We have argued that integration processes are generally driven by transnational activity, and by the autonomous behavior of supranational institutions that respond to this activity. Our model suggests that governments do not control the integration process in a determinative sense. We do not want to be misunderstood. Governments are hugely important actors in European politics. Their impact on integration is positive when they: (1) work to adopt, at the supranational level, Euro-rules, and (2) faithfully transpose, on the national level, European directives into national law. Governments have a negative impact on integration when they: (1) block secondary legislation (in their capacity as members of the Council), and (2) refuse to comply with Euro-norms that they do not like (in their capacity as national governments), and (3) revise the Treaty to limit integration or reverse its effects (in their capacity as members of the BC's constituent assembly). The crucial empirical questions are: to what extent have the member states been able to block or reverse integration processes?; and to what extent have the member-states, acting as principals, used their authority to control its agents? These questions can only be answered by process tracing, on a sector by sector basis.

In this section, we cross-check our model, examining concrete outcomes, over time, in two domains of European law: the free movement of goods and the Europeanization of social provisions. We therefore shift the focus from the broad relationships depicted in the statistical analysis to a more detailed examination of the institutional politics. Our choice of these two sectors, and not others, requires a defense.

Generally, case selection is defended in terms of representativeness and variance. While we can hardly claim a representative sample (it is a purposive sample based on only two cases), we do argue that free movement and social policy are quite distinct from one another. By avoiding two similar cases, we lessen the probability that the patterns we find -- both descriptive and theoretical -- are due to specific factors not likely to be replicated in other settings. We make no claim that we can statistically generalize our results to all issue areas or sectors. But we have taken the minimal steps to assure that our results are not completely idiosyncratic.

A second justification for our selection of cases has to do with robustness, defined as the persistence of uniformities in the face of theoretically extraneous variables. By specifying our theory in a certain way, we are not implying that other variables do not operate. Air pressure and wind are extraneous to the theory of gravitational attraction, yet certainly affect the flight of a baseball. But the stronger the theorized forces (the causal variables) in relation to what is left out, the more robust the theory. By introducing variation in the cases examined, we consciously open the door to other variables too. The free movement of goods domain is about the removal of national barriers to transnational exchange, whereas the Europeanization of social provisions is about the construction of Euro-rules that replace national ones. Not surprisingly, social policy has also led to more member-state conflict than has the free movement of goods, precisely because it raises distributional issues, sets constituencies off against one another (tax payers vs. welfare recipients), and threatens the institutional interests of the welfare state. As a result, it is difficult to make efficiency-based arguments in this area. We might also expect to see differences in the role of EU institutions in these two areas. Precisely because of these confounding influences, our design is a bit stronger. To the extent that patterns are invariant (or similar) across different conditions, the theory is better supported.

The third justification relates to process-tracing in general. We thought it desirable to corroborate our statistical analysis with a microsectoral account of the operation of the legal system in two sectors. This rationale does not imply the selection of specific cases. But by focusing on two sectors, we hope to understand better the mechanisms at work in the integration process. Detailed process-tracing will allow us to "see" the aggregate patterns in different and

we hope complementary ways.

Process Tracing
Our differences with intergovernmentalist must begin with the challenge of explaining the constitutionalization phenomenon (for our explanation, see part II). Recall that references are triggered by litigants who claim that a member-state is in non-compliance with EC law and requests, nevertheless, that a national judge enforce that law. As discussed above, the Court, activated by these references, constructed the EC legal system: it worked to diffuse the doctrines of supremacy and direct effect, and to provide national judges with the means to respond to member-state decisions not to comply with Community law. Further, it cannot be claimed, as some have, that constitutionalization took place without the knowledge of the member-states. ₃₄ In fact, governments file with the Court legal briefs -- what are called "observations" -- defending the legality of their own, or any other member-state's, rule, and instructing the Court on how it ought to resolve the conflict at hand. Thus, in advance of any important decision, the ECJ is well informed of member-states preferences. This dynamic would appear to belie intergovernmentalist assertions, although intergovernmentalists might be able to fashion a post hoc explanation to preserve their approach.

In moving from the dictates of our model to qualitative analysis, we can sharpen our differences with intergovernmentalists. Congruent with our theory, we expect art. 177 litigation to be patterned in predictable ways. These expectations can be stated as propositions. First, other things being equal, references will target -- disproportionately -- those national barriers to transnational activity which hinder access to larger markets relative to smaller markets. Second, other things being equal, art. 177 references will attack those national legal regimes that enshrine the least integrative rules relative to those in place in other member-states. Put differently, those states that represented the lowest common denominator position in the Council, or have sought to downgrade the effect and application of Euro-rules within national law, will be targeted -disproportionately -- by litigation. We further predict that the operation of the legal system will require those states to adjust, in an upward direction, their respective national rules; they will do so in order to conform to the progressive development of EC law. Our general claim is, therefore, that the EC legal system functions to dismantle barriers to transnational activity in place in the dominant member-states, and to ratchet upwards lowest common denominator Euro-rules. If we are right, another proposition follows logically: member state preferences will not have a significant impact on judicial outcomes. These claims and predictions conflict, fundamentally, with intergovernmentalist expectations. And they are testable.

The EEC Treaty fixed a time table for the completion of the common market, and two political processes to achieve completion. The first process is known as negative integration: the obligation of all member states to dismantle barriers to free movement within EC territory. It is negative because states renounce their authority to regulate a range of economic transactions within their borders. States were obliged by the Treaty to progressively reduce and ultimately eliminate (by 1969) all import tariffs and quotas, for example. The second process is known as positive integration: the creation of new rules, or coordinative norms, to regulate problems common to all member states. Negative and positive integration were meant to proceed concurrently. Negative integration would erase whole classes of national rules and regulations, leaving important "holes" to be filled by EC legislation. The kaleidoscope of disparate national laws that functioned to hinder trade in 1958 -- such as taxes, duties, border inspections, and rules governing health, licensing, and environmental protection standards -- would be replaced by uniform, "harmonized" Euro-rules. Beginning in 1966, some harmonization would proceed according to qualified majority voting rules.

This is not what happened. As the deadline approached, the French provoked a constitutional crisis which was resolved only by the "Luxembourg Compromise" (January 1966). The compromise -- an intergovernmental understanding among the member states -- enables any government to demand, after asserting that "very important interests are at stake," that legislation be approved by unanimity rather than by qualified majority voting. Thus, any member state could veto secondary legislation that required Council approval. The Luxembourg compromise, in theory at least, shifted final lawmaking authority away from a qualified majority of states, and to the member-state(s) in favor of the least integrative outcome in any given domain.

Nevertheless, the ECJ sustained negative integration during the 1970-85 period. In its most important line of decisions, on the free movement of goods, the Court constructed a transnational legal framework capable of sustaining increasingly higher levels of cross-border trade. This case law generated a politics that ultimately resulted in the 1986 Single European Act (SEA). Among other things, the SEA reinstated qualified majority voting as the dominant legislative process for achieving the common market. Generally stated, our argument is that -- as the legal system systematically removed national rules and administrative practices that operated to restrict trade, and as transnational exchange soared -- unanimity rules became increasingly costly to maintain. The Court and national judges were busily dismantling national rules; traders were busy pressing for Euro-rules; and the Council of Ministers was too often unable to produce those rules due to the unanimity requirement. In the end, member-state governments adjusted their interests to the interests of transnational society. The member-states, as the Court's principals, did not reverse the acts of their agent, nor did they try to maintain the Luxembourg compromise. Instead, they ratified the Court's moves, and reduced the intergovernmental aspect of the Council of Ministers. We will briefly trace this historical process, from free trade to Treaty revision, beginning with the Treaty of Rome.

