Elections and economic crisis in Southern Europe

Transcription

1 Elections and economic crisis in Southern Europe The EU at the ballot box? Working paper to be presented at the 5 th ECPR Graduate Student Conference in Innsbruck Abstract In the current context of economic crisis, the EU has stepped up and tried to play a major role in the design of the policies that tackle the economic crisis. This has highlighted the impact of multilevel governance and how it is a strain for democracy. When governments are not the only ones making decisions, who is to be accounted for in an election when outcomes are not what citizens expected? If European Parliament Elections are not used to evaluate European institutions (Hix and Marsh, 2007), could citizens express their views through their vote in general elections? This work attempts to give a tentative answer to the question. The countries selected, Southern European countries, are the worst hit by the economic crisis and where the role of the EU has been most salient. Therefore, it would be plausible for these citizens to be the first ones to express their concern. In these countries, survey data point to a deep change in attitudes towards the European Union. That party choice voiced these attitudes seems plausible. Analysis performed using post-electoral surveys from Greece (May 2012), Italy (February 2013), Portugal (June 2011) and Spain (November 2011) show a certain relevance of the EU in citizens voting decision. KEY WORDS: voting behaviour, party choice, European Union, Southern Europe, economic crisis. IRENE SANCHEZ VITORES Universidad Autonoma de Madrid Irene.sanchezvitores@gmail.com

2 Elections and economic crisis in Southern Europe: the EU at the ballot box? 1. Four pictures to compose a landscape. In 2010, the Financial Times popularized the term PIGS. Although the financial press had been using it since the 1990s, it was since this moment that it became of common use (Krouse, 2012). But what did this acronym stand for? It was a quick way of naming Portugal, Italy, Greece and Spain. Apart from being Southern European countries, they share a similar economic structure and, therefore, react in a similar way to the ups and downs of the economic cycles. At the beginning of the economic crisis, the governments of these countries were probably aware that hard times were coming. What seems hardly foreseeable is what actually came. A preliminary look at the results points to phenomena that need to be explained, however before going into the results it deems necessary to draw a picture of the context in which these elections were held. The international economic crisis that began in 2010 deeply affected the countries of the Eurozone and it later developed into a sovereign debt crisis that hit the hardest in Southern European countries, precisely those included in the PIGS club. Although this countries share many features in terms of economic, social and political structure, the crisis did not hit them all to the same extent and the elections did not take place always at the same [1]

3 stage (Bosco and Verney, 2012). Insofar as the consequences of the economic crisis had different nuances from one country to another, it seems plausible that there are variations in terms of the political consequences. Greece is probably the country where the economic downturn has materialized in the most dramatic consequences. For instance, it was in this country where the sovereign debt crisis set off. In a very short period, unemployment rates went from a 7.7 percent in 2008 to over 24 percent in 2012, the year when elections were held 1. Unemployment would still rise in 2013, to surpass the 25 percent mark for the general population. Amongst young people the statistics were even more worrying: youth unemployment was over 55 percent. In addition, to the depressed economic situation, the government was forced to cut public spending (over 2 billion in the period ) and accept two bailout packages to avoid defaulting. For over three years, Greek sovereign debt was out of the international markets due to the incredibly high borrowing costs they had to agree to. As a consequence, the standards of living of the Greek collapsed and the political system also suffered. The set back of the welfare state in Greece has been replaced by citizen solidarity and family support in most cases. However, it has also created a window of opportunity for the extreme right to cultivate a base of supporters as it had not been seen before (Dinas and Lamprini, 2012). The Troika, name given to the three institutions that intervened the Greek economy, namely the IMF, the ECB and the European Commission, also forced the PASOK government to resign and a new Grand Coalition government came to office. In this shaken context, the election held in May 2012 was an earthquake for the party system: Nea Democratia (ND) won the election closely followed by the radical left party 1 All the statistics on unemployment and public spending quoted in this section have been obtained from EUROSTAT. [2]

4 SYRIZA, the extreme right Golden Dawn (XA) performed extremely well and the socialist PASOK collapsed. The result was so inconclusive that new elections had to be held a month later and a government led by the conservative Antoni Samaras (ND) was appointed. Given that this second election consolidated what happened on the first, only the May election will be taken into account in this paper. Figure 1. Greek turnout for May and June 2012 elections June 2012 New Democracy (ND) Panhellenic Socialist Movement (PASOK) Communist Party of Greece (KKE) Democratic Left (DIMAR) may-12 Coalition of the Radical Left Unitary Social Movement (SYRIZA) Independent Greeks (ANEL) Golden Dawn (XA) Popular Orthodox Rally (LAOS) Source: Greek Ministry of Interior. Although Italy was not formally bailed out, the economic crisis had no less impact over the Italian political system. In a short period of time, the PdL had many open fronts. The combination of the economic downturn and the sovereign debt crisis caused Italy to borrow money at historically high prices, compromising the capacity of the Italian government to respond to its obligations. The crisis deepened when allegations of corruption and improper behavior on behalf of Berlusconi appeared on the media and [3]

