Regional Authority, Transnational Lobbying and the Allocation of Structural Funds in the European Union*

bs_bs_banner JCMS 2013 pp. 1 17 DOI: 10.1111/jcms.12038 Regional Authority, Transnational Lobbying and the Allocation of Structural Funds in the European Union* ADAM WILLIAM CHALMERS Leiden University Abstract The allocation of European Union (EU) structural funds is subject to fierce regional lobbying. This article examines the extent to which regions with greater political authority are better able to lobby for funds than their weaker counterparts. Existing research acknowledging the importance of regional authority in these processes has used inadequate indicators. This analysis, drawing on the Regional Authority Index, is the first to use regional-level data disaggregating between regional authority as self-rule and shared-rule. It also uses data that measure the lobbying capacity of regions Brussels-based lobbying offices. Controlling for a battery of competing and control variables, Tobit regression analyses of 181 regions receiving funds in the 2007 13 period suggest that regional authority expressed as shared-rule, but not self-rule, has a significant impact on the allocation of structural funds in the EU. Introduction European Union (EU) regional policy has sparked a veritable cottage industry of research, engaging the efforts of economists, regionalists and political scientists alike. Understanding this immense interest is easy. As a transnational redistributive strategy aiming to reduce economic disparities across Europe s many regions, EU regional policy is unprecedented. Its central mechanism the structural funds comprises more than 30 per cent of the EU s total budget or about 308 billion in the latest funding period. It also affects more than 450 million EU citizens across a massive and diverse geographical area in all 27 Member States. While scholars from different disciplines have variously examined the effectiveness of regional policy (Beugelsdijk and Eijffinger, 2005; Boldrin and Canova, 2001), how it has facilitated a form of multi-level governance in Europe (Hooghe and Marks, 2001; Marks, 1993) and how it has Europeanized different actors and institutions (Bache, 2008), little attention has been given to the actual politics of regional policy that is, to use Lasswell s famous formulation, we still know little about the who gets what, when and how underlying regional policy (Lasswell, 1950). This article examines the determinants of structural funds allocation to regions in the EU. Its starting point is the idea that the allocation of funds is subject to considerable pork-barrel politics, with powerful regions fiercely lobbying to effect financial transfers in their own favour (Wallace, 1977; Blom-Hansen, 2005). Strong regions (those that * I would like to thank the anonymous reviewers and numerous colleagues who provided insightful comments on earlier versions of this article (specifically those at the 2012 Netherlands Institute of Government Annual Work Conference in Leuven). I would like to extend special thanks to Juliet Johnson for her helpful guidance in the very early stages of this project., 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA

2 Adam William Chalmers exercise their own authority and have considerable decision-making powers) have long had a clear, independent presence in Brussels, whether through lobbying offices, the Committee of the Regions or, in some cases, even representation in the Council. Nevertheless, the impact of regional authority on structural funds has only been partially examined. Research has instead tended to focus on funds allocation as the result of intergovernmental bargaining and partisan politics. Where the effects of regional authority have been considered (for example, Dellmuth, 2011; Kemmerling and Bodenstein, 2006), the data used have been inadequate, measuring regional authority at the state rather than regional level and reducing regional authority to a single measure of decentralization. I argue that understanding the effects of regional authority on funds allocation requires regional-level data capturing the multidimensional nature of regional authority. This entails examining how different dimensions of regional authority have potentially different effects on the allocation process. To this end, this article uses regional-level data from Hooghe et al. s (2010) Regional Authority Index measuring regional authority across two dimensions self-rule and shared-rule using eight indicators. This article presents an empirical analysis of the determinants of structural funds allocation to 181 regions for the period 2007 13. Its central aim is to examine the extent to which regional authority measured as shared-rule and self-rule impacts the distribution of funds to regions. Its central hypothesis is that shared-rule, characterized by greater institutionalized regional participation in the national decision-making process, is more effective when it comes to lobbying for structural funds than the type of regional autonomy that characterizes self-rule. Additionally, this analysis conceives of this lobbying process within a multi-level framework where regions are engaged at the Member State level and at the EU level. As such, data on regional authority are coupled with data on the organizational characteristics of regions Brussels-based lobbying offices. Regression analysis is used to test these factors while controlling for competing explanations. My central finding provides evidence that shared-rule is an important determinant of funds allocation. The resources of lobbying offices, however, are not. Lobbying for structural funds takes place mainly between regions and central governments, and is facilitated by a region s institutionalized participation in state-level decision-making; EU-level lobbying is not primarily concerned with winning funds. I. The Politics of Structural Funds Allocation in the EU On paper, the allocation of EU structural funds follows a straightforward technicalfunctional logic. Regions with low economic growth and high unemployment rates relative to the EU average have long been the primary targets for regional policy funding. Redistribution was part of what the European Commission called an objective and transparent method of allocation (European Council, 2006, preamble 31). Recent research, however, has questioned these objective criteria. Indeed, empirical analyses have compellingly demonstrated that being poor is neither a strong nor a sufficient predictor of the amount of Structural Funds per head a region receives (Kemmerling and Bodenstein, 2006, p. 382). If need is not the main determinant of the allocation of these funds, then what is? Several scholars have explained funds allocation as the result of a larger intergovernmental bargaining process where funds act as side-payments to the losers of EU