Arts. 30-36 of the EEC Treaty constitute the normative context for free movement. Art. 30 prohibits the member-states, after December 31, 1969, from maintaining "quantitative restrictions [quotas] on imports' as well as "all measures having equivalent effect"; art. 33 empowers the Commission to adopt, on its own, secondary legislation to clarify and enforce art. 30; and art. 36 permits exceptions to the art. 30 prohibition, on the grounds of public morality, public policy, public security, health, and cultural heritage. These provisions can, of course, be interpreted variously. What exactly are "effects" that are equivalent to quotas, and what types of national measures produce them? What national measures, otherwise prohibited by art. 30, could not be justified with reference to an art. 36 exception? For that matter, what exactly is a "public policy" exception? Pursuant to art. 33, the Commission sought to resolve questions like these in a 1970 directive. The directive established a "discrimination test": national rules that treat national goods differently than imported goods are considered to be measures prohibited under art. 30. Almost immediately, however, the Court's jurisprudence rendered the Council's directive obsolete.

The leading decision is Dassonville. ₃₅ In 1970, Mr. Dassonville imported Johnnie Walker Scotch Whiskey into Belgium, after having purchased it from a French supplier. When Dassonville put the scotch on the market, he was prosecuted by Belgian authorities for having violated customs rules. The rules prohibited the importation from an EC country, in this case France, of spirits that originated in a third country, in this case Britain, unless French customs rules were substantially similar to those in place in Belgium. Dassonville was also sued by an Belgian importer who possessed, under Belgian law, an exclusive right to market Johnnie Walker in the country. Dassonville argued that, under the Treaty, goods that had entered France legally must be allowed to enter Belgium freely, and exclusive rights to import and market goods were not legally valid.

The case provided the Court with its first real opportunity to consider the meaning of art. 30. Dismissing the objections of the UK and Belgium, both of which argued that such rules were not prohibited under art. 30, the Court found for Dassonville, declaring the following:

All trading rules enacted by the Member States, which are capable of hindering, directly or indirectly, actually or potentially, intra-Community trade are to be considered as measures having an effect equivalent to quantitative restrictions (emphases added).

The ruling posed a delicate policy problem for the Court. Ordering the wholesale removal of national regulations would strip away legal regimes serving otherwise legitimate public interests, such as the protection of public health, the environment, and the consumer. Further, given the Council's seeming inability to produce harmonized legislation in a timely fashion (partly a product of the Luxembourg compromise), this lack of protection could well become quasi-permanent. The ECJ resolved the problem by declaring that member states could, within reason, continue to regulate the production and sale of goods in the public's interest, pending harmonization by EC legislation. The Court stressed that: (1) the burden of proof rests with the member state to prove it has acted reasonably; (2) such regulations -- as with national measures justified under art. 36 grounds -- could not "constitute a disguised restriction on trade between member states"; and (3) the European judiciary would review the legality of these exceptions to art. 30 on a case-by-case basis. The decision thus laid the foundations of what became an ongoing, principled discourse about the relationship of Treaty law (and its inherent purposes) and national regulatory regimes.

The "Dassonville principles" have animated the Court's free movement jurisprudence to this day. They enable the EC legal system to monitor member state behavior, but also to shape national law by progressively elaborating the permissible exceptions to art. 30. ₃₇ More generally, virtually every important domain of negative integration is today governed by judicially-constructed rules resembling, in their broad outline and logic, the Dassonville principles. That is, the Court works to remove barriers to the movement of persons, or to the provision of services, while requiring member states to justify such barriers on public interest grounds. The Court then interprets these justifications in terms of the Community's, rather than the member state's, priorities.

The purely judicial construction of a common market on a case-by-case basis, although theoretically possible, would have been a slow and inefficient process. The (re)construction of the Treaty law governing free movement, however, exerted a profound impact -- a feedback effect -- on the European polity as a whole. In 1979, the Court ruled, in Cassis de Dijon, that Germany could not prohibit the sale of a French liqueur merely because that liqueur did not conform to German standards. ₃₈ While a seemingly straightforward application of the Dassonville principles, the Court also declared that it could not divine:

[any] valid reason why, provided that they have been lawfully produced or marketed in one of the member states, alcoholic beverages should not be introduced into any other member states.

After Dassonville and Cassis de Dijon, levels of free movement litigation rose sharply, rulings of non-compliance proliferated, and national regulatory frameworks were placed in a creeping "shadow of the law." At the same time, the Commission, in alliance with transnational business coalitions, worked to convert member-state governments to the idea that mutual recognition could constitute a general strategy for breaking intergovernmental deadlock. They were successful at doing so. The political science literature on the sources of the Single Act has sufficiently demonstrated the extent to which governments were dragged along in this process. ₃₉ Governments acted, of course, in the form of a treaty that codified pro-integration solutions to their own collective action problems. But these solutions had already emerged, out of the structured interactions between transnational actors, the Court, and the Commission; and they were given urgency by globalization and the failure of "go it alone" policies to sustain economic growth.

We have argued that integration is an inherently dynamic, expansionary process which serves, among other things, to construct and reconstruct the contexts in which governmental choices and intergovernmental bargaining takes place. The impact of the Court's free movement case law on the European polity illustrates the point beautifully. Intergovernmentalists would presumably tell this story differently, highlighting the putatively proactive role of governments.

Leaving aside how best to interpret history, our confidence that we have got the story right is enhanced by our success at predicting the outcomes in free movement cases. Table 3 depicts, quantitatively, crossnational patterns of litigation in the legal domains of free movement of goods and social provisions. ₄₀ To compile the table, we separated art. 177 references into categories corresponding to subject matter, assigning to each country the number of cases within each category that would be predicted by that country's total proportion of litigation expressed in percentage terms. We then subtracted, for each category and for each member-state, the predicted number of references from the actual number of references. The percentage highlighted in bold is the difference between the actual and predicted level of references for each country standardized, by percentage, in terms of the total number of cases in each column. Cell entries consist of positive and negative percentages: a high positive value indicates that litigants are attacking the rules of a particular country in a particular legal domain relative to other countries and other areas; a negative value indicates that a country is not being dragged to the ECJ as often as we might expect based on overall litigation rates relative to other member-states and policy areas.

Note that in the free movement of goods domain, accusations of German non-compliance dominate EC litigation. Of 670 references concerning the free movement of goods, 265 (40%) target German laws. This does not imply that Germany has been more protectionist than other member-states. It does mean that the German market is the key prize of free traders. Further, it means that the matrix of trade-relevant rules in place in Germany has provided the predominant context for the Court's construction of an integrative case law.

Does the Court, as intergovernmentalist like Garrett have argued, defer to the revealed preferences of the dominant member-states, and especially Germany? Kilroy, ₄₁ who explicitly adopts the intergovernmentalist framework and P-A imagery, analyzed 122 free movement goods rulings, and assessed the relationship between observations -- the briefs filed by the Commission and the member-states in pending cases -- and the ECJ's rulings. She found that in 81 decisions (2/3 of her pool), the Court struck down national rules as treaty violations, and that in 41 cases (1/3 of her pool), the Court upheld national rules as permissible under EC law. She further found that in 98 of 114 cases in which the Commission intervened, the Court sided with the Commission. The Commissions position therefore predicted the Court's decision 86% of the time. Member-state interventions utterly failed to predict the Court's rulings, and German interventions were found to be particularly ineffectual in generating outcomes. Following the logic of Garrett, Kilroy found it "surprising that Germany has a relatively lower impact on the Court." ₄₂ We are not surprised. In our model, for reasons clearly stated, the Community's supranational institutions -- especially the Commission and the Court -- function to lower the transaction costs of transnational activity, not to codify the preferences of dominant member-states.

