Mapping European Law

Mapping European Law R Daniel Kelemen and Tommaso Pavone Paper presented at the Biennial Meeting of the European Union Studies Association, Boston, MA, March 5-7, 2015 ABSRACT This article constitutes the first systematic effort to analyze the subnational practice of European Union (EU) law across time and space We derive eight empirically falsifiable hypotheses about the spatiotemporal structure of EU litigation from the extant literatures on European legal integration and policy diffusion and assess their ability to explain temporal and spatial variation in the use of the preliminary reference procedure by Italian courts We then provide a series of empirical assessments of these hypotheses We begin with a description of the spread of preliminary references across Italy from 1957 through 2013 We then conduct a series of statistical analyses by pairing EU litigation data with subnational socioeconomic indicators to assess which variables best predict cross-regional variation in preliminary reference rates over the past two decades (1997-2012) Finally, we leverage GIS technology and elementary spatial statistics tools to more rigorously assess the spatial structure or clustering in the use of the preliminary reference procedure Introduction Governance is a process that unfolds across space and time The degree to which a government can govern a polity in practice ultimately depends on the geographical extent and temporal consistency of its authority over its dominion Does the central government exert its authority over the whole of its territory or over only a fraction of it? Does the central government exercise authority on an ongoing basis, or only sporadically? These are fundamental questions for any political system, particularly for federal or multi-level political systems in which multiple tiers of government may seek to exercise authority over the same territory The “presence” of a regime in space and time is a thick concept that may be measured along multiple dimensions, with indicators ranging from the proliferation of its bureaucratic offices and officials, to the maintenance of military garrisons, to the collection of taxes, to social behaviors that indicate broad recognition of the regime’s authority Just which indicators are most relevant will vary depending on the nature of the regime in question Certainly, for regimes that seek to rule principally through law, measures of the reach of the regime’s judicial authority are crucial Scholars, European policy-makers, the European Court of Justice and the EU Treaties themselves all declare that the European Union (EU) is a community based on the rule of law To the extent that the EU succeeds in governing its member states and private actors, it does so principally through judicial mechanisms As one of us has argued elsewhere (Kelemen 2011), the EU relies heavily on a judicialized mode of governance – termed ‘Eurolegalism’ - that enlists private litigants and national courts to help pursue the EU’s policy objectives In the broadest sense, the EU can be conceptualized as what medieval historian Joseph Strayer (1970) called a “law-state” – a state constructed principally through the progressive expansion of state judicial institutions Existing scholarly literature tells us a great deal about the crucial role the ECJ has played in promoting the process of European integration and about the politics surrounding the construction of the EU’s legal system Pioneering studies by Stein (1981), Weiler (1991), Burley and Mattli (1993), Slaughter, Stone Sweet and Weiler (1998), Stone Sweet (2000, 2004), Alter (2001), and others have demonstrated the importance of the ECJ’s expansive jurisprudence and of inter-court cooperation and competition in promoting legal integration These scholars, and others such as Cichowski (2007) and Kelemen (2011) also have highlighted the role of litigation brought by private parties in the construction of the EU’s legal order The effectiveness of European law has always depended on the ECJ’s relationship with national courts who send it cases via the preliminary ruling procedure, recognize the supremacy of European law and apply European law and ECJ rulings domestically As work by Alter (2001) and others has shown us, convincing courts in the original six member states to recognize the supremacy of European law and to send cases to the ECJ via the preliminary ruling procedure was a vital early step in the process of European legal integration Since that crucial founding period, the ECJ, and more generally the EU, have worked to ‘socialize’ national judiciaries of new member states into the interlocking system of national and EU-level courts Today thousands of judges across the twenty-eight states of the European Union are trained in European law, participate in EU-related judicial networks and are encouraged to engage with the EU courts in Luxembourg Recent work by scholars such as Bobek For scholarly views see for instance Kochenov (2009) and Pech (2009) For a view from EU leadership, see for instance, European Commissioner for Justice, Vivian