Governance Capacity and Regionalist Dynamics

Abstract The debate on the effects of regionalism and European integration on European nation states has been prominent for more than a decade. Regionalization of EU states has not brought with it genuine regional autonomy and regionalism has not emerged as a bottom-up public demand in European regions. It is contended here that to determine the future of regional devolution, whether as a result of bottom-up or top-down processes, the factors at play must be contextualized. This paper examines some determinants of regional political capacity, as identified in the policy literature, in tandem with a number of determinants of economic prospects and the existence of an economic milieu. This is done in a comparative context across 12 regions of the EU. It is suggested that the potential for regionalist pressures to emerge is dependent on regional governance capacity and the relative economic weight of a region.

legitimacy and effectiveness; [and] c. new processes, arrangements and methods" (Kooiman, 2002: 73). Regional governance is employed here to belong within the first category broadly comprising: rules and qualities of formal and informal regional political institutions as they relate to civic society and interact with other institutions and actors internal or external to the region. 2 In the 1980s a number of regional political actors started to mobilize when their respective regions faced the threat of economic marginalization by the core conurbation in Europe (the so-called London -Milan axis) and political marginalization by further European integration. They formed a number of transnational groupings, such as the Council of European Municipalities and Regions, the Assembly of European Regions and the Commission of Maritime Peripheral Regions. This mobilization has not been aimed at territorial autonomy but primarily at economic development as well as cultural and linguistic preservation. The reasons regional institutions mobilize therefore tend to be functional. Obviously, there are a number of regions endowed with political competence (autonomy of institutions), capacity (ability to devise and implement policies) and legitimacy (within and outwith the region), the stronger of which seek what Keating (2005: 134) has termed a "distinct role". Particularly those regions belonging to the association of Constitutional Regions and the Regions with Legislative Power. It does appear as a legitimate expectation for legislative regions to have an authoritative voice within EU governance. They are, after all, intricately entwined in the governance structure of their respective nation states.
Regional actors have benefited from EU initiatives promoting transregional 3 collaboration, evidence of which abounds (Cochrane, 1994;Morata, 1997). 4 Within the EU context such partnerships are formed between either metropolitan cities or regional governments. Within that framework encouragement by the institutionalization of a Committee of the Regions (CoR) has signalled regional political institutions to mobilize. The CoR has promoted awareness of EU institutional structures among regions with restricted regional autonomy but falls far short of the strong political platform anticipated by regions in federal states such as the German Lander. 5 Overall, European institutions can hardly be perceived as 'liberators of regions'. They, at best, anticipate a need to fill the vacuum left by the weakening of nation-state capacity for policy implementation and assessment. At the same time the Commission is unlikely to venture into institutional arrangements that may directly challenge member states.
To explore whether regionalism has been caused directly by a presumed Europeanization of policies and institutions within EU member states it is essential to create a picture of the interaction between institutions at all levels of EU governance-a task too ambitious for the present context. Interaction of sub-state with national and EU institutions within this complex "variable geometry" (Keating, 1992) invariably increases interdependence among these governance structures. Interdependence diversifies what traditionally would have been a domestic agenda and erodes any divisions between foreign and domestic policy that might have pre-existed (Keohane and Nye, 1988).
It is important at this juncture to distinguish the types of regionalist pressure apparent in contemporary Europe. Keating (1998: 14 -15) has broadly identified six ideal types: (a) conservative regionalism rooted on concepts of affective community; 364 D. C. Christopoulos (b) bourgeois regionalism based on pressure to the state from middle class regional elites; (c) technocratic regionalism based on a depoliticized view of economic development; (d) progressive regionalism that has attempted historically to counteract the effects of internal colonialism on nation states; (e) populist regionalism directed against fiscal subsidization from wealthy to poorer regions; (f) separatist regionalism prevalent within historic nations in Europe.
What Keating (1998: 15) described as "nation-states crises of territorial representation" is often behind most attempts by regional actors to take advantage of the relative political vacuum and exploit popular discontent.
Most accounts of regionalism focus on the effects of European integration-often as an assumed Europeanization of domestic policies-and the resonance of historic cultural identities that underpin regionalist demands. For expediency within the present exposition, regionalism is taken to signify a predominantly bottom-up pressure to devolve power from the centre and regionalization to be a state (topdown) attempt at decentralization. The qualitative difference between decentralization and regionalization lies in the role of governance structures in determining such a process. It is partly because non-state actors are integral within governance structures that regionalization becomes a systemic process and its dynamics can therefore be determined functionally.