Whereas negative integration results in the removal of barriers to integration, positive integration results in the construction of EC legal regimes that replace national ones. In positive integration processes, governments would seem to have enhanced means of controlling policy outcomes: a minority of Council members can veto secondary legislation; and, in consequence, lowest common denominator outcomes are common. If we restrict the empirical domain of positive integration to the process by which the Council of Ministers adopts secondary legislation, then governmental control of policy outcomes appears to be quite obvious. However, if we expand the empirical domain of positive integration to include what has gone on before and after Council intervention, then the Council's control of process and outcomes can only be an open empirical question. We will argue here, again, that the legal system both responds to and provokes the expansionary logic of integration processes. This dynamic undermines member-state control and enhances the power of private actors, national judges, and the EC1s supranational institutions within policymaking processes.

T.H. Marshall's classic definition of social policy is the use of "political power to supersede, supplement, or modify operations of the economic system in order to achieve results which the economic system would not achieve on its own." ₄₃ The definition does not make obvious that social provisions may involve policies that are unrelated to, or even counterproductive of, efficiency. For our purposes, we define social policy as the use of political power, institutional arrangements, and norms, to modify, extend, correct, restrain, and counteract the operation of the market. This definition makes it clear that social policy is not just a vehicle for correcting market failures.

That said, the formal treaty basis for social provisions (arts. 117-122) is thin: for the most part, it only establishes the competence of the Community's institutions to develop a social policy, without laying down content-based duties binding on the member-states. ₄₄ The exception is art. 119, which obliges the member-states to ensure that men and women receive "equal pay for equal work." In the negotiations leading to the Treaty of Rome, France insisted on inclusion of art. 119, in order to preserve its own equal pay for equal work rules, and to prevent social dumping. The rationale for art. 119, therefore, was related to market integration: labor is treated as a commodity and a factor of production; and member-states should not be able to obtain productivity advantages by allowing wage discrimination.

Economic arguments taken as given, art. 119 is also easily construed, as the Court did in Defrenne II(1976), ₄₅ as forming "part of the social objectives of the Community, which is not merely an economic union, but is at the same time intended... to ensure social progress and seek the constant improvement of the living standards and working conditions of [its] peoples." The Court, through the window of art. 119, envisioned a social Community. Three pieces of secondary legislation served to focus this vision: the Equal Pay Directive (1975), the Equal Treatment Directive (1976), and the Social Security Directive (1979). Each was adopted under unanimity voting rules. We note in advance that the operation of the legal system converted art. 119 and the directives into expansive, judicially-enforceable rights attaching to EC citizens, at a time when the concept of EC citizenship and EC rights was still in its infancy. Although we are examining the evolution of EC social provisions with other purposes in mind, this outcome marks an important shift in the complex relationship between market integration and the wellbeing of individuals that is now at the heart of EC politics.

We now turn to lawmaking and litigation in the area. It is a well documented fact that the UK government has constituted the crucial veto point in Council deliberations on the three major pieces of legislation that have been adopted. ₄₆ The Council adopted each under unanimity voting rules. Indeed, the UK government has never wavered in its intention, publicly declared, to veto any European proposal that would enshrine, in EC law, rules not already present in existing UK parliamentary statutes. Table 3 shows that litigation originating in the UK judiciary has driven the ECJ's docket in this area of the law. Fully 24% (40/167) of all references in the sector have attacked, as inconsistent with EC law, legal rules and administrative practices in the UK. The data confirm our proposition that litigants will attack national legislation that represents the lowest common denominator position on EC secondary legislation adopted by the Council.

We then tested our predictions concerning outcomes and the impact of observations made by the Commission and the member-state governments, by examining all of the Court's judgments pursuant to art. 177 references in the social provisions area, from 1970 (the date of the first reference) through 1992. ₄₇ Rulings were coded into one of two categories: either the Court had accepted a national rule or practice as consistent with EC law, or it had declared it to be in violation of EC law. We found that, in 91 judgments that could be unambiguously coded in this way, the Court made declarations of violation in 48(53%). The ECJ considered the lawfulness of UK practices in 24 rulings, declaring violations in 13. Aggregating results from litigation involving the "big four" -- France, Germany, Italy, and the UK -- the Court declared violations in 25 of 43 (58%) decisions. We also found, as Kilroy had in the free movement area, that the Commission's observations tracked results far better than did the observations filed by governments. The Commission's success rate is a whopping 88%: 73 of 83 observations predict the direction of the final ruling. The UK's rate of success was 58% (3 1 of 54 observations tracked final results). These numbers (tables 5 and 6 in the appendix) refute claims that the preferences of the most powerful member-states constrain the Court in a systematic manner. The legal system functions to push the integration project forward, not to give legal comfort to member-state preferences. ₄₈

A large body of scholarship exists on litigation of equal pay and non-discrimination, much of it focusing, for reasons that are obvious from the numbers, on the impact of ECJ's case law on the British legal system. ₄₉ This literature has shown that the ECJ has used its powers to expand the Community's jurisdiction over important areas of social policy, while shrinking member-state control. In the UK, British Conservative governments, who have staunchly opposed the Court's interpretation of the law in this area, have been forced by national court decisions to ask parliament, on successive occasions over the past fifteen years, to amend British statutes to conform to the ECJ's evolving case law. As these adjustments have been well documented, it would be more tedious than illuminating to trace all of them here. Instead, we will focus on how the UK's efforts to contain the development of EU law in this field have failed.

The UK has pursued a clear and consistent strategy in the Council of Ministers, one designed to anticipate and then seal off pathways that might lead to the preemption of national by supranational authority. It has systematically refused to support any secondary legislation containing provisions that would innovate, when compared with the state of existing British law, in the area of equal pay and non-discrimination. And it has written into directives detailed "exceptions" -- derogations to the proposed rules -- the purpose of which is to permit practices which, although lawful under British law, would be rendered, in the absence of the exception, unlawful under the proposed directive.

Despite these efforts, the Court has consistently interpreted the directives progressively, virtually obliterating the fire-walls erected by the UK and other governments. A few examples will suffice to make the point. During Council negotiations on the Equal Pay Directive, the UK insisted on writing into the Council's minutes its understanding of the relationship between art. I of the directive and British law, namely that the latter would conform to the former once the directive had been transposed into British law. Art. 1.1 of the Equal Pay Directive of 1975 calls for the "elimination of discrimination on grounds of sex" for "the same work" or "for work to which equal value is attributed." The provision employs language that is not contained in art. 119 of the Treaty, which speaks only of "equal work," and not "equal work of equal value." The relevant UK rules then in place defined "equal work" restrictively, as "like work," as when a man and a woman perform essentially the same job. Art. 1(2) then goes on to state that job classification schemes may be used to determine when "unlike work" might nonetheless be of equivalent value. British statutes on the matter recognized a woman's right to equal pay for work of "equal value," but only after, first, her employer had commissioned a study of the sources and effects of job classification in the workplace and second, that this study had determined that women were being systematically segregated into lower paying jobs. Under British law, however, women employees could not require their employers to commission such a study. In the Council, the UK worked to preserve this situation (pending changes in the British law that might be initiated by the British government itself).

In 1982, the Court ruled on the lawfulness of the UK rules. ₅₀ In its argument before the Court, the UK relied on a literal reading of the directive: according to art. 1.1, "equal value" must be "attributed" by some means; the only means to attribute equal value that is recognized in the directive is that which is found in British practice -- the evaluation of job classification practices); and, the directive does not confer a right on employees to demand -- or an obligation on employers to commission -- such an evaluation. Last, the UK argued that its support of the directive in the Council was contingent on the reading of the directive just given, and referenced the Council's minutes on the matter.

