Tested and Found Wanting: A Quantitative Analysis of Legal Integration in the Member States of the 2004 European Union Enlargement

Literature Review

The theories reviewed for this analysis seek to answer two basic questions: "How [did] the ECJ [come] to be an authoritative legal and political institution, and how [was] an international rule of law – which actually works – created within the EU?" (Alter, 1996, 459). Historically, these philosophies build upon two of the founding theories of European integration – neo-functionalism and intergovernmentalism. Recently, though, scholars have begun integrating international relations and European legal integration theories, creating neo-realist and social constructivist proposals of why and how European legal integration occurred. Neo-functionalism focuses on the factors that precipitate integration; arguing that the self-interest of different groups within member states, international and supranational institutions, and domestic and multinational organizations precipitate action that propels integration forward (Jensen, 2010, 176-188). On the other hand, neo-realist suppositions of the integral role played by member states guide integovernmentalist integration theory. Intergovernmentalism asserts that though the individual states may pool their sovereignty and dictate powers to the EU and its institutions, the member states of the EU retain their national sovereignty because they act based on their perceived best interests (Cini, 2010, 86-103).

Alter's version of neo-functionalism contends that judicial empowerment of national lower courts, competition between constitutional courts and courts of first instance within member states, and domestic political interests transformed European legal integration (Alter, 1996, 471-481). In her qualitative approach, the three independent variables presented incorporate judicial interests at two levels of national judiciaries and domestic political concerns, both of which directly influence the number of Article 177 references presented to the ECJ (Alter, 1996, 458-481). Not only do these variables contribute directly to possible reasons why the number of member state references differs, they also suggest that any comprehensive comparison between member states' trends of the number of Article 177 references made to the ECJ must take place on a state level, rather than an aggregated, multi-state level. However, Alter's analysis fails to take into account the referrals made by national courts other than courts of first instance or constitutional courts, a problem which stems from her conceptualization of lower courts as only courts of first instance (Alter, 1996, 466). Another flaw in this theory is that the only actors considered are national lower and higher court judges, domestic politicians and ECJ judges.

An additional neo-functionalist theory championed by Jonathan Golub purports that the 'traditional model' of European legal integration – any integration model that contends that the role of national judges is to provide the ECJ with Article 177 cases (as in Alter's theory) – assumes that all member states will have a high rate of preliminary references and that national judges only have incentives to make referrals (Golub, 1996, 360-385). Through a case-study analysis of environmental law in Britain, Golub demonstrates that national judges actually face many political disincentives for referring cases to the ECJ (Golub, 1996, 361). These disincentives differ between member states and topics (Golub, 1996, 375-381). These disincentives, the political structure of each judiciary, and the amount of time the nation has been a member state effects the number of preliminary references from each state (Golub, 1996, 375-381). In spite of Golub's narrow case study, his chosen independent variables imply future research should focus analyses on the trends in topics considered in Article 177 references across member states and include the length of EU membership and the political structure of the member state as independent variables.

Stone Sweet and Brunell test Alter and Golub's theories through statistical analyses of the EU-15, utilizing a dataset of preliminary references they assembled over many years of research in conjunction with the research offices of the ECJ. They also propose their own theory that the variance in the number of Article 177 references is caused by differing levels of transnational activity among the member states. Their analyses conclude that cross-national theories that "emphasize the reciprocal impact of litigants and transnational exchange, EC law, and judicial rulemaking" are better supported by the dataset, completely dismissing Alter and Golub's neo-functionalist theories (Sweet and Brunell, 1998, 95).

Furthermore, Stone Sweet and Brunell rank European legal scholars, Walter Mattli and Anne-Marie Slaughter, 'judicial empowerment' thesis as a close second behind their own proposition (Sweet and Brunell, 1998, 95). Mattli and Slaughter pursue a synthesis of neo-functional and intergovernmental legal integration theory (Mattli and Slaughter, 1998, 177-209). First and foremost, Mattli and Slaughter contend that the core assumption of neo-functionalist and intergovernmental legal integration theory that a state is a unitary actor needs to be rescinded (Mattli and Slaughter, 1998, 184-185). Rather, the state is a mixture of sub- and supra-national actors (Mattli and Slaughter, 1998, 184-185). The second half of the article puts forth the theory that the interests of each of these actors, specifically those of national judges and individual litigants, and the constraints facing national courts in pursuing their interests – conceptions of judicial legitimacy, variation in national policy preferences and legal cultures and doctrines – all contribute to European legal integration (Mattli and Slaughter, 1998, 184-204). Despite Mattli and Slaughter's encouragement to perceive the state as part of a large network of both sub- and supra-national actors, their suggestion to operationalize the variables of actor interests and constraints through a detailed collection of the course of ECJ cases across a range of topics and from different member states corresponds more directly with in-depth qualitative case studies than my quantitative research design (Mattli and Slaughter, 1998, 206).

