Party System Stabilisation during Times of Financial Crisis: The Paradox of the Baltic States

Transcription

1 Party System Stabilisation during Times of Financial Crisis: The Paradox of the Baltic States Mažvydas Jastramskis Institute of International Relations and Political Science Vilnius University Paper prepared for presentation at the 8th ECPR General Conference, Panel # 208, National Electoral Consequences of the Financial Crisis, Glasgow, Scotland, September 3-6, 2014 DRAFT: PLEASE DO NOT QUOTE WITHOUT AUTHOR S PERMISSION Abstract. Because of the salience of economic voting (against incumbents) and protest voting (against established parties), economic crises can transform the party systems and destabilize the electoral trends. However, there are cases that contradict such expectations: despite the increase of unemployment and recession, financial crisis did not push the party systems in the Baltic States towards increasing electoral instability. In contrary, some stabilization trends were observed: incumbents achieved satisfactory results and electoral volatility stayed at moderate levels. This paper argues that economic crisis provides an opportunity to break a vicious circle of electoral volatility, government instability and changing supply side of the party system. Budget cuts and the risk of electoral demise devalue the spoils of office: since little incentive exists for government reshuffling and foundation of new parties, political competition becomes more structured. Analysis of individual level data from 2008 and 2012 parliamentary elections in Lithuania supports these macro-level considerations.

2 Introduction Since Aristotle said that poverty is a father of revolution (Aristotle, 1962) economic downturns have been treated as a major factor contributing to the political instability. This is especially relevant in the context of democratic elections since economic voting theory (Lewis-Beck and Paldam, 2000; Lewis-Beck and Stegmaier, 2000) presupposes that voters hold the government responsible for the economy, sharper downturns have a power to destabilize the electoral trends and produce unbalanced electoral outcomes. Extent of larger losses of government parties, larger gains by opposition could go a long distance, as in the 2010 elections of Hungarian parliament: vote share of the governing socialists dropped by percent 1 and oppositional party FIDESZ got a super majority in the parliament. Economic crises can even transform the party systems through protest voting, if traditional parties are seen as similar and altogether blamed. Economic hardships in the Veimar Respublic and voting for antisystem parties (altogether with the rise of National Socialist German Workers Party) is probably a best known example of such a relationship, studied in history literature (Stogbauer, 2001), political science (Pelizzo and Babones, 2007) and even geography (O Loughlin et al., 1994). Though recent financial crisis in Europe (still) did not produce anti-democratic political revolutions, negative impact on the party systems is clearly visible in a handful of cases, particularly in the South countries. Economic voting, especially the valence questions contributed to (Lobo, 2013) or could strongly affect (judging from previous studies: Frail and Lewis-Beck, 2013) the losses of governing parties in Spain (socialists lost almost half of their support in 2011) and Portugal (in 2011 socialists got less than 30 percent of vote for the first time since 1991). Furthermore, Greece stands as as archetypical case of economic downturn induced protest voting and transformation: stable two-party system crumbled in the earthquake elections of 2012 May, when two dominant parties New Democracy and PASOK lost 45.3 percentage points of votes (in comparison to the 2009 elections). 1 If not specified otherwise, data on the election results is taken from European Election Database (Norwegian Social Science Data Services). Available from: [Accessed at 15 July 2014]. 2

3 However, in some countries different trend appeared. Latvia s and Estonia s prime ministers managed to stay in power even after the elections that were held during the crisis (Moulds, 2012). Despite of low ratings, the dominant government party in Lithuania HU-LCD (Homeland-Union Lithuanian Christian Democrats) won the lowturnout elections to European Parliament in Moreover, in 2012 parliamentary elections Lithuanian governing coalition managed to become the most successful government electorally since the transition this all happened despite that it ruled for four years since the beginning of the crisis in 2008 and had to apply unpopular austerity measures (Vilpišauskas et al., 2014). Finally, in all the three Baltic countries electoral volatility stayed at moderate levels. Such trends are even more striking when we take into account that post-communist democracies are infamous for the hyper accountability of their electorates (Roberts, 2008). Why Baltic party systems became more stable exactly during the period of the crisis, when in South countries different trends were visible? To answer this question, three aspects are needed: context, theory, some empirical prove. Three sections of this paper cover these aspects. First part is devoted to the political context and electoral trends in the Baltic States since the onset of economic crisis, Second section proposes a theoretical argument that economic crisis provides an opportunity to break a vicious circle of electoral volatility, government instability and changing supply side of the party system. Third, some empirical support is given to the argument at individual level in the analysis of 2008 and 2012 Lithuanian post-electoral surveys. Baltic Party Systems and Economic Crisis Any comparative analysis of the party systems in three Baltic States Lithuania, Latvia, Estonia should begin with a reservation that they are not a homogenous cluster. There are important differences to be considered, regarding the origins of party system, competition and cleavage structure. Unlike Lithuania, Latvia and Estonia avoided the emergence of successfully reformed ex-communist parties (Mikkel, 2006; Pabriks and Stokenberga, 2006). Even though significant parts of population were not awarded citizenship at the onset of independence, ethnic composition became the main 3

