Resumé

This article discusses factors effecting the implementation of European policies at the national level. Four main independent variables are distinguished: political institutions, the degree of corporatism, citizens' support for the EU, and political culture in the member states. The impacts of these variables on the success of implementing EU directives in ten policy areas are then tested with the multiple regression model. The results suggest that political culture and the design of political institutions in the member states had the most significant impact on implementation behavior. Countries with a high level of trust and political stability combined with efficient and flexible political institutions had the most success in implementing European policies.

Introduction

The aim of this article is to analyze the effects that political institutions, interest organization structure, political culture and mass opinion have on implementation rates in the twelve EU member states. This study has two main research hypotheses: first, it is presumed that countries with stable, effective political institutions and a corporatist system that integrates interest organizations into political decision making have the best capabilities to implement EU directives, and second that it is easier to implement EU directives in countries where the political system has high legitimacy, citizens are satisfied with democracy, the degree of social fragmentation is low, individual rights are highly respected, and the attitudes towards the EU are positive. This leads to a high overall willingness to obey the common rules.

Implementation matters are considered particularly important, also because they illuminate four very different aspects of European governance. First, they reflect some of the problems caused by the multi-level structure of European policy making. The EU has very limited chances to succeed in implementation unless it manages to persuade the relevant policy actors at the

---

* Risto Lampinen, University of Helsinki, Department of Political Science, PO Box 54. University of Helsinki, FIN-00014, Finland; and Petri Uusikylä, Ministry of Finance. Mariankatu 9, FIN-00170 Helsinki, Finland.

national and sub-national level to cooperate with its policy goals. Second, this cooperation is not possible unless all parties manage the complex policy networks or implementation regimes responsible for putting European policies into force. Third, the success of this process depends not only on the careful monitoring of lower level activities by the Commission, but also on the success of co-opting policy actors at earlier phases in the policy process. Fourth, in implementation, institutional factors of policy making play a major role. Institutions can be seen simultaneously as restrictions to successful implementation as well as opportunity structures that give certain individual actors or groups more possibilities to articulate their interests in the policy process.

Despite the purposeful attempts to develop the EU into a federalist type of political system, Euro-skeptics have often pointed out that the emerging shape of the European "non-state" or even post-Hobbesian state creates policy making processes that are too fragmented and dispersed and thus beyond the control of any single executive body (Schmitter 1996). It has been said that this highly sectorized multi-level system leaves much space for informal lobbying and bargaining and thus weakens the democratic control and guidance of European policy making (Andersen & Eliassen 1996).

It is apparent that the metamorphosis of the European arena also challenges academics to find new theoretical frameworks to complement the traditional neo-functionalism (see, e.g., Haas 1975; Keohane & Hoffmann 1991) versus intergovernmentalism (see, e.g., Waltz 1986) debate in order to better understand policy making in post-Maastricht Europe (Hix 1994). We believe that a policy approach combining traditional implementation analysis and the latest policy network tradition has potential to develop into the new paradigm of integration research. In this article, we concentrate more on the traditional implementation framework and treat policy networks only as exogenous factors that latently affect national as well as supranational policy making in the EU.

Decision making in the EU is not as simple as it would appear when we examine the official decision making process. As Kohler-Koch has noted, "throughout the different phases of agenda-setting, policy formulation, implementation and evaluation, issues are dealt with sometimes even simultaneously at the national, sub-national and supranational level" (Kohler-Koch 1996, 182). Depending on the issue and the decision making procedure, the arena shifts from the member states to the EU level and back again. At the European level, national administrations are represented during the whole preparation and decision making process, and they work closely with Union representatives. National interest organizations and other private actors who want to be aware of what is going on and make their voice heard have to be present at both European and national level. These levels and their continuously changing structures are by no means separate. Rather, they are

linked through complicated transnational networks, which tend to break up
established national interest intermediation structures and form new
transnational clusters (Schneider & Werle 1992).

In assessing the effectiveness of a political decision making system or of any other organization, it is essential to look beyond its record in law making to the question of how successful it is in enforcing those laws. Law making is an essential achievement, but it can also be ineffective if it lacks the machinery needed for enforcement (Harding & Sherlock 1995, 75.) The question of how to make actors at the lower levels (national and sub-national) observe and implement the results of political decision making, i.e., directives and regulations, is central in the governance of the EU. This is an especially important question, because nearly all EU laws have to be implemented at the national or sub-national level. Problems in implementation of EU laws have been present throughout the history of the European Community and the completion of internal markets has made the problem even more acute (Ekström 1994, 5).

