European Integration, Perverted Democracies and Rubber-Stamp Parliaments? An Empirical Analysis of Two Approaches

Abstract

Does European integration undermine national democracies by strengthening the executive branch to the detriment of the legislative branch? Prior research offers two competing theoretical approaches to answer this question: the abdication and delegation approaches. While the former predicts an increase in executives’ relative power, the latter envisions no significant change in the comparative powers of the executive and legislative branches in European Union member states. Despite the ample body of studies that purport to support either approach, the empirical record remains strikingly inconclusive about the relative merits of these approaches due to the problem of observational equivalence. By utilizing empirical evidence on cross-national and within-national variations in executives and legislators’ positioning on European integration, this study makes an effort to discriminate between the two approaches. The results lend a qualified support to the abdication approach.

Keywords:

• executive
• legislative
• delegation
• abdication
• national democracy and European integration

Notes

¹ Prior research offers a number of examples where analysts address broad institutional issues by invoking the ramifications of these issues for elite policy preference and behaviour, like Putnam (1973) in his study of parliamentarians in Britain and Italy, Searing (1994) in his study of legislative elites in Britain, Page and Wright (1999) in their analysis of bureaucratic elites and Hooghe (2005) in her analysis of European Commissioners.

² For the details of this survey, see http://ec.europa.eu/public_opinion/archives/top/top_en.htm. For a recent exemplary study that uses the data from this survey, see Gabel & Scheve (2007).

³ Data for these variables are extracted from the Survey. Education is measured with a question that asked: ‘How old were you when you stopped full-time education?’ Higher values indicate higher levels of education. Age is measured by a question that asks: ‘What is your year of birth?’ Gender is a dummy variable: ‘1’ if male, ‘0’ otherwise.

⁴ I relied on data from Eurobarometer Surveys 45 and 46.

⁵ The percentages are calculated from the data presented in Muller & Strom (2003).

⁶ The data for this variable are drawn from Lijphart (1994).

⁷ These calculations are based on the ratios of each variance component relative to the total variance in Support.

⁸ See Steenbergen and Jones (2002, pp. 225–226) for a general discussion of the difference between the two techniques and Kreft & De Leeuw (1998) for a more detailed elaboration. Additionally, I ran the analyses using the bootstrapping procedure to examine the robustness of the RMLE estimates. The results are substantially similar.