<<

Does Fiscal Deter or Spur Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 Secessionist Movements? Empirical Evidence from Europe

Martin Rode*, Hans Pitlik† and Miguel Angel Borrella Mas‡

*Universidad de Navarra; [email protected] †Austrian Institute of Economic Research; [email protected] ‡Universidad de Navarra; [email protected]

Popular and scientific contributions often call for increased regional decision-making power to manage secessionist conflict, assuming that fiscally federalized countries are less prone to disintegrate politically. From a theoretical standpoint this is not clear though, as federalism creates an institutional paradox where autonomous legislative and bureaucratic structures can potentially be used to make secessionism a viable strategy in the first place. In particular, the role of asymmetric territorial arrangements in this association is crucially underexplored at present. Using electoral data on separatist political movements from a large variety of European regions since the mid-1990s, our findings indicate that regions with comparatively higher fiscal and institutional autonomy are more prone to vote for secessionist parties. Accounting for possible endogeneity, asymmetric territorial self-governance seems to cause much stronger incentives to vote for secessionist platforms, while the association with fiscally more symmetric arrangements is surprisingly small.

The referendum on Scottish independence in September 2014 and the ongoing political conflict in over a plebiscite on Catalan independence are just the latest expressions of a European trend, where secessionist platforms are increasingly gaining ground in the political discourse (Griffiths 2016). This evolution constitutes a notable development, as it is all about creating smaller political entities with that are perceived to be closer to local populations, rather than creating bigger and politically more powerful nation states, as was the trend a century ago. Alesina and Spolaore (2003) and Desmet et al. (2011) argue that all nations essentially face a tradeoff between increasing returns in the provision of public goods and the costs of greater cultural heterogeneity. Economically, populous countries are better able to provide key public goods, such as security and monetary policy, at diminishing marginal costs to the individual citizen, giving

Publius:TheJournal of Federalism volume 48 number 2, pp.161^190 doi:10.1093/publius/pjx060 AdvanceAccess publication October 23, 2017 ß TheAuthor 2017. Published by Oxford University Press on behalf of CSF Associates: Publius, Inc. All rights reserved. For permissions, please email: [email protected] 162 M. Rode, H. Pitlik and M. A. Borrella Mas them a decisive edge over smaller states. Large political units are also better equipped to reap gains from internal trade. According to Lecours (2011) and Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 Keating (1999, 2012), the deepening of economic and political integration in the European Union (EU), with a common market and (for most member states) a joint currency, and as a provider of intra-European security may have facilitated regional secessionism by reducing the benefits of certain regions to pertain to a large nation state. To calm secessionist tendencies, of decision-making powers to the sub-state level is often called for: If regional governments have more fiscal autonomy, it is argued, this will ultimately reduce secessionism (Horowitz 1985; Bermeo 2002; Gurr 2000; Erk and Anderson 2009).1 Notwithstanding, enhanced federalism can create an important institutional paradox: The same legislative and bureaucratic structures that are supposed to manage separatist conflict can potentially be used by regional politicians to make secessionism a viable strategy in the first place (Horowitz 1985; Anderson 2004). Territorial self-governance could potentially serve as a springboard to eventual . Weingast (2013) claims we generally know too little about the impact of on secessionist conflict, where especially the role of is still underexplored (but see: Brancati 2006; Sorens 2004, 2005, 2012). Perhaps most importantly, the issue of symmetry in fiscal arrangements has not been addressed empirically in this context. It may make a significant difference though, whether all regions of a country have similar powers on taxation and service provision, or whether powers are distributed asymmetrically among jurisdictions (Congleton 2006; Zuber 2011). Obviously, these are vital questions for the constitutional design of federal systems. Using election data on separatist political movements from a variety of European countries since the mid-1990s, we empirically analyze effects of symmetric and asymmetric fiscal federal arrangements on secessionist movements. For reasons of comparability, we limit our sample to member countries of the EU and two closely associated countries (Norway and Switzerland), so that the external conditions faced by the individual states in terms of trade and security are broadly similar. As secessionism is not confined to one type of state organization, our country sample includes formally federal and non-federal states for which sufficient data is available on a regional level. Findings suggest that more fiscally autonomous regions are indeed more likely to vote for secessionist parties, but that results are highly dependent on the type of arrangement present: Asymmetric territorial self- governance appears to cause higher electoral support for secessionist platforms, probably because of the incentives that it creates for center-periphery bargaining, while the association of secessionism with symmetric fiscal federalist arrangements is close to negligible. Does Fiscal Federalism Deter or Spur Secessionism 163

The remainder of the article is organized as follows: The next section conducts a tentative review of the relevant literature and formulates the central hypotheses. Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 This is followed by a description of the data and our research strategy, succeeded by a section that presents preliminary findings from an Ordinary Least Squares analysis. The final section conducts a series of Difference-in-Difference estimations to firmly establish a causal relation in our findings, followed by the conclusions.

Literature Review and Hypothesis Development The literature mentions a number of possible sources and motives for secessionist tendencies within a country. Several authors note that interest and identity both play a major role in secessionist mobilization, and one is probably not sufficient without the other, but all find economic factors to play a crucial role in fueling secessionist demands (Collier and Hoeffler 2006; Sorens 2012; Grossman and Lewis 2014; Sambanis and Milanovic 2014). Since many wealthy EU regions with historically separate identities are nowadays subject to national and EU wide territorial redistribution, secessionism might be perceived as legitimate to defend tax revenues and limit membership to just one transfer union, instead of two (Scharnagl 2012). Regarding the possible fiscal causes that underlie secessionism, Buchanan and Faith (1987) argue that it can be a reaction to unjust taxation, and comparatively wealthy regions might choose this path to achieve a more equitable tax treatment through a that is closer to the local electorate.2 Secessionist support can be further spurred by the presence of large natural resource endowments, as highlighted by Collier and Hoeffler (2006), Sorens (2012), or Gehring and Schneider (2016), exacerbating conflicts over resources revenues between endowed sub-state territories and central governments. To calm secessionist tendencies that originate from conflicts over identity and interest, increased decision-making power at the sub-central level might provide a viable solution: If regional governments enjoy more fiscal autonomy and the means to provide public goods through its bureaucracy, it is argued, this will ultimately reduce secessionist motivations. Federalism, then, might be a formidable tool to accommodate territorially based conflicts, while maintaining the integrity of existing nation states (Erk and Anderson 2009). Horowitz (1985) and Bermeo (2002) both argue that the division of authority between states and regions has the potential to reduce inter- group conflict, while Bermeo (2002) further highlights that no violent separatist movement has ever succeeded in a federal democracy. Notwithstanding, the scholars cited above also highlight that this rationale is not as clear as it might seem at first sight: Theoretically, enhanced federalism, or territorial self-governance, can create an important institutional paradox: The same legislative and bureaucratic structures that are supposed to manage separatist conflict can potentially be used to make secession a viable strategy in the first place. 164 M. Rode, H. Pitlik and M. A. Borrella Mas