5 travelled around the world. The situation worsened when Berlusconi s government openly rejected to put in place the policies and spending cuts proposed by the European Commission. International pressure increased and the Berlusconi government went to Parliament for support in the form of a vote of confidence. However, Parliament was unwilling to back Berlusconi against the EU and even some members of the PdL left the party. To avoid the alleged instability that an election would bring, a technocratic government led by Mario Monti and full of experts in various fields was appointed. This new cabinet was expected to wade through the economic toughness, run the country and prepare it for the election. PdL and PD supported this government in Parliament while they harshly criticized its measures in the media (Vegetti et al., 2013). Unemployment climbing to an unseen 13% led to the development of widespread public dissatisfaction amongst the general public and numerous demonstrations while unemployment grew and living conditions worsened. The February 2013 General Election attracted everyone s attention due to not only the shaken context but also because they were to be held under new rules (Segatti, 2014). The results showed that no matter the cautions, voters still have the last say. The PD won the election by a couple of votes but they were so closely followed by the Movimento Cinque Stelle and the PdL that the majority price awarded to the party with the most votes did not configure a majority big enough to decide the government. After several rounds of talks, Paolo Bersani from the PD managed to get appointed to form a government although his support was weak. [4]

6 Figure 2. Turnout for Italian 2013 election. 30,0 25,0 20,0 15,0 10,0 5,0 0,0 PD Sinistra e liberta SVP* LN Movimento per l'autonomia All. Per il Sud Centro Unione di Centro Movimento association italiani a'll Estero Di Pietro Italia dei valori Centro democratico PdL Fratelli d'italia Movimento 5 Stelle Scelta civica per Monti Futuro e Libertà Autonomie, Liberté Democracie Source: Italian Ministry of Interior. Amongst the four countries considered, Portugal was the other one that suffered the worst economic situation, to the extent that they also required a joint IMF/ECB bailout package. As a result of the economic hardship, public spending was cut over a 25%, privatizations were accelerated and salaries and public benefits were cut at the same time that taxation increased. Unemployment rose to an unseen 14% amongst the general population and a 35% amongst young people. As described by Magalhaes (2012), public contestation to austerity measures became widespread in crowded demonstrations and numerous strikes. In the 2011 election that took place just after the signatures of the Memorandum by the three main parties and the rejection of the fourth austerity package proposed by the Portuguese government. In contrast to their Italian and Greek neighbors, the turmoil in the Portuguese political arena did not manifest in a complete change of the party system. The incumbent PS party fell to historically low results while his main competitor, the [5]

7 PSD, comfortably won the elections. As a result of this election a new government led by Pedro Passos Coelho was appointed in Portugal. Figure 3. Turnout for the Portuguese 2011 election. 45,0 40,0 35,0 30,0 25,0 20,0 15,0 10,0 5,0 0,0 Social Democratic Socialist People's Party Democratic Unity Coalition* Left Bloc Invalid votes Blank votes Source: Portuguese Ministry of Interior. As the countries previously described, Spain has been harshly hit by the economic crisis. The bursting of the housing bubble that had supported economic growth in the previous years, combined with the high price at which the Spanish government was borrowing money in the international markets were the main manifestations of the economic downturn. As a result, not only did public spending experience cuts, but also unemployment reached historically high rates, getting close to 25% amongst the general population and over 50% amongst young people. Living standards did not fall as much as in Greece, but the situation became tough for many families. Demonstrations and strikes increased as a response to the worsening situation. In this troubled context, the PSOE government advanced a couple of months the elections (Anduiza et al., 2013; Muñoz et al., 2014). The election was held in November 2011 and the results, if it were to be compared to the previously described countries, was more similar to the Portuguese than the Italian or the Greek. Volatility increased in comparison to previous elections and [6]

8 the incumbent had bad results but the party system stayed more or less the same. The collapse of the PSOE allowed for PP not only to win the election but also to obtain an absolute majority in Parliament. Small parties gained more votes due to discontent, although it was not enough to produce a realignment of the party system (Martín and Urquizu-Sancho, 2012). As a result of the election, Mariano Rajoy was appointed Prime Minister and asked to form government. Figure 4. Turnout for the Spanish 2011 election Partido Popular Partido Socialista Obrero Español Izquierda Unida Convergencia i Unió Partido Nacionalista Vasco Unión Progreso y Democracia Esquerra Republicana de Catalunya Source: Spanish Ministry of Interior. All four countries considered went, and still are going, through economic hardship. This is not a new situation by itself. Governments have resigned due to bad economic results, people have demonstrated against unwanted policies, and still there is something new to this context. The degree of intertwining of the economic and political systems has resulted in a stage that demands for a different way of understanding politics (Bardi et al., 2014: 244). The common currency and the EU institutions have resulted in a very complex decision-making process that forces governments to design policies coordinating with other member states (Fabbrini, 2013: 1004). This system, although slow, worked in the [7]