Regional authority, transnational lobbying and structural funds 3 integration. In other words, rich Member States channel funds to poor Member States in exchange for the latter s consent on other EU issues (Wallace, 1977; Allen, 2005). Other scholars, looking beyond interactions among Member States and the EU, have implicated regions themselves in allocation outcomes. Early proponents of multi-level governance focused on how the introduction of the partnership and programming principles to regional policy in 1988 empowered regions to outflank and even bypass central governments (Marks, 1992). How these newly empowered regions affected funds distribution in practice, however, fell far short of expectations for an emerging Europe of the Regions (Huysseune and Jans, 2008, p. 4). Several recent empirical studies have put assumptions about both side-payments and multi-level governance to the test. Kemmerling and Bodenstein (2006) assessed the determinants of funds allocation with a focus on the effects of party politics. In addition to finding some support for the side-payments thesis in the form of anti- and pro-eu cleavages in European Parliament elections, the authors also found visible but not robust (Kemmerling and Bodenstein, 2006, p. 374) evidence that regional authority had some bearing on how much funding individual regions received. A more recent article by the same authors (Bodenstein and Kemmerling, 2011) provides further evidence for the effects of partisan politics and regional power in the determination of funds allocation. Two articles assessing structural funds allocation have also weighed in on the sidepayments thesis, but with different results. Bouvet and Dall erba (2010) and Dellmuth (2011), examining side-payments in terms of the prevalence of Euroscepticism among EU citizens, found that funds are rarely allocated to the losers of integration. For Bouvet and Dall erba (2010), the main determinants of funds allocation are partisan politics (in particular, party ideology at the regional level) and additionality. Regions represented by left-wing parties receive more funding than regions represented by right-wing parties. Further, additionality the amount of matching funds regions and Member States are required to pay in order to receive structural funding was strongly correlated with getting more structural funding. This tends to favour rich regions that are able to provide additional funds over those poor regions that cannot. Dellmuth (2011) examines how regional authority acts on the Commission s logic for funds allocation and penchant for transferring funds to regions that have effectively managed structural funds in the past. However, regression results present a mixed picture showing that constitutionally weaker regions receive more funds than constitutionally stronger regions when they have proven themselves to be effective managers of funds in previous funding rounds. II. Regional Authority and Structural Funds Allocation This analysis assesses the extent to which regional authority impacts structural funds allocation. Its starting point is that regional policy presents an opportunity structure impacting a few well-positioned regions. Following Gary Marks (1993, p. 392), this does not entail regions outflanking the state, but rather regions taking part in negotiations within the framework of several tiers of nested governments at the supranational, national and regional levels. For this analysis, this amounts to regions lobbying for funds within a multi-level framework involving regional, national and EU actors. The rules governing regional policy define the scope of regional lobbying at these different lobbying levels. The regional policy process consists of two stages. The first

4 Adam William Chalmers stage takes place at the EU level and sees the Commission setting the Community strategic guidelines (CSG) and regional policy budget for a given funding period. CSGs define the Community s development priorities and are important in the second policy stage (discussed below). The budget is the main locus of lobbying at this first stage. While the Commission uses transparent criteria to determine allocation to Member States, these criteria are rather flexible, with the Commission retaining considerable discretionary power over who gets what and how much (Kemmerling and Bodenstein, 2006). While it would clearly be an exaggeration to say that regions fully determine their own individual allocations at this stage, the Commission s flexibility does present an opportunity for lobbying efforts. The second stage takes place at the Member State level and involves both national and regional governments. Following the funding priorities set out in the Commission s CSG, Member States develop a general national strategic reference framework (NSRF) that identifies the link between Community priorities on the one hand and national reform programmes on the other (Molle, 2007, p. 170). The individual operational programmes detailed in the NSRF are the main locus of lobbying here. Operational programmes are determined by Member States in consultation with regional authorities following the partnership principle. This principle ensures that Member State governments involve a broad range of regional and sub-state actors when elaborating on development projects and funding. Importantly, operational programmes outline exactly how much structural funding goes to which development projects in each region. Perhaps not surprisingly, the establishment of operational programmes is, to speak with Molle (2007, p. 171), often the result of fierce battles between the national government and regions. Two factors set certain regions apart from others when it comes to lobbying at both stages of the regional policy process: the allocation of political authority at the sub-state level; and the lobbying resources of regions Brussels-based lobbying offices. I discuss each factor in turn. As previous research suggests, regions in federal and/or semi-devolved states can be expected to have greater clout when it comes to lobbying for structural funds than regions in more centralized states. Politically powerful regions have the ability to exert greater influence over the formulation of EU policies, can speak for themselves at the EU level and can hold their own against national administrations. A central insight of this analysis, however, is that regional authority is a multidimensional concept and that we can expect different dimensions of regional authority to have different effects on the allocation of structural funds. Following Elazar (1991) as well as the more recent work of Hooghe et al. (2010), regional authority consists of two dimensions: self-rule and shared-rule. Self-rule is the authority exercised by a regional government over those who live in its territory ; shared-rule is the authority exercised by a regional government or its representatives in the country as a whole (Hooghe et al., 2010, pp. 7 8). In other words, self-rule refers to the autonomy of a region from the central government, while shared-rule refers to regional participation and co-decision-making at the state level (Auer, 2005). These differences have an important effect on regional preferences and the way that regions interact with the central government. Indeed, research shows that shared-rule gives regions a bigger stake in aggregate or countrywide outcomes, while self-rule fosters inward-looking preferences and creates tension among regions and central governments. For example, where shared-rule assuages the risk of ethnic mobilization, self-rule has been shown to incite related protest activities (Brown, 2009); where shared-rule raises

