Stop Making Excuses for the CRA

From late-2008 to mid-2009, banks and financial institutions had already written off or written down credit losses on U.S. mortgages by over $1.6 trillion according to the OECD.

Charge-offs on just the CDO-financed portion of the subprime lending debacle were recently estimated by economists at the Federal Reserve Bank of Philadelphia to approach a half trillion dollars. The Community Reinvestment Act of 1977 is still the law of the land and, if it was a root cause of the last financial crisis, as opponents conclude, then it likely will be again.

CRA remains a point of contention. Its proponents, most recently Ellen Seidman and Mark Willis in American Banker, attack research purporting to find a link between the CRA loans and ultimate credit loss. CRA supporters implicitly share the official explanation that greed got the better of lenders and regulators were powerless to stop them.

CRA purportedly reflected a concern that local bankers were not lending "enough” to good borrowers in their communities or neighborhoods, which were typically characterized by ethnic and/or racial concentrations. This systemic market discrimination supposedly persisted throughout the last century. During the last decade, while greedy mortgage lenders were originating about 20 million subprime loans to borrowers of dubious credit, CRA proponents implicitly argue that lenders would still have discriminated systemically against profitable CRA-qualifying borrowers but for the requirement. Is any of this credible?

Competition in the mortgage banking industry was first labeled "cutthroat” in 1944 and has been ever since. Because older inner-city neighborhoods often had a much higher percentage of African-Americans — and, later, other racial minorities — the implicit concern of regulators and economists was illegal racial discrimination. Because incomes were also generally much lower in older inner-city neighborhoods, and the risk of a systemic decline in property values much greater, it was generally difficult to distinguish illegal racial profiling from legal credit discrimination.

But, by the end of 1977, there were literally thousands of potential loan brokers who would profit from originating and selling loans made in older inner-city neighborhoods if they could be underwritten to the standards of the most liberal investors nationwide, including the Federal Housing Administration. Local branch offices of banks were rarely responsible for mortgage lending decisions. Moreover, banks decided whether to sell or to hold mortgages for lots of financial reasons, so while there was plenty of competition to make good loans, this was unlikely to be reflected in the portfolio of any "community” banker.     

In 2012, Oonagh McDonald found the CRA economic studies using Home Mortgage Disclosure Act of 1975 data to be seriously flawed, concluding that "the HMDA data did not serve a useful purpose throughout their history …” because the data didn't allow tracking the performance of loans subsequent to qualifying for CRA credit.

Subsequent statistical analysis of credit costs – whether low or most recently high – can always be criticized, but the same criticisms apply to studies purporting discrimination in the absence of CRA. You can buy McDonald's excellent book (for about $100) or ignore this ongoing methodological debate as irrelevant to what happened during the last decade. 

From the very beginning prudential regulators used poor grades on CRA assessment reports based on HMDA data to withhold approval for branching and merger applications in response to political advocacy group pressure to extract easier terms for select borrowers. How this got out of hand leading to $3.8 trillion of CRA-related commitments coming due between 2000 and 2007 is explained in a prior BankThink column.

These new CRA commitments – the majority agreed to by the "too-big-to-fail” banks – required that banks unleash their mortgage banking subsidiaries. Bank mortgage holdings doubled from $2.5 to $5.0 trillion, but almost $1 trillion of that were second mortgages. Banks could originate loans for CRA certification then sell "servicing retained" in pools of private-label mortgage securities, many of which were then sold to Fannie Mae and Freddie Mac to meet their affordable housing goals. Banks also bought about $0.5 trillion in PLS, likely for CRA credit with less than half the capital required of whole loans.

Holding the second loan while retaining servicing of the first probably continued to allow the loan to count for CRA, and second mortgage loans had a 100% risk weighting except when combined with a first mortgage lien, in which case it was only 50%., which would explain why the four CRA-driven TBTF banks held a disproportionate 45% of the stock of second mortgages.

Hence the bulk of CRA-related losses are likely found in CDO securities and second mortgage portfolios. That doesn't prove CRA was the cause, but CRA proponents have offered no convincing evidence of innocence or a need that justifies the risk. The Fed, our erstwhile systemic regulator, and the Federal Deposit Insurance Corp., our erstwhile prudential regulator, both shared CRA enforcement responsibilities and both failed massively and chronically in the last financial crisis.

I would amend the FDIC mission statement to "as an independent regulator, we will protect the FDIC insurance fund from the errors of bankers, the overconfidence of borrowers and the folly of politicians, political advocacy groups and other bureaucracies such as HUD and the CFPB" and hold them to it. 

Proponents support the status quo.

Kevin Villani, chief economist at Freddie Mac from 1982 to 1985, is a principal of University Financial Associates and an executive scholar at the Burnham-Moores Center for Real Estate of the University of San Diego.