equation, yet estimates of the income elasticity of housing demand derived from micro data are systematically lower than estimates obtained from grouped (metropolitan-wide) data. The explanation proposed was based on the effects of common specification errors related to the housing price term. It was demonstrated theoretically that, given certain assumptions, the omission of the price term biases the ungrouped income elasticity estimate downward (whether stratified by metropolitan area or not) and the grouped estimate upward. It was also shown that the inclusion of a metropolitan-wide housing price index in the micro equation worsens the downward bias of the income elasticity estimate (and biases the corresponding price elasticity estimate toward zero). At the time the paper was written, the only empirical evidence directly relevant to this price misspecification hypothesis was contained in a study by Maisel, Burnham, and Austin [1]. It was shown that their income elasticity estimates were entirely consistent with the proposed explanation. Although their price elasticity results were not consistent, it was suggested that this could easily be attributed to sampling error. The purpose of this note is to discuss the relevance of some new empirical estimates of housing demand for the price misspecification hypothesis. These results are part of a broader study of housing demand undertaken by the author with D. T. Ellwood [3]. Since that study discusses the data and the specification of the equations in detail, the focus here is on those aspects of the data and results which are especially relevant to the price misspecification explanation.

A. Mitchell Polinsky, David T. Ellwood, An Empirical Reconciliation of Micro and Grouped Estimates of the Demand for Housing, The Review of Economics and Statistics, Vol. 61, No. 2 (May, 1979), pp. 199-205