In direct contrast to Mattli and Slaughter's article is Geoffrey Garret, R. Kelemen and Heiner Schulz's neo-realist theory of European legal integration (Garrett, Kelemen and Schulz, 1998, 149-176). Garret, et. al., develop their theory around the core assertion that the ECJ and member states are rational strategic actors (Garrett, Kelemen and Schulz, 1998, 152). Through game theory analysis of three series of cases – agricultural imports, equal treatment of the sexes and state liability for violation of EU law – they demonstrate that the clarity of ECJ case law precedent, domestic costs to a member state of an ECJ ruling, the rate of activism of the ECJ, and the number of member state governments adversely affected by an ECJ ruling are all variables that influenced member state acceptance of European legal integration (Garrett, Kelemen and Schulz, 1998, 161-174). The problem inherent in this intergovernmentalist approach to European legal integration is that there are numerous other actors beyond the ECJ and the state, as noted by many other European legal scholars. Garrett published an article in response to his critics, which conceded the point that actors other than the member states and the ECJ are involved in European legal integration, but held to his belief that governments utilize cost-benefit analysis when accepting European legal integration (Garrett, 1995, 171-181).

Not to be deterred, Mattli and Slaughter published a critique of Garrett's response in which they agree with Garrett's theory, and refine their own theory of state preferences (Mattli and Slaughter, 1995, 183-190). They contend that the ECJ would not have advanced without the legitimacy derived from the national courts (Mattli and Slaughter, 1995, 187). In a later work, Mattli and Slaughter refined their neo-functionalist theory even further and specified the incentives and constraints to judicial empowerment (Mattli and Slaughter, 2007, 253-276).

The debate between Garrett and Mattli and Slaughter informs prospective research agendas by establishing the conceptualization of member states and the ECJ as amalgamates of interconnected groups, networks, and institutions, all with their own interests. Moreover, the further refining of both theories provides researchers with more concrete variables and concepts to test, including the national judicial preferences of judicial review, judicial competition and promotion of national policies that promote the policy interests of national judges (Mattli and Slaughter, 2007, 259-265). Kenneth Armstrong is another interdisciplinary scholar to incorporate international relations theory with his theory of European legal integration. He concludes that the independent variables espoused by institutional theory – the organization, procedures, and substantive and normative elements of national and European judicial systems – require researchers to reformulate their hypotheses to reflect the "differentiated and complex system of inter-linked regimes or networks" inherent in any given member state's judicial system, as well as in the ECJ (Mattli and Slaughter, 1995, 172). Armstrong's institutional integration theory imparts a specific independent variable that may have an effect on the trends of the number of Article 177 references made by member states: the structure and norms of both national and European courts. Still, the operationalization of these structures and norms presents an especially perplexing conundrum for researchers.

At the heart of Gregory Caldeira and James Gibson's social constructivist theory of legitimizing the ECJ is the assumption that a main component of legal integration is the acceptance of the ECJ as legitimate among the public (Caldeira and Gibson, 1995, 356-376). They build this theory through a quantitative analysis of independent variables that, they maintain, are typically not considered in legal integration theory (Caldeira and Gibson, 1995, 356). These variables include social class, unionized vs. nonunionized labor, gender, view of the EU, and how people perceive the law and knowledge of the ECJ (Caldeira and Gibson, 1995, 358-364). Caldeira and Gibson's independent variables demonstrate how to operationalize support for the ECJ as an independent variable, explaining the varied trends among member states. It also suggests that if a member state's population perceives the ECJ as legitimate, its number of Article 177 references will increase.

