But What's the Default Rate?
Steve Sailer discusses the alleged bias against minorities in subprime lending reported in the Times. The allegation is based on a report from The Furman Center for Real Estate and Urban Policy at New York University. In an accompanying editorial, the Times argues that "the burden of proof has to be on the lenders to show that no discrimination has occurred."
This is an astonishing statement given the contents of the "study" - really just a glorified homework assignment. If you look at the "analysis", it is a compilation of "Percent of HP Loans Issued by Subprime Lender" by neighborhood combined with demographic and median income data by neighborhood from census data. The analysis consists in eyeballing the top ten and bottom ten rates and noting that most subprime lending happens in minority neighborhoods and that the median incomes for some of them are not way different. For that the Times wishes to engage a witch hunt?
But finding discriminatory lending practices is really quite simple. There is a simple acid test you can apply to falsify the discrimination hypothesis. If minorities are indeed being discriminated against by disproportionately being given subprime mortgages, then we should find that their default rates are lower than that for non-minorities. It's pretty simple - the allegation is that too many minorities (compared to non-minorities) who get subprime loans are actuality better credit risks than than 'subprime' and should have gotten standard loans. If that is indeed the case, then the minorities should be defaulting at a lower rate. Is this what's happening? It's hard to imagine we wouldn't have heard about it if it were true, but I don't really know. I'll take a look later to see if there are any such stats available. But there's no need to speculate based on crude data as the Times is touting when a foolproof test (so rare in the real world!) exists.
Related: A 1999 study (Anderson/Vanderhoff) on national default rates disproving systematic discrimination in lending.
This is an astonishing statement given the contents of the "study" - really just a glorified homework assignment. If you look at the "analysis", it is a compilation of "Percent of HP Loans Issued by Subprime Lender" by neighborhood combined with demographic and median income data by neighborhood from census data. The analysis consists in eyeballing the top ten and bottom ten rates and noting that most subprime lending happens in minority neighborhoods and that the median incomes for some of them are not way different. For that the Times wishes to engage a witch hunt?
But finding discriminatory lending practices is really quite simple. There is a simple acid test you can apply to falsify the discrimination hypothesis. If minorities are indeed being discriminated against by disproportionately being given subprime mortgages, then we should find that their default rates are lower than that for non-minorities. It's pretty simple - the allegation is that too many minorities (compared to non-minorities) who get subprime loans are actuality better credit risks than than 'subprime' and should have gotten standard loans. If that is indeed the case, then the minorities should be defaulting at a lower rate. Is this what's happening? It's hard to imagine we wouldn't have heard about it if it were true, but I don't really know. I'll take a look later to see if there are any such stats available. But there's no need to speculate based on crude data as the Times is touting when a foolproof test (so rare in the real world!) exists.
Related: A 1999 study (Anderson/Vanderhoff) on national default rates disproving systematic discrimination in lending.
posted by ziel | 10/18/2007 08:29:00 AM
2 Comments:
Actually, I think you are probably right, but I feel constrained to point out there is a theory which may be able to account for the default statistics without excluding racism in loan origination:
Since racist lenders would (in effect) demand higher APRs from blacks, black borrowers might be more likely to default because they get in over their heads. To put that another way, higher APRs demanded from blacks by racist lenders could lead, rather than follow, black default statistics. Racist loan-underwriting might be a "self-fulfilling prophecy" problem. If blacks got more affordable loans, perhaps they would default on them less often.
This line of thinking provokes the question: do blacks default more often than whites with the same loan terms?
I think the default rate analysis would be based on the terms of the loan as well as the credit-worthiness of the applicants. My suspicion is, based on the historically higher default rate, is that blacks default at higher rates at every measure of credit qualification. You have a number of dynamics at play: mathematical aptitude, ability to search for preferable rates, conscientiousness in paying off loans, savings tendencies, "big-spender" syndrome - there's a whole constellation of traits where it's reasonable to assume group differences would manifest themselves in lending patterns.
Post a Comment
<< Home