This dissertation explores the impact of tax policy and institutions on decisions in the market for housing. The first essay is joint work with Andrew Hanson. In it, we estimate the sensitivity of mortgage interest deducted on federal tax returns to the availability of the Mortgage Interest Deduction (MID). Our primary results show that for every one percentage point increase in the tax rate that applies to deductibility, the amount of mortgage interest deducted increases by $303–590. The second essay simulates changes to average home prices in twenty-seven cities that would result were the MID reformed. I use local variation in housing parameters to simulate home price changes for three different reforms: eliminating the MID, converting the MID to a fifteen percent credit, and capping the MID at fifteen percent. City price changes vary in response to a single policy by as much as 12.8 percentage points. Spatial variation within cities is also notable, with areas high in income experiencing steeper price declines and areas of lower income experiencing shallow declines. The third essay is joint work with Andrew Hanson, Zack Hawley, and Bo Liu. We design and implement an experimental test for differential response by Mortgage Loan Originators (MLOs) to requests for information about loans. Our e-mail correspondence experiment is designed to analyze differential treatment by client race and credit score. Our results show net discrimination of 1.8 percent by MLOs through non-response.

This paper examines one of the most controversial items of the new GOP tax bill, the Mortgage Interest Deduction. The paper seeks to identify which groups would feel the greatest financial burden if the deduction is limited in the future tax code. The author identifies potential declines in mortgage interest rates and expensive home values as two key motivations behind the lobbying efforts for this deduction to remain untouched. Using data on mortgages originating in 2016, the author estimates a decline in mortgage interest rates between .039 and .043 percent for every $1,000 borrowed above the 2016 MID limits for taxpayers. The paper then goes on to discuss interest rate volatility implications for Mortgage Servicing Assets. The paper ends with a discussion on the downward pressure the new tax reform may have on expensive home values.

The debate over federal income tax treatment of home mortgage interest (HMI) has largely overlooked an important, and possibly unintended political and economic consequence of our federal income tax system. The distribution of the for home mortgage interest deduction tax benefit across states is a possible missing consideration. Specifically, this study offers a federalist1 perspective on the federal income tax benefit from the deduction for HMI - one of the largest personal federal tax expenditures on the books. This dissertation analyzes current national political rhetoric from a federalist perspective. Discussion also includes background, current status, and proposed changes to the tax code for of the HMI deduction. First, a Tobit regression is used to analyze the distribution of the HMI tax benefit across states and to test for disproportionate distribution across states in benefit derived from the federal income tax deduction for home mortgage interest beyond that which is explained by income. This initial part of the study is also the precursor to a hierarchical analysis seeking to identify significant factors affecting the distribution of the benefit of the HMI deduction across states. The Ernst and Young/University of Michigan Individual Model File of 1992 tax returns is the primary data source for this initial part of the investigation. The second part of the analysis examines the effect of sets of factors in a causal hierarchy on the HMI deduction benefit. By first controlling for the effects of personal and identifiable state characteristics on HMI deduction benefit, the possible existence of a residual socio-political force is tested. The primary data sources for this part of the study are the 1990 Census of Population and Housing 5% Public Use Microsample as well as tax data extracted from the Statistics of Income, Individual Public Use Tax File, Level III Sample, as well as others. Ridge regression is used for hypothesis testing. Results indicate the existence of a significant difference in the benefit from home mortgage interest deduction across states holding income constant. This study also finds that a set of personal as well as a set of state market, legal and tax characteristics significantly influence the taxpayer's HMI deduction benefit, and that a residual difference in benefit across states after controlling for personal and identified state attributes. Future study should examine the source of residual across state differences (attributed to socio-political differences between states). Federal housing goals may be frustrated as the effective subsidy actually helps support higher home prices in areas where high housing costs may already be a barrier to potential new homeownership. The concepts and techniques applied in this study could easily be applied to other provisions of federal tax, or to any other tax system in a federation for that matter.