The Court rejected these arguments, stating that the situation under UK law "amount[ed] to a denial of the very existence of a right to equal pay for equal of equal value where no [evaluation of] classification has been made." The judges then declared that, in national law, "a worker must be entitled to claim ... that his work has the same value as other work and, if that is the case, to have his rights under the Treaty and the Directive acknowledged by a binding decision." We need to emphasize that, in phrasing its ruling in this way, the Court had effectively: (1) conferred a judicially enforceable right on individuals (to have the value of their work evaluated for the purpose of determining the existence of discrimination in pay), and (2) anchored that right in the Treaty (although the Treaty says nothing about equal pay for work of equal value). The "constitutionalization" of the provisions of the directive -- interpreting the provisions of a directive in terms of Treaty law -- is a common technique in this area of EC law. ₅₁ Among other things, this technique displaces the Council of Ministers as the site of reversal: to overturn this decision, and many others like it, member-state governments would have to reassemble as a constituent assembly, and then revise the Treaty. In this case, the impact of the Court was swiftly registered: the UK amended its law to conform with the Court's ruling. ₅₂

The Court has also found ways to bypass many of the more explicit fire-walls built into directives, including those exceptions which reflect important, and unambiguously stated, national interests. The Equal Treatment Directive, which laid down the principles of nondiscrimination on the basis of sex in employment, promotions, and working conditions, but art. 1.2 excluded from the purview of the directive pension and retirement schemes in the member-states, pending future EC secondary legislation on the matter. A series of decisions pursuant to art. 177 references originating in Britain eroded the legal consequences of this exception. In Marshall (a case decided in 1986, discussed further below), the Court was faced with an alleged violation of art. 5 of the directive, which extends equal treatment to "conditions governing dismissal." Ms. Marshall, a dietitian working for a British Health Authority, had been forced to retire at 60 years of age, whereas the mandatory retirement age for male employees was fixed at 65. The retirement deprived Marshall of certain pension benefits. In its decision, the Court announced that the provision excluding of "social security matters from the scope of [the] directive, must be interpreted strictly [i.e., narrowly], in view of the fundamental importance of the principle of equal treatment," thus relegating art. 1.2 to a position of inferior status. The ECJ then grounded its decision on art. 5: "the dismissal of a woman solely because she has attained the qualifying age for a state pension, which age is different under national legislation for men and women, constitutes discrimination on grounds of sex, contrary to [the] Directive."

In its post-Marshall case law, the Court went on to preempt the Council as the Community's legislator in this field. In 1990, in Barber, ₅₃ the Court enacted, as a matter of Treaty interpretation, a directive designed to abrogate those provisions of the Equal Treatment Directive and the Social Security Directive which permitted the member states to derogate from principles of equal treatment in the provision of old-age pensions. First proposed by the Commission in 1987, France and the UK had blocked its adoption in the Council. ₅₄ Ruling that certain types of pension benefits constituted pay, and that therefore the provision of these benefits must conform to the principle of non-discrimination proclaimed by art. 119 (EEC), ₅₅ the Court ordered the UK government to compensate Mr. Barber for pension payments lost due to discriminatory rules permitting women to retire earlier than men. In an extraordinary example of a court freely admitting the policy consequences of its own activism, the Court sought to mitigate what the member states had characterized, at oral argument, as the "serious financial consequences" retrospective compensation for all victims of such policies would pose for national budgets. "The member states and other parties concerned," the Court declared, "were reasonably entitled to consider that art. 119 did not apply to pensions" before its ruling, and that therefore only pending and future litigation would be subject to the new rules.

The preemption of the Council in Barber is not an isolated incident in the domain of equal pay and non-discrimination. ₅₆ In Dekker (1990), the Court enacted, by judicial decision. the main elements of the pregnancy directive that had been opposed by the UK. ₅₇ The case, referred by a Dutch court, was brought by a woman who had been denied a teaching position at a youth training center. The center's hiring committee informed Ms. Dekker, three months pregnant at the time, that although she was the best candidate for the job, they could not hire her because its employment insurance did not cover sick leave for illnesses that (a) would occur during the first six months of employment, if (b) these illnesses could be anticipated at the time of hiring. At issue was the interpretation of art. 2 of the Equal Treatment Directive prohibiting employment discrimination on the grounds of sex, marital or family status. Dekker, supported by the Commission, argued that the Center's decision constituted sex discrimination, since women but not men can become pregnant. In its observation to the Court, the UK argued that the directive requires only that men and women be treated equally when they "become unable to work for health reasons. The ECJ agreed with the Commission, and ordered national judges to apply national laws as if they distinguished pregnancy from "illness." In 1992, the Council adopted the pregnancy directive, catching up to the Court.

Finally, we illustrate one of our core arguments, namely, that the progressive elaboration of EU rules generates an expansionary dynamic tending to recast policy processes and outcomes at the national level of governance. C6nsider the combined impact of two very different sex discrimination cases on the work of British courts. As noted above, in 1986 the Court agreed that existing British law requiring women to retire earlier than men violated the Equal Treatment Directive, and ordered British judges not only to apply British law in conformity with the ECJ's equal treatment case law, but to compensate Marshall for her losses. Compensation, however, proved to be a difficult task. Although the British industrial tribunal assessed damages at 19,405 pounds plus interest, the British sex discrimination act set a ceiling for such awards at 6 250 pounds, and excluded interest payments. Upon reference of the matter by the House of Lords, the Court ruled, in Marshall II (1993), ₅₈ that EC law requires "full and complete compensation" (including interest), and that national rules that did not meet this standard be set aside by national judges.

In the second case, Webb, ₅₉ the ECJ declared that provisions of the UK's Sex Discrimination Act of 1975 must be construed by British courts so that they would conform to Dekker (see above). In 1987, Mrs. Webb was hired by an air cargo company to replace a female clerk who was scheduled for maternity leave. During her training period, Webb discovered that she was pregnant. The company dismissed her, on the grounds that she was no longer able to do the job for which she was hired. Webb lost a succession of appeals, including a unanimous judgment of the House of Lords. British courts distinguished between Dekker, which prohibited the dismissal of a woman simply because she was pregnant, and Webb's dismissal for being unable to fulfill her contract. Further, the Lords, backed in oral argument by the UK government, had determined that discrimination was not at issue, since if a male employee had been hired to replace a pregnant woman and then requested an extended leave of absence, he too would have been dismissed. The Lords nevertheless agreed to refer the matter to the ECJ. The European Court, in a terse decision, found for Webb, implying that her case had already been decided by Dekker.

The Dekker and Webb line of cases, coupled with the Marshall II requirement for full compensation, opened a floodgate of claims from women discharged from the British armed forces. In the 1978-90 period, some 5,500 women were discharged on the grounds that pregnancy made them unable to perform their duties. Before Marshall, following the government's understanding of EC law on the matter, ₆₀ the Defense Ministry began settling claims at 3,000 pounds per woman. In settlements reached after Marshall, courts awarded one woman 33,000 pounds plus pension rights, and an air force pilot 173,000 pounds. As of Spring 1994, 1,800 compensation claims were pending. ₆₁

In summary, lowest common denominator outcomes in the social provisions field, as fixed by directives adopted by the Council of Ministers and by national legal regimes controlled by member-state governments and parliaments, do not stick. The ECJ has interpreted directives broadly, in terms of their effects on individuals, as bearers of rights guaranteed under EC law. Member-state interpretations of these directives, to the extent that they would reduce the effectiveness of individual rights, are pushed aside. More dramatically, the Court has supplanted the Council as the locus of lawmaking on more than one occasion, enacting legislative provisions that had stalled in the Council under unanimity voting. Lacking the unanimity necessary to reverse the Court in this area, the member-state governments have been forced to adjust to the Court's case law, by ratifying the ECJ's policy choices in Council directives, and by revising national legal regimes.