9 preceding years because prosperity masked differences between countries (Ayala, 2012: 46 48). However, when the hard times have come to test this institutional configuration, the problems have surfaced and the inequalities between countries have deepened (Laffan, 2014). In addition, the proposed solutions have not been very democratic, the relevant actors have preferred solutions articulated over technocracy and public spending cuts. As a result, the democratic deficit has reappeared as more than a merely theoretical problem as different authors had expected in previous years (Majone, 1998; Tillman, 2004: 603): the EU is having actual impact over the welfare states of citizens. Is it possible that this makes the EU a relevant issue for voters in national elections? 1. Making the EU matter for voters or not. EU issue voting can be defined as the process through which attitudes towards the European Union shape a citizen s voting decision (de Vries, 2007). Traditionally, the European Union has been pictured as a technical issue, distant from citizens everyday lives and worries and therefore distant from the voting decision. The EU has a complex institutional design and decision-making process that have contributed to the idea that this is a technical issue were governments are the only ones with the knowledge to make decisions (Carrubba, 2001; Henderson and Sitter, 2008). Despite the fact that the EU had competences over matters close to the everyday concerns of citizens such as safety rules in the working environment or food labelling regulations, they are not salient issues (Díez Medrano, 2003: 22). EU regulations shape a good number of the laws passed in member countries but they usually do not get credit for political decisions. However, this situation [8]

10 changed after The outbreak of the economic crisis has come accompanied by a change in the role given to the EU. The institution which had brought democracy and prosperity to Southern Europe is now promoting cuts in public expenditure, bailouts and other sorts of economic sacrifices. At the same time, the EU has tried to have a more predominant role in the design of economic policies in the EMU member States. This increase in prominence has not brought better evaluations of the EU, instead it has highlighted the lack of accountability and how it is an actual issue of concern for citizens (Loveless, 2010: 1084; Schwarzer, 2012). Despite the fact that Peter Mair (2000) in one of his works rejected the europeanization of the national political systems of member States, the truth is that the EU is playing an increasingly relevant role. The EU has developed into a peculiar international institution whose peculiarities have been enhanced after the passing of the Treaty of Lisbon and has taken integration into a further stage (Bickerton, 2013). This has meant that, compared to previous economic crisis, mainly due to the cession of monetary policies, the EU has had a more dominant role in the political arena (Fabbrini, 2013: 1008). Up until now, support for the EU has been built not only over ideological beliefs but also highlighting the benefits of membership as the key to economic development and a guarantor of democracy (Jackson et al., 2011). Even though its functioning was not strictly democratic, decisions where perceived as positive and there was consensus on what EU membership meant (Carrubba, 2001: 142). From 2011 onwards, citizens attitudes towards the EU start to show a change towards criticism and discontent with the decisions that were being made (Verney, 2011). The EU limiting the repertoires of governments is not a new phenomenon (Haughton, 2014: 82; Shu, 2009), however, after the beginning of the crisis it has taken a new face, demanding important sacrifices from the population. [9]

11 This discontent could affect the way in which partisan competition takes place, introducing the EU issue where it used to be excluded (Eijk and Franklin, 2004). Nonetheless, for any given issue to acquire relevance it is not enough for citizens to be willing to express an opinion additionally some other elements are needed (Carmines and Stimson, 1986: 902; de Vries, 2007). First, an issue needs to be presented as a contested issue. Three can be the possible ways in which this can be done: because parties are willing to introduce it into the public debate, because an actor may be interested on getting an issue under the public eye or because a focusing event takes place, centering the attention of the general public (Birkland, 2005: ). Second, it is not enough to present any given issue as contentious, citizens need to perceive that there is a defined problem affecting them in some way (Fossati, 2013; Lewis-Beck and Paldam, 2000: 15 17). This would mean that citizens would pay more attention to the EU and its decision when its proposed policies impact social benefits and public services. The combination of the economic crisis, austerity in public spending and the consequences of it both over citizens all across Europe, and particularly in Southern countries seems to have opened a window of opportunity for the issue to gain relevance (Tillman, 2004: ). Yet, political parties are always careful when they position themselves in this debate (Pennings, 2006), and they are not under the same incentive system, depending on the role they assume in the political system: whether they are government-seeking or minoritarian parties (Nanou and Dorussen, 2013). The European Union is a project of the political centre of the political centre, of the government-seeking parties across Europe (Kriesi et al., 2006; Marks, 2006). These parties, when they shape their electoral manifestos they are more constrained than nongovernment-seekers. These parties need to elaborate their manifestos keeping in mind the obligations that come from governing: internationally-agreed obligations, shared [10]

