National Interest Groups and Lobbying on European Union Legislation: The Case of Ireland Andreas Dür University of Salzburg [email protected] Gemma Mateo University of Salzburg [email protected]

Abstract What explains variation across national interest groups in their amount of lobbying on legislative proposals in the European Union (EU)? We present an argument that leads to the expectation that resource-rich associations and associations that are active in policy fields in which the EU has large competences engage in more lobbying on EU legislation than other associations. Moreover, business associations are expected to have privileged access to the European Commission and national governments; and citizens’ groups to parliaments. Using original data from a survey of interest groups in Ireland, we find support for these expectations. We conclude that national associations are heavily involved in EU lobbying, but that this lobbying is biased in favour of wellendowed business associations.

Introduction For some time, scholars and the public alike have discussed both the democratic legitimacy and quality of European Union (EU) decision-making. Interest group participation in the making of EU legislation may enhance the legitimacy and quality of decisions by allowing for different voices to be heard and expertise to be transmitted to decision-makers (Greenwood, 2007b). The involvement of interest groups, however, may also have the opposite effect, for example if civil society’s access to decision-makers is biased and the information that decision-makers receive from lobbyists is incomplete or even misleading. Any normative evaluation of the role of interest groups thus first requires an empirical assessment of which groups manage to gain access and be heard in EU decision-making. We approach this broad topic through an analysis of the lobbying activity by national-level interest groups (as opposed to European federations) on EU legislation. Many national-level associations such as chambers of commerce or citizens’ groups engage in lobbying not only on national but also EU-level decisions (Streeck and Schmitter, 1991: 137; Eising, 2009: 92).1 To analyze a potential bias in EU lobbying, we ask whether there is any variation in the amount of lobbying undertaken on EU legislation by these national associations. We develop an argument that leads to the expectation that resource-rich associations are disproportionally active with respect to advocacy on EU legislation. Moreover, our argument suggests a bias in favour of national-level business associations as compared to citizens’ groups and other types of associations in access to the European Commission and national governments, but not in access to parliaments. Furthermore, associations that are active in policy fields with a lot of EU legislation, such as agriculture and trade, should dedicate a larger share of their lobbying resources to the EU legislative process.

2

Data that originate from a survey of Irish associations that we carried out in late 2008 and early 2009 allow us to test these expectations. In our survey, we contacted 402 Irish associations, including agricultural associations, business associations, labour unions, nongovernmental organizations, and professional associations. Using a combination of online survey and hardcopy questionnaire, we received responses from 165 associations, amounting to a response rate of 41 percent. The quantitative analysis of the resulting data provides support for our hypotheses. Our overall findings suggest that contrary to the “compensation hypothesis” (for this term, see Beyers 2002), the possibility to lobby on EU legislation does not compensate for biases in access that may exist with respect to national legislation. The study nicely complements previous studies in the field. While in recent years an increasing number of studies has been written on the EU lobbying activities of national associations (Feld, 1966; Pappi and Henning, 1999; Richardson, 2000; Beyers, 2002; Constantelos, 2007; Geer et al., 2008; Eising, 2009; Klüver, 2010; Beyers and Kerremans, 2011), only a fraction of these studies address the question of variation across national associations in the amount of EU lobbying. The studies that come closest to our analysis focus on associations in Belgium, a small corporatist country (Beyers, 2002); firms and business associations in three large EU member states (France, Germany, and the United Kingdom) (Eising 2009); large professional associations in the health sector across the EU (Greer et al. 2008); agricultural associations from France and Germany (Klüver 2010); and groups aiming at influencing the EU’s external trade policy (Beyers and Kerremans, 2011). By looking at associations of all types and across all sectors in a small country with a pluralist system of interest intermediation, our study expands the empirical reach of this literature. Moreover, since our data is from 2008 and 2009, our analysis

3

allows for an update of findings reported in Beyers (2002) and Eising (2009), both of whom rely on data from the late 1990s. A further important difference to existing studies is that we are not solely interested in EU-level lobbying, but in lobbying on EU legislation, whether this takes place at the national or at the European level. The paper’s theoretical contribution lies in the integration of a series of hypotheses from the relevant literature in a coherent theoretical argument whose assumptions are clearly spelled out. The paper’s argument and evidence contribute to the broader debate on the role of civil society in international institutions and governance beyond the nation state. It allows for answers to questions such as: how do interest groups behave in a system of multiple levels of governance? What explains the extent to which interest groups from one level engage in lobbying at another level of governance? Our analysis suggests that interest groups are able to move beyond the level of governance at which they are predominantly organized to engage in lobbying at a higher level of governance, but that in this process pre-existing biases in favour of some groups are reproduced. Explaining EU Lobbying by National Associations The EU has an increasingly large impact on the chances of societal actors to achieve political outcomes in line with their preferences. The EU’s policy output has major implications for actors as diverse as import-competing companies that want to receive protection

against

foreign

imports,

chemical

producers

concerned

about

environmental standards, and development groups interested in the provision of aid to developing countries. Given the importance of the EU for these societal actors, the expectation is for a large number of them to try to influence EU legislation. These influence attempts can take place through at least three channels: through national

