Interplay Between Non-Partisan Presidents, Party System and Quality of Democracy in Lithuania

Transcription

1 Interplay Between Non-Partisan Presidents, Party System and Quality of Democracy in Lithuania Mažvydas Jastramskis, Vilnius University Paper prepared for presentation at ECPR General Conference 2017, University of Oslo, Oslo, September 6-9 PRELIMINARY DRAFT: PLEASE DO NOT QUOTE WITHOUT AUTHOR S PERMISSION Abstract. Lithuania has a semi-presidential political system with directly elected president, in the context of Central and Eastern Europe enjoying average presidential powers (Sedelius, 2006). There is a strong tradition to elect non-partisan presidents: since the restoration of state independence and transition to democracy, only two presidents had traceable background of belonging to a particular political party: Algirdas M. Brazauskas ( ) and R. Paksas ( ). If additional criterion of president coming from the established party system (hence not a new party) is added, this list shortens to only Brazauskas: the last time when Lithuania elected a truly partisan president was On other hand, Lithuania also has a semi-stable party system with average-to-high electoral volatility: traditional left-right parties persist, but are constantly challenged by newcomers. This poses two main questions: what are the conditions for such a success of non-partisan presidents and what is the political impact of this tradition, especially on the party system and quality of democracy? It is argued in the paper that there are three main groups of factors explaining the non-partisan presidents success in Lithuania: institutional, social and political. First, main institutional factors are the direct elections themselves and clause in the Constitution that president cannot be a member of any party (this applies only after the election, hence parties can raise their candidates). Second, social factors are very important, since they enable voting for the non-partisan candidates (also observed in the parliamentary elections): low degree of party identification (lots of floating voters), low trust in the political parties and lack of

2 ideological differences among the voters (the main cleavage being on the geopolitical orientation and historical memory). Finally, party strategies that try to maximise their political influence without antagonizing the potential supporters (in the parliamentary elections) also play the part: in 2009 and 2014 main right-wing parties (liberals and conservatives) chose to support popular non-partisan Dalia Grybauskaitė. Even though Grybauskaitė easily won both contests, quantitative analysis of her support base reveals both catch-all and right wing profile. This leads to a paradox: successful candidate in the Lithuanian presidential elections should be non-partisan, but also have the support of important parties. It is argued that such a symbiosis is actually damaging for the quality of democracy, since it reinforces the weakly structured party system and mistrust in parties: president enjoys the perks of being supported by the parties organizations in the campaign, but in order to stay popular has to keep the distance and exploit the nonpartisan side of the politics when elected. Moreover, direct presidential elections are usually very personalized and strengthen the role of populism and plebiscitarian linkages (issues that have little to do with the post of president are emphasized: for example, socioeconomic problems). Majority of Lithuanians would prefer the strengthening of presidential powers and support for the strong-hand politics is one of the highest in the EU.

3 Introduction There is a large increase of presidential elections in the democratic world since the end of 1980s world (Bormann and Golder, 2013). There are more than 50 countries in the world now with semi-presidential constitution (Elgie, Moestrup and Wu, 2011). This system is now the most popular in Europe (Amorim Neto & Strøm 2006: 623) and among the former post-communist countries (Sedelius and Mashalter, 2013), but not in the Baltic States. Latvia and Estonia are parliamentary political systems with indirect elections of president, and only Lithuania here has a semi-presidential political system with a popularly elected president and a cabinet responsible to the legislature (two main characteristics of semi-presidentialism: Elgie, 1999, 13). There is some disagreement over the pros and cons of semi-presidentialism, with the main dividing line drawn between those that see this governing system as a flexible, power sharing and those that present an argument of possible institutional conflict and political instability (Sedelius and Ekman, 2010; Sedelius and Mashalter, 2013). However, there are other important and less explored potential effects of semi-presidentialism. One of them is quite obvious: differently from parliamentary systems (predominant in Europe before the democratization of post-communist states), such constitutional system presents a possibility of non-partisan and/or populist presidents. Although there have been nonpartisan presidents under parliamentarism (e.g. in Latvia and Estonia), in semipresidential system they differ (even if their other powers are weak) in one crucial aspect: a popular mandate. According to Robert Elgie (2008, p. 52), one of main perils of semipresidentialism is that direct elections of president may encourage populist candidates and presidents that place themselves above the politics. Lithuania seems like a convenient case to test the impact of direct elections and independent presidents on the quality of democracy. Despite the young age of democracy, it already has a strong tradition to elect presidents with weak or no partisan attachments. Since the restoration of state independence and transition to democracy, only two presidents had traceable background of belonging to a particular political party: Algirdas M. Brazauskas ( ) and R. Paksas ( ). If additional criterion of president coming from the established party system (hence not a new party) is added, this

