Interest Groups in the European Union: How Powerful Are They?

Interest Groups in the European Union: How Powerful Are They? Andreas Dür School of Politics and International Relations University College Dublin Andreas.Duer@ucd.ie Paper forthcoming in West European Politics, Vol. 32 (2009), No. 1. Abstract The question of interest group influence is central for both scholars interested in studying policy-making in the EU and those concerned with the legitimacy of decision-making in that entity. Many hypotheses exist that stress a series of factors possibly shaping interest group influence. Nevertheless, only few studies have tried empirically to examine these hypotheses for the case of the EU. What is more, existing empirical studies report contradictory findings. To help researchers advance upon this state of the art, I discuss three obstacles that impede research dealing with interest group influence in the EU: defining the terms power and influence, accounting for different pathways to influence, and measuring influence. I also suggest ways to address these obstacles: by taking a pragmatic approach with respect to defining influence and power, by being conscious that actors may use different pathways to influence when drawing conclusions about interest group influence, and by combining different methods when measuring influence. Acknowledgments: I am grateful to Dirk De Bièvre, Richard Sherman, Arndt Wonka, the editors of the special issue, the participants of two workshops in Leiden in April 2005 and in Hagen in March 2006, and an anonymous referee for helpful comments on earlier versions of this article. Word Count: 8,400 Version: May 2008 Electronic copy available at: http://ssrn.com/abstract=1188504

1. Introduction The number of interest groups that are active in the European Union (EU) has increased sharply from the mid-1980s onwards (Greenwood 2007: 12). 1 Since one of the main objectives of these groups is to influence policy outcomes in their favour, the question how successful they are in doing so would seem to be of central concern for both scholars interested in studying policy-making in the EU and those apprehensive about the legitimacy of decision-making in that entity. Nevertheless, few attempts at evaluating interest group influence or power in the EU have been undertaken (see also Lowery et al., this volume). This state of affairs is particularly astonishing given that both of the grand theories of European integration, namely neofunctionalism and liberal intergovernmentalism, ascribe a major role to interest groups (Grossman 2004). The lack of studies on interest group influence in the EU also starkly contrasts with the sizeable literature on interest group power in national political systems, produced as a result of the community power (Mills 1956; Dahl 1961; Lowi 1969), the state autonomy versus pluralism (Evans et al. 1985; Almond 1988), and the weak versus strong state debates (Atkinson and Coleman 1989). Among the few studies that actually tackle the question of interest group influence in the EU, most do so with respect to specific policy decisions. The stated aim of many of these studies is to counter the simplistic view of all-powerful interest groups present in popular accounts (Balanyá et al. 2003) by mustering examples in which interest groups despite extensive use of their resources were unable to get what they wanted. In addition, at a middle level of abstraction, some studies deal with the impact of specific lobbying strategies, such as information transmission, on policy outcomes. Nevertheless, with the number of studies that empirically examine interest group influence in the EU quite limited, it is no wonder that many hypotheses, which stem from the broader literature on interest group power, have not yet been tested for the case of the EU. What is more, the few studies that 1 Electronic copy available at: http://ssrn.com/abstract=1188504

exist reach contradictory conclusions. 2 In this situation, an inventory of existing arguments and empirical findings may help advance the state of the art by drawing out areas of agreement and of disagreement. With this objective in mind, in the following I offer a critical review of existing research on interest group power in the EU. 3 I first sketch out a series of hypotheses about the role of interest groups in EU decision-making, which stress interest group resources, political institutions, issue characteristics, and interest group tactics as determinants of influence. Next, I outline the major findings of a series of studies that empirically examine some of these hypotheses for the case of the EU, and conclude that there is little agreement among these studies. I argue that what makes it difficult to arrive at more clear-cut findings are the problems of defining power and influence, taking into account different pathways to influence, and measuring influence. In the conclusion, I suggest some ways to advance beyond the current state of the art. 2. Theoretical Perspectives on Interest Group Influence in the EU The existing literature on the determinants of interest group influence provides a large number of hypotheses, many of them originally formulated for the case of the United States. When discussing some of these hypotheses as applied to the EU, I distinguish four broad clusters of determinants: interest group resources, political institutions, issue characteristics, and interest group strategies. 2.1 Interest Group Resources Most scholars agree that interest groups endowment with resources furthers their capacity to influence decision-makers and policy outcomes (Gerber 1999; Hall and Deardorff 2006; but see Burstein and Linton who take a sceptical stance). Interest groups resources mentioned in the literature include money, legitimacy, political support, knowledge, expertise and 2