4

political actors who then may defend the lobbyists’ interests in the Council of Ministers and vis-à-vis the European Commission and the European Parliament; through EU federations that mainly interact with the EU institutions; and directly with the EU-level institutions, in particular the European Commission and the European Parliament. While thus many national-level associations have a stake in EU legislation, in line with the existing literature we expect there to be systematic variation across associations in the amount of EU lobbying. Our argument starts with the assumption that each association has a fixed and limited amount of resources, where the overall endowment with resources varies across associations, that is, some associations have many resources and others few. Relevant resources are not only financial means, but also legitimacy and representativeness, knowledge, and expertise and information (Dür 2008: 1214). Such resources are necessary for influencing policy outcomes for several reasons: associations need resources to be informed about relevant policy developments, to organise lobbying activities such as manifestations and press conferences, and to use as currency in exchange for access and influence from decision-makers. Associations can decide to allocate a larger or a smaller amount of their resources to lobbying on EU legislation, but they cannot (in the short run) increase their overall amount of resources. The fact that the overall resources are limited introduces opportunity costs: the more an association invests on lobbying on national legislation, the less it can spend on EU legislation, and vice versa. A further assumption of the argument is that associations act with the aim of maximising their benefits from influencing legislation. These benefits concern both the direct gains from achieving a specific policy output that is in line with their interests (the logic of

5

influence, in the terminology of Schmitter and Streeck, 1981) and the indirect gains that arise if influencing policy allows them to increase their number of members and/or financial contributions (the logic of membership). Following on from these assumptions, we conjecture that the larger the share of overall resources that an association dedicates to EU legislation, the larger the gains it derives from EU legislation (and the lower the gains from national legislation, since fewer resources will be spent there). This strictly positive relationship between resources spent and gains can be understood in two different (but complementary) ways. On the one hand, it can be assumed that benefit is a continuous variable and the investment of resources on lobbying leads to an incremental increase in the value of this variable (that is, as an association increases the lobbying effort on a piece of legislation, it can expect an increasingly favourable policy output). On the other hand, such a strictly positive relationship also arises if lobbying (even if not always successful) increases the probability of getting a discrete benefit. At the same time, starting with a certain level of lobbying, there are diminishing returns to investing ever more resources on a specific piece of legislation.2 This is so because policy makers will find it unattractive completely to give in to the pressures from one interest group. By contrast, they will find it easy to give a small concession to a group that was heavily disadvantaged in an earlier proposal. Given these assumptions, variation in the resource endowments of associations should lead to variation in the level of lobbying on EU legislation (see also Klüver, 2010). We expect that there are many relatively small associations that find it difficult to get politically active independent of the level at which authority is located in a multi-level system. Others, by contrast, should be well organised internally and have the necessary resources to monitor and try to influence legislation at all relevant levels

6

of governance. Consequently, those associations that are well-endowed with resources should do more lobbying on EU legislation than poor associations. In form of a hypothesis, the more resources that an association has, the more active it is with respect to lobbying on EU legislation (Hypothesis 1). In line with the existing literature on lobbying in the EU, we also expect differences across types of associations in the amount of EU lobbying (see, for example, Schneider and Baltz, 2003; Coen, 2007: 335). This is so because business associations are more likely to possess information and expertise than other types of associations. Business associations have a much more well-defined constituency with concentrated material gains or losses from a specific piece of legislation than citizens’ groups (or professional associations that are composed of a large number of individuals). Collective action problems should therefore be less severe for business associations that protect the material interests of a relatively small number of members than for other groups that often defend a diffuse interest (Olson, 1965; Dür and De Bièvre, 2007). Collective action problems should hinder resource acquisition by diffuse groups that find it difficult to galvanize their members or followers in opposition to or in support of a specific legislative proposal. Activating members should be particularly difficult for citizens’ groups with respect to EU legislation, which tends to receive less public attention in the media than national legislation (Peter et al., 2003; Meyer, 2005). Business associations’ ability to supply information and expertise should provide them with access to political institutions that require these types of resources. With respect to EU legislation, the Commission and the national governments are generally considered to be most in need of these types of resources (Bouwen 2004). A first hypothesis with respect to differences between types of groups hence is that

7

national business associations have more contacts with the Commission and national governments with respect to EU legislation than other types of associations (Hypothesis 2a). At the same time, business associations should be at a disadvantage relative to citizens’ groups in supplying resources such as representativeness and legitimacy that are in relatively higher demand as compared to specialized expertise in the European and national parliaments. We thus also expect that citizens’ groups have more contacts with the European and national parliaments with respect to EU legislation than other types of associations (Hypothesis 2b). With citizens’ groups at a disadvantage compared to business associations with respect to inside lobbying, we expect them to engage in more outside lobbying. Inside lobbying refers to the process of establishing direct contacts with decisionmakers with the aim of influencing policy output, while outside lobbying refers to any activity that aims at either changing public opinion or mobilising latent public support for a position with the aim of influencing policy output (Kollman, 1998). In form of a hypothesis, citizens’ groups are more active in outside lobbying with respect to EU legislation than business associations (Hypothesis 2c). Finally, we expect variation across policy fields in the amount of EU lobbying (see also Mazey and Richardson, 2006). In fields with substantial EU authority, the potential gains from lobbying (and the potential costs from failing to lobby) on EU legislation should be higher than in other fields, therefore providing an incentive for associations to concentrate their lobbying effort on EU legislation. This is in line with V.O. Key’s (1956: 168) adage that “where power rests, there influence is brought to bear.” Our third hypothesis hence is that associations that are active in policy fields with high EU competencies are relatively more involved in EU legislative lobbying than other associations (Hypothesis 3).

