A Wall Street Journal commentary piece yesterday (regrettably, full online text is available only to WSJ subscribers) provided one (albeit sharply critical) perspective on how the US DoJ applies the disparate impact doctrine in the context of loan fee negotiations. (An alternative perspective from the New York Times, describing the Countrywide settlement, is here.) As the WSJ's Holman Jenkins describes, Countrywide "behaved much like a car dealer, setting "sticker prices" above the market price." Thus, customers were incented to bargain down from the sticker price while Countrywide's loan officers and brokers were incented to resist during negotiations.
Now let's turn to the data. According to Jenkins, DoJ found that "out of 4.4 million loans approved between 2004 and 2008, 525,000 went to African-American or Hispanic borrowers, of which some 210,000 paid higher fees or rates than the average paid by similarly situated "non-Hispanic White Borrowers"." Jenkins's claim of "statistical malpractice" flows from the obvious point that a "large numbers of white borrowers also paid higher than the average of all whites. It also goes without saying large numbers of minorities didn't pay higher rates, though Justice isn't interested in the average of what minorities paid, only that some minorities paid higher than the average of whites."