Regional authority, transnational lobbying and structural funds 5 barriers to political corruption, self-rule creates incentives (Neudorfer and Neudorfer, 2012); and where shared-rule has been shown to mitigate within-country spatial disparities, self-rule exacerbates these differences (Ezcurra and Rodriguez-Pose, 2011). We can theorize similar differences between self-rule and shared-rule when it comes to the allocation of structural funds. Of central importance here is the fact that the regional policy process is based on negotiations between regions, central governments and the Commission. These different governmental tiers co-ordinate to formulate CSGs, the regional policy budget and NSRFs, bringing together regions and central governments via the partnership principle. In this sense, autonomy from the central government the central feature of self-rule is not an advantage. Instead, participation and co-decision making central features of shared-rule become indispensable. Where self-rule has a tendency to create or exacerbate tension between different tiers of governments, shared-rule institutionalizes a region s interaction with the central government. Shared-rule, insofar as it gives regions a stake in countrywide outcomes, would even work to legitimize a region s efforts to lobby for structural funds. These insights lead to my first hypothesis: H1: Regions with greater shared-rule will receive more structural funds than regions with greater self-rule. Disaggregating regional authority by self-rule and shared-rule provides insight into a region s lobbying incentives for funds as well as a region s role in the regional policy process. This marks a considerable advance on existing research. Nevertheless, it is still only part of the story. Scholars have long speculated that lobbying for funds, especially at the EU level, also centres on Brussels-based regional lobbying offices. The establishment of the first lobbying offices in the mid-1980s was immediately linked to efforts to secure structural funds (Jeffery, 1997). Although the remit of these offices has broadened over the years and winning structural funding is no longer their raison d être, a strong link to regional policy remains (Mazey, 1995, p. 79). Indeed, these lobbying offices provide direct channels to European Commissioners and MEPs and can help regions exercise influence over the entire regional policy process. The real lobbying potential of these regional offices is their resources. Office finances, staff and credibility increase a region s lobbying capacity and thus help regions win funds. These assumptions lead to my second hypothesis: H2: The greater a regional lobbying office s resources, the greater the share of structural funding that region will receive. III. Research Design This analysis examines the determinants of EU structural funds allocation in 181 regions for the period 2007 13. Building on insights from the existing theoretical and empirical literature, this analysis examines funds allocation in terms of six explanatory variables: the EU s objective criteria; additionality funding; regional partisan politics; the sidepayments thesis; regional authority; and Brussels-based regional lobbying offices. In what follows I will discuss the operationalization of these variables as well as a further control variable: regions from new Member States.

6 Adam William Chalmers Structural Funds Allocation There are three main types of structural funds available in the period under consideration. Convergence funds comprise 81.5 per cent of the total regional policy budget and cover regions whose development is lagging behind (European Council, 2006, preamble 17). These funds follow the EU s transparent criteria: they are distributed to regions whose gross domestic product (GDP), measured in purchasing power parities, is less than 75 per cent of the EU average or, for cohesion regions, whose GDP is less than 90 per cent of the EU average. Competitiveness and employment funds comprise 16 per cent of the total regional policy budget and cover the territory of the Community outside the Convergence objective (European Council, 2006, preamble 18). Unlike earlier funding schemes, these funds have no specific criteria for allocation. Instead, they are channelled to regions that do not receive convergence funds. Finally, territorial co-operation funds comprise 2.5 per cent of the total regional policy budget and cover multiple regions spanning two or more Member States. The transnational character of this third type of funding cannot be directly linked to specific regions and will therefore not be considered in this analysis. This analysis examines the allocation of convergence and competitiveness and employment funding to all regions receiving funding in the 2007 13 period. This includes 181 regions from 17 Member States including regions in new Member States from the 2004 enlargement. Funding programmes covering entire countries (for example, Denmark, Estonia, Latvia, Malta, Cyprus and Lithuania) are not included in this analysis. Data for both funding types were gathered online at the EU s official website. Following existing research, the total amount of funding individual regions received was scaled by regional population to control for variation in the size of the regions under consideration (Bouvet and Dall erba, 2010). A full list of these regions and the type and amount of funding they received is available in the online appendix. EU Objective Criteria The EU s objective criteria for convergence funds are based on regional GDP: less than 75 per cent of the EU average in most cases or less than 90 per cent for so-called coherence regions. Data were derived from Eurostat and, following Commission procedures, were calculated on the basis of Community figures for the period 2000 to 2002 (European Council, 2006, Article 5). 1 There are no specific criteria for regions covered by the competitiveness objective. Nevertheless, these funds, as the predecessor of objective 2 funds in previous funding periods, have a distinct link with regional unemployment rates. As such, data on long-term regional unemployment relative to the EU average (derived from Eurostat and calculated as the average of values from the years 2006 and 2007) will be used as a control variable for competitiveness funding. Both variables were logged to normalize distribution. Additionality The additionality principle requires that regions and/or Member State governments provide matching funding for all development programmes receiving structural funds. 1 Due to data availability, GDP data for Austria, Hungary and Italy were calculated as the average of 2007 and 2008.