8

The Data We rely on a survey of Irish associations that we carried out in late 2008 and early 2009 to test these expectations. As with any study that relies on data from only one country, generalizations to other countries have to be made with much caution. Nevertheless, the single country approach seems less problematic in our case than in other studies, as with few exceptions, the causal relationships postulated in our hypotheses should not be influenced by variables that differ across countries, such as domestic political institutions. While some studies have suggested that the type of interest group system (pluralist, corporatist or statist) influences the ease with which national associations adapt to the challenge of Europeanization (for example, Cowles, 2001; Klüver, 2010; but see Eising 2009: 168), it is far less obvious that this variable impacts on the variation across associations that we tackle in this paper. For our survey, we asked 402 Irish associations to respond to a questionnaire on their lobbying activity (for a more detailed discussion of our approach, see authors). The population that we considered were all Republic of Ireland-based associations included in the Administration Yearbook & Diary 2008 (Institute of Public Administration, 2008), except sporting associations, political parties, and learned societies. We are confident that the yearbook’s listings are reasonably comprehensive and that our panel is a good representation of the population of associations that potentially engage in lobbying. What we did not consider were firms that may engage in lobbying independent of and in addition to (business) associations. Broken down by type of association, we contacted 155 business associations (38 percent of the total), 115 citizens’ groups (29 percent), 86 professional associations (21 percent), 17 agricultural associations (4 percent), 17 labour unions (4 percent),

9

and 12 (single-issue) associations that we could not classify in this typology (3 percent). We contacted all associations by e-mail, asking representatives of the associations to complete an online questionnaire.3 To maximise the response rate, we also sent hardcopies of the questionnaire to associations that did not respond to two email invitations. In the end, we achieved a response rate of 41 percent. This response rate, which we consider to be relatively high, should ensure that our findings are representative of the population of nationally-visible Irish associations in the categories distinguished above, even though it seems plausible that there is a slight bias in favour of responses from associations that are politically active. By type of associations, the response rates are as follows: 53 percent for labour unions, 47 percent for professional associations, 42 percent for citizens’ groups, 35 percent for agricultural associations, and 34 percent for business associations. The responding associations have a mean number of staff of 13 (when excluding one outlier with more than 1000 employees) and a mean number of staff working on advocacy of 2. In the sample, 86 percent of associations are membership based and 49 percent form part of a European federation such as the European Banking Federation. The survey data allow us to define two sets of dependent variables: one capturing the frequency of contacts between associations and a series of political institutions at the national and the EU level with respect to EU legislation and the other capturing the frequency of use of a series of lobbying tactics with respect to EU legislation (see Table 1). The data for the first set of dependent variables stem from the following questions: “How often have you been in contact with the following Irish [in a further question, replaced with “EU”] institutions with respect to EU legislative proposals over the last two years?” The Irish institutions that we listed were the top

10

level of government (ministers etc.), the government bureaucracy (officials in ministries etc.), Irish regulatory agencies, deputies in the Dáil (the lower chamber of the Irish parliament), deputies in the Seanad Éireann (the higher chamber of the Irish parliament), political parties, and committees of the Oireachtas (the Irish parliament). The EU-level institutions that we listed were the top level of the European Commission, desk officers (or equivalent) in the European Commission, EU regulatory agencies, national representatives in the COREPER, the Irish Permanent Representation in Brussels, members of the European Parliament, European Parliament committees, the European Commission representation in Ireland, and the European Parliament office in Ireland. Respondents could indicate that they had no, one, 2-5, 6-9, 10-19, or 20 and more contacts with these institutions over the last two years. For the following analysis, we recoded the variable to four levels, ranging from no contact to six or more, as only very few associations reported very high levels of contact.

TABLE 1 ABOUT HERE

The data for the second set of dependent variables result from a question on the frequency of use of various actions to influence EU legislation (“Please indicate how often your organisation uses the following actions to influence EU legislation”). The tactics that we listed were direct contacts with policymakers and/or public officials, participating in meetings organised by political institutions, preparing a detailed position paper, distributing folders and brochures, organising or participating in demonstrations and/or street actions, distributing a press release, organising a press conference, initiating a public debate on the internet, trying to mobilise other

11

associations/interest groups, and hiring a consultant. Respondents could indicate that they use these tactics never, less than yearly, once a year, about 2-5 times a year, about 6-9 times a year, about 10 to 15 times a year, approximately biweekly, or on a weekly basis. Again, we recoded the variable, this time using five levels, ranging from never to six or more times a year. Hypothesis 1 stresses variation in the resource endowment of associations. We use the number of staff working in public affairs and advocacy to operationalize this variable. In the absence of reliable information on the financial resources of associations, this operationalization seems preferable to one relying on the overall number of staff working for the association, as some associations (especially citizens’ groups and professional associations) have a large number of staff working on the provision of services to members or the broader community (for example, health groups that are active in the provision of specific health services). In fact, the association with the largest number of employees among our respondents is a citizens group providing health services. The distinction between types of association, relevant for hypotheses 2a-2c, is based on our own coding of associations. In some cases, this coding proved difficult. Especially the distinction between business and professional associations was not always clear-cut. For the purpose of this paper, we defined professional associations as groups for which membership is linked to a specific training, and where most members are individuals. Business associations, by contrast, mainly have companies as members. Finally, Hypothesis 3 refers to variation in EU competencies across policy fields. What we are interested in is a measure of “the relative importance of Community decision-making processes as compared with national processes” (Lindberg and Scheingold, 1970: 68). After considering several measures of the extent