information. Interest groups may dispose of financial means that they can use either to support an incumbent or a challenger in electoral contests. By dealing with certain political or bureaucratic actors (in particular officials that are not directly elected, such as those in the European Commission), interest groups may also be able to convey legitimacy upon them. Interest groups, moreover, can express their support for a politician in exchange for policies that favour their economic or other interests, which may influence the voting decisions of a rationally ignorant electorate. Most importantly, interest groups may have knowledge, expertise and information that can facilitate the task of decision-makers (Crombez 2002; Hall and Deardorff 2006). For example, groups may have information on market conditions, probable policy results, problems of implementation, and the support a specific policy will receive, which politicians need to assess the political feasibility of a policy proposal. Not all interest groups are equally endowed with these resources, leading to the expectation that some groups are more influential than others. Structural characteristics of a group such as size, the type of its membership, its internal organisation, and the degree of geographical concentration of the membership affect its resource endowment. Starting with size, the larger the pool of members, the more political legitimacy a group may be able to convey. The type of membership is also important, as one can expect groups composed of firms to be well equipped with monetary resources and groups made up of individuals to find it easier to mobilise personnel resources (Gerber 1999). In addition, business interests that control investment and production decisions in different countries may avail of specific information that other interests lack (McLaughlin et al. 1993: 200). With respect to internal organisation, hierarchically structured groups may be better at providing information than groups that have to live with conflicting interests within themselves (on this distinction, see also Beyers, this volume). The direction of the effect of geographic concentration on resource endowment is ambiguous: geographically concentrated interests may find it easier to overcome collective action problems and provide public goods. Geographically dispersed 3

interests that manage to organise, by contrast, have the advantage of being represented in several electoral districts at the same time. They may also possess more information and may be able to sway more voters. In a rational-choice exchange perspective, interest groups exchange their resources for political influence in an effort at maximising their own utility. In this view, how much influence interest groups can get in exchange for their resources depends on political actors demand for such resources (Kohler-Koch 1994: 170). Two factors have an effect on this demand: (1) the value of the resources for public actors in terms of achieving re-election (office-seeking politicians) or specific policy objectives (policy-seeking politicians) and (2) the availability of alternative sources of resources (either from other groups or from own sources). The value that resources have for public actors will vary depending on the type of uncertainty they are facing: incomplete information about the transformation of decisions into outcomes (policy consequences) or incomplete information about the strength of different constituencies (electoral consequences). Public actors will also only exchange influence for resources from a specific interest group if they are unable to produce these resources themselves, and if they have no other, lower-cost source for them. In short, according to this view interest groups capacity to supply and decision-makers demand for resources mutually determine the influence groups may gain over policy outcomes. Besides engaging in a mutually beneficial exchange of resources, interest groups may also use their resources to shape public actors beliefs, ideas, cognitive frames, and preferences (Collignon and Schwarzer 2003; Beyers, this volume). They may opt for arguing rather than bargaining, with arguing being aimed at changing the beliefs and preferences of decision-makers. Information then is used not in exchange for influence, but to convince decision-makers of the merits of a proposal. Arguing can also be directed at the broader public, with the objective of changing the framing of an issue in public debate. In this perspective, resources are also important, but what matters is not a group s overall 4

endowment with resources, but whether the group can come up with the right ideas at the right time, which change decision-makers or the public s interpretation of an issue. In the EU, the European Commission, which is characterised by an extensive policy agenda and limited policy resources (McLaughlin et al. 1993: 201), may depend more on interest group resources than other actors. The EU s alleged democratic deficit and contested legitimacy may also cause greater demand for interest group input by public actors. This leads to the expectation that interest groups that are well-endowed with resources should be particularly influential in this political system. In line with this idea, it has been argued that the Commission s dependence on interest group resources makes it offer access and influence to interest groups, leading to a symbiotic relationship between the two (Mazey and Richardson 2003: 212). Problematically, though, many of the resources supposed to be used by interest groups to gain influence cannot be measured empirically. How do we know how much information or knowledge an actor has? Neither is it possible to determine decisionmakers demand for resources. An empirical examination of the resource hypothesis thus appears difficult. 2.2 Political Institutions Political institutions affect the degree of interest group influence on policy outputs mainly by shaping societal actors access to policy-making processes (Immergut 1992; for a sceptical view, see Sheingate 2001). 4 No agreement exists, however, on whether the vertical and horizontal division of power in the EU facilitates or impedes the access of societal actors to decision-makers (Princen 2007). One argument is that intergovernmental cooperation at the European level strengthens the power of the state vis-à-vis domestic actors through four mechanisms: giving the executive control over domestic agendas; changing the procedures through which decisions are made; providing the executive with asymmetric access to information; and facilitating the ideological justification of policies (Moravcsik 1994). 5