Reding’s 2013 speech on the rule of law (Reding 2013) For references in EU law see Lisbon Treaty, Article 1a, and the ECJ’s ruling in Case 294/83 Les Verts v Parliament [1986] ECR 1339, para 23 Also see Fukuyama (2012 p.271) and Berman (1983, p.406) on the concept of states built primarily through the progressive subsumption of pre-existing local legal orders into the new overarching legal order of the larger polity being created See also Kelemen (2014) conceptualizing the EU as a ‘law-state’ also see Strayer 1970, pp.26-31) State builders sought to supplant the plurality of legal orders and judicial institutions that existed within their territories with uniform systems of law, and they largely succeeded As Harold Berman (1983, p.406) (2008) and Wind (2010) highlights the ongoing importance of – and challenges associated with – securing the participation of the judiciaries of new member states in the EU legal system And yet, for all that we know about the importance of the EU law to the process of integration, we know very little about actual penetration of the EU’s judicial order across space and time To be sure, as we discuss below, some studies examine variations in the rate of references for preliminary rulings across EU member states Chalmers and Chaves (2012) examine variations in the frequency of litigation before the ECJ across various areas of EU law More generally, many ‘impact studies’ seek to assess the actual extent to which EU law is applied across various member states But for all the insight offered by this literature, it does not address fundamental questions about the actual reach of the EU’s judicial order across space and time, such as: Is EU law litigation concentrated around national capitals or in other particular geographic pockets, or is it dispersed throughout the territory of EU member states? How has the geographic reach of EU law evolved over time? Has it expanded steadily or followed a path- dependent pattern remaining clustered in particular locales? Answering these questions is vital because doing so will shed light on the true reach of the EU judicial order across the Union This paper seeks to address these questions by harnessing a methodology entirely novel to the field of EU law and politics and an original dataset of all preliminary references submitted by Italian courts to the ECJ from 1964 through 2013 We geocode the city location of every court that submitted a preliminary reference to the ECJ and use Geographic Information Systems (GIS) technology to analyze and explain the pattern of diffusion of these references across space and time The aims of this paper are modest in some respects and extremely ambitious in others Empirically, we limit our investigation to the study of one member state – Italy – and emphasize one indicator of the reach of the EU judicial order – the geographic location of Italian courts sending references for preliminary rulings to the ECJ The conclusions we draw from the quantitative analysis of this data remain tentative But while the immediate claims we make based on the paper are modest, the methodological turn and the theoretical ambition embodied in the paper – as well as the broader project of which the paper is a part – hope to offer a transformative perspective on the study of EU law We contend that the time has come for a spatial turn in the study of European law and governance To understand the reach of the EU’s judicial order and of EU law more generally, scholars should consider the spatial dimensions of legal integration that have heretofore been ignored while remaining attuned to temporal changes in the practice of EU legal rules First of all, considering the reach of EU law across space and time will provide a far richer picture of the reach of EU governance ‘on the ground.’ Secondly, this approach offers an added source of data and analytic leverage with which to A review of this voluminous literature is beyond the scope of this article; see Treib (2008) http://europeangovernance.livingreviews.org/Articles/lreg-2008-5/ for a review of this literature assess existing theories of legal integration For instance, as we discuss below, many of the arguments in the literature that seek to explain variations in rates of references for preliminary ruling across EU member states generate observable implications regarding the litigation of EU law across regions within member states Specifically, the same factors (i.e trade exposure, social capital, wealth, domestic litigation rates) that are thought to lead some member states to generate more EU law litigation than others should correspondingly be expected to lead some regions within member states to generate more EU law litigation than others By taking into account variation across regions within states, scholars can dramatically increase the ‘n’ in quantitative studies of EU law from 28 member states to the hundreds of sub-national regions within them, thereby gaining analytic leverage that will render quantitative analyses more robust Though this paper is limited to the study of preliminary references from one