POLITICAL PARTIES IN MULTI-LAYERED SYSTEMS ★ Kris Deschouwer Vrije Universiteit Brussels, Belgium Abstract The language used to analyse party behaviour is intrinsically single-level in nature. Recent processes of institutional reform undertaken across Europe – prompting the development of meaningful party competition at the non-statewide level – have had little influence in altering this discourse. There is therefore an urgent need to develop nuanced frameworks suitable for understanding party activity in different electoral arenas. Most of the literature on party behaviour in multi-layered systems thus far has been very much case-oriented (one party, one country, one region) or has focused rather superficially The familiar language to describe and analyse the behaviour of political parties is – often only implicitly – a single-level language. That is not too surprising. Indeed, parties and party systems came into being as the result of a process of boundary closure, of the formation of national states and of a more or less simultaneous territorialization and democratization of politics. Major boundarycrossing societal conflicts were domesticated and consequently ‘frozen’ in national systems of partypolitical competition (Lipset and Rokkan, 1967; Bartolini, 1998; Flora et al., 1999). The institutional environment in which political parties have to function today, however, is becoming to an increasing extent, and in varying ways, multilayered. One of the reasons for this is the politicization of the European level of policy making. This has resulted in increased attention to public attitudes and thus voting behaviour in elections to the European Parliament. The increasing importance of the European Union has led to the labelling of the European system as ‘multi-level’ and to the rapid development of a wide body of research attempting to understand political activities and behaviour occurring in a system in European Urban and Regional Studies 10(3): 213–226 0969-7764[200307]10:3; 213–226;035822 only on the electoral dimension.This article explores the conceptual problems that arise in analysing such questions, suggesting some solutions for identifying the most appropriate unit of analysis and for identifying and eventually measuring aspects of the multi-layered political systems where we might expect to see some direct effects on the functioning of political parties. KEY WORDS ★ multi-layered systems ★ political parties which levels interact in complex ways. The literature on multi-level governance is, however, very much a party-free zone. The focus is on decision making and on implementation, and generally on non-party actors. One of the reasons for this is obviously the lack of real party politics at the European level. Even if the EU might be treated theoretically as one single – though multi-layered – political system (Hix, 1999), the role of political parties in the EU remains rather limited. Probably the strongest argument against the idea that the EU can be seen as a political system that can be compared with other (national) systems is the absence of party politics and especially of party government. Analysis of European elections has nevertheless generated a number of interesting concepts and ideas that can be very useful in the search for a more general conceptual language for dealing with party activities in multi-layered systems. The notion of ‘second-order elections’, that was coined in direct reference to the first European elections, has found its way into more general and comparative approaches (Heath et al., 1999). A second reason for increasing attention to party Copyright © 2003 SAGE Publications London, Thousand Oaks, CA and New Delhi, www.sagepublications.com Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016 214 EUROPEAN URBAN AND REGIONAL STUDIES 10(3) behaviour in multi-layered settings is undoubtedly the kind of administrative developments taking place in a number of European national states. The regionalization processes in Belgium, in Spain and recently also in the UK have created new multinational and also non-symmetrical regional and/or federal-type states.1 Parties and party systems differ between the regions and therefore also between the regions and the higher state level. Parties in these states are thus confronted with a very peculiar structure of opportunities, which has naturally attracted the attention of a number of scholars. Regional elections have become an interesting research area, and the relations between regional and national elections are generating questions relating to the strategic behaviour of parties in this new institutional environment. Most of the work done here follows a case-study approach, referring to one country or one region, but this can also be a good source of inspiration for identifying more general patterns for theory building and for comparative analysis. The latter is indeed the aim of this article. At this moment in time, we do not possess welldeveloped conceptual language that would allow us to set refined research questions that might enable us to systematically analyse political parties in a multi-layered political system. This article explores the conceptual problems that arise in this respect, and suggests some solutions for identifying the correct unit of analysis and for identifying and eventually measuring aspects of the multi-layered political system where we might expect to see some direct effects on the functioning of political parties. As was said above, interest in this phenomenon is increasing. This offers us two related but distinct sources of inspiration. The developments at the European level and the national regionalization processes are indeed two quite different phenomena, especially in view of the absence of classical party politics in the EU polity. The kind of dynamics developing in parties and the way in which they respond to the multi-layered institutional structure can be different in both cases. Probably the national level in regional states is the most interesting and fruitful area where the changing patterns of party politics can be explored. One can expect dual interactions here with the national level influencing the regional level and vice versa. In the European story the most interesting focus is on the way in which the European level (in a broader sense than only European party politics) is impacting on the domestic (national and regional) patterns of party competition (Ladrech, 1994; Guyomarch, 1995; Mair, 2000). The (potential) development of party politics in the regional states will thus be my first and most important reference point. The article will, however, also point at dynamics that are more typical for the interaction between the European and the national level. For the sake of clarity (and of parsimony) I will limit my reasoning to the interaction between only two levels of government. The European literature on multi-level governance takes as a starting point the development of a third level – the region – and concludes from this that a very new type of polity has seen the light of day (Hooghe, 1996; Marks and Hooghe, 2001). These three-way interactions are indeed interesting and also highly relevant for the functioning of political parties. It is, though, extremely difficult to produce a parsimonious theoretical framework for dealing with three levels at once. Limiting the conceptual exploration to only two levels is of course too easy a way out of the problem. We should keep in mind that interactions in practice can