Regional authority, transnational lobbying and structural funds 7 Doing so limits support to dubious development projects (Bouvet and Dall erba, 2010, p. 502) by ensuring regional and national investment. Unlike previous funding periods, there are no fixed rules for additionality in the 2007 13 period. Following Bouvet and Dall erba (2010), additionality requirements affect structural funds allocation in that they make it difficult for poorer regions to secure larger amounts of funding. This places poorer regions at a disadvantage when bargaining for funds. Additionality is measured as the total national contribution to each operational programme scaled by regional population. Data were gathered from the EU s official website. This variable was also logged to normalize distribution. Partisan Politics A consideration of the impact of partisan politics on funds allocation is central to most existing empirical studies (Bouvet and Dall erba, 2010; Kemmerling and Bodenstein, 2006; Bodenstein and Kemmerling, 2011; Dellmuth, 2011). Following this work, I consider three indicators: regional government ideology (left/right); alignment between regional and national leading parties; and the size of the leading regional party s electoral margin. Data for these variables were gathered from the 2006 Chapel Hill Expert Survey (Hooghe et al., 2008) and the Norwegian Social Science Data Services. 2 First, a party s ideological position on the left right spectrum has an impact on redistributive policies insofar as left-wing parties tend to favour more government spending than right-wing parties. For EU regional policy, following Dellmuth (2011, p. 1025), regions where left parties are strong should lobby more effectively for higher funds. Left parties would also find electoral support in increased government spending facilitated by structural funds. In this analysis, party ideology is assigned to leading regional party governments for 2006 using Chapel Hill Expert Survey data. These data place parties on a 10-point scale with 0 being the extreme left, 5 being centre and 10 being the extreme right. Second, a region s ability to impact funding outcomes also depends heavily on how regional governments interact with national governments. In particular, this interaction is facilitated by the coincidence of party alignment between the two levels (Bouvet and Dall erba, 2010). Is the regional government leading party the same as the national government leading party? Regions are expected to receive more funding when parties are aligned than when they are not. A dummy variable (1 = alignment) was used to capture alignment between regional and national government leading parties. And third, following Bouvet and Dall erba (2010, p. 512), the willingness of a regional or national government to bargain for more funds will also depend on how secure its current [electoral] majority is. Securing funds can be a way for regional governments to appease constituents and can increase their chances of re-election. Regional governments that are less secure in their chances for election (those with smaller electoral majorities) have a greater incentive to lobby for funds. The extent of a regional government s electoral majority is the difference between the leading regional party s electoral support and the second regional party s electoral support. Specifically, this is measured as the difference in the percentage of votes for the first party and the second party. The 2 Available at: «http://www.nsd.uib.no/nsd/english/index.html».

