and the Baltic states. These electoral trends raise a number of important theoretical questions about the political dynamics of postcommunist transitions and democratic transitions more broadly. Why did the consistent power alternation of the mid-1990s not result in party-system consolidation, as suggested by some earlier studies,2 but instead give way to a much more chaotic environment in which established mainstream political * I would like to thank Chris Achen, Nancy Bermeo, Ruth Berins Collier, Anna Grzymala-Busse, Keena Lipsitz, Kristin Michelitch, Conor O’Dwyer, Regina Smyth, Joshua Tucker, and four anonymous reviewers for helpful comments on earlier drafts of this paper. 1 Shafir 1997; Ganev 2007; Krause 2003. 2 Olson 1998; Miller et al. 2000; Tavits 2005. World Politics 62, no. 2 (April 2010), 221–60 Copyright © 2010 Trustees of Princeton University doi: 10.1017/S0043887110000043 222 world politics 3 Pacek 1994; Fidrmuc 2000; Tucker 2006. 4 Tucker 2005. 5 Tismaneanu 1998; Ramet 1999; Gallagher 2000. 6 Evans and Whitefield 1993. 7 Tavits 2005. 8 Filippov, Ordeshook, and Shvetsova 1999. 9 Moreover, most of the electoral changes since the early 1990s have tended to disadvantage the smaller unorthodox parties (e.g., raising electoral thresholds in countries like Poland and Romania), while the most prominent change in political system—Moldova’s switch from a presidential to a parliamentary system— should have also promoted stable party competition. parties lost considerable ground to new political formations based on personalist and populist appeals? Why did this reversal in Eastern Europe happen during a period of economic recovery, remarkable Western-integration progress, and broad acceptance of electoral democracy as the only game in town? What are the implications of these trends for the future of party politics and representative democracy in the former communist countries and beyond? The existing academic literature on postcommunist voting and political parties suggests a variety of possible explanations for the popularity of uops in Eastern Europe. A number of studies demonstrate the importance of economic voting in postcommunist elections.3 From that perspective, given the severity of the postcommunist economic trauma, the appeal of unorthodox parties can be explained at least in part as reflecting the search of impoverished voters for political alternatives to a disappointing status quo. However, as Joshua Tucker4 shows, such economic voting seems to affect support for communist successor parties but not for nationalist parties and candidates. Moreover, economic voting is not particularly useful for explaining the recent rise in unorthodox party support, given that this surge has happened against a backdrop of improving economic conditions in most East European countries. Some scholars rightly emphasize the appeal of ethnic politics to both postcommunist elites and average citizens,5 especially in ethnically diverse societies.6 But such an approach cannot account for the rising appeal of unorthodox parties in the region after the high ethnic tensions of the early 1990s had significantly declined. Other authors emphasize the importance of institutional arrangements such as district magnitude7 and direct presidential elections8 for electoral volatility. While institutional choices may explain cross-country differences in electoral patterns, they are less useful for explaining temporal variation because with a few exceptions, electoral institutions in Eastern Europe have been fairly stable for the last decade.9 Given the limitations of the existing approaches, this article proposes an alternative theory for the popularity of uops that focuses on throwing ou t the bums 223 10 The most frequent targets of such explanations are West European extreme-right and antiimmigration parties (Kitschelt 1995; Van der Brug, Fennema, and Tillie 2000), but the 1992 vote in the U.S. for Ross Perot has also been interpreted along these lines (Doherty and Gimpel 1997). 11 Van der Brug, Fennema, and Tillie 2000. 12 For an overview of the extensive literature on retrospective economic voting see Lewis-Beck and Stegmaier 2000. 13 Pop-Eleches 2001; Krastev 2002; Ucen, Gyarfasova, and Krivy 2005. 14 Mudde 2001. the role of protest voting in successive generations of postcommunist elections. Protest voting is the practice of voting for a party not because of the actual content of its electoral message but in order to “punish” other parties. It has been invoked as an explanation for a variety of electoral outcomes in Western democracies,10 but has been applied almost exclusively to extremist parties11 and has received much less theoretical attention than its parent concept, retrospective voting.12 In the East European context, protest voting has been used to explain individual election outcomes13 but there has been no systematic theoretical effort to understand the implications of protest voting for postcommunist party systems. This omission is particularly surprising since the postcommunist transition provided an ideal setting for protest voting due to significant voter discontent with incumbents and weak partisan ties between voters and parties after more than four decades of one-party rule. I argue that protest voting is crucial for understanding postcommunist electoral dynamics and that its implications are significantly mediated by electoral sequence. In particular, what sets apart the recent wave of elections is not the prevalence of protest voting but its main beneficiaries. Whereas in the first two generations of postcommunist elections, most voters disaffected with the status quo could opt for untried mainstream alternatives to the incumbents, in third-generation elections, which occur after at least two different ideological camps have governed in the postcommunist era, voters faced a shortage of untried mainstream alternatives. Therefore, the same basic political reflex of punishing incompetent and/or corrupt incumbents has produced both the initial “healthy” power alternation and the recent rise of uops in Eastern Europe. This article contributes in several ways to the literature on political parties and elections in the postcommunist context and beyond. First, given Cas Mudde’s complaint that “there is still a lack of clear-cut definitions [of populist parties] that could be used in empirical research,”14 I propose a definition and typology of unorthodox parties that allows a more integrated approach to understanding the political appeal of a wide variety of apparently very different political parties, ranging from 224 world politics communist to nationalist and new/centrist populist. Second, the article illustrates the importance of electoral sequence for the fortunes of different types of uops and thereby addresses an important theoretical gap in understanding the temporal dynamics of elections in postcommunist and other new democracies. Finally, the discussion contributes to a better understanding of the implications of retrospective and protest voting for the stability of party competition in new democracies. The first section of the article defines unorthodox parties and discusses the dimensions along which different uop subtypes deviate from the mainstream political party model. The next section introduces the concept of postcommunist election generations and provides a brief overview of the changing electoral dynamics of successive generations of postcommunist elections. The third section uses statistical evidence from seventy-six parliamentary elections in fourteen East European countries to analyze the electoral fortunes of governing and unorthodox parties and then illustrates the dynamics of protest voting on the basis of individual-level survey data from twelve postcommunist elections and a survey experiment in Bulgaria. The final section puts these findings in comparative perspective and discusses their theoretical implications for the understanding of electoral dynamics and democratic consolidation in Eastern Europe and beyond. I. Unorthodox Parties: Conceptualization and Classification Unorthodox political parties are not a unique East European phenomenon—such parties have made significant inroads in other developing countries (Venezuela and Peru, for example) and even in advanced democracies like Austria, Denmark, Holland, and Italy. While crossregional comparisons of this phenomenon are a promising intellectual enterprise, this article focuses on fourteen parliamentary democracies in ex-communist Eastern Europe: Albania, Bulgaria, Croatia, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Macedonia, Moldova, Poland, Romania, Slovakia, and Slovenia. This case choice is justified by the fact that these countries experienced comparable economic and political challenges during the postcommunist transition but differed along several crucial dimensions potentially associated with unorthodox party success (such as economic performance, ethnic diversity, and institutional configuration). To improve comparability and causal homogeneity, I exclude strongly authoritarian countries (Azerbaijan, Belarus, and the Central Asian republics); countries with short demo- throwing ou t the bums 225 cratic track records (Bosnia-Herzegovina, Montenegro, and Serbia);15 and countries where parliaments were dominated by extremely powerful presidents (Armenia, Georgia, Russia, and Ukraine). Analysts of postcommunist party systems have repeatedly emphasized the limited utility of concepts from the political party literature on established Western democracies in the unstructured political context of the postcommunist transition. Basic analytical categories such as the left-right distinction or classical measures of fractionalization and volatility have not traveled well to the new democracies.16 They have been at least partially replaced by more region-specific categories, such as communist successor parties17 or the distinction between radicalreturn and radical-continuity patterns among East European extremist parties.18 As a result, most analyses either restrict their geographical coverage to facilitate comparability between cases19 or focus on specific types of parties, most notably communist-successor parties20 and extreme-right parties.21 However, since the question at hand requires a more comprehensive classification of unorthodox parties, this article proposes such a typology and then uses it in the subsequent empirical analysis. Defining Mainstream and Unorthodox Parties Unorthodox parties are defined by what they are not—namely, orthodox or mainstream political parties.22 A political party is classified as mainstream if its electoral appeal is based on a recognizable and moderate ideological platform rather than on the personality of its leader and/or extremist rhetoric. In other words, a mainstream party represents an ideological orientation that can be mapped with reasonable accuracy onto the mainstream ideological spectrum of established Western democracies. By contrast, as illustrated in Figure 1, an unorthodox party is defined by the extent to which it deviates from the mainstream party 15 While Croatia is also vulnerable to this criterion given FranjoTudjman’s authoritarian governing style for much of the 1990s, Croatia nevertheless scored significantly higher than Slobodan Milosevic’s Yugoslavia in terms of Freedom House political rights and civil liberties, and was therefore included in the analysis. However, the statistical results presented in this article are not affected by this choice (see Table A3 in the electronic appendix at www.princeton.edu/~gpop). 16 Tismaneanu 1998; Kreuzer and Pettai 2003; Powell and Tucker 2008. 17 Grzymala-Busse 2002; Tucker 2006. 18 Shafir 1999. 19 Toka 1998; Kitschelt et al. 1999. 20 Ishiyama 1997; Orenstein 1998; Gryzmala-Busse 2002. 21 Ramet 1999; Shafir 2000; Beichelt and Minkenberg 2001. 22 Alternatively, these parties could also be labeled antimainstream or unconventional, but I settled on unorthodox as the least awkward alternative. Note, however, that in its present usage orthodoxy has neither religious nor strictly economic connotations. 226 world politics model along one or more dimensions. These deviations can take the form of adopting extremist political platforms in a number of issue areas and/or by sidestepping ideology and acting as political vehicles for their leaders. Unlike Giovanni Sartori’s discussion in which antisystem parties are explicitly defined in relation to other parties within a given national party system,23 this article defines unorthodox parties using an international reference point—the West European moderate programmatic party.24 The approach is justified for several reasons. First, the party systems of Western democracies provided an essential reference point for the nascent postcommunist democracies. Unlike their Latin American counterparts, where shorter and shallower authoritarian episodes25 were 23 Sartori 1976. 24 Of course, this does not mean that all West European parties are moderate programmatic parties. However, this does not invalidate the model; it just means West European democracies also have their fair share of unorthodox parties. 