12 competences with other institutions, etc. In other words, their proposals have to be flexible enough to adapt to the contingencies that may come up during the legislature but they also need to be able to incorporate decisions that have a different ideological inspiration such as policies coming from the EU (Nanou and Dorussen, 2013: 76 77; Neumayer, 2008: 155). In this parties it is essential to keep a coherent ideological discourse because their solvency as office-seekers builds precisely along this idea of being the best available manager of State affairs, in spite of all the existing constraints (Sánchez-Cuenca, 2008: 20 22). Therefore, the competition between these governmentseeking parties is limited and avoids criticizing certain policies under the warning that the other competing parties may be forced to undertake similar moves, betraying to some extent their ideological integrity to respond to the challenges of governance. Minoritarian parties, that is, parties who only aim to achieve seats in Parliament or, at best, to be coalition partners of other parties, have more flexibility when positioning themselves over difficult issues because they do not expect to be held by the constraints of governing (Taggart, 1998: 384). Credibility in these parties is not based in the fulfillment of electoral promises and the achievement of defined goals, instead it is more important for them to keep ideological integrity and loyalty to fundamental principles, allowing them to criticize the shadows of European integration. Moreover, this will not compromise their electoral results, rather it reinforces them, increasing their vote share (Nanou and Dorussen, 2013: 77 78). This flexibility and loyalty to principles allows these parties to exploit the weaknesses of the project, voicing those critical and discontent member of the electorate who do not feel represented by mainstream parties (De Vries, 2010; Verney, 2011). [11]

13 How should the context described be relevant for general elections when the EU has its own parliamentary elections? There are several considerations to be made before going further. The complex situation here described has developed in an environment of deep economic crisis. This Great Recession has been marked by an attempt of the EU to become the main institution when solving the economic difficulties. The reasoning behind was not very elaborate: the international origin of the Recession required and international response (Laffan, 2014). It is not the aim of this paper to address the effectiveness of the solutions adopted. What does come our as relevant is that this way of decision-making diffuses responsibility and it makes it harder for citizens to identify and evaluate institutions for the policy outcomes (Bingham Powell Jr and Whitten, 1993; de Vries, 2010). A very sophisticated citizens should be expected to use each election to analyze each institution s performance separately. However, the literature points precisely in the opposite direction (Bechtel, 2012; Weber, 2011). The national arena tends to concentrate most of the attention, no matter which institution is going to be appointed as a result of the election (Johns, 2010). This effect should be less intense when the outcome of the election is the direct appointment of a government (Henderson and McEwen, 2010). Furthermore, it could be expected that the decision is made taking into account all the institutions intervening. The literature on European Parliament elections has contributed to the idea that the EU could be materially unaccountable (Hix and Marsh, 2007; Reif and Schmitt, 1980). The literature has repeatedly found that European Parliament elections are resolved in terms of national politics, not in terms of EU politics (Hix and Marsh, 2011). In a context where [12]

14 the EU is more relevant in national politics, it could be expected that the EU was taken into consideration by citizens when deciding who to vote for in general elections. 3. Methodology 3.1 Hypothesis Assuming that citizens took into account the EU when deciding who to vote for in the last general election, to what extent did it influence the voting decision? This is the question I attempt to answer in this paper and there are different hypotheses to which the literature points as plausible. Hypothesis 1. Blaming the EU for the poor management of the economic crisis has no relation to the voting decision. When deciding their vote, citizens take into account the economic context and how governments have managed it (Duch and Stevenson, 2006). Their evaluation is not completely rational because few citizens have enough knowledge to understand autonomously the economic situation (Fraile, 2002, 2007). Therefore, they rely on cues and heuristics such as ideology or what their trusted media says to make this judgment (Lau and Redlawsk, 2006). The electoral campaign that preceded the elections here considered was practically monopolized by the bad economic situation and the poor governmental performance (Magalhães et al., 2012). However, there was such degree of agreement that the perception of the economic situation was no variable to explain the voting decision (Fraile and Lewis-Beck, 2013). Authors like Mariano Torcal (2014) have attempted alternative strategies although they are deeply complex and loose parsimony. [13]

15 On the other hand, using a variable of attribution of responsibility, a proxy could be created to measure how citizens perceived the outcome of governmental activity and who is to blame for the bad situation. This first hypothesis focuses on the idea that, despite all intervening institutions, the government is the one with the final say over what is to be done in terms of economic policy. The turnout results would reflect not only citizens punishment for the poor economic situation but also the party in government s incapability to formulate feasible plans of future to tackle the economic crisis (Maravall and Przeworski, 1999). Hypothesis 2. Controlling for partisanship, those citizens who blamed the European Union for the management of the economic crisis will probably behave differently to those who blamed the government. Systems of multilevel governance distribute decision-making amongst a wide variety of political actors which diffuse the responsibility that each assumes towards voters, blurring demands for accountability (De Vries et al., 2011; Hellwig, 2007; Parker-Stephen, 2013). This hypothesis explores how citizens are sensible to this complexity (Aguilar and Sánchez-Cuenca, 2005), while the following two subhypotheses propose the sense in which this influence would go. Hypothesis 2.1. Citizens feeling less close to the party in government and who blame more intensely the EU than their national governments, have a higher probability of voting for parties critical towards the European Union. Political parties carefully position themselves towards the EU issue, foreseeing that Brussels can limit their capacity of executing their manifesto. Alternatively, minoritarian political forces interpret coherence otherwise. This allows them to campaign shedding a light on the shadows of the integration project, an asset that should increase the [14]