Implementing European Policies

In the EU, it is necessary to distinguish between two types of national implementation procedures, namely legal implementation, which means the incorporation of EU legislation into national law, and final implementation, which means the administrative operations at the practical level, i.e., the actual application of EU laws (From & Stava 1993, 60). Since legal implementation is a prerequisite for actual implementation, it has been emphasized in the Union.

The Commission monitors application and legal implementation of Community Law in the member states and issues an annual report on the subject. Although the member states have made great efforts to improve implementation, unimplemented directives remain a serious problem. The Commission has even considered unimplemented directives the main obstacle to efficient enforcement of Community Law (Ekström 1994, 5).

Presumably, there are even greater problems in the final implementation. The final implementation is, however, hard to examine because there is no data available on it, and consequently this empirical study on application of Community law in the member states concentrates on legal implementation. Final implementation will be the subject of further research.

The reasons there are so many failures in implementing EU laws have been explained in several ways. First, some explanations emphasize the complexity of European policy formulation and the EU's dependence on member states when it tries to get Community laws implemented. Second, the role of sub-national governments, which are crucial actors in the actual implementation,

is seen as being too passive. According to this point of view, if these subnational governments were to become more integrated in the policy formulation phase, implementation would be remarkably facilitated. The Committee of Regions established by the Maastricht Treaty is one step in this direction (From & Stava 1993.) Third, it is believed that one of the main reasons for the "implementation deficit" lies within the nature and capabilities of EU institutions and their officials, especially within the Commission. Ludlow has written (1991, 107) about the bias within the Commission, the main implementor in the EU, in favor of policy formulation as opposed to policy execution. Most of these factors could be treated as statutory variables affecting the preconditions of the implementation process. In our empirical analysis, we shall complement this rather static approach by adding some non-statutory socio-economic conditions and public support variables.¹

Failures in Implementing Community Law in Member States

The number of EU directives in force has increased remarkably throughout the 19905. As Table 1 shows, the single most significant event that has increased the number of directives is the completion of the single market at the end of 1992, which caused an especially steep increase in directives concerning the removal of technical barriers. After 1992, there have been almost as many single market directives as all other directives put together. Most of the other directives concern either agricultural or environmental policies. The number of directives in force varies from year to year, because some directives have had fixed time periods, whereas others have been revoked by new directives or other regulations.

Table 2 shows the annual number of unimplemented EU directives in member states. If all categories of unimplemented directives are considered, Denmark is clearly the most successful implementor, and The Netherlands and The United Kingdom the second most successful countries. Implementation of directives has faced the biggest problems in Italy, Greece and Portugal and Belgium. However, it has to be mentioned that there is less variance among countries with regard to single market directives than other directives. This indicates that the differences in implementation rates among member states are mainly due to directives other than those related to single market.

The Enforcement of Unimplemented Directives

The formal preconditions for legal actions against a member state that has not succeeded in implementing EU directives will be briefly presented in order to help understand the context and the contents of the data used in the empirical part of this article. According to the Treaty on European Union, Article 169,

If the Commission considers that a member state has failed to fulfill an obligation under this Treaty, it shall deliver a reasoned opinion on the matter after giving the State concerned the opportunity (letter) to submit its observations. If the State concerned does not comply with the opinion within the period laid down by the Commission the latter may bring the matter before the Court of Justice.

Article 170 states that,

a member state which considers that another member state has failed to fulfill an obligation under this Treaty may bring the matter before the Court of Justice. Before a member state brings an action against another member state ... it shall bring the matter before the Commission . . .

After this, the action will follow the same procedure described in Article 169.
According to article 171,

... if the Court of Justice finds that a member state has failed to fulfil an obligation under
this Treaty, the State shall be required to take the necessary measures to comply with the
judgement of the Court of Justice. . . .