25 Bunce 1998. Extreme nationalism Nonethnic appeals Anticapitalist Procapitalist Nonideological/ personality driven Ideology/platform driven Radical left Extreme National Opportunist main Main nationalist populist stream stream New/centristpopulistFigure 1 Issue Space and Party Appeals throwing ou t the bums 227 usually followed by the revival of traditional political parties, after five decades of single-party rule Eastern Europe had only distant—and often not particularly compelling—domestic party system models to fall back on.26 Second, Western democracies used their considerable economic and political leverage to impose rather strict limits on the range of acceptable party platforms and to sanction any significant deviations from the basic tenets of economic and political liberalism.27 Third, using a national reference point would wrongly conclude that certain extremist parties (such as the nationalist Croatian Democratic Union [hdz] in Croatia or the Communist Party [pcrm] in Moldova) were mainstream, since they were the dominant political forces in their respective countries and therefore defined the domestic political mainstream.28 Using the Western model of the moderate programmatic party as a theoretical baseline is justified by the fact that most mainstream East European parties have made concerted efforts to fit into one of the established Western ideological traditions (socialism/social democracy, liberalism, conservatism/Christian democracy, and ecology parties).29 A Typology of Unorthodox Parties Depending on the type and the extent of their deviations from the mainstream, uops fall into several different categories (see Figure 1). The typology relies on three dimensions, which capture the two most important policy dimensions of postcommunist party appeal— economic policy orientation and reliance on ethnonationalist ap- peal30 —along with an organizational dimension reflecting the relative prominence of individual leaders. While postcommunist parties obvi- 26 Moreover, many of the more successful revivals of historical parties—such as the Czech Social Democratic Party (cssd) and the Romanian National Liberal Party (pnl)—fit the European model based on prewar attempts to emulate West European party politics. 27 Innes 2002. 28 This is also the reason why I have opted to use “unorthodox party” rather than the more widespread term of antiestablishment party (Schedler 1996; Abedi 2004); the latter has been applied to West European parties that, unlike their East European counterparts, remained outside of governing coalitions and were thus truly antiestablishment. 29 Of course, such efforts do not mean that mainstream postcommunist political elites have fully embraced Western liberal values. Nonetheless, mainstream leaders responded to international incentives and avoided or moderated their appeals on issues that conflicted with these international norms. 30 An obvious additional dimension is a party’s attitude toward European integration. However, opposition to Western integration is largely driven by either nationalism or opposition to capitalism (Hooghe, Marks, and Wilson 2002), and can therefore be captured as a combination of the two main policy dimensions. Following Sartori, I also consider support for the democratic regime as a policy dimension. However, since in post–cold war Europe democracy was (at least rhetorically) the only game in town, parties actively questioning the democratic regime had negligible political support among the countries in my sample. 228 world politics ously differ along a number of additional cleavages,31 for the purpose of identifying unorthodox parties, the two policy dimensions capture the most frequent and significant deviations from Western economic and political liberalism in Eastern Europe. A summary of party classifications can be found in Table 1. There are instances of radical parties on both extremes of the traditional political spectrum. Michael Shafir32 persuasively argues that the distinction between the radical left and the radical right is blurrier in the postcommunist context than elsewhere due to the enduring legacy of national communism for many countries in the region. Indeed many of these parties share a rejection of Western individualism and liberal democracy and are usually opposed to Western integration and neoliberal economic reforms. Nevertheless, a further division of the parties into radical-left and extreme-nationalist33 camps seems justified by the parties’ different international referents34 and by the dimension along which they reject the Western economic and political model. In the case of extreme-nationalist parties such as the Czech Republican Party (rsc-rpr), Romania’s prm, the Hungarian Life and Justice Party (miep), and the Slovak National Party (sns), this rejection is driven and accompanied by extreme nationalism usually combined with racism and anti-Semitism. Meanwhile, the radical left—composed almost exclusively of unreformed (or minimally reformed) communist successor parties such as the Czech and Moravian Communist Party (kscm), the Romanian Socialist Labor Party (psm), the Association of Workers of Slovakia (zrs), and the Hungarian Communist Workers’ Party (mkmp)—advocates a return to the golden days of the communist era. The second group of unorthodox parties, national populists, also feature nationalism as a prominent element of their electoral appeal and claim to represent the interests of an often mythical and idealized national collectivity. While their reliance on a Schmittian us-versus-them dichotomy places these parties, sometimes uncomfortably, close to the extreme nationalists (and their strength has undermined the emergence and consolidation of postcommunist democracy in Slovakia and the former Yugoslavia),35 national populists deserve a separate category 31 Kitschelt et al. 1999; Whitefield 2002. 32 Shafir 2001. 33 I chose the extreme-nationalist rather than the radical-right label to avoid mistaken associations with right (i.e., promarket) economic positions. 34 Both Istvan Csurka and Corneliu Vadim Tudor openly admit their intellectual indebtedness to the political ideas of the French National Front leader Jean-Marie Le Pen, whereas the Czech kscm looks to other communist parties as its potential international allies. 35 Bunce 2003. Table 1 Overview of Unorthodox Parties in Eastern Europe Country Party Name uop Type Elections Albania National Front Party (pbk) national populist 1996–2005 Albanian Republican Party (prsh) national populist 1996–2005 Bulgaria Bulgarian Business Bloc (bbb) new/centrist populist 1991–97 National Movement Simeon II new/centrist populist 2001   (ndsv) Ataka extreme nationalist 2005 Croatia Croatian Party of Rights (hsp) extreme nationalist 1992–2003 Croatian Democratic Union (hdz) national populist 1990–2000 Czech Czech and Moravian Communist radical left 1990–2006  Republic   Party (kscm) Republican Party (rsc-rpr) extreme nationalist 1990–2002 Estonia Committee for Defense of radical left 1990  Soviet Power Communist Party Free Estonia radical left 1990   Bloc Right Wingers’ Party (vkr) national populist 1995 Estonian National Independent national populist 1992   Party (ersp) Estonian Center Party (k) new/centrist populist 1992–2003 Estonian Country People’s Party   (eme/er) national populist 1999–2003 Res Publica (rp) new/centrist populist 2003 Hungary Hungarian Justice and Life (miep) extreme nationalist 1994–2006 Independent Smallholders (fkgp) national populist 1990–2002 Hungarian Socialist Workers’ radical left 1990   Party (mszmp) Workers’ Party (mp) radical left 1994–2002 Hungarian Communist Workers’ radical left 2006   Party (mkmp) Latvia Interfront radical left 1990 Fatherland and Freedom (tb) national populist 1993–2006 Latvian National Conservative national populist 1993–95   Party (lnnk) Pop Movement for Latvia (tkl-zp) extreme nationalist 1993–95 Latvian Unity Party (lvp) radical left 1993–98 Popular Harmony Party (tsp) radical left 1993–1998,   2006 Latvian Socialist Party (lsp) radical left 1993–95 For Human Rights in United radical left 2002–6  Latvia (pctvl) People’s Party (tp) new/centrist populist 1998 New Party (jp) new/centrist populist 1998 New Era Party (jl) new/centrist populist 2002 Table 1, cont. Country Party Name uop Type Elections Lithuania Independent Lithuanian radical left 1990   Communist Party Communist Party (sfsu) radical left 1990 Lithuanian National Union (lts) national populist 1992–96 Lithuanian National Party/Young extreme nationalist 1996–2000  Lithuanians (lnp/jl) New Union (ns) new/centrist populist 2000 Lithuanian Liberal Union (lls) new/centrist populist 2000 Order and Justice (tt) national populist 2004 Labor Party (dp) new/centrist populist 2004 Macedonia Internal Macedonian Revolutionary national populist 1990–2002  Organization (vmro) Moldova Communists’ Party of the Republic radical left 1990,   of Moldova (pcrm) 1998–2005 Alianta Electorala Braghis (beab) new/centrist populist 2001 Socialist Party/Unity (ps/edinstvo) radical left 1994–2001 Electoral Bloc Patria Rodina (bepr) radical left 2005 Poland Confederation Independent Poland extreme nationalist 1991–93   (kpn) Christian National Union (zchn) national populist 1991–93 Realpolitik Union (upr) extreme nationalist 1991–97 Law and Justice (pis) new/centrist populist 2001 Law and Justice (pis) national populist 2005 Self Defense (srp) radical left 1993–2005 Movement for Reconstruction of national populist 1997   Poland (rop) League of Polish Families (lrp) national populist 2001–5 Romania Greater Romania Party (prm) extreme nationalist 1992–2004 Romanian National Unity Party national populist 1992–2000   (punr) Socialist Labor Party (psm) radical left 1992–2000 New Generation Party (png) new/centrist populist 2004 Slovakia Movment for Demcratic Slovakia national populist 1992–2006   (hzds) Movement for Democracy (hzd) national populist 2002 Slovak National Party (sns) extreme nationalist 1990–2006 Right Slovak National Party (psns) extreme nationalist 2002 Association of Workers’ (zrs) radical left 1994–98 Slovak Communist Party (kss) radical left 1990–2006 Party of Civic Understanding (sop) new/centrist populist 1998 Direction—Social Democracy new/centrist populist 2002   (smer) Alliance of New Citizens (ano) new/centrist populist 2002 Slovenia Slovenian National Party (sns) extreme nationalist 1992–2004 New Slovenia (nsi) new/centrist populist 2000 throwing ou t the bums 231 because their nationalist and/or racist rhetoric tends to be less virulent, less direct, and less prone to incite physical violence against minorities. Moreover, in ideological terms, nationalism is often supplemented by a broader non-nationalist policy agenda aimed at specific groups (farmers in the case of Hungary’s Smallholders [fkgp] and Catholics in the case of Poland’s League of Polish Families [lpr], for example) rather than being the party’s only raison d’être. At the same time, it is important to distinguish national-populist parties from mainstream-opportunist parties (see Figure 1) that may occasionally resort to nationalist rhetoric or allies but that primarily pursue moderate ideological platforms.36 New/centrist-populist parties make up the final uop category. These parties do not adopt radical ideologies but rather attempt to sidestep ideology altogether by claiming to be nonideological antipolitical formations. The most prominent example of such formations is the Bulgarian National Movement Simeon II (ndsv), but similar parties have also surfaced in Lithuania (the Lithuanian Liberal Union [lls] of former Prime Minister Rolandas Paksas, and the New Union [ns] of Arturas Paulauskas); Slovakia (Robert Fico’s Direction-Social Democracy [smer] and Pavol Rusko’s Alliance of New Citizens [ano]); Latvia (the New Party [jp] of prominent composer Raimond Pauls; the People’s Party [t] of popular former Prime Minister Andris Skele; and more recently the New Era Party [jl] of former Central Bank head Einar Repse); and Estonia (Edgar Savisaar’s Estonian Center Party [k]). The vague electoral appeal of new/centrist-populist parties goes hand-in-hand with the strong presence of their prominent leaders who create the parties as vehicles for their personal political ambitions usually just prior to elections. Such parties are almost completely unencumbered by ideological constraints and are therefore free to tell the voters what they want to hear. Illustrative in this respect is Fico’s response to charges of populism: “While others are lying on beaches and resting, I go to the people. When you hear the same things from people at hundreds of political meetings, you’re obliged to repeat them on tv. 36 Of course, in practice the distinction may be more difficult to make. In such cases I rely on expert surveys where available and the party’s international affiliation as litmus tests. Thus, while there is some debate about whether pis should be labeled national populist in 2005, I decided to do so because of its decision to join the national conservative Alliance for Europe of the Nations (aen) and because the University of North Carolina (unc) expert survey codes pis as radical right (Steenbergen and Marks 2007). The Hungarian fidesz, despite some nationalist flirtations, joined the mainstream European People’s Party and is coded by the unc as conservative (and as opportunist mainstream in this article). Similarly, the Romanian pdsr and the Bulgarian bsp do not qualify as national populist in the mid-1990s because their use of nationalist appeals was largely opportunistic and at the same time they actively pursued membership in the Socialist International. 