Unfortunately, Europeanization can have a great number of conflicting definitional frames. It can be seen as a process or situation. Concurrently it can be shaped by both structural factors and agency. As a process it can be determined by incremental change or punctuated equilibria; it can be responsive, intended or unintended; coercive, mimetic or a result of negative integration. It can also be seen as governance transfer or attempt at misfit adaptation (Borzel and Risse, 2003). Keating (2005: 220) expressed a broad consensus in his analysis of Scottish policy making that "the trend right across Europe is for states to decentralise and Europeanise". Featherstone (2003) produced a comprehensive review of the literature where he identified institutional settings and policy processes as the two main formulations in which the term is employed. Featherstone and Radaelli (2003) suggested that the research agenda in the field should initially consider how EU integration alters the political equilibrium in member states. They summarized evidence from an edited volume that a wide range of actors is relevant to the process. These include among others: institutions as collective actors; national government administrations; subnational, regional and local authorities; interest associations; and a number of individual actors within the aforementioned associations. It appears that most types of actor can be relevant and most types of governance organization involved. This presents a number of specification problems. But, more importantly, for a term that describes a process undisputed by most specialists, there appears to be very little scope for operationalization. Radaelli (2003) and Borzel (2002), suggested that this process leads EU states to a clustered convergence. Although clustering has been employed in the context of nation-states, it is contended here that 'clustered' outcomes can be observed Governance Capacity and Regionalist Dynamics 365 among regions of different EU states. In that respect one can view regionalization as an attempt to conform to EU policy process requirements and therefore induced by a very broadly defined process of Europeanization. 6 There are a number of research questions that provide the background to the present research exercise. They concern institutional arrangements and policies of nation-states towards sub-state political representation in the EU. The resultant institutional architecture and relative political autonomy is assumed to be a response to challenges of EU integration, global competitiveness and administrative modernization. A number of assumptions underpin this research. Regionalization of larger EU states is the outcome of an attempt to conform to policy process requirements and broadly defined Europeanization pressures. The capability for bottom-up regionalist pressures depends on the governance capacity of regions. 7 This work has to be seen as an exploratory analysis of the consensus within the policy community (European Commission, 2001a;AER, 2003;CoR, 2004) juxtaposed to a number of economic indicators. This will allow critical examination of whether institutional autonomy is related to a region's economic prospects and relative economic weight.
The article proceeds by describing the institutional arrangements of sub-state political institutions in the three states selected, with particular reference to the specific regions selected from each state. A list of economic determinants of political power is then constructed and compared with the relative ranking of political institutions. In conclusion some insights are offered on the research questions that have guided this research exercise.

Regionalism vs. Regionalization in Spain, Poland and the UK
The choice of states in this analysis is informed by an attempt to locate cases where devolution has been contested and where the state has adjusted by introducing variable levels of regional devolution. For cases selected, the generic processes that affected all large nation-states are present but the sub-state bottom-up pressures vary.
Pressures for administrative re-organization (whether based on attempts at increasing efficiency, democracy, legitimacy or a combination of these) have been channelled by institutional structures and filtered through dominant policy paradigms. The case has been made more convincingly for larger states, where variations in political culture amongst peoples and regional elites, as well as wide differentials in economic development, have been behind many regionalist (bottom-up) demands for autonomy. As claimed earlier, these demands have often been identified with European integration and administrative devolution. Of the six larger EU states (Italy, Germany, France Spain, Poland, UK), three are selected for reasons of research expediency.
States excluded did not fit the research design criteria for a variety of reasons. A comparatively stable federal institutional structure and uniform regional autonomy is why Germany has been excluded. France has been excluded as a Unitary state that exercised administrative devolution (in the 1980s) lacking any strong regionalist (bottom-up) pressures to do so. Italy, like Germany, was endowed (post-WWII) with the makings of a Federal Constitution, but only started allocating capacity to regional institutions as an outcome of administrative rationalization in the 1970s. Regionalist pressures since the 1980s have led to an asymmetric devolution settlement that falls within the scope of the present study. Unfortunately, the data that could be collected on Italian political devolution were deemed unreliable for the sources employed here.