member state, it suggests important avenues for further research that we plan to pursue in subsequent work Indeed, the same methods we apply to Italy can be used to study the geographic dispersion of references to the ECJ over time within other member states Likewise, the use of geocoding and spatial analysis using GIS could be used to study other aspects of the spread of the EU’s legal order, including the proliferation across space and time of law offices specializing in EU law, law faculties offering degree courses in EU law, or national courts citing ECJ decisions or otherwise referencing EU law in domestic proceedings The remainder of the paper is divided into three sections Part I conducts a brief review of the existing literatures on European legal integration and policy diffusion and derives eight empirically falsifiable hypotheses about the spatio-temporal structure of EU litigation within Italy Part II then provides a series of empirical assessments of these hypotheses We begin with a description of the spread of preliminary references across Italy from 1964 through 2013 We then run a series of statistical analyses by pairing EU litigation data with subnational socioeconomic indicators to assess which variables best predict cross-regional variation in preliminary reference rates over the past two decades (1997-2012) Finally, we leverage GIS technology and elementary spatial statistics tools to more rigorously assess the spatial structure or clustering in the use of the preliminary reference procedure Our analysis uncovers several illuminating findings First, although lower courts were the first Italian judicial institutions to refer cases to the ECJ, high courts have a substantially greater propensity to refer Second, population levels and domestic litigiousness best explicate synchronic variation in preliminary reference rates since 1997, but there is evidence that the domestic litigation effect is causally heterogeneous We also find strong evidence that EU litigation is spatially clustered and exhibits important issue-specific variation that will be missed if EU law is treated as an undifferentiated corpus of rules Finally, Part III concludes I The Diffusion of EU Law: Literature Review and Testable Hypotheses A brief review of the literature on European legal integration and the diffusion of legal phenomena can serve as the basis to construct a number of empirically-testable hypotheses concerning the spread of EU law across space and time The first sizable wave of scholarship on European legal integration in the late 1980s and 1990s was catalyzed by the reinvigorated interest in explaining the origins and development of the EU.4 Spurred by the perceived jumps towards supranationalism represented by the 1986 Single European Act and the 1993 Maastricht Treaty, political scientists reawakened the dormant international relations debate between neofunctionalists, who highlighted the self-reinforcing “spillover” processes spurring regional integration outside state control (Haas 1968), and intergovernmentalists, who held that integration remained firmly under the control of member states, acting in pursuit of their self-interests (Hoffmann 1966) Yet increasingly, political scientists found it impossible to ignore the contributions of legal theorists such as Eric Stein, Hjalte Rasmussen, and Joseph Weiler, who highlighted the role played by the ECJ in spurring the integration process – particularly in periods when member states were reluctant to so (Stein 1981; Rasmussen 1986; Weiler 1986; 1991) Political scientists then developed a variety of political explanations of European legal integration focused on factors such as the ECJ's strategic agency, its interactions with national governments, its ability to forge allegiances with domestic courts to further legal integration, the challenges member governments faced in controlling ECJ activism, and functional forces that encouraged judicial empowerment Some scholars drew inspiration primarily from the neofunctionalist perspective (Burley and Mattli 1993, 1995; Stone Sweet 1999; 2000; Stone Sweet and Sandholz 1997) whereas others drew from a modified intergovernmentalist perspective (Geoffrey Garrett 1992; 1995) Still others built on the neofunctionalist perspective but framed their analysis in the language of historical institutionalism, which helped it to engage in broader disciplinary debates in comparative politics and international relations (Pierson 1996; Alter 1996; 2000; 2001) By the late 1990s debates organized along intergovernmentalist vs neofunctionalist lines seemed to exhaust themselves: On the neofunctionalist side, Slaughter and Mattli (1998: 179) wrote of their hope to close “the books on the neofunctionalist-intergovernmentalist debate” by promoting “a number of more specific studies on the role of individual litigants and national courts;'' On the intergovernmentalist side, Garrett, Kelemen, and Schulz (1998: 150) emphasized that they shared the neofunctionalists’ assumption that the ECJ is a “strategic actor” that is contingently able to “declare illegal national