4 force structuring party systems in these countries, with ethnic cleavage being dominant (Linz and Stepan, 1996; Saarts, 2011). On other hand, party system formation in Lithuania has been more similar to Poland and Hungary: cultural homogeneity and legacy of national communism allowed a successful reformation of the former communist party (Kitschelt, 1995; Kitschelt, 1999; Norkus, 2008; Norkus, 2011), with right left identities aligning among the anticommunist ex-communist cleavage (Ramonaitė, 2007). There have been a handful of mergers and splits since those times, and as in the neighbouring Poland ex-communist identities in Lithuanian party system are quite blurred. Despite this, evaluation of the former communist regime in comparison to current (directly connected to the changes in the social and financial status) is still the strongest factor in the explanations of voting behaviour (Ramonaitė et al, 2014). However, qualitative differences in the party system structure do not refute one similarity: party systems in Baltic States are quite unstable and volatile at the electoral level (Pabriks and Stokenberga, 2006; Mikel, 2006; Jastramskis, 2010; Saarts, 2011). Though the Estonian case was a relative exception in some elections (Saarts, 2011), on average it still exhibits more instability when compared to the majority of Visegrad countries (other significant region of post-communist EU members), not to mention the Western democracies. Party systems in Latvia and Lithuania could be regarded as antirecord makers in the European Union (EU), having volatility over 40 during the (Jungerstam-Mulders, 2006; Sikk, 2006) and at some elections even reaching 50 (Jastramskis, 2010; Saarts, 2011). Table 1 presents comparisons of electoral volatility (measured by Pedersen s index: Pedersen, 1979) levels (region averages of country means) across the various studies (to control for different assumptions about party splits and mergers) and different time periods (for judgement not to be time-dependent). On average, cluster of Baltic States is the most unstable regardless of the study or time period. To be certain that arithmetic mean is at least relatively representative, minimum (Baltics) and maximum (Visegrad and EU-15) cases were added to the table. It seems that the most stable party system (Estonia) in the Baltic states reaches similar level of the most unstable case (which interestingly varies across the studies) in Visegrad: this is not true only in the very recent study by Powell and Tucker (2014), which brings us to our argument why this 4

5 happened in the context of economic crisis? Generally, the upper bar of electoral volatility in non-soviet post-communist EU democracies is approximately in the same level where the lower bar of Baltic States (Estonia) lies. Without any doubts, Baltic States is the most electorally unstable region in the EU. Table 1. Electoral volatility in three regions: comparison of different authors and periods Lane and Mainwaring and Ersson Powel and Ersson (2007): Zoco (2007): from (2012) Tucker from 1990s the inauguration of (2014):1989- onwards democracy 2009 Baltic states Baltic states: 32.2 (Estonia) 45.4 (Estonia) (Estonia) min Visegrad Four plus Slovenia Visegrad 27.8 (Poland) 46.6 (Poland) (Slovakia) Four plus (Slovenia) Slovenia: max EU EU-15: max 24.3 (Italy) 15.3 (France) (France) - What is expected of the party systems, both stable (Western) and unstable (Baltics) when crisis hit? The obvious answer is given by economic voting theory, both in it s pure form (responsibility hypothesis) and extensions to the electoral volatility research. First, defeat of incumbents at after-crisis or during-crisis elections is expected. Since the post-communist states are known for the hyper accountability of their electorates (Roberts, 2008), larger losses of incumbents should be expected here. Second, previously well researched relationship between total electoral volatility and country s economic situation prompts to predict a rise of electoral instability (Mainwaring and Zoco, 2007; Tavits, 2005; Roberts and Wibbels, 1999). Third, since it is known that declining economy fuels protest voting and success of new and marginal parties, extrasystemic volatility (Powel and Tucker, 2014; Sanchez, 2009; Birch, 2003) can also be expected. As in the case of incumbent performance in elections, total volatility and 5

6 protest voting (extra-systemic volatility) should be of greater extent in the inherently unstable party systems, as in the Baltic States. Research on the electoral consequences of recent financial crisis is still in the making, as exemplified by this panel. However, existing attempts (Weschle, 2014; Kreis and Hernandez, 2013; Bosco and Verney, 2012) reveal that exactly those trends as were theorized in the previous paragraph appeared. First, across the Europe incumbents experienced heavier defeats (on average) than in normal times. Second, during and after crisis electoral volatility increased both in the Western and Eastern European countries. Thirdly, support of mainstream parties decreased with new parties gaining more electoral support, especially in the Central and Eastern Europe. Destabilizing trends are also quite emphasized in the case of Southern Europe (Bosco and Verney, 2012). Interestingly, until now Baltic States somehow escaped more thorough analysis of electoral consequences of the financial crisis. As mentioned in the introduction, this article tries to fill this gap and is based on an intriguing empirical observation Baltic States stand out as a relative exception to the European level trends, being a region where party systems actually became more stable during the crisis. This exception is best seen if we put these countries into a comparative perspective within their own party systems, not across the Europe. In order to emphasize this phenomenon, Table 2 sums up the most important indicators of electoral instability before, during and after the crisis in the three Baltic States. Regional averages before (up to 2008) and during (or after) the economic crisis ( ) indicate moderate increases in the electoral volatility however, this trend is heavily influenced by Latvia s indicators. To be more precise, in Estonia electoral volatility significantly decreased to a level (13.8) that would be considered quite normal even in the Western democracies (according to calculations by Kriesi and Hernandez (2013), electoral volatility has risen from 9.7 to 15.7 during the crisis in the Western Europe). In Lithuania, electoral volatility stayed at the same level as it was before crisis it was first time when Pedersen s index in this country remained at 30 for two parliamentary elections in a row. Finally, in the same Latvia where electoral volatility has risen, this increase is quite artificial: last pre-crisis parliamentary elections in 2006 were a notable exception (before them none of elections had volatility lower than 40 6