232 world politics . . . If these statements, which correspond to reality in Slovakia, are populism, then I want to be a populist.” 37 As will be discussed in more detail in the following section, the new/ centrist-populist parties have largely arisen as a reaction to the general disappointment of East European electorates with mainstream parties and the high cost of economic reforms. Unlike the radical and national-populist parties described above, however, new/centrist-populist parties avoid nationalist, anti-Western, and anticapitalist stances. By doing so, they promise weary electorates they will square the transition circle by pursuing Western integration and punishing mainstream elites widely associated with declining living standards and rampant corruption. Such appeals are most effective when the party is a newcomer or is in the parliamentary opposition; but once contaminated by power and constrained by the realities of governing, such parties are likely to reveal (or find) their true ideological colors and will thereby move into either the mainstream or the more radical party categories.38 As this discussion shows, there is no single element in electoral appeal and party structure that is shared by all unorthodox parties. Thus, while many of the unorthodox parties resort to nationalist or anticapitalist appeals, relatively few parties combine both types of appeals in their party platform. Similarly, while powerful and charismatic leaders are the driving force of many unorthodox parties, for others—such as the Moldovan communists—that is not the case. However, as David Collier and James Mahon’s discussion of “family resemblance” categories shows,39 the absence of such a central common element does not necessarily undermine the utility of a concept, as long as the members of the group can be recognized as belonging to the same family. In the case of unorthodox parties, this unifying principle is their significant deviation from the model of the moderate programmatic party ideal discussed above.40 37 Pisárová, Martina, and Tom Nicholson. “Robert Fico: If Speaking the Truth is Populism, then I Want to be a Populist.” Slovak Spectator 8, no. 28, July 22–August 4, 2002. 38 However, it should be noted that all new parties are not necessarily new/centrist populist (the Civic Platform [po] in Poland in 2001, for example). Some new parties explicitly adopt a moderate ideological platform and are not excessively dominated by a single personality. 39 Collier and Mahon 1993. 40 Another way to think about unorthodox parties is as a negative radial category. Unlike a typical radial category, which is anchored in a central subcategory/prototype and where noncentral subcategories share at least some of the prototype’s defining attributes (Collier and Mahon 1993, 848), a negative radial category is defined by the differentiation of group members from the central category (mainstream parties) along one of the attributes that define this center. Moreover, the central category does not belong to the group in a negative radial category. throwing ou t the bums 233 II. Electoral Dynamics in Three Generations of Postcommunist Elections This section discusses the most important region-wide trends in protest voting and uop support in three waves of postcommunist elections: first-generation founding elections, defined as the first at least partially competitive elections in the region since World War II; second-generation elections, which are elections that occur after the founding election but before center-right “willing” reformers and leftist or national-populist “unwilling” reformers have governed for a significant time period in the postcommunist period;41 and third-generation elections, which occur after two different ideological camps have had a significant shot at governing. For an overview of elections by generation see Table 2. Founding Elections: From The Party to Antiparties Formed in opposition to the Communist Party, the heterogeneous mix of individuals who joined forces against the communists in the founding elections were hardly parties in their own right. In fact, many groups avoided a party label—preferring vague names like Czech Civic Forum, Romanian National Salvation Front, Slovak Public Against Violence, and Polish Solidarity—and even proclaimed their reluctance to engage in factionalist party politics.42 While mainstream political parties emerged in time for the founding elections in a some countries, with the exception of Hungary (where prior political liberalization had facilitated party formation before the fall of communism), the first elections almost invariably pitted two broad camps against each other: marginally reformed communists and broad anticommunist coalitions. These coalitions usually included a mix of liberals, conservatives, nationalists, and traditional social democrats, whose only real commonality was their rejection of the communist regime. Despite their strong showing in the founding elections, most of these broad fronts barely outlived the demise of their common enemy—the communist regime—and usually split into a number of more ideologically homogeneous political parties in what looked like an important step toward the normalization of party politics in Eastern Europe.43 41 This category always includes the second election and for some countries may also include the third and fourth postcommunist election if earlier elections have not brought about two significant power alternations. 42 For example, the Czech Civic Forum claimed to be above party politics as indicated by its 1990 campaign slogan: “Parties are for party members, Civic Forum is for everybody.” 43 Miller et al. 2000. 234 world politics Second-Generation Elections: The Normal Years As a result of the breakup of the broad anticommunist coalitions, beginning with the second-generation elections, East European party systems became more differentiated within each country and began to display some recognizable cross-national differences. In those elections, radical-left parties lost a lot of ground compared to the founding elections, as most ex-communist parties adopted more moderate socialist/social democratic names and platforms. Unreformed communists had negligible support everywhere except in Moldova and the Czech Republic. At the same time, however, national-populist parties, which had established themselves as the strongest anticommunist challengers in the Croatian and Macedonian founding elections, emerged as crucial political players in Slovakia and Latvia and as a significant presence (and occasional governing-coalition partner) in Hungary, Poland, and Romania. Extreme-nationalist parties also reared their ugly heads in most East European second-generation elections, but generally with modest electoral results and limited political influence.44 The exceptions were the 15 percent vote share won by the Latvian Ziegerista 44 The main exceptions in this respect were the participation of the Slovak sns in the 1994–98 hzds government and the brief participation of the Romanian prm in the pdsr-led government for a few months in 1995. Table 2 Overview of Election Generations in Eastern Europe Initial Second-Generation Third-Generation Country Election Elections Elections Albania 1991 1992, 1996, 1997 2001, 2005 Bulgaria 1990 1991, 1994, 1997 2001, 2005 Croatia 1990 1992, 1995, 2000 2003 Czech Rep. 1990 1992, 1996, 1998 2002, 2006 Estonia 1990 1992, 1995 1999, 2003 Hungary 1990 1994 1998, 2002, 2006 Latvia 1990 1993, 1995 1998, 2002, 2006 Lithuania 1990 1992, 1996 2000, 2004 Macedonia 1990 1994, 1998 2002, 2006 Moldova 1990 1994, 1998 2001, 2005 Poland 1989 1991, 1993 1997, 2001, 2005 Romania 1990 1992, 1996 2000, 2004 Slovakia 1990 1992, 1994, 1998 2002, 2006 Slovenia 1990 1992, 1996, 2000, 2004 throwing ou t the bums 235 Party in 1995 and the 10 percent vote share won by the Slovenian National Party in 1992. The most consistent empirical regularity in East European democracies following the first-generation elections was a powerful anti-incumbency bias, as governing parties rarely managed to consolidate their electoral support in successive elections.45 Even though some of the governing parties occasionally escaped such retribution,46 in most cases the disenchanted voters punished incumbents by kicking them out of office.This led to an alternation of power between slow and fast reformers, usually represented by more or less reformed ex-communist successor parties and the democratic center-right opposition, respectively. While a detailed analysis of this remarkable incumbency disadvantage is beyond the scope of this article, a few common elements are worth mentioning here. First, despite a gradual economic recovery in most East European countries during the second part of the 1990s, living standards continued to fall for large portions of the population throughout the decade, which, combined with rising inequality, high unemployment, and welfare cuts, undermined the popularity of even the most capable governments.47 Second, the policy constraints imposed by Western conditionality (initially primarily from the World Bank and the International Monetary Fund and later increasingly from the European Union) left relatively little maneuvering space for governments trying to live up to their often unrealistic electoral promises regarding social protection and higher living standards.48 Third, the credibility of governments’ alternative explanations for economic difficulties was severely undermined in most countries by the widespread and blatant instances of corruption at all levels of government. Thus, in 2001, of the ten Eastern European EU candidates, Slovenia was the sole country where only a minority of respondents (42 percent) believed that most or almost all public officials are corrupt.49 Elsewhere in the region corruption perceptions ranged from 54 percent in Hungary to 89 percent in Romania, 92 percent in Latvia, and 95 percent in Lithuania,50 suggesting that after a decade of reforms East European 45 See for example, Roberts 2008. 46 Examples include the remarkably resilient Liberal Democracy of Slovenia Party (lds) and Latvia’s Way Party (lc), which participated in several successive postcommunist governments, as well as the national-populist hzds in Slovakia and hdz in Croatia, though all of these parties were eventually defeated at the polls. 47 Pacek 1994. 48 Innes 2002. 49 Not coincidentally, Slovenia has been the country with the most stable and moderate party system among the countries discussed here. 50 Rose 2002. 236 world politics electorates felt an overwhelming sense of having been betrayed by the political elite. Finally, frequent leadership defection, party splits, mergers, and name changes prevented most political parties (with the partial exception of the ex-communists) from building effective party organizations. Combined with the general weakness of organized interests at the mass level, the result was highly unstable constituencies for most political parties, and hence high fluctuations in vote shares compared to other regions. As a consequence, many former governing parties not only lost power in the subsequent elections, but saw their vote share drop to a fraction of its previous level and, in some cases, lost parliamentary representation altogether (the Polish Solidarity Electoral Alliance [aws] in 2001, the Moldovan Democratic Agrarian Party [pdar] in 1998, and the Romanian Democratic Convention [cdr] in 2000, for example). Following the disputable circumstances of the founding elections, the alternation in power between parties with different ideological and policy platforms as a result of competitive (and usually fair) elections represented an important step toward democratic consolidation. While governance still left a lot to be desired, it was at least possible to “throw out the bums” by electoral means even in countries governed by less committed democrats (such as the Party of Social Democracy [pdsr] in Romania, the Movement for Democratic Slovakia [hzds], and the Croatian Democratic Union [hdz]) Ironically, while these successive peaceful turnover of power signaled the acceptance of democracy as “the only game in town” by the main political actors, they also set the stage for a new (and somewhat unexpected) test of the robustness of East European democracies by reducing the menu of untried mainstream alternatives to the incumbent parties/ coalitions. Third-Generation Elections: The Return of the Antiparty? The new democratic challenge in third-generation elections must be analyzed within the broader trend of protest voting