12

of EU policy competencies by policy field, we relied on Tanja Börzel’s (2005) coding of the level and scope of European integration, using provisions as set down in the Treaty of Nice (which entered into force in 2003 and remained the relevant treaty basis until November 2009). Börzel applied a five-point scale ranging from exclusive national competencies (score of 1) to exclusive European competencies (score of 5) for level of authority and a six-point scale ranging from no coordination at EU level (score of 0) to supranational centralization (score of 5) for scope of authority.4 We multiplied scope with level to get a single measure of the EU’s competencies by policy field. A potential drawback of this measure is that it only captures formal competencies, but not whether the EU institutions actually make use of these competencies. As a robustness check, we thus also used the number of policy proposals, recommendations and consultations for which a Directorate General had primary responsibility in 2007 and 2008.5 One problem that we faced in elaborating this measure was to map directorates to policy fields.6 Nevertheless, the resulting values are intuitive and largely coincide with the values using the EU’s formal competencies: trade, transport, environment, and agriculture turn out to be the policy fields with most EU authority, whereas the EU has least competence for cultural, health, and research policy. For the analysis, we used this classification of policy fields in combination with data from the survey in which we asked associations to specify the importance of thirteen policy fields for them. Associations could indicate that a policy field was very important, important, somehow important, little important, or not at all important in their work. Combining the classification of policy fields with this data allowed us to calculate two variables: the first adds the values for each association across the four policy fields with the highest degree of EU competencies (giving a score of 0 to

13

associations that responded that all four policy areas were not at all important in their work and a score of 16 to associations that considered all of them very important) and the second sums the values for the three policy fields with the lowest degree of EU competencies (for a variable that varies from 0 to 12). The Findings Having discussed data sources and operationalization, we first present bivariate tests of our hypotheses, before examining the hypotheses by way of regression analysis. Bivariate analysis In the first hypothesis, we have suggested that those associations that are well endowed with resources should be particularly active on EU legislation. For this bivariate test, we recoded the variable number of staff working on public relations into four levels: no staff working in the area of public relations; up to and including one full-time staff; more than one but no more than two; and more than two full-time staff (we had asked respondents to report staff that only partially work on advocacy in decimals, for example 0.5 if a person dedicates 50 percent of her time to public relations). This categorization makes sense given that the minimum value on this variable is 0, the median 1, the mean 1.9, and the maximum 18. For each of the four levels, we then calculated the mean frequencies of contact on EU legislation with six political institutions. The results, which are shown in Figures 1a through 1f, clearly support our expectation: an increase in the number of staff working in the field of public relations boosts an association’s activity on EU legislation. The patterns are very similar across political institutions.7

FIGURES 1a-1f ABOUT HERE

14

Hypotheses 2a and 2b suggest that national business associations have an advantage over other national associations in getting access to executive institutions and citizens’ groups over other national associations in getting access to parliaments with respect to EU legislation. We again calculate mean frequencies of contact, this time for three types of associations: business associations, citizens’ groups, and professional associations.8 As before, the results are in line with our expectations (see Figures 2a through 2f). Business associations clearly have better access to executive institutions than other associations. By contrast, citizens’ groups have more contacts than business associations with both members of the European Parliament and deputies in the lower chamber of the Irish parliament. The finding of little lobbying on EU legislation by professional associations is in line with existing research that suggests that professional associations engage in less EU lobbying than both business associations and citizens’ groups (Greenwood, 2007a).

FIGURES 2a-2f ABOUT HERE

Hypothesis 2c distinguishes between inside and outside lobbying tactics. In Figures 3a-3f we show by type of association the mean use of six different tactics, three of which can be considered inside (direct contacts, participation in meetings, and preparation of position papers) and three outside (press releases, launching of an internet debate and participating in a demonstration). While the differences between business associations and citizens’ groups are small with respect to direct contacts and participation in meetings, Figures 3d-3f provide support for our expectation that citizens’ groups engage in more outside lobbying on EU legislation than other types of associations.

15

FIGURES 3a-3f ABOUT HERE

Finally, in hypothesis 3 we link activity in policy fields with large EU competencies with associations’ amount of EU lobbying. For this bivariate test, we distinguished four levels of activity in these four policy fields (agriculture, environment, trade, and transport): no activity (these policy fields are not at all important for an association), low activity, medium activity, and high activity (an association considers the four policy fields at least important on average). Again, the results provide support for our hypothesis (Figures 4a to 4f). The effect is weakest, however, for contacts on EU legislation with the top level of the European Commission. Overall, the bivariate analysis has corroborated our hypotheses. In the following regression analysis, we assess whether these results are robust in a multivariate setting.

FIGURES 4a-4f ABOUT HERE

Regression Analysis To test the hypotheses proposed above in a multivariate setting, we use ordinal logistic regression.9 This is the appropriate approach given the ordinal nature of our dependent variables (from 1=no contact to 4=six or more contacts in the last two years and from 0=does not use a tactic to 4=uses the tactic six or more times a year). An assumption underlying ordinal logistic regression is that one set of coefficients can capture the relationship between any category and all other categories. This means, for example, that the same coefficients have to apply to the difference between no

16

contact and one or more contacts and to the difference between six or more contacts and fewer than six contacts. We checked that this proportional odds assumption holds in our models by running a series of binary logistic regressions and comparing differences in coefficients across levels (for this approach, see Harrell 2001: 331-44). The independent variables included in the model are those described above and summarized in Table 1. A dummy variable for sub-national chambers of commerce (of which there are seven in the database) serves as control variable, as such associations can be expected to do less lobbying on EU legislation than other business associations. Because of missing values, the number of cases slightly varies across the four models that we estimate. In Models I to VI, the dependent variable is the frequency of contact with a series of EU and national-level political institutions with respect to EU legislation. The models’ findings are supportive of our hypotheses (see Table 2). Looking at Hypothesis 1 first, the coefficients for number of staff working in public relations are statistically significant at the 95 percent level in four of the six models and at the 90 percent level in a further model. As we standardized all inputs to the regression (by subtracting the mean and dividing by two standard deviations), the size of the coefficients can be directly compared across models. This comparison shows that the effects are strongest for deputies in the Dáil and members of the European Parliament. By contrast, the effect is weakest (and not statistically significant) in the case of contacts with the top levels of the Irish government.