occur between more than two levels, and Europeanization means an impact on both the national and the regional (and eventually also lower) levels. The unit of analysis: parties or party systems? The adding of one or more levels to the classic single-level political systems that have guided our conceptual language for the analysis of parties and party systems poses serious conceptual problems. The first one is the unit of analysis. In a single-level system it is easy to identify the party system (the interactions between the parties of that political system) and easy to identify the party actors within it. With more than one level, things become very much more complex. In Figure 1 I present a simple map of a two-level political system with four ‘regions’. At both levels we can imagine parties being active. I have given them numbers per party or party family. P1 might for instance refer to the social-democratic party at European Urban and Regional Studies 2003 10(3) Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016 DESCHOUWER: POLITICAL PARTIES State level Regional level P1 P1 P2 P3 | P1 P2 P2 P4 215 P3 | P1 P3 P4 | P1 P2 P5 Figure 1 Parties and party systems over two levels the state level and to its four regional units. In that case P1 is a single party. If P1 refers to a party federation, it consists of four member parties of the same ideological family, rather than a single party. This way of presenting the parties actually assumes that we are able to make these classifications and that they are meaningful. The problem is not so much the identification of ‘social democracy’ or of any other meaningful label, but the implicit choice that is being made for a unit of analysis. If we say a party and its ‘units’, we assume that there is indeed one single party, and that the units are the composing parts of the national party. Empirical reality can be very different. We will come back to this later. In Figure 1 we can see that there are a number of possible interactions between party actors. All the possible interactions should be identified. There are five different types: 1. Vertical interactions within parties: relations between P1 at the federal level and all the regional units of P1. 2. Vertical interactions between parties: the actions undertaken by P2 at the central level (e.g. forming a coalition with P3) might have direct effects on the strategies to be chosen by P1 in one of the regions. That can of course also be the other way round. If the Volksunie joins the regional government in Flanders, in a coalition with Greens, Socialists and Liberals, while it stays out of the same coalition at the federal level, the party is faced with a number of difficult strategic choices. The federal government might engage in institutional changes that are criticized by the Volksunie as being too limited, while their partners in the regional government support the institutional reform. This example is not a virtual one: the Volksunie indeed broke apart in 2001 as a result of exactly this kind of strategic discussions. 3. Horizontal interactions between parties of one system: these are the normal or classical interactions of a party system, that can be measured with the classical indicators referring to number of parties, polarization, Left–Right orientation, volatility, etc. 4. Horizontal interactions within parties between systems: the actions undertaken by P1 in the first region might affect the strategies of P1 in one of the other regions. If the Austrian Christian Democrats enter a coalition with the populists of the FPÖ, other Christian Democratic parties in Europe are facing new challenges and choices (and the European-level organization is also affected – another example of Type 2). These horizontal interactions occur and become relevant because the higher level is a meaningful level of decision making, really pulling the regions together into one political system, creating some sort of common political community. European integration is a good example of the way in which formerly more isolated national parties and party systems are being mutually affected by what happens in other countries. The same should be the case in countries going through a process of regionalization, since one can assume that the existing common political community still survives. (Variations are of course possible, and the extreme case of the total breakdown of the common community is not unthinkable.) 5. Horizontal interactions between parties: since the party systems belong to one common political system, this interaction can also be conceived as becoming more frequent and more relevant. The relations between Plaid Cymru and the Labour Party in Wales might for instance European Urban and Regional Studies 2003 10(3) Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016 216 EUROPEAN URBAN AND REGIONAL STUDIES 10(3) have an influence on the relations between the Scottish Labour Party and the SNP. These five possible types of interactions could be used to define in general a multi-layered party system. The characteristics of such a party system would then be the properties of all these relations, and also the differences and similarities between the party systems at one level and between levels. If we furthermore take into account that these relations might differ between the electoral, the parliamentary and the governmental arenas, this multi-layered party system would probably be the least parsimonious concept ever produced. Complexity becomes so huge that nothing meaningful can be said any more. There is, however, a way to simplify the model, or at least to reduce the challenge of trying to say something about everything. If we look at all these possible interactions, we realize that some are more familiar than others. The third category is, as we said above, the normal category of interactions of parties in a party system. Differences between these systems – both horizontally and vertically – do have to be looked at, however, because they can explain why parties behave differently on this or on that level, in this or in that region. The first and the fourth categories are also a bit more familiar. They deal in fact with vertical relations inside political parties. They thus refer to intra-organizational characteristics of political parties. Rather than belonging to the party system, these relations belong to party organization. What we witness here is in fact a strange and disturbing blurring of the difference between parties and party systems. This is, though, a consequence of our decision to treat a party of the same family being active at different levels as one single party unit. As was already said above, this is a theoretical assumption that needs to be qualified. It strongly depends on the internal cohesion and vertical integration of that party family. We would probably not hesitate to treat a Land section of the German SPD as belonging to the federal SPD, and analyse its vertical relations as belonging to the internal organization of that federal party. Yet the European level or party organization is much looser (or ultimately even too loose to qualify as a real party), which means that the relations between SPD in Germany and the other Social-Democratic parties are actually relations between parties rather that within the European-level Socialist Party. That is not only because the European Socialists are weakly integrated (other families have even looser ties), but because of the nature of the European political system itself. It is itself less