A controversial legal theory on discrimination threatens to disrupt consumer lending and poses a potential liability for anyone securitizing these loans.
Called "disparate impact," it is being used by the Department of Housing and Urban Development to cite lenders for unintentional discrimination against minorities and others under the Fair Housing Act. The Consumer Finance Protection Bureau has also used disparate impact to cite finance companies under the Equal Credit Opportunity Act, or ECOA.
The theory is being challenged: in November, a U.S. District Court blocked HUD's use of disparate impact in a fair housing case, and this has ignited industry hopes that the Supreme Court will take a similar stance on a separate, larger case pending before it. Should the high court support the use of disparate impact, however, regulators would apply it even more aggressively.
The risk: Originators of loans deemed to have been discriminatory would have to track down where the loans reside and securitizations' representations and warranties would likely require the seller to buy them back, assuming it is still solvent and hasn't chosen instead to make lump sum payments to the borrowers.
"Let's say a bank charges a higher rate of interest on a loan, that loan is securitized, and the CFPB or DOJ [Department of Justice] comes back later and says the bank or finance company overcharged the borrower in violation of fair lending laws," said Jeffrey Taft, a partner at Mayer Brown. "If it securitized that loan and doesn't own it anymore, there's some concern about liability emerging after selling those loans, and potentially less flexibility in any remediation efforts."
In June 2013, two insurance-industry trade groups filed a lawsuit contesting a rule enacted by HUD to codify a longstanding disparate impact doctrine. HUD sought the rule in part because the two earlier cases contesting the doctrine were settled before the Supreme Court had a chance to review them.
"There was a feeling HUD was going through this rulemaking proceeding both as a prelude to greater enforcement activity and to bolster judicial deference to its position, since having a formal rule out there would codify its interpretation of the statute," said John Culhane, a partner at Ballard Spahr.
A district judge recently decided in favor of the insurers. On Nov. 3, Judge Richard Leon of the District of Columbia ruled that the use of disparate impact violates the Fair Housing Act, which he said only prohibits disparate "treatment," or intentional discrimination. That case is expected to be appealed and could eventually reach the Supreme Court.
The nation's highest court has already agreed to hear a case between the Texas Department of Housing and Community Affairs and the Inclusive Communities Project that also turns on whether disparate impact is valid under the Fair Housing Act.
Neither of the two cases extends directly to ECOA, which applies to all kinds of consumer credit including auto loans and credit cards. However, observers think that a ruling limiting the use of disparate impact on housing finance could only entice companies to challenge the CFPB and other regulators on using the same theory under ECOA.
"It's likely the CFPB would say it's a different statute and will continue to pursue disparate impact theories of liability until there's a controlling court decision that specifically rejects the viability of that theory under ECOA," Culhane said.
The CFPB's use of the doctrine is more problematic for lenders and securitizers than HUD's use. That's because, unlike the Fair Housing Act, ECOA forbids lenders from collecting personal information such as race and gender from borrowers. So the CFPB uses another data-intensive, and controversial, method that combines borrowers' surnames and geographical information to determine, by proxy, the probabilities of their races or ethnicities.
"Research has found that this approach produces proxies that correlate highly with self-reported race and national origin and is more accurate than relying only on demographic information associated with a borrower's last name or place of residence," the CFPB said in a report issued over the summer.
A separate study published by the American Financial Services Association in November comes to a different conclusion: It found that the CFPB's proxy methodology exaggerated the price disparity between what dealers charged minorities and non-minorities. "AFSA's results are much lower than what the CFPB alleges as problematic in the marketplace, because the association's study factored in complexities of the automotive market that the CFPB did not consider, and errors associated with CFPB methodology," AFSA President and Chief Executive Chris Stinebert said in a statement.
AFSA tested the methodology's probabilities against a test population of mortgage data, in which race and ethnicity are known. Among its findings were that when the proxy methodology finds an 80% probability that a person is African-American, this determination was correct less than 25% of the time.
Beyond accuracy, there are concerns about the impact that the use of disparate impact will have on underwriting. Paul Compton, a partner at Bradley Arant Boult Cummings, said disparate impact claims in the mortgage market have resulted in much more rigid underwriting standards, potentially inhibiting originations. These claims are also one of many factors damping private-label mortgage securitization.
The auto loan market is currently booming and underwriting continues to loosen. However, applying the disparate impact doctrine to non-mortgage consumer loans could have a similar impact, Compton said, noting that it both lowered credit score requirements and down payments at Fannie and Freddie beginning in 1999 and limited mortgage credit criteria, resulting in lower quality mortgages. For non-mortgage lenders, however, the impact could be even more problematic.
"Lenders can't even test themselves because they don't have the data, and the data the CFPB creates may be gravely flawed," Compton said. He noted that courts have generally rejected disparate impact in the context of employment discrimination claims. "Lenders are left to either accept the theories of CFPB or be prepared to litigate against allegations of racial discrimination based upon invisible regulations enforced by statistical regression formulas."
The CFPB recently acknowledged the disparate impact conundrum that lenders may face in the context of mortgage loans. As the regulator developed rules for qualified mortgages, which provide lenders with certain legal protections for following specified lending standards, concerns arose that lenders choosing to offer only QM could later face disparate impact claims if such a decision had an adverse impact on potential minority borrowers.
The CFPB and banking agencies responded to those concerns in October 2013, saying " the agencies do not anticipate that a creditor's decision to offer only [QM] would, absent other factors, elevate a supervised institution's fair-lending risk."
Robert Ledig, a Washington partner at Dechert LLP, said that lenders have long sought an indication of what types of business justifications regulators would find to be potentially acceptable in a disparate impact situation, making the government's response very significant.
In their statement, the agencies pointed to a number of such potential business justifications for limiting their lending to only QM loans, including credit risk, secondary market opportunities, capital requirements and liability risk. The statement provides some clarification, and would likely be used by lenders in other asset categories. Still it doesn't prevent the government from pursuing disparate impact claims stemming from the use of those business justifications or other data points.
"While the agencies noted certain potential business justifications, they left themselves flexibility to deal with individual situations as they arise," said Ledig.
Ironically, the concept of disparate impact appears to have played a critical role in watering down requirements for qualified residential mortgages, reintroducing risk that Dodd-Frank had sought to eliminate. Originally, QRM required borrowers to make a 20% down payment and to have a debt-to-income ratio of no more than 36%. However, in the final rule published in October, the down-payment requirement disappeared and the debt-to-income ratio rose to 43%.
Adhering to the initial and much more conservative QRM standards relieved securitization sponsors of the requirement to keep retain exposure to their deals. When the government eased the requirements for QRM in the final risk retention rule, it referred to concerns about the adverse impact of a very narrow rule on minority borrowers, in language that resembled concerns associated with disparate impact theory.