4 list shortens to only Brazauskas: the last time when Lithuania elected a truly partisan president was On other hand, despite the average-to-high electoral volatility Lithuania has a party system with persisting core of two main left-right parties. They are constantly challenged by newcomers (or previously marginal parties with revamped image), but competition patterns remain largely unchanged (Ramonaitė et al., 2014). This poses two main questions that are dealt with in this paper. First, what are the conditions for such a success of non-partisan presidents? Second, what is the political impact of this tradition, especially on the party system and quality of democracy? It is argued in the paper that there are three main groups of factors explaining the nonpartisan presidents success in Lithuania: institutional, social and political. We cover the institutional factors in the first section of the paper, where we introduce the semipresidentialism in Lithuania and discuss the powers of president. In the second section we analyse the social factors that enable voting for the non-partisan candidates, mainly trust in institutions, partisanship of voters and cleavages. Third, we discuss the political strategies of the parties and present individual-level regression analysis of support base of incumbent Lithuanian president Dalia Grybauskaitė ( , re-elected in 2014). This provides additional insight into what makes a successful non-partisan candidate in the Lithuanian presidential elections. Finally, we discuss the possibility of vicious cycle: semi-presidential system and anti-party sentiments in the electorate enable success of non-partisan presidents that maximise their popularity by keeping the distance from the parties and by exploiting the non-partisan side of the politics when elected. Thus trends of non-partisan and personalized politics are reinforced, with damaging consequences for the party democracy. Semi-presidential system and president powers in Lithuania According to widely accepted definition by Elgie (1999, p. 13), semi-presidentialism has two main qualities: 1) directly elected fixed-term president and 2) cabinet (prime minister) that is collectively responsible to the legislature. This definition is quite

5 inclusive (Sedelius and Mashtaler, 2013) and easier to apply for the purposes of classification purposes compared to the previous definitions that employ rather ambiguous characteristics such as considerable powers of the president (Duverger, 1980; Sartori, 1994). Elgie s definition also draws a clear line between the semi-presidentialism and two other forms of democratic government: 1) parliamentarism (only parliament is directly elected, government formed and accountable for the parliament; 2) presidentialism (president is directly elected, forms and commands the government). Of course, since definition is parsimonious and inclusive, there is a considerable institutional variation among countries with such constitution (Elgie, 2011, p. 19). Shugart and Carey (1992) distinguished between two main forms of semi-presidentialism that were later adopted by further research (see Elgie, 2011; Elgie, Moestrup & Wu, 2011): president-parlamentarism and premier-presidentialism. Under presidentparliamentarism, prime minister and cabinet are collectively responsible both to the legislature and president. In a premier-presidential system prime minister and cabinet are collectively responsible solely to the legislature. Essentially, premier-presidentialism is a form of semi-presidential system with a weaker president vis-a-vis the cabinet. In a president-parliamentary system, president usually has more power in the government formation and cabinet is responsible to him (as to the legislature). Lithuania has had a premier-presidential form of semi-presidentialism since the adoption of Constitution in President is elected for fixed five-year term and can be removed from the power only by legal process of impeachment (happened once in 2004). President also nominates prime-minister that has to be voted-in by legislature. It must be emphasised that according to the 1998 ruling of the Constitutional Court, president is restricted when nominating a prime-minister by the confidence of the governing coalition in Seimas (Krupavičius, 2008). After the approval of prime minister, governing coalition selects cabinet members that have to be approved by the president. Apart from the involvement in the government process, other significant powers of Lithuanian president are: initiation of legislation, suspensive-package veto (overridden by 71 out of 141 parliament members) and nomination of several important state officials (such as chief justice of the Supreme Court, chief justice of constitutional court, attorney general).

6 There is some disagreement over the role of Lithuanian president in the academic literature. Daunis Auers (2015) evaluates the presidential powers of Lithuanian president as quite weak and compares them to the heads of states in two parliamentary Baltic countries, Latvia and Estonia. Algis Krupavičius (2014, p. 214) emphasizes that parliamentary element is very strong in the premier-presidential form of Lithuanian semipresidentialism. On other hand, according to index by Thomas Sedelius (2006) index (that includes appointive powers, political powers and method of election) Lithuanian president enjoys average powers in the context of Central and Eastern Europe: weaker than his/her Polish counterpart but significantly stronger than the heads of states in Estonia or Latvia. Diverging evaluations possibly arise due to the nature formal presidential powers and trend of non-partisan presidents. Laimonas Tallat-Kelpša (1996, p. 96) describes presidential powers in Lithuania as responsive: president maximises influence acting like an arbiter in the political system and/or using the popular mandate. President may be passive for a longer term and activate his/her powers when a political opportunity presents itself. Illustrative cases are when Algirdas M. Brazauskas (in 1995) and Valdas Adamkus (in 1998) went into conflict with less popular prime ministers (both conflicts resulted in the resignations of prime ministers) or when Dalia Grybauskaitė vetoed very unpopular new version of Labour Code recently, in 2016 (later adopted in Seimas with revisions). Hence the influence of president in the Lithuanian political system is fluctuating however, not in the exact sense that Giovanni Sartori (1994) described when analysing the French Fifth Republic. The main difference is that according to Sartori, president maximizes his (informal) power when his party dominates the parliament. In contrary situation, the de facto role of the first executive is transferred to prime minister and the president is forced to accept the rules of cohabitation. Formally, due to the success of non-partisan candidates Lithuanian president is in a permanent state of cohabitation. However, especially due to his/her non-partisanship president exerts some power over the parliamentary majority: mistrust that Lithuanians have in parties do not damage his/her authority which stems from a popular mandate.

