Bargaining for Self-Determination: The Secession Game in the EU

Transcription

1 Bargaining for Self-Determination: The Secession Game in the EU Adam Holesch presented on the 7 th September 2017 at the ECPR General Conference in Oslo 1

2 Abstract: This article develops a game-theoretical model of negotiation over secession in the multilevel framework of the European Union (EU). After extending the traditional calculation of the secession costs to include the supranational level and the control over the monopoly of legitimate violence, this article explains the importance of credible threats in centre-periphery bargaining. By modelling two games, it shows that since democratic limitations do not allow for military confrontation the sub-state unit has few possibilities to develop a credible exit threat. In contrast, the central state can build a reputation as a tough player, which allows it to control the game. This outcome can change if the EU engages in the conflict and makes the sub-state s threat credible. I test this model with the Spanish State of Autonomies. This plausibility probe confirms that if the setting of this game does not change, European sub-state units have no possibility of seceding. Keywords: secession, Spain, EU, Catalonia, game theory, threat Secession in Western democracies is rare (Dion 1996). Nonetheless, secessionism, which I define following Wood (1981:110) as the demand for formal withdrawal from a central political authority by a member unit [ ] on the basis of a claim to independent sovereign status is frequent. In the last few decades, we have witnessed in many substate units or regions of the European Union (EU) such as Catalonia, the Basque Country, Scotland, Flanders, Brittany, Corsica, Val d Aosta, Padania, Veneto, and South Tyrol the emergence of a new political phenomenon: secessionism within the multilevel framework of EU. Even if the literature has done much to explain why secession movements occur all over the world (Hechter 1992; Alesina and Spolaore 2003; Colomer 2007; Keating 2001; Sambanis and Milanovic 2014) and when these groups are more likely to succeed (Sorens 2005; Siroky and Cuffe 2015), there is barely any empirical discussion on secession in an exclusively democratic context (Young 1994, 2004). With the exception of Griffiths et al. (2015), we still do not know much about how the secession conflict is 2

3 played when a democratic state is a member of European structures. The politics of secession in the EU is still for the most part an enigma. This article will close this gap in the literature by asking: Under which conditions can a sub-state unit secede within the EU s multilevel framework? It will also examine how the European level of governance can influence the secession game. I claim that in a democratic state the centre has a more privileged position in the secession game for endogenous and exogenous reasons. The central state internally controls the monopoly of legitimate physical violence (Weber 2004 [1919]: 33), which gives it an important advantage in any conflict on that scale. Exogenously, the international community has no incentives to support secession. At first glance, this incentives structure does not change in the EU s multilevel context so without any internal or external support, sub-state units cannot exit. This can change if all three levels become involved and a supranational party starts to support the sub-state unit. Then this actor is able to pose a credible exit threat and a within-eu secession becomes more probable. This article will first clarify why a game-theoretical approach is appropriate for explaining conflicts over secession. Later, it will analyse with cost calculation when the sub-state units have the incentive to secede and what changes if we include the EU as a third party. I will first design a secession game without any credible threat of the substate unit. Then I will build a second version of the game, where the support of the EU as a third party can render the threat credible. In a narrative, I will test if this model can explain interactions in the Spanish State of Autonomies. The contribution of this paper is to develop a model of a secession game, which includes the supranational party and offers a better understanding of how secession conflicts occur in the European multilevel framework. By identifying the existing 3

4 options and establishing the preferences of the actors, I help to clarify the absence of secession in the multinational states of the EU. An additional novelty lies in analyzing the subject of threats not only in the traditional centre-periphery modelling but also in the European setting. I use the game-theoretical model as a conceptual tool where I design the game in terms of a political centre and a sub-state unit bargaining. For simplification, I assume that both are independent unitary players. This design can be subject to some objections. First, the assumption of a unitary collective actor is risky. A central or a sub-state unit usually consists of different interests, expressed by different actors such as governments, political parties, social movements, and voters. In order to capture all coalitional activity within the structures of one actor, a two-level model (Putman 1988) or even a multilevel model, which includes all this activity would be necessary. Second, there are no fixed territorial divisions between both actors. Even if the sub-state unit is modelled as a fixed territory, it is territorially part of the central state at least until it exits. Nonetheless, in game-theoretical models, the preferences of the actors can be over-simplified and many possible streams within an actor reduced to one. Additionally, when using the terms central state or sub-state unit, I will refer to the central and the sub-state governments in the times of the conflict. This article does not aim to provide post hoc explanations; it aims to elucidate on present political interactions. The answers to the questions raised can enrich research on European integration, secession, territorial politics, and regionalism. Theoretical framework Some scholars, building on the seminal works of Riker (1964, 1975), have applied game theory to model the strategic interaction between the central state and the sub-state 4