Other factors fueling secessionism in such a context are the presence of powerful group grievances (Horowitz 1985; Anderson 2004), the generalization Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 of majority ethnic symbols and political values (Coakley 1990), the granting of comparatively little territorial autonomy in the aftermath of conflict (Cederman et al. 2015), a lack of symbolic recognition for the minority (Basta 2018), institutionally weak states (Bakke 2015), or minority groups with unusual political leverage (Jenne et al. 2007). That is, the “freezing” of an inter-group conflict over territorial secession through enhanced autonomy potentially reduces the formal cost of actual secession radically, once institutions and quasi-state structures have been successfully set up (Lecours 2011). In addition, regional policy platforms give secessionist politicians an arena for identity-based electoral strategies (Collier and Hoeffler 2006), and incentives to secure the allocation of public resources via secession (Grossman and Lewis 2014). So in this view, enhanced federalism serves as a type of springboard, or slippery slope, to eventual secession, and not as a means to mitigate the underlying conflict (Horowitz 1985; Bunce 2003).3 Much of the literature on federalism and secessionism explicitly treats the former as an entirely political concept, even though many of the functional arguments on the provision of public goods actually apply much better to the concept of fiscal federalism. Following Rodden (2004) and Sorens (2016),we define fiscal federalism in competitive terms, where sub-state level jurisdictions compete for a mobile tax base in a common market by exercising autonomous regulatory and taxation powers. On the one hand, a more traditional view states that competitive federalism fosters income disparities between sub-national jurisdictions if it is not supplemented by an inter-regional transfer scheme within the (e.g., Prud’homme 1995), thus fueling secessionist sentiments. On the other hand, competitive federalism can generate stronger incentives for growth- promoting economic regional policies and reduce intra-regional differences in per capita-income (e.g., Oates 1999; Weingast 2013). In that case competitive fiscal federalism could, on average, reduce secessionist support in the long-run. Also, for most of the world’s contemporary secessionist conflicts, some form of competitive fiscal federalism remains the only viable institutional option available, despite potential drawbacks (Anderson 2016). There is not a whole lot of quantitative empirical research that has dealt with the question whether fiscal decentralization helps mitigate secessionist conflict. Some studies explore the effect of decentralization on ethnic conflict and secessionism, but either employ case studies to demonstrate their arguments (Lecours 2011; Keating 2012; Bakke 2015), or focus on determinants of ethnic conflict in developing and democratizing countries (Collier and Hoeffler 2006; Alonso and Ruiz Rufino 2007; Bakke 2015; Cederman et al. 2015). Exceptions are Brancati (2006) and Sorens (2004, 2005, 2012), who investigate the role of fiscal federalism for secessionist voting in developed democracies: Sorens establishes the Does Fiscal Federalism Deter or Spur Secessionism 165 determinants of regional vote shares for secessionist parties in high-income democracies with time-series (Sorens 2004), and cross-sectional data (Sorens 2005). Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 These papers do not find that fiscal autonomy impacts the share of secessionist votes, but rather find countries with significant secessionist parties to be more likely to decentralize. This highlights the tremendous difficulty of empirically establishing a causal relationship. More recently, Sorens (2012) investigates the role of fiscal federalism in influencing secessionism in the long run, finding that a substantial devolution of competences does not promote secessionist parties’ electoral success on balance. For a cross-section of thirty democracies, Brancati (2006) reports that political and fiscal decentralization actually reduce secessionist conflict, but also induce the growth of regional parties. Regarding the impact of fiscal federalism on secessionist electoral success in Europe, we formulate the following hypothesis: H1: In European regions with a higher degree of fiscal autonomy, secessionist parties are less likely to gain electoral support, on average. Many authors further divide the universe of democratic federal states into symmetrical and asymmetrical systems. The former is defined as a system where all regional units have the same constitutional rights and obligations, while embedded special rights exist for some distinct members of a federation in the latter (Stepan 2001; Benedikter 2007). Asymmetric arrangements that give some regions a higher degree of fiscal policy leverage can even exist in formally non-federal states. McGarry and O’Leary (2009) argue that pluri-national are often much more stable in practice than suggested by many political theories of federalism. According to McGarry (2007) and Benedikter (2007), asymmetric autonomy is especially well-suited to accommodate ethnic minorities. Wolff (2013) proposes territorial self-governance, referring to the transfer of certain powers from a central government to a self-governing regional entity, as a form of successful secessionist conflict management. As such struggles are about group autonomy Gurr (2000) also finds that negotiated territorial self-governance is an effective antidote for conflicts of ethno-national secession.4 Lustick et al. (2004) report experimental evidence from a simulation study, finding that secessionist challenges can effectively be mitigated by semi-autonomous institutions. Finally, Sorens (2016) argues that central governments may try to appease regional secessionist movements by selectively decentralizing political powers, most notably over policies that are essential for identity formation and welfare, such as education, cultural policies, environment, and health. Congleton (2006) and Zuber (2011), however, show with theoretical models that communities with special fiscal autonomy can compete much more effectively, while regions that lack these unique rights may fall behind economically, creating a political system that is inherently unstable and will eventually gravitate towards 166 M. Rode, H. Pitlik and M. A. Borrella Mas secessionism (cf. Hechter 2007). Sorens (2016) also emphasizes that an asymmetric regional devolution of tax autonomy may be counterproductive to combat Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 secessionist sentiments. It is perfectly possible that granting special rights to certain regions actually motivates secessionist tendencies in the long run: First, due to the fact that inter-regional redistribution is seen as increasingly unjust by the autonomous region itself (Collier and Hoeffler 2006). Second, because the concentration of economic resources in self-governing territories increases the leverage of the minority group (Jenne et al. 2007). Hale (2004) argues that ethno- federalism is particularly unstable if federations are geographically divided along ethnic lines, regions have a single dominant group, and concentrated minority groups enjoy unusually high levels of autonomy. Interestingly, there are currently no quantitative empirical studies that assess the effect of asymmetric federalism and autonomy arrangements on secessionism. Regarding the impact of asymmetric fiscal federalism on secessionist electoral success in Europe, we thus formulate the following second hypothesis: H2: In European regions with an asymmetrically large degree of fiscal autonomy, secessionist parties are more likely to gain electoral support, on average.

Data and Variables In Europe, the phenomenon of regional secessionism is closely linked to the existence of parties that demand formal independence of a corresponding territory. Fortunately, the era of violent secessionist movements in Europe seems to be over, and objectives to create new states are persecuted rather peacefully by regional party platforms. Their existence is, in turn, often associated with a local minority language that is not spoken as a first language by the country’s majority population, but this is not always the case.5 For our purposes, it is sufficient that parties state the (long-term) goal of regional secession from an existing nation as part of their program, disregarding the underlying motivation. In turn, regionalist parties that simply demand greater autonomy for a certain territory within the existing country, such as the Christian Social Union (CSU) in Bavaria, are not counted as secessionist parties. In some cases, there is a gray area between both, and the word “independence” is circumscribed with other formulas, such as “complete political autonomy in a united Europe,” converting it into a bit of a judgment call whether a party is to be labeled secessionist or regionalist.6 Fortunately, these cases are few and most regionally based parties in Europe are either clearly secessionist or regionalist in nature. Languages spoken in minority enclaves in one country might be the dominant language of nearby countries, as is the case with the majoritarian German-speaking South Tirol in Northern . For these regions, the political phenomenon of Does Fiscal Federalism Deter or Spur Secessionism 167 is probably a more adequate description of the political goals expressed by local secessionist parties (Sorens 2005). Due to the fact that these parties clearly Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 aim at changing the territory of an EU member country by having a certain region secede to join another country, we also count them as secessionist in this paper. Since the existence of secessionist parties is not limited to federal states only, we also consider countries in this study that do not fit the classical definition of federalism. Our sample thus includes formally federal and non-federal states. This is further necessary because many fiscally asymmetrical states with minority regions that exhibit territorial self-governance actually exist in formally non-federal countries, such as Finland, Italy, or the . A list of secessionist (and irredentist) parties is presented in table 1, showing country of origin, party name, ideological orientation, the state that it proposes to create (or join), the regions where it is politically present, and the years that this party has been active. This listing is by nowhere a complete catalogue of all secessionist movements throughout Europe. Coverage is limited to EU and associated countries, where sufficient data is available on a regional level, and all entities are subject to the phenomenon of supra-national public goods that might create a common incentive for secession (Alesina and Spolaore 2003; Keating 2012; Muro and Vlaskamp 2016). Only parties that propose secession of an existing administrative region are counted, due to the fact that administrative lines substantially shape secessionist outcomes (Griffiths 2015). For example, the Republican Left of (ERC) also operates as a party in the French department of Pyre´ne´es-Orientales, which is partly Catalan-speaking. As this department is only a small part of the region Languedoc-Roussillon, vote shares for the ERC in southwestern France are not counted in our dataset. Likewise, we do not count vote shares for parties that propose the secession of entire meta-regions that currently do not exist as administrative units, such as the Northern League (LN) in Italy with its proposed fictive state of Padania.7 In order to be classified as relevant, parties had to participate in national parliamentary elections individually, or as members of an electoral coalition, and receive at least an average of 0.5 percent of all votes at the regional level, otherwise they were not counted. The reason for this is twofold: First, we only want to count platforms of some significance. Second, from an electoral standpoint, it is arguably a somewhat more pronounced revelation of secessionist preferences to vote for such a party in national elections. While voting for an openly secessionist movement in regional elections can have many reasons, it can safely be argued that doing so in a nationwide parliamentary election is a clear support of secessionist goals and the desire to give it a voice at the national level. The latter might also be the case in regional elections, but this is unquestionably less clear. According to Sorens (2012), vote shares for secessionist parties are usually higher in regional elections, meaning that our dataset might actually be a lower Table 1 Secessionist parties in Europe (1995–2015) 168