7 There have been some signs of growing president s influence during the two terms ( , 2014 till now) of Dalia Grybauskaitė. When after 2012 parliamentary election she had to accept Labour Party (DP also see the appendix for party acronyms) as a part of governing coalition, this was interpreted as a sign of parties predominance in Lithuanian semi-presidentialism as they forced Grybauskaitė to nominate center-left government (Auers, 2015, p. 51). However, it must be emphasized that president reacted by being very picky with candidates to the minister posts, thus establishing a precedent of president s strong involvement in the personal composition of the government. She even approved the government without two ministers (of Social Security and Education, both delegated by Labour party), explaining that candidacies were not competent enough. When Electoral Action of Poles in Lithuania (LLRA) quit the coalition in 2014 and strategically important energy ministry was formally transferred to the Labour Party, it was de facto the president that selected the candidate Rokas Masiulis. Her influence over the personal composition of government grew even more after 2016 parliamentary elections, when election-winning LVŽS (Lithuanian Farmers and Green Union) together with LSDP (Lithuanian Social Democratic Party) agreed to form mostly technocratic government (only 10 out of 14 party ministers belong to parties). Political pundits argue that as many as half of the minister could have been hand-picked by Dalia Grybauskaitė. It is also important that Lithuanian president is more active in the veto arena than heads of state in Estonia and Latvia (Auers, 2015), indicating higher degrees of activity and independence from the parliament. Hence, tradition of non-partisan presidents along with the semi-presidential design contribute to the reactive type of president in Lithuania. However, it could be argued that institutional aspects that are part of this design contribute to the success of nonpartisan presidents. First main factor are direct elections themselves obviously: even though there are non-partisan presidents under parliamentary regimes, direct elections (especially under weakly institutionalized party system) create a possibility for nonpartisans to enter the race. Especially under a weakly institutionalized party system (such as Lithuania), voters tend to choose independent candidates when possible. After the introduction of direct mayor elections in 2015, in 5 out of 60 municipalities non-partisans were elected to the post, including 4 out of 6 major cities.

8 Second, parliamentary and presidential elections in Lithuania are not concurrent. The main reason behind this are different terms of parliament (4 years) and president (5 years). Besides the facilitation of cohabitation periods (Krupavičius, 2008, p. 77), nonconcurrence of these elections is also important for the interplay between the nonpartisanship of presidents and political parties. Concurrent elections could be regarded as a factor favourable for the formation of presidential parties. Without the parliamentary elections, presidential candidates have less incentives to invest into organization building and make themselves partisan : with a partial exception of Rolandas Paksas whose party was formed by splitting from the Liberal and Center Union. There have been instances that former candidates try to capitalize on the name recognition gained through the election and found new parties (or join the older, marginal ones) in the subsequent elections. The most notable case is of Artūras Paulauskas that came close second in the 1997 presidential race. In 1998 he found an anti-establishment reform party New Union (for the classification of such parties see Hanley and Sikk, 2016) that later finished first in 2000 local elections and third in the parliamentary elections of the same year. In this case, direct presidential elections encouraged party system instability in the subsequent parliamentary elections. On other hand, non-concurrent elections mean that established parties are in a less favourable position to raise their own candidates. It is argued that on average governments reach their lowest point of popularity in the middle of term (Marsh, 1998). In Central and Eastern Europe countries where hyper-accountability is observed (Roberts 2008), government parties as a general rule become less popular after they move into power. Moreover, protest voting against government tend to encompass voting against all the parliamentary parties (Pop-Eleches, 2010). In combination with weak partisanship (see second section), that means that established parties cannot compete in Lithuanian presidential elections when they are the strongest: trust in parliament and the parties is higher only immediately after the parliamentary elections. Finally, article 83 in the Constitution states that elected president has to suspend her/his activities in the political parties and political organization. Even though parties still can support and de facto raise candidates, it could be argued that the establishment of a nonpartisan president institution in the Constitution has strong suggestive power both for the