5 unit(s), usually in federal states. Young (1994) has drawn secession games in Canada and De Figueiredo and Weingast (2005) have explained federalism as an equilibrium institution, which to be successful must be self-enforcing for political officials at all ranks of government. Bednar (2009) developed a robust federal design by showing how complementary institutions could regulate and preserve the allocation of authority between state and federal units. Filippov et al. (2004) showed the importance of an integrated political party system to the success of the federal design with games and Zuber (2011) developed a game-theoretical framework for two-level games in which she explains the complexity of the asymmetrical development of multinational states. More recently, Siroky, Mueller, and Hechter (2016) suggested a game-theoretical approach to centre-periphery bargaining. Using the rational-choice approach, these contributions argue that the logic of the market can explain political bargaining, where actors try to achieve their targets by exchanging rational demands backed by credible threats and promises. To matter these threats and promises, however, must involve some notion of commitment real or fake if it is to be anything (Schelling 1960: 127). We can distinguish between at least two ways threats can be used in collective bargaining. Schelling (1960, 1966) explained the threat of the power to punish or hurt when explaining mutual nuclear deterrence in the Cold War. Research on federalism and secession, building on Hirschman s work on loyalty, voice, and exit (1970), focuses on the potential of an exit threat (Bednar 2007; Siroky et al. 2016). In both cases, the exploitation of potential threats, not their application, has been identified as decisive. Even if there is a resilient caveat namely that the boundaries between the threat of punishment and threat of exit are fluent, this distinction is helpful when explaining political bargaining in a multinational state. While the sub-state unit can use the threat 5

6 of exit, the central state can threaten with the use of its monopoly of legitimate violence to punish the exit or even the intent to exit. The question of the disposal of these threats becomes even more puzzling in the European structure. Secession in the EU Due to a lack of European law and precedent, secession in the European multilevel framework needs to be explained. It is not clear which policy fields would or could be included within this process. The question of how sovereignty and authority would be reorganized is still unanswered (Closa 2016; Keating 2017). Most authors agree that when analyzing questions of sovereignty in the EU, we should distinguish between sovereignty and sovereign/public powers (Grimm 2015; Scharpf 2015). Even if we would assume that sovereignty is still indivisible and remains at the level of the central state 1, there is no question that the different public powers are shared and pooled between the supranational, national, and regional (sub-state) level of authority. This varying allocation of public powers has created a system of differentiated integration (Schimmelfennig et al. 2015). The EU is also highly involved in core state powers such as fiscal and monetary policy, taxation, and policing, which usually have been protected from external influence (Genschel and Jachtenfuchs 2014). The banking union is one of the most recent advances in supranational economic integration (Schimmelfennig 2016). Even if different policy fields are shared or pooled, we find one characteristic, which is held exclusively in the hands of the central state: the monopoly of legitimate violence. Aspects of this monopoly become important when analyzing secession in the EU. When the sub-state unit demands to leave the national framework, it could be interested in transferring (or leaving) some of its public powers to the supranational level. Nonetheless, it is eager to gain control over the monopoly of legitimate violence 6

7 (MON v ) from the central state. Because it is not shared or pooled at the supranational level, it is a decisive achievement coming from secession, which should be included into the cost calculation of secession. The cost of secession in the EU When rational-choice authors model political conflict, the concept of equilibrium institutions (Shepsle 1989; North 1990; Calvert 1995) is fundamental. Institutions stay in equilibrium as long as the incentive environment remains constant and the players do not have a motivation to change their behavior unilaterally (Morrow 1994). Citizens or organizations tend to benefit from using existing institutions and institutional improvements offer increasing returns. Changing those institutions comes at a particular cost and a player chooses to change its preferences if the cost of substituting the existing institutions is lower than the disadvantages of playing by the current rules. In a multinational state, the sub-state unit when governed by a nationalist/secessionist government can successfully threaten with exit only if it is backed by popular support. 2 It has to convince its citizens that they would be better off if they were to leave the common state. The credibility of exit threats depends largely upon perceptions about the costs of exit (Slapin 2009). When the cost of secession increases, the sub-state units threat to exit has fewer possibilities of success (de Figueiredo and Weingast 2005, Young 2004, Zuber 2011). My calculations of the costs of secession are based on the following assumptions. The central state usually receives all revenue (R) from its territory and exercises full authority, which includes the monopoly of legitimate violence (MON v ). If in a state there is a certain degree of decentralization or federalism, the sub-state unit can receive the revenue (R) after subtracting the share that belongs to the central state 7

8 (S). Some public powers can be shared, even if MON v is not. The sub-state units as rational players are continually interested in reducing the state share (S) by gaining additional revenue and public powers. The central state has the opposite interest, either trying to gain hold of a bigger state-share (S) or at very least by maintaining the same level. If the sub-state unit were to secede then that would lead to full revenue (R). It would also obtain the monopoly of legitimate violence (MON v ). However, secession does not come for free and has a cost (C) and for that reason the central state wants to keep the cost of secession (C) high. If necessary, it can also lower its state-share (S), which would make secession less attractive. To the cost of secession, we could also add the cost of polarization (p), which I define as the growing social and political tension between the central state and the substate unit. Polarization has a symbolic, not an economic, dimension. Even so, polarization can have economic consequences. Polarization can create negative emotions in a society, which can reduce the importance of the cost calculation because of the non-rational idealistic behavior of the actor (Young 2004). Even if polarization does not have a numeric value and we cannot include it into the cost calculation, we can use it as an additional theoretical variable, which can better explain some of the strategies in the secession game. Following my preliminary estimate, secession would be the superior strategy for the sub-state unit only if: (R + MON v ) - C > R - S - MON v 8