integrated than ‘national’ federal-type systems (though Belgium or Switzerland might come closer to the European degree of integration). The unit of analysis can thus not be defined in an abstract way. It depends on where we are and on what we look at. We will therefore need to identify criteria with which we can make a clear difference between intra-party and inter-party relations. We will have to search for indicators that can tell us where the core of a party can be found. The second and fifth type of relations that we identified above are ‘new’, in the sense that they are typical and unique for multi-layered systems: interactions between parties are not confined to one single level of competition. These relations are theoretically the most challenging. It is here that we should try to imagine what the relations might look like and which variables can influence their nature. The core of the party: the level of reference In the previous paragraph we were confronted with the fact that party and party system were not always easy to disentangle as the vertical relations between party units (thus within parties) can also be interpreted as relations between autonomous party actors. The fundamental question is: where exactly is the party? Is it at the federal level or is it at the regional level? There is obviously, as has already been pointed out, some empirical variation to be seen. In what follows I try to imagine a number of indicators that might allow us to identify the core of a party, its central headquarters, serving as a reference for the other sections. We should certainly not assume that the highest level (in terms of territory) is naturally the core level. I have already mentioned the European party federations as actually less important than the national party organizations, but other non-typical hierarchies, especially in very decentralized countries, can certainly occur. Though it again increases complexity, we should European Urban and Regional Studies 2003 10(3) Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016 DESCHOUWER: POLITICAL PARTIES not assume that there is one single and unique core level for a party. For the party system in general it is clear that there is no unique core level. First, this core level might be different, depending on the point one is observing from. If there is a region or a country with a very distinct nature, its core level might be the region or the country itself, while for the other substates of the multi-layered polity the central level is the most crucial point of reference. Second, the point of reference in a multi-layered system might not be the same for all the parties. A party focusing on only one region and not engaging in federal politics has a different point of reference than a party that is present and active in all the substates. Third, the core level might not be the same for the electoral, parliamentary and governmental arena. Election results There are many different ways and many different indicators that could be used to identify the core or the centre of a political party in multi-layered systems. The first is the interpretation of election results. The reading and interpretation of results gives an idea of the place where the political will of the people is being aggregated, and as such it tells something about the societal context in which parties act and in which they see and place themselves. In elections for the European Parliament, the electoral results are always presented per country and per national party (in Belgium it is done per language group, as for other elections). At the European level the votes are never aggregated. The main reason for that is that it is not clear which parties can be considered as belonging to one European core party. At the European level the results are aggregated in seats, only after the complex puzzle of putting national parties in European parliamentary groups has been completed. This aggregation furthermore always changes during the term, because of the constant reshuffling (especially on the conservative and Christian-Democratic side) of group memberships. What we have here is an indicator that can tell us whether the larger territory is a meaningful level of party competition. The European example is obvious, but Belgium is another very striking example. Election results, both for national and regional elections, are always aggregated at the level of the language communities. Yet Belgium is a very particular case, in the sense that parties operate at only one level: there is actually no Belgium-wide party system. Unilingual parties compete within their own community for the seats available in and for that community in the regional, national and European parliaments (Deschouwer, 1997). Another interesting example is Switzerland, where electoral competition also occurs at the cantonal level and between cantonal parties. These are, however, aggregated at the federal level in federal party organizations. Elections results have meaning at both levels, but the importance of the lower cantonal level is evident. The indicator only works for elections at the higher level. Then it can tell us whether that higher level is more meaningful than the regional one. For regional elections, of course, the results are read and interpreted at the regional level. Regional results, though, can have consequences for the higher level: they can be seen as a test for an upcoming national election, or even – as in Germany – affect the power balance in the second house of the parliament. Yet this is already taking us further in the direction of the interpretation of party strategies and of the meaning of the interaction between levels. Both aspects are, however, closely related, and depend on other variables. We will also discuss this below, but we can already note here that when regional (or local) elections are being held simultaneously (i.e. in all regions on the same day), the way in which the results are read can again be meaningful. If the aggregation matters, the lower level can be seen as a second-order level, and thus as a confirmation of the fact that the national party is the core. Party discipline A second possible indicator for identifying the core level is the degree of internal party discipline, or more concretely voting discipline, in the party groups (Hadley et al., 1989). We assume here that discipline will be higher at the core level, because that is the level where more is at stake. In doing so we are also measuring – as with the other indicators – at the same time a characteristic of both the European Urban and Regional Studies 2003 10(3) Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016 217 218 EUROPEAN URBAN AND REGIONAL STUDIES 10(3) political system in general and the political parties in particular. In the European Parliament party discipline is lower than at the national level, even if general patterns of voting behaviour by MEPs can be identified. The European level, because of its peculiar role and because of the peculiar (i.e. very loose) nature of the parties, does not come near the national level for the degree of party discipline. The Belgian case is again a bit awkward. Party discipline is high at both levels, indicating the nonhierarchical relationship between the levels that are separated but intimately linked. In Switzerland the level of party discipline is clearly higher at the level of the cantons. The federal parties lack the central organization to impose clear discipline on the behaviour of the MPs. Measuring