According to these articles of the Maastricht Treaty, the Commission, the Court of Justice or another member state can bring action against a member state that has failed to fulfill the obligations of the EU law. Because the Treaties of Rome and Maastricht remain silent in many areas, the ECJ has often been called upon to fill the gaps. This was the case when the Court explicitly decided in the van Duyn case that directives were capable of producing direct effect and that an individual citizen was entitled to bring action against a member state on the basis of a directive, even if it had not been implemented. As a result of the van Duyn case, unimplemented directives are also a source of individual rights (Ekström 1994.)

On the basis of elaborate subsequent jurisprudence of the ECJ, citizens of member states have been provided with a number of tools for securing their rights. In addition to direct effect, these include the duty of courts and administrative authorities to interpret national law in accordance with EC liability, and the liability of the member states for not fulfilling their duties in the implementation of directives. This "decentralized" aspect of ensuring compliance will not, however, be examined in this article.

Table 3 is divided into three time periods. The first period covers time before the Single European Act. The second period covers the time when the single market was "under construction," and the third describes the situation after the completion of the single market. The table shows clearly that the number of letters and reasoned opinions have steadily increased from 1983 to 1995. In contrast, the number of referrals to the Court has decreased during the same period. The combination of these facts suggests that countries tend to neglect the demands of the EU legislature more often. However, countries try to avoid serious conflicts with the Union.

As for individual countries, it seems that countries like Greece, Italy and Portugal have had the most difficulty implementing EU legislation on time, whereas The Netherlands, the UK and especially Denmark have obeyed EU demand most conscientiously. However, it must be kept in mind that Spain and Portugal became members of the EC as late as 1985, and Portugal in particular has had long transitional periods in several sectors which is reflected in the number of referrals to the Court. All in all, overstatements should be avoided when interpreting these figures, since implementation failures are caused by multiple factors affecting both countries' ability and willingness to comply with EU rules. In the following, we shall try to elaborate on these two components behind implementation failures.

The Conditions of Successful Implementation

Implementation failures cannot be understood by looking only at institutional political factors or at the overall EU support in member countries. Implementation success is more likely to be the result of more latent cultural factors and multi-actor bargaining processes within the member states. Although policy networks are embedded in the institutional setting, they function and operate according to their own dynamics.

The process of implementing European policies at the national level has been rather problematic, suggesting that there is still much to learn about how to design effective implementation structures. Conflict resolution theories suggest that the likelihood of cooperative behavior in the implementation process is linked to the degree of conflict of interests between participants in a collective decision (Stoker 1991). Next we shall briefly discuss the role and impact of some contextual factors on the conditions of successful implementation. However, one restriction has to be made clear at this point: since we have no data on implementation networks or on the interests and strategies of implementors in each member state, we can only compare some of the general institutional factors of each state and analyze the association between the preconditions of effective implementation and the success of implementation. The usefulness of this analysis is that it helps us to determine what kind of context or mechanisms promote cooperation and policy coordination. It should also serve as the first step in creating an understanding of which institutional factors have the greatest influence on the structure and functions of policy networks.

The analytical framework of our implementation analysis is presented in Figure 1. It sets out the four factors which we contend are the key elements in creating conditions for cooperation and policy coordination in member states, and which thus help to explain failures in implementing European policies. The main independent variables in our model are: 1) the political culture of the nation state, 2) mass opinion and attitudes towards the EU in member states, 3) the role of interest groups in the state, and finally 4) the overall political system or political institutional settings in member states. All these factors affect strategic behavior, either by creating opportunities by putting restrictions on coordination and policy cooperation. Thus successful implementation depends on two different conditions: the willingness of relevant policy actors to cooperate (the behavioral dimension) and the capability to implement policies once the overall consensus is reached (the technical dimension). We assume that political culture and mass opinion primarily affect willingness to implement, and that the design of the political system (or institutions) affect the technical capacity to implement. Interest group structure falls into both of these categories, as it can be seen as being part of the formal governance structure (for instance in the role of collective bargaining) and it contains behavioral elements as well (for instance informal lobbying and other modes of pressure politics).