Turning around the two-level-game expectation that domestic constraints increase the bargaining power of chiefs of government in international negotiations, it can also be hypothesised that politicians use cooperation at the European level as a device to loosen the constraints imposed on them by interest group pressures in the domestic arena (Grande 1996). Finally, the complexities created by the division of power in the EU may make it difficult for interest groups to design an efficient lobbying strategy and thus inadvertently thwart their lobbying effort. Alternatively, some authors emphasise that the European multi-level system strengthens interest groups (even those representing diffuse interests) by increasing the number of potential access points (Pollack 1997). Societal actors have many access points in the EU since they can lobby at both the domestic and the European level. In addition, at the European level institutions such as the European Parliament and the European Commission seem to be exceptionally accessible to interest groups. A particularity of the EU is also that private actors can avail of a supranational judicial channel to achieve their desired policy outcomes (Smith 2001; Bouwen and McCown 2007). The existence of such additional venues at the European level may even lead to the break-up of established policy communities at the domestic level (Richardson 2000). The lack of transparency created by the complex institutional structure of the EU, however, may strengthen concentrated interests to the detriment of nongovernmental organisations because it decreases the latter s control over policy outputs (Schneider and Baltz 2003). The idea behind this reasoning is that public actors can give in to welfare-reducing demands from special interest group only to the extent that voters do not punish them for doing so. Increasing information asymmetries due to higher costs of monitoring, in this view, encourage slippage by public actors. It may also be that whether public actors are appointed or elected has an impact on interest group influence. For example, the bureaucrats in the European Commission do not have to fight for re-election; this may make them less receptive to interest group demands 6

simply because they do not rely on resources from interest groups for a re-election campaign. Moreover, the Commission s lack of direct electoral legitimation may make it particularly eager to appear even-handed (McLaughlin et al. 1993: 202), further limiting its capacity to give in to interest group demands. At the same time, the lack of electoral oversight may increase the officials scope for giving in to interest groups. For the case of the European Parliament, the need to gain elections limits parliamentarians ability to yield to interest group demands for two main reasons. First, once elected politicians have publicly taken position on an issue, they find it difficult to change their position following interest group demands since they have to appear ideologically consistent to appeal to voters (Downs 1957: 110-12). Second, in the case that rent seeking by specific groups goes to the detriment of overall economic efficiency, ex-post voters may punish elected officials for decisions taken. In summary, political institutions are an important determinant of interest group influence. What is problematic about this argument, though, is that institutions themselves are partly expressions of past interest group lobbying (see also Princen and Kerremans, this volume). Often, the rules and procedures governing decision-making may reflect a specific (past) balance of power among societal interests; powerful interests may have been able to shape the rules in a way that favours their interests. Only in the short run, therefore, can political institutions be treated as if they were exogenous factors determining interest group influence. 2.3 Issue Characteristics Issue characteristics, such as policy type, degree of technicality, and public salience, also have an effect on interest group influence. Whether an issue is of a regulatory, distributive or redistributive kind (Lowi 1964) plays a role in determining interest group influence. In distributive politics, where the benefits of action are concentrated on a specific group, and the costs are widely dispersed, groups representing concentrated interests tend to win out over 7

diffuse interests. The main advantage of the former is that they can better monitor the implementation of policies that affect their economic well-being than diffuse interests that face collective action problems (Lohmann 1998). Regulatory decisions generally entail concentrated benefits for one and concentrated costs for another group. In this situation, at least two groups with opposing interests should clash, imposing a limit on the power of individual groups (Wilson 1980). Interest group influence should be rather small with respect to redistributive policies, as such policies produce diffuse costs and benefits. Some scholars also stress that interest groups have more influence on technical issues than on issues of high politics (Smith 2000; Greenwood 2003: 20; see also Beyers, this volume). This difference may partly be explained by arguing that politicians demand for information and hence dependence on interest groups is highest in technical fields (Schumann et al. 2005). In more technical areas, moreover, interest group influence on policy outcomes may be higher because decision-makers have to rely on domestic actors cooperation in the implementation phase of the policy cycle (Wolf 2005). Finally, interest group influence may depend on the public salience of an issue (Mahoney 2007). Interest groups should find it difficult to influence policy outcomes when the public is highly attentive to an issue. On such issues, legislators should fear electoral punishment if they implement policies that favour concentrated interests to the detriment of diffuse interests. Salience may, however, be endogenous to the policymaking process. For example, a group may try to increase the salience of an issue by way of campaigning to advance its interests. This may be a particularly effective strategy for groups that want to block policy changes advocated by other groups. 2.4 Interest Group Strategies Strategies are a final factor shaping interest group influence. To maximise influence, interest groups have to employ their resources effectively given the opportunities provided by the 8

