Racial and ethnic disparities in mortgage lending have long been a concern, with minority groups experiencing lower approval rates compared to Whites. This paper examines mortgage lending practices in Boston, focusing on the influence of race/ethnicity on loan approval decisions. Drawing on data from the Home Mortgage Disclosure Act (HMDA) for the year 1990, we explore the factors affecting mortgage approval, including marital status, credit history, other financial obligations, and race/ethnicity. Our analysis aims to shed light on potential discrimination in mortgage lending practices and inform policy interventions to promote fair access to credit.
Controlling for relevant characteristics, is race/ethnicity associated with the outcome of a mortgage loan application?
Econometric Model and Estimation Method: We employ both logit and probit regression models to analyze the relationship between loan approval (a binary outcome) and various predictor variables. The models include marital status, credit history, other financial obligations, loan amount, and race/ethnicity as explanatory variables. Logistic regression estimates the log odds of loan approval, while probit regression estimates the marginal probability of approval. Sample selection criteria were applied to ensure data integrity, and summary statistics by race/ethnicity were computed to examine disparities in loan approval rates.
Both logistic and probit regression analyses reveal significant associations between race/ethnicity and mortgage lending decisions. Controlling for other variables, Blacks and Hispanics exhibit lower odds of loan approval compared to Whites. Moreover, credit history meeting guidelines, marital status, and loan amount are significant predictors of loan approval, with higher odds associated with favorable credit history and marital status. The predicted probabilities of loan approval for prototypical individuals further underscore disparities in lending practices, with Whites consistently having higher approval rates. The findings suggest systemic discrimination against minorities in mortgage lending practices.
The analysis underscores the persistent racial and ethnic disparities in mortgage lending, highlighting the need for policy interventions to promote fair and equitable lending practices. Both logistic and probit regression models provide robust evidence of discrimination against minorities, emphasizing the importance of addressing systemic biases within the lending industry. Policy initiatives aimed at increasing transparency, enforcing anti-discrimination laws, and expanding access to credit for minority borrowers are essential steps toward fostering inclusive homeownership opportunities.
While this study sheds light on mortgage lending disparities in Boston, its findings are subject to certain limitations. The analysis is based on data from a specific time period and geographic location, limiting generalizability. Omitted variable bias and unobserved confounders may also affect the estimated results. Future research should aim to expand the sample size, incorporate income data, and consider neighborhood-level factors to provide a more comprehensive understanding of mortgage lending disparities.