in the postcommunist context. A protest vote (or antivote) is an electoral option driven less by the positive appeal of the chosen party’s ideological/policy platform than by the rejection of other possible political choices. As suggested above, protest voting played an important role even in early East European elections. In the founding elections the most consistent electoral appeal of the democratic challengers was their anticommunism. In subsequent rounds their messages became more varied, and to some throwing ou t the bums 237 extent more positive,51 but the general appeal to “vote for us because we are not the government” retained the primary emphasis of negative voting. Encouraged by mainstream opposition parties and embraced by large parts of the disillusioned electorate, protest voting at first appeared to be an easy fix to the dilemma of how to sustain democracy in poor countries with corrupt elites and weak civil societies. After all, the system ensured two crucial preconditions for democratic consolidation: first, it reduced the danger of popular revolts by giving electorates regular opportunities to vote the incumbents out of office; and second, it gave losers a stake in the democratic system due to the implicit promise that they would be returned to office by popular mandate in subsequent elections. However, whereas in prior elections voters disillusioned with the incumbents could opt for the mainstream opposition (i.e., reformed ex-communists in countries where reformers had won the founding elections, and center-right reformers in countries where ex-communists or national populists had won the first elections), in third-generation elections both main camps already had well-established and usually not-altogether-positive track records, thus reducing the likelihood that ideologically uncommitted voters would give them the benefit of the doubt. Therefore, the protest-vote mechanism backfired on the established political elites, and the seemingly consolidated party systems emerging from the second-generation elections suffered significant transfigurations during third-generation elections.52 Rather than recount individual country experiences with thirdgeneration elections, I will note a few relevant regional trends that emphasize the distinctive nature of this new generation of postcommunist elections. First, and most importantly, third-generation elections have marked significant electoral gains by unorthodox parties, often accompanied by serious or even devastating electoral losses by incumbent parties (such as in Romania in 2000, Moldova in 2001, and Poland in 2001 and 2005). Second, while the particular nature of predominant uops varied significantly across countries, the rise of these unortho- 51 By “positive” I do not mean to indicate any normative judgments but rather specific policy proposals about how to govern the country or what the party stands for rather than what it opposes. 52 The mechanism was a bit more complicated in Slovakia and Croatia, which were governed by national-populist (and, hence, unorthodox) parties for most of the 1990s. However, in both countries the broad coalitions of mainstream parties that succeeded them included reformed ex-communists and a mix of liberal and conservative anticommunist parties. By the time of the first third-generation elections, Slovak and Croatian voters had largely exhausted untried mainstream alternatives, thereby creating electoral opportunities for new unorthodox challengers. 238 world politics dox challengers took two main forms: an increase in the vote share of existing but previously marginal uops from either the nationalist53 or radical-left camps,54 and the emergence and spectacular electoral rise of entirely new political parties.55 These newcomers—usually new/centrist populists—rose from quasi-anonymity to national prominence in the months preceding parliamentary elections and, in a few instances,56 succeeded to outpoll their established mainstream competitors. Finally, while third-generation elections led to significant party-system reconfigurations in many East European countries, they did not necessarily do so each time, as illustrated by the solid performance of incumbents and the marginal role of uops in several third-generation elections.57 Therefore, the present argument is inherently probabilistic in that it predicts that third-generation elections have a greater likelihood of producing strong uop showings because of the lower availability of untried mainstream alternatives, but the actual election outcomes greatly depend on national and temporal circumstances such as the recent performance of mainstream parties (both government and opposition) and the supply of plausible unorthodox challengers. III. Statistical Tests This temporal overview of protest voting in postcommunist Eastern Europe suggests that after declining in second-generation elections, support for unorthodox parties once again increases during third-generation elections as voters find fewer acceptable mainstream parties to vent their frustration.This section tests whether the anecdotal evidence presented above can be confirmed by systematic cross-national statistical tests, especially when controlling for alternative explanations of unorthodox party support. I have constructed a data set of seventy-six postcommunist parliamentary elections from 1990–2006 that,while too small for more elaborate simultaneous equation models, nevertheless yields a sufficiently large number of observations to be analyzed by traditional statistical methods for time-series cross-sectional data. 53 Examples include the prm in Romania in 2000, the fkgp and miep in Hungary in 1998, and Self-Defense (srp) in Poland in 2001. 54 See for example, the Communist Party (pcrm) in Moldova in 2001, srp in Poland in 2001, and the kscm in the Czech Republic in 2002. 55 As Powell and Tucker (2008) argue, electoral volatility driven by such new parties has potentially much more serious economic and political repercussions than volatility driven by vote shifts among existing parties. 56 The most dramatic example was the ndsv in Bulgaria in 2001, but this also applies to the jp in Latvia in 2002 and the Labor Party (dp) in Lithuania in 2004. 57 Examples include Poland in 1997 and 2007, Albania in 2001, and Hungary in 2002 and 2006. throwing ou t the bums 239 In line with the theoretical discussion above, the statistical tests focus on two types of indicators of postcommunist electoral dynamics. First, for each election I coded the combined vote shares of governing parties during the most recent and preceding election58 and calculated the change in voter preferences for parties with significant involvement in the outgoing government.59 Second, to assess the drivers of electoral success for unorthodox parties I used odds-ratio transformations60 of the vote shares of six different types of uops shown in Table 3: new/centrist-populist parties (model 1); radical-left parties (model 2); extreme-nationalist parties (model 3); nationalist parties, comprising extreme-nationalist and national-populist parties, (model 4); extremist parties, comprising radical-left and extreme-nationalist parties (model 5); and an aggregation of all different types of unorthodox parties (model 6). Since the data set is cross-sectionally dominated and displays temporal dependence between successive elections, I ran random-effects Prais-Winsten regressions with heteroskedastic panel-corrected standard errors and a correction for first-order autocorrelation.61 Given the theoretical emphasis on temporal trends in postcommunist elections, the main independent variables are dummy variables indicating whether a given election was a first- or third-generation election (using the definitions discussed in Section I). The excluded category—second-generation elections—allows for a more direct test of whether the dynamics of third-generation elections differ significantly from those of preceding electoral contests. In particular, from a protest-voting perspective, there should be stronger support for uops during third-generation elections. Meanwhile, the party-system consolidation hypothesis predicts increasingly normal results in successive generations of postcommunist elections—namely a reduction in the incumbency disadvantage of the early transition era and declining vote shares for uops.62 58 Powell and Whitten 1993. 59 These calculations account for mergers and splits in governing parties but I exclude parties whose participation in the government was less than a year, since such parties are less likely to be held responsible by voters for the government’s performance. 60 For a justification of this approach see Mosteller and Tukey 1977, 109. 61 Random-effects models have the advantage of capturing the important cross-national variation in my sample. Running fixed-effects models does not change the main findings (see electronic appendix at http://www.princeton.edu/~gpop) but it requires dropping several interesting time-invariant independent variables, so they are omitted here. 62 Statistical tests using either a simple time trend or an indicator of the age of electoral democracy (Tavits 2000) produced significantly weaker model fits (even when also including squared terms), which suggests that the more theoretically grounded approach of focusing on election generations also performs better empirically. Table 3 Drivers of Electoral Support for Different Types of UOPs Model 5 Radical Model 7 Model 9 Model 1 Model 2 Model 3 Left and Incumbent uops New/Centrist Radical Extreme Model 4 Extreme Model 6 Vote Model 8 (Instrumental Populist Left Nationalist Nationalist Nationalist uops Change uops Variable) First-generation election –.367 1.499** –.776* –.135 .332 .846 –24.10** 1.219 1.581 (.494) (.522) (.352) (.417) (.546) (.520) (8.510) (.778) (1.329) Third-generation election 1.021** .449 –.419 –.385 .539# .716* –2.249 –.374 –.391 (.343) (.355) (.270) (.340) (.373) (.347) (3.869) (.402) (.720) EU member –1.448# –1.117 1.107# 2.052** –1.126 –.362 –4.757 .024 –.257 (.854) (.820) (.581) (.682) (.861) (.830) (8.733) (.776) (.853) First-tier EU candidate –.358 –1.571** 1.364** 1.653** –.943 –.405 –6.333 –.427 –.449 (.549) (.595) (.411) (.543) (.593) (.557) (6.267) (.503) (.546) Second-tier EU candidate –.317 –1.679** .853* 1.199* –1.269** –.964* –4.754 –.900* –.930* (.435) (.463) (.341) (.468) (.463) (.443) (4.470) (.405) (.430) Inflation –.258* .026 .066 .399** .051 .190 –1.573 .265* .221# (.126) (.125) (.095) (.141) (.127) (.121) (1.188) (.112) (.125) Unemployment .004 –.026 –.007 .078** –.042* .027 –.149 .020 .024 (.024) (.021) (.014) (.019) (.021) (.021) (.190) (.019) (.021) gdp change .005 .017 .004 –.005 .035# .013 –.065 .015 .007 (.018) (.019) (.014) (.017) (.021) (.018) (.152) (.017) (.019) gdp as percent of 1989 –.008 –.001 –.007 .006 –.003 –.010 .216* –.008 –.003 (.009) (.010) (.006) (.008) (.010) (.010) (.096) (.009) (.010) gdp per capita .210* .081 .021 –.074 .100 .133 .024 .131 .106 (.093) (.103) (.070) (.096) (.110) (.102) (.989) (.095) (.104) Number of governments .155 –.239 –.189 –.299# –.185 –.252 –1.658 –.313# –.338# (.181) (.169) (.133) (.156) (.176) (.186) (1.953) (.167) (.198) Ethnic minority share .030* .021 –.008 .012 .015 .036** .109 .035** .035** (.014) (.015) (.010) (.012) (.014) (.012) (.130) (.011) (.012) Urban-rural cleavage .010 .032 –.046** –.057** –.007 –.004 –.312 –.013 .001 (.021) (.021) (.016) (.017) (.021) (.020) (.218) (.018) (.020) pr system .297 1.376** .076 .147 1.013** .883** –3.447 .766** .876* (.337) (.308) (.224) (.266) (.309) (.336) (3.700) (.296) (.342) Elected president .194 –.551# –.026 –.461# –.393 –.248 –.542# –.142 (.325) (.325) (.235) (.276) (.324) (.332) (.311) (.335) Centrist populist vote share –.036* (.016) Incumbent vote share (t–1) –.625** (.156) Incumbent vote share change .011 .018 (.010) (.022) Third-generation election* –.074** –.073#   incumbent vote change (.018) (.040) Country dummies no no no no no no no no no N 76 76 76 76 76 76 75 75 75 R2 .39 .52 .43 .54 .41 .47 .77 .52 .46 Prais-Winsten regression coefficients with standard errors in parenthesis; # p<.1, ** p<.01, * p<.05; one-tailed where appropriate 242 world politics Second, the regressions attempt to capture the important influence of the EU integration process on East European party politics. While a detailed discussion of these complicated dynamics is beyond the scope of the present article, it is nevertheless useful to assess the electoral repercussions of the strict constraints placed by EU conditionality on the policy options of East European governments. At the elite level these constraints arguably contributed to a more moderate electoral dis- course,63 and even though the limited degree of policy differentiation for mainstream political parties64 may have left wider policy space (and more voters) at the disposal of uops, this effect was probably muted by the fact that postcommunist voters were broadly (and in some countries overwhelmingly) supportive of EU membership65 and were probably more reluctant to endorse political parties whose electoral success could endanger EU accession. These effects should be the strongest in countries where EU membership was possible but by no means assured at the time of the election, whereas countries that were either already EU members or those with no immediate EU integration prospects should not be affected by such constraints. To test these hypotheses, the analysis builds on the Alexander Pacek, Grigore Pop-Eleches, and Joshua Tucker66 classification of postcommunist countries into three EU integration tiers67 and adds a fourth category—EU member. Third, the regressions include several economic performance indicators that are important elements of retrospective voting. In line with a well-established set of findings about the role of economic voting in East European elections,68 the models include two key economic performance measures: the unemployment rate in the year preceding the election and the logged average inflation rate for the two years preceding the election. The models also include a third measure of shortterm economic performance: gross domestic product (gdp) change in the year before the election. Worse performance along any of these short-term measures should undermine the popularity of incumbents and increase the electoral appeal of unorthodox parties. Since voters may also respond to longer-term economic trends, I include an index of the country’s gdp as a percentage of pre-transition output levels (in 63 Vachudova 2008. 