“Institutional Racism and Black Mortgage Denials: What Google Searches Reveal”
A presentation by Prof. Nolan Kopkin of the Department of African and African Diaspora Studies
April 6th, 2018
Mitchell Hall 206
Past research uses survey evidence of racial prejudice to show that black-white gaps in both conventional mortgage denial rates and conventional subprime mortgage originations are conditionally correlated with prejudicial attitudes towards blacks at the county level using counties covering approximately 40% of US housing; this research also provides evidence that these positive associations do not exist in the Federal Housing Administration (FHA)-insured market.1 In this study I use a different measure of racial prejudice – the proportion of Google searches including the ‘N’ word and its plural measured by media market. Using nearly 90% of all applicable mortgage applications in the US, I show that the proportion of Google searches including the ‘N’ word is conditionally correlated with the black-white conventional mortgage denial gap such that a 50 percentage point reduction in the search proportion is associated with a 3 percentage point narrowing in the black-white conventional mortgage denial gap. However, associations between the search proportion and black-white gaps in FHA-insured denial rates and conventional and FHA-insured subprime mortgage originations are not robust to the inclusion of social and economic characteristics. Results largely corroborate past findings and imply reducing barriers to outside competition and implementing stricter enforcement of anti-discrimination laws in jurisdictions where racial prejudice is measurably high may help correct the situation.
1 Kopkin, N. (2018) The conditional spatial correlations between racial prejudice and racial disparities in the market for home loans. Urban Studies, Online First: 1-19.