ÌìÃÀ´«Ã½

Newswise — The  (SERC) at the University of Maryland’s Robert H. Smith School of Business announces the rollout of a pair of mortgage credit risk indexes to guide lenders, servicers, credit investors, regulators and other market participants and inform their view of changes in credit risk of GSE-eligible mortgages.

The separate online resources, the Mortgage Credit Risk Index (MCRI) and Mortgage Redtail Risk Index (MRRI), represent “the newest advancement in SERC’s mission to promote the dissemination of leading risk management practices and tools to industry and government,” says Professor of the Practice and SERC Director.  

The MCRI measures the expected 3-5-year default risk for a given loan aggregated to each quarter for loans sold to Freddie Mac and Fannie Mae. Higher scores signify greater credit risk with every 40 points of MCRI score doubling the odds of default. The MCRI ranges between 300 and 900.  

“Mortgages are a huge portion of household debt, accounting for about $13 trillion and 70% of household debt according to the ,” Rossi says. “Understanding the trajectory of mortgage credit risk is critical to maintaining a healthy U.S. economy.”

The MRRI measures the degree of risk layering or potential adverse selection of loans sold to the GSEs. “Specifically, loans with greater combinations of high-risk factors such as low credit scores, higher loan-to-value ratios and high debt-to-income ratios signify risk layering is present,” Rossi says. “While the MCRI provides an assessment of the overall credit risk of new loans being originated, the MRRI provides an estimate of the changes in the share of new originations with the highest combinations of risk factors.”

Rossi, along with graduate students in Smith’s  program, developed both indexes based on a set of proprietary algorithms leveraging machine learning technology and well-established mortgage credit risk analytics.

The MCRI was developed from four million GSE-eligible loans originated between 2000 and 2018 with loan performance through Q224, says Rossi, a former risk executive with deep mortgage credit experience.

“The multivariate model underlying the MCRI takes into account key borrower, property, loan and other factors to develop a forward-looking comprehensive view of mortgage default risk,” he says, adding: “The model has been validated out-of-sample across origination years and along key risk attributes to establish its predictive quality.”  

SERC will update both indexes quarterly, as the GSEs release new data. 

The indexes are accessible , with more information forthcoming via the .