Quantifying Racial Prejudices with Housing Transaction Data

Abstract

This paper aims to quantify racial prejudices using transaction-level housing market data. We pin down the impact of a marginal change of racial composition in a narrowly-defined neighborhood on the price appreciation between repeated sales of a house, and we find that an additional nonwhite household within a radius of 0.2 miles reduces the price appreciation by 0.127 percentage points. The effects are weaker in neighborhoods with a thicker housing market and a higher income level, suggesting that the role of racial prejudices is limited by market forces. The effects are also associated with voting behaviors and public anti-racism statements made on social media, indicating that they are likely to be driven by taste-based racial preferences. While we also find evidence of racial price discrimination effects as in the literature, these effects tend to be a form of statistical discrimination driven by other factors that correlate with race.

Keywords: Housing market, repeated-sales, racial prejudice

JEL Classification: J15, R23, R30, H00