Here we refer to political culture in the traditional Eastonian sense. For Easton (1965), political culture is the set of beliefs, ideas, norms and values that prescribe particular patterns of thinking, feeling, and acting that in turn affect the political input and performance of the system which members perceive or potentially experience as having an authoritative or binding force (Patrick 1984, 271). Political attitudes influence the types and extent of policies carried out, the strength or weakness of political parties, the level of integration/fragmentation of the policy making process, the degree of trust in society, the set of informal rules of bargaining processes, and finally the general context in which policy networks operate. Thus, our first hypothesis is that political stability and a high degree of system commitment and social integration enhance the conditions of policy cooperation and thus increase the probability that a nation state will implement European policies.

The attitude towards the EU is a more specific contextual variable that plays an important role in determining the cooperative conditions during the process of implementing European policies in member states. Since politicians often make policy choices that promote their re-election, it can be assumed that the lower the overall mass support for the country's membership in the EU, the higher the probability that a member state will face difficulties in implementing European policies. Despite the well-known exceptions (namely Denmark and Italy), our second hypothesis is that the higher the overall support for the country's EU membership, the greater its success in implementing European policies.

As discussed earlier, interest groups play an important role in decision making and in the implementation of European policies. The role of the interest groups in member states refers to the corporatist settings of interest intermediation. In general, neocorporatism describes a cooperative relationship between government and certain interest groups, a constellation which is necessary for stability and predictability when European policies are implemented. Therefore, corporatist arrangements increase the stability and degree of institutionalization of policy networks at the national level and set more rigid rules for interorganizational bargaining (Streek 1991). Thus we assume that a high degree of corporatism improves the conditions for successful implementation.

Political institutional factors are crucial in understanding governments' policy making capacity. This institutionalist emphasis focuses on the formal decision making structures and processes of democratic parliamentary systems, and on the systems for separations of power. Weaver & Rockman (1993) conceive the political-institutional arrangements of advanced capitalist states in terms of a number of discrete institutional variables, such as executive-legislative relations, electoral rules, party systems and patterns of government formation, center-periphery relations, the role of the judiciary, and the character of the bureaucracy. By capability of government, the

authors mean a pattern of government influence on its environment that produces substantially similar outcomes across time and policy areas. We assume that the more efficient the political institutions are in the member state, the better it will succeed in implementing European policies. We shall test these hypotheses by applying a multiple regression model, where the dependent variable is the average implementation rate.

Results of the Empirical Analysis

Data

To test our hypotheses we have collected comparative implementation data from The Annual Report on Monitoring the Application of Community Law produced by the Commission. This report contains statistics about the implementation ratios of directives and the infringement of treaties in each member state. This data is also available from each policy sector. From these reports we have constructed an average implementation score for each of the "old member states" (members prior to the 1995 enlargement of the Union). That score is the average percentage of implemented directives between 1990 and 1995.

Table 4 shows the construction and the alpha scores of our four scaled variables. The negative and positive impacts of the individual variables refer to the direction of each variable on the scale. Only the extreme alternatives of the scale are presented in order to make the table more understandable. All individual indices are coded by the 0-100 scale.

Our first independent variable is the form and role of the political institutions. Here we have chosen variables indicating the stability and efficacy of the political institutions. The first group consists of variables based on institutional design that reflect the preconditions of efficacy, i.e., type of parliamentary chambers, parliamentarianism/presidentialism, type of government coalition, and so-called institutional autonomy score.² The real government efficacy component contains five variables, number of parties in parliament, number of parties in government, effective number of parties score based on the Laakso & Taagepera index, average government support score and party government efficacy score. The last subcomponent of the political institution scale contains variables that reflect government stability. Here we have used the so-called polarization index, volatility of voter support for political parties, irregular changes of the government and average government durability score. In order to test that each of the items selected really belongs to the scale, we first analyzed the bivariate correlations between the variables selected and then tested the overall reliability of the scale by using Cronbach's alpha test.

The interest group scale includes three variables, which are degree of unionization, percentage of support for the moderate left from 1991-94,⁴ and the degree of corporatism/consociationalism.⁵ Degree of unionization reflects citizens' trust in interest organizations and the importance of such organizations. Several researchers have noted that strong support for the moderate left is likely to increase and institutionalize negotiations between state and interest groups.⁶ Whereas the previous variables are mainly concerned with the preconditions for strong interest group representation, degree of corporatism/consociationalism measures the actual appearance of collective bargaining systems and the significance of interest organizations in society and in the political system.

