House Prices and the Labor Force Composition of Cities

This dissertation tests a theoretical framework relating house prices and the population composition of cities. If the income elasticity of demand for a primary residence is less than unity, the ratio of skilled to unskilled worker wages falls as housing cost rises. This changing wage ratio induces substitution in production across all industries causing the skill intensity of the labor force to rise. Secondly, any household type for which share of income spent on housing is smaller than average will tend to live in cities with higher than average housing prices. These are termed the income elasticity of demand and housing preference hypotheses, respectively, regarding the relation between house prices and the population composition of cities.;The first two empirical analyses in the dissertation test the income elasticity hypothesis by demonstrating that the skilled wage ratio varies inversely, and the skill intensity ratio directly, with different measures of housing cost. Furthermore, the estimated elasticities of the wage and intensity ratios with respect to housing cost agree closely with values derived from calibration of the theoretical model.;A third empirical analysis examines why the spatial distribution of Hispanic households by skill-level varies from that of non-Hispanic whites. The difference in spatial distribution provides a test of the effect of house prices on population composition applying both the income elasticity and housing preference hypotheses. Demand for housing by unskilled Hispanic households is substantially less than that of similar non-Hispanic white households. This difference in demand suggests an inverse relation between skill intensity and house prices for Hispanic workers based on the housing preference effect. Theory shows that this effect could even reverse the positive relation between skill intensity and house prices for Hispanics predicted by the income elasticity effect. Empirical tests confirm this prediction and show that house prices have different effects on Hispanic and non-Hispanic white households across U.S. cities.;This dissertation calls into question models of urban population determination that assume either homothetic or homogeneous preferences. It shows that heterogeneity among cities in relative wage and skill intensity may not represent real differences in relative productivity. Instead, apparent productivity differences can result from skilled and unskilled workers requiring different compensating differentials in wages in response to spatial variation in relative prices of housing and amenities.