Housing, and Racial Equity

Abstract

In this dissertation I assess the impact of various government policies on employment, wages, health, and housing. Studying the impact of government policy is important not only to conserve resources but also to ensure an equitable distribution of such resources. One major emphasis in this dissertation to accurately measure the impact of government policy. I attempt to achieve this in two ways. First, I used spatial econometric methods to reduce bias in standard econometric models. Second, I use better data by incorporating a larger set of control variables and covering longer periods of time. A standard assumption in mainstream cross-sectional and panel data models is the indepen dence of the sample’s observations across units of analysis. This assumption ensures that the resulting error terms are spatially independent. However, this assumption is often violated in the assessment of government policy where data is often collected at the census tract, zip code, county, or state level, and such spatial units are usually interdependent. In particular, neighboring spatial units are often similar in characteristics, which is a violation of the independence assumption, resulting in estimates that are biased and inconsistent (Anselin and Bera 1998). In addition to accounting for spatial autocorrelation, I also take advantage of geographic discontinuity in govern ment policy to reduce confounding in standard regression models. Moreover, I incorporate many relevant control variable in my models to reduce omitted variable bias. This dissertation is comprised of three essays. The first tackles a government policy which had reaching socioeconomic and demographic consequences, namely the role of Home Owners’ Loan Corporation (HOLC) in shaping the US housing market. HOLC’s residential redlining policy in the 1930s denied racial and religious minority groups access to mortgage financing and home improvement loans as well as access to credit for starting new businesses. This limited education 1 and employment opportunities for African Americans for generations to come. In this regard, I ask, Did historically redlined Zip Codes experience worse COVID-19 outcomes? To answer this question, in Chapter 2, I assess the impact of residential redlining on COVID-19 outcomes. Using georeferenced COVID-19 data and historical redlining maps of Baltimore, I find that COVID-19 cases are higher in redlined neighborhoods while vaccination rates are lower. I measure redlining using historical neighborhood grades produced by Home Owners’ Loan Corpo ration (HOLC) in the 1930s, which resulted in redlining. I also use the Home Mortgage Disclosure Act (HMDA) data to estimate a redlining index which predicts the difference in odds of loan ac ceptance by race after controlling for income, debt, and loan size. Using spatial autoregressive models, I find that historically redlined neighborhoods have significantly higher COVID-19 cases and significantly lower COVID-19 vaccination rates, even after controlling for a host of social vulnerability and residential mobility measures. My second essay investigates the effect of state tax polices on job creation, job loss, and wages. In this context, I used discontinuity in tax regimes to compare business activity and taxes in coun ties that share a state border. My research question is: do cross-border differentials in state taxes affect firms’ job creation, job destruction, and average monthly wages in state-border counties? And if so, does the effect vary by race and ethnicity? To answer this question, in Chapter 3, I asses the effect of cross-border state tax differentials in counties straddling various state borders on job creation, job destruction, and wages. I find that higher top marginal state corporate income tax rates (compared to the state’s neighbors) are associated with lower job creation, higher job destruction, and lower average monthly earnings for small- and medium-sized firms operating in counties along the state’s border. I also find that higher top marginal state personal income tax rates reduce job creation, increase job loss, and have no effect of average monthly wages. In addition, higher sales taxes are associated lower job gains. In my third chapter, I investigate the effect of affordable housing projects on current home prices in areas where there projects locate. This topic is related to both redlining and state tax policies in that the lack of jobs combined with low wages limit affordable housing options and per 2 petuate residential redlining. In this essay I ask, Does the presence of affordable housing projects affect the median area home price in DC? And if so, does the effect depend on area income level? To answer this question, in Chapter 4, I assess the effect of affordable housing on home values in DC at the census tract level. Given that there is significant spatial autocorrelation in median home prices, median income, the number of affordable housing units, and the percentage of non white residents per tract, I employ spatial econometrics methods. I run a spatial Durbin model which regresses the median home price on its own spatial lag, the number of affordable housing units, the median income, the percentage of nonwhite residents, and the spatial lag of the indepen dent variables. The results indicate that the total effect of a unit increase in affordable housing units on the log of median home prices is 0.035 with a p-value of 0.007. This indicates that affordable housing enhances home values in the neighborhood.