There is a clear case for including Spain and the UK as there is a long-standing process of administrative devolution from above, as well as regionalist/sub-state nationalist pressures from below, that are responsible for the current institutional settlements. Moreover, in both cases there are asymmetric arrangements of regional devolution that have led to debates on the extent, conflict and range of powers of political elites and regional institutions, exemplified in the so-called, West Lothian question in the UK. In the case of Poland, the state was led to an administrative devolution soon after democratization without any substantial bottom-up agitation or pressure. The significant variations in economic development and strength of identity amongst different regions in Poland provide the context for a potentially volatile devolution arrangement. From within each of the states selected, four regions were identified that would represent the range of strength of identity, political autonomy and economic significance for the respective nation-states. 8 As could be expected, the recent devolution process has followed different 'evolutionary' paths in the three states examined. In the UK a series of regional referenda has led to the introduction of legislation in 1998 to devolve; in Spain variable autonomy arrangements are confirmed in its constitution since 1983, although they remain contested; and, in Poland, regional elected assemblies were created in 1998 prior to the accession of the country to the European Union.

Poland
Poland has to deal with pent-up demands for autonomy inherited from the complex history of the modern state. Polish state elites have to contain bottom-up regionalist demands from Slaskie and may eventually have to deal with economic devolution demands from the faster-growing economic regions, such as Pomorskie and Wielkopolskie. A substantial reform of the territorial organization of Poland implemented in 1998 has opened a new chapter in the history of regional devolution in Poland (see Kuklinski and Olejniczak, 2002). Three elements of this reform are relevant in the present context: (a) the changed scale of the region as the pre-existing 49 voivodships (provinces) were replaced by 16 much larger units. Larger voivodships are better equipped to foster regional identity that can be linked to the salience of historical traditions in these regions; (b) the new 16 territorial units were recognized not only as a subdivision of the administrative power of the central government but also as a unit of regional self-government with a regional Sejmik (regional assembly) elected directly by the population of the province and electing in turn a 'Marshal' as leader; (c) the Voivod (governor), an official appointed by central government and representing central government to the region.
The prospect of Polish participation in the EU (with the relevant share in Cohesion Policy funding) accentuated an impetus for reform. What could be seen as a

Governance Capacity and Regionalist Dynamics 367
Europeanization effect has facilitated the reform of public administration and the emergence of autonomous regional structures. One of the stated objectives of the 1998 reform was to elevate regions into legitimate interlocutors within the enlarged European Union. 9 The role some regional administrations play in policy implementation within the EU and, most importantly, their role in fund absorption provided a compelling motive for the administrative reform of the Polish regions. The economic development gap with the rest of Europe and the extreme inter-regional disparities within Poland could create the background for more substantial devolution demands. This has the potential to affect both the formation and the implementation of policy as implied in a number of EU Commission accession reports. 10 A comprehensive analysis of the political and historical background to regional identities in a number of Polish regions was provided by Zarycki (2005). He identified the regional dimension of the transformation of the Polish state to entail a cultural and an institutional component. The former mainly involved the transformation of the logic of a centralized to a decentralized state, while the latter a reform from centrally controlled local units to regionally elected administrations. The present debate on the reform of the regional administration of Poland dates back to at least 1989, long before the prospect of accession to the EU could have played any role. Of the variant of models proposed in the decade of that debate, the number of regions ranged from 6 to 17. Some political actors even argued for retaining the 49 administrative regions. 11 From the regions chosen for this study, Upper Silesia (regional capital Katowice) has "a very clear regional identity, understood both in terms of self-identification of the inhabitants as well as in terms of the external image of the region" while it is "the only region with a clearly accentuated movement demanding autonomy for the region" (Zarycki, 2005: 14). The region attained 51% of EU average per capita measured in PPS and is the region with the second largest GDP in Poland as a percentage of the national economy (14%). 12 Mazurskie (regional capital Olsztyn) is a "relatively 'weak' voivodship . . . an excellent example of the new regional identity on the basis of the mixed heritage of settlers from different regions and of different national identities" (Zarycki, 2005: 14). The region has the lowest per capita income from all those examined here (34% of EU). Pomorskie (regional capital Gdansk) is a region with above average economic potential (46% of EU GDP in PPS), while Mazowieckie (regional capital Warsaw) can be seen as the political and economic heartland of Poland, a region dominated by a prosperous capital city juxtaposed with an underdeveloped periphery. 13 The domination of this region is evident in its per capita income (70% of EU), while it contributes 20% of the national GDP, furthermore attracting 23% of business and 49% of government research and development (R&D) investment. Lack of regional statistics does not allow for a discussion of the Polish regional economic milieu in more detail.