This last point deserves to be nuanced somewhat. In response to the Barber decision, the member-states, for the first time ever, addressed themselves as a constituent assembly to a decision of the Court. In the so-called "Barber protocol," attached to the Maastricht Treaty on European Union, the member states declared that:

the direct effect of art. 119 may not be relied upon in order to claim entitlement to a pension with effect from a date prior to that of this judgment [the ruling in Barber, May 17, 1990], except in the case of workers ... who have before that date initiated legal proceedings or raised an equivalent claim under the applicable national law.

We believe that the EC legal system operates according to a generalizeable dynamic. Individuals ask national judges to void national rules in favor or Euro-rules within a particular domain of activity. The interaction between national courts and the ECJ recasts the law governing that domain of activity, and therefore recasts the policymaking environment. As new rules are generated and existing rules are reinterpreted, member-states, whose national rules are now out of step with EC rules, are placed "in the shadow of the law": they can expect to be the target of litigation until they act to comply with the dictates of the ECJ's case law. Further, that case law enhances the capacity of individuals to initiate future litigation, by providing potential litigants with more precise information about the content and scope of European law. In process tracing, we have seen, again, the self-sustaining dynamic at the heart of our model.

In emphasizing regularities in how the legal system operates, we do not mean to imply that market integration and the Europeanization of social policy have proceeded at the same rate. On the contrary, we have argued that transnational activity drives integration processes. If we are right, it must be that the construction of supranational governance does not proceed evenly. Instead, it must be that levels of integration vary, across sectors, as a function of differential rates of transnational exchange. It follows that the competence to govern in areas in which exchange across borders is relatively low, compared to areas in which exchange across borders is relatively high, will be more resistant to being transferred to the supranational level. If this is so, the EC will tend to respond better to the interests of manufacturers and investors, especially to holders of mobile capital, rather than the interests of labor unions and welfare recipients. Because of this, the EC will continue to rely heavily on existing welfare state regimes, operating at the national level, to supply the bulk of what we normally mean by social policy.

That said, considerable controversy exists as to how far the EC has departed from its market perfection project, whether that project is defined narrowly as exploiting Opportunities for economic exchange, or more broadly to include devising the "correct" rules to facilitate not only economic but also political exchange. Unlike free movement, which raises the prospects of large collective gains, social policy implies redistribution and market-limitations. We do not mean to say that free movement (of goods and productive factors) does not redistribute income, wealth, and jobs, among other things. But the efficiency logic of free movement is that a portion of the collective gains can be used to compensate the losers. The rationale for social policy rests partly on a logic of legitimation, which is tied more closely to merit goods or provision based on need. Market integration might only be tolerable to the extent that the social problems that it inevitably generates are mitigated by public policy.

Here we revisit our points of departure in this paper and then draw out the consequences of our research for the broad themes of this volume.

The paper lays out a theory of legal integration that does not itself rely on the P-A construct. Nevertheless, we chose to employ the framework for two reasons. First, the approach helps us to frame initial questions about how institutions are established and operate, and it provides auxiliary concepts (delegation, slack, control, oversight) that can be used to guide the formation and evaluation of hypotheses. Second, intergovernmentalists have used the P-A framework to organize their research on the institutional politics in the EC, and especially on the Court. In order to standardize the terms of the debate, it seemed useful to nest our theory within the P-A framework.

Let us recall some basics. The member states wrote a Treaty which created the ECJ to adjudicate disputes among member states over its proper interpretation. However, a great deal of discretion was given, even in the formal provisions of the Treaty itself. Judges are appointed for six years (and can be reappointed); their opinions are secret; votes are not published, and dissenting opinions are prohibited. Judges can be very independent, but the degree of their independence is not always directly observable. The Court's work is constantly monitored by the member-state governments, and governments regularly intervene in the judicial process in the form of observations. But the Treaty, not the governments, per se, generates the context for the Court's work. And, as we have seen, the Court interprets the Treaty and, therefore, the legal obligations of the member-states, expansively.

We have found that in legal integration processes, as Scharpf has argued more generally, the "status quo default position" increasingly favors pro-integration forces. ₆₃ Overturning Court judgments requires unanimous consent of the member states. The rules governing the interaction of institutions once the judicial process is in motion are different from the rules established by the initial delegation Finally, we should note that the activities of the ECJ are in some ways not effectively captured by a framework based on agency and control. While discretion, degree and kind of oversight, and compatibility of incentives between principals and agents provided useful points of departure, some of the dynamics of the Court outgrew this process. The Court's effective constitutionalization of the Rome Treaty and its revisions have created structural change reconstituting relationships among the ECJ, national courts, and private and public actors at the national and transnational levels that the treaty-makers did not anticipate. The actor-oriented nature of the P-A framework is not well equipped to capture the structural changes caused by the Court's role in Europe.

How do our findings relate to the Group's project? We have shown that movement from left to right, along the continuum (from pure intergovernmentalism to polity) has taken place in two quite different policy sectors; we have provided a theoretical account of why this movement takes place; and we have tested our theory quantitatively, and cross-checked it with a more intensive analysis at the sectoral level. Further, we have demonstrated that the intergovernmentalist understanding of the Court and of the operation of the legal system is deeply flawed, if not indefensible altogether. We will not recapitulate all of our arguments here. Instead, we will raise a number of crucial issues about the broader relationship between market integration and social policy that is implied by our research.

While the outcomes per se in free movement and social policy could be squared with highly general versions of intergovernmentalism and neo functionalism, when we examine the policymaking process in detail, intergovernmentalism immediately runs into problems. The ECJ displays high levels of autonomy in both sectors. Since social policy is more controversial and is more closely tied to domestically powerful welfare-state constituencies, intergovernmentalism should predict a more tightly reined-in Court in this sector. This is not what we find. Second, and related, the ECJ appears to be following a "logic of rules" oriented toward fostering international economic exchange. To be sure, this logic is more broadly applied in free movement than in social policy, since this is where the greater opportunities for international exchange exist. While the scope of the mandate is larger in free movement, the Court's activity in social policy is no less intense, nor less faithful to the Treaty. Thirdly, the Court is not setting down a pattern of jurisprudence that follows (or reflects) the positions of the most powerful states. Indeed, some of the most powerful states were on the other side of the Court's decisions most of the time. Thus, the predictions of intergovernmentalism are not supported by the evidence.

A second axis of debate, one not anticipated by our initial formulation, brings neofunctionalists and Keynesian social democrats into confrontation with one another. Since this controversy is not the central subject of this chapter, some clarification is in order. As we see it, our research has identified two separate dynamics. The first dynamic, on which we have focused in this paper, concerns the Court-led processes of adjudication and inter-institutional politics. Given the initial delegation of authority in the Treaty, how far and in which directions has the Court gone? The second dynamic "frames" the first by asking how important the activities of the Court are when assessed within the context of national policymaking, or put another way, in relation to the complete scope of activities on which the Court could, in principle, adjudicate. The ECJ could be highly activist and expansionary in its jurisprudence within a very narrowly circumscribed area. This distinction, between intensive and extensive jurisprudence, is important for understanding the second axis of debate.

The most fundamental criticism of the Community's social provisions centers on the limited aims of social policy and either the outright defeat or dilution of the major proposals in this area. Struck is quite right to remind us of "... the institutional limits for social policy at the European level." ₆₄ He points out that the Social Charter and the associated Action Programme (which proposed forty-seven items for the Commission 5 consideration) failed to say anything at all about social protection and collective bargaining. To Streeck, the lively jurisprudence of the ECJ in social policy amounts to intensive activity within a very restricted domain.