Country Party name Ideology Proposed state Regions active Years active

Belgium New Flemish Alliance (NVA) center-right Flanders Flanders, Brussels 2001–now Flemish Block (VB) right Flanders Flanders, Brussels 1979–2004 Flemish Interest (VB) right Flanders Flanders, Brussels 2004–now Rally Wallonia France (RWF) left Wallonia/France Wallonia 1999–now Czech The Moravians (Moravane´) right Moravia South Moravia, Olomouc, Zlın 2005–now Republic Finland Future of A˚ land (AF) center-right A˚ land A˚ land 2001–now France Alsace First (Alsace d’Abord) right Alsace Alsace 1989–now Our Land (Unser Land) center-right Alsace Alsace 2010–now Breton Democratic Union (UDB) left Brittany Brittany 1964–now Breton Party (PB) center-left Brittany Brittany 2000–now

Corsican Nation (CN) left Corsica Corsica 1992–2009 Mas Borrella A. M. and Pitlik H. Rode, M. Union of the Corsican People (UPC) center-left Corsica Corsica 1977–2002 Party of the Corsican Nation (PNC) center-left Corsica Corsica 2002–now Free Corsica (CL) center-left Corsica Corsica 2009–now Germany Bavaria Party (BP) right Bavaria Bavaria 1946–now Italy Sardinian Action Party (PSdAz) center-left Sardinian Republic 1922–now Sardinia Nation (SN) left Sardinian Republic Sardinia 1994–now The Libertarians (Die Freiheitlichen) right Free-state South Alto Adige – Su¨dtirol 1992–now Tirol/Austria Citizens Union for South Tirol (BU) center-right Free-state South Alto Adige – Su¨dtirol 1989–now Tirol/Austria South Tirolean Freedom (STF) right Free-state South Alto Adige – Su¨dtirol 1997–now

Tirol/Austria Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 May 27 on user Salzburg of Library University by https://academic.oup.com/publius/article-abstract/48/2/161/4562682 from Downloaded osFsa eeaimDtro prSecessionism Spur or Deter Federalism Fiscal Does Venetian League (LV) center-right Venetian Republic Veneto 1979–now Venetian Republic League (LVR) center-left Venetian Republic Veneto 1998–now Venetian Independence (IV) center Venetian Republic Veneto 2012–now Poland Movement for Silesian Autonomy (RAS) right Silesia/Germany Silesia 1990–now Spain Aragonese Union (CHA) left Aragon Aragon 1986–now Basque Nationalist Party (PNV-EAJ) center-right Basque Country Basque Country, Navarra 1895–now Aralar left Basque Country Basque Country, Navarra 2001–now Popular Unity/Unity (HB) left Basque Country Basque Country, Navarra 1986–2011 Gather (Bildu) left Basque Country Basque Country, Navarra 2011–now Basque Solidarity (EA) left Basque Country Basque Country, Navarra 1986–now Cantabrian Nationalist Council (CNC) left Cantabria Cantabria 1995–now Convergence and Union (CiU) center-right Catalonia Catalonia 1978–2015 Republican Left of Catalonia (ERC) left Catalonia Catalonia, Valencia, Balearic 1931–now Islands Socialist Party of Mallorca (PSM) left Catalonia Balearic Islands 1976–now Valencian Nationalist Bloc (BLOC) left Catalonia Valencia 1999–now Valencian Peoples Union (UPV) left Catalonia Valencia 1982–1999 Galician Nationalist Bloc (BNG) left Galicia 1982–now Switzerland Ticino League Lega dei Ticinesi right Ticino/Lombardy (Italy) Ticino 1991–now United We Ourselves (Sinn Fein) left /Ireland Northern Ireland 1905–now Kingdom Social Democratic and Labour Party left Northern Ireland/Ireland Northern Ireland 1970–now (SLDP) Scottish National Party (SNP) center-left Scotland 1934–now Scottish Socialist Party (SSP) left Scotland Scotland 1998–now Party of (Plaid Cymru) center-left Wales Wales 1925–now

Sources: Dandoy (2010), Sorens (2005, 2012), European Free Alliance (EFA) http://www.e-f-a.org/whos-who/member-parties/, Parties and Elections in 169

Europe http://www.parties-and-elections.eu/, and individual party websites. Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 May 27 on user Salzburg of Library University by https://academic.oup.com/publius/article-abstract/48/2/161/4562682 from Downloaded 170 M. Rode, H. Pitlik and M. A. Borrella Mas bound measure of secessionist potential across European regions. Parties might decide not to participate in national elections for policy reasons, so their non- Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 participation is not necessarily a sign of political weakness. Still, when using regional elections as an outcome variable, we may face at least three important data selection problems: First, the fact that regional elections within a country are frequently not be in the same year, because electoral cycles are different across regions. Second, if regional autonomy is only very weak (or absent), regional elections lack substantial relevance, and artificially multiply observations for some regions. Third, it is very difficult to create a representative sample of regional elections that is consistent with a level playing field. In particular, it is not possible to measure the performance of a regional party X in regional elections in region Y, since they are not participating at all in these elections. In contrast, when using national elections, we can observe the performance of all regional parties in the same countrywide election. These problems associated with employing regional elections could potentially create an upward bias for our empirical investigation. Faced with this tradeoff, we prefer a conservative identification approach. Information on secessionist parties was collected from diverse sources. Both Dandoy (2010) and Sorens (2012) present a very ample catalogue of secessionist parties in Europe, but some of these either do not participate in national parliamentary elections, or when they do, receive a vote share of less than half a percent. Therefore, our listing of secessionist parties in table 1 is shorter. In fact, our party catalogue is rather similar to the members’ directory of the European Free Alliance (EFA), which is the official European parliamentary group for secessionist and strongly regionalist parties in EU member states. Note that some of these parties participate in intra-regional coalitions during national parliamen- tary elections that can change over time, but individual candidates (or lists) are clearly identifiable to the regional electorate in all cases.8 Electoral data for these parties was principally collected from the European Election Database, which publishes regional election results according to the Nomenclature of Territorial Units for Statistics (NUTS). As our principal interest lies in the outcomes of symmetric and asymmetric fiscal federalism for secessionist parties’ electoral success, we need a comparative indicator of regional economic decision-making power to capture the former. We employ an indicator by Sorens (2011), which is based on an institutional view of fiscal federalism, measuring the scope of regional economic policy autonomy, tax- raising powers, and institutionalization. Regions are assigned scores in each of these categories, according to the scope of competences that they have. The summary regional level fiscal federalism index is a simple multiplicative product of all three categories, with scores ranging from 0 to 48, to account for complementarities of its sub-dimensions.9 Here, 0 represents no autonomous decision-making power, while a score of 48 represent maximum possible Does Fiscal Federalism Deter or Spur Secessionism 171