institutional structure, the characteristics of the issue, the preferences they advocate, and their past strategies. If groups always adopted ideal strategies, strategies would only be an intermediary variable that is perfectly explained by a group s resources, the institutional framework and so forth, and hence could be ignored in attempts at explaining interest group influence. Yet, while very little research exists on this issue, it seems plausible that groups sometimes fail to pick the most effective lobbying strategy. This may be due to the EU s complex institutional system, which compels interest groups to make even more choices than in national political systems to maximise their influence: lobby at the national or at the EU level (Lindberg 1963); lobby the Commission, the Parliament, or the Council of Ministers (Bouwen 2004); lobby directly, via national associations, or via European associations (McLaughlin et al. 1993; Bennett 1999); and lobby at the time the Commission becomes active, when a directive is discussed in the Council, or when it is implemented and enforced at the national level (Crombez 2002). Only if a group manages to choose the ideal strategy among these many options will it be able to maximise its influence on policy outcomes. By learning how to use opportunity structures effectively (Tarrow 2001: 238), interest groups thus may increase their influence on policy outcomes. Learning may not be sufficient for the adoption of effective tactics, however. The pursuit of objectives other than influencing public policy such as maintaining or enlarging the group s membership may lead groups to adopt specific tactics even when knowing that they are ineffective in shaping policies (Dür and De Bièvre 2007; Lowery 2007). 3. Contradictory Empirical Findings Only recently have some scholars started to empirically examine a few of these hypotheses for the case of the EU. Contradictory findings characterise the resulting literature. Some studies, for example, suggest that concentrated interests exert substantial influence over EU decision-making. One study comes to the conclusion that large firms had substantial influence 9

on policy outcomes in the biotechnology and electrical energy sectors (Bandelow et al. 2000). Firms engaged in package deals with the European Commission that provided the former with influence and the latter with an expansion of its powers. Another study found that business and agricultural groups wield considerable influence on the EU s approach to trade negotiations (Dür 2008a). This finding is based on a comparison of the demands voiced by concentrated interests and the position adopted by the EU in trade negotiations in the framework of the World Trade Organisation. For other authors, concentrated interests influence on EU decision-making is far more circumscribed. According to them, the Commission is quite autonomous in developing its preferences, and then seeks support from groups that have a similar preference. Irina Michalowitz (2007), for example, argues that groups can only exert what she calls technical influence; they cannot change the core aspects of proposals presented by the Commission. Interest groups influence has even been seen as depending on the goodwill of relevant policymakers in the Commission (Coen 1997). The inter-institutional dynamics between the Commission and the Council of Ministers may also limit interest group influence. In trade policy, for example, the Commission may be reluctant to embrace calls for protectionism, if the officials in Directorate General Trade suspect that broad support among member states for such a policy is lacking (Woll forthcoming). In this view, the Commission only adopts policies that it feels do not create deadlock among member states. Even with respect to agricultural policy, where farmers associations are often assumed to be highly powerful, it has been argued that societal actors were not particularly influential (Rieger 1994). Farmers supposedly only achieved the protectionist policies they preferred because the Commission used agriculture as leverage in international trade negotiations. Contradictory findings also characterise the literature on the question of how much influence diffuse interests can wield in the multilevel system of the EU. On the one hand, some authors find that diffuse interests are surprisingly influential in the EU (Pollack 1997; 10

Warleigh 2000; Tarrow 2001: 238; Ruzza 2002). In support of this position, it has been argued that diffuse interests were successful in using the EU to get member states to tighten protection for the environment, consumers and women (Pollack 1997). On the other hand, other authors contend that diffuse interests are largely unable to influence EU policy outcomes (Schneider and Baltz 2003; Dür and De Bièvre 2007). In trade negotiations, for example, nongovernmental organisations representing diffuse interests had little impact on the positions taken by the EU with respect to issues that were highly salient to them (Dür and De Bièvre 2007). Despite organising major campaigns, they were unable to significantly influence the EU s approach to the negotiations concerning the Economic Partnership Agreements with African, Caribbean and Pacific countries and cheaper medicines for poor countries. Finally, there has hardly been any empirical research regarding the question which strategies allow interest groups to maximise influence. It is well established that interest groups use both national and European channels of influence (Eising 2004). In addition, we know that groups rely on both access and voice to exert influence (Beyers 2004). These findings, however, cannot answer the question whether groups use of these strategies is always effective. The strategy of lobbying at different levels and of combining access and voice does not necessarily maximise groups influence at all times. For example, if it is important to frame an issue as being of a European kind to gain influence with the Commission (Gardner 1991, 67-68), lobbying at both the national and the European level may be counter-productive. With respect to strategies, therefore, the problem is not one of contradictory findings, but that existing hypotheses linking strategies to influence have hardly been tested at all. 11