Let us try to put this argument into a broader context. As Hooghe and Marks point out, ₆₅ Keynesianism refers not just to a body of economic theory but also to a style of national decisionmaking that buffers domestic constituencies from the global economy. Thus, the defeat of Keynesian policies at the domestic level implies more than the decline of economic doctrine. It calls into question the social compromises underlying domestic governing coalitions. While most of the achievements of national welfare states have proven difficult to reverse, ₆₆ the freer movement of capital at the European level has not been balanced either by a regional reorganization of labor, or by the elaboration of a system of rules governing the relationship between citizens and markets at the transnational level. While many comparativists have written on the decline of corporatism domestically, ₆₇ few have related these "domestic" developments to the emerging structure of political authority at the European level. Streeck and Schmitter provide an important exception. ₆₈ Their analysis starts from the premise that domestic corporatist arrangements have been replaced by a looser, more fluid, transnational pluralism. Thus, centralized wage bargaining and concerted decisionmaking among peak associations of employers, workers, and consumers have given way to a more decentralized and at the same time transnational system of interaction among key social groups.

However, this emerging pluralism is not one where all groups are represented equally or even in proportion to their domestic weight. Highly mobile forms of transnational capital, because of their superior organizational capacity and ability to "exit" (to leave the national bargaining table), gain advantage at the expense of labor and owners of fixed assets. Therefore, what is entailed in the process of European integration is not just a technical shift in the location of production and commerce from local (national) units to the European level. Integration also involves a reconfiguration of class forces. In Streeck's view, the reason social policy is far behind free movement is that social policy is largely the concern of labor, the unemployed, and citizens with unclear locations with respect to the market. Until the fundamental forces having to do with the domestic weakness of labor change, social policy at the European level will be weak.

Streeck's analysis forces us to confront the following question: what is the appropriate reference point for judging EC progress? A Hobbesin anarchy would lead us to expect very little progress at all. Pervasive conflict, as well as the absence of trust and central (or common) institutions, leads states to a preoccupation with relative gains. Since social policy is inherently redistributional, cooperation in this sector never gets off the ground. Streeck starts from a very different baseline, these "domestic" developments to tthe historical achievements of fiscally powerful modern welfare states. ₆₉ If the measure of the EC's success is the degree to which it has duplicated, replaced, or internationalized the functions of mature national welfare states, the conclusion must be that it has failed or barely begun. And it is unlikely that the EC's modest movement can be explained by its embryonic nature, or "early stage" on the development continuum. A highly interdependent, institutionally thick conception of Europe provides us with an intermediate prediction. ₇₀ While Western Europe is not an anarchy, it is nevertheless characterized by the continuing power of nation states. We expect more cooperation than realism would allow, but less than if free movement and social policy were completely the creatures of nation states.

Our view is that it is unrealistic to expect the EC to replicate the achievements of existing welfare states. The evolving transnational system of social provisions -- EC social policy if you like -- represents an evolution in the political division of labor by which certain aspects of social policy have been internationalized. Both free movement and social policy are developing in ways pressed by the expansion of the transnational market but within the overall context of a decentralized international system. Free movement of goods and factors more obviously displays efficiency characteristics. It is easier for the EC to move ahead in this "sector". We have argued that social policy is different since it includes not only aspects that lend themselves to market perfection (e.g. the removal of competitive distortions that can lead to social dumping) , but also the provision of goods that imply redistribution on the basis of need or merit.

However, there is no a priori reason why social policies should always be thought of as "regional public goods" marked by non-rivalness, inability to exclude, and pervasive externalities. Thus, there is no inherent economic logic in centralizing all aspects of social policy at the European level. However, as we have tried to emphasize, we do not think the market is a limited institution. Thus, we don't find the phrase "social policy only tied to the market" very restrictive. The market is pervasive, intrusive, and disrupts along many fronts. As the market deepens, many corrections (not perfections) will be called for. We are in complete agreement with Leibfried and Pierson that "... the tidy separation between market issues and social issues" cannot be maintained. ₇₁ In Europe today, as in the development of national economies, the market is increasingly "an instituted process." Europe will continue to struggle to fashion a social policy, which the Court will continue to interpret expansively.

ECJ 1963. Van Gend en Loos, Case 26/62 ECR 1963: 1.

ECJ 1964. Costa, Case 6/64 ECR 1964: 585.

ECJ 1974a. Van Duyn, Case 41/74, ECR 1974: 1337.

ECJ 1974b. Dassonville, Case 8/74, ECR 1974: 837.

ECJ 1976. Defrenne (II), Case 43/75, ECR 1976: 455.

ECJ 1978. Simmenthal, Case 106/77, ECR 1978: 629.

ECJ 1979. Cassis de Dijon, Case 120/78, ECR 1979: 649.

ECJ 1981. Jenkins, Case 96/80, ECR 1981: 911.

ECJ 1982. Commission v. UK, Case 61/81, ECR 1982: 2601.

ECJ 1984. Von Colson, Case 14/83 ECR 1984:1891.

ECJ 1986. Marshall (I), Case 152/84, ECR 1986: 723.

ECJ 1990a. Barber, Case 262/88, ECR 1990: 1889.

ECJ 1990b. Marleasing, Case C-106/89 ECR 1990:1-4135.

ECJ 1990c. Hertz, Case, 179/88, IRLR 1990: 532.

ECJ 1991a. Francovich, Case C-6 & 9/90, ECR 1991: 1-5357.

ECJ 1991b. Dekker, Case 177/88,IRLR ECJ 1993. Marshall (II), Case 271/91, WLR 1993: 1054.

ECJ 1994. Webb, Case 32-93, ICR 1994: 770.

Beck, Nathaniel, and Jonathan N. Katz. 1995. "What to Do (and Not to Do) with Time Series Cross-Section Data." American Political Science Review 89: 634-47.

Burley, Anne-Marie and Walter Mattli. 1993. "Europe Before the Court: A Political Theory of Legal Integration." International Organization 47: 41-76.

Caldeira, Gregory A., and James L. Gibson. 1995. "The Legitimacy of the Court of Justice in the European Union: Models of Institutional Support." American Political Science Review 89: 356-76.

Caporaso, James A. 1996. "The European Community and Forms of State: Westphalian, Regulatory, or Post-Modern?" Journal of Common Market Studies (Spring).

Craig, P. P. 1991. "Sovereignty of the UK Parliament After Factortame." Yearbook of European Law 1991: 221-56.

Current Survey. 1994. European Law Review 19: 221.

Curtin, Deirdre. 1990. The Province of Government: Delimiting the Direct Effect of Directives. European Law Review 15: 195-223.

Dehousse, Renaud. 1994a. "Comparing National and EC Law: The Problem of the Level of Analysis." EUI WorkingPaper inLaw No. 94/3. Fiesole, Italy: European University Institute.

Dehousse, Renaud. 1994b. La Cour de justice des Communautes europeennes (The European Court of Justice). Paris, France: Montchrestien.

Deutsch, Karl W (et al.). 1957. Political Commlmity and the North Atlantic Area. International Organization in the Light of Historical Experience. Princeton, N.J.: Princeton University Press.

Ellis, Evelyn. 1991. European Community Sex Equality Law. Oxford, UK: Oxford University Press.

Fernandez Esteban, Maria Luisa. 1994. "La Nocion de Constitucion Europea en la Jurisprudecia del Tribunal de Justicia de las Comunidas Europeas" (The Notion of a European Constitution in the Case Law of the European Court of Justice). Revista Espanola de Derecho Constitucional 40.

Garrett, Geoffrey. 1992. "International Cooperation and Institutional Choice: The EC's Internal Market." International Organization 46: 533-560.

Garrett, Geoffrey, R. Daniel Kelman, and Heiner Schultz. 1996. The European Court of Justice: Master or Servant? Legal Politics in the European Union. Unpublished manuscript.

Gormley, Laurence. 1985. Prohibiting Restrictions on Trade within the EEC. North Holland, the Netherlands: Elsevier-TMC Asser Instituut.