16 year Fed. Score Sec. Vote (in%)

Sec.1995 Vote (in%) 18 9 15 Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 15 14 45 Fed.1996 Score 18 9 14 14 13 13 1997 22 10 13 40 1998 22 10 12 12 11 11 1999 2311 10 35 11 11 10 10200010 10 23 10 10 9 9 2001 25 11 30 2002 25 11 25 8 2003 25 11 2004 25 12 20

vote share in % 6 2005 25 11 regional autonomy 2006 27 11 15 2007 28 13 4 2008 27 14 10 2009 28 13 2 2010 28 13 5 2011 28 15 0 2012 28 15 0 1995 1996 1997 19982013 1999 2000 2001 2002 28 2003 2004 2005 14 2006 2007 2008 2009 2010 2011 2012 2013 year

Figure 1 Secessionist party support and fiscal self-governance in active regions (1995-2013). decision-making power in all categories. As the indicator by Sorens (2011) is available only until 2007, we create an updated version for the regions in question that takes into account institutional changes all through 2013. To assess asymmetric federalism effects, we further create a dummy variable for autonomous regions that takes the value of one for regions that have an above average score with respect to the country mean score, and zero otherwise.10

Preliminary Evidence Before discussing the results of our formal tests, figure 1 shows the average secessionist vote share of separatist minority regions and their average score on the fiscal federalism index. Both have changed quite substantially over the course of these eighteen years, where increases in fiscal self-governance are matched by a corresponding rise in secessionist party support, especially from 2007 onwards. Obviously, these figures say nothing about a causal relationship and it is perfectly possible that regions with comparatively higher secessionist support can pressure more easily for territorial self-governance. This important point will be further discussed in the following section. To test our hypotheses H1 and H2, we perform simple Ordinary Least Squares (OLS) regressions of the vote share for all secessionist parties on the regional level of fiscal federalism, introducing a range of regional-level and country-level controls: 172 M. Rode, H. Pitlik and M. A. Borrella Mas

Reg: sec :Voteijt ¼ b0 þ b1fiscal federalismit1 þ b2auton: regionit1 Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 þ b3regional level controlsit1 þ b4country level controlsjt1 þ eijt

The dependent variable is the sum of secessionist party vote shares in a national election in region i and country j during the 1996–2013 period, only taking into account years in which a national election has taken place. This gives us a total of 1,240 observations for 273 European regions from eighteen countries. Of these eighteen countries, sixteen are full EU members (eleven are members of the Eurozone), plus two associated countries, Norway and Switzerland.11 A list of countries and the corresponding sub-national observational units can be found in the appendix (see table A1). All control variables are lagged by one year and include a full array of regional characteristics and some country-wide covariates. We introduce time fixed effects that control for the growth of secessionist party vote shares common to all regions. To account for the Moulton-bias that can cause standard errors of macro- covariates to be far too small (Moulton 1990), we cluster standard errors at the regional level. To save space and focus the analysis on our principal points of interest, exact definitions, underlying reasons for inclusion, and sources of remaining independent variables are described in table 2, where descriptive statistics are also reported. Results of our baseline regressions are shown in table 3, where we also display the full set of controls. There, we find a couple of noteworthy details: Regional per capita income has negative and statistically significant coefficients, while regional per capita income in relation to the national average shows the expected positive sign, but does not systematically explain variation in secessionist voting across regions. Similar to findings by Sorens (2005) and Brancati (2006), wealthier regions in Europe seem to have a reduced presence of secessionist movements. In turn, we do not find evidence for the idea that the relatively wealthy regions of a country are the ones that want to secede (cf. Buchanan and Faith 1987; Collier and Hoeffler 2006; Scharnagl 2012; Sorens 2012; Sambanis and Milanovic 2014). Since we lose about 22 percent of our observations by introducing the regional unemployment rate to our model, we only do so in columns with even numbers. Results show that a comparatively higher regional unemployment rate is negatively associated with vote shares of secessionist parties, indicating that net recipient regions seem to shy away from the risk of losing these transfers in a secessionist experiment. We find no statistically significant impact of the total regional population, but coefficients on the relative size of a region are sizable and statistically significant in most models. This lends some support to the idea that relatively larger regions are perceived by the electorate to be more viable as a country of its own. Findings also Table 2 Variables and descriptive statistics Secessionism Spur or Deter Federalism Fiscal Does

Variable Description Source Obs. Mean SD Min Max

Regional characteristics Secessionist vote (see section 3) EED 1240 1.21 5.54 0 52.1 Federalism (see section 3) Sorens (2011) 1240 14.0 16.2 0 48 (regional) Autonomous region (see section 3) Sorens (2011) 1240 0.04 0.20 0 1 GDP per capita Following Brancati (2006) and Sorens (2005), we use the log of regional OECD 1162 3.13 1.34 1.14 11.6 (log) GDP per capita and regional GDP per capita in relation to the GDP per capita, national average measure relative within-country income of a region. OECD 1162 0.94 0.26 0.42 2.52 relative to the Higher economic development should generally reduce secessionist country average party support, as it might makes voters less likely to support radical solutions. If development is highly concentrated in certain regions though, secessionism might be fueled by economic motivations (cf. Buchanan and Faith 1987, Collier and Hoeffler 2006, Scharnagl 2012, Sorens 2012, Sambanis and Milanovic 2014). Unemployment rate On the other side of intra-regional redistributive conflicts are regions OECD 929 8.5 5.4 0 28.7 that receive relatively more transfers, for example through unem- ployment compensation. We introduce regional unemployment rates, even though its effect on the growth rate of secessionist party votes is somewhat ambiguous: On the one hand, high unemployment rates might capture local discontent, which can ultimately induce voters to support secessionist platforms. On the other hand, it can be an expression of being net recipient of intra-regional transfers and lower support for secessionist parties (cf. Sorens 2016).

(continued) 173 Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 May 27 on user Salzburg of Library University by https://academic.oup.com/publius/article-abstract/48/2/161/4562682 from Downloaded 174 Table 2 Continued

Variable Description Source Obs. Mean SD Min Max

Total population As Collier and Hoeffler (2006), Sorens (2004) and Sambanis and OECD 1240 1653 2298 15 18074 (in 1000) Milanovic (2014) argue, larger regions of a country are more likely to Share of national be secessionist, due to the fact that the formation of an own state OECD 1240 0.07 0.06 0.002 0.58 population seems to be a more realistic possibility for sizable political units. The region’s total population (in 1000), as well as the regional share of total population are introduced as control variables. History of Brancati (2006), Jenne et al. (2007), and Sorens (2012) argue that there CIA World 1240 0.11 0.32 0 1 Independence is a path dependent element to the degree of secessionism. Some Factbook regions may have a history of being independent (or politically autonomous territories), and the existence of an oppression narrative justifies secessionism in the eyes of voters (Lecours 2011). We create a dummy variable with a value of one if the region has a history of Mas Borrella A. M. and Pitlik H. Rode, M. being independent, zero otherwise. Country variables Majoritarian/Mixed Electoral systems significantly influence vote shares for secessionist DPI: Beck 273 0.32 0.47 0 1 electoral system parties in national elections if they discriminate against smaller et al. parties. Following Brancati (2006), we create a dummy that takes the (2001) d’Hondt rule value of one if a majoritarian or mixed electoral system at the DPI: Beck 273 0.47 0.5 0 1 national level is used, and zero otherwise. In case a country uses a et al. proportional representation system, the d’Hondt rule might place (2001) small regional parties at a relative disadvantage (Gallagher 1991). We therefore create a dummy that takes the value of one, if dHondt rule to allocate seats in national parliamentary elections is employed, and

zero otherwise. Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 May 27 on user Salzburg of Library University by https://academic.oup.com/publius/article-abstract/48/2/161/4562682 from Downloaded osFsa eeaimDtro prSecessionism Spur or Deter Federalism Fiscal Does Number of seces- Sorens (2012) highlights that the mere presence of a secessionist Sorens 273 2.0 3.1 0 11 sionist parties discourse in one region of a country can motivate similar tendencies (2012), in other regions that previously had no inclinations as a type of EFA spillover effect. To control for this fact, we introduce the total number of secessionist party movements that exist all over the country on a yearly basis, as shown in table 1. Linguistic Ethnolinguistic diversity is often cited as a reason for the appearance of Alesina et al. 273 0.21 0.17 0.02 0.54 fractionalization secessionist parties (Brancati 2006, Erk and Anderson 2009, Sorens (2003) 2012, Sambanis and Milanovic 2014). We employ data on linguistic fractionalization of countries by Alesina et al. (2003). It measures the probability that two citizens picked at random from the population will be of different linguistic groups, reflecting basic diversity in the population, and potential for secessionist conflict.