16 probability of getting votes from discontent citizens (Gramacho and Llamazares, 2007; Nanou and Dorussen, 2013). Hypothesis 2.2. Citizens feeling close to the party in government and blaming the EU for the poor management of the economic situation more than they blame their national government, should have a higher probability of exonerating the incumbent, voting for this party. Voters feeling sympathetic for the incumbent party should be more sensible to the cues provided by the government to interpret reality (Lau and Redlawsk, 2006). Thus they would more willingly accept the message proposed by governments that the hardships and sacrifices and the decisions that enacted them come as direct demands from the European institutions (Menéndez, 2012; Stokes, 2001). This blame out argument would allow citizens to put behind them the cognitive dissonance that blaming the incumbent and voting for him would mean. In other words, convincing themselves that the decisions came from somewhere else allow these citizens to keep voting for the party they feel close to. 3.2 Case selection As it was states in a previous section, European Parliament elections should be the ones where it was measured at best what citizens think about the EU s management, not only of economic issues, but also of the different areas of competence. However, and still reiterating what has already been said, the literature on these elections has found that they were discussed in national terms (Hix and Marsh, 2007, 2011). The last European Parliament elections have been held so recently (May 2014) that it is too soon to make any reasonable affirmation of what this turnout means for the theory. The fact that the [15]

17 European Parliament elections do not serve this purpose does not necessarily mean that citizens do not evaluate the EU s political outcome in some way. In this sense, it would be likely to think that they used national or general elections to voice their concerns over the way in which the EU is managing issues, given that they perceive a direct consequence over their everyday life (Marks, 2006). The role of the European Union in this crisis has been highlighted by governments across Europe. Ideally, and future developments will do it, this study should cover every European country. In one way or another, every country has been affected by the crisis so there is no reason to reject any country. However, this piece of research is at an explanatory stage, constrained by formal issues such as time and extension. As a consequence, in this study only Southern European countries have been included, namely Greece, Italy, Portugal and Spain. I am aware that some may criticize this choice because it introduces a bias. Anyways, this bias is intentional insofar as the presence of the EU in national elections is a topic for which there is a limited amount of literature. If the EU were to be a relevant issue in general elections of any country, it is reasonable to think that where the economic crisis has hit the worst it would be more visible. That is to say, Southern European countries should be the first countries where the presence of the EU as a relevant issue in national elections could be observed. After the beginning of the crisis, the compromises that are a result of EU membership have become heavier and Brussels is no longer just a source of wealth and prosperity (Fabbrini, 2013: ). This change has been particularly intense towards Southern Europe. These countries were in particularly bad shape when it came to facing the economic crisis and, as a consequence, the results have been especially salient (Serricchio et al., 2013: 58). Socially, conditions have worsened to a bigger extent. But also, politically, these countries have suffered the consequences of giving up their sovereignty. [16]

18 Governments have had their hands tied by the compromises voluntarily attained. Additionally, these governments have been constrained and their sovereignty weakened because of their poor economic situation (Laffan, 2014). All in all, the differences between countries inside the EU have increased, hurting the existing consensus over EU issues and turning it into a matter over which parties could compete (Bosco and Verney, 2012; Verney, 2011). It could also be argued that an EU with strong claims over what any given policy of a country should be is not as new as it could be thought of. If the scope of analysis widened from economic policies to other sorts of political decisions in which the EU imperatively demands for precise policies should also be taken into account. In this sense, the democratic conditionality, that is, the exigencies the EU negotiates with countries that are applying for membership on topics such as the protection of human rights or frontier relations with neighboring countries little has to do with mere advice (F Schimmelfennig et al., 2003: ). It could be likely that these countries would have experienced similar phenomena. In spite of this, I have decided to exclude these countries from my analysis. On the one hand, democratic conditionality refers to delicate issues, intimately linked with parts of sovereignty that countries usually do not share willingly. On the other hand, party systems do not align in the same way when the country is a member State than when it is an applicant. Before joining the club, positioning over this issue is just another way of stating their world view (Jackson et al., 2011; Marks et al., 2006). Whereas once they become members of the club things change. Membership is seen as a one-way trip and exit means a political cost that no actor is willing to assume (Haughton, 2014). Countries inside this group can again be grouped into two categories: those countries who joined the EU in the last enlargement (Central and Eastern Europe) and applicant members of the Mediterranean arch (Turkey and Cyprus). [17]

