“Estimating Property Tax Capitalization: A Critical Comment”, 1977-04-01 (; similar):
In recent years, a number of studies have appeared, investigating the capitalization of the local property tax have appeared…Since much of the variation observed in property taxes occurs because of assessment errors which may not persist indefinitely, the basis for tax capitalization may be different from the presently observed tax. Since the accuracy and frequency of assessment varies substantially from place to place, capitalization of observed tax payments to the same extent in all jurisdictions is unlikely.
…Because Wallace Oates kindly made his data available to me, it is possible here to compare the results of his well-known study of capitalization and the Tiebout model (196955ya, 1973) with and without the misspecification. It appears that in his case, the misspecification of the tax effect has created an upward bias of about 40% in the estimated extent of tax capitalization
Many studies of property values have used a clearly incorrect specification of the tax effect, one likely to introduce a bias of uncertain direction and magnitude. In the one case I have been able to examine closely, it appears that the original estimate of capitalization is about 40% too great. What bias exists in other studies is quite uncertain, and I do not attempt to imply any general pattern relevant to other studies from the present empirical results. The principal point here is that the demonstrated sensitivity of the results in equations (I) and (II) to the specification choice makes it doubtful that studies using data similar to these and mis-specifying the tax effect in the same way could yield reliable evidence about the extent of tax capitalization. Consequently, I believe that our knowledge of the extent of tax capitalization is very much less than is commonly supposed.