9 This cost estimate, however, is more complicated when both players are members of the EU, which can have a considerable influence on the cost calculation. In general, the EU is an institutional framework, which supplies incentives, opportunities, and constraints. One of the most relevant examples is economic and monetary integration. The economic framework guarantees the free movement of goods, capital, services, and people in the Single Market. The monetary framework offers stability with a common currency (the Euro), even if it takes some of the regulatory tools of monetary policy away from the central states. In most EU states, this supranational economic and monetary framework can replace the national framework, and the sub-state units automatically benefit from participating in it. When including the supranational level into the cost calculation, we have to extend the previous calculation by dividing the variable cost of secession (C) in two. The central state as an EU Member State can emphasize the cost of secession from the national framework (C n ) and the costs of leaving the supranational structures of the EU (C eu ). The sub-state unit expects the opposite. It hopes that by staying in the EU, it can reduce C eu to next to 0. With that, the sub-state unit would only have to assume the cost of leaving the national framework (C n ). In this case, exit would be a superior strategy for the sub-state unit only if: (R + MON v ) (C n + C eu ) > R S MON v This inequality is equivalent, after simple rearrangements, to (*) 2 MON v + S - (C n + C eu ) > 0. Since MON v 0, as a corollary of the above inequality, we conclude that if S > (C n + C eu ), then exit is always the superior strategy. However, if S < (C n + C eu ) then we cannot conclude anything without knowing the exact value of MON v. 9

10 If exit is the superior strategy, then the sub-state government can credibly convince its citizens that it is rational for them to exit and it can use that threat in its bargaining with the central state. Threats in centre-periphery bargaining The literature highlights two ways to make a threat credible: commitment and reputation. Schelling (2006:3), while developing the concept of commitment, has claimed that, [A] threat [ ] to be effective must be persuasive, and persuasiveness has two dimensions that may be called "potency" and "credibility". Potency refers to the (manifest) ability of the person making the promise or threat to carry it out on a scale sufficient to make compliance attractive. Credibility refers to the correspondent's belief that what is promised or threatened will indeed be carried out. Commitment refers to credibility. While commitment and credibility appear to be congruent, the concept of reputation follows a slightly different logic. In the tradition of the game-theoretical Chain Store Model (Selten 1978), reputation builds on the belief about the type of player in a game, that you are the type that always carries out your promise or threat (Morrow 1994: 281). By developing a larger strategic plan or set of interactions, a player can build up a reputation, typically called toughness. States which are tough are expected to face fewer threats and have more probability of being believed when they say that they will resist than those branded weak or irresolute. One of the most convincing ways to make a threat credible is to have threatened and carried through in the past. Taking into consideration these characteristics and assumptions, which potential do the threats of both actors in centre-periphery bargaining have? 10

11 Table 1: Players threat potential in the secession game Central state: The threat to punish (intents of) secession Sub-state unit: The threat of exit Player reputation: Tough Player reputation: Weak The potency of the threat: Rule of law, monopoly of violence The potency of the threat: Possible majority support for secession Showing commitment: By rule of law and reputation Showing commitment: By civic and political disobedience, e.g. general strike Threat credible?: Credible Threat credible?: Depends on potency and commitment The interests of the player: - Lower the level of conflict / polarization (p) - Threat better not to be carried out The interests of the player: - Make its own threat credible - Center should carry out its threat - Increase the level of conflict/polarization (p) Source: Author s elaboration On the one hand, the central state can threaten with punishing any intent to secede. As a Western democracy, the threat potential of the central state is significant and because of the control over rule of law and legitimate violence this player can repeatedly reject the secession demands of the sub-state unit. Besides that, the central state decides if the referendum takes place or not. The central state rapidly gains the reputation as a tough player in the secession conflict. This reputation is enough to make the threat of the central state credible. It also has the instruments to carry it out, even if it usually prefers not to. If the central state decides to punish, it is increasing the level of conflict/polarization (p), which could, as mentioned, reduce the cost-calculation of the sub-state citizens, also due to non-rational idealistic behavior. 11

12 On the other hand, the sub-state unit has the possibility to threaten with exit from the common state. In history, this exit was usually forced by military confrontation but today the democratic nature of the state and the EU poses limitations on the sub-state actor, which exclude violent conflict. The reputation of the sub-state unit is reputedly weak from the impossibility of violent conflict, and the lack of control over the monopoly of legitimate violence. To check if that threat can be made credible, we have to connect it to the concepts of potency and commitment. In the case of the sub-state unit, both are strongly linked to the interplay between the government and its population. The sub-state unit can show the potency of its threat mainly by having a majority of popular support for secession. However, because potency and commitment belong together, it would be helpful if this expressed support would be strengthened by some commitment, such as civic disobedience by calling a general strike, which would be expensive for both players and could change the cost calculation of secession. The main interest of the substate player is to make its threat credible but it is also interested in the central state carrying out its threat, which could increase the level of conflict/polarization (p) and reduce the importance of the cost calculation of secession. The Games When using a game theoretical approach, we usually define the game as a triplet which consists of a set of players, a set of strategies, and a set of preferences with payoffs for each player. All those are influenced by the game s characteristics. Accordingly, before playing the secession game, we have to establish the preferences of the actors with their payoffs. These payoffs will be ranked where figure 4 symbolizes the best outcome from the perspective of a player and 1 the worst possible outcome. 12

