Racial discrimination and housing outcomes in the United States rental market
What is this research about and why did you do it?
Racial discrimination continues to constrain the housing decisions of minority households in the US. Research design and measurement challenges have limited our ability to learn about implications of discrimination for broader patterns of racial inequality. We conduct the largest correspondence study conducted to date in the rental housing market, encompassing 50 major cities in the U.S. Our analysis is powered to document patterns of discrimination across U.S. regions, and to explore relationships between discrimination, segregation, and economic opportunity.
How did you answer this question?
Using a software bot developed at the National Center for Supercomputing Applications, our study examines more than 25,428 interactions between property managers and fictitious renters engaged in search on an online rental housing platform, revealing patterns of discrimination encountered in the initial stage of a search. The experimental evidence enables statistically powered analysis at neighborhood (census tract) or city scales.
What did you find?
Hispanic and African American renters continue to face constraints from discrimination in most metropolitan markets. We find evidence of relative response rate differentials greater than -30%. Moreover there is substantial heterogeneity in discrimination facing renters from different racial/ethnic groups even when interacting with the exact same property managers in a given market. Discriminatory constraints are also stronger in cities with higher levels of residential segregation. Outcomes of our experiments predict actual outcomes in rental markets with respect to the race of the tenants who end up in these units.
This figure plots relative response rate differentials for the 50 markets in our sample for African Americans (top) and Hispanic/LatinX (bottom) identities. Relative response rate differentials divide the percentage point difference in the probability of a response for each of the groups of color by the baseline response rate to inquiries from white identities and subtract 1 from this number. For example, a white response rate of 0.6 along with an African American response rate of 0.54 would yield a relative response rate differential of -10%. We find evidence of differentials greater than -30% in magnitude.
What implications does this have for the research on wealth concentration or economic inequality?
We analyze how discrimination varies with an indicator of economic mobility derived by Chetty et al. (2020), and find evidence of a strong relationship for African American but not Hispanic/LatinX households. This is strongly suggestive of a link to neighborhood sorting patterns, given that the intergenerational income mobility gap has declined in recent decades for Hispanic/LatinX but not for American American households.
What are the next steps in your agenda?
In completed work, we examine the role of discrimination in steering households into polluted neighborhoods. In related ongoing work, we combine these methods with the structure of a residential sorting model to measure the welfare consequences of discrimination.
Citation and related resources
This paper can be cited as follows: Christensen, P., Sarmiento-Barbieri, I., and Timmins, C. 2021. 'Racial discrimination and housing outcomes in the United States rental market.' NBER Working Paper no. w29516.
- Christensen, P., and Timmins, C. 2021. 'The damages and distortions from discrimination in the rental housing market.' NBER Working Paper no w29049
- Christensen, P., Sarmiento-Barbieri, I., and Timmins, C. 2022. 'Housing discrimination and the toxics exposure gap in the United States: Evidence from the rental market.' Review of Economic and Statistics, 104(4), pp. 807-818