Homeowner tax preferences, housing demand, and vertical equity: Theoretical, empirical, and policy analyses

Owner-occupied housing benefits from several forms of tax preference in the United States. The non-taxation of implicit rental value, deductibility of mortgage interest and property taxes, and capital gains preferences are all specific forms of housing tax preferences. The purpose of this dissertation is threefold: (1) to investigate the relationship theoretically as well as empirically between homeowner tax preferences, in particular, nominal mortgage interest and property tax deductions, and housing demand, (2) to examine empirically the vertical equity consequence of removing housing tax deductions, (3) to develop an interval estimate of the change in Suits Indices between the current federal income tax and an alternative income tax that excludes housing tax deductions to test the significance of the estimated vertical equity effect of eliminating housing tax deductions. In this dissertation, I develop a theoretical model for housing demand to analyze the impact of housing tax deductions on the demand for housing. A comparative static result is derived and it indicates that the housing tax deductions have an ambiguous effect on housing demand. A deterministic simulation of the effect of housing tax deductions on housing demand is performed using time series data for the period 1974-1993. Simulation results reveal that housing tax deductions have positive influences on housing demand. I also simulate the effect of housing tax deductions on net per capita single family housing expenditure utilizing a linear expenditure system for housing demand. Simulation results reveal a direct relationship between housing tax deductions and housing demand. I am able to assess both the non-revenue and revenue neutral vertical equity effects of removing housing tax deductions. Results in both cases show that eliminating housing tax deductions increases the vertical equity of the tax system. I then test the significance of the estimated vertical equity effect using both the bootstrap-t and percentile methods. The results reveal that the estimated vertical equity effect is statistically significant