laws, regulations, or practices.” Just as the intergovernmentalist vs neofunctionalist debate over EU legal integration subsided, a new wave of scholarship began to shift focus towards explaining cross-national variation in the domestic demand for EU litigation Despite these studies’ country-level focus, they nonetheless suggest a There were of course a handful of earlier studies of EU legal integration from a social science perspective, such as those by Scheingold (1965, 1971) number of hypotheses capable of explaining variations in engagement with the EU legal order (via the preliminary ruling procedure) across the subnational regions of EU member states, to which we now turn Judicial Empowerment Weiler (1991) famously posited that the preliminary reference procedure was activated by the willingness of low-level domestic courts to litigate European law “The legally driven constitutional revolution,” argued Weiler, “was a narrative of plain and simple judicial empowerment [… as] the lower courts made wide and enthusiastic use of the Article 177 procedure [ thus] the constitutionalization of the Treaty of Rome [ ] meant an overall strengthening of the judicial branch vis-à-vis the other branches of government” (Weiler 1991: 2426) Burley and Mattli (1993) echo Weiler’s thesis: The reason why domestic judges began to refer cases to the ECJ is that “the E.C system gave judges at the lowest level powers that had been reserved to the highest court in the land” (Burley and Mattli 1993: 63) In a later article, Weiler further noted how the bourgeoning consumption of the preliminary reference procedure by lower courts spurred the gradual and reluctant acceptance of the ECJ’s supremacy doctrine by the French Conseil d’Etat, the Italian Constitutional Court, and the German Federal Constitutional Court (Weiler 1994: 521-524) Indeed, Stone Sweet and Brunell (1998) and Alter (2001) both emphasize that the delayed referrals by domestic courts of final appeal is consistent with Weiler’s claims, since “the logic of empowerment did not apply to all national judges, such as those sitting on constitutional courts” (Stone Sweet and Brunell 1998: 69) To this end, Alter (1998) develops an inter-court competition argument suggesting that lower level courts used the preliminary reference procedure to circumvent superior courts in their own national judicial hierarchies who might otherwise have overturned their rulings on appeal Despite the reiteration of similar arguments in later studies (see Conant 2002; Tridimas and Tridimas 2004; Bobek 2008), the judicial empowerment thesis has not found unconditional acceptance: Golub (1996: 360), for example, argues that it cannot explain the comparative reluctance of British lower court judges to refer cases to the ECJ For our purposes, although the judicial empowerment argument does not suggest particular hypotheses with regard to cross-regional variation, it does suggest that in broad terms we should expect lower level courts to be more active users of the preliminary ruling procedure than higher level courts within national judicial hierarchies Analysis of the courts that were the sources of preliminary references across Italy affords us with the opportunity to assess whether the judicial empowerment thesis, and suggests our first hypothesis: H1A: Lower court judges across Italian regions should be the most active consumers of the preliminary reference procedure Second, the judicial empowerment thesis also suggests that we should observe a particular pattern across time This argument suggests that in the early days of legal integration, bold lower court judges who embraced the opportunity to review domestic statutes in light of supreme EU law should have been the first judicial actors to participate in the preliminary ruling system, whereas higher courts should have only conceded to participation in the EU legal system later, after the pattern of references from lower courts had been established This causal narrative suggests a corollary to our first hypothesis: H1B: Over time, the share of Italian references coming from high courts should increase Trade In their pioneering study, Stone Sweet and Brunell (1998) posited that increased national levels of transnational economic exchange lead to increased demand for ECJ dispute resolution, activating a “virtuous” cycle Specifically, “levels of preliminary references will vary” posit Stone Sweet and Brunell, “across member states, court systems, and legal domain – as a function of the intensity of transnational activity,” where “intra-EC trade [serves] as a proxy for transnational exchange” (Stone Sweet and Brunell 1998: 75) Their finding of a positive correlation between the magnitude of a country's exports and its preliminary reference rates was later corroborated by Fligstein and Stone Sweet (2004), Tridimas and Tridimas (2004), and Carrubba and Murrah (2005) Importantly, the functional logic behind the argument linking increased levels of trade to increased rates of references from national courts to the ECJ should extend to the subnational level: The same causal logic