13 I will draw two different games, which start when the sub-state unit stops being loyal and starts demanding exit. Both games are based on the assumption that the central state would not allow a referendum on secession. 3 In the first game, the sub-state unit wants to secede from the central state but does not have a credible threat to make. This secession game can be found at least in two scenarios. In the first, the institutional framework is an advanced Western democracy. In the second, this advanced Western democracy is in a multilevel framework but the supranational party does not participate in the game, accepting the secession game as an internal matter of the Member State. Considering actor preferences, the sub-state unit would enlarge its strategy space by introducing the exit threat of a referendum (REF) into the preference set. Institutional Reform (IR) would be its second best option. The central state s preference would be to maintain the Status Quo (SQ) or even enlarge its revenue through stronger Centralization (C). Preference set in the first game REF: Referendum IR: Institutional Reform SQ: Status Quo C: Centralization As shown in the cost calculation, the central state is interested in increasing or at least maintaining its revenue R and its public powers. The central state s first preference would be the Status Quo (SQ), followed by Centralization (C), Institutional Reform (IR), and Referendum (REF). There is the valid caveat that Centralization (C) outcome could also be the central state s preference. However, considering the institutional and 13

14 political obstacles, I claim that during a period of secession conflict the central state would soften its demands and opt for Status Quo (SQ) as its first preference. The preference set of the sub-state unit would be Referendum (REF), followed by Institutional Reform (IR), Status Quo (SQ), and Centralization (C). Central state: SQ>C>IR>REF Sub-state unit: REF>IR>SQ>C In game theory strategies are the sources of action (or sequences of moves) available to the players. Each player has two strategies. The central state as the row player chooses between the upper and lower row, the sub-state unit as the column player between the left and right column. Under the assumption of these simple ordinal preference rankings, when playing this game, the dominant strategy for the central state would be to defect, choosing the lower row with C/SQ. The dominant strategy for the sub-state unit would be to defect, selecting the column REF/SQ. The game equilibrium is the Status-Quo (SQ). Figure 1: The game without a credible threat Sub-state unit Cooperate Defect Central state Cooperate IR (2/3) REF (1/4) Source: Author s elaboration Defect C (3/1) SQ (4/2) 14

15 How can we explain this equilibrium? I claim that the central state can build a tough reputation due to limitations imposed on the sub-state unit, which cannot use the threat of military confrontation. The sub-state unit is aware of its weakness and the lack of a credible threat. Even if the majority of its population supports it a population which would commit to carrying out the threat in the form of general strike, the central state could for the most part ignore those events if there is no external support for the substate unit. The sub-state unit chooses to defect also because if it cooperated, it would end up with the Centralization (C) outcome. The influence of the EU on the credibility of the sub-unit s threat The result of this game in a multilevel framework could be different if we include an active third party. Brams (2003:30) has foreseen this option in game-theoretical constellations when claiming that, Although the issue of coalition formation is irrelevant in two-person bargaining games, under some of the procedures to be analyzed a third party [ ] can influence the agreement between the bargainers. This third party, however, is never a player in the sense of being an actor who has preferences and makes strategic choices, though its actions may be informed (e.g. by certain information) or constrained (e.g. by limited resources or self-imposed restrictions) in certain ways. Even if Brams (2003) here is using the idea of an appraiser in an economic buyer-seller setting, I will borrow the concept and apply it to the EU setting, where the European level of governance could be considered as a third party. Let us assume that the EU becomes involved due to one of the following scenarios. The sub-state units can force the central state s hand and make it carry out its threat of punishment. The central state might withdraw some of the public powers it gave to the sub-state unit during the decentralization process or even go so far as to apply military law. As a response, the 15

16 sub-state unit could lodge a complaint within the multi-level framework. It could argue that retaliation for a general strike or that the suspension of elements of sub-state selfgovernment contradict the democratic values of the EU. In the dominant remedial right only theory of secession, secession can be justified when facing persistent and serious injustices (Buchanan 1997). The EU could also see itself involved if there were permanent majority support for secession in the sub-state unit. We could expect that in a democratic multilevel framework, the EU as a third party could not remain indifferent to a situation where a sweeping majority of citizens asks for secession. We can assume that in both cases, the EU would signal support for the sub-state unit. Second game Let us suppose that because of the involvement of the EU as a third party, the central state begins to consider the exit threat of the sub-state unit as credible. In this case, the central state would find itself in a much weaker bargaining position. The credible threat of exit would change the set of preferences for the sub-state unit, creating a new possible outcome: within-eu secession (S-EU). 4 Additionally, the sub-state unit would maintain its second option: referendum (REF). The central state, which is highly integrated within the EU and could be asked by the EU to accept within-eu secession, would choose a more conservative set of preferences with Institutional Reform (IR) and Status Quo (SQ). 16

