Portal de Periódicos da CAPES

An Empirical Reconciliation of Micro and Grouped Estimates of the Demand for Housing

1979; The MIT Press; Volume: 61; Issue: 2 Linguagem: Inglês

10.2307/1924587

1530-9142

Maria Polinsky, David T. Ellwood,

Fiscal Policy and Economic Growth

T is not uncommon for estimates of the income elasticity of housing demand based on metropolitan-wide averages to be 60% to 90% higher than estimates based on individual household data. Price elasticity estimates also vary greatly among micro and grouped studies, although no pattern has emerged. Such differences cannot be explained by the unit of observation since the grouping of observations in a correctly specified equation does not by itself affect the expectations of the estimates. Therefore, it seems likely that the housing demand equations have been misspecified, possibly in a number of different ways. The purpose of this paper is to show empirically that frequently made specification errors can account for the very different estimates obtained from micro and grouped equations, and that the correctly specified equations are reconcilable.1 Previous studies have not been able to examine these issues properly because of the absence in the micro demand equation of a housing price term that varies with each observation.2 A common practice has been to use micro housing expenditure and income data from one source in conjunction with a metropolitan-wide index of housing price from a completely different source. Aside from disparities in the data bases, this practice does not take into account the fact that the unit price of housing (for a standardized bundle of land and structure) varies substantially by location within a city. Such variation occurs because of differences among sites, for example, with respect to workplace accessibility. Thus, individuals with similar incomes residing at different locations would be expected to purchase different quantities of housing because they face different unit prices of housing. The present study is able to overcome this problem by developing a new technique for estimating the appropriate unit price of housing for each micro observation and then applying it to a unified data base. Section II describes the correctly specified and the misspecified micro and grouped demand equations to be estimated. In section III the technique for calculating micro housing prices is derived. Section IV discusses the data. Section V presents and analyzes the empirical results. Finally, in section VI the results are summarized and qualified.