party discipline actually also means looking at the level in the party where the power is, where actions of individual MPs can be steered and – most importantly – ultimately also sanctioned. The availability of these sanctions is thus a very similar indicator of the centrality of a level. One of these is the selection of candidates. Selection of candidates A third indicator of the centrality of a party level could thus be the place where candidates are selected for the next election. Again this is an indicator that can give an idea whether the higher level is indeed a more central one. If the higher level interferes in the candidate selection at the lower level, or if the lower level can dictate its will for higher-level elections, that level is clearly the core. The strength and validity of this indicator varies, unfortunately, with the electoral system. If the national level functions with national lists (which is exceptional), the indicator is pretty clear. But if national elections require the formation of lists at a lower level, the fact that the constituency level decides fairly autonomously does not necessarily mean a loss of centrality for the central party organization. The European example is again fairly clear: candidate selection does not occur at the European level, since the European parties themselves are not really participating in the elections. The centrality of the national level (in the Belgian case the language group level) is then obvious. Political careers Career patterns can offer an interesting insight into the hierarchies of a political system and of its party system. Indeed, the analysis of career patterns in multi-layered political systems is a new and developing research area (e.g. Stolz, 2001). The question that might be relevant for us is: which level is perceived as the highest prize in a political career? That would then be the core level. We assume that parties recruit the top personnel from the hierarchically lower levels to fill up the places at the higher level (which can of course be the region). Evidence from federal systems shows quite some variation in career and recruitment patterns. In Germany the federal chancellor often comes from below, while in Belgium there is so far no clear career pattern to be seen, but a lot of shifting between levels and in both directions (Deschouwer, 2000; Fiers, 2001). It does make the matter more complex, but there can be differences between parliamentary and governmental careers with respect to the perceived value of a position at one of the levels. Being a member of the European Parliament is not a very high prize, but a position in the European Commission is very high indeed, and can be put at least at the same level as membership of a national government. Money and staff A fifth possible indicator is the level where the financial resources and the staffs of a party are concentrated (Katz, 1999). Again this is intimately linked to the nature of the political system itself. Political parties in most countries are very dependent on state funding, and one could try to see where the most important funding flows into the party. The latter is not just a function of the system itself, since the parties themselves decide on how and where they are subsidized by the state. The decisions they take in this respect can thus be an indicator of what they perceive as the core level for their own functioning. European Urban and Regional Studies 2003 10(3) Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016 DESCHOUWER: POLITICAL PARTIES 219 SYSTEMIC FEATURES TYPE OF PARTY PARTY BEHAVIOUR Figure 2 The framework for analysis What is at stake? The framework for analysis This is the notion used by Reif (1980) to define second-order elections. European elections are second-order because there is no government to be chosen. Therefore the stakes are lower, the election is less important and gets meaning from the more important level for which it can be a ‘test’ election. This approach is, however, somewhat limited, and only applicable to systems in which elections – even at the national level – are meaningfully linked to government formation. That might be the case in majoritarian or in bipolar systems, but certainly not in multi-polar multi-party systems. There the election can be about which will be the biggest party, which then can choose its coalition partners. The direct election of the head of the executive – at the local, regional or national level – might also determine the stakes of the election. In sum, the idea that some elections are more important than others needs further elaboration. Heath et al. (1999) also suggest that second-order is a matter of degree: local elections are less second-order than European elections. Using these kinds of indicators should at least help us to identify our units of analysis in a more or less explicit and informed way. Once we know where the party actors are – the core level of a party family – we can move to the analysis of the interactions between this core level (to be seen as the level of reference) and the other levels. We did, however, see that when trying to identify the core of the party, several of the suggested measurements were directly affected by the characteristics of the political system itself. A framework for analysis of party behaviour in multi-layered systems should therefore take very explicitly into account that the system – the institutional environment – both shapes the party and its potential behaviour. Parties make choices in their competition with other parties in the same system. This includes party actors at different levels (vertical interactions) and party actors in other party systems at the same level (horizontal interactions). As a very general way of looking at these choices, we suggest the framework in Figure 2. This means actually that a party choice or strategy is a function of the interaction of two variables (or sets of variables). The first is the party type. Different kinds of parties have different goals and opt for different strategies. This can be analysed in a very broad sense (e.g. Müller and Strøm, 2000), but for our purpose the type of party refers rather to its position in the multi-layered system. That explains the double arrow between ‘type of party’ and ‘systemic features’. For the latter, we also suggest a limited and specific meaning. Not all aspects of the political system are relevant. Only those that are directly related to its multi-layered aspect can provide clues for both the way in which a party is positioned (on this see further below) and for the way in which the system itself offers or limits opportunities for action. A party typology for multi-level polities There are many typologies of parties available. There is no urgent need to develop yet another one. But for the analysis of parties in a multi-layered system, a number of characteristics might be important enough to highlight them explicitly. The typology or classification of party actors should allow for a differentiation in the kind of problems with which it can be confronted and in the ways in which it can or will deal with it. Party ideology is a first and important element to European Urban and Regional Studies 2003 10(3) Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016 220 EUROPEAN URBAN AND REGIONAL STUDIES 10(3) be looked at. This is rather obvious. Regional parties especially (see the further classification below) show some variation in the degree to which they claim fuller autonomy or even