Introduction

Social theorists have recently given considerable attention to the problem of designing institutions which ensure that individuals seeking their own interests maximize welfare or serve the interests of some principal, such as the government. By institutions is meant the rules that cover such things as decision-making procedures, property transfers, bargaining processes, or voting schemes. So far, however, very little attention has been given to some of the implications of problems of the second best for institutional design. Such problems concern the optimal policy when all the conditions required for a first best solution are not present. They are well known in the literature on welfare economics (Mishan 1981; Ng 1983; Bos and Seidel 1986). The main point of interest in this literature is that, contrary to our intuitions, when the conditions required for a first best do not apply in one sector of the economy, the next best solution may not be to get all other sectors as close as possible to the conditions that were required for a first best solution. It may be better to have all other sectors also deviate from the first best settings (Lipsey and Lancaster 1956), moreover, in ways that are difficult to predict (Mishan 1981). The purpose of this chapter is to explore some of the more general implications of this aspect of the second best. It is particularly concerned with the question of whether, in political institutions, it can be assumed that small changes in initial conditions or rules can be ignored.