Referendums in the EU’s constitution building process

Abstract

In ten member states of the European Union (EU) the new constitutional treaty was supposed to be ratified by referendum. A growing number of theoretical models predicts that such additional ratification hurdles result in an advantage for negotiators in the bargaining game. The impact such a referendum constraint can exert, however, depends on the timing of its announcement, the remaining ratification rules as well as the preference constellations. If parliament and voters are actually in favor of the new treaty, ratification constraints may cease to affect the bargaining outcome. After presenting the theoretical foundation of these arguments, we present empirical evidence much in line with the theoretical implications. More specifically, especially for the issues changed during the intergovernmental conference, those governments gained that had scheduled a referendum and voters had a stronger preference for the status quo

Reforming the equilibrium? Veto players and policy change in the European constitution-building process

“…15 Overall our findings support the Schelling-conjecture (Schelling 1960): Cohesive member states on the status quo achieved higher gains than those member states on the status quo whose voters and/or parliaments held a diverging position. 15 Using a different operationalization Hug and Schulz (2007) conclude that the correlation between reform sceptic voters and governmental gains is slightly stronger among those member states holding referenda. However, the operationalization suggested by Hug and Schulz (2007) assumes that negotiators are able to identify the 'voter space' including all 25 median voters separately from the 'government space' including all 25 governmental positions.…”

“…15 Using a different operationalization Hug and Schulz (2007) conclude that the correlation between reform sceptic voters and governmental gains is slightly stronger among those member states holding referenda. However, the operationalization suggested by Hug and Schulz (2007) assumes that negotiators are able to identify the 'voter space' including all 25 median voters separately from the 'government space' including all 25 governmental positions. Instead our operationalization suggests that negotiators are able to identify the major conflict dimensions on the basis of voters' and governments' positions.…”

“…In order to identify the distance between the governmental veto players and their voters, we used a procedure to calculate the median voter's position on 40 questions related to the EU constitution asked in the Eurobarometer (EB) 60.1 and Candidate Country EB (CCEB) 2003.4 (Hug and Schulz 2007). For each of the 40 questions it was possible to code the provisions under the Treaty of Nice (status quo), the conventional draft and the constitutional treaty (outcome).…”

“…A further methodological problem is to bridge the two datasets, and we used the status quo as the anchor point. Unlike Hug and Schulz (2007) 6 we apply a joint estimation of the reduced two-dimensional issue space via Bayesian item response analysis which allows us to locate each issue's reform option, the median voter's position and the governmental position into a common issue space. 7 This allows us to calculate issue-specific gains for each type of actor and distances of each actor to every reform option (Euclidian distances).…”

“…Party positions in these countries were imputed with the mean across all European parties belonging to the same party family. 6 Difference between Hug and Schulz's (2007) separate and our joint approach concern the relative position of SQ, Draft and Outcome as well as the position of voters to those of member states. Comparing the results the correlation within both subgroups of actors (voter and member states) is very high (about 0.89).…”