In the years leading up to the housing boom, the government was very busy promoting “hard data” mortgage underwriting standards that dangerously under-estimated the riskiness of loans. Risk models that were built on data where lenders used “soft-information” no longer accurately predicted defaults. It’s another way the Community Reinvestment Act and other government fair lending laws required and encouraged lax lending standards that contributed to the later mortgage boom.
Research from the Stephen M. Ross School of Business demonstrates that banks used statistical models in an attempt to predict loan defaults that failed to warn lenders about risky borrowers. The problem was that the models relied too much on hard information such as credit scores and loan-to-value ratios. Old standards, so called “soft information” that came from personal contact with borrowers or knowledge of local circumstances, were abandoned in favour of the “hard data.” As a result, risk models built on historical data that incorporated periods in which soft information was used under-estimated defaults.
This change was driven in part because of the need for banks to satisfy CRA examiners who looked almost exclusively at “hard data” and the lending data disclosures the banks were required to make under the Home Mortgage Disclosure Act. When combined with securitization, where investors and ratings agencies also looked to the “hard data,” lenders were incentivized to rely increasingly on hard data.
The HMDA requires banks report mortgage application acceptances and rejections by the race, age, sex, income and the geographic location of the applicants. This “hard data” driven regulations forced banks to abandon or highly discount the use of “soft information” to avoid adverse publicity.
Early HMDA statistics showed a lower acceptance rate for low income and minority mortgage applications, embarrassing the banks and contributing to the demand for a strengthen CRA.
To quote from the studies author Uday Rajan:
“In addition to collecting hard data about a borrower, such as a credit score, a lender also has an incentive to verify undocumented information, or soft information, about the borrower,” Rajan says. “In particular, the lender screens out borrowers who are poor credit risks based on their soft information.”
But when the incentives change, so that hard information is emphasised and soft information downgraded, this check no longer applies.
“As a consequence, borrowers who are poor credit risks on the dimension of soft information, but apparently creditworthy based on their hard information, also receive loans. Thus, when one examines loans that have been approved, the same hard data have very different implications for borrower creditworthiness with and without securitization. That is, the hard information can mean something very different across these two worlds.”
While Rajan emphasises the role of securitization, the HMDA-CRA nexus also contributed to the downplaying of soft-information.
This was not some theoretical possibility. Banks actively upended their mortgage lending structure in pursuit of CRA loans. At one point, Bank of America took away the power of any of its local branch offices to decline loans. It then sent senior managers from Charlotte to supervise every single mortgage application. This occurred as early as 1988, long before securitization took off. Soft knowledge was downgraded in favour of hard knowledge, all in pursuit of CRA compliance.