outright independence for their region. It goes without saying that their position affects the strategic choices that they will make, since their position affects also the attitudes of the other parties and therefore the kind of competition in the system, both at the regional and at the national level. Regionalism is not the only relevant ideology that can be looked at. We can see that in general the socialist parties in federal countries defend a higher degree of centralization of policy making (especially in social and economic matters), which also means that they will build more strongly integrated federal party organizations. This actually applies most clearly to the ‘classical’ federations and their Leftist parties; more recently, in the newer federal-type countries, the Leftist parties (Labour in the UK, the Left in France, the PSOE in Spain) seem to be more inclined actively to defend claims for decentralization and regional autonomy. This is, at the least, an element of variation that deserves to be explored in more detail. Next to the ideological dimension, the position of a party actor in the multi-layered system can be used as a way to classify the parties and to organize the thinking about their possible strategic choices. The following party types can then be defined: • A regional section of a party that has its core at a higher territorial level (e.g. the Land-level parties in Germany, the Labour Party in Scotland, the regional branches of the Spanish PSOE, PP, IU, etc.). • A regional section of a party that is itself the core. This ‘section’ is thus a party in itself (e.g. the cantonal parties in Switzerland; the nationalstate parties in the EU). • A regional party that has no organization at a higher territorial level and that: (a) also participates in elections and conceivably government formation at the higher level (e.g. the Catalan CiU, the Basque PNV, the CSU in Bavaria, the parties in Belgium); (b) is only active at this lower level. • A party at the national level that has only loose ties with the party organizations at the lower levels (e.g. the Swiss federal parties, the European party organizations). These parties are not the cores of their organization. They could also be labelled party federations, to make clear where exactly the party is to be situated. Such a party can be present in all or in some of the regional party systems. • A party at the national level that is a wellintegrated federation of the party sections at the lower levels (the Austrian federal parties, the German federal parties). It can be present in all or in only some of the regional party systems. Systemic features: federalism The shape of the system influences the positions of the parties and the kind of choices they can or want to make. The general idea is that a multi-layered system offers more and more complex possibilities for political parties to play their role (Lancaster, 1999). The way in which a multi-layered system differs from a single-layer system in this respect is, however, not that easy to pin down. One way of exploring the matter is by looking at what has been said so far about parties and party systems in federal or federal-type polities in general. Parties are engaged in the recruitment of political personnel, they participate in elections and seek governmental power, they mobilize voters, refer to communities and identities, defend or contest policies. If they govern, they engage in intergovernmental relations, both horizontally (between regions) and vertically (between region and federal level). Or to put it differently: political parties are one of the very important political actors that produce the linkages between the political institutions. Whereas in a unitary state the linkage function of parties can be mainly situated in the interaction between state and society (Lawson, 1980), their role in federal polities is more complex. They have to adapt to the institutions, while their actions and the differences between levels also affect the way in which the federal institutions can function. One would expect that this special role of parties in federal polities has been extensively described and analysed. Very surprisingly this is not the case. There is no conceptual language available to grasp the role of political parties in complex and multi-layered polities. European Urban and Regional Studies 2003 10(3) Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016 DESCHOUWER: POLITICAL PARTIES There are certainly a number of interesting and inspiring case-studies on Germany (e.g. Lehmbruch, 1976; Gabriel, 1989; Jeffery, 1999), Canada (Chandler, 1987; Dyck, 1996; Thorburn, 1996), Switzerland (Girod, 1964), Spain (Hamann, 1999) or Belgium (Deschouwer, 1997; 1999), but these are very much free-standing. There is some literature on voting behaviour and on party identification (e.g. Uslaner, 1989; Cotter and Stovall, 1992; Stewart and Clarke, 1998). Of course most analyses of parties and party systems in federal states do mention the federal logic and do tell something about the interaction between levels, but there is a striking absence of cross-references and thus no common language. In his handbook of the Swiss political system, Kriesi (1995) spends only six pages in the chapter on parties discussing the impact of federalism. In a recent account of party politics and territorial representation in Germany, Jeffery (1999) updates the work of Lehmbruch (1976). He produces a long, detailed, welldocumented and very interesting account of party politics in Germany and of the interaction between the federal state and the Länder, but it is merely descriptive, even introducing different concepts and models per historical period. This literature does, however, provide us with some clues in relation to some aspects or dimensions of federalism that need to be brought into the picture when analysing the role of parties. Regional autonomy, whether it is strong or weak, allows for differentiation. That is a very obvious starting point. Power is not centralized and can therefore be shared. The federal opposition parties have opportunities for regional access to power. This diffusion of power is conducive for the legitimacy of the system, since minority groups have less chance of being totally excluded (Lijphart, 1984; 1999; Chandler, 1987; Hodge, 1987; Gabriel, 1989). There are more access points and thus more possibilities for opposing the central government. Parties at the local level can ‘run against Washington’ (Katz, 1999). But the diffusion of power and of competencies in federal systems affects the representational role of parties, especially the governing parties; since responsibility can be diffused, the blame can be shifted (Gabriel, 1989; Tuschhoff, 1999). Since two different games are being played, one at the federal and one at the regional level, one can expect an internal differentiation of the political parties, a relatively loose link between the regional and the federal sections. That is, however, not a general rule. It depends on the degree of autonomy of the regions, on the type of autonomy (see below) and also on the degree of asymmetry in the federation. If the regions are very different from each other, the regional games are very different and the parties have to allow internally for this variation (Hodge, 1987; Hadley et al., 1989). If the federation is more homogeneous, then the development of large catch-all parties, able to integrate the regional differences, is more likely (Chandler, 1987). Though all these