19 In the case of the countries from Central and Eastern Europe, even if this grouping simplifies a very rich reality, the issue of EU membership has been very relevant to the political arena. Moreover, this issue has been crucial to party competition whereas it has not been so competed in the selected countries. In these countries the issue presents itself as a choice: either they joined the EU or they continued under the sphere of influence of the Russian Federation. In these countries, the EU has been presented with the most intensity as a source of economic development and democracy (Marks et al., 2006). In addition. There are government-seeking parties with a negative perception of the EU membership, willing to support a close collaboration with the Russians (Taggart and Szczerbiak, 2013). Besides, the differences in party s configuration and the economic situation, these States have recently entered the EU and their status is still peculiar. Cyprus and Turkey, the parties included in the second group, could be more easily compared to Southern European countries. Even if many differences separate them, there are also many similarities that would allow a viable comparison. Even there is no regional power altering what membership means, as it did happen in the previous group of countries. Previous studies include strategies to compare all of these countries (Bellucci et al., 2012; Verney, 2011). In the first stages of the design, there was an attempt to consider them, they have finally been discarded because membership is far more relevant for these countries and negotiations have been stagnated for so long that in both cases it does not seem that membership is going to be granted any time soon. In the case of Cyprus, some of their mainstream parties have positioned themselves not only critically towards EU membership but also plainly against its entry (Christophorou, 2007). In Turkey, parties place themselves in a similar way with an addition: negotiations for the application procedure started over ten years ago and they have remained stagnated to an extent that the Turkish political parties have the feeling that they will remain candidates [18]

20 forever (Frank Schimmelfennig et al., 2003). As a consequence, the way in which they agitate the issue in elections or how they position, even amongst mainstream parties (AKP and CHP) are inconsistent and they change with each party s leadership (Baris Gulmez, 2013). In the last couple of years the leaders of the main parties appear to favor the entry in the EU, even if the agenda of the Erdogan government seems to evade from truly fulfilling the goals set for them in the entry negotiations as part of the democratic conditionality. By contrast, the immediate predecessors of Erdogan at the head of the AKP were openly Eurosceptic. 3.3 Variables This section will present the variables used in this paper to elaborate the multinomial logistical regressions used to test the hypotheses, further descriptions can be found in the Figure 5 in the Annex. The dependent variable here chosen captures for citizens voting decision. Not only so, accountability is also to be operationalized in order to look into how citizens administered punishments and rewards for the governmental outcome. To achieve both aims the variable has been operationalized taking into account the voters that voted for the incumbent or one of his main competitors. Operationalizing party choice is not a decision without consequences, particularly for the selected set of elections. All four cases have in common multiparty systems that have suffered a profound shock in the last election. In Spain and Portugal, incumbents have been intensely punished and the effective number of parties has increased. By contrast, the Italian and Greek party systems have suffered an earthquake that may or may not remain in coming elections. As a consequence, choosing the three best performing parties in these elections will, on the one hand exclude [19]

21 a significant number of voters, and, on the other hand, reduce the reliability of results. In figure 6 (see Annex) the scheme of party competition considered. Two are the independent variables considered: the responsibility attributed to the EU against the responsibility attributed to governments and closeness or sympathy for a given political party. The first variable considered confronts the responsibility attributed to national governments against the responsibility attributed to the EU for the management of the economic crisis. There are several reasons to choose this variable configuration rather that other possible coding. First of all, an ipsative question, that would something such as Who do you think is the main responsible for the economic crisis? or some equivalent wording would not suit the purposes here aimed for several reasons. Technical issues would impede comparison. Even if the question is coded in different ways in the surveys considered, respondents are only asked to choose one institution as the main responsible for the economic crisis in the Spanish and Portuguese questionnaires. Furthermore, other institutions such as the banking system and the international economic situation are included amongst the options. As a result, most of the respondents choose financial institutions such as banks or international markets rather than political ones (governments, etc.). This should work against the relevance of this piece, however, I think that the picture drawn would be too simplistic and it deserves to look into it in depth. In all four questionnaires, respondents are asked to evaluate how much to blame are a number of institutions (banking systems, government, EU, citizens, etc.) for the economic crisis. The hypothesis aim to look into how the game of responsibilities has played and it was not enough to include the variables corresponding to how much each political institution is accountable. This is the reason behind including an index, a more complex variable. In this index, the responsibility attributed to the EU is subtracted to the [20]

22 responsibility attributed to the government. The resulting variable permits to order respondents according to the differential of responsibility attributed to the two main institutions involved in decision-making during the economic crisis. In other words, this allows a measure of whether blaming more the EU than the national government for the poor economic situation has an impact on voting behavior. The second variable included is party identification or closeness to a political party. Although it is not the main explanatory variable, it is relevant to the hypothesis because it is a way of capturing those voters that could be more sensible to cues by the party in governments and, thus, to use the EU as a fall guy for the poor performance of their party while in government. At some point, it could play as a means to justify why they keep voting for their preferred party. Whereas those who don t have this party identification but still feel that the EU is responsible for what is going on should be more prone to vote for eurocritical or Eurosceptic parties. The control variables here chosen are the ones that the literature on voting behavior usually choses as controls (Fraile, 2002): gender, age and education. These studies also introduce usually ideology as a control, however the lack of this variable in the Portuguese questionnaire has prevented me from introducing. However, this is not an issue because ideology was just a complementary control to partisanship. That is, that citizens would put the blames of the economic crisis according to their ideological self-placement rather than to partisan cues. This could be it but since the relationship between ideology and party identification is complex and the latter has been said to be the main component of the first. Therefore, the absence of ideology in the model should not cause relevant changes in terms of the results to be observed. [21]