9 presidents that try to keep a distance from parties (even from their former, as Brazauskas in ) and for the voters. It defines a normative ideal of a president that is independent and above the parties. Krupavičius (2008, 77) argues that the effect of this article was negative, as it weakened the role of parties in the system while also providing presidents an opportunity to act as a free-rider against other institutions and sometimes in quite populist manner. Trust in president and other institutions In his analysis of the political systems in three Baltic States, Aeurs (2015, p ) provides three explanations for the success of former émigré presidents in mid-2000s. Two of these explanations (knowledge of English and experience in democratic politics) are specific for the cases of émigré presidential candidates, but the third, lack of domestic political networks (or low involvement in them) also applies for the non-partisan presidents in general. If parties are as seen as corrupt organizations by the voters and not trusted, non-partisan president will have an advantage. Quite similarly to Italy, where non-partisan governments are presented as the consequence of the failure of political parties (Pasquino and Marco Valbruzzi, 2012), in the case of Lithuania we can attribute part of the non-partisan presidents success to the inability of political parties to lay stable roots in society. President and political parties are respectively the most and least trusted political institutions in Lithuania (see Figure 1). Data from 2012 and 2016 post-electoral surveys also reveal that low confidence in the parties is quite immobile: even though trust in other institutions increased quite considerably, the change of trust in the parties is insignificant. And vice versa, the largest increase (by 1.3 on average in a scale from 0 to 10 scale) is recorded for the president institution. It must emphasized that this happened during the second term of non-partisan president: Dalia Grybauskautė was re-elected in 2014: (first re-election of president in Lithuanian modern history). It appears that president came out the largest winner out of the previous conflict between parties and non-partisan Grybauskaitė (see previous section).

10 <Figure 1> There are relatively strong correlations (Pearson coefficients from 0.5 to 0.8) between the voters trust in different institutions: the ones that trust the president more also tend to express more confidence in the government, parties and parliament. However, correlations between the trust in president institution and parties are the weakest (Pearson coefficient is 0.51 in 2016 and 0.48 in 2012). From the Figure 2 we can see that this relationship is non-linear. Voters that do not trust the president (0-3 in the eleven point scale), almost equally mistrust the parties. However, the ones that give the most confidence to the president (7-10), place significantly less confidence in the parties: even though both variables are positively correlated, the actual gap between institutions in both surveys become higher with every higher level of trust in president. Even more, there is some prove that at the highest point of trust in president (9-10 in 2016 and 10 in 2012), confidence in parties drops down. Hence it could be argued that complete trust in president slightly decrease the confidence for the parties. <Figure 2> Even in the context of general mistrust for the parties as an institution, party organizations and cleavages could stabilize the voter-party relationship (Bartolini and Mair, 1990). Previous research (Jurkynas 2004; Ramonaitė 2007; Ramonaitė et al., 2014) reveals that one important division in the politics of Lithuania resembles a cleavage: between those with more positive attitudes towards the communist past and cooperation with Russia on one camp and the ones that evaluate communist times negatively and see Russia as a threat on the other. This cleavage of geopolitical orientation and historical memory correlates with the left-right in Lithuania best, other ideological differences among voters and parties (such as economic left-right scale or social liberalism) being weakly pronounced (Ramonaitė et al., 2009; Ramonaitė et al., 2014).

11 The main right-wing party TS-LKD (Homeland Union-Lithuanian Christian Democrats) capitalized on this cleavage in the long run, creating an electoral base of partisans with strong anti-communist attitudes (and strong perception of Russia s threat) that find themselves and their party on the right spectrum of political competition (Ramonaitė 2007, Ramonaitė et al., 2014). However, voters on the other side of the spectrum (prosoviet nostalgia and cooperation with Russia) did not rally around one political organization and voted for different parties: for example, in 2012 their preferences (altogether with those voters that are at the centre of this political division) split between previously mentioned DP, center-right LSDP and party Order and Justice (TT). There is also a minor ethnic cleavage that separates polish voters of LLRA: however, differently from Estonia and Latvia this cleavage in Lithuania it is only of second importance (Saarts, 2011). Consequentially, the overall partisanship of Lithuanian electorate is quite low and longterm relationships between the parties and voters are weak. According to the calculations by Scott Mainwaring, Carlos Gervasoni and Annabella Espana-Najera (2016), Lithuania is at 6 th place out of 67 countries on the criteria of electoral volatility. Even though relative stabilization was observed in 2008 and 2012 parliamentary elections (Novagrockienė, 2012; Jurkynas, 2014), in 2016 volatility increased again to 35 per cent (workings of Lithuanian mixed-majoritarian electoral system turned this into 58 per cent turnover of parliamentary members). Of course, aggregate volatility of election results have roots at the individual level: according to data from post-electoral surveys, only around 45 per cent of people that turn to vote are loyal (voted for the same party in previous election). The remaining either changed party or did not participate in previous election (average turnout in Lithuanian parliamentary elections is rather low, around 50 per cent). From Figure 3 (data from 2016 post-electoral survey) we can see that stability of voting behaviour is related to the trust in parties and president. The stable voters are most trusting in both institutions and the ones that did not vote in 2012 or 2016 parliamentary elections are most distrusting. However, there is a stable gap of trust between the parties and president in all five main groups: even though stable voters and the ones that did not participate in previous elections (40 per cent of them are of years old) put more