TABLE 2 ABOUT HERE

17

The data also back our expectations with respect to variation across types of associations. In support of Hypothesis 2a, business associations engage in more lobbying on EU legislation with executive institutions than other types of associations even when controlling for other variables. Business associations’ advantage is slightly larger with respect to contacts with the Irish government (both the top level and the administration) than with the European Commission. As expected by Hypothesis 2b, business associations do not have more contacts with members of parliament. While our reasoning leads us to the expectation that agricultural associations should also engage in more lobbying of executive institutions than other associations, the data only weakly back this expectation. The reason for the lack of statistical significance of the coefficients for this variable, however, is most likely to be found in the low number of agricultural associations in the sample (only six). The final variable related to the type of association, namely sub-national chamber of commerce, has the expected sign in all models and is statistically significant in five of the six models. The models also provide support for Hypothesis 3, namely that associations that are active in policy fields with large EU competencies engage in more lobbying on EU legislation than other associations. The coefficients for the variable capturing activity in highly integrated policies have the right sign and are statistically significant in all of the models (though weakly so in Model VI). The effect of this variable is stronger for the EU institutions than for the national ones. The mirror variable capturing activity in three policy fields with low EU competencies (culture, health and research), by contrast, is not statistically significant in any of the models. Across all models, the pseudo R2 varies between 0.15 and 0.22, indicating that the fit is decent, but that also many factors not captured in the models influence groups’ decisions on how much to lobby on EU legislation.

18

The substantive effect of the key variables is large. Table 3 shows the changes in probabilities of an association falling into a specific category of the dependent variable for a change in the business variable from 0 (not a business association) to 1 (a business association) and in the advocacy staff variable from 0 (the minimum) to 18 (the maximum). For the case of the top level of the European Commission, business associations have a predicted probability of 33 percent of having had two or more contacts in the last two years, while an association with the same characteristics that is not a business association only has a predicted probability of 18 percent of having had as many contacts. The difference is even larger for the desk level in the Commission, with the predicted probabilities being 27 percent and 45 percent respectively. The effect of the number of staff working on advocacy variable is even stronger: for the top level of the Commission, for example, the probability of having had two or more contacts increases from 16 to 54 percent when moving from the minimum to the maximum on this variable. The probability of such a resource-rich association having had two or more contacts on EU legislation with the top level of the Irish government is even 76 percent (as compared to 46 percent for an equivalent association without staff working on advocacy).

TABLE 3 ABOUT HERE

To check the robustness of these findings and to provide a test of Hypothesis 2c, we rely on a second dependent variable, capturing associations’ frequency of use of a series of tactics to influence EU legislation. The tactics that we look at are those mentioned before (three inside tactics, namely direct contacts, participation in meetings, and preparation of position papers, and three outside tactics, namely press

19

releases, launching of an internet debate and participating in a demonstration). The results are encouraging.10 As before, Hypothesis 1 receives considerable support across five of the six models (see Models VI to XII in Table 4). In support of Hypothesis 2c, business associations are less likely to participate in demonstrations or engage in internet debates than citizens’ groups and professional associations. By contrast, business associations are more active with respect to position papers than other types of associations (and here the effect is only weakly statistically significant). The type of association does not have an effect on the frequency of use of direct contacts and the participation in meetings to influence EU legislation.11 In support of Hypothesis 3, the coefficient for activity in highly integrated policy areas is positive and (at least weakly) statistically significant in three of the six models. In the other three models, we cannot reject the null hypothesis of no difference between groups active in highly integrated policy areas and other groups.

TABLE 4 ABOUT HERE

Conclusion We have assessed Irish associations’ lobbying activity on EU legislation. The analysis has demonstrated that there is a substantial amount of lobbying on EU legislation by Irish associations. Lobbying on EU issues clearly has not been fully outsourced to EU federations, a finding that potentially has far-reaching consequences for the assessment of the democratic legitimacy of EU policies. The large number and variety of groups that engage in lobbying on EU legislation may allow the EU to claim input legitimacy. The detailed analysis of the data, however, paints a less benign picture of the EU’s democratic legitimacy, at least as far as access by organized groups is

20

concerned. In particular, we see our findings about variation in the amount of lobbying across associations as a confirmation of the view that sees the EU’s system of interest intermediation as a form of elite pluralism (Coen, 1997). Since wellendowed groups find it easier than other groups to advocate their interests, and business groups have better access to the Commission and national governments, we expect there to be a bias in the views heard by political decision-makers. Two potential caveats to these broader interpretations of our findings are important to note. First, access does not necessarily translate into influence (Dür, 2008). It is possible that when assessing influence, the findings reported above are inverted. We do not think that this is plausible, however. In fact, we rather suspect that assessing influence would expose an even greater bias in favour of resource-rich business associations. This conjecture is backed by a previous study of trade policy lobbying that found that citizens’ groups have relatively good access to EU policymakers, but that they had little impact on outcomes (Dür and De Bièvre, 2007). Second, we neglected firm lobbying in our analysis (for lobbying by firms in the EU, see for example McLaughlin et al., 1993; Coen, 1997; Bennett, 1999; Bernhagen and Mitchell, 2009). Again, however, we would argue that when also considering lobbying by firms, the difference between business associations and citizens’ groups is largely to appear even larger than in this study. Overall, therefore, it seems plausible that the current system of interest group representation is not a panacea for the EU’s democratic deficit.