23 4. A look at the results: scapegoats and party competition. The first hypotheses proposes that, despite the complexities of modern governance, citizens expect their government to exercise the power they supposedly have (Artes, 2011). No matter their knowledge of the actors involved, these citizens are aware that the final decision is made by governments. Therefore they would expect their governments to include in their political calculations all possible constraints to get done what needs to be done. Many difficulties may arise, but governments should count on them when they decide which course of action to take. As a result, the logical consequence would be that only governments are to be accounted for when the outcome is not satisfactory. The results of the multivariate analysis reject this hypothesis in the four countries considered. The fact that the index of responsibility attributed is a significant variable is in itself meaningful. Authors such as Fraile and Lewis-Beck (2013) have pointed towards the paradox of the relevance of the economy in the last set of elections and its lack of explicative value as an independent variable for vote choice. Maybe the answer to this paradox is simpler: the economy is very important but politics still matter more. Citizens have historically low levels of trust in political institutions but they still expect these institutions to be the ones who find a way out of the crisis. The economy has undoubtedly been salient in these elections, but politics or political beliefs are still the key to whom citizens choose to trust with governmental responsibilities. [22]

24 Figure 7. GREECE General election, May 2012 (outcome based on PASOK) ND SYRIZA Gender a 0.54 (0.63) 0.34 (0.48) Age (0.03)* (0.02) Education b Elementary and basic secondary ( ) 9.50 ( ) Advanced secondary and vocational training ( ) 9.63 ( ) University ( ) 9.47 ( ) Responsibility (0.27) (0.24)* Party identification ND 33.5 ( ) 3.62 (1.44)* Other party identification ( ) 5.15 (1.07)*** Constant ( ) (n) = 276 Pseudo r 2 = 0.61 Notes: *** Statistically significant for a la level of trust of 0,00; ** statistically significant for a level of trust under 0,010; * statistically significant to a level of trust under 0,050. Standard errors have been included between parentheses. a Reference category: Female. b Reference category: without studies. Source: ELNES 2012 post-electoral survey. Figure 8. ITALY General election, February 2013 (outcome based on PdL) Monti per l Italia PD Movimento Cinque Stelle Gender a (0.52) (0.49) (0.49)* Age 0.03 (0.02) 0.03 (0.02) (0.02) Education b Elementary and basic secondary ( ) 5.1 (1.82)** 1.91 (1.81) Advanced secondary and vocational training ( ) 6.09 (1.86)** 2.23 (1.84) University ( ) 7.00 (2.02)** 3.31 (2.00) Responsibility (0.16)** (0.15)*** (0.15)* Party identification PD ( ) 9.79 (1.45)*** 6.56 (1.29)*** Movimento 5 Stelle ( ) (640.41) (640.41) Other party identification ( ) 5.39 (1.08)*** 4.81 (0.82)*** Constant ( ) (2.51)*** (2.32) (n) = 705 Pseudo r 2 = 0.62 Notes: *** Statistically significant for a la level of trust of 0,00; ** statistically significant for a level of trust under 0,010; * statistically significant to a level of trust under 0,050. Standard errors have been included between parentheses. a Reference category: Female. b Reference category: without studies. Source: ITANES 2013 post-electoral survey. [23]

25 Figure 9. ITALY General election, February 2013 (outcome based on PD) PdL Monti per l Italia Movimento Cinque Stelle Gender a 0.60 (0.48) (0.35) (0.31) Age (0.02) 0.01 (0.01) (0.01)*** Education b Elementary and basic secondary (1.75)** ( ) (1.47)* Advanced secondary and vocational training (1.79)*** ( ) (1.48)** University (1.96)*** ( ) (1.52)* Responsibility (0.16) 0.46 (0.14)** (0.11) Party identification PD (1.46)*** ( ) (1.28)** Movimento 5 Stelle (888.39) ( ) 2.10 (1.31) Other party identification (1.08)*** ( ) (1.26) Constant (2.53)*** ( ) 7.06 (2.13)** (n) = 708 Pseudo r 2 = 0.62 Notes: *** Statistically significant for a la level of trust of 0,00; ** statistically significant for a level of trust under 0,010; * statistically significant to a level of trust under 0,050. Standard errors have been included between parentheses. a Reference category: Female. b Reference category: without studies. Source: ITANES 2013 post-electoral survey. Figure 10. PORTUGAL General election, June 2011 (outcome based on PS) PPD-PSD CDU (PCP-PEV) Gender a 0.07 (0.32) 0.44 (0.41) Age 0.00 (0.01) 0.01 (0.02) Education b Elementary and basic secondary 0.16 (0.64) 0.40 (0.93) Advanced secondary and vocational training (0.72) 0.92 (1.01) University 0.26 (0.14) 0.20 (1.21) Responsibility (0.14)* (0.17) Party identification PPD-PSD 6.89 (1.07)*** (743.09) Other party identification 3.18 (0.40)*** 4.39 (1.03)*** Constant (1.25) (1.90)* (n) = 438 Pseudo r 2 = 0.38 Notes: *** Statistically significant for a la level of trust of 0,00; ** statistically significant for a level of trust under 0,010; * statistically significant to a level of trust under 0,050. Standard errors have been included between parentheses. a Reference category: Female. b Reference category: without studies. Source: TNS-ICS 2011 post-electoral survey. [24]