Spain
The historical rights of the Galicians, Basques 14 and Catalans date back to the formation of the Spanish State in the fifteenth century. Under the Bourbons in the eighteenth century there was an attempt at curbing those rights which, however, were eventually recognized in the 1930s by the Republican Government (Heywood, 1995). The Franco regime that prevailed after the civil war maintained a concerted attempt at eradicating regional identities or any regionalist demands. Regionalism and sub-state nationalist movements were, to some degree, associated with resistance to centralism and the Franco regime.
The 1978 Spanish Constitution is attempting a tightrope between a regionalist and federalist model. The Constitution pronounces an Estado de las Autonomias while at the same time the predominant logic among national political elites is one of an Estado Unitario. Democratic consolidation allowed national elites to reassert their confidence in the nation-state by endeavouring to increase their authority within the political system. The introduction in 1981 of the Ley Organica de Armonizacion del Proceso Autonomico can be seen within this context as "an attempt on the part of the state to recover its powers" (Keating, 1998: 16). However, in 1983, the Constitutional Court adjudicated that large segments of the new law were unconstitutional. This led to a speeding-up of the process of approval for autonomy statutes with regional elections held in May 1983. 15 The Constitution provides for national legislative power to be shared by two chambers-the Congreso de los Diputados and the Senado. The latter of these chambers "provides a measure of territorial representation" (Morata, 1998: 101) at the nation-state level. These two chambers jointly (Cortes) provide a fulcrum for the Spanish political elite and, in some respect, the counter-point to political power exercised by regional assemblies. The constitutional arrangement, between the state and autonomous communities, has invoked references to "variable geometry state-order" as exemplified in the seminal work by Keating (1992). Williams (1994) suggested that initially there were broadly three levels of relative autonomy from state institutions corresponding to: (a) the high level of autonomy of the historic nationalities of Catalonia, the Basque Country and Galicia; (b) the intermediate level of autonomy of regions with exceptional claims due to distinctive linguistic, cultural or political institutions, namely Andalusia, Valencia, Navarre and the Canary Islands; (c) the low level of relative autonomy enjoyed by the ten remaining regions of the Spanish state.
In the contemporary context Morata (1998) pointed to a dichotomy between a two-tier system. The seven regions with full autonomy responsible for public health, social security, local administration and police affairs are now responsible for disposing over a quarter of total public expenditure and over a third of total public investment. While the other ten regions are responsible for a variety of policy areas, such as education, regional public works, regional development, agriculture and tourism (Morata, 1998: 101).
The position of regions, although constitutionally guaranteed, was not integral to policy making towards the EU. In 1992 the Constitutional Court pronounced that regions have the constitutional right to be present in negotiations at the EU level "provided that such activity does not interfere with the defense of national interests" (Morata, 1998: 115 fn4). Main beneficiaries of these new governance arrangements Governance Capacity and Regionalist Dynamics 369 have been the nationalist parties of Catalonia, the Basque Country and the Canary Islands. Apart from historical rights, a basic differentiation of Spanish regions relates to differences in cultural and linguistic background. Euskera the language spoken in the Basque Country has no Indo-European roots. Catalan and Galician are also distinct languages, while dialects are spoken in Andalusia, Asturias and the Canary Islands. Differences in the rate of industrialization and development in the nineteenth century has also accentuated these variations. The Basques and Catalans attained the highest economic development, while national elites were predominantly Castillian and Andalusian. Genieys (1998) suggested that elites in less autonomous regions proceed to emulate institutionally those elites in regions of higher autonomy. He provides evidence that this has happened to the Andalusian regional elites who have emulated their Catalan counterparts. Galicia (regional capital Santiago de Compostela) and Catalunya (regional capital Barcelona) are regions with historic nationalities and entrenched autonomous traditions. Variation between the two is predominantly economic as the latter is one of the main motors of the Spanish economy (112% of EU), while Galicia is comparatively underperforming (75% of EU). Catalunya has a strong economic milieu, with 1.9 the times the average patents in high tech compared with the rest of Spain. 16 Evidence of innovation exists in all high tech sectors with obvious concentration in semiconductors (four times the national average) and communication technologies (1.1 times the national average). Andalucia (regional capital Seville) is a region of major disparities in economic development (71% of EU), having a comparative high weight in the national economy (14% of national GDP) with some evidence of the existence of an economic milieu in the aviation (2.0 times the national average of patents) and the communications industry (1.3 times the national average of patents). Aragon (regional capital Zaragosa) is a region with high per capita GDP (102% of EU) but no evidence of concentration of economic innovation.