175 Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 May 27 on user Salzburg of Library University by https://academic.oup.com/publius/article-abstract/48/2/161/4562682 from Downloaded 176 M. Rode, H. Pitlik and M. A. Borrella Mas

Table 3 Secessionist voting, regional autonomy and fiscal federalism Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020

Dep. Variable Secessionist vote

Methodology OLS OLS OLS OLS OLS OLS (1) (2) (3) (4) (5) (6)

Federalism (reg.) 0.124 ** 0.162 *** 0.081 ** 0.071 * 0.166 *** 0.065 (0.053) (0.054) (0.039) (0.044) (0.052) (0.086) Auton. Region 7.134 ** 10.572 ** 5.900 * 11.326 ** (3.374) (4.606) (3.671) (5.406) Log GDP pc (reg.) 0.673 * 1.295 *** 0.527 * 0.750 ** 3.535 * 1.673 (0.354) (0.454) (0.290) (0.326) (1.801) (2.318) GDP pc share 1.743 1.1189 0.981 0.395 2.431 1.298 (1.315) (1.377) (1.117) (1.245) (1.633) (0.517) Unemployment (reg.) 0.184 *** 0.103 * 0.121 ** (0.062) (0.054) (0.062) Pop. (reg.) 0.000 0.000 0.000 0.000 0.000 0.000 (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Pop. Share 17.767 20.498 19.946 * 24.122 * 16.521 ** 21.619 * (11.383) (12.873) (11.333) (13.233) (7.858) (9.239) Indep. History 2.109 * 1.992 * 1.967 * 1.688 2.416 ** 2.046 ** (1.234) (1.171) (1.161) (1.107) (1.124) (1.081) Maj/Mix elec. sys. 0.194 0.386 0.056 0.336 (1.020) (1.018) (0.983) (0.959) d’Hondt rule 0.762 * 1.384 ** 0.873* 1.401 *** (0.461) (0.558) (0.462) (0.508) No. Sec. parties 0.417 ** 0.418 ** 0.406 ** 0.417 ** 0.279 0.201 (0.187) (0.181) (0.181) (0.177) (0.351) (0.360) Ling. fract. 1.157 2.039 2.768 4.356 * (1.157) (2.102) (1.964) (2.395) N 1,162 906 1162 906 1162 906 Regions 272 270 272 270 272 270 Countries 18 18 18 18 18 18 Country effects No No No No Yes Yes Period effects Yes Yes Yes Yes Yes Yes Adj. R2 0.20 0.23 0.26 0.31 0.29 0.36

Note: Standard errors in parenthesis; *significant at 10%; **significant at 5%; ***significant at 1%; except for columns 5 and 6, all regressions include a constant term. Does Fiscal Federalism Deter or Spur Secessionism 177 lend some support to the idea that regions with a history of independence are more likely to be secessionist. Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 While the electoral system itself seems to have no impact, as expected the use of the d’Hondt rule in proportional representation systems significantly reduces the success of secessionist parties in most models. In contrast, findings strongly support the hypothesis that there are rather important spillover effects from the presence of secessionist movements, as shown by positive and statistically significant coefficients on the number of secessionist parties. Finally, we encounter only weak evidence that linguistic fractionalization drives secessionist electoral success. Coming to our main variables of interest, it can be observed in columns 1 and 2 of table 3 that the regional level of fiscal federalism has a positive and statistically highly significant association with secessionist voting. However, coefficients indicate that the effect is not very large. A one standard deviation increase of regional fiscal federalism is associated with an expansion in secessionist voting of only 0.12 to 0.16 percentage points. Still, the implication of the result is that a higher level of economic territorial self-governance might create small but positive incentives to vote for secessionist parties in minority regions, rather than reducing it, thus contradicting H1. This finding is somewhat put into perspective when introducing a dummy variable for autonomous regions with special rights in columns 3 and 4. While coefficients for fiscal federalism remain positive and significant, the variable capturing asymmetric territorial self-governance enters the equation with a positive sign, and it is statistically significant at the 5 percent level. Coefficients for regional autonomy are substantial in size, showing that these entities have secessionist vote shares that are, on average, seven to eleven percentage points larger. This is a sizable effect, amounting to almost two standard deviations in secessionist voting. The country fixed effects estimates of columns 5 and 6, as well as further robustness checks strongly support this finding.12 So taking into account the impact of asymmetric regional autonomy, our results suggest that regions with a special fiscal status are also more likely to vote for secessionist parties in the next national elections.

Causal Identification Although the preliminary evidence from OLS analysis is appealing, it may be plagued by omitted variable bias and/or reverse causality. For instance, it is possible that our previous results are endogenous to the point that more secessionist regions are also more likely to receive an autonomy status in the first place. Among other factors, secessionist party support is a very conceivable reason for granting certain regions more territorial self-governance. 178 M. Rode, H. Pitlik and M. A. Borrella Mas

In this section, we apply a Difference-in-Differences (DiD) strategy to identify possible causal effects of asymmetric territorial self-governance on secessionist voting. Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 By distinguishing a treatment effect of having regions with a special autonomy status and a time effect, this methodology allows us to exclude endogeneity concerns on our main findings. Specifically, we estimate the following model:

Reg: sec :Voteijt ¼ b0 þ b1Auton:regionit1 þ b2Eurojt1 þ b3Auton:regionEuroijt1 þ kt þ eijt where the dependent variable is again the regional vote share of secessionist parties in national elections, and the treatment variable is our autonomous region dummy. The time effect is captured by a Euro dummy, which is equal to 1 for all observations from the year 2002 onwards, corresponding to the physical introduction of the Euro. We use the introduction of the common currency as an exogenous event, because it clearly represents the most visible creation of a supranational public good during our observation period. Following Alesina and Spolaore (2003), Desmet et al. (2011), and Keating (1999, 2012), introduction of the Euro should reduce the subjective benefits of pertaining to a large nation state inside the EU. This would even count for secessionist regions in countries that are not part of Euro zone, as there is a credible alternative for joining a different monetary area that is now available to regional policymakers. This is well illustrated by the recent calls for Scotland to join the Euro area, if it becomes an independent state without access to the British Pound (Johnson 2016). In line with this idea, Muro and Vlaskamp (2016) find in a recent survey analysis for Scotland and Catalonia that secessionist support is significantly increased by prospects of direct EU membership. However, all results hold if we restrict the estimated sample to countries of the Eurozone or the EU. In some specifications we also include year dummies (kt Þ and the full set of regional and country level controls. Standard errors are clustered at the regional level.

