Implementing Cost-Benefit Analysis When Preferences are Distorted

Abstract

Cost-benefit analysis is routinely used by government agencies in order to evaluate projects, but it remains controversial among academics. The standard defense appeals to the Pareto standard or the Kaldor-Hicks standard, and assumes that agencies should respect people's actual preferences, as opposed to informed or otherwise restricted preferences. This paper argues that cost-benefit analysis is best understood as a welfarist decision procedure, and its most plausible defense is that use of cost-benefit analysis is more likely to maximize overall well-being than is use of alternative decision-procedures. The paper focuses on the problem of using cost-benefit analysis when preferences are "distorted." A person's preferences are distorted when their satisfaction does not enhance that person's well-being. Preferences typically thought to be distorted in this sense include disinterested preferences, uninformed preferences, adaptive preferences, and objectively bad preferences; further, preferences may be a poor guide to maximizing aggregate well-being when wealth is unequally distributed. We argue that government agencies currently recognize these problems but respond to them in an ad hoc way, and that a more systematic treatment of these problems is warranted. The paper describes conditions under which agencies should (or should not) correct for distorted preferences, for example, by constructing informed or non-adaptive preferences, discounting objectively bad preferences, and treating people differentially on the basis of wealth. Institutional and political constraints - the inability of agencies to make lump sum transfers, the need for transparency - are also considered.