that suggests that states that trade more should generate more references to the ECJ would in turn suggest that regions within states that trade more should generate more references This leads to the following hypothesis: H2: The greater the share of trade in a region’s economy, the higher the rate of references to the ECJ originating from courts in that region Population and Domestic Litigiousness In a recent paper, Vink et al (2009) posit that the strong correlation between trade levels and reference rates uncovered by Stone Sweet and Brunell (1998) suffers from omitted variable bias Specifically, “we suggest two more intuitive factors: population size and litigation rates First, bigger member states will make more references than small member states simply because there are more court cases as such and also more courts and judges that can refer questions Second, some states are more litigious than others, due to legal culture, the availability of systems of alternative dispute resolution, or legal aid” (Vink et al 2009: 13) The observable implication of this argument, then, is that any statistically significant correlation between trade levels and reference rates will disappear when controlling for population size and domestic litigation rates Vink et al (2009)’s country-level analysis finds empirical support for this hypothesis, where both population size and litigation rates exhibited a strong and positive effect on use of the preliminary reference procedure On the other hand, Golub (1996: 368-375) underscores that if population size is the primary predictor of preliminary reference rates, then British judges should refer substantially more cases to the ECJ than they in practice, and he further highlights that several qualitative studies have “found that differences in societal litigiousness and internal national judicial structure were either minimal or inconsequential” in explaining reference rates In this light, a more finegrained, subnational analysis may allow us to assess the validity of Vink et al (2009)’s claims, leading to the following two hypotheses: H3: The more populous a region, the higher the rate of references to the ECJ originating from courts in that region H4: The greater the overall rate of litigation within a region, the higher the rate of references to the ECJ originating from courts in that region Social Capital and Civic Engagement Cichowski (2007)’s analysis of domestic actors’ use of the preliminary reference procedure builds upon Epp (1998) to offer the important caveat that the litigation of EU law is contingent upon the plaintiffs’ ability to garner the necessary civil society support Underlying these arguments is a longstanding claim, popularized by Putnam (1994), that patterns of civic engagement and associational life bolster the effectiveness of democratic institutions Importantly, Putnam’s empirical study revealed substantial subnational variation in social capital levels across the Italian peninsula, which offers us an excellent opportunity to assess whether civic engagement may explain inter-regional variation in preliminary reference rates This leads to our fifth hypothesis: H5: The greater the levels of civic engagement within a region, the higher the rate of references to the ECJ originating from courts in that region Material Resources The resource mobilization approach to the study of social movements emphasizes that the mobilization of material resources is a necessary condition for social movement formation (McCarthy and Zald 1977; Jenkins 1983) Drawing on this approach, Epp’s (1998) comparative study of social activists’ ability to leverage law to promote progressive social change emphasizes the degree to which prospective plaintiffs are dependent on the existence of a “litigation support structure.” Cichowski (2007) incorporates this perspective within her analysis of EU law litigation and emphasizes that, in addition to the presence of social capital, material resources and organizational structures are necessary to support litigation In part, the concept of litigation support structures is based on the existence of civic organizations that can support litigation, and thus overlaps with measures of social capital and civic engagement discussed just above and captured in hypothesis But the concept of material resources also has a more purely economic dimension: Ceteris paribus, wealthier individuals should be better able to invoke EU legal provisions in court Indeed, the resource mobilization literature stressed the support of economic elites for the advancement of any social movement, for they possess the resources necessary to support a movement’s organizational and litigation strategies Conant (2002) and Tridimas and Tridimas (2004) make similar claims: “private litigants who have access to the required financial resources… will be privileged” in using the procedure (Tridimas and Tridimas 2004: 133) We believe that GDP per capita can serve as a plausible measure of the “resource stock” available to prospective EU litigants, and this provides the basis for our sixth hypothesis: H6: The