Drafting constitutions: A comparative institutional analysis of constitutional conventions in the European Union and Germany

This dissertation examines why politicians participate in constitutional conventions. One the one hand, a constitution is a higher order legal document, and the drafting process ought to involve as much constitutional expertise as possible. On the other hand, constitution makers have institutional preferences that are induced from their policy preferences. How can we determine whether constitutional experts, ideologues, or expert ideologues hold the pen? I address this puzzle by investigating the institutional organization of two constitutional conventions: the European Union Convention (2002-2003) and the West German Parliamentary Council (1948-1949). I argue that delegate ideology should predict membership in important agenda control committees, even when these committees also include the most experienced delegates.;I employ a novel approach to measure preferences in constitutional conventions. Rather than defining the issue space ex ante, I use revealed delegate behavior to infer latent positions by applying scaling techniques to available data. For the European Convention, I use proposal co-sponsorships, recorded in several thousand documents, and for the German Parliamentary Council, I rely on plenary speeches delivered by the delegates in order to locate actors on a dimension using a new technique to estimate preferences from political texts.;The empirical analysis first tests conjectures from legislative theories, logrolling and informational theories. The committee composition results show that there is no strong evidence for these theories. The second analysis predicts membership in drafting committees on a delegate level, jointly testing the expertise and ideology hypotheses. While the most important drafting committees have the highest average expertise, this effect disappears when both the ideology and expertise hypotheses are tested jointly. The German convention is an example for a partisan control assembly. Party groups are the key organizing units and appoint extremist delegates. This results in a heterogeneous committee in which political disputes are strong and prolonged. In contrast, the European convention is an example for non-partisan leadership control. Parties are less influential, and an exogenous agenda control committee holds strong formal powers. Its central location and homogeneous composition, combined with its agenda setting powers, are a key element in assuring a successful outcome.