17 Preference set in the second game S-EU: Within-EU secession REF: Referendum IR: Institutional Reform SQ: Status Quo Following that line of argument, the central state's first preference would put Institutional Reform (IR) option in the first slot, followed by Status Quo (SQ), Referendum (REF), and within-eu secession (S-EU). The sub-state unit would choose within-eu secession (S-EU), followed by Referendum (REF), Institutional Reform (IR), and the Status Quo (SQ). Central state: IR > SQ > REF > S-EU Sub-state unit: S-EU> REF> IR > SQ Figure 2: The game with a credible exit threat Sub-state unit Cooperate Defect Cooperate IR (4/2) REF (2/3) Central state Source: Author s elaboration Defect SQ (3/1) S-EU (1/4) 17

18 Under the assumption of these simple ordinal preference rankings, the dominant strategy for the sub-state unit would be to defect, choosing the column REF/S-EU The dominant strategy for the central state would be to cooperate, choosing the row IR/REF. The equilibrium of this game would be the Referendum (REF) outcome. The narrative: The game between Spain and Catalonia The following narrative will connect this game-theoretical model to the conflict over authority in the Spanish State of Autonomies. I will analyse if this plausibility probe can rightly predict most of the strategic interaction between the Spanish central state and the Catalan sub-state unit in the 2010s. I have chosen the Spanish State of Autonomies because it has all the characteristics necessary for applying this model. 5 It is a Member State of the European Union with two strong nationalist sub-state units: Catalonia and the Basque Country. Nationalist/secessionist governments of both sub-state units have requested exit and have already looked for the approval of the European Union. While the secession requests from the Basque sub-state unit are stagnant after a period of intense conflict which culminated in the Ibarretxe-Plan ( ), the Catalan government has been pushing since 2012 for secession. Catalonia did not always want to become an independent state. After a short period of autonomy during the Second Spanish Republic ( ), ended by the Civil War, Franco s regime suppressed most of the rights of the Catalans, such as their language. It also had withdrawn autonomy. After establishing democracy in the second half of the 1970s, Catalonia wanted to get most of that autonomy back. The Catalans received some special rights and a faster track to autonomy within the Spanish State of Autonomies. However, some asymmetrical differences were established in the dual 18

19 taxation system, based on exceptions made for the historical fueros. Fiscal instruments were only given to the Basque Country and Navarre (special regime) and not to Catalonia (common regime). Other Autonomous Communities (AC) challenged this faster track to autonomy and the other asymmetrical competencies from the beginning. They, in turn, also saw their autonomy upgraded, often described as café para todos (coffee for everyone). Every time asymmetry was created in an AC with a nationalist minority, a resymmetrisation through a competence upgrade for the other regional AC followed, which led to renewed requests from the AC with a nationalist minority for more asymmetry (Linz and Montero 1999; Balfour and Quiroga 2007; Maíz and Beramendi 2004; León-Alfonso 2007; Requejo and Nagel 2011). Before the 2004 elections, the Partido Socialista Obrero Español (Spanish Socialist Workers Party or PSOE) proposed to federalize Spain and to recognize the plural character of the Spanish state, giving Catalonia special status. When the PSOE candidate José Luis Rodríguez Zapatero was elected, Catalonia started to elaborate suggestions for reform, which would include a fiscal solution similar to the special regime governing the Basque Country and Navarre. 6 This would have given Catalonia the opportunity to raise taxes and transfer money for the services offered by the central state. Catalonia also asked for a symbolic upgrade in the form of recognition of the Catalan nation (Colino 2009). The reformed Catalan Statute of Autonomy was passed in 2006 by referendum, with 48.8% participation in Catalonia, 73.2% of which was in favour. From 2006 to 2010, however, the Catalan Statute was challenged by nearly all political actors, even if for different reasons. There was the outright opposition of the Partido Popular (Popular Party or PP). The PSOE promised to accept any statute approved by the Catalan Corts, 19

20 however, later around 80% of the text was amended by the parliament in Madrid. The Catalan party Esquerra Republicana de Catalonia, (Republican Left of Catalonia or ERC), while initially supporting the process in the end did not accept Madrid s amendments. The Constitutional Court interpreted some of the articles of the new Catalan Statute as unconstitutional and others were declared null and void 7 (Rico and Liñeira 2014). In the aftermath of that ruling and the shadow of an economic crisis, the Catalan government formed by Convergència i Unió (Convergence and Union or CiU) started to change its preference set by including requests for self-determination in its political program. It was backed by a robust social movement, which organized massive demonstrations for the right to decide during the national Catalan day La Diada from 2012 to The governing Partido Popular as well as the main opposition parties such as PSOE or Ciudadanos (Citizens) have rejected the demand for an agreed referendum between the central state and the sub-state unit. Only the leftist party Podemos (We Can) is in favour. The Catalans, faced with this refusal, held a non-binding consultation on secession in November 2014, with 37-41% participation. Slightly more than 80% of the voters were in favour of Catalonia becoming an independent state. Even if most of the Catalans who oppose secession did not vote in this consultation, the result confirmed the surveys, which showed the support for independence in Catalonia at about 2,300,000 of 6,300,000 citizen (El País 2014). Nonetheless, the fact that the Catalan people voted in a unilateral and nonbinding consultation did not put Catalonia in a better bargaining position. After the event, the Catalan government claimed to have a democratic mandate which should be accepted by the central state and supported by the EU. They did not, however, start any 20