United Kingdom
The contemporary British state is based on the principles of nineteenth century liberalism and is therefore distinct from the French centrist state tradition dominant in most of mainland Europe. In this context national and sub-national political institutions of government have been seen as providers of specific services to their constituents rather than as foci of political identity or political representation (Loughlin, 2001). This has obvious implications for the ability of sub-national actors to mobilise. The government and Parliament exercise overall control over local and regional government as the lack of a formal constitution implies that the UK Parliament is the ultimate political arbiter in the British political system. Loughlin (2001: 43 -48) expounded on the limitations of local democracy in the British state and on criticisms relating to democratic accountability and remoteness from the average citizen.
The election of a Labour government in 1997 gave impetus to the devolution agenda. Referenda in 1997 and 1998 provided a mandate for devolution in the case of Scotland (74.3% in favour), Wales (50.3% in favour), Northern Ireland (71% in favour) as well as an elected Mayor for London (72% in favour). The creation of a Scottish Parliament, a Welsh Assembly, a Northern Ireland Assembly (although intermittently suspended) and a devolved Greater London Authority swiftly ensued. As Leeke et al (2003) recognized, the different arrangements for administrative, executive and legislative arrangements between the devolved regions and the eight English regions have created an asymmetric settlement.
Asymmetric arrangements are not only the result of the different functions devolved to the different territorial units but can also be related to the West Lothian Question (Bogdanor, 1999) of whether Scottish MPs should have the right to vote on English affairs; Sewel motions (Leeke et al., 2003;Keating, 2005), which is an arrangement for the Scottish Parliament to authorize the UK Parliament to legislate in reserved policy areas; and the powers of local government (Christopoulos, 1997;Herbert, 2003;Berry and Herbert, 2005) that can be seen as relegated to a secondary role of mere policy implementation. Furthermore the presence of a strong political culture underpinning a distinctive political system (Kellas, 1989;Keating, 2005) also significantly distinguishes some of the devolved regions.
The assumption within the policy community had been that this asymmetry would be rectified by devolving planning, implementation and eventually legislative power to the English regions. 17 However, in a referendum (4 November 2004) in the North East of England, the electorate rejected establishing an elected regional assembly by an overwhelming majority (78%). The British government duly abandoned plans for similar referenda in the North West and Yorkshire & Humber. Bond and McCrone (2004) have suggested that one should not conclude that regional identity is an insignificant factor in regionalist pressure for devolution. The same authors cannot discount instrumental factors of "perceived economic and social benefits" (2004: 23) as significant to creating demand for regional autonomy.
The question of what went wrong with the campaign is dealt with by Elcock (2005). He listed the reasons offered for this failure to include: mishandled campaigning; insufficiently supportive media; a negative political climate; extreme public cynicism about politics and politicians; media stories about MP expenses; media stories about the expensive Scottish Parliament building; limited powers and the small size of the regional assembly; distance from the people; and the role of euroscepticism. He found weak evidence in favour of the hypotheses that there are some significant interregional variations in support for devolution and resistance to devolution in areas where local government reorganization was required. Elcock further suggested that voters cannot relate to the strategic functions of a regionally elected body. 18 Part of the failure could lie with Labour government policy, which comprehends regionalism in functional terms of increasing efficiency in the redistribution of resources. Or, as Sandford (2006: 91) has argued, "regionalism is presented as a purely technocratic solution to various problems of governance of certain regions of England".
If, indeed, the electorate cannot relate to the core functions of regional policy making, this implies that they do not perceive a democratic deficit to exist. It is reasonable therefore not to expect any grass-roots demand for increasing democratic accountability. It may be the case that regional devolution is not just an issue of adverse political circumstance or educating the electorate to the benefits of devolution of power, but lack of a functional imperative for the existence of regional governance structures. 19 It has been argued elsewhere that devolution may have less to do with a sense of shared local identity and more with a sense of shared common purpose Governance Capacity and Regionalist Dynamics 371 (Christopoulos, 2005). Obviously, notions of community would be fundamental in such a case, although it can be argued that citizen preferences are affected by the neo-liberal discourse and determined by citizens' evaluation of personal utility. Sandford (2006) suggested that sub-state national identity plays some role in the failure of the English regions to capture the popular imagination.