12 trust in the parties, they still trust the president by average of two points more (the same as the people that do not vote). Also, even most stable voters tend more to distrust the parties on average. From those that voted in parliamentary elections, switching voters are the most distrustful in parties. How discussed social factors are important for the success of non-partisan presidents? First, when president institution is trusted almost two times more than the parties (and the gap of trust between parties and president increase with the higher levels of trust in president), party candidates have a permanent disadvantage in direct presidential elections (and vice versa, non-partisan candidates have an advantage). Second, lack of strong cleavages (with one relative exception) and weak ideological structuring of the political competition leave more space for personal, charismatic politics. Third, low partisanship and high volatility of Lithuanian electorate means that parties cannot transplant large bases of voters across the parliamentary and presidential elections: at least not enough to secure a win in the latter. <Figure 3> Strategies of parties in the presidential elections It is argued that institutional format (parliamentary, presidential, semi-presidential system) influences the recruitment of executive candidates and the capacity of political organizations in new democracies (Samuels and Shugart, 2013). However, one could also leave some space for the agency since there are other important incentives: building of party organization is crucial for the party survival in the Central and Eastern Europe (Ibenskas, 2014). Parties may also chose different strategies of presidential candidate recruitment under the same system, as exemplified by the example of Lithuania. In this section and further in the paper, I concentrate on trends in the recent decade that includes the 2009 and 2014 presidential elections. For a comprehensive overview and discussion on presidential elections see Krupavičius, Organizationally, Lithuanian parties are quite weak. Even though the party membership has been increasing in the last decade, percentage of electorate in Lithuania that belong

13 to parties is still significantly less than the average of Europe: 2.71 per cent and 4.65 per cent respectively (Biezen, Mair and Poguntke, 2012). There is a joke among Lithuanian political scientists that the per cent of people that trust in parties is equal to per cent that belong to parties: implying that only party members have confidence in the institution. Interestingly, the low membership is explained not only by the society s anti-party sentiments. According to the qualitative and quantitative sociological analysis by Ramonaitė (2010), there is a part of electorate that has an inclination towards the involvement into party organizations but does not know how to do it (especially young voters). State financing dominates the budgets of all the relevant parties, but little of it goes for the organizational building. In 2014, the party with the most staff was TS-LKD, employing 26 party workers (author s collected data). Hence the parties are also partly to blame (or maybe lack of state financing) for the low partisanship of the voters: they do not invest enough in the long-term to the organization building and rather focus on the electoral campaigns. <1 table> In the context of weak parties relationship with society and weak political organizations, the popularity of non-partisan candidates in direct elections is not surprising. However, parties may choose different candidate recruitment and endorsing strategies thus reinforcing or, contrary, trying to mitigate the trends of non-partisan democracy. In Table 1 we can see the per cent of votes cast for the candidates raised in parties. Both of these contests were dominated by non-partisan Dalia Grybauskaitė. In 2009, she won in the first round with 68.2 per cent of voters. In 2014 Grybauskaitė was also was a clear favourite, but fell short of majority in the first round with 45.9 per cent of votes. However, in the second she won quite easily against the LSDP candidate Zigmantas Balčytis that received 40 per cent of votes. Only three parties nominated their candidates for the presidential election in both elections: centre-left LSDP, party of the polish minority LLRA and populist DP. Two main parties of the right (liberal LRLS, Liberals Movement of the Republic of Lithuania

14 and conservative TS-LKD) did not raise a candidate in neither of election (actually, the last time conservatives raised their candidate was 1997 presidential elections). However, also only three parties raised their candidates in the last presidential elections preceding Grybauskaitė era (held in 2004). Moreover, the outcome of the first round in 2014 elections retrospectively showed that competitiveness of elections was higher than (possibly) expected. It could be that party strategies that try to maximise their political influence (or minimise the damage) without antagonizing the potential supporters (in the parliamentary elections) also play the part. Since president in Lithuania has mostly responsive powers, the main executive power remains in the hands of the government and prime minister. Also knowing the anti-party sentiments of the society and specifics of direct elections (favourable for non-partisans), parties may see the candidacy for the presidential post as bargaining chip in future support for the executive policies (or at least minimising the risk of conflict). In , during the economic crisis centre-right government let by TS-LKD applied unpopular austerity measures and having a critical president besides could have been detrimental. Conservatives and two liberal parties of the coalition chose not to raise candidates in the presidential race, supporting Grybauskaitė that silently approved the centre-right economic policy. In 2014 presidential elections liberals and conservatives again supported Grybauskaitė. However, the move by the government-leading LSDP to nominate a candidate also could have been a strategic move. Instead of nominating Algirdas Butkevičius, prime minister ( ) and then one of the most popular politicians in the country, party raised candidacy of less popular European parliament member Zigmantas Balčytis. He barely got to the second round with 13.6 per cent of votes and did not pose a serious competition for Grybauskaitė. Former was rather critical of the centre-left government and Butkevičius possibly chose a safe scenario, opening the way for the second Grybauskaitė term.