4. Three Obstacles and How to Address Them This very brief overview of the (relatively small) empirical literature on interest group influence in the EU has demonstrated that conflicting findings persist. Three obstacles seem to be particularly important in inhibiting this literature from arriving at more clear-cut findings: the difficulty of defining influence, the need to consider several pathways to influence, and the problem of measuring influence. In the following, I discuss these obstacles and also offer some suggestions concerning ways to overcome them for the literature to advance beyond the current state of the art. 4.1 Defining Power and Influence A first obstacle crippling empirical research on interest group power and influence is the lack of an agreed definition of these two terms. Power is not only one of the most important, but also one of the most contested concepts in the field of political science. It can be understood both as a property or capability of an actor and as a causal concept (Hart 1976). Depending on which conceptualisation is chosen, researchers are pushed either to study the resources of an actor or the impact of an actor on outcomes. To complicate matters further, in the community power debate of the 1960s and 1970s scholars distinguished several faces of power. The first face of power is the traditional one of who wins and who loses on a certain issue (Dahl 1961). The second face of power refers to an actor s ability to set the agenda, in particular to keep unwelcome issues off the agenda (Bachrach and Baratz 1962). The third face of power, finally, has to do with an actor s capacity to prevent other actors from recognising their genuine interests; weak actors are driven to prefer things that go against their fundamental interests (Lukes 1974). Scholars that want to assess interest group power thus confront at least two major problems. On the one hand, they have to make explicit their position in this rather unwieldy debate. This implies investing some time in legitimising a specific definition of power and 12

sketching out how this definition compares to others used in the literature. Inattention to this requirement creates problems for the cumulation of research as it makes later generalisations of findings difficult. On the other hand, studying the second and especially the third face of power is extremely difficult. How should we know what the genuine interests of an actor are, when it is already difficult to determine the actual preferences? Rather than conceding defeat in the face of these difficulties, I propose that future research should consciously focus on specific aspects of the concept, therefore making it amenable to empirical research. For example, it may be helpful in some studies to assume that preferences are genuine and stable, thus by assumption excluding the working of the third face of power. Other studies may only concentrate on assessing whether the second (or third) face of power is at work. A potential drawback of this approach is the difficulty of reintegrating the various aspects of power examined in different studies in an aggregate analysis. Nevertheless, this path is likely to be more productive than were the endless quarrels about defining and operationalising an overly complex concept that characterised past approaches to the question of interest group power. 4.2 Considering Different Pathways to Influence A second obstacle for research on interest group influence is the need to take into account different pathways to influence, namely access, selection, voice, and structural coercion, to avoid running the risk of underestimating interest group power. Access refers to interest groups direct expression of demands to decision-makers. In the EU, this can take place in a number of different forms, including informal communication and testifying before expert commissions or public hearings (Beyers 2002; 2004; Bouwen 2004; Wessels 2004). Although interest groups may find it easier to influence policy outcomes whenever they have a seat at the table, access does not necessarily translate into influence. Opposing groups may have equal access and political actors can reject the demands made by interest groups. Public actors 13

may even use access as an instrument to co-opt societal interests. Taking access as a proxy for influence thus is likely to lead to erroneous results (Dür and De Bièvre 2007). Interest groups are not limited to influencing policy outcomes at the time of the passage of a bill (or, for that matter, before a court issues a ruling), but can also bias the selection of politicians (or of judges and bureaucrats) that have formal decision-making powers (Fordham and McKeown 2003). They can do so by contributing to election campaigns, by influencing public opinion, and by putting pressure on political principals who appoint agents. The rationale behind this strategy is that actors policy decisions are likely to be influenced by their policy preferences and not only by re-election concerns. Policy preferences, in turn, are often shaped by an actor s past experiences and socio-economic background. By influencing the selection of decision-makers, interest groups can make sure that actors with preferences close to their own are in key decision-making positions. If this strategy is successful, it may be difficult to distinguish between public actors on the one hand and the governed on the other. Although quite compelling, so far this idea has received little attention in the literature on interest group influence in the EU. The process of influencing selection is likely to be different from the case of the United States, where interest groups extensively contribute to the electoral campaigns of politicians. Nevertheless, it is to be expected that interest groups at least try to have their voice heard in the selection of decision-makers also in the European case. Interest groups may in particular try to influence the decisions on which Directorate- General of the Commission is dealing with an issue and which Member of the European Parliament is selected as rapporteur when the Parliament deals with a specific Commission proposal. One study, for example, highlights the success of the biotechnology industry in reducing the control of DG XI (Environment) over biotechnology policy in the EU (Bandelow et al. 2000: 19). 14