To operationalize EU support in member states, we have used an index which consists of twelve variables⁷ that reflect overall confidence, satisfaction and attachment to the EU among citizens in the member states. The group of confidence includes the following variables: EU optimism, i.e., trust in the future of the EU; attitude towards abolition of inner borders; and confidence in the Council of Ministers, the European Parliament, and the Commission. The variables in the satisfaction group are attitude towards EU membership, satisfaction with democracy in the EU, belief that national benefit is being gained, and overall satisfaction in what the EU has gained. Also included as variables in this index are citizens' ideas of national and individual influence on EU decision making and citizens' attachment to the EU. The alpha score of the EU support scale was rather good (0.84), indicating the consistency and unidimensionality of the variables chosen.

Political culture is a concept that is extremely difficult to operationalize, and it is often used as a catch all for phenomena and developments that cannot be explained otherwise. We have constructed a political culture index by rescaling some of the variables reported in numerous comparative research projects, reflecting commitment, stability and democratic performance at the aggregate level. The first group of variables in the political culture index reflect legitimization and overall satisfaction with democracy as seen in electoral participation and in satisfaction with democracy. The second group of variables reflects social fragmentation and social cleavages in society. Here we have applied the social fragmentation index in a broad sense⁸ along with post-war protest scores from the countries. The third group of variables in the political culture index contains variables based on individual values and autonomy. Here we use post-material value scores in each country based on Inglehart (1990) and the so-called individual autonomy score.⁹ We have also added a trust variable based on 1993 Eurobarometer data. This index shows how the trustworthiness of citizens in one country is evaluated by citizens from other countries. We also applied the test for unidimensionality using the correlation analysis and the test for reliability using Cronbach's alpha test, and both clearly suggested that all the variables belonged to the same scale

(alpha = 0.76). We interpret the dimensions of this scale as indicating that
high scores on our political culture scale reflect high societal stability and
trust in democracy.

Results

Since the size of the data (N = 11) seriously restricts causal analysis and statistical inference, we will first present some scatterplots to illustrate the relationships between our independent variables and the implementation failure variable. We begin by analyzing the relationship between the type and form of the political governing institutions and implementation failures. The correlation coefficient between these two factors remained relatively low at only 0.33. To elaborate the relationship between political institutions and implementation behavior, we have collapsed the institution scale into to subscales, government efficacy and stability. Countries are ranked differently according to government efficacy and government stability. Denmark is a good example of this case. Its government efficacy score is extremely low, while the stability of the country remains quite high. The former position can be explained by the large number of political parties in Denmark. Despite the low efficacy, Denmark's implementation failures remain very low. When discussing the Danish case, we must also consider the important role of the Parliamentary European Committee in Denmark. This committee is an important co-optative body which links the bureaucratic preparation process

of the Danish negotiation positions with the final acceptance of decisions made by the Council of Ministers. This institution was included in the efficacy scale since its counterparts could not be found in the other member states included in the analysis.

This can partly be explained by the political stability of the country. Italy, on the other hand, has the lowest implementation score, which suggests that part of the explanation might be the inefficacy of the country as well as its political instability. Although the political systems in the UK, Germany and Greece received relatively high efficacy and stability scores, there are still remarkable differences as far as their implementation is concerned; while the UK and Germany can be characterized as successful implementors, Greece does not belong to this category.

Next, we shall analyze the association between degree of corporatism in a country and its implementation behavior. We found a moderately positive correlation (0.40) between degree of corporatism and implementation success. Most of the countries seemed to support our working hypothesis quite well, according to which a high degree of corporatism increases the likelihood of successful implementation. However, two interesting exceptions, namely the UK and France, both of which can be considered deviant cases, have succeeded in implementing EU directives rather well despite their low degree of corporatism.

Figure 4 shows, somewhat surprisingly, that there is a modestly negative correlation (-0.46) between overall support for EU membership and implementation success. One reason could be the exceptional character of countries like Portugal, Denmark and the UK. This finding suggests that the overall mass opposition towards EU membership does not seem to increase problems in implementing common European policies, and similarly does high support not directly increase the probability for higher implementation rates.