Concepts

The theories outlined above share five core concepts that are pertinent to my research due to their widespread use and incorporation across disciplines and theories of European legal integration: European legal integration, member states, disaggregate sovereignty and national courts. European legal integration is conceptualized as the synthesis of member states' national laws and legal systems with European Union laws as brought about by the constitutionalization of the Treaty of Rome by the ECJ through the adoption and promotion of the doctrines of direct effect and EU legal supremacy (Sweet and Brunell, 1998, 67). This integration has taken place within EU member states, defined as those countries that include Weber's "coercive monopoly of the use of force:" a bounded territory within which a population of people share a common culture, language and rights under a specific legal system, are governed by the same political institutions and international recognition, and have relinquished some of their sovereignty to the European institutions upon entry into the EU. The 2004 member states are those member states that joined the EU through the 2004 enlargement: the Czech Republic, Cyprus, Estonia, Hungary, Latvia, Lithuania, Malta, Poland, Slovakia and Slovenia. All of the theories of European legal integration tested in the forthcoming statistical analysis consider the member states to exhibit disaggregate sovereignty. I adopt Mattli and Slaughter's definition of this concept as "an image of different governmental institutions interacting with one another, with individuals and groups in domestic and transnational society and with supra-national institutions" (Mattli and Slaughter, 2007, 255).

Although Stone Sweet and Brunell conceptualize national courts in a three-tiered hierarchy – lower, intermediate and higher courts – that takes into account more actors than Alter's dual conceptualization – lower and higher courts – the limits of the publicly accessible information only permits the researcher to distinguish references between lower and high courts (Sweet and Brunell, 1998 and Alter, 1996, 471-481). Lower courts are conceptualized as only courts of first instance within national judicial systems. High courts are considered to be any national court that is not a court of first instance. High courts then include appellate courts, constitutional courts and courts of last instance.

Variables

As stated above, the variables considered in the subsequent analysis are drawn from the previously discussed theories of European legal integration as measured by the number of Article 177 references made by each 2004 member state. Thus, the dependent variable being measured in each of the theories is the total number of Article 177 references made by a member state from 2004 to 2010. The independent variables under consideration are the member states' national legal systems, judicial empowerment, publicly perceived legitimacy of each EU institution, and transnational activity. Unfortunately, many of the independent variables identified in the aforementioned theories are not plausible variables for this research project due to time and fiscal constraints, as well as the quantitative and generalizing nature of this project.²

Research Design

The research undertaken for this project is modeled on Stone Sweet and Brunell's quantitative analysis of the differences in the number of Article 177 references between the member states of the EU-15 (Sweet and Brunell, 1998). Their comprehensive analysis of various European legal integration theories constructs a sound methodological foundation upon which my own analysis is built. It draws upon different areas of European legal integration theory to create its tests and hypotheses, effectively testing the theoretical literature against empirical data and statistical analyses. It is through these tests that I hope to replicate and expand in order to determine which variables explain the variance in the number of Article 177 references made by those member states that joined the EU in 2004. The first theory tested is Armstrong's proposal that the type of legal system a member state possesses influences the number of Article 177 references the state will make in any given year (Armstrong, 1998, 155-174). The independent variable tested under this theory is the type of national legal system. To test this theory, all of the 2004 member states were divided into two groups: states with civil law systems and states with common law systems (coded as 1 and 2, respectively) (see Appendix One). , the total numbers of references made between 2004 and 2010 were tallied for each member state. The hypothesis is that states with civil law systems will request more references than those states with common law systems.

The second theory assessed is Alter's version of the judicial empowerment thesis (Alter, 1996, 458- 487). The independent variables found here are the number of preliminary references sought by lower courts and the amount of references coming from higher courts for all of the 2004 member states from 2004-2010. According to Alter's theory, the number of Article 177 references from the lower courts should be greater than the number from the higher courts.

Next, the alternate model of judicial empowerment proposed by Mattli and Slaughter is tested. The specific relationship being examined is the effect of the type of national legal system and the amount of competition between national courts on the number of Article 177 references (Mattli and Slaughter, 2007, 77-131). The type of legal system is operationalized as above in my evaluation of Armstrong's theory, and the competition between courts is measured as a function of the number of references made by low and high courts. The hypothesis posited by Mattli and Slaughter's theory is that if the total number of Article 177 references increase as the number of lower court references increase, and if the number of lower court references is greater in member states with civil law systems, then judicial competition and the type of national legal system both directly influence the number of preliminary references produced by a member state.