15 Non-partisan president with a partisan base How important was the support of right-wing parties for Grybauskaitė? Although the electorate of liberal parties is quite fluid (as the liberal parties themselves: there have been series of splits and mergers), the conservatives (TS-LKD) have a faithful voter base, built on the previously discussed cleavage of geopolitical orientation and historic memory. In 2012 and 2016 parliamentary elections, 86 and 77 per cent of voters respectively were loyal among the TS-LKD electorate (voted for the same party in previous election). Despite losing both parliamentary elections, conservatives command the largest number of partisan voters in Lithuania: 32.4 per cent from all the stable voters ( ). Right-wing supporters also tend to appear at the elections more frequently (Ramonaitė et al., 2014). Unfortunately, no post-electoral surveys were made in Lithuania after the presidential elections. However, we can analyse the supporting base of Grybauskaitė trough the sympathy question from the 2016 post-electoral survey (conducted after the parliamentary elections). Respondents were asked to rate politicians on a scale from 0 to 10 where 0 means antipathy (respondent does not like a politician) and 10 means sympathy (respondent likes the politician very much). As can be seen from the Figure 4, distribution of answers to the question about sympathy towards Grybauskaitė is negatively skewed: more Lithuanians like her than do not like. This evaluation correlates strongly with the trust in the president institution (Pearson coefficient is 0.73). In order to explore the partisanship of Grybauskaitė s supporting base, we use the independent variables that measure propensity to vote for parties. In this way, we could have different independent variables measuring support for different parties: PTV is a measure that is not bounded to a fixed sum (van der Eijk et al. 2006, Angelis and Garzia 2013). Question given for respondents was formulated like this: Would you indicate for each party how probable it is that you will ever vote for that party? The scale is from 0 to 10 where 0 means that voter would never vote for the party and 10 that is very probable that he/her would vote for the party in the near future.

16 We used an ordinary least squares regression model, also controlling for the education (three level variable), gender, age, left-right scale (0 left, 10 right) and evaluation of soviet times (reference category agree that soviet times were better). The latter variable is important in Lithuanian context, capturing the main political division (cleavage) among voters. We added six propensity-to-vote independent variables, each for every parliamentary party. All of them were statistically significant in the model that explained 0.23 per cent of the variance in the dependent variable (adjusted R squared). In Figure 5 we present the main findings (average predictions of dependent variable by propensity-to-vote for parties) of the regression analysis regarding the relation between partisanship and Grybauskaitė s support base. First, the right-wing profile of her (potential) voter base is confirmed. High levels of propensity to vote for liberal and conservatives (upper part of the figure) in the near future are associated with high levels of sympathy towards Grybauskaitė. Coefficients of propensity-to vote for both parties (0.19 for TS-LKD and 0.21 for LRLS) are positive and indicate quite steep, linear relationship that is represented in the Figure 5. <Figure 5> However, propensity-to-vote for center and center-left parties that formed a governing coalition after the 2016 parliamentary elections (liberals and conservatives remaining in opposition) also have positive a positive relationship with Grybauskaitė s evaluation: see the middle part of Figure 5. Although the coefficients are smaller (0.15 for LSDP and 0.11 for LVŽS) and relationship is less steep, but supporters of these two parties still tend to evaluate Grybauskaitė more positively than non-supporters. This indicates a catch-all profile: although right wing voters are apparently the most hard-core supporters Grybauskaitė s support base, her positive evaluations extend significantly to the centre and centre-left of Lithuanian party system. Only the parties with a voter base that is mostly pro-soviet and pro-russian (LLRA and TT, lower graph) had negative relationship with the dependent variable.

17 The result of analysis points to a paradox. Successful candidate in the Lithuanian presidential elections should be non-partisan, since the trust in parties is very low. However, he/her also needs to have the support of party electorates potential voters of the main four Lithuanian parliamentary parties and supporters of Grybauskaitė largely overlap. On one hand, this again points to the weakness of parties to establish an ideological, loyal voter base. LSDP supporters like Grybauskaitė more than nonsupporters, even though party s candidate was her opponent in the second round of 2014 presidential elections. On other hand, victory of Grybauskaitė in 2014 would be not so guaranteed without the support of the right wing parties that also provided the support of their political organizations (for example, important for the collection of 20,000 signatures needed for the candidacy in the presidential elections). Discussion: (vicious) cycle of non-partisan presidents and weak party system? It is difficult to separate the impact of semi-presidential system on the Lithuanian democracy from the underdeveloped parties, volatile electorate, weak civic society and other factors (Krupavičius, 2008, p. 83). However, if we turn the focus on the tradition of non-partisan presidents in Lithuania (one of the effects of semi-presidential systems) a possibility of (vicious) circle emerges. In this paper we covered three main groups of factors that account for the non-partisan presidents success in Lithuania: institutional, social and political. However, for the latter two groups the relationship is not so straightforward: what if weak, underdeveloped parties (party system) and mistrust in parties are not only the conditions enabling, but also (at least partially) themselves an outcome of the non-partisan presidential politics in Lithuania? At centre here is the paradox of the interplay between non-partisan presidents and parties: presidential candidate needs to be a non-partisan and have a catch-all base to win the elections, but he/her also relies on the support of parties and their electorate. It could be argued that such a symbiosis is actually damaging for the quality of democracy, since it possibly reinforces the weakly structured party system and mistrust in parties. President enjoys the perks of being supported by the parties organizations in the campaign, but in