The regions chosen from the UK try to capture the asymmetric devolution of the UK state. Two regions have strong identity and institutions that underpinned the bottom-up regionalist demands and two regions have weaker regional identity and regional institutions which have resulted from top-down regionalization. Wales (capital Cardiff) is a region with low per capita income within the UK (90% of EU), with some indication of concentration of economic innovation in biotech and the aviation industry (1.6 times the national average of patents). Scotland (112% of EU) has an innovative cluster in semiconductors (1.3 times the national average in patents), presence in the biotech and laser innovation industries, as well as a good rate of R&D investment from government (13% of national expenditure). The North West of England (regional capital Manchester) is a region with average per capita income (103% of EU) and some evidence of an innovative milieu in the aviation industry (1.9 times the national average of patents), with presence also in the computer and biotech industries. By comparison the North East (regional capital Newcastle) has low per capita income (91% of EU) with no exceptional innovative activity or evidence of exceptional business or government R&D spending. Overall, the experience of industrial decline is a common factor in all these UK regions, while the asymmetric devolution of their political institutions provides the context of their different responses to economic dilemmas of regeneration.

Embeddedness of Regional Institutions
The article proceeds by constructing an array of measures based on legal and published policy documents. A classification of nation-states from unitary to federal, as available in the policy literature (European Commission, 2001a;AER, 2003), provides the first such measure. Then the right for direct representation at the Committee of the Regions is examined, signifying the relative power of regions within a national administration. This measure tries to capture the relative importance of regional bodies vis-à-vis local government. The degree to which local government is involved directly or, indeed, has the power to block or effect the nomination of candidates from regional bodies merits examination. In most non-federal EU states local authorities are given rights of codecision on the appointment of CoR delegations (CoR, 2004). Administrative procedures for nomination in the states examined are indicative of the administrative autonomy of regional interests.
Delegation nominations in England are made by the Local Government Association (LGA), the English Regional Assemblies and the greater London Authority. The LGA, together with the Local Government International Bureau (a department of the Foreign and Commonwealth Office), are "responsible for ensuring the final nominations are politically, geographically, ethnically and gender balanced" (CoR, 2004: 60). Similar procedures are followed in Scotland and Wales, although the Scottish Parliament and Executive and the Welsh Assembly are ascribed equal status with local government in the nominations process. 20 The Polish delegation to the CoR is nominated after the joint committee of Government and Local Government decides the distribution of seats between "leading corporations of the local government according to their size and importance" (CoR, 2004: 45). In the case of the Spanish delegation a Senate Motion (serie I, no34, 28 October 1993) determines that each Spanish region will have a seat in the CoR, with the four remaining reserved for local government.
The CoR (2004: 65) report identifies these variations in selection criteria and ascribes them to variations in institutional arrangements: in federal countries or countries with strong regional systems, such as Germany, Austria, Belgium, Spain or Italy, the representative role of the regions is clearly acknowledged, and the selection procedures for members of the Committee of the Regions are expressly stated in their legal texts . . . on the other hand, in those countries with weaker regional systems, or which have no regional system at all the representatives are above all-or even exclusively-local representatives.
A number of other variables augment this index. The legitimacy of regional administrations in their capacity as bodies representing regional interests at EU institutions crucially depends on the existence of an elected regional assembly and on the competence of such a body for key areas of policy in which it can lobby the EU.
The overall legitimacy of a region at the EU level would also crucially depend on the resources the region can command and on its autonomy from the national administration. Scotland, Galicia, Catalunya and, to a lesser degree, Wales have some legislative autonomy that includes a number of exclusive areas of competence. These vary among regions but generally include education, health, spatial planning, roads, transportation, culture and economic development (AER, 2003: Appendix 1).
Constitutional status represents the strength of regional autonomy and it could depend on a constitutional provision or a legislative act. This measure attempts to present a picture of entrenched rights of regions by gauging the difficulty in amending the institutional competence of regional political institutions.
On all dimensions (Table 1) the Polish regions exhibit consistency of values with very low standard deviation. It is self evident that the Polish regions have a low level of autonomy. Low autonomy is also evident for the two English regions, but here the standard deviation of values is rather large. This signifies the equivocal place of these regions within a state with assymetric devolution. The four Spanish regions, Scotland and Wales exhibit high average scores for these measures of autonomy.
