Do Black and Hispanic Homeowners Earn Lower Housing Returns Than White Homeowners?

This study asks why Black and Hispanic homeowners earn lower housing returns than White homeowners. The authors use nationwide administrative data linking race, home purchases, and later sale prices for more than 13 million ownership spells. They find that minority homeowners earn about 2.3 percentage points lower unlevered annual returns mainly because they face far higher rates of distressed sales, such as foreclosures and short sales. When a sale is distressed, homeowners lose large amounts of value—about $8,800 for Black owners and $15,700 for Hispanic owners over ten years relative to White owners. Without distressed sales, racial gaps in returns nearly disappear.

Why This Article Was Selected for The Policy Scientist

This article addresses an important component of long-standing racial wealth gaps by quantifying how distressed home sales shape realized returns, a topic with broad relevance for understanding asset accumulation in the United States. Its timeliness reflects renewed attention to household balance-sheet fragility and the uneven effects of economic downturns. The authors' dataset is unusually large and detailed, enabling precise measurement of housing returns and distress. Although the study relies on observational methods, the empirical strategy is transparent and robust, and the findings appear generalizable to other U.S. housing markets with similar credit structures.

Full Citation and Link to Article

Kermani, A., & Wong, F. (2024). Racial Disparities in Housing Returns. American Economic Review (forthcoming). https://www.aeaweb.org/articles?id=10.1257/aer.20220789

Central Research Question

This article investigates why Black and Hispanic homeowners realize substantially lower housing returns than White homeowners, despite broadly similar rates of property appreciation when homes are not sold under distress. The authors ask whether racial differences in distressed sales—foreclosures and short sales—statistically account for observed gaps in realized returns. The broader research question concerns the mechanisms through which financial fragility, income volatility, leverage, and liquidity constraints shape racial disparities in a core asset class that underpins household wealth accumulation. The study also considers the relative contribution of purchase timing, location, and borrower characteristics in explaining return differentials.

Previous Literature

The article contributes to longstanding research on racial wealth inequality, homeownership dynamics, and housing-market frictions. Prior studies have documented stable racial wealth gaps since the mid-twentieth century (Kuhn, Schularick, and Steins 2020) and large racial disparities in default risk, liquidity, and income volatility. Earlier work measuring racial differences in housing returns relied on neighborhood-level price indices (Anacker 2010; Kahn 2024) or homeowners’ self-reported valuations (Flippen 2004; Faber and Ellen 2016). These approaches do not capture the substantial discounts associated with distressed sales, which Campbell, Giglio, and Pathak (2011) show destroy considerable homeowner wealth. Research on refinancing behavior (Gerardi et al. 2020), discriminatory lending, and credit constraints has documented systematic differences in borrowing costs and mortgage outcomes, while studies of racial gaps in foreclosure during the Great Recession (Rugh and Massey 2010; Gascon, Ricketts, and Schlagenhauf 2017) highlight differential exposure to housing downturns. This article extends the literature by using administrative data to measure realized returns from actual purchase and sale prices, incorporating distressed transactions directly. This approach differs from Diamond and Diamond (2024), who find no racial gaps using a property-year methodology that assigns more weight to homes that avoid distress.

Data

The authors assemble a comprehensive dataset linking homeowner race and ethnicity from HMDA mortgage records to a large national repository of individual property transactions covering purchases mostly between 2000 and 2014, with returns measured through March 2020. This dataset includes purchase prices, sale prices (including distressed sales), sale dates, and ownership durations. Additional administrative data enhance the analysis: McDash mortgage servicing records supply information on mortgage payments and leverage; Equifax credit bureau data provide measures of borrower financial conditions; and loan performance records from Fannie Mae, Freddie Mac, and ABSNet allow for improved measurement of leverage, delinquency, and mortgage modifications. Approximately three-fourths of home purchases in the United States occur with leverage, and racial identification from HMDA records is available only for mortgaged transactions. Cash purchases are studied separately using a larger dataset in which race is imputed using name and Census block information. Ownership spells with complete purchase and sale information span 40 states, with expected undercoverage in non-disclosure states. The dataset is large and detailed: more than 13 million ownership spells are analyzed. The authors benchmark representativeness by comparing key sample characteristics to the American Housing Survey and by confirming that distressed sale rates align with external sources. This level of administrative linkage allows the authors to compute unlevered, total, and levered returns with high precision and to directly examine the role of distressed sales in shaping realized returns.

Methods

The authors compute three measures of annualized housing returns: unlevered returns (ratio of sale price to purchase price annualized over the holding period); total returns (which add imputed net rental flows and account for closing costs); and levered returns (internal rate of return incorporating mortgage cash flows). Distressed sales are identified through foreclosure and short-sale designations. The empirical strategy centers on descriptive and regression-based decompositions that isolate the contribution of distressed sales to racial gaps in returns. Race-specific average returns are compared across distressed and non-distressed sales, and the authors adjust for differences in purchase timing, leverage, income, family structure, and neighborhood by adding covariates progressively. The study uses multiple robustness checks, including reweighting the sample to external benchmarks, imputing race for cash purchases, examining alternative return definitions such as net present value, and estimating returns both with and without implicit rents. The authors also assess the effects of modifying assumptions about implicit rental flows. The study is observational and does not employ causal identification strategies, though it uses causal methods to estimate the effects of mortgage modifications in one section. The analysis leverages the large sample and detailed administrative coverage to minimize measurement error and mitigate concerns about sample selection. The methodological emphasis is on precise accounting of realized returns rather than causal inference regarding the sources of distress.

Findings/Size Effects

The central finding is that Black and Hispanic homeowners earn substantially lower annual housing returns than White homeowners, and these differences are statistically accounted for by higher rates of distressed sales. Among mortgaged purchases, Black and Hispanic homeowners realize unlevered annual returns that are each about 2.3 percentage points lower than White homeowners. Total returns show gaps of 1.3 percentage points (Black–White) and 2.5 percentage points (Hispanic–White), while levered returns show even larger gaps—2.2 percentage points for Black homeowners and 7.6 percentage points for Hispanic homeowners. These disparities largely vanish in the subsample of non-distressed sales: the Black–White gap in unlevered returns shrinks to less than 40 basis points, and the Hispanic–White gap reverses. Distressed sales impose large wealth losses: in the first ten years of tenure, distressed sales erase an average of $8,814 more home value for Black homeowners and $15,676 more for Hispanic homeowners compared with White homeowners. The likelihood of distress is substantially higher for minority homeowners, consistent with racial differences in liquidity, income volatility, credit access, and leverage. The authors show that purchase timing and location explain roughly one-third of the Black–White gap. Income, leverage, gender, and family structure explain another one-sixth. However, even among high-income, low-leverage couples in similar neighborhoods, racial gaps in distress and returns persist. Robustness checks confirm the central mechanism: racial gaps among cash buyers are negligible, and returns inferred using alternative methods or samples are consistent with the baseline results.

Conclusion

This study provides the first large-scale, transaction-level evidence that racial disparities in realized housing returns arise primarily from differential exposure to distressed sales rather than differential rates of home price appreciation. These findings refine the understanding of how housing contributes to persistent racial wealth disparities, demonstrating that wealth erosion during financial distress—not appreciation rates—is the central mechanism. The study extends the literature by showing that distressed sales impose large dollar losses and that racial differences in liquidity and income volatility strongly mediate these outcomes. Although the analysis is descriptive rather than causal, the administrative data’s coverage and precision make the conclusions credible. The results have broad relevance for understanding household financial stability and the role of housing in wealth accumulation across racial groups.