References Bennett, R. J. (1999) ‘Business Routes of Influence in Brussels: Exploring the Choice of Direct Representation’, Political Studies 47(2): 240-57.

21

Bernhagen, P. and N. J. Mitchell (2009) ‘The Determinants of Direct Corporate Lobbying in the European Union’, European Union Politics 10(2): 155-176. Beyers, J. (2002) ‘Gaining and Seeking Access: The European Adaptation of Domestic Interest Associations’, European Journal of Political Research 41(5): 585-612. Beyers, J. and B. Kerremans (2011) ‘Domestic Embeddedness and the Dynamics of Multi-level Venue-Shopping in Four EU Member-States’, Governance: An International Journal of Policy and Administration (forthcoming). Börzel, T. A. (2005) ‘Mind the Gap! European Integration between Level and Scope’, Journal of European Public Policy 12(2): 217-36. Bouwen, P. (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. Coen, D. (1997) ‘The Evolution of Large Firm Political Action in the European Union’, Journal of European Public Policy 4(1): 91-108. Coen, D. (2007) ‘Empirical and Theoretical Studies in EU Lobbying’, Journal of European Public Policy 14(3): 333-45. Constantelos, J. (2007) ‘Interest Group Strategies in Multi-level Europe’, Journal of Public Affairs 7(1): 39-53. Cowles, M. G. (2001) ‘The Transatlantic Business Dialogue and Domestic BusinessGovernment Relations’, in M. G. Cowles, J. Caporaso, and T. Risse-Kappen (eds) Transforming Europe: Europeanization and Domestic Political Change, Ithaca: Cornell University Press, pp. 159-79. Dür, A. (2008) ‘Interest Groups in the European Union: How Powerful Are They?’, West European Politics 31(6): 1212-30.

22

Dür, A. and D. De Bièvre (2007) ‘Inclusion without Influence? NGOs in European Trade Policy’, Journal of Public Policy 27(1): 79-101. Eising, R. (2009) The Political Economy of State-Business Relations in Europe: Modes of Interest Intermediation, Varieties of Capitalism, and the Access to EU Policy-making, London: Routledge. Feld, W. (1966) ‘National Economic Interest Groups and Policy Formation in the EEC’, Political Science Quarterly 81(3): 392-411. Greenwood, J. (2007a) Interest Representation in the European Union, Houndmills: Palgrave. Greenwood, J. (2007b) ‘Review Article: Organized Civil Society and Democratic Legitimacy in the European Union’, British Journal of Political Science 37(2): 333-57. Greer, S. L., E. Massard da Fonseca and C. Adolph (2008) ‘Mobilizing Bias in Europe: Lobbies, Democracy and EU Health Policy-Making’, European Union Politics 9(3): 403-433. Harrell, F. E. (2001) Regression Modeling Strategies: With Applications to Linear Models, Logistic Regression, and Survival Analysis, New York: Springer. Institute of Public Administration (2008) Administration Yearbook & Diary 2008, Dublin: IPA. Key, V. O. (1956) American State Politics: An Introduction, New York: Knopf. Klüver, H. (2010) ‘Europeanization of Lobbying Activities: When National Interest Groups Spill Over to the European Level’, Journal of European Integration 32(2): 175-91. Kollman, K. (1998) Outside Lobbying: Public Opinion and Interest Group Strategies, Princeton: Princeton University Press.

23

Lindberg, L. N. and S. A. Scheingold (1970) Europe’s Would-be Polity: Patterns of Change in the European Community, Englewood Cliffs: Prentice Hall. Mazey, S. and J. Richardson (2006) ‘Interest Groups and EU Policy-Making: Organisational Logic and Venue Shopping’, in Jeremy Richardson (ed.) European Union: Power and Policy-Making, London: Routledge, pp. 247-68. McLaughlin, A. M., G. Jordan and W. Maloney (1993) ‘Corporate Lobbying in the European Community’, Journal of Common Market Studies 31(2): 191-212. Meyer, C. O. (2005) ‘The Europeanization of Media Discourse: A Study of Quality Press Coverage of Economic Policy Co-ordination since Amsterdam’, Journal of Common Market Studies 43(1): 121-148. Olson, M. (1965) The Logic of Collective Action, Cambridge: Harvard University Press. Pappi, F. U. and Ch. H. C. A. Henning (1999) ‘The Organization of Influence on the EC’s Common Agricultural Policy: A Network Approach’, European Journal of Political Research 36(2): 257-81. Peter, J., Semetko, H.A., de Vreese, C.H. (2003) ‘EU Politics on Television News: A Cross-National Comparative Study’, European Union Politics 4(3): 305-327. Richardson, J. (2000) ‘Government, Interest Groups and Policy Change’, Political Studies 48(5): 1006-23. Schmitter, P. C. and W. Streeck (1981) ‘The Organization of Business Interests: A Research Design to Study the Associative Action of Business in the Advanced Industrial Societies of Western Europe’, Wissenschaftszentrum Paper IIM/LMP 81-13. Schneider, G. and K. Baltz (2003) ‘The Power of Specialization: How Interest Groups Influence EU Legislation’, Rivista di Politica Economica 93(1-2): 253-83.

24

Streeck, W. and P. C. Schmitter (1991) ‘From National Corporatism to Transnational Pluralism: Organized Interests in the Single European Market’, Politics & Society 19(2): 133-64.

