Use of Distributional Weights in Cost-Benefit Analysis: a Survey of Schools

Abstract

It is now almost accepted practice that distributional weights be incorporated into cost-benefit criteria. While there are still major dissenters on this issue, notably Harberger, the relevant questions now involve the nature of the weights themselves. A number of alternative formulations have been recommended, ranging from "a priori" to "revealed preference" schools. Recognition of these options allows one to select a framework of assumptions in which one may feel comfortable, without being encouraged to return to the traditional position of excluding distributional considerations. Thus, Harberger's influential pur ported defense of the traditional school, which includes some legitimate reservations, should be interpreted as a criticism of certain schools, not the use of weighting per se. By going back to first principles of what distributional weights are trying to achieve, this survey attempts to dispel various widespread misconceptions concerning distributional weights. It argues that redistribution in kind is the major theoretical justification. This provides a perspective from which to judge the type of distributional weighting that is most consistent with some recent advances in the welfare economics literature.