7 (Saarts, 2011) 2. Therefore, judging from Pedersen index, electoral volatility in the Baltic States during the crisis stayed at (relatively) moderate levels (contrary to the theoretical expectations). Table 2. Electoral instability in the Baltic states, before and during the crisis. Before-crisis elections: 2006 in Latvia, 2007 in Estonia, 2008 in Lithuania. During and after-crisis elections: 2010 and 2011 in Latvia, 2011 in Estonia, 2012 in Lithuania. Lithuania Latvia Estonia Region average Electoral volatility Before crisis During crisis (1) During crisis (2) ** Difference * Incumbent share of votes (loss) Before crisis (-12.32)* (-3.07) 61 (4.9) (-3.5)*** During crisis (1) (-20.17) (-1.93) 49.1 (3.4) During crisis (2) - 31 (-19) (-9.4)** Difference (-7.85) (-7.39) (-1.5) (-5.9) Success of new parties Before crisis During crisis (1) During crisis (2) ** Difference * * Average of elections during the crisis minus elections before crisis. ** Average of all four elections in the Baltic States during the crisis. *** Number in brackets show the gains (losses) of incumbents comparing with previously held elections. Sources: author s calculations, election data from The Central Electoral Commission Of The Republic Of Lithuania, Central Election Commission of Latvia, Estonian National Electoral Committee. Another important indicator of electoral instability is the success of new parties, both as a measure of extra-systemic volatility and protest voting against established parties. On average, vote share of new parties decreased in the Baltic States during the crisis. Estonia stands out at 2011 parliamentary elections none of the votes were cast for the new parties there (to be more exact, there were none of new parties to cast a vote for). In Lithuania new parties achieved moderate success (12.3 percentage of votes), but it still was a significant decrease in a country where each of the previous three elections 2 These elections cease to be an exception it we treat Concord Center (founded in 2005, splinter from Human Rights in United Latv) as a new party in such a case volatility in 2006 Latvian parliamentary election rises to

8 featured a new party winning either a second place (2000 and 2008) or even first (2004) according to the vote share. Finally, Latvia stands out as a moderate deviation of the region with an increase of vote share received by new parties. This trend is mostly due to the premature parliamentary elections of 2011 where ex-president s Valdis Zatlers s newly founded Reform Party finished second according to the received votes and parliament seats. Lastly, performance of incumbents in the Baltic States varied quite starkly depending on country and election. In Estonia, coalition of two centre-right parties got 9.4 percentage of votes less than previous incumbents however, they managed to increase their own vote share (by 3.4) and such a performance ensured them majority in the parliament. Oversized coalition of five parties in Latvia managed to win the 2010 elections, but shortly after that cabinet was reshuffled. Two remaining parties (Unity and Union of Peasants and Greens) lost 19 percent of votes in the premature event of It is interesting that despite the victory of Concord Center (party supported by Russian speaking minority), prime minister Valdis Dombrovskis managed to remain in the position. The fact that Latvia s and Estonia s prime ministers which were in power during the crisis got re-elected got some attention internationally (Moulds, 2012). Our data suggests that genuine re-election happened only in the case of Estonia - in Latvia, art of coalition crafting ensured the continuing reign of Dombrovskis. Still, on average incumbents in Latvia achieved at least satisfactory results, especially if compared to Lithuanian counterparts who at first glance appear as the biggest losers in the Baltic States: gathering only percent of votes in the 2012 parliamentary elections and losing percentage points, if compared to their performance in However, this impression does not hold when several factors are taken into account. First, cabinet of Andrius Kubilius was the only one in the Baltic States that stayed in power for a full four year term, sworn in before the crisis. Second, losses were inflated by the abrupt electoral demise of Rising Nation Party, a newcomer whose vote share shrunk (mainly due to the scandals inside the party) from percent of votes in the 2008 parliamentary to 1.04 percent in the 2009 European elections. Lastly, centre right coalition managed to become the most electorally successful Lithuanian government 3 Rising Nation Party merged with Liberal and Center Union in 2011, effectively reducing the number of coalition partners from four to three. 8