However, the political culture of a country seems to be an important factor in creating conditions for successful implementation at the national level. We found that the relationship between political culture and implementation success was surprisingly strong compared to the other factors (r = 0.61). Figure 5 clearly shows that countries with a stable political culture, as we assumed, seem to succeed well in the task of implementation. Most countries in the plot are characterized by medium level of stability in terms of political culture and manage quite well to implement EU directives. Denmark, as previously pointed out, is a very good implementor and in addition, its political culture is the most stable among the countries studied. Italy and Portugal have difficulties implementing EU policies, and their scores on the political culture scale are fairly low. These results seem to support our hypothesis that there is a positive relationship between cultural stability and

trust and successful implementation. What is interesting, however, is the finding that despite the fairly high political culture scores, Greece has not been able to meet its requirements concerning the implementation of European policies. A contrasting case is Spain, which receives the lowest political culture scores and nevertheless has succeeded fairly well in the implementation of European policies. The deviancy of these two countries could be partly explained by the short time of their membership.

In order to conclude that political culture is the main determinant of national failure to implement European legislation, one should firstly simultaneously control for the other variables and take sectoral variation in implementation ratios into consideration. To control for the main independent variables along with the main policy sectors, we shall now introduce a multiple regression model in which implementation rates are treated as the dependent variable.

Regression Analysis

The data used in the regression model is collected from the same Commission
reports as in the previous model. This time only the dependent variable

consists of average annual implementation percentages¹⁰ by the old member states (excl. Luxembourg) in nine different policy areas. Thus the size of our data (N = 110) equals the number of member states times the number of policy sectors in the model. Each country and each policy sector were then controlled separately by dummy variables. All sectors were taken into the model and countries were then added step by step to control for the unit effect heteroscedasticity (see Stimson 1985; Clarke et al. 1993). In the first step we applied a model without country dummies. In the second step, using the residual diagnostics of the first step, regression countries were then added sequentially, beginning with those which the residual diagnostics indicate are most likely to need to be controlled for unit heteroscedasticity. Additional country dummies are added in subsequent analyses if residual diagnostics suggest that they are required. The main independent variables (EU support, degree of corporatism, political culture and institutional settings) are all rescaled so that they vary between 0 and 1.

The results of the regression analysis are presented in Table 5, which shows
that the fit of the model can be considered as more than sufficient with an
adjusted R2R2 of 0.66. However, according to our main hypothesis, the overall

fitness of the model is not our main concern. Instead, we are concerned with the question as to whether our main independent variables have an empirically significant impact on the implementation ratio. As we can see, only political culture and the type of political institutional settings have the expected positive (b's - 2.13 and 0.64) and statistically significant impact on implementation success. That is, the higher the degree of political culture (stability, satisfaction and trust) and the more efficient the institutional architecture of the politico-administrative system in a member state, ceteris paribus, the higher the ratio of successful implementation of the European legislation at the national level. In contrast overall support for the EU and the degree of corporatism do not seem to be good predictors of national implementation behavior. Employment and social policy, energy policy and technical barriers are the sectors facing the biggest problems in implementing European policies. All sectoral dummies are statistically significant and have lower implementation percentages compared to the tax barriers used as a baseline sector.

Conclusion

In this article, we have studied the effects that political institutions, interest organization structure, political culture and mass opinion have on implementation rates in the twelve EU member states. This study had two main research hypotheses. First, it was presumed that countries with effective and stable political institutions and a corporatist system that integrates interest organizations into political decision making, would have the best capability to implement EU directives. Our second hypothesis was that it is easier to implement EU directives in countries where the political system has high legitimacy, people are satisfied with democracy, the degree of social fragmentation is low, individual rights are highly respected, and attitudes towards the EU are positive. In these circumstances, the overall willingness to obey rules, including EU rules, is high.