Haas, Ernst B. 1958. The Uniting of Europe. Political, Social. and Economic Forces, 1950-57. Stanford, CA: Stanford University Press.

Haas, Ernst B. 1961. "International Integration: The European and the Universal Process." International Organization 15: 366-92.

Hervey, Tamara K. 1994. "Legal Issues Concerning the Barber Protocol." In Legal Issues of the Maastricht Treaty, David O'Keefe and Patrick M. Twomey, eds. London, UK: Chancery.

Hooghe, Liesbet, and Gary Marks. 1996. "The Birth of a Polity." Unpublished manuscript.

Iverson, Torben. 1996. "Power, Flexibility, and the Breakdown of Centralized Wage Bargaining: Denmark and Sweden in Comparative Perspective." Comparative Politics 28: 399-456.

Kenney, Sally J. 1992. For Whose Protection? Reproductive Hazards and Exclusionary Policies in the United States and Britain. Ann Arbor, MI: University of Michigan Press.

Kenney, Sally J. 1994. "Pregnancy and Disability: Comparing the United States and the European Community." The Disability Law Reporter Service 3: 8-17.

Kenney, Sally J. 1996. "Pregnancy Discrimination: Toward Substantive Equality." Wisconsin Women 's Law Journal 10: 3 51-402.

Keohane, Robert 0. 1989. International Institutions and State Power. Essays in International Relations Theory. Boulder, CO: Westview.

Kiewiet, Roderick, and Matthew D. McCubbin. 1991. The Logic of Delegation. Congressional Parties and the Appropriations Process. Chicago, IL: University of Chicago Press.

Kilroy, Bernadette A. 1996. Member State Control or Judicial Independence?. The Integrative Role of the European Court of Justice, 1958-1994. Unpublished manuscript.

Landfried, Christine. 1984. Bundesverfassungsgericht und Gesetzgeber (The Federal Constitutional Court and Legislation). Baden-Baden, Germany: Nomos.

Landfried, Christine. 1992. "Judicial Policymaking in Germany: The Federal Constitutional Court." West European Politics 15: 50-67.

Leibfried, Stephen, and Paul Pierson. 1995. "Semisovereign Welfare States: Social Policy in a Multitiered Europe." In European Social Policy. Between Fragmentation and Integration, Stephen Leibfried and Paul Pierson, eds. Washington, D.C.: Brookings Institution.

Lenaerts, Koen. 1990. "Constitutionalism and the Many Faces of Federalism." American Journal of Comparative Law 38: 205-64.

Levitsky, Jonathan E. 1994. "The Europeanization of the British Style." American Journal of Comparative Law 42: 347-80.

Mancini, Federico G. 1991. "The Making of a Constitution for Europe." In The New European Community, eds. Robert O. Keohane and Stanley Hoffmann. Boulder, CO: West View.

Marshall, T.H. 1975 (4th ed.). Social Policy. London. UK: Hutchinson.

North, Douglas R. 1981. Structure and Change in Economic History. New York, NY: Newton.

North, Douglas R. 1990. Institutions, Institutional Change, and Economic Performance. Cambridge, UK: Cambridge University Press.

Oliver, Peter. 1988. Free Movement of Goods in the EEC. London, UK: European Law Centre.

Pierson, Paul. 1993. "When Effect Becomes Cause: Policy Feedback and Political Change." World Politics 45: 595-628.

Pierson, Paul. 1994. Dismantling the Welfare State? Reagan, Thatcher and the Politics of Entrenchment. New York: Cambridge University Press.

Pillinger, Jane. 1992. Feminising the Market. London, UK: Macmillan.

Pollack, Mark. 1995. "Obedient Servant or Runaway Eurocracy? Delegation, Agency, and Agenda Setting in the European Community." Working Paper No. 95-10. Cambridge, MA: Harvard University Center for International Affairs.

Prechal, Sacha, and Noreen Burrows. 1990. Gender Discrimination Law of the European Community. Hants, UK: Gower-Dartmouth.

Sandholtz, Wayne, and John Zysman. 1989. "1992: Recasting the European Bargain." World Politics 42: 95-12~.

Scharpf, Fritz. 1988. "The Joint Decision Trap: Lessons From German Federalism and European Integration." Public Administration 61: 239-78.

Shapiro, Martin, J. 1980. Courts. A Comparative and Political Analysis. Chicago, IL: University of Chicago Press.

Shapiro, Martin, J. 1992. The European Court of Justice. In Euro-politics, ed. Alberta M. Sbragia. Washington, D.C.: Brookings Institution.

Shapiro, Martin J., and Alec Stone. 1994. "The New Constitutional Politics." Comparative Polilical Studies 26: 397-420.

Slaughter, Anne-Marie, Joseph Weiler, and Alec Stone, eds. Forthcoming. The European Court and the National Courts. Legal Change in Its Social Context.

Stein, Eric. 1981. "Lawyers, Judges, and the Making of a Transnational Constitution." American Journal of International Law 75: 1-27.

Steinmo, Sven, Kathleen Thelen, and Frank Longstreth, eds. 1992. Structuring Politics. Historical Institutionalism in Comparative Analysis. New York: Cambridge University Press .

Stimson, James. 1985. "Regression Across Time and Space." American Journal of Political Science 29: 914-47.

Stone, Alec. 1992. The Birth of Judicial Politics in France. New York, NY: Oxford University Press .

Stone, Alec. 1994a. "Judging Socialist Reform: The Politics of Coordinate Construction in France and Germany." Comparative Political Studies 26: 443-69.

Stone, Alec. 1995. "Constitutional Dialogues in the European Community." EUI Working Paper RSC No. 95/38. Florence, Italy: European University Institute.

Streeck, Wolfgang. 1994. "European Social Policy After Maastricht: The 'Social Dialogue' and 'Subsidiarity."' Economic and Industrial Democracy l5: 151-77.

Streeck, Wolfgang. 1995a. "From Market-Making to State-building? Reflections on the Political Economy of European Social Policy." In European Social Policy Between Fragmentation and Integration, Stephen Leibfried and Paul Pierson (eds.). Washington, D.C.: Brookings Institution.

Streeck, Wolfgang. 1995b. "Neo-Voluntarism: A New Europe Social Policy Regime?" European Law Journal 1: 31 -59.

Weiler, Joseph. 1981. "The Comrnunity System: The Dual Character of Supranationalism." Yearbook of European Law 1: 268-306.

Weiler, Joseph H.H. 1987. "The European Court, National Courts, and References for Preliminary Rulings - the Paradox of Success: A Revisionist View of Article 177 EEC." In Article 177 EEC: Experiences and Problems, eds. Henry G Schermers, Christiaan W.A. Timmermans, Alfred E. Kellerman, and J. Stewart Watson. North Holland, the Netherlands: T.M.C. Asser Institute.

Weiler, Joseph. 1991. "The Transformation of Europe." Yale Law Journal 100: 2403-83.

Weiler, Joseph. 1994. "A Quiet Revolution: The European Court and Its Interlocutors." Comparative Political Studies 26: 510-34.

Note *: Alec Stone Sweet, Dept. ofPolitics and Society, University of California, Irvine. Back.
Note **: James A. Caporaso, Dept. of Political Science, University of Washington. Back.

Note 1: Although the "European Union" is now commonly used to denote the European polity, we use "European Community" throughout the paper. Formally, the "European Community" remains the most inclusive term for how the organization functions most of the time. Back.

Note 2: See Kiewet and McCubbin 1991; and Moe 1987. Back.

Note 3: Pollack 1995. Back.

Note 4: Garrett 1992; Garrett, Keleman and Schultz 1996; and Kilroy 1996. Back.

Note 5: Garrett 1992, 556-59. Back.

