Several authors have applied the hedonic model using before-after data
to estimate the willingness to play for a change in public expenditur
e or amenity. This article shows that attempts to interpret the result
s of such estimation as a measure of willingness to pay for a local am
enity or public expenditure in this case are flawed. The implicit basi
s for such estimation is a single-jurisdiction partial equilibrium mod
el, whereas the underlying Tiebout model is multijurisdictional and ge
neral equilibrium. It is argued that the results of such estimations i
ncorporate biases, the direction and magnitude of which cannot be pred
icted. If service-tax packages are exogenous, then the hedonic model u
sing before-after data will likely provide incorrect estimates of the
willingness to pay.