Credit bureau TransUnion and mortgage data provider First American CoreLogic have joined forces to provide more robust information regarding the quality and value of residential mortgage-backed securities (RMBS).
Instead of relying simply on pool-level home price appreciation (HPA) and initial loan-to-value (LTV), potential investors will have access to complete adjustable-rate mortgage exposure and a consumer’s capacity to repay the loans in question.
“Billions of dollars in mortgage securities were traded without visibility into the risk of the underlying borrowers of the loans backing the securities,” said Jeff Hellinga, president of TransUnion’s U.S. Information Services division, in a release.
“This was sufficient as long as property values continued to rise. But now, with the collapse of the housing market, direct insight into the actual risk of the underlying borrowers is critical.”
In other words, home price appreciation no longer masks borrower risk, one of the many factors that led to this unprecedented mortgage crisis.
The so-called “TransUnion Consumer Risk Indicators for RMBS” link individual loans within non-agency mortgage-backed securities to consumer credit profiles of the specific borrowers, instead of basing value on “aggregate” characteristics such as loan size and loan-to-value.
As a result, two seemingly similar pools of loans with like qualities are being broken down further, factoring in things like the credit utilization and total debt outstanding of individual borrowers, which may result in very disparate default rates and subsequent valuations.
“This is particularly relevant given the recent creation of the Public-Private Investment Program (PPIP), which is designed to draw private capital into the market to facilitate price discovery of legacy assets and the expansion of other government programs, such as the Trouble Assets Relief Program (TARP) and the Term Asset-Backed Securities Loan Facility (TALF).”
An initial TransUnion analysis found that improvement in default prediction was up more than 15 percent compared to current predictive methodologies.
(photo: wetwebwork)