assumptions do sound very plausible, we have not found any comparative empirical data that might be able to support them. In extreme cases of asymmetry, when in some regions one party clearly dominates the game, the federal logic gives that party a double role. It becomes then at the same time the regional governing party, linked to the regional electorate, and the party than can speak for the region in the horizontal and vertical intergovernmental relationships (Wolinetz, 1999). It is both agent and principal. The best example of this is probably the Party Québécois, but the Catalan CiU is also clearly playing this double role. Another good example is the Bavarian CSU. All this assumes that we are indeed looking at a federal system in which the governments and/or the levels are closely interconnected. If they are not, the regions and their governing parties do not have to engage in intergovernmental relations, and only play the role of principal to a very limited extent. Much depends on the nature of the federal structures. If the levels are connected, if competencies overlap to a certain extent, either jurisdictionally or functionally, the institutions become extremely relevant for the parties. Federal and intergovernmental politics are then party politics and vice versa (Lehmbruch, 1976; Chandler, 1987; Rydon, 1988; Jeffery, 1999). That is even more the case if the regional level is directly incorporated into federal decision making, as in the German federation (Chandler, 1987; Wolinetz, 1999). ‘It creates the possibility of federal-provincial relations being defined in partisan terms and provides opposition forces with an incentive for using regional arenas as a means of challenging the legitimacy of an existing federal majority party or coalition’ (Chandler, 1987: 155). If the two levels are closely connected, and European Urban and Regional Studies 2003 10(3) Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016 221 222 EUROPEAN URBAN AND REGIONAL STUDIES 10(3) especially if the regional level is present and necessary for federal decision making, the regional level tends to lose its political autonomy. Regional elections then become relevant for federal policy making, and will also be framed in these terms. Regional elections gain a meaning with reference to the higher level. They become federal mid-term elections (Gabriel, 1989; Jeffery ,1999; Wolinetz, 1999). With reference to Australia, Rydon (1988) makes the interesting point that the proliferation of elections with relevance for the federal level leads to an increasing de-ideologization. Unpopular decisions constantly get postponed. This brief overview of the literature on parties in federal systems (at rather high altitude and not doing full justice to it) certainly reveals that we are confronted with a very complex and multidimensional political system. That is nothing to be surprised at. But it does show clearly that it is not very useful to try to find clear and predictable patterns about the role and function of parties in federal polities in general. Much depends on the type of party one looks at (that is why we have this aspect explicitly in our conceptual scheme), and on the type of federation one looks at. This is where multi-dimensionality enters the picture. Federal systems can vary along many different lines, and some of these are extremely relevant for the way in which the parties can play their role. From the brief overview above, four very relevant variables can be distilled: 3. The degree of asymmetry. If the competencies given to one region differ from those given to another one, this affects the party-political game. It means that region-based parties are confronted with a different opportunity structure, depending on where (territorially speaking) in the system they are located. It means that regional sections of state-wide parties are playing in different institutional contexts depending on their position. It also means that the strategies of the state-wide or national parties vis-a-vis their regional sections become different for different regions. 4. The homogeneity of society. That can actually be seen as an aspect of asymmetry. It refers to differences in identity between different territorial units of the system. While the unequal distribution of competencies can be labelled ‘institutional asymmetry’, the latter can be labelled ‘societal’ (see also Watts, 1999). The consequences for the parties and their strategic choices are the same as those listed in the previous point: strategies have to be differentiated. In this case of societal asymmetry the differentiation will probably be even higher, since societal asymmetry will produce party-political asymmetry: regionspecific parties and party systems and thus a region-specific pattern of competition. 1. The interconnectedness of the two (or more) levels and the way in which the lower level is incorporated in the higher level. That tells us a lot about the degree to which the things that happen at one level or in one region can affect the political life of the others. This is very relevant for parties, because their choices then have an impact on the political life of more than one level at a time, and the kind of choices they make and the way in which they come about are then extremely relevant aspects for analysis. 2. The degree of autonomy. That seems to be the important indicator of the degree to which the parties at the different levels can and will play a different role. If the regions are very different from each other, because their autonomy allows them to make different political choices, regional games are very different and the parties have to allow internally for this variation. Systemic features: electoral rules and laws Next to aspects that are directly linked to the regionalized state institutions, the multi-layered aspect of a polity also accounts for variations in other very relevant characteristics that can vary between levels and that can offer or take away opportunities for party action, in particular electoral rules and electoral laws. Elections at different levels and in different regions or countries are not necessarily contested under the same conditions. The right to vote In the first place the composition of the electorate can be different, i.e. the rules on the right to vote can be different for elections at different levels. That might at first sight be something that is more European Urban and Regional Studies 2003 10(3) Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016 DESCHOUWER: POLITICAL PARTIES appropriate for historical analysis, since today all citizens are enfranchised at all levels. Yet variations in the definition of citizenship can still be relevant. It is clear that this type of differentiation between two levels is not likely to appear in regional states. But for Europe it can certainly be relevant. In local elections and for elections to the European Parliament citizens can vote in the country or in the local municipality where they reside. In countries with a very high proportion of European foreigners (like Luxembourg), this can really affect the results. In Brussels and in the area around Brussels, where a high concentration of Europeans live in an area where linguistic balances are extremely delicate, the vote of these Europeans has definitely influenced the strategies of both the Francophone and the Dutch-speaking Belgian parties. Electoral rules and formulas If the electoral formula should differ, some parties might not have the same chances to elect