9 since the transition to democracy in 1990 (Ramonaitė et al., 2014). Figures of vote share received by incumbents, opposition and other parties in the Lithuanian parliamentary elections are presented in the Figure % 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Incumbents Other Opposition Figure 1. Vote share of incumbents, opposition and other parties (marginals and new parties) in Lithuanian parliamentary elections, Source: author s calculations, The Central Electoral Commission Of The Republic Of Lithuania. Paradoxical success of Lithuanian crisis government becomes more apparent if we consider the previously discussed nature of Baltic party systems and also the hyper accountability of post-communist electorates. Differently from Estonia and Latvia, Lithuanian party system lacks an ethnic cleavage and strong Russian population-based party on the left there is no soft cushion for the right wing politicians to land on and no justification for the almost permanent centre-right coalitions. Also, Lithuania has always been a paradigmatic case of electoral hyper accountability: none of the governing coalitions until 2012 succeeded in receiving at least a quarter of votes. Figure 1 suggests that trend of decreasing hyper accountability could be observed before the 2012 elections: especially the difference between 2008 and 2012 elections (Figure 1) looks insignificant, with Gediminas Kirkilas government ( ) receiving and Andrius Kubilius government percent of votes. However, it must be again emphasized that this happened after the four year of governing under the economic 9

10 crisis: previously discussed theory suggest that incumbents should be crushed and protest voting should rise. None of this happened in Lithuania. To sum up the discussed electoral trends in the Baltic States during the crisis: stabilization in Estonia, staying at (relative to the context) moderate levels in Lithuania and slight decrease of stability in Latvia. We see that the pattern is of either reduced instability or staying in the similar levels as before the crisis. Even if it could be argued that presented numbers do not indicate party system stabilization in it s pure form (Estonian case being the closest), the fact that Baltic party systems did not became more unstable is worth exploring. These observations are in odds with the previously discussed theory of relationship between elections and economic downturns unless the Baltic States would be the ones untouched by the financial crisis. Interestingly, quite the opposite is true. Lithuania, Latvia and Estonia experienced sharpest declines of economy not only in EU, but in the whole world. In all three Baltic States were among the top five nations in the world according to the GDP contraction (Vilpišauskas and Kuokštis, 2010): Latvia having the largest decline in the world, Estonia coming second and Lithuania (fourth place) only is being surpassed by Ukraine. To counter the crisis all three countries chose a policy of internal devaluation (Kuokštis, 2013) which is another negative factor for the incumbents, since it leads towards the cutting of wages and indirect costs for employers. Of course, there were some important differences to consider. Crisis in Latvia was the deepest, considering largest contraction of GDP and highest unemployment among the Baltic States (Austers, 2014). Estonia successfully managed to keep the government debt at bay, restored the trust of investors and even joined the Eurozone at the end of 2011 (Veebel, 2014). Of course, these differences have to be taken into account while explaining the national electoral consequences of the financial crisis. However, because of the share magnitude of economic decline in the Baltic State, the economic argument is not sufficient it must be emphasized that stabilization trends appeared under extremely unfavourable conditions. To explain this paradox, some correction of theory is needed: next chapter presents an argument that precisely the economic crisis provides an opportunity to break a vicious circle of electoral volatility. 10

11 Stabilization during the Crisis: Theoretical Argument In order to explain some trends of electoral stabilization, it must be discussed why the analysed party systems are (or were, but it is still too soon to tell) unstable in the first place part of the answer may be found exactly there. This section of the paper begins with a summary of previous research on the electoral volatility in the postcommunist and Baltic States. Theoretical argument for the discussed paradox emerges both as a part of these considerations and re-evaluation of them in the context of economic crisis. There are two approaches to explaining instability of the post-communist states, including the Baltic party systems. First approach emphasizes demand side of the electoral politics, namely the peculiarities of social structure and economy. On one hand, post-communist states and especially three ex-soviet democracies inherited flattened societies, with no substantial differentiations according to the economic and ideological lines (Linz and Stepan, 1996; Whitefield, 2002; Norkus, 2008). Accordingly, major part of voting population could not develop any partisanships or ideological orientations and the civil society was weak (Birch, 2003; Miller et al., 1998; Wyman et al., 1995). On other hand, triple transformation state building, democratization, transition from communism to capitalism also meant unstable political and especially economic climate, contributing to the estrangement of significant population segments from politicians and institutions (like the impoverished or ethnic minorities). Lower levels of economic modernisation than in West and general deterioration of economic conditions (Lane and Ersson, 2007; Tavits, 2005), high unemployment (Roberts, 2008) contributed to electoral volatility. Mistrust in the political institutions added to the protest voting and non-voting. The second approach emphasizes that electoral instability in the post-communist Central Eastern Europe is greatly influenced not just by the demand side (social structure and political-economic views), but also from the supply side of the party system (Tworzecki, 2003; Rose and Munro, 2003). Ethnic cleavages in Latvia and Estonia, post-communist division in Lithuania and increasing economic differentiation in all three countries under free market created formidable conditions for the party system structuring: however, contrary to expectations there was some puzzling 11