26 Figure 11. SPAIN General election, November 2011 (outcome based on PSOE) Gender a 0.22 (0.12) 0.41 (0.15)** Age 0.01 (0.004)* 0.00 (0.01) Education b Elementary and basic secondary 0.69 (0.28)* 0.68 (0.47) Advanced secondary and vocational training 0.83 (0.32)** 1.51 (0.50)** University 0.70 (0.30)* 1.79 (0.48)*** Responsibility (0.07)*** (0.07) Party identification PP 7.78 (0.45)*** 2.32 (1.12)* Other party identification 3.68 (0.19)*** 3.90 (0.30)*** Constant (0.49)* (0.73)*** (n) = 3334 Pseudo r 2 = 0.46 PP Notes: *** Statistically significant for a la level of trust of 0,00; ** statistically significant for a level of trust under 0,010; * statistically significant to a level of trust under 0,050. Standard errors have been included between parentheses. a Reference category: Female. b Reference category: without studies. Source: TNS-ICS 2011 post-electoral survey. IU The different regression models shown in Figures 7 to 11 treasure more information to be taken into account. Indeed, the role of governments and the EU are relevant to citizens voting decision. However, which way did these considerations take? Citizens that blame more the EU than their government for the poor economic situation were expected to behave differently than citizens that blame the government. Those who blame the government are expected to punish the incumbent parties to a larger extent. However, amongst those who blame the EU, behavior is not expected to be homogeneous either. Incumbent parties have repeated in their speeches that unpopular decisions are a demand from Brussels to grant the needed help to escape default. Bad management of previous administrations has become almost a taboo, while they repeated as a mantra that Brussels came to help after years of wasting the wealth of prosperous years and spending above everyone s possibilities. However, citizens do not feel strongly this sense of shared responsibility, so here party identification seems relevant. Those who feel close to the [25]

27 incumbent party are said to be more sensible to its cues, while those who do not have this sympathy would be eager to search amongst the available party supply for other options. To make it clearer, I will propose an example of how a voter would behave if these hypotheses were to be true. Feeling disappointed with how the EU has managed the economic crisis, a Greek citizen that sympathized with PASOK, in May would have a higher probability to vote for this party than another voter that put the blame more intensely on the Papadopoulos government. On the other hand, if this same citizen did not have sympathy of any kind for PASOK, he would be more prone to have looked for a party critical with how the EU is being managed, or even Eurosceptic, which should result in an increase of his chances of voting for SYRIZA. Before going into a further description of the results, it is remarkable how party supply is unequal in the countries considered. While in Italy and Greece there are openly critical parties with the EU such as Movimento Cinque Stelle (Beppe Grillo s Five Star Movement) or SYRIZA, even hard eurosceptics such as Golden Dawn (XA). Portuguese and Spanish parties have remained quite silent over the topic and although some voices have publicly criticized the path taken by EU s decision, they have remained shy to say the least, fearing traditionally pro-european societies (Ares Castro-Conde, 2013; Nanou and Dorussen, 2013). The results of the four models show a consistent trend across the four countries, which seems a finding in itself. Citizens of the countries worst hit by the economic crisis have behaved in an equivalent way. Controlling for partisanship, citizens that blame the EU more than their national government for the bad economic situation have a higher chance of voting for the incumbent than its main competitors. However, the first subhypothesis is not confirmed under the light of the results obtained. [26]

28 All in all, in these countries the EU seems to be the scapegoat for the incumbent parties. But before going into the concluding remarks, I would like to detain over the Spanish survey. With over 5000 respondents, against the that the other surveys have, this survey has the advantage that it allows for more sophisticated analyses to be performed. In particular, it permits the construction of a term of interaction with the differential attribution of responsibility and partisanship. To allow for the results to be statistically significant, the variable of differential attribution of responsibility has been simplified into three coding categories: 1.- those who blame to a larger extent their government than the EU, 2.- those who blame both institutions to the same extent and 3.- those who blame the EU to a larger extent than their national government. This allows for a stronger control: it allows testing for what happens when both conditions take place at the same time and it allows for a precise picture of precisely how party sympathizers behave. Figure 12. Graph for the interaction between differential attribution of responsibility and party identification. Spain, Vote for PP vs. PSOE Vote for IU vs. PSOE Pr(Iupsoe) Government Both EU Differential attribution of responsibili Government Both EU Differential attribution of responsibili PSOE Other PP PSOE Else IU Source: CIS 2011 post-electoral survey. [27]