candidates in regional or in national elections. They might even not run in elections in which they are not likely to reach the threshold. Different district magnitudes can have this effect, as has been demonstrated by Lutz (1998) for Switzerland. The European Parliament is an interesting place to see MEPs from different countries playing their roles differently, because they are elected in different ways. There is already some convincing evidence on the way in which electoral rules affect the role orientations of members of parliaments (Bowler and Farrell, 1993; Hamann, 1999; Katz, 1999; Wessels, 1999). It is clear that when one level uses a proportional formula and the other a majoritarian one, parties have to make different choices, make different alliances. The good scores for Eurosceptic parties in France in PR-based European elections (and their decision to participate in the first place) contrasts with their poor results or even absence at the polls in the majoritarian elections for the French Assemblée. If the logic of preference voting differs, the choice of candidates can be different. If a head of executive can be elected directly at one level or the other, this would also affect the choice of candidates and the organization of party-political careers in general. Timing of elections: cycles and simultaneity The literature on European elections has shown very nicely how the way in which the European and the national elections relate to each other is a matter of timing. The elections can coincide; they can come just before or just after one another, or at mid-term. Electoral cycles seem thus to be an important indicator of the way in which the levels interact (Reif, 1984; Van der Eijk et al., 1996; Lohman et al., 1997). This is something that has also been found for the German federal and Land elections – see Jeffery (1999) for a good summary – and for Australian state and federal elections (Rydon, 1988). Not only the interconnection of electoral cycles is important. Simultaneity of elections also plays a role. Two sorts of simultaneity can be distinguished: horizontal or vertical. Horizontal simultaneity means that elections for all the regions take place on the same day. If this happens, their ‘nationalization’ is very likely. A very striking example of this was the regional elections in Italy in 2000. Not all the regions went to the polls (the special statute regions and the islands did not have elections), but the fact that a very large number and in fact the most important regions were all organizing elections allowed the opposition leader Berlusconi to turn the regional elections into a national test (i.e. into second-order national elections). The opposition won most of the regions, and as a result of that Prime Minister D’Alema resigned, as if the electoral defeat was a defeat in national elections. Vertical simultaneity means that elections at different levels coincide, for instance national and European elections. The first-order election is then likely to ‘absorb’ the second-order election in terms of stakes and campaigning. This is something that is very visible (so far) in Belgium. Regional parliaments have been elected directly twice so far, always on the same day as the national elections. Combined with the fact that all Belgian parties are regional parties, and that exactly the same party actors compete in the elections at both levels, the two campaigns, voting behaviour, the results and then coalition formation made the two levels actually collapse into one (Versmessen, 1995; Deschouwer, 2000). European Urban and Regional Studies 2003 10(3) Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016 223 224 EUROPEAN URBAN AND REGIONAL STUDIES 10(3) Conclusion Developments in the institutional environments of political parties affect the way in which they are organized and behave. This proposition lies at the heart of this article. Both regionalization and European integration – often interacting with each other – make the institutional environment more complex. We focused on one crucial aspect of this complexity: the multi-layered nature of the political systems resulting from the processes of regionalization and European integration. The consequences for party theory are farreaching. Concepts for describing and analysing parties have indeed to a large extent been developed for the study of parties in single-level systems. Parties in federal systems have of course received attention, but – also as a consequence of some variation in the way in which different federal systems work – this has not led to a set of analytical tools that allows us to analyse in a systematic and comparative way the behaviour of political parties in multi-layered systems in general. We have discussed some problems of definition. In particular, the search for the most appropriate unit of analysis has proved difficult, because the interaction between levels tends to blur the analytic difference between party units and party systems. We have suggested – relying on existing empirical evidence – a number of possible indicators to identify the core level of a party in a political system. This core level depends to a large extent on the nature of the political system itself. That is why we have suggested a general framework of analysis in which the characteristics of the parties and the characteristics of the political system interact with each other. At the level of the political system, we have relied on the literature on federalism to identify relevant dimensions of variation that can have effects on political parties. We have also pointed at the importance of variation in electoral rules between levels. Not only the electoral formula but also the timing of elections at different levels seem to be crucial elements of the environment in which political parties function. The major aim of this article was to show that we do indeed need to reflect on the way in which we can analyse political parties in these complex systems. Just as for parties in federal systems, the literature so far has been very much case-oriented (one party, one country, one region) or has focused on the electoral dimension only. The notion of second-order elections offers food for thought and invites a more comparative exploration of differentiated electoral behaviour. But party behaviour itself also deserves a more comparative approach. That is why we have tried to present this framework for a comparative analysis of political parties in multi-layered systems. Note 1 The term ‘regional state’ refers to a national state in which a number of regional subunits have received farreaching political autonomy. With this broader notion of ‘regional state’ we cover both the classical federations like Germany, the regionalized states like Italy or Spain and the states that have produced fairly complex and often non-symmetrical multi-layered political institutions, like Belgium or the UK. References Bartolini, S. (1998) ‘Exit Options, Boundary Building, Political Structuring. Sketches of a Theory of Largescale Territorial and Membership “Retrenchment/Differentiation” Versus “Expansion/Integration” (with Reference to the European Union)’, EUI, SPS Working Paper No. 98/1. Florence. Bowler, S. and Farrell, D. (1993) ‘Legislator Shirking and Voter Monitoring: Impacts of European Parliament Electoral Systems upon Legislator–Voter Relationships’, Journal of Common Market Studies 31 (1): 45–69. Chandler, W.M. (1987) ‘Federalism and Political Parties’, in W.M. Chandler and H. Bakvis (eds) Federalism