18 order to stay popular has to keep the distance and exploit the non-partisan side of the politics when elected: partly due to the institutional (constitutional article on the nonpartisanship of the president), partly due to the social reasons (mistrust in parties). During her presidency Grybauskaitė on various occasions accentuated her non-partisan status and even criticized the institution of parties: the most notable example is her interview for the German TV Deutsche Welle where president declared having allergy to membership in political parties. 1 This trend was reversed only in the 2017: after the 2016 parliamentary elections that were won by LVŽS list dominated by non-partisan candidates (and formation of mostly non-partisan, technocratic government), president addressed the need to develop strong parties in her annual report. At , Lithuania was among the leading countries in the Central and Eastern Europe according to support for a strong authoritarian leader (Haerpfer, 2002, p. 34). It could be argued that tradition of directly-elected president with reactive powers that is above the parties may be a factor favourable for the persistence of such views. Direct presidential elections are usually very personalized and create space for populism. Moreover, in Lithuania issues that have little to do with the post of president (according to constitution, president executes only foreign policy altogether with the government) are emphasized, such as socio-economic problems. Two decades and more into the independent statehood and democracy Lithuanian society still strongly favours personal, presidential power over the party democracy. In the 2008 post-electoral survey, 75.9 percent of respondents preferred the strengthening of presidential powers. In a 2016 post-electoral survey, 39.7 per cent strongly agreed and somewhat agreed that having a strong leader in government is good for Lithuania even if the leader bends the rules to get things done; in comparison, only 27.8 per cent strongly disagreed and somewhat disagreed. Comparison with other European countries only strengthens the argument. In the 2012 wave of European social survey, Lithuania was only third-to-last among 29 covered countries according to the statistical mean of the variable, measuring how important is to the respondents to live in a democratically governed state (7,49 on a scale from 0 to 10; only Russia and Ukraine had lesser means). 1 15min.lt, Lithuanian President to Deutsche Welle: I have allergy to membership in political parties. [Accessed ] <

19 In the 2008 European values study, Lithuania was 14 out of 46 countries according to the support for political system with a strong leader with 52 per cent evaluating it fairly good and very good.

20 Figures and tables 7,0 6,0 5,7 5,0 4,0 3,0 4,5 3,6 4,4 3,0 3,9 3,2 3,5 2,0 1,0 0,0 President Government Parliament Parties Figure 1. Trust in political institutions. Source: Lithuanian national electoral studies, 2012 and Mean in a scale from 0 to 10, where 0 is not trust at all and 10 is complete trust.

21 1 table. Vote share of the candidates by the main Lithuanian parties in 2009 and 2014 presidential election TS-LKD LRLS 11,7 13,6 LMLR TT 6,1 LLRA 4,7 8,2 LVŽS 4,1 DP 3,6 12 Non-partisans and nonparliamentary parties 72,7 60,5

22 Not trust at all Complete trust in president institution Figure 2. Trust in president institution (horizontal scale) and average trust in political parties. Source: Lithuanian national electoral studies, 2012 and Scale from 0 to 10, where 0 is not trust at all and 10 is complete trust.

23 Stable Only 2016 Switching Only 2012 Do not vote Parties President institution Figure 3. Trust in political institutions and stability of voting behaviour in parliamentary elections. Source: Lithuanian national electoral study Mean in a scale from 0 to 10, where 0 is not trust at all and 10 is complete trust.