8 Adam William Chalmers smaller the difference, the more a regional government will feel the need to lobby for more funds. Data for this indicator were derived from the Norwegian Social Science Data Service and taken for the most recent regional-level elections prior to 2007. 3 Side-Payments Thesis Side-payments explain structural funds allocation as a form of compensation to the losers of EU integration. Following existing research, we should thus see more funding going to regions with widespread anti-eu sentiments. For Kemmerling and Bodenstein (2006) this amounts to the presence of strong anti-eu parties. In this analysis, the side-payments thesis is measured as a regional government party s position toward EU integration. Data are derived from the Chapel Hill index and measured on a1to7scale, with 1 being strongly opposed and 7 being strongly in favour. Lower scores correspond to greater anti-eu sentiments. Side-payments can also be captured in terms of anti-eu sentiments directly measured at the level of the EU citizen (Bouvet and Dall erba, 2010; Dellmuth, 2011). A second indicator for the side-payments thesis is derived from the 2006 standard Eurobarometer survey. The question used to measure Euroscepticism asks respondents if EU membership is: (1) a good thing, (2) a bad thing or (3) neither a good nor bad thing. Following Bouvet and Dall erba (2010), the greater the response for EU membership as a bad thing, the more widespread anti-eu sentiments are and thus the more structural funds we can expect these regions will receive. Regional Authority Existing research has not only given short shrift to the effects of regional authority on structural funds allocation but has also used inadequate indicators measuring regional authority. The few studies that have examined regional authority use Lijphart s (1999) federalism index, which is limited by the fact that it only provides scores at the national level and is therefore unable to capture variation among regions within a state. This is particularly problematic when state decentralization is not uniform across all regions (like in Spain, Belgium, the United Kingdom and Italy) and even more so when some regions have special autonomous status within a state (like Åland, Azores and Madeira). A central advance made in this analysis is the use of the regional authority index, which provides scores measured at the regional level. It also assesses the multidimensional nature of regional authority by disaggregating it across two main domains self-rule and shared-rule and on eight dimensions. Self-rule is the sum of four indicators: institutional depth ( the extent to which a regional government is autonomous rather than deconcentrated ); policy scope ( the range of policies for which a regional government is responsible ); fiscal autonomy ( the extent to which a regional government can independently tax its population ); and representation ( the extent to which a region is endowed with an independent legislature and executive ) (Hooghe et al., 2010, p. 8). Scores for these indicators range from 0 to 3 or 4, with lower scores indicating less self-rule and higher scores indicating more self-rule. Aggregated self-rule scores run from 0 to 15 with 3 Data for Italian regions were not available for the period before the 2007 13 funding period. Thus, data were instead derived from the next most recent parliamentary elections.

Regional authority, transnational lobbying and structural funds 9 higher scores indicating more self-rule than lower scores. Shared-rule is the sum of: law-making ( the extent to which regional representatives co-determine national legislation ); executive control ( the extent to which a regional government co-determines national policy in intergovernmental meetings ); fiscal control ( the extent to which regional representatives co-determine the distribution of national tax revenues ); and constitutional reform ( the extent to which regional representatives co-determine constitutional change ). Scores for these indicators range from 0 to 2 or 3 with lower scores indicating less shared-rule authority and higher scores indicating more shared-rule authority. Aggregated shared-rule scores run from 0 to 9, with higher scores indicating more shared-rule than lower scores. In addition to self-rule and shared-rule, the index accounts for a region s special autonomous status vis-à-vis central government. Such regions, like Åland and Madeira, are sui generis in that they receive special treatment in the constitution and in statutory law (Hooghe et al., 2010, p. 29). Special status combines strong shared-rule with a more general goal of influencing national legislation in a way that primarily benefits the region itself. As such, these regions often face a sharp trade-off between deciding their fate and co-determining that of the country (Hooghe et al., 2010, p. 31). These mixed preferences make special status regions lobbying potential somewhat more difficult to define. However, the importance of co-decision-making and the institutionalized interaction between these regions and central governments suggest that special status regions should have some of the same advantages of regions with high shared-rule scores. Special autonomous status applies to just 14 regions analyzed here and was measured as a dummy variable (1 = special autonomous status). The following analysis examines the effects of both aggregate as well as individual self-rule and shared-rule indicators on funds allocation. This will allow us to not only test the effects of self-rule and shared-rule on the allocation process, but to better and more accurately explain the results. In other words, how do specific aspects of self-rule and shared-rule influence the allocation of funds? Table 1 provides descriptive statistics for all regional authority indicators. Table 1: Regional Authority Descriptive Statistics Mean Standard deviation Minimum Maximum Self-rule 8.96 3.66 1 14 Institutional depth 3.30 0.63 1 3 Policy scope 2.07 0.94 0 4 Fiscal autonomy 1.37 1.28 0 4 Representation 3.22 1.26 0 4 Shared-rule 1.78 2.87 0 9 Law-making 0.32 0.61 0 2 Executive control 0.38 0.65 0 2 Fiscal control 0.47 0.74 0 2 Constitutional reform 0.59 1.13 0 3 Special status 0.23 0.80 0 1 Source: Author s own calculations.

