IN contrast to the extensive empirical work on the income elasticity on demand for housing services, the relation between tenure choice and income has received relatively little attention. Those studies which have been completed have generally involved only a single housing market, and the method, data used, and the markets studied have been so diverse as to make comparison of the resultant estimates extremely difficult. Further, one of the major shortcomings of earlier efforts has been the lack of a proper theoretical framework underlying the estimated models.' The purpose of the present paper is to develop and test a model of aggregate tenure choice based on the household tenure decision. The tenure-choice estimates are based on a sample of metropolitan areas and are obtained for six household types by race using income data for eight income intervals. The principal value of the research is the rigorous definition of the aggregate tenure-income relation and the provision of estimates that can serve as benchmarks with which to compare similar estimates for individual housing markets.' The estimates might also serve as an input into the evaluation of public policies involving home ownership, such as the effect on home ownership rates of changes in the treatment of certain ownership related expenses under the Internal Revenue Code. The main findings are the estimates of the income elasticity of demand for owner occupancy. The elasticity at the mean for all households combined is about 0.18. Wide variation among household types was found: elasticities larger than the mean were obtained for husband-wife family types with heads under age 45 and for non husband-wife families and primary individuals. There are substantial differences between races in the elasticities, although the overall pattern for household types just described remains generally valid for both races. The mean income elasticity of all black households combined is somewhat larger than for all white households, 0.25 vs. 0.18. The remainder of the paper consists of two sections. In section II the theoretical model is developed and its empirical specification given. Section III presents the estimates of the aggregate cross-cities model and contrasts them with fairly comparable estimates for a single housing market.