and the Role of the State, pp. 149–70. Toronto: University of Toronto Press. Cotter, P.R. and Stovall, J.G. (1992) ‘Party Identification and Level of Government: the Validity of Mixed Party Identification’, Midsouth Journal of Political Science 13: 515–34. Deschouwer, K. (1997) ‘Une fédération sans fédérations de partis’, in S. Jaumain (ed.) La réforme de l’Etat ... et après? L’impact des débats constitutionnels en Belgique et au Canada, pp. 77–83. Brussels: Editions de l’ULB. Deschouwer, K. (1999) ‘In der Falle gefangen. Belgiens Parteien und ihre Reaktionen auf abnehmende European Urban and Regional Studies 2003 10(3) Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016 DESCHOUWER: POLITICAL PARTIES Wählerloyalitäten’, in P. Mair, W. Müller, and F. Plasser (eds) Parteien auf komplexen Wählermarkten. Reaktionsstrategien politischer Parteien in Westeuropa, pp. 281–314. Vienna: ZAP. Deschouwer, K. (2000) ‘Belgium’s Quasi-regional Elections of June 1999’, Regional and Federal Studies 10 (1): 125–32. Dyck, R. (1996) ‘Relations Between Federal and Provincial Parties’, in B. Tanguay and A.G. Gagnon (eds) Canadian Parties in Transition, pp. 160–89. Toronto: Nelson. Fiers, S. (2001) ‘Carrièrepatronen van Belgische parlementsleden in een multi-level omgeving’, Res Publica 43 (1): 171–92. Flora, P., Kuhnle, S. (eds) (1999) State Formation, Nationbuilding and Mass Politics in Europe. The Theory of Stein Rokkan. Oxford: Oxford University Press. Gabriel, O.W. (1989) ‘Federalism and Party Democracy in West Germany’, Publius 19 (4): 65–80. Girod, R. (1964) ‘Geography and the Swiss Party System’, in E. Allardt and Y. Littunen (eds) Cleavages, Ideologies and Party Systems. Helsinki: Westermarck Society, Vol. 10. Guyomarch, A. (1995) ‘The European Dynamics of Evolving Party Competition in France’, Parliamentary Affairs 48 (1): 100–24. Hadley, C., Morass, M. (1989) ‘Federalism and Partyinteraction in West Germany, Switzerland and Austria’, Publius 19 (4): 81–97. Hamann, K. (1999) ‘Federalist Institutions, Voting Behaviour, and Party Systems in Spain’, Publius 29 (1): 111–37. Heath, A., McLean, I. et al. (1999) ‘Between First and Second Order: a Comparison of Voting Behaviour in European and Local Elections in Britain’, European Journal of Political Research 35 (3): 389–414. Hix, S. (1999) The Political System of the European Union. London: Macmillan. Hodge, C.C. (1987) ‘The Supremacy of Politics: Federalism and Parties in Western Europe’, West European Politics 10 (2): 253–68. Hooghe, L. (1996) Cohesion Policy and European Integration. Building Multi-level Governance. New York: Oxford University Press. Jeffery, C. (1999) ‘Party Politics and Territorial Representation in the Federal Republic of Germany’, in J. Brzinski, T. Lancaster (eds) Compounded Representation in Western European Federations, pp. 130–66. London: Frank Cass. Katz, R.S. (1999) ‘Parties in Europe and Parties of Europe’, paper presented at the conference ‘Multi-level Party Systems: Europeanisation and the Reshaping of National Political Representation’ (Dec.), Florence. Kriesi, H-P. (1995) Le système politique Suisse. Paris: Economica. Ladrech, R. (1994) ‘Europeanization of Domestic Politics and Institutions: the Case of France’, Journal of Common Market Studies 32 (1): 69–88. Lancaster, T.D. (1999) ‘Complex Self-identification and Compounded Representation in Federal Systems’, in J. Brzinski, T. Lancaster (eds) Compounded Representation in Western European Federations, pp. 59–89. London: Frank Cass. Lawson, K. (ed.) (1980) Political Parties and Linkages: a Comparative Perspective. New Haven and London: Yale University Press. Lehmbruch, G. (1976) Parteienwettbewerb im Bundesstaat. Stuttgart: Kohlhammer. Lijphart, A. (1984) Democracies: Patterns of Majoritarian and Consensus Government in Twenty-one Countries. New Haven, CT: Yale University Press. Lijphart, A. (1999) Patterns of Democracy. Government Forms and Performance in Thirty-six Countries. New Haven, CT: Yale University Press. Lipset, S.M. and Rokkan, S. (eds) (1967) Party Systems and Voter Alignments. New York: The Free Press. Lohmann, S., Brady, D. and Rivers, D. (1997) ‘Party Identification, Retrospective Voting, and Moderating Elections in a Federal System: West-Germany, 1961–1989’, Comparative Political Studies 30 (4): 420–49. Lutz, G. (1998) ‘The Interaction of National and Regional Electoral Systems and Their Influence on Representation in Switzerland’, paper presented at the ECPR Joint Sessions of Warwick. Mair, P. (2000) ‘The Limited Impact of Europe on National Party Systems’, West European Politics 23 (4): 27–51. Marks, G. and Hooghe, L. (2001) Multi-level Governance and European Integration Oxford: Rowman and Littlefield. Müller, W. and Strøm, K. (2000) Policy, Office, or Votes? How Political Parties in Western Europe Make Hard Decisions. Cambridge: Cambridge University Press. Reif, K-H. (1980) ‘Nine Second Order National Elections. A Conceptual Framework for the Analysis of European Election Results’, European Journal of Political Research 8 (1): 3–44. Reif, K-H. (1984) ‘National Electoral Cycles and European Elections’, Electoral Studies 3 (3): 244–55. Rydon, J. (1988) ‘The Federal Structure of Australian Political Parties’, Publius 18 (1): 159–71. Stewart, M.C. and Clarke, H.D. (1998) ‘The Dynamics of Party Identification in Federal Systems: the Canadian Case’, American Journal of Political Science 42 (1): 97–116. Stolz, K. (2001) ‘The Political Class and Regional Institution-building: a Conceptual Framework’, Regional and Federal Studies 11 (1): 80–100. Thorburn, H.G. (1996) ‘The Development of National Political Parties in Canada’, in H. Thorburn and G. European Urban and Regional Studies 2003 10(3) Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016 225 226 EUROPEAN URBAN AND REGIONAL STUDIES 10(3) Scarborough (eds) Party Politics in Canada, pp. 5–12. Englewood Cliffs, NJ: Prentice-Hall. Tuschhoff, C. (1999) ‘The Compounding Effect: the Effect of Federalism on the Concept of Representation’, in J. Brzinski, T. Lancaster (eds) Compounded Representation in Western European Federations, pp. 16–33. London: Frank Cass. Uslaner, E.M. (1989) ‘Multiple Party Identifiers in Canada: Participation and Effect’, Journal of Politics 51: 993–1003. Van der Eijk, C., Franklin, M.N. (1996) ‘What Voters Teach Us About Europe-wide Elections: What Europewide Elections Teach Us About Voters’, Electoral Studies 15 (2): 149–56. Versmessen, E. (1995) ‘In the Kingdom of Paradoxes: the Belgian Regional and National Elections of May 1995’, Regional and Federal Studies 5 (2): 239–46. Watts, R. (1999) Comparing Federal Systems. Montreal: McGill-Queens University Press. Wessels, B. (1999) ‘Whom to Represent? Role Orientations of Legislators in Europe’, in H. Schmitt and J. Thomassen (eds) Political Representation and Legitimacy in the European Union, pp. 209–34. Oxford: Oxford University Press. Wolinetz, S. (1999) ‘Parties in Multilevel Systems of Governance: a Working Paper on Parties and Party Systems in the European Union and Other Multilevel Systems of Governance and the Ways in Which They Might Develop’, paper presented at the European Community Studies Association Biennial Convention, Pittsburgh, Pennsylvania. Correspondence to: Professor Kris Deschouwer, Vrije Universiteit Brussel, Vakgroep Politieke Wetenschappen, Pleinlaan 2, 1050 Brussels, Belgium. [email: [email protected]] European Urban and Regional Studies 2003 10(3) Downloaded from eur.sagepub.com at PENNSYLVANIA STATE UNIV on May 18, 2016
© Copyright 2026 Paperzz