10 Adam William Chalmers Regional Lobbying Offices Regional lobbying at the European level also involves regional lobbying offices. Like any lobbying firm, a regional office s lobbying potential is largely linked to its resources. For this analysis, a regional lobbying office s resources are measured by two indicators: the number of staff working in the lobbying office; and the organizational age of the lobbying office. Having more staff allows a region to better monitor EU activities, conduct research, generate policy-relevant information and pursue multiple lobbying strategies. The organizational age of a lobbying office reflects its experience in lobbying at the EU level as well as its credibility among EU decision-makers and other Brussels operatives (Mahoney, 2004, p. 452). Regional offices with a long history in Brussels have had more time to learn the rules of the game, develop effective lobbying practices and establish important contacts among other lobbyists and decision-makers. Data for both number of staff and age of the lobbying office were derived from the Committee of Regions 2008 Regional Offices Contact Directory 4 and the author s own research. 5 New Member State Regions Regions from new Member States are those that joined the EU in the 2004 enlargement. These regions can be expected to differ from others in that many still lack the administrative capacity to effectively use funds, have struggled to regionalize during the accession process and have little experience negotiating for funds at the EU level (Allen, 2005). A dummy variable was created (1 = new Member State region) in order to control for these differences. IV. The Determinants of Structural Funds Allocation in the EU The following uses regression analysis to test the determinants of structural funds allocation in the EU. From the 181 regions tested, 73 receive convergence funding while 108 receive competitiveness and employment funding. Since the EU s transparent criteria for funds allocation differs across the funding schemes, the two were necessarily tested in separate models. This meant that in each model a large number of regions are recorded as having received zero funds. Tobit regression is a common estimation method for such a truncated or limited dependent variable as it allows us to censor certain observations that would otherwise skew regression results (Bouvet and Dall erba, 2010, p. 515). Table 2 shows the results from Tobit regression analyses in four models: two for convergence funding and two for competitiveness funding. Models 1 and 3 use aggregate self-rule and shared-rule indicators, while models 2 and 4 use the individual indicators. A test for collinearity between these eight individual regional authority indicators as well as special status was run prior to the Tobit analysis. This test showed a very high level of collinearity (with a variance inflation factor of 11.57) with the policy scope indictor. This indicator was therefore dropped from the analysis. A similar test of the three partisan politics indicators showed acceptable tolerance levels. 4 This directory can be found at: «http://ec.europa.eu/regional_policy/conferences/od2008/doc/pdf/catalogue_en.pdf». 5 This research comprises data gathered from regional offices websites as well as expert interviews conducted with regional offices in 2009.

Regional authority, transnational lobbying and structural funds 11 Table 2: The Determinants of Structural Funds Allocation, 2007 13 Variables Convergence Competitiveness and employment Model 1 Model 2 Model 3 Model 4 GDP (log) 1249.61 (372.66)*** 739.47 (187.26)*** Unemployment (log) 79.04 (22.88)*** 114.04 (28.74)*** Additionality (log) 639.62 (149.47)*** 484.85 (102.69)*** 44.29 (16.14)** 51.50 (16.35)** New Member State 457.41 (435.21) 385.21 (245.42) 159.68 (67)* 126.90 (71.88) Party ideology 26.05 (54.24) 31.52 (30.43) 3.37 (8.60) 1.60 (8.50) Party alignment 468.70 (160.04)** 219.89 (86.86)** 13.61 (30.47) 8.04 (30.41) Regional majority 4.12 (10.05) 3.97 (5.70) 1.58 (1.69) 1.23 (1.63) Party position on EU integration 189.18 (89.36)** 13.86 (51.90) 9.02 (16.10) 9.06 (16.18) EU a bad thing 6134.48 (2282.40)** 8057.00 (2070.84)*** 63.90 (360.69) 682.10 (453.51) Self-rule 16.22 (27.47) 4.82 (5.37) Shared-rule 91.12 (35.29)** 5.89 (6.71) Special status 242.17 (85.82)** 1477.53 (126.22)*** 78.28 (24.31)** 92.12 (30.84)** Institutional depth 1295.21 (311.80)*** 142.26 (61.32)* Fiscal autonomy 244.04 (90.45)** 3.08 (22.30) Representation 292.17 (80.13)*** 46.13 (27.37) Law-making 3163.88 (320.86)*** 10.51 (64.64) Executive control 1221.99 (222.57)*** 82.96 (42.01)* Fiscal control 221.47 (200.97) 120.07 (47.66)* Constitutional reform 2325.31 (264.12)*** 65.48 (46.08) Number of staff 37.66 (19.15)* 13.90 (11.60) 2.11 (2.73) 1.33 (2.99) Age of office 27.46 (13.39)* 24.28 (9.24)** 0.54 (2.82) 0.08 (2.90) Constant 1203.49 (2690.14) 3825.99 (1503.63)** 10,88 (192.43) 17.74 (211.99) Observations (censored) 126 (81) 126 (81) 126 (45) 126 (45) Log-likelihood 352.45 319.28 532.14 527.29 LR Chi 2 132.49 198.82 52.48 62.19 Probability > Chi 2 0.00 0.00 0.00 0.00 Pseudo R 2 0.16 0.24 0.04 0.06 Source: Author s own calculations. Notes: Standard errors in parentheses. * p < 0.05; ** p < 0.01; *** p < 0.001.