Furthermore, interest groups may use the voice channel to influence political processes. Groups can make noise by way of manifestations, rallies, petitions, statements in the media, and participation in public debates. The voice channel also includes influencing referenda or citizen initiatives, for example by way of campaigning (Gerber 1999). Groups can try to attain two objectives with such outside lobbying (Kollman 1998): on the one hand, voice may be aimed at influencing public opinion in favour of the demands of certain groups. For this, an initial coincidence between public and private interests (or at least the adequate framing of an issue as advancing the public interest) may be necessary. On the other hand, outside lobbying can be used to transmit information on policy preferences to politicians. In the case of the EU, as in other political systems, interest groups seem to use the voice channel less frequently than the access channel (Beyers 2004). Possible explanations for this finding abound: making noise may be more costly than gaining access; it may be little useful when trying to influence decision-makers that are not directly electorally accountable such as European Commissioners; and it may only work if an issue has some prior salience with voters. Finally, especially business interests may avail of structural power because of their capacity to decide when and where to invest (Lindblom 1977; Bernhagen 2007). Businesses can invest in polities where they feel that their interests are best served, and exit from polities where they face high taxation and regulatory obstacles. These investment decisions influence policy-making if politicians rely on a good business climate for their re-election. Alternatively, politicians may depend on investments by firms for financial reasons: economic prosperity increases tax revenues, and an increase in tax revenues means that politicians have more leeway to implement desired policies. In either case, politicians are likely to instigate policies that favour business interests with structural power, even if these businesses do not actively engage in lobbying. If politicians lack the necessary information to decide which political decisions hurt business, and especially how much a decision hurts business, however, 15

they may require business interests to send a costly signal whenever they want to veto a decision (Bernhagen 2007). In such a situation, active lobbying combines with structural power to secure business groups influence over policy decisions. What is important about this discussion of pathways to influence is that since groups are likely to use several pathways to achieve their objectives, the analysis of only one may lead to a serious underestimation of overall group impact. For example, a group does not need to gain access to decision-makers if it manages to influence their selection. Indeed, in such a case, researchers should see hardly any conflict between societal and public actors. Equally, without any overt political activity, some societal actors may exert influence by way of structural power. In future research on interest group influence in the EU, therefore, it will be essential to consider the existence of distinct pathways to influence. This does not mean that each and every study has to assess all pathways to influence; instead, it means that scholars have to make explicit the limitations of the findings obtained in studies that only consider one or a few of these pathways. In particular, claims about groups lack of influence will have to be made more sparingly unless all possible pathways to influence have been considered. 4.3 Measuring Influence Research on interest group influence, both within and outside of the EU, is also hindered by the difficulty of measuring influence (see also the more detailed discussion of influence measurement in Dür 2008b). Although researchers have made some advances, using processtracing, measures of attributed influence, and assessments of the degree of preference attainment, to determine interest group influence, all of these methods come with certain drawbacks. An awareness of these limitations should improve future research by allowing researchers to design projects that can offset the disadvantages of any one approach. For one, using process tracing, a qualitative assessment of interest group influence based on detailed knowledge of a case may be possible. Applying process-tracing to the study 16

of interest group influence, researchers examine the preferences of societal actors, their political activities, the way decisions are changed in the course of the political process, the extent to which interest group preferences are reflected in policy outcomes, and groups statements about their satisfaction with outcomes. Most studies of interest group influence in the context of the EU follow such an approach (see for example Cowles 1995; Moravcsik 1998; Warleigh 2000; the contributions in Pedler 2002; Dür and De Bièvre 2007; Michalowitz 2007). Alas, the resulting narrative case studies suffer from a series of problems. First, for such studies it is very difficult to assess the degree of influence, as a clear yardstick is missing. Findings are often framed as group X had (had no) influence on outcome Y, when actually the group may have had some influence, and it would be interesting to know exactly how much influence it had. Second, given the data intensity of process tracing, researchers using this method can only study few cases. Generalisations across issues and across policy fields from these very specific studies tend to be difficult. For example, it seems quite plausible that the European Roundtable of Industrialists had substantial influence over the passage of the Single European Act in the 1980s (Cowles 1995). But what do we learn from this about more general questions, such as the relative influence of concentrated and diffuse interests? Alternatively, researchers can try to measure attributed influence (March 1955), by asking groups to assess their own and/or other groups influence, or by inviting experts to gauge the influence of different groups. While quite widespread in the United States, this method, which measures perceptions of influence rather than influence as such, has been applied sparingly in the context of the EU. As an exception, Janet Edgell and Kenneth Thomson (1999) asked some twenty British interest groups to evaluate their influence on the Common Agricultural Policy. Andreas Dür and Dirk De Bièvre (2007) questioned interest groups on their influence on the EU s trade policy. The results attained from such studies, though, have to be treated cautiously as interest groups may have good reasons to either 17

