This paper by Michael Malecki looks like an tremendous breakthrough in the study of judicial decision-making. It uses changes in the composition of the European Court of Justice — more particularly, in the composition of the subdivisions of the court in which the judges hear cases — to estimate judges’ ideal points, pro- or anti-integration.

In a sense, this paper is heir to a paper written by David Robertson many years on judicial decision-making in the House of Lords. Although the Lords, unlike the ECJ, does permit dissenting opinions, they’re so rare that stable ideal points can’t be extracted from them (at least, not when I’ve tried).  So, Robertson looked at changes in composition, and, via linear discriminant analysis, worked out what could be interpreted as ideal points. Putting the whole shebang in a Bayesian framework allows one to get confidence intervals, which would be great. Only problem is that one would have to stipulate the dimensionality of each decision — fairly easy in the ECJ context, where the dimensionality is pro- or anti-integration, and the Commission’s opinion is taken, axiomatically, to represent the pro-integration opinion –