12 empirical evidence that cleavages did not stabilize the party systems (Tavits, 2005). Part of the answer could be found in the factor of party supply: widespread mergers and splits, unstable coalitions and springing of new party organizations contributed to high levels of electoral volatility (Olson, 1998; Sikk, 2006; Jastramskis, 2010). Openness of party systems and weak establishment of competition patterns could not be avoided without pre-existing party systems (Birch, 2003) and created vicious cycle: openness led to electoral instability (Mair, 1997) which in turn maintained party system openness, creating possibilities for new party success and innovative coalitions. Some research also pointed out the effects of institutions (Birch, 2003) and party system fragmentation (Lane and Ersson, 2007). Argument that the electoral instability in the post-communist states is driven by political elite (Tavits, 2008; Van Biezen, 2003; Mair, 1997) serve as a basis for our explanation of electoral trends in the Baltic states during the economic crisis. It is quite interesting that prognosis for the electoral stabilization in the new democracies was quite bleak, since previous research showed that it is not the age but birth time of democracy that mattered (Mainwaring and Zoco, 2007). However, it could be that the influence of political elite and especially patterns of interaction between the parties were overlooked: why not search for a possible chance of stabilization in the supply side? According to Tavits, elite choices in designing the party system may help stabilise electoral alignments (Tavits, 2008): since the stabilizing of electoral trends during the economic crisis refutes the demand approach, precisely factor of party competition has a potential both to provide an explanation of discussed phenomenon and also a hope of party systems stabilization in the Baltic states. Peter Mair argued (2008) that some degree of electoral stability can be achieved due to many factors, of which three groups are the most important: first, cleavage structure and social segmentation, second, institutions and third, party system itself as a patterns of competition that provide (or do not provide) strategic constraints on electoral behaviour. Though the latter factor is sometimes overlooked (or used as dependent variable of party system institutionalization Casal Bertoa, 2014), it could explain us the push towards the electoral stabilization (or at least not destabilization) in the Baltic Systems during the crisis. We cannot expect for the cleavage structure to change in a short period, and electoral institutions remained the same for the analysed period (

13 2012). However, the patterns of interaction could have some influence on the electoral behaviour: to cite Mair, in the absence of a cleavage structure certain rules and norms in the process of party competition can anchor voters into place (Mair, 2008). Degree of party system closeness and predictability of competition patterns define these rules and norms. Open competition between parties facilitates electoral instability: more there are of innovative coalitions, partial changes in the party composition of government and easier the access for new parties to the government, less strategic constraints for the vote switching between parties, ergo, volatility at the macro level. An vice versa, in a context there competition is structured along two blocks of parties and competition is predictable, some electoral stability could emerge because electoral choices are quite constrained: if you vote for a party, you vote for a government. To talk more specifically, familiar coalitions, wholesale (or none) changes in party composition and restricted access for new parties to the government are favourable for the stabilization of party system at the electoral level (Mair, 1997; Mair, 2008). But why competition of structure in the party system should become closed during the times of economic crisis? Previous research reveals that competition structure is quite open in the Baltic states (as in all post-communist democracies) and got even more unpredictable in the last decade (Casal Bertoa and Mair, 2010). However, Lithuanian case serves as a great empirical example of presented theoretical argument: precisely during the crisis Lithuanian party system aligned into two clear opposing blocks, in contrast to the period when partial changes in government and innovative governments were norm. When HU-LCD supported the minority government of Lithuanian Social Democratic Party (LSDP) led coalition in 2006, it appeared that the structure of the party system has become completely unpredictable exactly these two parties were only two continuing elements of the party system, on right and left respectively. After 2008 parliamentary elections none of the Lithuanian observers expected for the centre right coalitional government of Kubilius to retain power for the full term. However, the economic crisis indirectly facilitated the stability of government and opposition. Under the prolonged period of high unemployment and inevitable strict budget policies, oppositional parties led by LSDP (largest parliamentary faction in the 13

14 opposition) saw few incentives for the political manoeuvres designed to bring the government down and become part of the governing coalition. It must be emphasized that weak links in the cabinet existed, foremost in the form of Rising Nation Party (RNP): splinter members established new oppositional parliamentary faction and after the merging with Liberal and Centre Union (LCU), RNP ceased to exist as an independent organization. However, the opposition initiatives were not strong enough and those showcase (like talks about the motion of no confidence for the prime minister in 2010) efforts that existed were easily countered by dominant government party HU- LCD (for example, with disproportional number of ministries handed to junior partner LCU - with only 6 seats in parliament they had 2 ministers in the government in 2010). Quite paradoxically, crisis cabinet of Kubilius became the first since the restoration of Lithuanian independence to serve a full term between two parliamentary elections even though in its majority minimal (71 out of 141 parliament members) and in the last year of term (2012) it existed as a minority government (Krupavičius, 2010; Krupavičius, 2013). Lack of incentives for political manoeuvres during the times of economic crisis could be extended to the establishment of new parties. In a milieu of weak civil society, electorally viable new parties could not emerge and sustain without the support of business, as exemplified by the only survivor from the previous array of new parties 4 in Lithuania - populist Labour party, led by influential businessman Viktoras Uspaskichas. However, context of economic crisis creates two restrictions for the creation of such organizations. First is financial, as the contraction of GDP naturally reduces the possibilities of capital that could be used for creation of party organization and electoral campaigning. Second restriction covers the lack of incentives as in the case of parliamentary opposition: creation of new party under such economic and political climate is an irrational action, since the budget cuts and bleak chances of country s recovery seriously undermine the political gains in the near future. It is interesting that the only new political power that was established during the crisis period in Lithuania, Way of Courage (WC), was an ideological-issue party with no relation to the business. To sum up presented theoretical argument, economic crisis in new democracies provides an opportunity to break a vicious circle of electoral volatility, government 4 If Order and Justice is treated as a splinter from the Liberal Union. 14