overstate if they want to legitimate their existence vis-à-vis their members or understate to avoid the creation of counter-lobbies their policy impact. This problem can only be resolved if results are crosschecked with the assessments made by other groups and public actors. Expert assessments may also be unreliable, since these assessments may be shaped by prominent cases, or (even worse) by existing studies of influence that suffer from the shortcomings that cripple the other two methods to assess interest group influence presented here. Finally, research can assess the distance between actors preferences and policy outcomes across a series of cases. It may then be possible to infer influence from the observation that an actor repeatedly finds herself close to the policy outcome (Mahoney 2007). This is a dubious indicator of influence, though, since different causal effects may explain closeness. Consequently, some studies employ a more sophisticated method that uses the preferences of the public actors that have formal power of decision and decision-making rules to predict an outcome. The distance between this initially expected and the actually observable outcome then serves as an indicator of interest group power. In a study of the national positions of four EU members (Finland, Germany, Great Britain, and the Netherlands) concerning 15 Commission proposals, Gerald Schneider and Konstantin Baltz (2003) apply this method. They use the ideal point of the lead ministry in the policy area as the hypothetical outcome and then calculate the difference to the actual policy outcome. The larger the move towards the ideal point of a specific actor, the larger this actor s power is supposed to be. Again, this method comes with a series of drawbacks. First, when constructing hypothetical outcomes, the results depend on the correct modelling of the decision-making process based on preferences and institutional structures. This is difficult, not least because measuring the preferences (rather than positions) of public actors is very demanding. Second, studies in this vein tend to blackbox the process by which interests influence outcomes. 18

This is an advantage in the sense that it makes sure that even structural power can be taken into account; it is a disadvantage in the sense that these studies cannot explain how influence was exerted. Third, selection effects through which interest groups may determine the preferences defended by public actors are excluded from the analysis. Last, researchers using this approach find it difficult to consider the potential for counteractive lobbying (Austen- Smith and Wright 1994) and for piggybacking on the influence of other actors (in which case, a move towards the ideal point of a societal actor may be luck rather than influence; see Barry 1980). They hence are forced to argue that the inclusion of many cases in the analysis, which were selected randomly, should cancel out such influences (Schneider and Baltz 2003: 5; Mahoney 2007: 38). In sum, measuring interest group influence is difficult, and most methods that have been tried are problematic in one way or the other. Some of the contradictions in the empirical findings reported above may result from the use of different methods to measure influence. It appears plausible, for example, that studies using process-tracing tend to underestimate influence, while studies of the degree of preference attainment rather overestimate influence (Dür 2008b). Nevertheless, the drawbacks of different methods should not deter scholars from tackling the question of interest group influence. Indeed, the recent literature, by giving more attention to questions of method, has already provided some advances in measuring influence. Future studies should build on these advances and explicitly discuss the strengths and weaknesses of their methodological approach to allow for cumulative research. 5. Conclusion In the EU, as in other political systems, interest groups seem to be influential at some times and ineffective at other times. What are the factors that make them succeed or fail? I have summarised the literature tackling this question, showing that many hypotheses exist that stress such factors as interest group resources, political institutions, issue characteristics, and 19

interest group strategies as determinants of interest group influence. The review also has made evident that more research is needed to test these hypotheses. What are the lessons learned from the existing literature for how this should best be done? First, at a theoretical level, future work has to bear in mind that groups often counteract each others lobbying effort. Ignoring such countervailing pressures leads to the underestimation of the strength of different groups. Moreover, researchers should consider that the link between state and society is not a one-way street; public actors sometimes manage to manipulate societal actors because of the latter s need for financial resources (Mahoney 2004). Further theoretical work is also particularly important with respect to interest group strategies, since few hypotheses exist that link specific strategies to influence. In this respect, it will be essential to consider that interest groups may pick strategies that maximise influence in the long-term rather than the short-term. Second, although research on interest group influence has made important advances concerning measurement over the last decade, in the future scholars will have to resolve some problems that continue to create problems for existing measures. For example, there are very few studies that try to measure interest group influence at the agenda-setting stage, a problem captured by the idea of a second face of power. Furthermore, it will be important to find ways of measuring influence in situations in which different issues are linked to each other by way of vote trading, a practice that seems to be quite widespread in the EU. Combining different methods in single studies may be a strategy that helps scholars overcome shortcomings of individual methods (for such an approach in the context of the United States, see Smith 2000; see also the EAR method proposed by Arts and Verschuren 1999 that links attributed influence to process-tracing). Finally, while the multiplication of single-case studies has provided some insights into interest group influence in the EU, additional research should give more importance to comparative and large-n studies. One advantage of large-n studies is that they allow for 20

assessments of hypotheses that stress issue-specific factors; by contrast, in single-case studies such issue-specific factors have to be treated as constants. The work by Frank Baumgartner and others on lobbying success in the U.S. provides for a nice model that large-n research in the EU context could draw upon (Baumgartner et al. 2001). Cumulative research may also become easier if researchers clearly define and specify the concepts that they use in their studies. Hopefully, then, with researchers addressing the current problems in the literature on the influence of interest groups in the EU, a future review will be able to add new and interesting empirical findings to the ones mentioned in the present study. 21

Notes: 1 I accept a broad definition of interest groups including not only the traditional membership-based trade associations, but also firms and other organisations that try to influence policy outcomes in the EU. 2 Similarly, Frank Baumgartner and Beth Leech (1998: 13) single out the question of interest group power and influence in the United States as an area of confusion within the broader interest group literature. 3 For reviews of the broader literature on interest group influence in different national settings, which complement my focus on the EU, see for example Smith (1995) and Burstein and Linton (2002). Quantitative studies of interest group influence are reviewed in Potters and Sloof (1996). Van Winden (2003) provides an overview of formal models of interest group influence. 4 Institutions may also shape the chances for mobilisation (see also Lowery et al., this volume). I disregard the question of mobilisation here. 22