Fourth, Caldeira and Gibson's theory on the influence of public legitimacy of the ECJ on the total number of Article 177 references is also tested (Caldeira and Gibson, 1995). The independent variable investigated under this theory is publicly perceived legitimacy of the ECJ. It is operationalized by aggregating Eurobarometer data from 2004 to 2010 on the public's trust in the ECJ for each 2004 member state and comparing it to the level of public trust in other EU institutions. The Eurobarometer collects this information as the percentage of the people who answered "tend to trust," "tend to not trust" or "don't know" on the poll (European Commission). The hypothesis derived from Caldeira and Gibson's theory is that if a member state's population perceives the ECJ as legitimate, its number of Article 177 references will increase.

Finally, I evaluate the theory proposed by Stone Sweet and Brunell in their 1998 article: there is a direct correlation between transnational activity and the number of Article 177 references member states solicit (Sweet and Brunell, 1998). As per their article, transnational activity is operationalized as the amount of intra-EU trade and the number of direct actions and regulations passed each year by the EU.³ Stone Sweet and Brunell hypothesize that if intra-EU trade and the number of direct actions and regulations increases, then so too does the total number of Article 177 references made by all of the member states from 2004 to 2009.

My research seeks to test the above hypotheses derived from scholarly literature on European legal integration against the actual historical numbers and trends of Article 177 references for specific member states. A core assumption of my research is that states are disaggregate entities consisting of numerous actors, and that these national actors then interact with the international and supranational actors of the ECJ and other international organizations. This assumption directly effects the variables that I have chosen to test. Rather than only testing variables that measure national court to ECJ interactions, such as the number of lower and higher court references from each 2004 member state or the type of national legal system of the member states, I have chosen to include sociological and economic factors such as public opinion data and intra-EU trade measurements. I believe that including factors outside of strictly court-to-court interactions will provide a better analysis of possible factors contributing to the success or failure of European legal integration, as measured by the reference trends of each member state. I have chosen to focus on those member states that joined the EU through the enlargements of 2004 and 2007 due to the lack of scholarly research on the number and trends of the Article 177 references made by these member states, in addition to the time and fiscal constraints of this project.

Data Analysis

To assess Armstrong's theory, a comparison of means test was run, using the total number of Article 177 references from each member state over the period 2004 to 2010 as the dependent variable and the type of legal system (either civil or common law) as the independent variable. Table One shows the results of the comparison of means as created in SPSS. According to this analysis, those member states with civil law legal systems referred an average of 14.63 references between 2004 and 2010, compared to an average of only 1.5 references made by member states with common law systems over the same seven-year period.

These results speak directly to Armstrong's thesis. This analysis supports the above stated hypothesis derived from Armstrong's theory. Since the comparison of means proves that those 2004 member states with civil law legal systems referred over nine times more cases to the ECJ for preliminary ruling than those with common law legal systems, the hypothesis must be accepted that member states with civil law systems will refer more cases to the ECJ under Article 177 than member states with common law systems. Still, this test does not answer, and was not designed to answer, why this variance in the number of preliminary ruling references occurs between member states with differing legal traditions.

On the other hand, Alter's theory of judicial empowerment is not supported by the data collected by the ECJ. This data is presented in Figure One as a bar graph measuring the number of lower and higher court references for each member state from 2004 to 2009. The actual values for each year per member state are displayed in Table Two. Figure One signifies that Alter's theory only holds in two of the ten 2004 member states (the Czech Republic and Hungary). Thus, the hypothesis derived from Alter's theory of judicial empowerment, that the number of lower court references will be higher than those requested by high courts, must be rejected, at least as it pertains to the 2004 member states as an aggregate entity.

However, the data does suggest that Alter's thesis is valid in certain member states. This result poses a new puzzle for researchers: what conditions are present, or missing, in Hungary and the Czech Republic that cause lower courts to refer more cases under Article 177 than high courts? This is a puzzle worth exploring in the future through in-depth qualitative analyses that could only take place in conjunction with field research.

The alternative brand of judicial empowerment proposed by Mattli and Slaughter is partially represented by the comparison of means test previously discussed under Armstrong's theory (see Table One). Mattli and Slaughter argue that the number of Article 177 references is directly correlated to the type of legal system within a member state. As discussed in consideration of Armstrong's theory, there is a trend for member states with civil law systems to refer more cases, on average, than those with common law systems. This finding supports Mattli and Slaughter's argument, and the researcher accepts the same hypothesis as under Armstrong.