15 instability and changing supply side of the party system. Budget cuts and the risk of electoral demise devalue the spoils of office: since little incentive exists for government reshuffling and foundation of new parties, political competition in the short term becomes more structured. It is what happened in Lithuania: lines between opposition and government remained clear cut during the period, resulting in polarisation between left (oppositional left and populist parties) and right (governmental right parties). Despite being the first elections held after deep crisis (and one of the largest contractions of GDP in the world), outcome in 2012 was the best performance of incumbents in country s history, same level of electoral volatility as before crisis and diminished success of new parties. After the 2012 elections wholesale alternation of government happened, with two surviving parties of crisis cabinet HU-LCD and Liberals Movement of the Republic of Lithuania (LMRL) going to opposition and three previously oppositional parties forming a new centre-left government. Could the argument be extended for other two Baltic countries? Table 2 delineates the patterns of competition, operationalized as governmental changes in all three Baltic States before, during and after the crisis. Similar pattern as in Lithuania could be observed in case of Estonia: even though social democrats exit the government at 2009, Andrus Ansip s centre-right coalition of two parties stayed in power during all the crisis period and even after it. However, here it must also be stressed that Estonians generally more trust in political institutions (hereby Estonia s party system being the most stable from three Baltic countries) and as discussed in the first section, Ansip government was the most successful among the Baltic countries in countering the negative effects (such as budget deficit and decline of investor trust) of economic crisis (Kuokštis, 2013). However, these two factors alone are not sufficient in explaining why Estonia s party system approached such low level of electoral volatility and zero success of new parties in the Yes, it has generally been more stable than in other two Baltic countries (Saarts, 2011; Casal Bertoa and Mair, 2010) and this could be partly accounted by higher political trust: but this characteristic of society does not explain why party system actually achieved Western standards of stability exactly during the crisis. Again, restoring the trust of investors and curbing the budget deficit tell a story of effective governmental policy, but it is not exactly the logic of economic voting theory: 15

16 GDP per capita in Estonia reached pre-crisis level of 2008 only in 2013 (according to World Bank data). It could be argued that the missing component is structuring of party competition during the crisis. As in the case of Lithuania, pre-crisis party competition in Estonia was quite open (Casal Bertoa and Mair, 2010). In Ansip s Reform party was in coalition with pro-russian Estonian Centre Party and agrarian People s Union of Estonia: such government cross-cut the main ethnic cleavage and ideological lines. However, in competition blocks in Estonian politics were crystalized as in case of Lithuania, with two right parties being in government during the crisis and two main left parties (social democrats left the government exactly because of their opposition to budget cuts) criticizing the right wing policies from the opposition. In combination with quite successful anti-crisis policies, Estonian party system achieved new level of stability despite the second-largest in the world GDP contraction. Table 3. Before, during and after crisis governments in the Baltic States: structure of competition, Cabinet Year Government: party Opposition: party Government composition composition alternation Lithuania Kubilius II HU-LCD, LMRL, LSDP, OJ, LP LCU, RNP Butkevičius LSDP, OJ, LP, HU-LCD, LMRL, Wholesale EAPL WC Estonia Ansip II ERP, PPRPU, SDP ECP, PUE, EG Ansip II ERP, PPRPU ECP, PUE, EG, Partial-none 2011 SDP Ansip III ERP, PPRPU ECP, SDP None Latvia Godmanis LFP, FFF, PP, UFG HC, FHRUL, CU, NEP, SOP Dombrovski I CU, NEP, UGF, PP, FFF HC, FHRUL, SOP LFP Partial Dombrovski Unity (CU, NEP, HC, FGL (LFP, Partial II 2011 SOP), UGF PP), FFF Dombrovski Unity, RP, NA HC, UFG Partial III 2014 (FFF, AFL) Source: author s data, collected from various news publication sources (baltictimes.com, delfi.en, etc.). Blue colour: right parties, red colour: left parties, green colour: centre parties (classification according to the coalitional behaviour since 2007). For acronyms, see the appendix. 16