64 Innes 2002. 65 Pridham 2002. 66 Pacek, Pop-Eleches, and Tucker 2009. 67 Thus first-tier countries are countries with a high likelihood of being admitted in the next EU expansion wave, second-tier countries are less advanced but have been identified as credible potential candidates, and third-wave countries have minimal accession prospects at a given point in time. 68 Tucker 2001, 2006; Pacek 1994. throwing ou t the bums 243 1989). This index captures the country’s cumulative postcommunist economic trajectory; worse outcomes should affect not only current incumbents but also other participants in post-1989 governments, which should translate into greater electoral support for uop attacks against mainstream elites. Given Tucker’s69 finding of prospective voting for communist successor parties in poor areas, the regressions include a control for the country’s economic development level, measured as gdp per capita in the previous year. However, economic performance is only one of the dimensions along which voters may judge the governing parties. One prominent noneconomic source of voter disillusionment with the political elite’s handling of the transition is the prevalence of political infighting in parliamentary politics.The spectacle of endless, frivolous, and often inconclusive parliamentary debates can fuel demands for a strong leader, thereby strengthening the electoral appeal of populist leaders and parties. Therefore, the models include a proxy for the political instability that often results from such conflicts—namely the frequency of major cabinet changes between elections.70 While higher political instability should trigger greater popular discontent and therefore higher electoral support for uops, it complicates the task of assigning blame and may channel the protest vote toward the “less guilty” mainstream parties. Fourth, there are good theoretical reasons for expecting that institutional choices would affect the development of political parties and electoral dynamics. An important institutional design question is whether countries with proportional representation (pr) electoral systems experienced the radicalization and fragmentation of party politics predicted by the theoretical literature. Since with the exception of a few initial elections none of the countries in the sample maintained the communist-era single member district (smd) system, the measure used in these tests is a dummy variable coded 1 for pr systems and 0 otherwise (primarily mixed systems).71 Similarly, given earlier findings that direct presidential elections contribute to postcommunist party fragmentation,72 the tests include a dummy variable indicating whether the country’s constitution prescribed direct elections for the presidency at the time of a given parliamentary election. We should expect di- 69 Tucker 2006. 70 I measure political instability as the number of cabinets since the previous elections. 71 I also test the effect of pr thresholds, but since it was not statistically significant, I exclude it from the models presented here. 72 Filippov, Ordeshook, and Shvetsova 1999. 244 world politics rect presidential elections to strengthen the appeal of personality-based politics and therefore lead to higher uop vote shares. Finally, the regressions include two indicators of potential important political cleavages. Given Geoffrey Evans and Stephen Whitefield’s expectation that ethnic cleavages are more likely to be politicized in ethnically diverse societies,73 the regressions include a measure of the total share of ethnic minorities. However, I obtain very similar results with alternative measures such as the share of the largest minority and the degree of ethnic fractionalization.74 In addition, following Margit Tavits,75 I use the absolute difference between the share of urban and rural population as a proxy of urban-rural cleavages that may be exploited by populist politicians.76 Discussion of Results The statistical results in Table 3 reveal significant temporal differences in the electoral support for different types of unorthodox parties. The statistical patterns confirm that the evolution of postcommunist party systems did not result in a region-wide convergence toward the Western ideal type of political competition between moderate and ideologically differentiated political parties. To interpret the coefficients in Table 3, it is important to recall that the reference category is second-generation elections. Thus, model 1 in Table 3 confirms the significant rise in the electoral appeal of new/centrist populists during third-generation elections. However, the half-life of success is likely to be short for most centrist-populist parties, since greater public visibility following their initial success often triggers either the adoption of a more concrete political agenda—thereby moving them into either the mainstream or the more radical party categories—or a gradual loss of electoral appeal. Even though some countries have already witnessed the rise of successive new/centrist-populist parties (in Latvia, the New Party (jp) in 1998 and the New Era Party (jl) in 2002, for example), the marginally significant negative effect of EU membership in model 1 suggests that the appeal of such civilized challenges to mainstream parties may decline more broadly once EU accession constraints are removed. Meanwhile, model 2 shows that support for radical-left parties de- 73 Evans and Whitefield 1993. 74 I calculate these ethnic diversity indicators using census data from the different countries, a variety of data sources presented in Eberhardt 2003. 75 Tavits 2005. 76 Since this measure is very highly correlated with urbanization, it also taps into the broader issue of modernization and its relationship to party system and democratic consolidation. throwing ou t the bums 245 clined dramatically after the initial elections as voters predominantly rejected unreformed communists in favor of their reformed counterparts since neoliberal economics were effectively the only game in town. However, radical-left parties started to rebound during thirdgeneration elections (though the effect fell just short of statistical significance), as seen in the solid performance of the radical left in Latvia, the Czech Republic, and Poland. On the other hand, the negative coefficients for the first- and third-generation election dummies in models 3 and 4 suggest that the vote for extreme-nationalist and nationalist parties peaked during the second-generation elections and declined afterwards, perhaps in response to the fairly consistent Western stance against nationalist parties and politicians such as Slovakia’s Vladimir Meciar. However, these effects also fell short of achieving statistical significance, and the downward trend may be reversed after EU accession, as suggested by the positive and significant effect of EU membership on electoral support for nationalist parties in models 3 and 4. The electoral performance of radical political parties in third-generation elections is arguably understated by the regressions in models 2 and 3, since these models do not account for the significant rise of new/centrist-populist parties that compete with radical parties for the support of protest voters. This substitution hypothesis is confirmed by model 5, which shows that higher vote shares for new/centrist populists are associated with a significant drop in the combined electoral support for radical-left and radical-right parties. More importantly, for the current analysis the combined electoral support for radical political parties experienced a marginally significant increase in third-generation elections once we control for new/centrist-populist support. Given that this article is primarily concerned with the overall evolution of protest voting in postcommunist democracies, what matters more than the predominant type of uop is the evolution of the overall level of support for unorthodox parties. Once the different types of uops are aggregated, there is a much clearer picture of the temporal patterns of protest voting in Eastern Europe. Thus, according to model 6, the combined electoral support for unorthodox parties declined significantly after the founding elections but rebounded significantly during third-generation elections. These findings confirm the initial observation that East European party systems have not only stopped consolidating but are actually in the process of deconsolidation, as mainstream parties have lost ground against their unorthodox competitors. To understand the mechanisms of this deconsolidation process, models 7, 8, and 9 focus on the roots and consequences of what is 246 world politics arguably the most important driver of protest voting: citizen dissatisfaction with the governing style of mainstream political parties. As mentioned above, postcommunist elections stand out for the severity with which voters punish incumbent parties. Therefore, it is reasonable to ask whether the rising appeal of uops in recent elections is simply due to the fact that voters have started to lose patience with the shortcomings of their governments even as economic conditions have generally improved. The negative coefficient for third-generation elections in model 7 provides some support for this assertion, but the effect is not statistically significant, suggesting that the recent uop surge is not primarily due to greater voter disappointment with current incumbents. What has changed during third-generation elections is not the prevalence but the destination of protest votes. During the first two election generations, voters disappointed with the status quo could turn to an untried mainstream alternative in their quest to punish the incumbents. However, starting with the third-generation elections, the availability of such untried mainstream alternatives declined dramatically. Voters faced a dilemma: they could either vote for the parties linked to the first generation of postcommunist governments or they could reject all parties associated with the disappointments of the transition and vote for an untried alternative. Since such untried alternatives were much more likely to be either extremist parties excluded from previous governing coalitions or political newcomers without clear ideological platforms, protest votes in third-generation elections should benefit uops to a much greater extent than in earlier electoral contests. To test this theoretical prediction, model 8 includes an interaction term between the change in the governing parties’ combined vote share and the third-generation election dummy. The powerful negative interaction effect in model 8 provides strong empirical support for the central theoretical claim of this article. As illustrated in Figure 2, which plots the predicted uop vote shares based on the regression in model 8, electoral losses by incumbent parties translated into greater support for unorthodox parties during third-generation elections (significant at .001), whereas during the first two election generations such discontent had a substantively small and statistically insignificant effect. Moreover, the results suggest that third-generation elections produce stronger results for uops only when incumbents suffer significant electoral losses: the conditional effect of third-generation elections is significantly negative (at .05) only in elections where the incumbents lost at least 14 percent of total votes77 compared to the previous election.78 In other words, these findings suggest that uops only benefit when incumbents are unpopular and voters have few untried mainstream alternatives to vent their frustration. The results above raise concerns of reverse causation and endogeneity. One may argue that incumbent losses are the result rather than the cause of uop popularity. However, such reverse causation concerns are alleviated by the fact that the unpopularity of incumbents is theoretically prior to the success of uops and by the fact that uop-vote-share and incumbent-vote-share change are only very weakly correlated.79 77 The model specification and the calculation of conditional effects for the interactions in these models follow the prescriptions of Brambor et al. 2006. 