In this setting, b3 is our parameter of interest, measuring the causal effect of asymmetric fiscal federalism on the regional vote share of secessionist parties, thus delivering a direct test of our H2. Note that the inclusion of control variables in DiD estimators does not affect the validity of estimates in general, but only increasing their precision, giving them a somewhat reduced importance (Blundell and Costa Dias 2000). According to the common trend assumption in any DiD methodology, the outcome in treatment and control group would follow the same time trend in the absence of treatment. Although this assumption is not testable for us, because the history of fiscal federalism is obviously not the same in both groups of regions, we can nonetheless implement several tests to reassure the validity of our estimation strategy (Blundell and Costa Dias 2000). Does Fiscal Federalism Deter or Spur Secessionism 179 Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020

Figure 2 Secessionist voting and asymmetric fiscal federalism.

First, taking advantage of the availability of several years of data, we can graphically analyze whether the outcome variable followed parallel trends over the period analyzed. Visual inspection of trends in figure 2 for control and treatment groups clearly suggests that the parallel trend assumption is not violated. In the absence of treatment, the change in the outcome of the treated group would have been similar to the change for the non-treated group. Second, we can formally test the evolution of pre-trends by interacting the treatment variable with time dummies (Autor 2003). To explore these dynamics, we estimate our main specification described by the DID equation, augmented with leads and lags of having regions with a special autonomy status. The estimated model is as follows:

X1 Reg: sec :Voteijt ¼ b0 þ dsAuton:regioniyear2002þs s¼5

X10 þ csAuton:regioniyear2002þs þ kt þ eijt s¼0 where the dependent variable is the same as before, kt are year dummies, and we include the interactions of time dummies and the treatment indicator for all 180 M. Rode, H. Pitlik and M. A. Borrella Mas

Table 4 Secessionist voting and asymmetric fiscal federalism (DID) Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020

Dep. Variable Secessionist vote

Methodology DiD DiD DiD DiD (1) (2) (3) (4)

Auton. Region 4.497 3.744 3.142 5.295 (2.739) (2.550) (2.512) (3.480) Euro 0.149 1.883* 0.548 0.012 (0.198) (1.007) (0.781) (0.661) Auton. region*Euro 11.690 ** 11.971 *** 12.103 *** 10.094 ** (4.571) (4.451) (4.098) (4.156) N 1240 1240 1162 906 Regions 272 272 272 270 Countries 18 18 18 18 Country effects No No No No Period effects No Yes Yes Yes Control variables No No Yes Yes Unemployment No No No Yes

Note: Standard errors in parenthesis; *significant at 10%; **significant at 5%; ***significant at 1%;. pretreatment periods (“leads”) and all posttreatment periods (“lags”).13 The intuition is as follows. If the outcome trends between treatment and control groups are the same, then all ds should be insignificant, i.e., these coefficients should not be significantly different between both groups in pretreatment periods. Table A3 of the Appendix shows the estimated impact of having regions with a special autonomous status for all years before-after the physical introduction of the Euro. Consistent with this assumption, the leads are not significant, providing strong evidence for no anticipatory effects and support for the validity of the common trend assumption. Finally, another possible way of analyzing whether the treatment and the control group are comparable is to implement a placebo DiD. Given that the Euro was physically introduced at the beginning of 2002, we should not expect a significant change in years prior to 2002. Table A4 of the Appendix replicates DiD estimates, using as threshold years 2000 and 2001. Small and non-significant estimates of the “fake” interaction terms provide renewed support for our identifying strategy. Table 4 shows the estimation results of our baseline DiD specifications, where we also employ the full set of regional and countrywide control variables. We find Does Fiscal Federalism Deter or Spur Secessionism 181 autonomous regions within special fiscal competences to have a positive and significant effect on the vote shares of secessionist parties in all cases, obtaining Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 very similar coefficients to those of the OLS analysis of the previous section. Furthermore, results are robust to the inclusion of year fixed effects (column 2), year fixed effects and control variables (column 3) and year fixed effects, time- variant control variables, and unemployment rates (column 4). Consistent with econometric theory, the inclusion of control variables does not affect our estimates, but only increases their precision. Finally, we also examine whether the results are robust when focusing only on countries in which the Euro was introduced, which does not overturn our findings (table A2). Consistent with our H2, we thus find strong support for a substantive positive and causal effect of asymmetric territorial self-governance on secessionist party voting, despite the fact that much support for secession is determined by pre- existing sources of support. This is a highly interesting finding, because it sheds a somewhat different light on the practice of granting an unusually ample autonomy to minority regions in the framework of an asymmetric fiscal federalism: By giving these territories more economic competences than other regions, secessionist sentiments are encouraged, and not reduced.

Conclusions The recent surge of secessionist tendencies in various EU member states has produced a flood of reports, which often recommend a more ample federalism to combat the of minority regions. The findings in this article suggest that a is only marginally associated with the growth of secessionist platforms during the last twenty years. Given the well-documented economic benefits of fiscal federalism, our results provide certainly no case against a further devolution process in European countries, also because the viable alternatives to a federal system for group conflict management are non-existent (Anderson 2016). Still, European governments concerned with secessionism should probably consider the design of the devolution path and the associated incentives. An asymmetric fiscal federalism, where special autonomy rights exist for some (secessionist) regions means that these competences can easily be expanded, if the individual region is just able to create sufficient political pressure on the central government, effectively creating a dual-power conflict between regional authority and central government (Hale 2004). If these regions also observe a superior economic performance, as a consequence of their greater fiscal competences (Hechter 2007), their political leverage vis-a-vis the central government is enhanced, possibly leading to more radical demands of minority regions (Jenne et al. 2007). In such a setting, secessionism can become a negotiation strategy, once regional electorates realize that voting for secessionist parties is an excellent 182 M. Rode, H. Pitlik and M. A. Borrella Mas possibility to signal a desire for enhanced power and raise the stakes of the game (Eerola et al. 2004). According to Grossman and Lewis (2014), a secessionist Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 discourse will also incentivize regional elites to mobilize, since it offers them the best possibility to hold office and control the allocation of public resources in the future. In addition, it is also likely that political entrepreneurs of other historically distinct regions will pick up a secessionist discourse in such a climate, making asymmetrical federal rules inherently unstable (Zuber 2011). Under these circumstances, a symmetric devolution is probably preferable to the selective granting of territorial self-governance. According to the ideas of Keating (2012), an alternative to national conflict management is nowadays offered by the EU, where an effective form of multilevel governance could make a different concept of sovereignty possible, ultimately making secession pointless. This would be more in line with a Europe where the differences between domestic and external policies are increasingly hard to draw. Still, it is questionable how concepts of supra-national governance and “independence light” are supposed to work in practice, especially when the EU is currently being confronted with the United Kingdom seceding from its own governance architecture, and the continuing threat of a total break-up lingers on in the Euro-crisis. Finally, future research on the topic could potentially tackle two limitations of the present study: First, to see whether our focus on national elections makes significant differences for the findings, as it is possible that we underestimate the true effect of federalism on secessionist voting. Second, by corroborating whether the results hold for federal and quasi-federal countries with important secessionist movements outside Europe. The latter is very well possible and would constitute an important test for the findings forwarded in this article.

