Cost-benefit analysis is routinely used by government agencies in order to
evaluate projects, but it remains controversial among academics. This paper
argues that cost-benefit analysis is best understood as a welfarist decisi
on procedure and that use of cost-benefit analysis is more likely to maximi
ze overall well-being than is use of alternative decision procedures. The p
aper focuses on the problem of distorted preferences. A person's preference
s are distorted when his or her satisfaction does not enhance that person's
well-being. Preferences typically thought to be distorted in this sense in
clude disinterested preferences, uninformed preferences, adaptive preferenc
es, and objectively bad preferences; further, preferences may be a poor gui
de to maximizing aggregate well-being when wealth is unequally distributed.
The paper describes conditions under which agencies should correct for dis
torted preferences, for example, by constructing informed or nonadaptive pr
eferences, discounting objectively bad preferences, and treating people dif
ferentially on the basis of wealth.