The Relative Weight of the Regional Economy A measure of the relevance of the regional economy is constructed based on two broader criteria. First, two variables are considered as proxies for a region's gross economic weight. These are the relative size of the population of a region and the proportion of national economic activity based in the region, as a percentage of national GDP. 21 Together they present a fair measure of the relative economic significance of a region for its respective nation-state. 22 Relative performance and therefore prospects of regions are then considered through some statistics on investment. A measure of the vitality of a regional economy is the gross investment in tangible goods. Manufacturing investment provides a good telltale; a high rank here typically signifies the existence of a regional industrial milieu. Research and Development (R&D) funding can be seen as the most crucial signifier of economic prospects and sentiment within an economic milieu. 23 This partially relates the presence of a developed service sector since high levels of R&D can be seen as an indication of an entrenched service economy in a region. The distinction between private and public R&D can be of interest, since high public funding would indicate a region with high political clout, while high private funding could indicate a region with good long-term economic development prospects.
These measures are put together in a series of ranked coefficients presenting an intriguing picture of regional economic relevance between the regions examined (Table 2). In Spain the economic prominence of Catalunya and Andalucia is evident, the gross investment and R&D expenditure in the former far surpassing any other region examined. In Poland the dominance of Mazowieckie, indicates a strong core -periphery structure, although Slaskie is still an important part of the national economy and Pomorskie shows excellent long-term prospects indicated by the business investment in R&D. In the UK the dominance of the North West seems to be underpinned by a massive disparity in R&D investment by the public sector and universities. Scotland shows healthy figures of gross manufacturing investment but a lack of confidence from the private sector as indicated from their weak R&D investment.
Overall there are wide disparities between these regions. A higher degree of autonomy 25 is associated with better economic prospects, as signalled by gross investment in the regions. However, relative autonomy for regions such as Galicia or Wales has not been linked to exceptional economic performance (European Commission, 2006). Against a common assumption, there is also very weak evidence that higher autonomy is correlated with more public investment in R&D.
Comparing the two variables of economic weight with the three indicators of relative performance/prospects, small variations in average scores can be observed. More than one point rank difference is observed for Galicia and Pomorskie who appear to rank as having better prospects than their relative economic weight, while Andalucia and Scotland have worse prospects than their relative economic weight. These variations are not substantial. Two-third of regions examined have almost identical ranks in economic weight and prospects.
To explore the interaction between these factors further, the relationship between the relative economic weight of a region with the institutional variables that give a measure of autonomy for regional governance is illustrated in the following section.

Exploring the Political Geography of Regionalism
In Figure 1, membership in the different quadrants can be explained with reference to the relative autonomy and degree of economic prominence of each region within their Governance Capacity and Regionalist Dynamics 375 respective nation-state. The dynamism of regional elites as agents of change is not examined directly within the present framework (it is explored elsewhere by Christopoulos, 2005). A typology is created to capture the different positions within which the regions examined can be classified. In Table 3 the different quadrants are assigned values relating to the different locations within the political geography of regionalism. In the case of the first quadrant, regions have high economic weight and good prospects but have low governance capacity. One can observe regions where low governance capacity is coupled with above average economic prospects and relative economic weight. These regions appear dependent on nation-state regionalization. Such top-down regionalization would presumably attempt to introduce administrative efficiency gains by increasing governance capacity for the regions. Examples of this are most of the Polish regions and the English North West region. 26 Obviously, Europeanization as an external pressure that can lead to administrative modernization could play a significant role, although as Radaelli (2003) argued, macro level top-down Europeanization effects are unlikely.
In the case of the second quadrant, a high degree of governance capacity is coupled with relatively high economic weight. Most Spanish regions and Scotland end up in this category. In this quadrant are found political elites that appear emancipated In the fourth quadrant are the regions with low relative governance capacity coupled with low relative economic prominence and prospects, such as the North East region of England or Mazurskie of Poland. It is difficult to see regional autonomy developing without an external shock introducing a change in either the relative weight or the political capacity of these regions. Their peripherality, in political as well as relative economic weight, indicates that it is unlikely for a significant change in their political autonomy to occur. It can be assumed that changes in institutional architecture would be the outcome of processes external to the region.
In the third quadrant, regions with high governance capacity but low economic relevance can be seen as the obvious 'hotbed' of regionalist or bottom-up demands for regional devolution. Here the political institutions are present but regions are peripheral and relatively underperforming, as compared with other regions within the state. This hypothesis is premised on the assumption that regional agents can mobilize effectively at the sub-state level or foster effectively the public disaffection with the inefficiencies of centralization. Demands for devolution could lead to a further increase of governance capacity at the regional level.