These research hypotheses were tested by analyzing a set of data set which included the number of unimplemented directives as well as indices reflecting factors that affect countries' capability and willingness to implement EU directives. The twelve member states can be divided by implementation rate into three categories: Denmark, The Netherlands and the UK are by all measures the best implementors, while Italy, Greece, Portugal and Belgium have constantly failed in implementing EU legislation. Other countries like Germany, France, Spain and Ireland receive average implementation scores. The results of the regression analysis support two of our main hypotheses: they show that stable political culture combined with efficient and flexible institutional politico-administrative design are the best predictors of successful

implementation of common European policies. Examples are Denmark and the Netherlands, where European legislation is efficiently implemented, even if these countries have had a somewhat critical attitude towards European institutions and the contents of joint policy objectives. It was obvious that critical mass opinion towards the EU did not have much to do with implementation behavior. The same goes for the role of key interest organizations Countries traditionally labeled as corporatist did not succeed better than non-corporatist ones. This may be explained by the faet that in corporatist countries powerful labor unions can easily block certain European directives and thus cause many problems in the implementation process. The faet that employment and social policy together with agriculture, fishing and energy (all sectors with strong meso-corporatist structures) were policy areas with relatively low implementation scores confirms the hunch that interest groups are the important actors in national bargaining over compliance with European rules.

The new member states Austria, Finland and Sweden, were not included in the analysis because they they only joined the EU in 1995, and thus their annual variances could not be controlled. However, it can be said that even during the first year of membership, the implementation rate in Sweden was higher than the average among "old" members. This also applies to Finland, if we take into account that most of the unimplemented directives in Finland are due to the delays caused by the slow implementation process in the autonomous region of Ahvenanmaa. Excluding Finland, Austria has the lowest implementation rate in the EU, mainly due to difficulties and delays in implementing agricultural directives.

All in all, unimplemented directives remain a serious threat to the development of common European policies and thus to the strengthening of the integration process. A key problem seems to be that despite its massive monitoring body, the Commission, the EU lacks an effective control system and appropriate enforcement strategies to control member states. The fact that almost all member states have some areas in which their legal implementation rates fall far below the average raises the question of what happens during the final implementation process, i.e., the actual administrative application of EU laws. This could, in a worst case scenario, lead to a situation in which the EU produces political decisions with no real impact on every day life in the member states

---

NOTES

1. For the distinction between statutory and non-statutory preconditions of the implementation process, see Mazmanian & Sabatier (1989, 20-39).

2. Institutional autonomy score adds up for the scores for the four dimensions: degree of federalism, territorial autonomy, regional and local government discretion, functional autonomy.

3. Polarization is the distribution of the electorate along the classical right-left scale.

4. Sources: EJPR Political Data Yearbooks 1992-95 and Katz & Mair (1992).

5. Sources: Lane & Ersson (1994); Schmitter (1981); Lembruch (1984); and Schmidt (1982).

6. It is thus assumed that corporatist arrangements like "policy communities" (Jordan & Schubert 1992, 23) are likely to facilitate the implementation, while e.g. "issue network" (Heclo 1978, 102-5) bargaining can even be an obstacle to effective implementation.

7. These variables are coded from the "Continuous Tracking Survey Data" gathered by the Commission.

8. Social fragmentation is an index on social heterogeneity consisting of an average for the ratings of the various nations on religious structure, ethnic structure and class-based cleavages (Lane & Ersson 1994, 99).

9. Individual autonomy is a rescaled index comprising the civil rights score, the political rights score, political and economic discrimination score and the human rights index (Lane & Ersson 1994, 204).

10. Implementation rate (IR) was transformed by LN(IR/(1-

ACKNOWLEDGMENTS

The authors wish to express their thanks to Olli Mäenpaä and Mikko Mattila for valuable
comments on this article.

REFERENCES

Andersen, S. S. & Eliassen, K. A. 1996. "Introduction: Dilemmas, Contradictions and the
Future of the European Democracy," in Andersen, S. S. & Eliassen, K. A., eds., The
European Union: How Democratic Is It? London: Sage.

Clarke, H. D., Dutt, N. & Kornberg, A. 1993. "The Political Economy of Attitudes toward
Polity and Society in Western European Democracies," The Journal of Politics 55: 998-1021.

Commission of the European Communities. Annual Report on Monitoring the Application of
Community Law. Years 1983-95. Official Journal of the European Communities, C series.

Easton, D. 1965. A Framework for Political Analysis. Englcwood Cliffs, N.J.: Prentice-Hall.

Ekstrom, M. 1994. The Role of the European Court of Justice in the Enforcement of
Unimplemented Directives. Publications by the Institutc for European Law at Stockholm
University, No. 21.

European Journal of Political Research, Political Data Yearbooks. Various years. London:
Kluwer Academic Publishers.

