The Institutionalization of Bias: Tracing Interest Groups Populations from the National to the Global Level

Interest groups react to policy-making activities; new laws, government spending, international negotiations, all trigger the mobilization of new groups or attract the policy interest of existing interest groups. The contours of populations of interest groups consequently partially mirrors the policymaking ‘energy’ of governments (Leech et al, 2005; Gray and Lowery, 1996). However, the strength of this effect, i.e. how many and what type of groups are attracted, depends on the institutional strength of government in a given domain such as bureaucratic structures, party political conflict lines and legal policy competences. In an institutional vacuum such as on non-decision issues (Bachrach and Baratz, 1961/2) and in nascent or weak policy environments such as in some transnational arenas, interest groups have more room to set the agenda themselves and act as creators of institutions (Kohler-Koch, 1994; Wessels, 2004, 199-203; Kirchner, 1978, 4; Hanegraaff 2015). This all means that the shape of interest group populations at different levels and domains of policy making should substantially differ. One classic hypothesis derived from this logic is that at weaker institutional venues we should expect specific interests to dominate the interest group population more than at venues with strong institutional structures. This leads to a paradox in which the strengthening of political institutional structures leads to less democratically legitimate input of interest groups. While theoretically appealing, there has not been much empirical work to substantiate this claim because it requires a comparison of population across a variety of venues (but see Berkhout et al. 2017). In this paper we combine datasets constructed in various research projects to overcome this obstacle. This allows us to compare domains of interest groups at different governance levels and examine whether the ‘strength of government’ critically shapes the numbers and types of interest groups present, and so affects the representative nature of groups active at these venues. We use data from several sources and compare domains in the WTO, the UN, the EU, Germany, France, the Netherlands and the United Kingdom to see how biases in populations travel across political levels.