Applying ideal point estimation methods to the Council of Ministers

Abstract

This research note addresses an increasingly popular topic in the EU literature, namely the measurement of policy preferences in the Council of Ministers. It aims to provide conclusions on three issues: (1) what data are in fact available from the Council, (2) how preferences are measured in other legislatures, and (3) whether these methods would be suitable for analyses of Council members' preference positions given the available data. Applying the popular scaling method NOMINATE and a Bayesian MCMC model to a data set consisting of all legislation adopted by the Council in 1999-2004, it is found that, although the two methods show similar voting patterns at the general level, the failure to report standard errors by NOMINATE in particular proves to be a severe problem when trying to identify individual governments' policy location. Conversely, the Bayesian approach provides a convincing method for analyses of Council decision records and is easily extended to include more advanced empirical information than merely the governments' decisions to support or oppose a proposal.