Another component of Mattli and Slaughter's theory is that it is expected that "the relationship between the European Court and the national courts [will] grow tighter over time, measurable in part by increasingly higher levels of art. 177 activity" (Sweet and Brunell, 1998, 69). As Stone Sweet and Brunell did for 1961-1994, I tested this theory by graphing the number of references for all of the 2004 member states over the time period 2004-2010 in Figure Two. The graph illustrates that, indeed, the total number of preliminary references does increase over time.⁴ This general trend is further demonstrated through an examination of each member states' annual references, as seen in Figure Three. Therefore, the data suggests that the ECJ and national courts have generally formed a closer relationship over the course of the seven years since the 2004 member states joined the EU. In order to test the theory advanced by Caldeira and Gibson that as the percentage of the public in each member state who perceives the ECJ as legitimate increases, so too will the number of Article 177 references. I compiled information on the amount of trust each member states' population had in the EU institutions over the seven-year period of 2004-2010 using data collected by the public opinion surveys of the Eurobarometer (European Commission). I then ran bi-variate regressions on the average trust, lack of trust and uncertainty all of the 2004 member states had for the ECJ, Council of Ministers, European Commission and European Parliament between 2004 and 2010.⁵ The results of this correlation test are seen in Table Three.

Even though Stone Sweet and Brunell originally tested Caldeira and Gibson's theory using a multivariate regression on diffuse support, population and GDP, this statistical avenue makes little sense for my study for two reasons (Sweet and Brunell, 1998, 76). First, Stone Sweet and Brunell employed the information found in Caldeira and Gibson's index of diffuse support as the data for their diffuse support independent variable (Sweet and Brunell, 1998, 76). Unfortunately, this index only contains information for the EU-15 (Caldeira and Gibson, 1995, 358-360). Thus, I had to rely on published Eurobarometer data, which bring the discussion to the second reason why a multivariate regression would not be appropriate in this case. Due to the multi-colinearity of the Eurobarometer polls, any R2 value formulated from this information would be artificially enflamed. In other words, public perception in one year is influenced by public perception the year before, which causes a falsely high percentage of the total number of Article 177 references to be explained by public legitimacy of the ECJ. However, by running a bi-variate regression, it is possible to determine the strength and direction of the relationship between the total number of preliminary references made by the 2004 member states between 2004 and 2010 and public legitimacy of various EU institutions by examining the resulting Pearson's-r for each regression.

As seen in Table Three, the Pearson's correlation value for each of the EU institutions is very close to zero. At best, there is a weak correlation between the total number of Article 177 references and public legitimacy of the EU institutions. Considering the ECJ exclusively, the Pearson correlation of the average percentage of the populations of the 2004 member states that trust the ECJ is -0.143. This indicates that there is a weak, negative relationship between the level of trust the populations of the 2004 member states have in the ECJ and how many references were made to the ECJ by these member states. This result directly contradicts the hypothesis derived from Caldeira and Gibson's theory. Instead of rising public legitimacy of the ECJ (measured as trust in the ECJ) increasing the total number of preliminary references, increased legitimacy actually slightly decreases the number of references. Therefore, the original hypothesis must be rejected.

Unlike the above analyses, the analysis for Stone Sweet and Brunell's transnational activity was completed using two different methods. This is due to the fact that the independent variables used by Stone Sweet and Brunell to measure transnational activity – intra-EU trade levels and the number of directives and regulations published by the EU each year – are on two different levels: intra-EU trade is on a state-by-state level, while directives and regulations measure the EU as a whole on a supra-national level. Hence, I first performed a bi-variate regression with the total amount of intra-EU trade for each 2004 member state as the independent variable and the total number of Article 177 references as the dependent variable (See Table Four). I then graphed the level of total intra-EU trade for the 2004 member states for each year and the number of EU directives and regulations for each year (See Figures Four and Five).

As recorded in Table Four, the Pearson correlation for the total intra-EU trade for the 2004 member states between 2004 and 2009 and the total number of preliminary references made by these same member states over the same time period is 0.340. This signifies that the relationship is fairly weak and positive. A positive relationship matches Stone Sweet and Brunell's hypothesis and findings, but the weak relationship contrasts with the strong correlation they find for the EU-15 (Sweet and Brunell, 1998, 75).

A comparison of Figure Four and Figure Two further indicates that Stone Sweet and Brunell's conclusions may not apply to the 2004 member states. Contrary to what these scholars hypothesize, the aggregate level of intra-EU trade of the 2004 member states (measured as the sum of the trade balances of the 2004 member states for each year in question) decreases between 2004 and 2007, while the total number of Article 177 references increases. However, from 2007 to 2009, the relationship is as predicted by Stone Sweet and Brunell's theory with the number of preliminary ruling references and total intra-EU trade increasing for the 2004 member states. On account of this incongruity between the patterns of the two variables, it is not possible to positively accept or reject the first portion of Stone Sweet and Brunell's hypothesis as it pertains to the 2004 member states.