78 However, given the strong incumbency disadvantage, this applied to more than half the cases in the sample. 79 Moreover, in the absence of interaction effects with election generation, neither variable is a significant predictor of the other in multiple regressions (results omitted). Figure 2 Incumbent Losses and UOP Voting in Postcommunist Electionsa a Statistical signifance for predictor variable: continuous line (p<.05), dotted line (p>.05). Statistical significance for modifying variable: striped area (p<.05). 62 57 52 47 42 37 32 27 22 17 PredicteduopVoteShare First and second generation Third generation –50 –45 –40 –35 –30 –25 –20 –15 –10 –5 0 Incumbents Vote Share Change (%) throwing ou t the bums 247 248 world politics To address the endogenous nature of incumbent-vote-share change, model 9 uses an instrumental-variable version80 of the regression in model 8, and produces very similar results. Most of the alternative explanations of uop voting tested in Table 3 receive fairly modest and inconsistent support. The negative and statistically significant effect of second-tier EU candidacy status provides some evidence of the impact of EU political conditionality, but the effects are uneven across different uop subcategories. Confirming Ana Milada Vachudova’s81 findings, the picture is even more mixed with respect to the post-accession period, given that the boost for nationalist parties is balanced by the marginally lower support for the radical left and new/centrist populists. pr systems are characterized by greater support for radical-left parties and uop parties more broadly, which may help explain the relative stability and moderation of the political party system in Hungary and (somewhat surprisingly) in Albania. However, other institutional features, such as directly elected presidents, have modest effects,82 and a higher government turnover appears to have actually reduced uop support slightly (possibly because it reduced the clarity of governmental responsibility). Nor is there consistent evidence that uop voting is driven by worse economic conditions, though inflation and unemployment are associated with higher votes for nationalist parties and, to a lesser extent, with higher overall uop support. Finally, ethnolinguistic fractionalization was associated with higher overall uop vote shares but—surprisingly—this effect was driven by greater support for centrist populists rather than for nationalists, as an ethnic cleavage explanation would have predicted. Individual-Level Survey Evidence While the analysis so far suggests that the rising electoral support for unorthodox parties in third-generation postcommunist elections is due at least in part to the redirection of protest votes from mainstream opposition parties toward unorthodox challengers, this section resorts to more fine-grained, individual-level data to provide detailed evidence about the dynamics of protest voting. I use survey data from two dif- 80 In the first-stage regression I use the lagged vote share of incumbents as an instrument for incumbent vote-share changes. This choice is justified by the fact that parties receiving very large vote shares in the preceding election have a greater loss potential (correlated at -.81). Such past popularity has little effect on current uop appeals (correlated at .05). 81 Vachudova 2008. 82 However, this does not necessarily mean that strong presidencies do not affect the shape of the party system given that the region’s strongest presidential systems are not included in the present analysis. throwing ou t the bums 249 ferent sources. The first part analyzes surveys from twelve postcommunist parliamentary elections included in the Comparative Study of Electoral Systems (cses). While cses only includes about one-sixth of the elections in the data set used for the cross-national regressions, it nevertheless covers a reasonable range of second- and third-generation elections from seven East European countries, and thus represents the most appropriate set of comparably worded and designed electoral surveys on which to test the temporal dimension of protest voting and uop support. The second part presents an alternative approach for illustrating the microdynamics of protest voting and describes the result of a survey-based experiment I administered in Bulgaria in July 2008 as part of a nationally representative survey of twelve thousand Bulgarian citizens. Cross-National Survey Findings Since cses includes only twelve postcommunist elections, which undermines the utility of more elaborate hierarchical linear models, the discussion below is based on a simple comparison of voting trends among different subsets of voters in second- and third-generation elections.83 As a first step, the top three rows in Table 4 look at all voters and confirm the broad regional trend whereby the declining electoral fortunes of governing parties seem to benefit uops to a greater extent than mainstream opponents during third-generation elections. However, such a shift to uops does not by itself constitute evidence of protest voting; after all, it could be possible that voters genuinely embraced the electoral platforms of uops. While in the context of a single party it may be possible to detect protest voting by analyzing the extent to which a party’s platform matches the values and preferences of its voters,84 such an approach would be problematic when applied to multiple parties across different countries given the variety and ambiguity of uop electoral appeals.85 Therefore, I chose a different approach and instead tried to identify groups of voters who may be particularly prone to cast a protest vote, i.e., a vote informed primarily by the rejection of other alternatives rather than the endorsement of a given party. If uops scored disproportionate gains among such likely protest voters during 83 See the notes to Table 4 for details about the elections included and the wording of survey ques- tions. 84 For such an analysis in the context of the 2000 Romanian elections, see Pop-Eleches 2001. 85 Moreover, cses does not include sufficiently detailed survey questions tapping into economic policy preferences and ethnic tolerance to allow for such an analysis. 250 world politics third-generation elections, then that would constitute additional support for the theoretical claims of this article. One such category of potential protest voters consists of respondents who report that they are not close to any political party. To the extent that such citizens choose to cast a vote for a given party, it is reasonable to suspect that they do so because they dislike other parties more, not because they strongly endorse the electoral platform as their actual choice. The voting behavior of such nonpartisans reveals a dramatic shift between election generations. During second-generation elections such voters are somewhat less likely to endorse uops than the average voter, and their electoral support for governing and mainstream opposition parties is statistically indistinguishable from the overall electorate. Table 4 Electoral Choices of Disaffected Votersa Second- Third- Generation Generation Probability of Elections Elections Difference Voter type Voting for… (%) (%) (%) All voters government parties 37.1 26.9 –10.2 largest mainstream 28.3 26.1 –2.3   opposition party unorthodox parties 7.4 17.9 +10.5 Voters who do government parties 36.8 21.8 –15   not feel close largest mainstream 28.8 18.8 –9.9   to any partyb   opposition party unorthodox parties 5.2 23.7 +18.4 Voters who think government parties 34.1 17.8 –16.3   that it makes no largest mainstream 28.9 22.0 –6.9   difference who   opposition party   is in powerc unorthodox parties 7.2 22.6 +15.4 Voters who are government parties 31.6 12.0 –19.5   dissatisfied with largest mainstream 30.3 26.9 –3.4   democracyd   opposition party unorthodox parties 9.7 27.2 +17.5 a Elections included: Albania 2005, Bulgaria 2001, Czech Republic 1996, 2002; Hungary 1998, 2002; Poland 1997, 2001; Romania 1996, 2004; and Slovenia 1996, 2004. b Do you usually think of yourself as close to any particular political party? c Some people say it makes a difference who is in power. Others say that it doesn’t make a difference who is in power. Using the scale on this card, (where one means that it makes a difference who is in power and five means that it doesn’t make a difference who is in power), where would you place yourself? d On the whole, are you very satisfied, fairly satisfied, not very satisfied, or not at all satisfied with the way democracy works in [country]? throwing ou t the bums 251 By contrast, in third-generation elections nonpartisans are much more likely than other voters to vote for uops, and this large uop gain comes at the expense of both governing and opposition mainstream parties. Given that nonpartisans represent about 40 percent of voters in thirdgeneration elections and 60 percent of voters in second-generation elections, a simple calculation suggests that nonpartisans account for about half the overall uop voters in third-generation elections and for 60 percent of the uop vote gains between the two election generations. Table 4 also reflects the voting behavior of a second group of respondents with high protest voting potential. This group, which accounts for about 11 percent of the electorate, consists of voters who believe that it makes little or no difference who is in power. Since they presumably do not care about the election outcome for instrumental reasons, it seems reasonable to assume that when such citizens vote, they do so primarily for expressive/rhetorical purposes. Once more, the electoral patterns are consistent with the protest-vote hypothesis, given that uop electoral gains in third-generation elections are significantly greater among such cynical voters, whereas mainstream parties suffer considerably greater losses than among the rest of the electorate. Finally, the last three rows in Table 4 reveal a very similar pattern of shifting electoral support among voters who are dissatisfied with the way democracy works in their country.86 The survey evidence again suggests that whereas in second-generation elections disaffected voters are generally willing to give either governing- or opposition-mainstream parties the benefit of the doubt, in third-generation elections they turn in much larger numbers toward unorthodox challengers. Overall, then, the individual-level public opinion data suggests that the surge of uop support in third-generation elections is concentrated among disaffected voters with weak partisan ties, which provides additional evidence for the protest-voting hypothesis advanced in this article. Experimental Findings While the combination of macro- and micro-level cross-national evidence presented so far provides consistent empirical support for the protest-voting hypothesis, cross-national statistical tests nevertheless 86 A survey question about the performance of the incumbent government would have been more appropriate, but unfortunately this question was only asked in the second cses wave, which included only one second-generation election. However, satisfaction with democracy was rather highly correlated with government satisfaction (at .42), so this question provides the best available proxy. Moreover, satisfaction with democracy was only correlated at .29 with support for democracy, which suggests that the question is not primarily capturing antidemocrats (who may be more prone to vote for uops). 