(Georgia Institute of Technology, 2022-11)
Calhoun, Rylee; Xia, Maggie
This paper seeks to uncover potential mortgage loan discrimination related to family size for Boston
mortgage loan applicants in 1989. We utilize mortgage loan amounts as the primary dependent variable
and an applicant’s number of dependents as the primary independent variable to explore this relationship.
The dataset analyzed in this paper, loanapp, comes from Dr. Jeffrey Wooldridge’s introductory
econometric data repository that records data from the 1989 Boston Federal Reserve Bank’s study of the
Boston metropolitan area’s mortgage practices. After implementing multiple regression models, we find
no evidence of statistical discrimination concerning family size, suggestive by the insignificance of our
primary independent variable, dep, in regression analysis. Specifically, we find that a single increase in an
applicant’s number of dependents raises mortgage loan amounts by about 0.381% overall and raises loan
amounts by 0.137% for applicants who have at least one dependent. Ultimately, we reject our null
hypothesis that increases in family size negatively influence mortgage loan amounts.