17 Judging from the governmental changes in Latvia during the (Table 3), the experience of country runs against the argument that economic crisis supresses the cabinet reshuffling: Latvians experienced four cabinets during the crisis and none of governmental changes were other than partial. However, exactly this lack of closing competition helps to pinpoint its relationship to the electoral stability - Latvia was the only country in the Baltics where electoral volatility actually has risen. Moreover, 2011 election that entailed big losses of incumbents and rise of new parties were not a direct result of economic crisis, but rather of intrigues of political elite and precisely the unprecedented actions by then-president Valdis Zatlers. Facing the threat of not being re-elected for a second term, Latvian president Zatlers initiated referendum for the dismissal of Saeima (Latvian parliament) that ultimately resulted in new parliamentary elections. Capitalizing on these events and image of a fighter against the oligarchic parties, Zatlers created new Reform Party which finished second (according to the received votes and parliament seats) in the elections, entered the coalition with Unity (third Dombrovskis cabinet during the crisis!) and again changed the competition structure of Latvia s party system. For presented argument to be more clear, comparisons of the party systems in the Baltic States and various relevant explanatory variables is given in the Table 4. It is apparent that economic voting could not explain the differences between Baltic States, not to mention the general trends of stabilization (or not destabilization). In 2011 (elections in Latvia and Estonia, one year before election in Lithuania) none of the Baltic States had the same level of GDP per capita as in 2008, before the financial crisis. Latvia lagged behind according to the GDP s recovery, but unemployment (according to previous research it is best correlated with election results in post-communist states Roberts, 2008) in Lithuania was highest though indicators of electoral volatility in the latter country improved more. 17

18 Table 4. Comparison of party systems and relevant indicators in the Baltic States during and after the crisis. Estonia Latvia Lithuania GDP per capita (2011), as a percentage of 2008 Unemployment (2011) Dominant cleavage Ethnic/anticommunist Ethnic Anticommunist Trust in parties (2011) Electoral System PR list PR Mixed list Competition before crisis Open Open Open Government structure during the crisis Closing Open Closing Electoral stabilization during the crisis + - +/- Source: Casal Bertoa and Mair, 2010 (openness of party system), The World Bank (GDP per capita), Eurostat (unemployment), Saarts, 2011; Ramonaitė, 2007 (cleavages), Eurobarometer (trust in parties). It appears that the only shared characteristic (high unemployment and nearly recovered GDP put aside) between two countries where electoral stabilization (again, relative to the country s context) was observed Lithuania and Estonia is closing competition between the parties during the crisis period. All other characteristics are either (theoretically) negative for the electoral volatility or different (as cleavages, electoral system and trust in parties). Logic of one similarity dictates that proposed theoretical argument about the relationship between the party competition and economic crisis in the Baltic States is justified by the comparison of countries. In the last section of this paper, some empirical evidence at the individual level is presented to support it. Competition of Parties during the Crisis: Individual Level Analysis Interestingly, theoretical approaches that link party system as an independent variable to the electoral behaviour (Mair, 2008; Sartori, 2005) are rarely adapted to the micro level analysis differently from demand side theories, namely the sociostructural and economic voting. The obvious obstacle here is operationalization. Even though electoral constraints of closed party system could be defined in theory, it is hard to measure them empirically on the individual level: should we analyse the distance between parties, their placement on left and right or maybe voting for distinct blocks? 18

19 To provide some support for paper s theoretical argument, several techniques were employed. Using data from two post-electoral surveys in Lithuania one pre-crisis (2008) and second after-crisis (2012) the following steps of analysis were conducted. First, simple descriptive statistics and significance tests were calculated in order to determine if perceptions of left and right space shifted during the crisis. Second, multidimensional scaling was applied for available thermometer questions about the main parties before and after the crisis: in order to determine if structure of party system was viewed differently after the analysed period. If theorized incentives and clear-cut competition of crisis period really made influence on the perceptions of voters, some changes should be observed in the micro-level statistical analysis. Figure 2 presents the first indicator of competition space perception: respondent s subjective placement in the political left and right scale. Though it is foremost a measure of individual ideological predisposition, there are serious considerations that it s substance varies according to the analysed context and is not necessarily ideological (Fuchs and Klingemann, 1990; Inglehart and Klingemann, 1976). In the political context of weak ideological traditions (as in Lithuania see Ramonaitė, 2007; Ramonaitė; 2008) it is very probable that these positions are set according to the supported party s self-image and coalitional behaviour: in Lithuania, those voters that support (more or less) ideologically defined parties more easily place themselves in the left-right scale. Accordingly, supporters of the parties with obscure ideological image either do not find their placement, or their positions are very atypical if compared to other countries (Ramonaitė et al., 2014). 19

20 Figure 2. Respondents left and right placement in 2008 and 2012 post-electoral surveys: histograms with fitted normality curves and density plots. Data from post-electoral surveys supports expectation that placements along the leftright scale in Lithuania after the crisis became more structured this could be a direct result of crystallization of two competing blocks, namely the right parties of the crisis government vs. centre-left opposition of the austerity measures. First, in 2012 postelectoral survey only 19.7 percent (31 valid percent) of Lithuanians placed themselves in the centre in opposition to the 30.7 percent in 2008 (43.1 valid). Also, positions to the left and right from the centre became more frequent in 2012 as it is clearly visible from the histograms with the fitted normality curves and density plots (Figure 2). Finally, these considerations were supported by F-test which resulted in statistically significant difference between the variances of right-left placement in 2008 and 2012 (F = 0.72, p-value = 0.00, 95 percentage confidence interval: [0.63:0.82]). 20