252 world politics suffer from several well-known limitations, including the cross-national validity of concepts and indicators, and endogeneity concerns. One way to address such issues is through the use of experiments that attempt to create a fictional electoral environment and then vary certain parameters for different randomly selected groups of respondents in an effort to isolate the effects of these factors. To test the dynamics of protest voting in postcommunist countries I designed a survey experiment that was administered as part of a broader public opinion survey administered to twelve thousand Bulgarian citizens in July 2008 by Alpha Research, a Bulgarian survey research firm. The respondents were randomly assigned to four groups, a, b, c, and d (each with three hundred respondents); each group was given a variant of the survey question as follows: Group A: Suppose that after next year’s elections your favorite party became part of the government and several of its ministers are accused of corruption by the media. Then a few months before the following elections a new party appears and its leader promises to clean up Bulgarian politics. There is little information about what the new party stands for otherwise except that its leader supports Bulgaria’s further EU integration. What would your most likely action be in that election? Group B: Suppose that after next year’s elections your favorite party became part of the government and several of its ministers are accused of corruption by the media. Then a few months before the following elections a new party appears and its leader promises to clean up Bulgarian politics. There is little information about what the new party stands for otherwise except that its leader is critical of Bulgaria’s further EU integration. What would your most likely action be in that election? Group C: Suppose that after next year’s elections your favorite party became part of the government and several of its ministers are accused of incompetence by the media. Then a few months before the following elections a new party appears and its leader promises to clean up Bulgarian politics. There is little information about what the new party stands for otherwise except that its leader supportsBulgaria’s further EU integration. What would your most likely action be in that election? Group D: Suppose that after next year’s elections your favorite party became part of the government and several of its ministers are accused of incompetence by the media. Then a few months before the following elections a new party appears and its leader promises to clean up Bulgarian politics. There is little information about what the new party stands for otherwise except that its leader is critical of Bulgaria’s further EU integration. What would your most likely action be in that election? throwing ou t the bums 253 Each group was to select responses from the following choices: 1. I would vote again for my old party choice. 2. I would vote for the new party. 3. I would vote for one of the existing opposition parties. 4. I would not vote. 5. DK/NA The rationale for this question was to test voter response to alleged shortcomings of their preferred parties once those parties are in government.The experiment varies the nature of the shortcoming (corruption versus incompetence) and one element of the otherwise intentionally vague platform of the new challenger party (support versus criticism of greater EU integration). The second variation largely mimics the difference between the generally pro-EU new/centrist populists and the generally Euroskeptic radical unorthodox parties. Two additional elements are common to all four versions of the question. The first— media allegations against the party’s ministers—is a common occurrence in Bulgarian politics and thus hardly represents a very powerful treatment (unlike, say, having party ministers convicted of corruption charges). The second—the promise to clean up Bulgarian politics—is a quasi-universal part of the appeal of new political parties. Table 5 summarizes for each of the four versions the proportion of likely voters who indicated that in the subsequent election they would switch to either the new challenger party or to one of the preexisting opposition parties. While overall percentages need to be taken with a grain of salt since such survey experiments offer a very stylized version of an electoral campaign environment, it is worth noting the fairly high potential support for the new challenger party, which attracted significantly more previous governing party supporters than did existing opposition parties in two of the four scenarios. This powerful appeal is all the more remarkable since respondents were given very few positive reasons to support the new party beyond the usual anticorruption promises (there was no mention of charismatic leaders or populist spending promises, for example). However, this is not to say that actual platforms are irrelevant. As the difference between groups a and c on one hand and b and d on the other hand suggests, new challengers fare significantly better if they do not adopt a radical platform that goes against the preferences of the majority of voters. In fact, the responses from group A suggest that as long as the newcomers adopt vague and safe platforms 254 world politics and their anticorruption rhetoric is reinforced by media allegations of high-level corruption, they can woo the vast majority of disgruntled voters. Thus, the survey experiment findings reveal a high degree of risk-seeking attitudes by Bulgarian voters, which may explain the strong performance of extraparliamentary parties in recent elections, including the spectacular victory of the ndsv in 2001 and the nationlist party Ataka’s strong showing in 2005. IV. Discussion and Conclusion In this article I develop a theoretical framework for understanding the puzzling surge in voting for unorthodox parties in recent East European elections. I argue that in order to understand the overall electoral support for such parties there is a need to analyze the widespread practice of protest voting in the context of different generations of postcommunist elections. The empirical patterns of uop support in seventy-six parliamentary elections in fourteen countries from 1990–2006, combined with micro-level evidence in twelve second- and third-generation elections, confirm the importance of paying closer theoretical attention to election sequence following re-democratization. Whereas in first- and second-generation elections the anti-incumbent bias of East European voters generally benefitted mainstream opponents and thereby contributed to healthy power alternation, in third-generation elections (i.e., elections occurring after two or more distinct political camps have governed in the postcommunist period) voter disaffecTable 5 Vote Choices in Bulgaria Support for Alleged Preexisting Governance EU Integration Support for New Opposition Problem of Stance of New Party among Parties among Respondent’s Challenger Likely Voters Likely Voters Group Favorite Party Party (%) (%) A corruption supportive 22.3 6.4 B corruption critical 11.1 11.1 C incompetence supportive 15.8 6.1 D incompetence critical 5.6 11.2 A and C any supportive 18.8 6.2 B and D any critical 8.3 11.2 Source: Author’s public opinion survey in Bulgaria 2008. throwing ou t the bums 255 tion with incumbents starts to benefit unorthodox parties. Therefore, the apparent party system stabilization of second-generation elections gives way to renewed instability as mainstream political parties lose significant ground to unorthodox challengers despite the region’s economic and European integration progress. The implications of these findings for an assessment of the future of democracy in Eastern Europe are mixed. The good news is that, contrary to the impression conveyed by some recent elections, East European voters have not suddenly en masse turned into fascists and communists. Rather, many voters have merely endorsed uops to punish mainstream elites for their often incompetent and corrupt governing style. Indeed, the main beneficiaries of voter discontent during thirdgeneration elections are neither radical-left nor nationalist parties, but a new breed of centrist-populist parties that avoid extremist policy positions in favor of vague or even explicitly anti-ideological appeals focusing on the personal qualities of their leaders. Moreover, the findings from the 2008 Bulgarian survey experiment suggest that East European voters may continue to endorse such centrist-populist newcomers even after EU accession, in which case the dynamics of instability without radicalization are likely to continue as long as there is a sufficient supply of centrist-populist leaders. However, the new/centrist-populist solution, which allows voters to punish mainstream parties without endorsing various brands of political extremism, is inherently unstable and its long-term implications for postcommunist party systems depend on the subsequent political trajectories of such parties. Given the self-defeating nature of electoral success for the unorthodox appeal of such parties, their leaders have two main options if they hope to escape the return to electoral marginality in subsequent elections. They can either enter the realm of mainstream political competition by adopting a recognizable and moderate ideological platform to complement the personal appeal of the leader, or they can decide to move toward the more radical uop categories by adopting nationalist or anticapitalist themes as key elements of their political platforms. To the extent that the latter strategy prevails—as in the case of the increasing national-populist turn of the Polish Law and Justice Party (pis) since 2005—then new/centrist-populist parties may turn out to be Trojan horses through which radical politicians can position themselves as credible alternatives to the discredited mainstream elite. These concerns are particularly relevant following the reduction of external pressures that comes with EU accession and that could translate into greater leeway for radical political adventures. In 256 world politics the more optimistic scenario, in which new/centrist populists complete their transformation to mainstream political parties, the future of electoral and democratic stability in the region depends on their ability to improve on the governing record of their predecessors and thereby avoid becoming victims of the very anti-incumbency bias that brought them to power in the first place. Unfortunately this seems to be a challenging task, as illustrated by the Bulgarian ndsv: it successfully rebranded itself as a mainstream liberal party during its eight-year stint in government, but in the process suffered a dramatic electoral decline and eventually failed to cross the electoral threshold in the 2009 elections. While developed in the specific context of postcommunist transitions, several aspects of this theoretical approach are applicable to other regions and political contexts. First, the definition of unorthodox parties as deviations along a number of dimensions from the ideal type of the moderate programmatic party can be applied to other regions. In the West European context even the main policy categories—anticapitalism, ethnic nationalism, and perhaps Euroskepticism—should provide a useful analytical starting point. In the Latin American context, ethnic nationalism would have to be reconceptualized given the differences in ethnic and race politics in the region. Moreover, the concept of new/centrist populism may deserve greater attention outside the postcommunist context, since it arguably provides a more accurate description of the initial appeal of populist leaders such as Peru’s Alberto Fujimori than the more widespread neoliberal-populist label.87 Furthermore, while differences in the specific electoral appeal of uops obviously matter, the importance of protest voting in the electoral rise of such parties suggests that these parties should be treated as part of a broader phenomenon of voter rejection of mainstream elites. Second, the concept of election generations can be fruitfully applied to electoral politics in a wide range of new democracies. While some earlier studies have analyzed the dynamics of first-generation elections,88 this article suggests that in assessing the stability of party systems in new democracies closer attention should be paid to the dynamics of third-generation elections. While the very fact of peaceful democratic power turnovers is often considered a mark of democratic consolidation, elections taking place after a rapid succession of turnovers create political opportunities for unorthodox parties that can attract voters disenchanted with the governance failures of different 87 Roberts 1995; Weyland 1999. 88 Darden and Grzymala-Busse 2006. throwing ou t the bums 257 camps of mainstream political parties. While such disenchantment with the mainstream may have been exacerbated by the traumatic nature of the postcommunist transition, it is a rather common phenomenon among third-wave democracies, many of which have struggled with poverty, economic crises, and poor governance. Indeed, Fujimori’s spectacular rise to power in in 1990 occurred precisely in the context of a third-generation election, after both the center-right Accion Popular (ap) and the leftist Alianza Popular Revolucionaria Americana (apra) had disappointed Peruvian voters in the decade since the return to de- mocracy. Finally, this article contributes to a more systematic understanding of protest voting in the postcommunist context and beyond. While protest voting has usually been invoked to explain surprisingly strong support for extremist parties, I argue that protest voting has been a much more widespread phenomenon in Eastern Europe than is usually recognized because when protest votes benefit mainstream political parties, they are interpreted as endorsements of the moderate ideological platforms espoused by such parties. The evidence in this article suggests that what has changed in recent elections is not the extent of protest voting but its particular outlet. Because postcommunist voters had fewer untried mainstream alternatives in third-generation elections, they were more likely to endorse unorthodox parties. Overall, this article’s emphasis on the interaction between protest voting and election sequence calls for a reassessment of postcommunist electoral outcomes. It suggests that the democratic power alternations of the first two generations of postcommunist elections are not really indicative of positive and durable links between mainstream political parties and their social constituents and that the more recent surge in unorthodox party support is not a sign of voter radicalization. Rather, these results indicate that East European countries (along with many other third-wave democratizers) still have a long way to go in pursuit of institutionalized representative democracy along the lines of their (admittedly imperfect) Western role models. References Abedi, Amir. 2004. Anti-Political Establishment Parties: A Comparative Analysis. London: Routledge. Beichelt,Tim, and Michael Minkenberg. 