21 action which might be taken by the central state as commitment and the central state stopped short of polarizing the situation by bringing charges against individual secessionist politicians only. The EU and the international community took notice of the secession referendum but no further actions were taken. The Catalan journey for independence did not stop after this referendum. The Catalan government then declared the regional elections in September 2015 as a plebiscite on self-determination. In those regional elections, two separatist platforms Junts pel Sí (Joined for a Yes, including Convergència i Unió and the leftist-separatist Esquerra Republicana de Catalonia), as well as Candidatura d'unitat Popular (Popular Unity Candidature or CUP), won a majority for independence in seats (72 of 135) but not in votes (47.7%). Nonetheless, both sides have interpreted these results as a victory (Guirao 2015). The Catalan government consequentially began a process to separate from Spain, which the Spanish Constitutional Court declared illegal and the EU widely ignored. In the following years up to 2017, neither actor changed its position. The central state focused on strengthening its reputation as a tough player and the sub-state unit tried to involve the supranational party and create a credible threat. In the following, I will analyse why so far it has failed. When applying Schillings s (2006) distinction between potency and credibility of a threat, we see that considering its potency, the measured support of Catalans for independence is strong but barely a majority. It oscillated in between 35% and 55%, depending not only on the type of research institute conducting the survey, be it Centro de Investigaciones Sociológicas (Sociological Research Centre or CIS) in Madrid or Centro de Estudis d Opinio (Opinion Studies Centre or CEO) in Barcelona, but also on the question asked and the answers/choices offered. Also a possible 21

22 exclusion from the EU influences the vote intention (Muro and Vlaskamp 2016). The Catalan government has problems convincing not only the central state but also the EU as a third party that it is following the preferences of the majority of its citizens. The credibility or the commitment of this threat, which in a democratic state excludes military confrontation, could include an act of civil disobedience in the form of a general strike. Even if we have no data on a general strike, a GESOP (2015) survey showed that only 35% of the Catalan population is in favour, that the Catalan parliament encourages some acts of disobedience, and that the majority of Catalans (52%) is against such actions. The missing potency of the threat has surely been one of the reasons why the Catalan sub-state unit did not get the EU s support. The Spanish central state was always eager not to raise the level of conflict/polarization and leave the rational cost of secession high. The central state lobbied successfully against secession among international partners, which have no incentives to support secession for two reasons: it could boomerang into their territory when possible national minorities claim exit and supporting secessionists could worsen relations with the partner state. What does this game-theoretical model predict? If support for independence does not rise and the EU does not start to support Catalonia's secession goals, the outcome of this secession game will continue to be the status quo outcome. It would change only if the central state rearranges the order of its preferences. That could occur if: a statewide party that is in favour of a Catalan referendum on self-determination were to win elections; there was further asymmetric decentralization; or the Spanish population changed its view on holding a statewide referendum on the status of Catalonia. With strong support from the entire Spanish population, a Catalan referendum and within-eu secession would be more likely. 22

23 Conclusions In this paper, I have modelled the dynamics of a secession game in an advanced Western democracy embedded in the EU. It includes not only the interaction between the central state and the sub-state unit but also analyses the influence of the EU as an alternative institutional framework as well as an additional third party. By drawing two games, I have explained the condition for the successful exit of a sub-state unit, which lies primarily in the EU s support as a third party. The first game shows that in the standard democratic framework the central state can build a reputation as a tough player. When the supranational party is not included or neutral and the central state chooses to ignore the secession demands, the sub-state unit is hard pressed to express a credible exit threat. However, I show in the second game that this dynamic can change when the EU supports the sub-state unit. In this case, the Catalans could decide a possible secession through a referendum. This article contributes new conceptual ideas and empirical observations. By applying questions of centre-periphery bargaining to a multi-level framework, it widens bargaining models and redefines some standard applications. It is pioneering in including the variable of the monopoly of legitimate violence to a cost calculation of secession. Analyzing the puzzle of threats in bargaining situations, it shows that due to the impossibility of military confrontation, questions of a credible threat during political conflict in advanced Western democracies and in democratic multilevel frameworks have to be reformulated. This paper comes to different conclusions regarding the secession conflict between the Spanish State of Autonomies and Catalonia than previous research by Griffith et al. (2015), which start their game after a referendum won by Catalonia. Unlike these authors, I claim that analyzing how a sub-state unit could gain a referendum on self-determination is more relevant and should be explained first. 23