21 Discussed comparison reveals that more people in 2012 chose to be on left and right: however, it says little about the changes in the perceptions of the party system structure. In order to evaluate this aspect we employed multidimensional scaling (MDS), applying this technique for the questions of party thermometer (how particular party is liked, measurement in an ordinal scale) in 2008 and 2012 postelectoral surveys. Such a method does not impose any predefined scheme on the voters perception and is quite widespread practice, used for the mapping of competition space between parties or candidates (Weisberg and Rusk, 1970, Damgard and Rusk, 2010, Rusk and Boore, 2010). It serves two purposes in the context of this paper: first, helps to map the perceived competition of parties and second, enables to compare these perceptions before and after the crisis. In order to compare the changes in the perceptions of party competition, we applied classic metric MDS for both pre-crisis and post-crisis electoral surveys. For calculations of distance between the parties, Euclidian distance was used. In addition to pre-crisis and after-crisis distinction, different MDS analysis was run for two sets of parties (to be more precise, sets of thermometer questions on particular parties). First set entailed all the parties that were in parliament either before or after (main parties) respective election. Second set of parties entailed only the big five of Lithuanian politics: those parties that survived both 2008 and 2012 elections, remaining relevant during all the crisis period and after it. Results of MDS analysis that included all the main parties from 2008 elections (Figure 3, upper left) reveals picture of fragmented competition with no signs of leaning towards the structuring along two camps. Four parties of pre-crisis Gediminas Kirkilas minority government (LSDP, NU, LPPU, LCU) appear to form one cluster of centre parties. However, as in real coalitional patterns of then Lithuanian politics two oppositions could be observed: right parties (LMRL and HU-LCD) and populist wing (OJ and LP). In addition, it seems that newly established RNP alone formed a fourth cluster of parties. Removing those parties that ceased to be relevant during the crisis (Figure 3, upper right) does not make the picture clearer: supposedly leftist LSDP is in the centre of horizontal dimension and opposed to all other main parties on the vertical dimension. This quite accurately reflects coalitional policies of those times LSDP 21

22 avoided siding with either right or populists and created a coalition with small centreleaning parties in Figure 3. Perceptions of Lithuanian party competition before and after the crisis: results of MDS analysis, 2008 and 2012 post-electoral surveys. As theorized in the previous part, party competition actually become more structured in 2012: lower left picture depicts clear alignment of the six pre-2012 parliamentary parties into two blocks, first comprising the oppositional LSDP, LP and OJ and second including all three parties of Kubilius centre-right government. Newly established WC forms another cluster (as in the case of RNP, it is very similar to niche of protest voting) but it must be emphasized that party s electoral gains were quite modest (7.9 percent of votes and only 7 seats in the parliament). One trend is particularly striking complete dissolution of the former centre cluster, with LSDP joining the populist-left and LCU drifting towards their partners of the crisis 22

23 government. Removing the less significant parties (lower right, Figure 3) reveals larger difference between two right parties. This could reflect ideology of LMRL, since it is the only party from the big five that seeks to represent the moral individualism in the relatively conservative Lithuanian electorate (Ramonaitė et al., 2014). Despite this, horizontal dimension quite clearly depicts division between left and right blocks even when WC and LCU are removed. Conclusions Theory of economic voting promised bleak electoral scenarios for the governments and party systems in the Baltic countries during the crisis. However, despite being among the leaders in the world according to the scales of experienced economic recession, Baltic states were not pushed towards the brink of electoral destabilization or even transformation of party systems. In contrary to theoretical expectations, electoral volatility stayed at moderate (relative to the context) or decreased (in Estonia it even reached level of Western countries), success of new parties diminished (with exception of Latvia) and Estonians even re-elected their government. The main aim of this paper was to explain this paradox of the Baltic States: why financial crisis did not bring party systems in these countries to chaos and even some stabilization trends were observed? Paper proposed a theoretical argument that proposes a correction to the established wisdom of relationship between economic downturns and electoral results. It is argued that in new democracies an opportunity to break a vicious circle of electoral volatility and openness of party competition presents itself during the economic crisis. Spoils of office are devaluated during the crisis, especially if austerity measures are hardly avoidable: rational political actors do not see incentives for the attempts of cabinet reshuffling (opposition) or establishment of new parties. Political competition becomes more structured over time, and division between political blocks is indirectly strengthened by the crisis context: polarisation between the crisis government and the parliamentary opposition criticizing former s policies emerges. Of course, further research is needed to expand on what social and political conditions enable such scenarios that happened in Estonia and Lithuania. Part of answers could lie in the broader context of growing body of research on the economic crisis, especially studies 23