Notes We would like to thank Aziz Berdiev, Christian Bjørnskov, Shai Dothan, Kai Gehring, Monika Ko¨ppl-Turyna, Niklas Potrafke, Panu Poutvaara, Stephan Schneider, Jason Sorens, three anonymous referees, and the editor John Dinan for helpful comments on earlier drafts of the article. All remaining errors are naturally our own. 1. This might have motivated the government of David Cameron to propose a more amplified autonomy for Scotland, just days before the 2014 referendum on secession came out in favor of remaining part of the UK. Also, in Spain, an amplified autonomy for Catalonia, similar to the far reaching self-rule of the Basque Country and Navarra, is frequently discussed as an alternative to an independence referendum, and it is portrayed as a possible way of reducing conflict with national authorities in Madrid. 2. Recently, Vaubel (2013) has argued on similar grounds that the right to secession should actually be part of the EU treaties. 3. As Griffiths (2015) highlights, the secession of existing administrative territories also has a much higher possibility of being conducted in a peaceful manner. Does Fiscal Federalism Deter or Spur Secessionism 183

4. Keating (1999) makes the important point that all pre-modern states were essentially Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 asymmetrical, and that asymmetry remains an important feature of many present European states, including the EU itself. 5. For example, the Basque nationalism of the Basque National Party (PNV) is intimately connected to the existence of Basque as a minority language, while the Scottish nationalism of the Scottish National Party (SNP) is not significantly based on the defense of a minority language. 6. This terminology is probably employed in an attempt to maximize the party’s vote share, because it is judged to be more acceptable for large parts of the local electorate. 7. To be exact, we do count one regional members of the Italian Northern League (LN) alliance in our dataset, which is the Venetian League (LV). This is due to the fact that the LV formally keeps pursuing the goal of Venetian independence. In contrast, we do not count other regional members, such as the Lombardy League, because this LN member does not seek the creation of an independent Lombardy. 8. This is mostly the case for secessionist parties in Finland, France, and Poland (and to a lesser degree in Italy and Spain). Excluding these three countries from the estimations has absolutely no effect on our findings. These are available from the authors upon request. 9. Sorens (2011) argues that multiplication better reflects the fact that undermining autonomy in one of these three areas basically undermines the whole objective of regional autonomy. 10. National averages in fiscal federalism are given by Sorens (2011) as a population proportion-weighted average of the regional scores. 11. Switzerland has developed an ample network of bilateral agreements with the EU since the early 1970s that lets it participate in the common market, having even joined the Schengen treaty in 2008. We can therefore safely assume that the external conditions for Switzerland, in terms of security and trade, are very similar to those of its immediate neighbors. 12. Additional control variables, multilevel estimation methods, and an alternative political federalism measure, based on membership in the Regions with Legislative Power in Europe association (http://www.regleg.eu/index.php), all provide practically identical results. These are not shown, but available from the authors upon request. 13. Specifically, we include five out of six years before the introduction of the Euro, and eleven out of twelve years after the creation of the common currency, for reasons of perfect multicollinearity.

References Alesina, Alberto, Arnaud Devleeschauwer, William Easterly, Sergio Kurlat, and Romain Wacziarg. 2003. Fractionalization. Journal of Economic Growth 8 (2): 155–194. Alesina, Alberto and Enrico Spolaore. 2003. The Size of Nations. Cambridge, MA: MIT Press. Alonso, Sonia, and Rube´nRuız Rufino. 2007. Political representation and ethnic conflict in new democracies. European Journal of Political Research 46 (2): 237–267. 184 M. Rode, H. Pitlik and M. A. Borrella Mas

Anderson, Liam. 2016. Ethnofederalism and the management of ethnic conflict: Assessing the alternatives. Publius: The Journal of Federalism 46 (1): 1–24. Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 Anderson, Lawrence M. 2004. The institutional basis of secessionist politics: Federalism and secession in the United States. Publius: The Journal of Federalism 34 (2): 1–18. Autor, David. 2003. Outsourcing at will: The contribution of unjust dismissal doctrine to the growth of employment outsourcing. Journal of Labor Economics 21 (1): 1–42. Bakke, Kristin M. 2015. Decentralization and intrastate struggles. Chechnya, Punjab, and Que´bec. Cambridge: Cambridge University Press Basta, Karlo. 2018. The state between minority and majority nationalisms: Decentralization, symbolic recognition, and secessionist crisis in Spain and . Publius: The Journal of Federalism 48 (1): 51–75. Benedikter, Thomas. 2007. The world’s working regional autonomies. London: Anthem Press. Beck, Thorsten, George Clarke, Alberto Groff, Philip Keefer, and Patrick Walsh. 2001. New tools in comparative political economy: The database of political institutions. World Bank Economic Review 15 (1): 165–176. Bermeo, Nancy. 2002. The import of institutions. Journal of Democracy 13 (2): 96–110. Blundell, Richard, and Monica Costa Dias. 2000. Evaluation methods for non-experimental data. Fiscal Studies 21 (4): 427–468. Brancati, Dawn. 2006. Decentralization: Fueling the fire or dampening the flames of ethnic conflict and secession?. International Organization 60 (3): 651–658. Buchanan, James M., and Roger L. Faith. 1987. Secession and the Limits of Taxation: Toward a Theory of Internal Exit. The American Economic Review 77 (5): 1023–1031. Bunce, Valerie. 2003. Rethinking recent democratization: Lessons from the postcommunist experience. World Politics 55 (2): 167–192. Cederman, Lars-Erik, Simon Hug, Andreas Sch€adel, and Julian Wucherpfennig. 2015. Territorial autonomy in the shadow of conflict: Too little, too late?. American Political Science Review 109 (2): 354–370. Coakley, John. 1990. National minorities and the government of divided societies: A comparative analysis of some European evidence. European Journal of Political Research 18 (4): 437–456. Collier, Paul, and Anke Hoeffler. 2006. The political economy of secession. In Negotiating self-determination, ed. Hannum, H. and E. F. Babbitt, 37–59. Lanham: Lexington Books. Congleton, Roger D. 2006. Asymmetric federalism and the political economy of decentral- ization. In Handbook of fiscal federalism, ed. Ahmad, E. and G. Brosio, 131–153. Cheltenham: Edward Elgar. Dandoy, Re´gis. 2010. Ethno-regionalist parties in Europe: a typology. Perspectives on Federalism 2 (2): 194–220. Desmet, Klaus, Michel Le Breton, Ignacio Ortuno~ Ortın, and Shlomo Weber. 2011. The stability and breakup of nations: A quantitative analysis. Journal of Economic Growth 16 (3): 183–213. Does Fiscal Federalism Deter or Spur Secessionism 185

Eerola, Essi, Niku M€a€att€anen, and Panu Poutvaara. 2004. Citizens should vote on secession. Topics in Economic Analysis and Policy 4 (1): 23. Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 Erk, Jan, and Lawrence M. Anderson. 2009. The paradox of federalism: Does self-rule accommodate or exacerbate ethnic divisions?. Regional and Federal Studies 19 (2): 191–202. European Election Database. 2016. Norwegian Social Science Data Services. http://www.nsd. uib.no/european_election_database/about/ Gallagher, Michael. 1991. Proportionality, disproportionality and electoral systems. Electoral Studies 10 (1): 33–51. Gehring, Kai, and Stephan A. Schneider. 2016. Regional resources and democratic secessionism. Working Paper, University of Zurich. Gurr, Ted R. 2000. Peoples versus States: Minorities at Risk in the New Century. Washington DC: United States Institute of Peace Press. Griffiths, Ryan 2015. Between dissolution and blood: How administrative lines and categories shape secessionist outcomes. International Organization 69 (3): 731–751. ———. 2016. The age of secession. The international and domestic determinants of state birth. Cambridge: Cambridge University Press. Grossman, Guy, and Janet I. Lewis. 2014. Administrative unit proliferation. American Political Science Review 108 (1): 196–217. Hechter, Michael. 2007. Asymmetrical federal systems: Self-determination, cultural identity, and political and fiscal decentralization. Ethnopolitics 6 (1): 125–127. Hale, Henry E. 2004. Divided we stand. Institutional sources of ethnofederal state survival and collapse. World Politics 52 (2): 165–193. Horowitz, Donald L. 1985. Ethnic groups in conflict. Berkeley: University of California Press. Jenne, Erin K., Stephen M. Saideman, and Will Lowe. 2007. Separatism as bargaining posture: The role of leverage in minority radicalization. Journal of Peace Research 44 (5): 539–558. Johnson, Simon. 2016. Nicola Sturgeon: SNP will discuss using euro if Scotland independent after Brexit. The Telegraph, June 21. Keating, Michael. 1999. Asymmetrical government: Multinational states in an integrating Europe. Publius: The Journal of Federalism 29 (1): 71–86. ———. 2012. Rethinking sovereignty. Independence-light, devolution-max and national accommodation. Revista D’Estudis Autonomics i Federals 16: 9–29. Lecours, Andre´. 2011. Sub-state nationalism in the western world: Explaining continued appeal. Ethnopolitics 11 (3): 268–286. Lustick, Ian S., Dan Miodownik, and Roy J. Eidelson. 2004. Secessionism in multicultural states: Does sharing power prevent or encourage it?. American Political Science Review 98 (2): 209–229. McGarry, John. 2007. Asymmetry in federations, and unitary states. Ethnopolitics 6 (1): 105–116. 186 M. Rode, H. Pitlik and M. A. Borrella Mas

