Counting organized interests in the European Union: a comparison of data sources

ABSTRACT

In contrast to large-n research on US interest organizations using lobby registration data, European Union (EU) scholars have used a variety of data sources, including registries maintained by the European Commission, directories of interest organizations active in Brussels, and data on access to the European Parliament (EP). This diversity of sources raises important questions about their comparability. To what extent do these different EU sources actually measure the same population of interest organizations? We find that the several data sources vary markedly over EU institution and publisher. The paper discusses the methodological and substantive causes of these differences and their implications for further development of large-n research on interest organization politics in the EU.1

Keywords:

• Business groups
• European Commission
• European Parliament
• European Union
• interest groups
• lobbying

Notes

1. An earlier version of this paper was presented at the Annual Meeting of the Southern Political Science Association in New Orleans, 8 January 2006. We wish to thank Marianne Ananyeva, Veerle van Doeveren, Willem Masman and Anne Messer for their assistance in this research.

2. Official list (1992, 1996) or CONECCS (2003–): Balme and Chabanet 2002: 45–9; Broscheid and Coen 2007; Eising 2004: 219, 237; Eising 2005; Eising and Kohler-Koch 2005: 18; Greenwood et al. 1992: 1–2; Kohler-Koch 1994: 170; Mahoney 2004; Mazey and Richardson 1993: 6; Wessels 2004. EP register (2003–): Bernhagen and Mitchell 2006.

• Official list in combination with Landmarks (1990–) or Euroconfidentiel: Aspinwall and Greenwood 1998: 2–3; Beyers 2002, 2004; Greenwood 1997, 2003b. Landmarks (1990–): Lahusen 2002: 699; Lahusen 2003; van Schendelen 2002: 46. Eising combines his sample with national sources: CBD Research Ltd 2005; Le Conseil National du Patronat Français 1997; Vogel et al. 1953–.

• DG Trade: Beyers and Kerremans 2007b. Eising combines his sample with national sources: CBD Research Ltd 2005; Le Conseil National du Patronat Français 1997; Vogel et al. 1953–.

3. Andersen and Eliassen (1991) cited, along with Commission numbers, an earlier publisher survey (Morris et al. 1986). They cited numbers of lobbyists (multiple lobbyists may be employed by one organization) and (national and European) organizations ‘deemed to be of relevance’ in this 1986 publication: 659 European associations and 6,000 national associations and persons. References to these figures can also be found in Kohler-Koch (1994: 174). Others relied on newspaper estimations (Grant 1995: 98), figures that are again cited to confirm similar estimates by the Commission (Greenwood 1997: 3). By then, there were an estimated 3,000 interest groups, 500 European or international federations and 10,000 individuals employed by these or other organizations.

4. There are some other, less important sources on European interests, and similar directories in EU member states and internationally. Some of these include the following. EU: Euroactiv 2006 (used by Greenwood) and comparable to member lists of PA organizations such as the Society of European Affairs Professionals (SEAP); Europa Publications 1995; Morris et al. 1986; Routledge 2005; Seingry 2003. International: Centre for Civil Society (LSE) 2004; Saur 1995; Union of International Associations 2006; Zils 2004. US: (Thomson/Gale, 1st edn: 1956 described by Baumgartner (2005). UK: CBD Research Ltd 2005 used by, among others, Bennett (1999).

5. Editions of all published directories lag behind by one year, so that a 1996 edition reports 1995 data.

6. There is a project by the Danish International Centre for Analytical Reporting using several of these sources to compile an online database of EU groups (Bækgaard and Svith 2004).

7. Number of organizations in 15 March 2006.

8. Slob and Smakman 2006: 31.

9. 3,683 lobbyists/cardholders.

10. Dating from spring 2007, the website contained the following statement: ‘On 21 March 2007, the Commission adopted the Communication on the follow-up to the Green Paper “European Transparency Initiative” (COM(2007)127). As a result of the Communication, the Commission will create and launch in spring 2008 a voluntary register for interest representatives. This register will replace the current database, CONECCS.’

11. Since early 2006 the online registration form of this list contains a box: ‘I agree that the information I entered will be shared with the Central EU Civil Society Database.’

12. The sampling procedures were similar in all cases. For the online directories, we copied the complete list to a database and took a random sample. For the Philip and Euroconfidentiel directories, we sampled the index of names, taking every x number to achieve a sample of 50. For the Landmarks directory, we numbered all listed organizations and used a random number generator to compile a sample of 50 registrants. For these publications, we selected certain sections of the directories. These are specified in Table 1 under ‘Reviewed section(s)'.

13. All variables: ID number, coder name, name of organization (different languages), date, abbreviation, website address, number of sample, does the organization still exist? (yes, no, merged, other), HQ country, type of organization (membership group, association of membership groups, association of associations, association of institutions, institution, mixed, lobby firm, other), size (if membership group or association number of members, if institution number of employees), substantive interest (categories similar to Gray and Lowery (1993: 90)), funding sources (public, private, EU, other), establishment date, number of staff, main locus of business (one, several, or most of EU, Africa, North America, South America, Middle East, South Asia, East Asia, Eurasia, world, EFTA, NATO, near abroad, other), type of organization (employers' peak association, employers' sectoral association, labour union, small and medium enterprise associations, agriculture producers, agriculture processors, consumer NGOs, environmental NGOs, developmental NGOs, other public interests), contact information, comments.

14. Twice: Landmarks 2005; Euroconfidentiel 2002 and Philip 1996.

15. In coding the websites, the four coders working on the project regularly discussed difficult cases among themselves and with the project directors. And a subset of cases were coded by multiple coders to identify ambiguities in the coding scheme. The ‘response rate’ – that is, whether we could find a website of an organization in a directory and adequately code the most important substantive variables on the type of organization and its substantive interest – varied between 93 per cent for CONECCS and 86 per cent for the DG trade list for the official directories. The rates for the published directories varied between 88 per cent for Landmarks (2005) and 78 per cent for the Philip directory (1996). This relatively high success rate in coding organizational websites was essential both because the information provided by the directories alone, especially from the parliamentary list, was insufficient to code many of the variables on the nature of the organizations and because the more detailed information from the websites was needed to avoid ambiguities (names, abbreviations) in how organizations were identified across the official lists and published directories.

16. In estimating the overlap between each source based on our samples, we also took some extra samples to reduce the error term (i.e. the difference in the two estimates of the overlap). Information on these additional samples can be obtained from the authors.

17. The bars in Figure 1 have been calculated using these numbers. For example, the CONECCS–EP register Venn diagram: we found 43 per cent of our CONECCS organizations in the EP register, so our first estimation of the overlap is 735 × 43% = 316; our second estimation is 2119 × 10% = 212; this leaves us with an estimated average of 264 organizations in both lists.

18. (European) associations of (national) membership groups are composed of national organizations with individuals (not professionals) as members (e.g. labour unions, consumer groups, tourist groups). (European) associations of (national) associations represent national trade and professional associations (lawyers, farmers, tourist agencies, employers). (European) associations of (public/private) institutions directly represent organizations with companies, agencies or governments (banks, municipalities, public transport, energy companies, consultants). Institutions are public/private organizations directly representing their own interests (non-EU governments, companies, international organizations). Lobby firms are organizations (e.g. law, consulting companies) representing any other organization. Mixed/other types include organizations without a dominant type mode of membership (certain development foundations, some sector associations with both direct membership and associational membership).