24 This article argues that we will not see a Catalan referendum or secession as long as factors like the level of popular support for independence, the unwillingness of the majority to engage in civil disobedience, and the position of the central state and the EU stay the same. Like all research, the model has significant limitations. Unfortunately, the theoretical arguments, though decisive, are uncertain because of the ambiguity surrounding the EU s position. Due to lack of European law and precedent, we cannot be sure how the EU would influence the cost calculation as an alternative institutional framework. There is also uncertainty about its position regarding within-eu secession. This limitation should be seen as an opportunity for future research. It could analyse the secession game, emphasizing not only to what degree the EU influences the cost calculations, but emphasizing the EU as an active actor, analyzing its reputation, toughness, preferences, and threat potential. Note 1 The question of who is the holder of sovereignty in the EU is an object of dispute. Scholars are 2 The exit option could be more than a threat; it can also be a real preference of a sub-state unit, which is not interested in maintaining union with the central state anymore like after a moment of high polarization between both parties. It is difficult to distinguish between exit as a threat and exit as a real preference because some sub-state actors can prefer the former while others the latter. Qvortrup (2014) claims that usually referendums on ethnic and national issues have been held for strategic and not for idealistic reasons. 3 These conditions could be found until now in Spain in the Basque Country ( ) and Catalonia (from 2012 on). 4 The EU Treaties are silent on secession within their Member States. Article 4.2 underlines the respect for state functions and constitutional identity. This offers space for different and even contradictory interpretations by lawyers when analyzing the Scottish and Catalan cases. Some argue that the Treaties do not prohibit separatist processes, only because EU Member States do not consider it necessary to do so, since an implicit prohibition can be 24

25 deduced from the values, principles, and objectives of the Union, as contained in Articles 2, 3 and 4 of the TEU (Closa 2016). Others argue that the EU would take a more pragmatic approach in case of secession (Avery, 2014, Tierney & Boyle, 2014). 5 In 2017 another possible candidate would be Scotland, even if the United Kingdom is based on the logic of a union. 6 It was similar but not identical because the Catalans suspected that the Constitution does not allow transfer of the recognition of historical rights of the disposición additional primer to territories other than Navarre and the three Basque provinces. 7 Article 5 of the Catalan Statute about historical rights was said not to be equivalent to the disposición adicional primera of the Constitution. With that, it did not give the Catalan government the right to claim the same institutions as the Basque Country. Another important feature was also the clear refusal of Article 8, which dealt with symbolic de jure national asymmetry. The Constitutional Court s interpretation cemented the institutional status quo, the Catalan statute cannot speak of a Catalan nation and only of nationality, which should not question the indissoluble unity of the Spanish Nation. 25

26 26

27 References Alesina, Alberto, and Enrico Spolaore (2003). The Size of Nations. Cambridge, MA: The MIT Press. Avery, Graham (2014). Independentism and the European Union, Policy Brief, Brussels: European Policy Centre. Balfour, Sebastian and Alejandro Quiroga (2007). The Reinvention of Spain: Nation and Identity since Democracy. Oxford: Oxford University Press. Bednar, Jenna (2007). Valuing Exit Options, Publius, 37:2, Bednar, Jenna (2009). The Robust Federation. Cambridge: Cambridge University Press. Buchanan, Allen (1997). Theories of Secession, Philosophy & Public Affairs, 26:1, Brams, Steven J. (2003). Negotiation Games: Applying Game Theory to Bargaining and Arbitration (Vol. 2), New York: Routledge. Calvert, Randall (1995). Rational Actors, Equilibrium, and Social Institutions, in Jack Knight and Itai Sened (eds.), Explaining Social Institutions. Ann Arbor: University of Michigan Press, CIS , Centro de Investigaciones Sociológicas, Madrid. Closa, Carlos (2016). Secession from a Member State and EU Membership: the View from the Union, European Constitutional Law Review, 12:2, Colino, Cesar (2009). Constitutional Change Without Constitutional Reform: Spanish Federalism and the Revision of Catalonia's Statute of Autonomy. Publius, 39:2, Colomer, Josep (2007). Great Empires, Small Nations. London: Routledge. De Figueiredo, Rui J., and Barry R. Weingast (2005). Self-Enforcing Federalism, Journal of Law, Economics, and Organization, 21:1, Dion, Stéphane (1996). Why is Secession Difficult in Well-Established Democracies? Lessons from Quebec, British Journal of Political Science, 26:2, El País (2014). 1,8 millones de personas votan por la independencia catalana en el 9- N, El País, available at tml (accessed on 11 January 2017). 27

28 Filippov, Mikhail, Peter C. Ordeshook and Olga Shvetsova (2004). Designing Federalism: A Theory of Self-Sustainable Federal Institutions. Cambridge: Cambridge University Press. Genschel, Philipp. and Markus Jachtenfuchs (2014). Conclusion: the European Integration of Core State Powers. Patterns and causes, in Philipp Genschel and Markus Jachtenfuchs (eds.). Beyond Market Integration. Analysing the European Integration of Core State Powers, Oxford: Oxford University Press, GESOP (2015). Gabinet d'estudis Socials i Opinió Pública, 9. November Griffiths, Ryan D., Pablo Guillén Alvarez, and Ferran Martinez i Coma (2015). Between the Sword and the Wall: Spain's Limited Options for Catalan Secessionism, Nations and Nationalism, 21:1, Grimm, Dieter (2015). Sovereignty in the European Union, in Jeffrey Ellsworth and Johan van der Walt (eds.), Constitutional Sovereignty and Social Solidarity in Europe, Baden-Baden: Nomos, Guirao, Fernando (2015). An independent Catalonia as a member state of the European Union? :Terra ignota, In Xavier Cuadras - Morato (ed.), Catalonia: a new independent state in Europe? A debate on secession within the European Union, London: Routledge, Hechter, Michael (1992). The Dynamics of Secession, Acta Sociologica, 35:4, Hirschman, Albert O. (1970). Exit, Voice, and Loyalty. Cambridge, MA: Harvard University Press. Keating, Michael (2001). Plurinational Democracy. Stateless Nations in a Post- Sovereignty Era. Oxford: Oxford University Press. Keating, Michael (2017). The European Question, in Michael Keating (ed.), Debating Scotland: Issues of Independence and Union in the 2014 Referendum, Oxford: Oxford University Press, León-Alfonso, Sandra (2007). The Political Economy of Fiscal Decentralization: Bringing Politics to the Study of Intergovernmental Transfers. Barcelona: Institut d'estudis Autonòmics. Linz, Juan J., and José R. Montero (1999). The Party Systems of Spain: Old Cleavages and New Challenges, Working Paper, Madrid: Centro de Estudios Avanzados en Ciencias Sociales, Juan March Foundation. MacCormick, Neil (1999). Questioning Sovereignty. Oxford: Oxford University Press. 28

