The temptation to apply power indices to decision-making in the European Un
ion should be resisted for two reasons. First, power index approaches eithe
r ignore the policy preferences of relevant actors in the EU or incorporate
them in ways that generate unstable and misleading results. Second, no mat
ter how sophisticated, power indices cannot take into account the strategic
properties of the procedures that govern Europe's legislative processes, e
specially concerning changes in the institutional location of agenda-settin
g power. Proponents have responded to our criticisms of earlier power index
research with ingenious efforts to include functional substitutes for inst
itutions and preferences. The problems with power indices, however, are con
genital and cannot be adequately addressed without moving to a non-cooperat
ive game theoretic framework.