greater the levels of GDP per capita within a region, the higher the rate of references to the ECJ originating from courts in that region Issue-Specific Variation One of the most illuminating claims of Stone Sweet and Brunell (1998: 88) that has yet to be systematically assessed is that intra-issue variation in country-level reference rates exists With respect to social provisions, for example, “UK judges refer far more cases per year to the ECJ than judges in any other member state,” despite the fact that five states – Germany, France, Italy, the Netherlands, and Belgium, in this order – refer more cases to the ECJ overall (Stone Sweet and Brunell 1998 88; 74) Sigafoos (2012: 504) corroborates this thesis by positing that the UK’s higher level of social-policy related references compared to France can be traced to the latter’s comparatively more generous “Bismarckian welfare state,” which is “very generous to insiders, who are seldom motivated to challenge the system.” While the organization of the welfare state is likely to engender variation focused at the country level, it is certainly plausible that analyzing the litigation of an undifferentiated corpus of EU law would mask important issue-specific variation at the subnational level For example, we might expect trade-rich regions to litigate EU common market provisions more frequently than EU social policy provisions, and we might expect the opposite pattern of litigation in unemployment-prone regions This logic suggests a seventh empirically-testable hypothesis: H7: The subnational use of the preliminary reference procedure will exhibit issue-specific variation across time and space Spatial Clustering Although the EU integration literature has yet to systematically engage with the burgeoning scholarship on the diffusion of judicial practices and public policy, we believe this is an opportune time to “bring diffusion in,” as it were On the basis of a survey of over thirty different diffusion mechanisms uncovered by the extant literature, Elkins and Simmons (2005: 35) conceptualize diffusion as “a set of processes characterized by interdependent, but uncoordinated, decisionmaking [ ] In other words, uncoordinated interdependence.” The authors are joined by Berry and Baybeck (2005) in uncovering two broad classes of mechanisms posited by the literature: (1) learning and (2) adaptation to a more competitive environment.5 Amongst public law scholars, these mechanisms have been discussed most extensively in the American literature With regards to learning mechanisms, scholars highlight that inter-court proximity, greater judicial professionalism/prestige, the presence of shared issue-specific knowledge, integrated legal reporting mechanisms, and common cultural linkages facilitate the diffusion of legal precedent and litigation via various forms of emulation (Canon and Baum 1981; Caldeira 1985; Moser and Turbek 2006; Linos 2006) Evidence of competition-induced legal development has been somewhat scarcer Shipan and Volden (2008) interpret their finding that American cities are reluctant to adopt antismoking laws if neighboring cities have not already done so as supportive of competition-induced diffusion, and Moser and Trubek (2006: 1471) broadly claim that “agents in competitive environments are more likely to innovate and imitate than agents facing weaker competitive pressures.” Linos (2006), however, leverages a case study of the development of American corporate law to argue that legal innovations not converge towards more efficient focal point solutions via competition but are instead spurred by learning-induced mimicry While scholars clearly disagree with respect to which underlying causal mechanisms best explain the spread of governance practices across space and time, most of the diffusion scholarship generates the same observable implication: Namely, that proximity facilitates the diffusion of policy and judicial practices The causal mechanisms posited to explain diffusion, such as competitive pressures, information Elkins and Simmons (2005) and Shipan and Volden (2008) also note that diffusion can be induced by a hegemonic power via coercion; we not consider this mechanism here as the lack of a supranational coercive apparatus in the EU precludes this from being a germane diffusionary mechanism or Cichowski (2007)’s emphasis on civic associationalism can be leveraged to explicate systematic subnational variation in EU litigation rates Fourthly, logged GDP per capita, which proxies the presence of the material resources necessary to support litigation, appears uncorrelated with EU litigation in all models except Model Perhaps the most interesting result, however, concerns the relationship between domestic litigation before regional administrative courts and EU litigation rates (as posited by H4) Although no significant relationship appears in Model when data is aggregated at the national level (excluding the Lazio region), this is a byproduct of causal heterogeneity engendered by the Italian north-south divide Specifically, Model reveals a positive and statistically significant (p