From, J. & Stava, P. 1993. "Implementation of Community Law: the Last Stronghold of
National Control," in Andersen, Sv. S. & Eliassen, K. A., eds., Making Policy in Europe -
The Europeification of National Policy making, London: SAGE Publications.

Haas, E. 1975. "Turbulent Fields and the Theory of Regional Integration," Internationa!
Organization 30.

Harding, C. & Sherlock, A. 1995. European Community Law - Texts and Materials. London:
Longman.

Heclo, H. 1978. "Issue Networks and the Executive Establishment," in King, A., ed., The New
American Political System. Washington: AEI.

Hix, S. 1994. "The Study of European Community. The Challenge to Comparative Politics,"
West European Politics, 17: 1-30.

Inglehart, R. 1990. Culture Shift in Advanced Industrial Society. Princeton: Princeton
University Press.

Jordan, G. & Schubert, K. 1992. "A Preliminary Ordering of Policy Networks," European
Journal of Political Research, 21: 7-28.

Katz, R. S. & Mair, P. (eds.). 1992. Party Organizations -A Data Handbook. London: Sage
Publications.

Keohane, R. & Hoffmann, S. 1991. "Institutional Change in Europe in the 1980s," in Keohane,
R. & Hoffmann, S., eds., The New European Community. Boulder CO: Westview Press.

Kohler-Koch, B. 1996. '"The Strength of Weakness: The Transformation of Governance in the
EU," in Gustavsson, S. & Lewin, L., eds., The Future of the Nation State. Stockholm: Nerius
& Santerus.

Lane, J.-E. & Ersson, S. O. 1994. Politics and Society in Western Europe. London: Sage
Publications.

Lehmbruch, G. 1984. "Concertation and the Structure of Corporatist Networks," in
Goldthorpe, J., ed., Order and Conflict in Contemporary Capitalism. Oxford: Clarendon
Press.

Ludlow, P. 1991. "The European Commission," in Keohane, R. & Hoffmann, S., eds., The
New European Community. Boulder, CO: Westview Press.

Mazmanian, D. & Sabatier, P. A. 1989. Implementation and Public Policy. New York:
Lanham.

Patrick, G. M. 1984. "Political Culture," in Sartori, G., ed., Social Science Concepts. A
Systematic Analysis. Beverly Hills: Sage Publications.

Schmidt, M. G. 1982. "Does Corporatism Matter? Economic Crisis, Politics and Rates of
Unemployment in Capitalist Democracies in the 19705," in Lehmbruch, G. & Schmitter, P.
C, eds., Patterns of Corporatist Policy Making. London: Sage Publications.

Schmitter, P. C. 1981. "Interest Intermediation and Regime Governability in Conteporary
Western Europe and North America," in Berger, S., ed., Organizing Interests in Western
Europe. Cambridge: Cambridge University Press.

Schmitter, P. 1996. "If the Nation-State Were to Wither away in Europe, What Might Replace
It?," in Gustavsson, S. & Lewin, L., eds., The Future of the Nation State. Stockholm: Nerius
& Santerus.

Schneider, V. & Werle, R. 1992. "Networks and Concertation in European Policy Making:
The Cases of Chemical Control and Telecommunicatons," Paper, presented at the ECPR
joint sessions in Limerick, 1992.

Stimson, J. A. 1985. "Regression in Space and Time: A Statistical Essay," American Journal of
Political Science 29: 914-47.

Stoker, R. P. 1991. Reluctant Partners: Implementing Federal Policy. Pittsburgh, PA:
University of Pittsburgh Press.

Streek, W. 1991. "Interest Heterogeneity and Organizing Capacity. Two Class Logics of
Collective Action," in Czada, R. & Windhoff-Heritiér, A., eds.. Political Choice.
Institutions, Rides and the Limits of Rationality. Campus Verlag: Frankfurt am Main.

Waltz, K. 1986. "Reflections on Theory of International Theory of Politics: A Response to My
Critics," in Keohane, R., ed., Neorealism and its Critics. New York: Columbia University
Press.

Weaver, R. K. & Rockman, B. A. 1993. Do Institutions Matter? Governments Capabilities in
the United States and Abroad. Washington, D.C.: Brookings Institution.