| This dissertation specifies and tests a model of the determinants of interurban and intertemporal variations in house rents and prices. Using Capozza and Helsley's land pricing model as a starting point, I develop urban house pricing models. The theoretical models include local income and its growth rate, the after-tax interest rate, construction cost, agricultural land rent, urban amenities, and urban spatial size. Holding the fraction of distance between CBD and urban boundary constant across areas and over time, the first derivatives of house price with respect to agricultural land rent, non-land construction cost, local income and its growth rate, and urban amenities are positive; while those with respect to the discount rate, level of household utility and its growth rate are negative.; The sample covers 65 urbanized areas from 1982 to 1990. The estimation methods allow for serial correlation and cross-equation correlation. In the best regression result, all of the explanatory variables have their expected signs and the adjusted R-square for the house price estimation is 0.77. In the house price equation, local real income, climate mildness, agricultural land rent, and construction cost are positive and significant, while the local after-tax interest rate and national income growth rate are negative and significant.; Using the regression results, I measure how much of the variation in regional or urban house prices is explained by the model. While changes in the explanatory variables contribute to changes in house prices over time, much of the variation is caused by a serially correlated error. |
