Default Rates Aren't The Way To Determine Bias In Boston
Gary Becker criticizes the Federal Reserve Bank of Boston's study showing that minority mortgage applicants are more likely to be turned down than whites ("The evidence against banks doesn't prove bias," Economic Viewpoint, Apr. 19). As Becker acknowledges, the study considered each applicant's credit history, including past defaults. But he faults the study for failing to examine the default rates and profitability of loans to minority and white borrowers.
The argument that discrimination would be indicated by lower average default rates for minority borrowers depends on three assumptions: (1) Lenders are good predictors of default; (2) discrimination requires minority applicants to meet a higher standard of creditworthiness; and (3) the distribution of economic characteristics among white and minority borrowers is similar enough that setting a higher hurdle for minorities results in a pool of accepted minority applicants who are less likely than accepted whites to default.
Whatever one's views about the first two assumptions, the third is invalid for applicants in Boston, and probably elsewhere. The pool of accepted white applicants had a larger fraction with very low obligation and loan-to-value ratios relative to successful minority applicants. The data suggest that minority borrowers may be concentrated just above their relatively higher acceptance threshold, while the group of white applicants who clear their threshold includes a larger proportion at the upper end of the creditworthiness spectrum. Such a disparity means the whites as a group could show a lower default rate, despite a lower threshold for acceptance.
As a final point, the similarity in foreclosures in Boston provides little insight, since many minorities purchase houses in predominantly white neighborhoods and whites also buy in minority areas.
Lynn E. Browne
Acting Director of Research
Federal Reserve Bank of Boston