Note 6: Logically, in this situation, it is the Community legislator (the Commission, the EP, and the Council of Ministers interacting in fixed procedures) that constitutes the Court's principal, not the Council of Ministers But, to the extent that the Council remains the crucial cog in the Community's legislative machinery, the view that the Council is itself the legislator, and therefore the principal, is only partly, if meaningfully, wrong. Back.

Note 7: Anticipatory reactions are part and parcel of any strategic context characterized by high levels of interdependence between two or more actors, such as that comprised by the principal-agent dyad. Back.

Note 8: North 1981,1991. Back.

Note 9: Haas 1958, 1961. Back.

Note 10: Deutsch 1957. Back.

Note 11: Landfried 1984, 1992; Stone 1992, 1994; and Burley and Mattli 1993. Back.

Note 12: Pierson 1993; and Steinmo, Thelen, and Longstreth, eds., 1992. Back.

Note 13: Lenaerts 1990; Mancini 1991; Shapiro 1992; and Weiler 1981, 1991. Back.

Note 14: Fernandez Esteban 1994. Back.

Note 15: Shapiro and Stone 1994. Back.

Note 16: ECJ 1964. Back.

Note 17: ECJ 1978. Back.

Note 18: ECJ 1963. Back.

Note 19: ECJ 1974a. Back.

Note 20: ECJ 1984. Back.

Note 21: ECJ 1990b. Back.

Note 22: ECJ 1991a. Back.

Note 23: Stone 1995. Back.

Note 24: Stein 1981. Back.

Note 25: We are grateful to Thomas Brunnel, a graduate student in political science at the University of California, Irvine, for his assistance in coding and analyzing this data. Back.

Note 26: Although most references are limited to a single subject matter of EC law, some references contain claims based on as many as five subject matters. This accounts for the difference between the total number of references and the total number of subject matters invoked in references in the data presented below. Back.

Note 27: Dehousse 1994; and Slaughter, Stone, and Weiler, eds., forthcoming. Back.

Note 28: Caldeira and Gibson (1995,364) have constructed an index of diffuse support for the Court of Justice based on Eurobarometer public opinion polls. We hypothesized that national litigants and judges broadly share attitudes toward the EC and the ECJ held by national populations, and that higher levels of diffuse support would result in higher levels of references. Diffuse support is positively correlated with our dependent variable (references), but weakly (r=.36, n=11). Back.

Note 29: We hypothesized that larger populations would generate greater numbers of references. Population (aggregated as the average population for each member-state, 1961-93) is positively correlated with references, but weakly (r=.36; n=11). See the pooled, cross sectional, times series analysis below. Back.

Note 30: We hypothesized that larger economies would generate greater numbers of references. Gross domestic product (aggregated as the average GDP for each member-state, 1961-93) is positively correlated with references (r=.63, n=11). See the pooled, cross sectional, times series analysis below. Back.

Note 31: Beck and Katz 1989; and Stimson 1985. Back.

Note 32: We expect that as the European polity matures, the litigation of EC legal disputes will increase. We do not expect that art. 177 references will continue to rise indefinitely. The capacity of the ECJ to process references is limited. We predict that national judges themselves will increasingly resolve EC legal disputes on their own, without a prior reference. For a discussion of the problem of art. 177 and the overloaded docket of the ECJ, see Weiler 1987. Back.

Note 33: We coded references by subject matter and relevant provision of the EEC Treaty. Each of the legal subject matters listed in the first column correspond to the sections of the EEC Treaty following in parentheses: agriculture (arts. 38-47); free movement of goods (9-37); social security (51); taxes (95-99); competition (85-94); approximation of laws (100-102); transportation (74-84); establishment (52-66); social provisions (117-122); external (a miscellaneous category including all EC economic policies affecting the European Free Trade Area. the CATT, food aid, and special agreements with non-EC states); free movement of workers (48-50); environment (13OR-T); commercial policy and dumping (110-116). Back.

Note 34: As Burley and Martli (1993) do. Back.

Note 35: ECJ 1974b. Back.

Note 36: See Gormley 1985, 22. Back.

Note 37: Gormley 1985; and Oliver 1988. Back.

Note 38: ECJ 1979. Back.

Note 39: Alter and Meunier 1994; Dehousse 1994b; and Sandholtz and Zysman 1989. Back.

Note 40: The table excerpts material reproduced in full in table 4 of the appendix. Back.

Note 41: Kilroy 1996. Back.

Note 42: Kilroy 1996, 23. Back.

Note 43: Marshall 1975. Back.

Note 44: Other Treaty domains touching on social policy boradly conceived include rights of establishment (articles 52-58), free movement of workers (articles 48-51), and the European Social Fund (articles 123-128). Back.

Note 45: ECJ 1976. Back.

Note 46: E.g., Pillinger 1992, 85-101. Back.

Note 47: We are grateful to Rachel Cichowski, a graduate student in political science at the University of California, Irvine, for her invaluable research assistance. We were forced to exclude data that could otherwise have been collected from ECJ judgements rendered in 1993 (14 cases); as of August 1996, the European Court Reports for 1993 were unavailable, having been recalled to correct for errors. Back.

Note 48: Relative to other European judiciaries, one would expect the UK courts to enforce EC law only with great difficulty (Stone 1995). The doctrine of parliamentary sovereignty formally prohibits judicial review of legislation, on any grounds, and doctrines governing the resolution of conflicts between treaty law and parliamentary statutes conflict with the ECJ's doctrine of supremacy. Both of these long-lived orthodoxies have been swept aside, in areas governed by EC law, as the UK judiciary has incorporated as national law the doctrines of supremacy, direct effect, and indirect effect (Craig 1991; Levitsky 1994). Back.

Note 49: E.g., Ellis 1991; Prechal and Burrows 1990; Kenney 1992, 1994, 1996; and Pillinger 1992. Back.

Note 50: ECJ 1982. Back.

Note 51: E.g., ECJ 1975; ECJ 1981. Back.

Note 52: The Equal Pay Regulations of 1983. For a fuller account, see Ellis 1991, 99-101. Back.

Note 53: ECJ 1990a. Back.

Note 54: Curtin 1990. Back.

Note 55: Article 119 proclaims the principle that men and women should receive equal pay for equal work. Back.

Note 56: We are aware of yet another example. In Hertz (ECJ 1990c), the Court enacted the main terms of the "burden of proof" directive, proposed by the Commission, and designed to place the burden of proof on the member states in certain cases involving maternity and sex discrimination. The directive had been blocked by a UK veto. Back.

Note 57: ECJ 1991b. Back.

Note 58: ECJ 1993. Back.

Note 59: ECJ 1994. Back.

Note 60: In Marshall I, the UK government had argued that EC law required only that compensation be possible under national law, not that it be flill and complete. ECJ 1986. Back.

Note 61: Current Survey" 1994,221. Back.

Note 62: The legal status of the Barber Protocol is ambiguous. Even if we accept that the protocol enjoys a rank equivalent to that of Treaty provision, the problem of how to resolve a potential conflict between the terms of the Protocol and conflicting Treaty provisions remains. The Court's ruling in Barber is based on an authoritative interpretation of art. 119 of the Treaty, and the Protocol may well enshrine a policy solution that is contrary to art. 119. For an analysis of these issues, see Hervey 1994. Back.

Note 63: Scharph 1988. Back.

Note 64: Streeck 1994, 152, 161. Back.

Note 65: Hooghe and Marks, 1996, 7. Back.

Note 66: Pierson 1994. Back.

Note 67: E.g., Iversen 1996. Back.

Note 68: Streek and Scmitter 1991. Back.

Note 69: Streek 1994, 1995a, 1995b. Back.

Note 70: Caporaso 1996; and Keohane 1989. Back.

Note 71: Leibfried and Pierson 1995, 44-45. Back.