Nonetheless, a side-by-side examination of Figure Two and Figure Four supports the second clause of Stone Sweet and Brunell's hypothesis. Figure Four depicts the annual number of EU directives and regulations. When compared to Figure Two, a clear relationship emerges. For the years 2004 to 2008, both the number of Article 177 references and EU directives and regulations rise. But in 2008, the number of directives and regulations decreases dramatically while Article 177 references continue to increase. Although, as noted in the previous analysis of Mattli and Slaughter's judicial empowerment theory, the number of preliminary references does drop significantly between 2009 and 2010. Unfortunately, data on the number of EU directives and regulations published in 2010 has not yet been made available to the public. So, it is impossible to tell whether the decrease in the number of preliminary ruling references is a response to the fall in the number of directives and regulations in the previous year.

In light of the limited information available (and the restricted modes of analysis thus available to the researcher), it is not possible to make a definitive decision about the relationship between the annual number of directives and regulations produced by the EU and the annual total number of Article 177 references made by the 2004 member states. Yet there is enough evidence to tentatively argue that intra-EU trade does not directly influence annual totals of preliminary reference rulings as much as the annual number of EU directives and regulations within the 2004 member states. In other words, based on my research, the level of intra-EU trade does not necessarily have a strong positive correlation with the number of preliminary references in the 2004 member states, but it is likely that the number of EU directives and regulations has a strong direct correlation with the number of Article 177 references made by the 2004 member states. Furthermore, it is likely that the number of EU directives and regulations has a direct correlation with the number of Article 177 references made by the 2004 member states, though the strength of that direct correlation is not yet possible to determine given the lack of data currently available.

Conclusion

The purpose of this research project was two-fold. First, as my research question states, it sought to determine what factors explain the variance in the number of preliminary ruling references made by the member states that joined the EU through the 2004 enlargement. This was accomplished through quantitative analyses of five of the leading theories of European legal integration modeled on the analyses published in 1995 and 1998 by Stone Sweet and Brunell (Sweet and Brunell, 1998). Second, it sought to determine if the theories that held true in the EU-15 remain valid or proved false for the 2004 enlargement states by comparing Stone Sweet and Brunell's findings for the EU-15 with the conclusions drawn in this project about the 2004 member states.

In regard to the latter of these two purposes, only two out of five conclusions matched those found for the EU-15. In both Stone Sweet and Brunell's analysis and my own research, Alter's theory of judicial empowerment was proven false (Sweet and Brunell, 1998). However, Mattli and Slaughter's version of judicial empowerment as a direct, positive correlation of the national legal system and level of judicial competition within member states on the number of preliminary ruling references made by a member states is demonstrated to be valid for both the EU-15 and the 2004 member states.

One theory where my findings differ from Stone Sweet and Brunell's conclusions is Armstrong's proposition that the type of national legal system will influence the number of Article 177 references. Stone Sweet and Brunell state that disparities in the type of national judicial systems in the EU-15 "have no systematic effect that is measurable by [their] data." (Sweet and Brunell, 1998, 73). Yet my analysis states that over nine times the number of references originate in member states with civil law systems (see Table One). Instead, I found that there is a very weak, negative relationship between public legitimacy of the EU institutions, not simply the ECJ, and the total number of Article 177 references. It is important to point out, though, that I was not able to run a multivariate regression like Stone Sweet and Brunell did to test this theory because the data used by Stone Sweet and Brunell is not available for the 2004 member states (Sweet and Brunell, 1998). The Eurobarometer data I utilized is not viable for a multivariate regression due to its property of co-linearity. This difference in data selection and analysis is a plausible explanation for our opposite conclusions. Further data collection using Gibson and Cladeira's index of diffuse support is needed before a difference in methodology and data can be ruled out as an explanation for the contradicting findings.

