Geography, Housing Prices, and Interregional Migration

This dissertation consists of three papers that explore the intersection between geography and housing markets. The research examines both how geographic context shapes housing prices and how house prices influence geography through household location decisions.The first paper explores the spatial structure of housing prices within Tucson, Arizona. Hedonic house-price studies typically assume that housing attribute prices are constant over space. The research tests this assumption and compares two methods of incorporating spatial-varying parameters into house-price models: geographically weighted regression and the spatial expansion method. The results provide evidence that housing attribute prices do indeed vary with geographic context and suggest that more reliable parameter estimates and better house-price estimation accuracy can be achieved through the use of these techniques.The second paper builds upon the first by examining how more realistic conceptions of housing market spatial structure influence the hedonic price estimates of location-specific externalities. The empirical analysis examines how two key spatial effects, spatial dependence and spatial heterogeneity, impact the marginal price estimates for proximity to the Rillito River, within Tucson, Arizona. Both spatial effects are found to influence the resulting estimates, but spatial heterogeneity is of greater practical importance as the price estimates vary widely with geographic context. This research highlights the importance of considering both spatial effects in hedonic externality valuations.The final paper explores how housing prices influence interregional migration patterns, and more specifically, how their influence varies with both stage in the life course and educational attainment. The research models metropolitan migration within the United States during the period 1995 to 2000. The results indicate that housing prices play an important role in driving regional demographic change, as their influence varies with both demographic characteristics. High housing prices deter individuals in their late twenties and early thirties, but their influence wanes during middle age. House prices become more important as individuals near retirement. The results also provide evidence that college graduates respond more to house price differentials than do persons with lower levels of educational attainment