First Things First? The Agenda Formation Problem for Multi-issue Committees

It is often argued that multi-issue committees should discuss issues simultaneously to avoid ineciency. However, in practice, parties can be constrained to discuss issues sequentially and in this case, existing game-theoretical models give inconclusive results: either parties have dierent preferences over agendas or they are indierent. We show that when there is an important issue, parties have the same preferences over agendas, in particular they prefer to discuss the most important issue Þrst. Moreover, when an issue is dicult/urgent (in the sense that the rejection of a proposal on this issue implies a game breakdown with a positive probability) parties prefer to postpone the negotiations over the dicult/urgent issue. We highlight several incentives that players need to take into account in forming their preferences over agendas. Since theseareoften in conßict, the existence of a Pareto optimal agenda is of particular interest.