12 Adam William Chalmers Table 2 shows mixed results for the EU s objective criteria for funding. First, regions with lower GDPs relative to the EU average (poorer regions) do appear to receive more convergence funds. This result is expected given that in the 2007 13 period regional GDP is the EU s sole criteria for distributing convergence funds to regions. At the same time, however, regions with lower unemployment rates (richer regions) seem to receive more competitiveness funds. Need is not a consistent determinant of funds allocation across the different funding schemes. In contrast to earlier funding periods, unemployment is no longer part of the EU s official objective criteria in the 2007 13 funding period. Indeed, the EU has no official criteria for the allocation of competitiveness funds in the most recent period. Some caution needs to be exercised in comparing results for convergence funds and competitiveness funds. As the pseudo R 2 scores in models 3 and 4 suggest, very little of the variation (less than 6 per cent) in the allocation of structural funds is actually being explained in these models. This needs to be taken into consideration even when the indicators in models 3 and 4 show significant differences. By contrast, the pseudo R 2 scores for convergence funds in model 1 and 2 show far more promise, indicating that nearly 16 per cent of the variation is being explained in model 1 and about 24 per cent in model 2. Taken together, however, the regression results for EU objective criteria make it difficult to say that being poor is a sufficient condition for structural funds allocation. What is more, regression results for the other variables tested in Table 2 give purchase to the notion that funds allocation is driven by political considerations as well as objective, need-based criteria (see Kemmerling and Bodenstein, 2006). The results for additionality support similar findings to Bouvet and Dall erba (2010). Regions receiving more funding also commit larger amounts of national funds for development projects. What makes this finding interesting, especially in the 2007 13 funding period, is the relative absence of specific criteria for how much additionality funding recipients of structural funds would need to commit to development projects. As Bouvet and Dall erba (2010) suggest, poorer regions are at a considerable disadvantage in this regard: they are potentially excluded from receiving greater amounts of funding because they are simply unable to provide the requisite matching funds. The finding therefore provides a partial explanation for why need is not a stronger determinant of funds distribution: richer regions are in a better position to contribute more to additionality funding and can therefore offset the EU s objective criteria for funds allocation. Being a new Member State region is not a significant determinant of the allocation of convergence funds. At the same time, however, this variable is negatively correlated at a statistically significant level for competitiveness funds. Again, while we need to be cautious in interpreting the results in models 3 and 4, this finding is consistent with the fact that most new Member State regions are not eligible for competitiveness funding in the first place (only three from 33 new Member State regions receive competitiveness funding). Unlike previous research, the results shown in Table 2 provide mixed support for the effects of partisan politics on structural funds allocation. Of the three partisan politics indicators, only regional party alignment shows significant differences (at the 0.01 level or higher) and only for convergence funding that is, more convergence funding tends to be allocated to regions when the leading regional party is the same as the national leading party. Regional government party ideology (left/right) and the magnitude of a regional government s electoral majority show no significant differences in any of the models.

Regional authority, transnational lobbying and structural funds 13 These results stand in stark contrast to the findings of previous research showing that left-wing governments and regional governments without secure electoral majorities tend to receive more funding (Bouvet and Dall erba, 2010; Kemmerling and Bodenstein, 2006). Findings for the side-payment thesis are similarly equivocal. Anti-EU regional parties measured in terms of their support for EU integration only show significant differences on the allocation of convergence funds and only in model 1. There is thus some evidence that convergence funds are used to compensate the losers of EU integration. The sidepayments thesis tested using Eurobarometer data on citizen dissatisfaction with the EU, however, is statistically significant in models 1 and 2 but negatively correlated with funds allocation. While anti-eu parties are the subjects of side-payments, widespread anti-eu sentiments at the level of the citizen are not. The regression results presented in Table 2 provide significant evidence supporting H1, especially for the allocation of convergence funding. Model 1 shows that with each unit increase in a region s shared-rule score we can expect a 91.12 million increase in funding. By contrast, model 1 shows that self-rule is not a statistically significant determinant of the allocation of convergence funds. This finding supports the expectation that autonomy from the central government does not help a region secure a greater level of funding. Instead, the more that a region participates in co-decision-making processes, has routine and institutionalized interactions with the central government and is invested in countrywide or aggregate outcomes, the more convergence funding it will receive. Table 2 also shows that regional authority expressed as special autonomous status is an important determinant of funds allocation, showing significant differences in models 1 and 2. A shift from regions without special status to special status regions marks a dramatic increase in convergence funds: over 242 million in model 1 and more than 1 billion in model 2. The results are particularly interesting given that such special status only applies to four of the 181 regions tested in this analysis. Special status regions might face a sharp trade-off between deciding their own fate and co-determining that of the country, but high levels of shared-rule in these regions, bringing regions into close contact with central governments, provides some further support for H1. Importantly, special status is negatively correlated with the allocation of competitiveness funding (as shown in models 3 and 4) at a statistically significant level. These results and the fact that H1 is only borne out when it comes to convergence funds highlight an important difference between the two funding schemes. Competitiveness funds comprise only 16 per cent of the total regional policy budget while convergence funds comprise 81.5 per cent. In other words, the size of the reward matters. Regions use their political capital when it counts namely in their efforts to secure greater amounts of convergence funding. Where previous studies have drawn out broad conclusions about regional authority in general, this analysis pinpoints the specific dimensions of regional authority impacting the allocation of structural funds and provides a theoretical explanation as to why regional authority matters. This becomes even more evident when we consider the impact of regional authority on funds allocation using the individual regional authority indicators as presented in models 2 and 4. Model 2 shows that law-making and executive control two shared-rule indicators are important determinants of the allocation of convergence funds. Regions with greater law-making capacities are not only given majority representation in the national legislature (as opposed to having representation determined by regional weights), but these