25

1

We use the term citizens groups to refer to groups whose members or supporters do not have a direct

and concentrated material stake in the policies advocated by the association. 2

Initially, however, there may be increasing returns to investing more on a piece of legislation, as an

association has to achieve a certain amount of visibility to effectively defend a stance. 3

For this online survey, we used an open-source tool called LimeSurvey.

4

For four policy field mentioned in our questionnaire (development cooperation, health, migration, and

transport) we had to do the coding ourselves (relying on the same criteria as those used in the original study). The values that we assigned for these policy fields are 3, 1, 3, and 3 (level) and 2.5, 2, 2, and 4.5 (scope). 5

The data for this measure are from the register of Commission documents available at

http://ec.europa.eu/transparency/regdoc/recherche.cfm?CL=en (as of October 2009). 6

In particular, where a Directorate General (DG) has competencies for several of the policy fields

distinguished here (for example, consumers and health), we manually coded the proposals as belonging to a specific policy field. For migration, we looked at the proposals listed as emanating from the DG Justice, Freedom and Security, and selected only proposals dealing with migration, defined broadly as covering immigration, asylum, human trafficking, and border control. 7

The pattern is also similar when using tactics as dependent variable (not shown for space reasons). A

positive relationship with the number of advocacy staff exists for all tactics. 8

Because of the small number of agricultural associations and labour unions in our dataset, making

claims about these types of groups is not meaningful. 9

We used R Version 2.12.0 for the statistical analysis.

10

In Models XI and XII we had to exclude the dummy variable for sub-national chambers of

commerce, as these chambers did not show any variation on the dependent variable, making the models unstable. 11

This finding is not astonishing considering the results presented above: business associations have

more contacts with the European Commission and government actors, but fewer contacts with members of parliament. In the question on frequency of use of tactics, these two effects offset each other, leading to a zero effect.

26

Table 1: The data Dependent variable I (institutions) Dependent variable II (tactics) Staff Type (Agricultural, business, labour union, citizens group, professional, and other) Sub-national chambers Highly integrated policies

Little integrated policies

Frequency of contacts on EU legislation across a series of EU and Irish institutions (ranges from 1=never to 4=more than 6 times in the last two years) Values across a list of tactics used on EU legislation (ranges from 0=never to 4=more than 6 times a year) Number of staff working on advocacy in an association Dummy variables based on our own coding of associations. Takes the value 1 for sub-national chambers of commerce; 0 otherwise. Sum of values an association gives to the policy fields agriculture, environment, trade and transport (policy fields based on Börzel 2005 and number of legislative acts per directorate) [ranges from 0=none of the four policy areas of any importance to 16=all four policy areas very important]. Sum of values an association gives to the policy fields culture, health and research (policy fields based on Börzel 2005 and number of legislative acts per directorate) [ranges from 0=none of the three policy areas of any importance to 12=all three policy areas very important].

27

Table 2: Explaining variation in EU lobbying across associations, by institution I II III IV V VI Comm. Comm. European Irish Irish Dáil Predictor top desk Parliament government admin. Staff 0.55* 0.63** 0.83** 0.39 0.78** 1.01*** (0.29) (0.32) (0.34) (0.33) (0.37) (0.37) Business 0.77** 0.78** 0.12 0.96*** 0.91*** -0.06 associations (0.34) (0.34) (0.32) (0.34) (0.35) (0.33) Agricultural 0.57* 0.60* 0.41 0.41 0.27 0.15 associations (0.30) (0.31) (0.31) (0.29) (0.30) (0.28) Subnational -0.97** -1.19** -0.36 -0.81** -0.72** -0.82** chambers (0.49) (0.49) (0.34) (0.34) (0.36) (0.39) Active in highly 0.83** 1.12*** 1.14*** 0.73** 0.79** 0.68* integrated policies (0.40) (0.40) (0.36) (0.36) (0.37) (0.36) Active in little -0.17 0.13 -0.26 0.01 -0.26 0.14 integrated policies (0.40) (0.38) (0.36) (0.36) (0.37) (0.35) Intercept 1 -0.41** -0.13 0.72*** 1.01*** 1.49*** 0.85*** (0.18) (0.18) (0.18) (0.19) (0.21) (0.18) Intercept 2 -1.3*** -0.81*** -0.06 0.28 1.06*** 0.5*** (0.21) (0.20) (0.17) (0.17) (0.19) (0.18) Intercept 3 -3.74*** -1.88*** -1.57*** -1.78*** -0.56*** -1.05*** (0.19) (0.48) (0.25) (0.22) (0.23) (0.18) N 152 148 142 153 152 154 2 Pseudo R 0.16 0.22 0.15 0.15 0.16 0.15 Note: ordinal logistic regression coefficients with standard errors in parentheses. All input variables are standardized. *** p<0.01; ** p<0.05; * p<0.1 (two-tailed tests) Table 3: Prediced probabilites of contact frequencies for selected institutions No One 2-5 6 or more contacts contact contacts contacts Commission Business=0 65 17 16 2 top Business=1 45 21 29 4 Advocacy staff=0 69 15 14 2 Advocacy staff=18 26 20 45 9 Commission Business=0 58 15 16 11 desk Business=1 37 17 23 22 Advocacy staff=0 63 14 14 9 Advocacy staff=18 17 12 25 46 Irish top Business=0 33 18 39 11 Business=1 15 12 47 26 Advocacy staff=0 36 18 36 10 Advocacy staff=18 13 11 47 29 Note: the table reports predicted probabilities (in percent) for an association falling into one of the levels of the dependent variable. The probabilities are calculated while keeping all other variables constant (continuous variables at the mean and nominal ones at 0).