Finally, Stone Sweet and Brunell "found broad support" for their own transnational activity theory in the EU-15 (Sweet and Brunell, 1998, 95). Unfortunately, detailed data needs to be collected on the annual number of EU directives and regulations before it will be possible for me to positively say that the number of EU directives and regulations have a positive, direct influence on how many requests for preliminary rulings a 2004 member state makes. Also, there is the simple matter of time. The trends discussed and analyzed by Stone Sweet and Brunell occur over a period of thirty-four years, whereas my data only covers seven years. As time goes on and more data is collected in greater detail, I will be able to definitively conclude whether Stone Sweet and Brunell's transnational activity theory applies to the 2004 member states. At this time, my findings suggest that the amount of intra-EU trade a member state partakes in has little bearing on the annual number of Article 177 references made by that member state, while the annual number of directives and regulations published by the EU has a direct, positive causal relationship with the total annual number of preliminary references made by the 2004 member states.

Beyond the future research agendas already discussed, it would be incredibly beneficial for future researchers to focus on enhancing the dataset on preliminary references already started by Stone Sweet and Brunell (Sweet and Brunell, 2007). In particular, a systematic categorization of preliminary references by member state, year of referral, year of completion, subject matter, court of referral and name and type of the litigants (i.e. NGO, member state, individual, lobbying group) needs to be accomplished. Such a dataset, if it were kept up to date, would provide scholars with irreplaceable information that would reveal new dynamics of European legal integration and aid researchers in the testing of previously established theories in new situations.

For example, I could not test the theory that the interests of numerous actors, including national and ECJ judges, domestic and European politicians, NGOs and private businesses, that affect the number of Article 177 references because the cross-referenced data is not available publicly, and the only dataset that does this only has incomplete data (at best) after 1998 (Sweet and Brunell, 2007). With the continuing enlargement of the EU – geographically, politically, legally and economically – such a dataset would be invaluable to researchers.

In summary, I conclude that the variance in Article 177 references in the 2004 member states is not explained by the same factors as in the EU-15. I found extensive support for Armstrong's proposal and Mattli and Slaughter's version of judicial empowerment. Nevertheless, Alter's judicial empowerment thesis and Gibson and Caldeira's emphasis on public legitimacy for the ECJ fall flat against my findings. Stone Sweet and Brunell's transnational activity theory produces mixed results within my analysis. In order to discover the full explanation for the differing trends in Article 177 references within the 2004 member states, a dataset that categorizes preliminary references by numerous variables, including member state, subject matter and the type of litigants, needs to be created for the EU-27. Overall, the type of national legal system found within a member state, competition between national judges and the number of directives and regulations produced by the EU each year partially explain the inconsistency in Article 177 references requested by the member states that joined the EU through the 2004 enlargement.

It is my hope that the findings presented in this research will cause European legal scholars to rethink their approach of treating legal integration in each member state as a set, formulated process, and cause scholars and policymakers alike to develop a method and mode of legal integration for new member states that takes into account the individual needs and problems of each new member states' legal system and its relation to its citizens. Because of the unique nature of the European Union's legal system and its interaction with the national legal systems of its member states, I would caution scholars in applying the details of this research to legal relations within other quasi-supranational entities, such as the United Nations or the Association of Southeast Asian Nations. However, I would encourage legal scholars of other areas of the world to utilize the key point of my research: legal integration is not an uniform process for all member states, and, thus, supranational entities need to develop programs for legal integration taking into account the unique nature of each state's legal system and socio-political relations.

References

Alter, Karen J. 1996. "The European Court's political power" West European Politics 19 (1996): 458-487.

Armstrong, Kenneth A. 1998. "Legal Integration: Theorizing the Legal Dimension of European Integration," Journal of Common Market Studies 36 (June): 155- 174.

Caldeira, Gregory A. and James L. Gibson. 1995. "The Legitimacy of the Court of Justice in the European Union: Models of Institutional Support," The American Political Science Review 89 (1995): 356-376.

Cini, Michelle. 2010. "Intergovernmentalism," In European Union Politics, edited by Michelle Cini and Nieves Perez-Solorzano Borragan, 86-103. New York, NY: Oxford University Press.

Egan, Michelle, Neill Nugent and William E. Paterson. 2010. "Introduction: Researching the European Union." In Research Agendas in EU Studies: Stalking the Elephant, edited by Michelle Egan, Neill Nugent and William E. Paterson, 1-13. London, UK: Palgrave Macmillan.

European Commission. "Eurobarometer Interactive Search System," European Commission. http://ec.europa.eu/public_opinion/cf/index_en.cfm (accessed April 9-17, 2011).

