Taxes, Spending, and Property Values: A Comment and Further Results

Abstract

In a recent issue of this Journal, Matthew Edel and Elliott Sclar (1974) present an interesting criticism and extension of Wallace Oates's wellknown article (1969). Edel and Sclar argue that the cross-section relationship between property values and local public budgets reported by Oates does not confirm the Tiebout (1956) hypothesis of the existence of an approximation to an efficient allocation of local public goods and services. Rather, Oates's results merely confirm that families have preferences regarding local service levels and act accordingly. Moreover, Edel and Sclar continue, the presence of a significant negative relationship between residential property values and the local effective property tax rate (property tax capitalization) implies that Tiebout's solution is not achieved -at least not for the period covered by the data. Nevertheless, Edel and Sclar suggest that, by observing changes in the magnitude of property tax capitalization over time in a given housing market, we may be able to discern movement toward the Tiebout result as supplies of local public services are adjusted to their long-run equilibrium values. They conclude by presenting results for the Boston area which seem to suggest movement toward Tiebout equilibrium for local public education between 1950 and 1970. This paper attempts to demonstrate that although Edel and Sclar have made an important contribution in the theoretical investigation of this issue their empirical methodology suffers from two major shortcomings. First, they have failed to account adequately for several factors resulting in intercommunity differentials in local property taxes and in residential property values; second, the structure of their empirical model, and its interpretation, is inappropriate to the hypothesis being considered. Finally, it is shown that when these deficiencies are corrected in a reexamination of the New Jersey sample utilized by Oates, the results strongly suggest an approximation to the Tiebout result in both 1960 and 1970.