24 Figure 4. Voters sympathy towards Dalia Grybauskaitė. Source: Lithuanian national electoral study 2016.

25 Figure 5. Average predictions of sympathy towards Grybauskaitė by propensity to vote for parliamentary parties.

26 References Amorim Neto, O.& Strøm, K. (2006). Breaking the parliamentary chain of delegation: Presidents and nonpartisan cabinet members in European democracies. British Journal of Political Science 36(4), pp Auers, D. (2015). Comparative Politics and Government of the Baltic States. Palgrave Macmillan, New York. Bartolini S. ir Mair P. (2007). Identity, Competition and Electoral Availability: The Stabilisation of European Electorates Essex: ECPR Press, 2007 (first published in 1990 by Cambridge University Press). Biezen, I., Mair P. & Poguntke T. (2012). Going, going gone? The Decline of Party Membership in Contemporary Europe. European Journal of Political Research, 51(1), pp Bormann, N. & Golder, M. (2013). Democratic Electoral Systems around the world, Electoral Studies 32, pp Duverger, M. (1980) A New Political System Model: Semi-Presidential Government. European Journal of Political Research, 8(2), p Elgie, R. (1999). The Politics of Semi-Presidentialism. In Robert Elgie (ed.), Semi- Presidentialism in Europe, Oxford: Oxford University Press, pp Elgie, R. (2008). The Perils of Semi-Presidentialism. Are They Exaggerated? Democratization, 15:1, pp Elgie, R., Moestrup, S. & Wu, Y. (2011), Semi-Presidentialism and Democracy. Palgrave Macmillan, New York. Haerpfer, C.W. (2002). Democracy and Enlargement in Post-Communist Europe. Routlege, New York and London. Ibenskas, R. (2014). Activists or money? Explaining the electoral success and persistence of political parties in Lithuania. Party Politics, 20(6), pp Jurkynas, M. (2004). Emerging cleavages in new democracies: The case of Lithuania. Journal of Baltic Studies, 35 (3), pp Jurkynas, M. (2014). The Parliamentary Election in Lithuania, October Electoral Studies 34,, pp Krupavičius, A. (2008) Semi-presidentialism in Lithuania: origins, development and challenges. In Elgie R. & Moestrup S. (eds.), Semi-presidentialism in Central and Eastern Europe. Manchester University Press, pp

27 Krupavičius, A. (2013). Lithuania s president: A formal and informal power. In V. Hlousek, Presidents above parties? Presidents in Central and Eastern Europe, their formal competencies and informal power/ Brno: Masaryk University, pp Mainwaring A., Gervasoni C. and Espana-Najera A. (2016). Extra- and within-system electoral volatility. Party Politics, published early online, January, doi: / Marsh M. (1998). Testing the Second-Order Election Model after Four European Elections. British Journal of Political Science 28 (4), pp Novagrockienė, J. (2012). Lietuvos partinės sistemos raida. Stabilizacijos problemos. Paper presented at VU TSPMI and LPA annual conference Lietuva po Seimo rinkimų Pasquino G. & Valbruzzi, Marco. (2012). Non-partisan governments Italian-style: decision-making and accountability. Journal of Modern Italian Studies, 17:5, pp Pop-Eleches, G. (2010). Throwing out the Bums: Protest Voting and Unorthodox Parties after Communism. World Politics, 62 (02 ), pp Ramonaitė A. (ed.) (2009). Partinės demokratijos pabaiga? Politinis atstovavimas ir ideologijos, Vilnius: Versus Aureus. Ramonaitė, A. (2007). Posovietinės Lietuvos politinė anatomija, Vilnius: Versus aureus. Ramonaitė, A. (2010). Kodėl žmonės nestoja į partijas? Lietuvos gyventojų požiūrio į partinę narystę analizė. Politologija, 58(2), pp Ramonaitė, A. et al. (2014). Kaip renkasi Lietuvos rinkėjai? Idėjos, interesai ir įvaizdžiai politikoje, Vilnius: Vilnius universiteto leidykla. Roberts, A. (2008). Hyperaccountability: Economic voting in Central and Eastern Europe. Electoral Studies, Vol. 27 (3), p Samuels, D. J. & Shugart, M.S. (2013). Party capacity in new democracies: how executive format affects the recruitment of presidents and prime ministers. Democratization, 21(1), pp Sartori, G. (1994). Comparative Constitutional Engineering. An Inquiry into Structures, Incentives and Outcomes. New York University Press. Sedelius, T. & Mashtaler, O. (2013). Two decades of semi-presidentialism: issues of intra-executive conflict in Central and Eastern Europe East European Politics 29 (2), pp Sedelius, T. & Ekman, J. (2010). Government and Opposition, 45 (4), pp

28 Sedelius, T. (2006). The Tug-of-War between Presidents and Prime Ministers: Semipresidentialism in Central and Eastern Europe. Örebro Studies in Political Science 15. Shugart, M.S. & Carey J.M. (1992), Presidents and Assemblies. Constitutional Design and Electoral Dynamics. Cambridge: Cambridge University Press. Sikk, A. & Hanley Sean. (2016). Economy, corruption or floating voters? Explaining the breakthroughs of anti-establishment reform parties in Eastern Europe. Party Politics, 22(4), pp Talat-Kelpša, L. (1996), Pusiau prezidentizmo link: Lietuvos institucinės reformos analizė. Politologija, 7(1). Appendix: Party Acronyms Acronym Party Party name in Lithuanian TS-LKD Homeland Union-Lithuanian Christian Democrats Tėvynės Sąjunga-Lietuvos Krikščionys Demokratai LICS Liberal and Centre Union Liberalų ir Centro Sąjunga LRLS Liberals Movement of the Republic of Lithuania Lietuvos Respublikos Liberalų Sąjūdis LSDP Lithuanian Social Democratic Lietuvos Socialdemokratų Partija Party DP Labour Party Darbo Partija TT Order and Justice Tvarka ir Teisingumas LLRA Electoral Action of Poles in Lietuvos Lenkų Rinkimų Akcija Lithuania LVŽS Lithuanian Farmers and Green Union Lietuvos Valstiečių ir Žaliųjų Sąjunga