Advocates of regulatory benefit-cost analysis are right in theory, alb
eit somewhat impractical. In a sense, they have put the cart before th
e horse. Right now we cannot satisfactorily estimate the costs of gove
rnment regulation, let alone its benefits. We must understand what dri
ves cost before we can make satisfactory cost estimates. But, first of
all, we need good cost measures to figure out what drives cost. We do
n't now measure the costs of regulation Accurate measurement of the co
st of regulation is, therefore, the first step to better regulatory pr
actice. This means accounting for the regulatory burdens on firms and
subordinate governments and allocating those burdens to the regulatory
policies that induced them.