14 Adam William Chalmers regions can veto ordinary legislation (Hooghe et al., 2010, p. 20). High scores for executive control is an indication of routine meetings between regional and central governments with authority to reach legally binding agreements. Both indicators speak to the importance of institutionalized interactions between regions and central governments and the stakes that regions have in aggregate or countrywide outcomes. Having considerable shared-rule powers in terms of law-making capabilities and executive control enhances a region s negotiating position when it comes to formulating CSGs and determining the details of NSRFs for regional policy. Results in model 2 are less clear on two further shared-rule indicators: fiscal authority and constitutional reform. Fiscal authority shows no significant differences in model 2 and constitutional reform is negatively correlated with funds allocation at a significance level of 0.00. The results for fiscal control show that a region s authority over the distribution of tax revenues (something that might speak to a region s capacity for handling and absorbing funds) has little bearing on a region s ability to win greater amounts of convergence funds. Constitutional reform, endowing a region with the ability to veto constitutional change and hence shape the rules of the game, appears to lead to the receipt of less funding, not more. This result is difficult to interpret given the central assumptions made in this analysis and the findings regarding law-making capacity and executive control discussed above. A central finding presented in model 2 is that representation a self-rule indicator appears to be a statistically significant determinant of the allocation of convergence funds. Regional representation refers to the capacity of regional actors to select office holders (Hooghe et al., 2010, p. 20). Do regions have regional assemblies and are they directly elected? Is the regional executive government appointed by the central government or is it appointed by a regional assembly or directly elected? Representation, considered on both of these points, would seem to affect funds allocation primarily via a region s links with, and responsibility to, its constituents. These regions that score high on representation are in the business of pleasing citizens via large-scale spending policies, like EU regional policy. While regional authority explains part of the causal story of structural funds allocation in Table 2 the same cannot be said of regional lobbying offices, whose resources, measured in terms of staff and age, show inconsistent results across the four models. Number of staff is negatively correlated at a significance level of 0.05 in model 1. Age, while significant in models 1 and 2, shows a positive effect in model 1 and a negative effect in model 2. Neither indicator shows significant differences in models 3 and 4. There is thus little evidence supporting H2. How can we explain these results? Part of the answer is that regions are really not in the business of trying to outflank state interests at the EU level. Indeed, research finds little evidence of sub-state mobilization against national governments using EU channels (Moore, 2008). What is more, Marks et al. s (1996, p. 178) survey of regional lobbying offices found little or no confirmation of these lobbying offices being established on the basis of winning structural funds. This is not to say, however, that regional offices do not lobby for funds. Rather, as Moore (2008) argues, an important distinction needs to be made between politically powerful and politically weak regions. It seems that only the offices of politically weak regions tend to lobby for structural funds. While politically weak regions might even employ a dedicated EU funding officer in their lobbying

Regional authority, transnational lobbying and structural funds 15 offices, for instance, for politically powerful regions direct funds-seeking remains only a marginal activity (Moore, 2008, p. 526). Comparing results for regional lobbying offices and regional authority suggests that much of the lobbying for funds takes place primarily between regions and central governments over the contents of the NSRF and specifics about operational programmes. Lobbying offices of powerful regions based in Brussels are not concerned with winning funding. Conclusions This article has examined the determinants of structural funds allocation in the EU. Its point of focus was the role that regional authority played in this process while also testing an array of competing explanations. The central finding is that the greater a region s shared-rule, the more funding it receives. This was mainly the case, however, for convergence funds. This difference can be explained with regard to the amount of actual funding available under the convergence objective compared to the competitiveness objective. Results for competing explanations were mixed, showing differences in some models and not others. Nevertheless, it does appear that funds allocation can at least be partially explained as a function of regional party alignment with the national party and as compensating the losers of integration. Two central advances were made in this analysis. First, whereas existing studies used Lijphart s federalism index to measure regional authority this analysis used the regional authority index. This index not only provides authority scores at the regional level, but makes a more fine-grained distinction between different aspects of authority namely, shared-rule and self-rule. The second advance was the inclusion of indicators for regions Brussels-based lobbying offices. Combining data on these lobbying offices with regional authority was important for understanding how regions vie for funds within a multi-level lobbying framework. While shared-rule showed significant differences in the regression models, the lobbying capacity of regional offices did not appear to have an impact on the receipt of funds. This finding supports the idea that regional lobbying offices no longer prioritize lobbying for structural funds. This analysis has provided further support for the idea that being poor is not a sufficient condition for receiving structural funds. Where scholars have found evidence for this trend in previous funding periods, I have demonstrated a continuation of this trend in the 2007 13 funding period. The politics of EU regional policy do not only favour regions that are most in need of funds. The slow erosion of the EU s objective criteria for funds allocation has tipped the balance in favour of politically powerful regions. Indeed, much more so than need, regions with considerable political authority determine the allocation of structural funds in the EU. Correspondence: Adam William Chalmers Institute of Political Science Leiden University Wassenaarseweg 52 2333 AK Leiden The Netherlands email: a.w.chalmers@fsw.leidenuniv.nl