28

Table 4: Explaining variation in EU lobbying across associations, by tactics VII VIII IX X XI XII Predictor Direct Meetings Position Press Internet Demonlobbying papers releases debate strations Staff 0.88*** 1.10*** 1.18*** 0.97*** 0.30 0.70** (0.31) (0.30) (0.35) (0.33) 0.34 0.34 Business 0.00 -0.01 0.56* -0.22 -0.74* -1.62*** associations (0.32) (0.32) (0.32) (0.32) 0.42 0.62 Agricultural 0.36 0.42 0.53* 0.21 -0.26 -0.06 associations (0.32) (0.31) (0.30) (0.28) 0.44 0.44 Sub-national -0.36 0.13 -0.57* -0.22 chambers (0.30) (0.31) (0.32) (0.31) Active in highly 1.01*** 0.50 0.63* 0.81** -0.13 -0.03 integrated policies (0.35) (0.35) (0.34) (0.34) 0.42 0.50 Active in little 0.04 -0.09 0.21 0.01 0.83* 0.06 integrated policies (0.33) (0.34) (0.34) (0.35) 0.43 0.48 Intercept 1 1.51*** 0.61*** 1.53*** 1.03*** -1.07*** -1.71*** (0.21) (0.17) (0.21) (0.19) 0.19 0.26 Intercept 2 0.16 -0.30* -0.07 0.05 -2.20*** -2.71*** (0.17) (0.17) (0.17) (0.17) 0.27 0.34 Intercept 3 -0.46*** -1.08*** -0.68*** -0.62*** -2.86*** -3.11*** (0.17) (0.19) (0.18) (0.18) 0.34 0.38 Intercept 4 -1.80*** -2.51*** -2.39*** -1.60*** -3.84*** -4.66*** (0.23) (0.29) (0.28) (0.22) 0.52 0.73 N 160 159 158 157 156 155 2 0.16 0.14 0.18 0.13 0.08 0.12 Pseudo R Note: ordinal logistic regression coefficients with standard errors in parentheses. All input variables are standardized. *** p<0.01; ** p<0.05; * p<0.1 (two-tailed tests)

29

Figure 1: Resource endowment and extent of EU lobbying

>0<=1

>1<=2

>2

4 3 2 0

1

Frequ. of contacts

3 2 0

1

Frequ. of contacts

3 2 1

0

>0<=1

>1<=2

>2

0

>0<=1

>1<=2

1d.) Irish government

1e.) Irish administration

1f.) Dáil

>1<=2

>2

3 2 0

0 >0<=1

0

Number of staff in advocacy

1

Frequ. of contacts

3 2

Frequ. of contacts

3 2 1 0

0

>2

4

Number of staff in advocacy

4

Number of staff in advocacy

4

Number of staff in advocacy

1

Frequ. of contacts

0

0

Frequ. of contacts

1c.) European Parliament

4

1b.) Commission desk

4

1a.) Commission top

>0<=1

>1<=2

>2

0

Number of staff in advocacy

>0<=1

>1<=2

>2

Number of staff in advocacy

Figure 2: Type of association and extent of EU lobbying

Citizens

Professional

4 3 2 0

1

Frequ. of contacts

3 2 0

1

Frequ. of contacts

3 2 1

Business

Citizens

Professional

Business

Citizens

Professional

2d.) Irish government

2e.) Irish administration

2f.) Dáil

Citizens

Professional

Type of association

3 2 0

0 Business

1

Frequ. of contacts

3 2

Frequ. of contacts

3 2 1 0

4

Type of association

4

Type of association

4

Type of association

1

Frequ. of contacts

0 Business

Frequ. of contacts

2c.) European Parliament

4

2b.) Commission desk

4

2a.) Commission top

Business

Citizens

Professional

Type of association

30

Business

Citizens

Professional

Type of association

Figure 3: Type of association and use of inside and outside tactics

Citizens

Professional

4 3 2 0

1

Frequ. of contacts

3 2 0

1

Frequ. of contacts

3 2 1

Business

Citizens

Professional

Business

Citizens

Professional

3d.) Press releases

3e.) Internet debate

3f.) Demonstrations

Citizens

Professional

3 2 0

0 Business

1

Frequ. of contacts

3 2

Frequ. of contacts

3 2 1 0

4

Type of association

4

Type of association

4

Type of association

1

Frequ. of contacts

0 Business

Frequ. of contacts

3c.) Position papers

4

3b.) Meetings

4

3a.) Direct contacts

Business

Type of association

Citizens

Professional

Business

Type of association

Citizens

Professional

Type of association

Figure 4: Activity in highly integrated policy fields and extent of EU lobbying

Low

Medium

High

4 3 2 0

1

Frequ. of contacts

3 2 0

1

Frequ. of contacts

3 2 1

None

Low

Medium

High

None

Low

Medium

High

4d.) Irish government

4e.) Irish administration

4f.) Dáil

Low

Medium

High

Activity in integrated policy fields

3 2 0

0 None

1

Frequ. of contacts

3 2

Frequ. of contacts

3 2 1 0

4

Activity in integrated policy fields

4

Activity in integrated policy fields

4

Activity in integrated policy fields

1

Frequ. of contacts

0

None

Frequ. of contacts

4c.) European Parliament

4

4b.) Commission desk

4

4a.) Commission top

None

Low

Medium

High

Activity in integrated policy fields

31

None

Low

Medium

High

Activity in integrated policy fields