29 Maíz, Ramón, and Pablo Beramendi (2004). Spain: Unfulfilled Federalism, in Ugo M. Amoretti and Nancy Bermeo (eds.). Federalism and territorial cleavages, Baltimore: John Hopkins University Press, Muro, Diego, and Martin C. Vlaskamp (2016), How do Prospects of EU membership Influence Support for Secession? A Survey Experiment in Catalonia and Scotland, West European Politics, 39:6, Morrow, James D. (1994). Game Theory for Political Scientists. Princeton, NJ: Princeton University Press. North, Douglass C. (1990). Institutions, Institutional Change and Economic Performance. Cambridge: Cambridge University Press. Qvortrup, Matt (2014). Referendums and Ethnic Conflict. Philadelphia: University of Pennsylvania Press. Putnam, Robert D. (1988). Diplomacy and Domestic Politics: The Logic of Two-Level Games, International organization, 42:3, Requejo, Ferran, and Klaus-Jürgen Nagel, (2011). Federalism beyond Federations. Burlington: Ashgate. Rico, Guillem, and Robert Liñeira (2014). Bringing Secessionism into the Mainstream: The 2012 Regional Election in Catalonia. South European Society and Politics, 19:2, Riker, William H. (1964). Federalism: Origin, Operation, Significance. Boston: Little, Brown. Riker, William H. (1975). Federalism, in Fred I. Greenstein and Nelson W. Polsby (eds.), Handbook of Political Science. Vol 5, Reading, MA: Addison-Wesley, Sambanis, Nicholas, and Branko Milanović (2014). Explaining Regional Autonomy Differences in Decentralized Countries, Comparative Political Studies, 47:13, Scharpf, Fritz W. (2015). Sovereignty and Democracy in the European Polity: Reflections On Dieter Grimm s Essay Sovereignty In The European Union, in Jeffrey Ellsworth and Johan van der Walt (eds.), Constitutional Sovereignty and Social Solidarity in Europe, Baden-Baden: Nomos, Schimmelfennig, Frank, Dirk Leuffen, and Berthold Rittberger (2015). The European Union as a System of Differentiated Integration: Interdependence, Politicization and Differentiation, Journal of European Public Policy, 22:6,

30 Schimmelfennig, Frank (2016). A Differentiated Leap Forward: Spillover, Pathdependency, and Graded Membership in European Banking Regulation, West European Politics, 39:3, Schelling, Thomas C. (1960). The Strategy of Conflict. Cambridge, MA: Harvard University Press. Schelling, Thomas C. (1966). Arms and Influence. New Haven, CT: Yale University Press. Schelling, Thomas C. (2006). Strategies of Commitment and Other Essays. Harvard: Harvard University Press. Selten, Reinhard (1978). The Chain Store Paradox, Theory and decision, 9:2, Shepsle, Kenneth A. (1989). Studying Institutions. Some Lessons from the Rational Choice Approach, Journal of Theoretical Politics, 1:2, Sorens, Jason (2005). The Cross-Sectional Determinants of Secessionism in Advanced Democracies, Comparative Political Studies, 38:3, Siroky, David S., and John Cuffe (2015). Lost Autonomy, Nationalism and Separatism, Comparative Political Studies, 48:1, Siroky, David S., Sean Mueller, and Michael Hechter (2016). Center-Periphery Bargaining in the Age of Democracy, Swiss Political Science Review, 22:4, Slapin, Jonathan B. (2009). Exit, Voice and Cooperation: Bargaining Power in International Organisations and Federal Systems. Journal of Theoretical Politics 21:2, Tierney, Stephen, and Katie Boyle (2014). An Independent Scotland: The Road to Membership of the European Union, Briefing Paper, Edinburgh: Centre on Constitutional Change. Weber, Max (2004 [1919]). Economy and Society. Vol. I and II, Berkeley, CA: University of California Press. Wood, John R. (1981). Secession: A Comparative Analytical Framework, Canadian Journal of Political Science, 14:1, Zuber, Christina. I. (2011). Understanding the Multinational Game: Toward a Theory of Asymmetrical Federalism, Comparative Political Studies, 44:5, Young, Robert A. (1994). The Political Economy of Secession: The Case of Quebec, Constitutional Political Economy, 5:2,