McGarry, John, and Brendan O’Leary. 2009. Must pluri-national federations fail? Ethnopolitics 8 (1): 5–25. Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020 Moulton, Brent. 1990. An illustration of a pitfall in estimating the effects of aggregate variables in micro units. Review of Economics and Statistics 72 (2): 334–338. Muro, Diego, and Martijn Vlaskamp. 2016. How do prospects of EU membership influence support for secession? A survey experiment in Catalonia and Scotland. West European Politics 39 (6): 1115–1138. Oates, Wallace E. 1999. An essay on fiscal federalism. Journal of Economic Literature 37 (3): 1120–1149. OECD. 2016. OECD Statistics Regional Database. http://stats.oecd.org/Index.aspx? datasetcode¼REG_DEMO_TL2 Prud’homme, Remy. 1995. The dangers of decentralization. World Bank Research Observer 10 (2): 201–220. Rodden, Jonathan. 2004. Comparative federalism and decentralization: On meaning and measurement. Comparative Politics 36 (4): 481–500. Sambanis, Nicholas, and Branko Milanovic. 2014. Explaining regional autonomy differences in decentralized countries. Comparative Political Studies 47 (13): 1830–1855. Scharnagl, Wilfried. 2012. Bayern kann es auch allein: Pladoyer€ fu¨r den eigenen Staat. Cologne: Bastei Lu¨bbe (Quadriga). Sorens, Jason P. 2004. Globalization, secessionism, and autonomy. Electoral Studies 23 (4): 727–752. ———. 2005. The cross-sectional determinants of secessionism in advanced democracies. Comparative Political Studies 38 (3): 304–326. ———. 2011. The institutions of fiscal federalism. Publius: The Journal of Federalism 41 (2): 207–231. ———. 2012. Secessionism. Montreal: McGill-Queen’s University Press. ———. 2016. Secession risk and fiscal federalism. Publius: The Journal of Federalism 46 (1): 25–50. Stepan, Alfred 2001. Arguing comparative politics. Oxford: Oxford University Press. Vaubel, Roland. 2013. Secession in the European Union. Economic Affairs 33 (3): 288–302. Weingast, Barry R. 2013. Second generation fiscal federalism: Political aspects of decentralization and economic development. World Development 53: 14–25. Wolff, Stefan. 2013. Conflict management in divided societies: The many uses of territorial self-governance. International Journal on Minority and Group Rights 20 (1): 27–50. World Factbook. 2017. Central Intelligence Agency. https://www.cia.gov/library/publications/ the-world-factbook/index.html Zuber, Christina I. 2011. Understanding the multinational game: Toward a theory of asymmetrical federalism. Comparative Political Studies 44 (5): 546–571. Does Fiscal Federalism Deter or Spur Secessionism 187

Appendix Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020

Table A1 List of countries and observational units

Country No. of Sub-state units Name of units NUTS level

Austria 9 Lander€ 2 Belgium 3 Re´gions/Gewesten 1 Czech Republic 14 Kraje 3 Finland 20 Maakunta 3 France 22 Re´gions 2 Germany 16 Lander€ 1 Hungary 19 þ capital Megye´k 3 Ireland 8 Regions 3 Italy 21 Regione 2 Netherlands 12 Provincies 2 Norway 19 Fylke 3 Poland 16 Wojewodztwa 2 Portugal 7 Regioes~ 2 Slovak Republic 8 Kraje 3 Spain 17 (þ2) Comunidades Autonomas 2 Sweden 21 Lan€ 3 Switzerland 26 Kantone 3 UK 12 Regions 1 188 M. Rode, H. Pitlik and M. A. Borrella Mas

Table A2 Secessionist voting and asymmetric fiscal federalism (DiD) for the Eurozone Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020

Dep. Variable Secessionist vote

Methodology DiD DiD DiD DiD (1) (2) (3) (4)

Auton. Region 5.643 4.805 5.950 6.569 * (3.865) (3.747) (3.788) (3.864) Euro 0.480 2.284 * 1.064 0.487 (0.310) (1.357) (1.204) (1.163) Auton. region*Euro 8.210 * 8.899 ** 8.646 ** 8.906 ** (4.899) (4.492) (3.560) (3.978) N 639 639 639 515 Regions 137 137 137 137 Countries 10 10 10 10 Country effects No No No No Period effects No Yes Yes Yes Control variables No No Yes Yes Unemployment No No No Yes

Note: Standard errors in parenthesis; *significant at 10%; **significant at 5%; ***significant at 1%;. Does Fiscal Federalism Deter or Spur Secessionism 189

Table A3 Evolution of pre-trends (DiD) Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020

Dep. Variable Secessionist vote

Methodology DiD DiD DiD (1) (2) (3)

1997 8.616 9.687 * (5.417) (5.405) 1998 0.000 0.000 (0.000) (0.000) 1999 8.114 5.635 (5.399) (5.277) 2000 11.677 10.597 1.290 (9.653) (9.693) (10.955) 2001 2.943 4.836 13.994 (5.669) (5.491) (8.494) 2002 7.573 4.902 13.909* (5.386) (5.193) (7.818) 2003 1.729 0.223 8.777 (5.404) (5.217) (7.835) 2004 19.588 ** 18.516 ** 9.258 (7.747) (7.732) (9.294) 2005 0.975 0.471 9.870 (8.252) (7.018) (7.669) 2006 0.000 0.000 0.000 (0.000) (0.000) (0.000) 2007 0.341 1.473 7.677 (5.419) (5.221) (7.836) 2008 21.067 *** 18.580 ** 9.495 (7.443) (7.339) (9.024) N 1,240 1,162 906 Regions 272 272 270 Countries 18 18 18 Country effects No No No Control variables No Yes Yes Unemployment No No Yes

Note: Standard errors in parenthesis; *significant at 10%; **significant at 5%; ***significant at 1%. 190 M. Rode, H. Pitlik and M. A. Borrella Mas

Table A4 Placebo test using “fake” threshold years (DiD) Downloaded from https://academic.oup.com/publius/article-abstract/48/2/161/4562682 by University Library of Salzburg user on 27 May 2020

Dep. Variable Secessionist vote

Methodology DiD DiD DiD DiD (1) (2) (3) (4)

YEAR 2000 Auton. Region 4.575 3.867 3.516 17.607 (3.582) (3.345) (3.320) (12.335) Euro 0.162 1.864 * 0.491 0.077 (0.203) (0.965) (0.820) (0.686) Auton. region*Euro 4.677 4.014 7.662 *** 6.271 (2.984) (2.923) (2.605) (10.809) YEAR 2001 Auton. Region 4.593 4.037 3.169 5.825 (2.890) (2.717) (2.693) (4.260) Euro 0.250 1.839 * 0.505 0.067 (0.246) (1.000) (0.785) (0.654) Auton. region*Euro 5.099 5.130 5.833 4.351 (4.798) (4.601) (4.073) (4.592) N 1,240 1,240 1,162 906 Regions 272 272 272 270 Countries 18 18 18 18 Country effects No No No No Period effects No Yes Yes Yes Control variables No No Yes Yes Unemployment No No No Yes

Note: Standard errors in parenthesis; *significant at 10%; **significant at 5%; ***significant at 1%.