European Court of Justice. 2010. "2010 Annual Report of the Court of Justice," European Court of Justice. http://curia.europa.eu/jcms/jcms/Jo2_11670/ (Accessed April 9-17, 2011).

Eurostat. 2010. "External and intra-EU trade – statistical yearbook," European Commission.

Garrett, Geoffrey. 1995. "The Politics of Legal Integration in the European Union," International Organization 49 (Winter), 171-181.

Garret, Garrett, R. Daniel Kelemen, Heiner Schulz. 1998. "The European Court of Justice, National Governments, and Legal Integration in the European Union" International Organization 52 (Winter), 149-176.

Golub, Jonathan. 1996. "The politics of judicial discretion: Rethinking the interaction between national courts and the European court of justice" West European Politics 19 (April), 360- 385.

Jensen, Carsten Strnby. 2010. "Neo-functionalism." In European Union Politics, edited by Michelle Cini and Nieves Perez-Solorzano Borragan, 176-188. New York, NY: Oxford University Press.

Mattli, Walter and Anne-Marie Slaughter. 1995. "Law and Politics in the European Union: A Reply to Garrett," International Organization 49 (Winter), 183-190.

Mattli, Walter and Anne-Marie Slaughter. 1998. "Revisiting the European Court of Justice," International Organization 52, 1 (1998), p. 177-209

Mattli, Walter and Anne-Marie Slaughter. 2007. "The Role of National Courts in the Process of

European Integration: Accounting for Judicial Preferences and Constraints." In The European Courts and National Courts – Doctrine and Jurisprudence: Legal Change in its Social Context, edited by Anne-Marie Slaughter, Alec Stone Sweet and Joseph Weiler, 253-276. Oxford, UK: Hart University Press.

Stone Sweet, Alec and Thomas L. Brunell. 1998. "The European Court and the national courts: a statistical analysis of preliminary references, 1961-95," Journal of European Public Policy 5, 66-97.

Stone Sweet, Alec and Thomas L. Brunell. 2007. Data Set on Preliminary References in EC Law, NEWGOV Project, Robert Schuman Centre, European University Institute (San Domenico di Fiesole, Italy).

The United States Central Intelligence Agency, "The CIA World Factbook," The United States Central Intelligence Agency. https://www.cia.gov/library/publications/the-world-factbook/fields/2100.html?countryName=Afghanistan&countryCode=af®ionCode=sas&#af (Accessed April 9-17, 2011).

Endnotes

1. The above research question is feasible in the sense that much of the information required is accessible through the internet. The social importance and scientific relevance of attempting to provide an explanation for the trends of the number of preliminary ruling references made by member states who joined in the 2004 enlargement is that such an explanation enters into dialogue with much of the already existing European legal literature, and the final product could easily produce research on the 2004 member states that has yet to be thoroughly researched and documented. Also, through a focus on only the 2004 enlargement, the data becomes manageable for an undergraduate researcher. Finally, the assumption inherent in this question is that the number of preliminary ruling references each member state submits to the ECJ form measurable trends.
2. The excluded variables discussed in the literature review are the incentives and disincentives for national judges to refer cases to the ECJ, the interests of numerous suband supra-national actors and the constraints facing these actors. Through further research, I hope to measure national judges' incentives and disincentives by examining whether judges are appointed or elected to their positions and how appointed judges are chosen for their positions, which will measure the level of political influence in this process. Also, as discussed in my conclusion, a dataset modeled on the one published by Stone Sweet and Brunell needs to be created and kept up-to-date. This dataset would include information on the subject matter of each preliminary reference and the type of litigants involved with each preliminary reference, among other basic information.
3. For a full discussion of how these variables were originally operationalized, please see Alec Stone Sweet and Thomas L. Brunell. "The European Court and the National Courts: a Statistical Analysis of Preliminary References, 1961-95" Journal of European Public Policy 5, no. 1 (1998).
4. Although the sharp decrease in the annual number of references between 2009 and 2010 does not strictly undermine Mattli and Slaughter's thesis, it should be noted and analyzed in conjunction with data from subsequent years, as the information becomes available.
5. The European Central Bank was not considered in my analysis because not all of the countries under consideration are members of the Eurozone.

Appendix

Appendix One: Code Book for Quantitative Analyses Done in SPSS

Appendix Two: Tables and Figures Table One: Results of Comparison of Means Test for Armstrong's Theory (Information used to categorize member states' legal systems found at: United States Central Intelligence Agency)