2002. “Right Wing Extremism in Transitional Societies—Incubator Conditions and Explanatory Model.” Osteuropa 52, no. 3: 247–62. 258 world politics Brambor,Thomas, William Roberts Clark, and Matt Golder. 2006. “Understanding Interaction Models: Improving Empirical Analyses.” Political Analysis 14, no. 1: 63–82. Bunce, Valerie. 1998. “Regional Differences in Democratization: The East versus the South.” Post-Soviet Affairs 14, no. 3: 187–211. ———. 2003. “Rethinking Recent Democratization: Lessons from the Postcommunist Experience. World Politics 55, no. 2: 167–92. Capoccia, Giovanni. 2002. “Anti-System Parties: A Conceptual Reassessment.” Journal of Theoretical Politics 14, no. 1: 9–35. Collier, David, and James E. Mahon, Jr. 1993. “Conceptual ‘Stretching’ Revisited: Adapting Categories in Comparative Analysis.” The American Political Science Review 87, no. 4: 845–55. Darden, Keith, and Anna Grzymala-Busse. 2006. “The Great Divide: Literacy, Nationalism, and the Communist Collapse.” World Politics 59, no. 1: 83–115. Evans, Geoffrey, and Stephen Whitefield. 1993. “Identifying the Bases of Party Competition in Eastern Europe.” British Journal of Political Science 23, no. 4: 521–48. Fidrmuc, Jan. 2000. “Economics of Voting in Post-Communist Countries.” Electoral Studies 19, no. 2/3: 199–217. Filippov, Mikhail, Peter Ordeshook, and Olga Shvetsova. 1999. “Party Fragmentation and Presidential Elections in Post-Communist Democracies.” Constitutional Political Economy 10, no. 1: 3–26. Fiorina, Morris P. 1981. Retrospective Voting in American National Elections. New Haven: Yale University Press. Gallagher, Tom. 2000. “Nationalism and Democracy in South-East Europe.” In Geoffrey Pridham and Tom Gallagher, eds., Experimenting with Democracy: Regime Change in the Balkans. London: Routledge. Ganev, Venelin. 2007. Preying on the State: The Transformation of Bulgaria after 1989. Ithaca, N.Y.: Cornell University Press. Grzymala-Busse, Anna. 2002. Redeeming the Communist Past: The Regeneration of Communist Parties in East Central Europe. New York: Cambridge University Press. Hellman, Joel S. 1998. “Winners Take All: The Politics of Partial Reform in Postcommunist Transitions.” World Politics 50, no. 2: 203–34. Hooghe Liesbet, Gary Marks, and Carole Wilson. 2002.“Does Left/Right Structure Party Positions on European Integration?” Comparative Political Studies 35, no. 8: 965–89. Innes, Abby. 2002. “Party Competition in Postcommunist Europe—The Great Electoral Lottery.” Comparative Politics 35, no. 1: 85–104. Ishiyama, John. 1997. “The Sickle or the Rose? Previous Regime Types and the Evolution of the Ex-Communist Parties in Post-Communist Politics.” Comparative Political Studies 30, no. 3: 299–33. Kitschelt, Herbert. 1992. “The Formation of Party Systems in East Central Europe.” Politics and Society 20, no. 1: 7–5. ———. 1995. The Radical Right in Western Europe: A Comparative Analysis. Ann Arbor, Mich.: University of Michigan Press. Kitschelt, Herbert, Zdenka Mansfeldova, Radoslaw Markowski, and Gabor Toká. 1999. Post-Communist Party Systems: Competition, Representation, and InterParty Cooperation. Cambridge: Cambridge University Press. throwing ou t the bums 259 Krastev, Ivan. 2002. “Democracy Without Choices.” Journal of Democracy 13, no. 3: 39–53. Krause, Kevin. 2003. “Slovakia’s Second Transition.” Journal of Democracy 14, no. 2: 65–79. Kreuzer, Marcus, and Vello Pettai. 2003. “Patterns of Political Instability: Affiliation Patterns of Politicians and Voters in Post-Communist Estonia, Latvia, and Lithuania.” Studies in Comparative International Development 38, no. 2: 76–98. Lewis, Paul. 2000. Political Parties in Post-Communist Eastern Europe. New York: Routledge. Lewis-Beck, Michael, and Mary Stegmaier. 2000. “Economic Determinants of Electoral Outcomes.” Annual Review of Political Science 3: 183–219. Miller, Arthur, Gwyn Erb, William Reisinger, and Vicki Heslie. 2000. “Emerging Party Systems in Post-Soviet Societies: Fact or Fiction.” Journal of Politics 62, no. 2: 455–91. Mosteller, Frederick, and John W. Tukey. 1977. Data Analysis and Regression. Reading, Mass.: Addison-Wesley. Mudde, Cas. 2000. “Populism in Eastern Europe—Part I.” East European Perspectives 5, no. 8: 1–10. ———. 2001. “In the Name of the Peasantry, the Proletariat, and the People: Populisms in Eastern Europe.” East European Politics and Societies 14, no. 2: 33–53. Nadeau, Richard, and Michael Lewis-Beck. 2001. “National Economic Voting in U.S. Presidential Elections.” The Journal of Politics 63, no. 1: 159–81. Olson, David. 1998. “Party Formation and Party System Consolidation in the New Democracies of Central Europe.” Political Studies 46, no. 3: 432–64. Orenstein, Mitchell. 1998. “A Genealogy of Communist Successor Parties in East Central Europe and the Determinants of their Success.” East European Politics and Societies 12, no. 3: 472–99. Pacek, Alexander. 1994. “Macroeconomic Conditions and Electoral Politics in East Central Europe.” American Journal of Political Science 38, no. 3: 723–44. Pacek, Alexander, Grigore Pop-Eleches, and Joshua Tucker. 2009. “Disenchanted or Discerning: Voter Turnout in Post-Communist Countries.” The Journal of Politics 71, no. 2: 473–91. Pop-Eleches, Grigore. 2001. “Romania’s Politics of Dejection.” Journal of Democracy 12, no. 3: 156–69. Powell, Eleanor N., and Joshua A. Tucker. 2008. “New Approaches to Electoral Volatility: Evidence from Postcommunist Countries.” Annual Meeting of the Midwest Political Science Association, Chicago, April 3–6. Powell, G. Bingham, and Guy D. Whitten. 1993. “A Cross-National Analysis of Economic Voting: Taking Account of the Political Context.”American Journal of Political Science 37, no. 2: 391–414. Pridham, Geoffrey. 2002. “EU Enlargement and Consolidating Democracy in Post-Communist States: Formality and Reality.” Journal of Common Market Studies 40: 953–73. Ramet, Sabrina P. 1999. The Radical Right in Central and Eastern Europe since 1989. University Park, Penn.: Pennsylvania State University Press. Roberts, Andrew. 2008. “Hyperaccountability: Economic Voting in Central and Eastern Europe.” Electoral Studies 27, no. 3: 533–46. 260 world politics Roberts, Kenneth. 1995. “Neoliberalism and the Transformation of Populism in Latin America: The Peruvian Case.” World Politics 48, no. 1: 82–116. Rose, Richard. 2002. “Advancing into Europe? Contrasting Goals of Post-Communist Countries.” In Adrian Karatnycky, Alexander Motyl, and Amanda Schnetzer, eds., Nations in Transit 2001–2002. New Brunswick, N.J.: Transaction Publishers. Sartori, Giovanni. 1976. Parties and Party Systems. Cambridge: Cambridge University Press. Schedler, Andreas. 1996. “Anti-Political-Establishment Parties.” Party Politics 2, no. 3: 291–312. Shafir, Michael. 1997. “Romania’s Road to ‘Normalcy.’” Journal of Democracy 8, no. 2: 144–58. ———. 2001. “Reds, Pinks, Blacks and Blues: Radical Politics in Post-Communist East Central Europe.” Studia Politica: The Romanian Political Science Review 1, no. 2: 397–446. Steenbergen, Marco, and Gary Marks. 2007. “Evaluating Expert Surveys.” European Journal of Political Research 46, no. 3: 347–66. Tavits, Margit. 2005. “The Development of Stable Party Support: Electoral Dynamics in Post-Communist Europe.” American Journal of Political Science 49, no. 2: 283–98. Tismaneanu, Vladimir. 1998. Fantasies of Salvation: Democracy, Nationalism, and Myth in Post-Communist Societies. Princeton: Princeton University Press. Toka, Gabor. 1998. “Party Appeals and Voter Loyalty in New Democracies.” Political Studies 46, no. 3: 589–610. Tucker, Joshua. 2001. “Economic Conditions and the Vote for Incumbent Parties in Russia, Poland, Hungary, Slovakia, and the Czech Republic from 1990– 1996.” Post Soviet Affairs 17, no. 4: 309–31. ———. 2005. “Red, Brown, and Regional Economic Voting: Russia, Poland, Hungary, Slovakia, and the Czech Republic from 1990–99.” Annual Meeting of the Midwest Political Science Association, Chicago, April 7–10. ———. 2006. Regional Economic Voting. New York: Cambridge University Press. Ucen, Peter, Olga Gyarfasova, and Vladimir Krivy. 2005. “Centrist Populism in Slovakia from the Perspective of Voters and Supporters.” Slovak Foreign Policy Affairs 6, no. 1: 28–46. Vachudova, Ana Milada. 2008. “Tempered by the EU? Political Parties and Party Systems before and after Accession.” Journal of European Public Policy 15, no. 6: 861–79. Van der Brug, Wouter, Meindert Fennema, and Jean Tillie. 2000. “Anti-Immigrant Parties in Europe: Ideological or Protest Vote?” European Journal of Political Research 37, no. 1: 77–102. Weyland, Kurt. 1999. “Neoliberal Populism in Latin America and Eastern Europe.” Comparative Politics 3, no. 4: 379–401. Whitefield, Stephen. 2002. “Political Cleavages and Post-Communist Politics.” Annual Review of Political Science 5: 181–200. T H E C O N T R I B U T O R S Eric C. C. Chang is an associate professor of political science at Michigan State University. He is a coauthor (with Mark Kayser, Drew Linzer, and Ronald Rogowski) of Electoral Systems and the Balance of Consumer-Producer Power (forthcoming), and he has published articles in numerous journals. He can be reached at echang@msu.edu. Miriam A. Golden is a professor of political science at the University of California at Los Angeles and serves as chair of the American Political Science Association’s Organized Section in Political Economy.  She is a coeditor (with David Austen-Smith, Jeffry Frieden, Karl Ove Moene, and Adam Przeworski) of Selected Works of Michael Wallerstein: The Political Economy of Inequality, Unions, and Social Democracy (2008) and the author of numerous recent journal articles. Her current research is on distributive politics, political corruption, and political capture, especially in less developed nations. She can be reached at golden@ucla.edu. Seth J. Hill is a doctoral candidate in the department of political science, University of California, Los Angeles. He is completing a dissertation on the effects of changes in who turns out to vote on American election outcomes, and he has published his work in various journals. He can be reached at sjhill@ucla.edu. Grigore Pop-Eleches is an assistant professor of politics and international affairs at the Woodrow Wilson School at Princeton University. He is the author of From Economic Crisis to Reform: IMF Programs in Latin America and Eastern Europe (2009). He is currently working on a book about the impact of communist legacies on post­communist attitudes and political behavior, as well as on a project about the interaction between elections and structural conditions in driving political liberalization and deliberalization episodes since the end of the cold war. He can be reached at gpop@princeton.edu. Steffen Hertog is a professor at the Chaire Moyen-Orient Méditerranée at Sciences Po Paris and a lecturer in political economy at Durham University. He is the author of Princes, Brokers and Bureaucrats: Oil and State in Saudi Arabia (2010) and (with Diego Gambetta) of Engineers of Jihad (forthcoming). He can be reached at shertog@gmx.de. Carlos Gervasoni is an assistant professor of political science and international studies at Universidad Torcuato Di Tella in Buenos Aires and a doctoral candidate in political science at the University of Notre Dame. He is completing a dissertation on measuring and explaining variance in levels of subnational democracy in Argentina. Other recent and current research includes work on party discipline in Brazil (with Frances Hagopian and Juan Andrés Moraes), extra system electoral volatility (with Scott Mainwaring and Annabella España Nájera), and the dampening effects of democracy on economic crises (with Leslie Elliott Armijo). He can be reached at cgervasoni@utdt.edu. Monika Nalepa is an assistant professor of political science at the University of Notre Dame. Her research interests include institutions of transitional justice and political transitions. She is the author of Skeletons in the Closet: Transitional Justice in Post-Communist Europe (2010) and is crurrently working on a project on postcommunist legislative institutions. She can be reached at mnalepa@nd.edu. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.