Comment on a comment

Abstract

Gordon Tullock's comment (Tullock, 1986) on my essay on optimal majorities (Wickstr6m, 1986) raises several interesting questions. I will choose to comment on a few of them. One can broadly divide my remarks into two categories: the relationship between normative and positive theory on the one hand and the technical formalizing of the problem of choosing an optimal constitution on the other hand. In his comment Tullock lists three reasons for a reinforced majority in constitutional matters. The first two of these, in my opinion, are mainly technical points which have to do with formalizing the problem, whereas the third one touches upon the relationship between normative and positive analysis. I agree, of course, with Tullock that it is a natural condition that a constitutional change require at least as large a majority as any action regulated by the constitution. This is simply a consistency requirement. But it can also easily be derived from the problem of obtaining an optimal majority. By nested voting rules, one just has to look at the smallest majority necessary in order to make and effectuate a given decision. This effective majority can then be chosen optimally. However, if the decision has to be made through several steps, such as constitutional amendments, this involves unnecessary transaction costs and the procedure will as a rule no longer be optimal. Hence, in a first-best world, the consistency requirement is a consequence of cost minimization and follows as a result of the theory. One could, however, in a world with constraints on the set of possible institutional structures, imagine that a non-consistent nested voting rule could be the best choice among the available alternatives. This, however, goes beyond the theory for optimal majorities, which in essence is a first-best normative theory. Tullock's second reason for a reinforced majority is that certain decisions require considerable discussion and therefore ought to be delayed. The relevant questions here are on the one hand why the constitutional changes should be thoroughly debated and on the other hand whether one sees the optimal majority as a normative or positive concept. If the reason for a