Bibliography Almond, Gabriel A. (1988). The Return to the State, American Political Science Review, 82:3, 853-74. Arts, Bas and Piet Verschuren (1999). Assessing Political Influence in Complex Decision-Making: An Instrument Based on Triangulation, International Political Science Review, 20:4, 411-24. Atkinson, Michael M. and William D. Coleman (1989). Strong States and Weak States: Sectoral Policy Networks in Advanced Capitalist Economies, British Journal of Political Science 19:1, 47-67. Austen-Smith, David and John R. Wright (1994). Counteractive Lobbying, American Journal of Political Science, 38:1, 25-44. Bachrach, Peter and Morton S. Baratz (1962). Two Faces of Power, American Political Science Review, 56:4, 947-52. Balanyá, Belén, Ann Doherty, Olivier Hoedeman, Adam Ma anit, and Erik Wesselius (2003). EUROPE INC: Regional & Global Restructuring and the Rise of Corporate Power. London: Pluto Press. Bandelow, Nils C., Diana Schumann, and Ulrich Widmaier (2000. European Governance by the Emergence of a New Type of Package Deals, German Policy Studies, 1:1, 8-38. Baumgartner, Frank, Jeffrey M. Berry, Marie Hojnacki, David C. Kimball, and Beth L. Leech (2001). Issue Advocacy and Interest-Group Influence, Paper presented at the First General Conference of the European Consortium for Political Research, 6-8 September 2001, Kent. 23

Baumgartner, Frank R. and Beth L. Leech (1998). Basic Interests: The Importance of Groups in Politics and Political Science. Princeton: Princeton University Press. Bennett, Robert J. (1999). Business Routes of Influence in Brussels: Exploring the Choice of Direct Representation, Political Studies, 47:2, 240-57. Bernhagen, Patrick (2007). The Political Power of Business: Structure and Information in Public Policymaking. London: Routledge. Beyers, Jan (2002). Gaining and Seeking Access: The European Adaptation of Domestic Interest Associations, European Journal of Political Research, 41:5, 585-612. Beyers, Jan (2004). Voice and Access: Political Practices of European Interest Associations, European Union Politics, 5:2, 211-40. Bouwen, Pieter (2004). Exchanging Access Goods for Access: A Comparative Study of Business Lobbying in the EU Institutions, European Journal of Political Research, 43:3, 337-69. Bouwen, Pieter and Margaret McCown (2007). Lobbying versus Litigation: Political and Legal Strategies of Interest Representation in the European Union, Journal of European Public Policy, 14:3, 422-43. Burstein, Paul and April Linton (2002). The Impact of Political Parties, Interest Groups and Social Movement Organizations on Public Policy: Some Recent Evidence and Theoretical Concerns, Social Forces, 81:2, 380-408. Coen, David (1997). The Evolution of Large Firm Political Action in the European Union, Journal of European Public Policy, 4:1, 91-108. Collignon, Stefan and Daniela Schwarzer (2003). Private Sector Involvement in the Euro: The Power of Ideas. London: Routledge. 24

Cowles, Maria Green (1995). Setting the Agenda for a New Europe: The ERT and EC 1992, Journal of Common Market Studies, 33:4, 501-526. Crombez, Christophe (2002). Information, Lobbying and the Legislative Process in the European Union, European Union Politics, 3:1, 7-32. Dahl, Robert A. (1961). Who Governs? Democracy and Power in an American City. New Haven: Yale University Press. Downs, Anthony (1957). An Economic Theory of Democracy. New York: Harper and Row. Dür, Andreas (2008a). Bringing Economic Interests Back Into the Study of EU Trade Policy-Making, British Journal of Politics and International Relations, 10:1, 27-45. Dür, Andreas (2008b). Measuring Interest Group Influence in the EU: A Note on Methodology, European Union Politics, 9:4. Dür, Andreas and Dirk De Bièvre (2007). Inclusion without Influence: NGOs in European Trade Policy, Journal of Public Policy, 27:1, 79-101. Edgell, Janet M. and Kenneth J. Thomson (1999). The Influence of UK NGOs on the Common Agricultural Policy, Journal of Common Market Studies, 37:1, 121-31. Eising, Rainer (2004). Multilevel Governance and Business Interests in the European Union, Governance: An Interdisciplinary Journal of Policy, Administration, and Institutions, 17:2, 211-45. Evans, Peter, Dietrich Rueschmeyer and Theda Skocpol, eds. (1985). Bringing the State Back In. Cambridge: Cambridge University Press. 25