Quadrants one and three can be seen as comparatively unstable institutional locations where economic or political forces could drive devolutionary dynamics. Quadrants two and four can be perceived as comparatively stable locations where regional autonomy has reached a relative equilibrium commensurate with the degree of political capacity and relative economic weight of a region. The tentative conclusion here is that regionalism and regionalization can be the outcome of a disequilibrium between political capacity and a regions' relative importance in the national economy. Obviously this analysis, based on a static snap-shot picture of these parameters, has limited utility as a predictive tool.

Towards Regional Clustered Convergence?
The data employed have obvious limitations. It is important to reiterate that this analysis is sample specific and, although the regions where chosen as representative of the varying conditions of regionalist pressures by large states in the EU, these results cannot be assumed universally generalizable. In the construction of the two indexes, variables have also been assumed to have equal weight, although it is recognized that this is unlikely. For all the weaknesses in the data, the results provide novel insights into a number of assumptions in the field. 27 In the twelve regions within the present sample, those with high economic relevance but low institutional autonomy are candidates for regionalization. This appears to be conditioned on whether the respective nation-state would respond to broader Europeanization processes and/or EU policy process requirements by further devolving. Conversely, in those regions with some autonomous institutional capacity but low economic relevance, the ground is fertile for bottom-up devolutionary pressures to emerge. 28 This exploratory analysis attempts to sketch a layer of the 'political geography of regionalism' and place regional autonomy within the context of economic and institutional parameters. The existence of an economic milieu cannot be uncoupled from the other economic indicators employed here. Regions with evidence of a milieu ranked high on economic prospects. No clear evidence of the role of economic milieu in fostering regional devolution is offered here. However, the existence of an economic milieu is likely to underpin claims for political autonomy as it is highly correlated with the relative economic weight and prospects of a region.
It is suggested that pressures for regionalism would be strongest when governance capacity exists but the region is relatively lagging within their nation-state (quadrant 3 hypothesis). The caveat is that if regional elites are weak, or fail to mobilize, then regionalist demands are unlikely to emerge. It is reasonable to assume that the same Europeanization dynamics that lead states to clustered convergence (Borzel, 2002;Radaelli, 2003) will force regions, ceteris paribus, to clusters of similar policy choices and governance structures. It is argued here that the parameters likely to lead to such clusters will relate to functional economic imperatives and political capacity.
There has been no examination of the role of political actors as political entrepreneurs or agents of change-a study that would require sophisticated data on the role of political actors within regional and EU governance. 29 Neither can the effect of cohesion policies on regions with higher governance capacity be explored yet. 30 Overall, the analysis indicates a nuanced relationship between levels of autonomy and the role regional elites could play in furthering or constraining it. Future analysis should use multivariate and path analysis models to explore the complex causal relations among variables. Obviously, comparative data do not allow modelling for processes of change, so many of the more exciting hypotheses in the field will not be empirically tested before longitudinal data become available. 21 Labour as a resource of locational advantage is explored by Bradley and Taylor (1995). 22 Including more variables on the regional economy would confound this analysis. Multiple measures are likely to have multiple causes and, since this is exploratory work, maintaining a lean (parsimonious) model allows for stronger causal assumptions. 23 Camagni (1995) presented the relevance of innovation to the creation of economic milieu, while Krugman (1991) presented a potent argument for the relevance of location to economic development. 24 Note that ranks are given for all NUTS 2 regions in a national economy, and that only ranks of the regions under examination are reported here. Ranks represent a relative place among the 17 Spanish, 16 Polish and 12 British regions. The comparative context is one of relative rank. 25 Variable levels of autonomy in Spain and the UK imply that some regions have more political power and policy competence than others in the same state. 26 If this hypothesis can be confirmed the North West would have presented a better case at regional devolution than the North East region and, therefore, would concur with those that perceived it as a better prospect for regional devolution within the UK (see Elcock, 2005). 27 Data limitations not specifically mentioned so far reflect generic weaknesses of exploratory analysis. 28 I do not make the assumption that regional devolution has positive economic effects. As McGregor and Swaves (2005) argued, devolution of development policy benefits regions because of information advantages but has detrimental effects by relegating the effects of regional economic interdependence. 29 See Christopoulos (2006) for a definition of political entrepreneurship and an application of the concept to EU policy making. 30 Indeed, it can be argued that the European Commission favours a corporatist agenda for regional actors. Commissioner for regional policy Danuta Hubner has recently claimed that "The Lisbon strategy must draw inspiration from the partnerships that cohesion policy has set up between the national, local and social players. . .. The regions need Lisbon" (European Commission, 2005). Cawson (1985) in his analysis of local corporatism identified the significant role not only of regional but also national and supranational elite actors.