--Joy Wiltermuth
A sharp increase in large, delinquent Alt-A residential mortgage loans is anticipated in California over the coming year, adding to already impaired credit fundamentals for the residential mortgage-backed securities. “We’ve addressed the subprime market, but there is no government cure for the Alt-A’s, said Jonah Green, director of mortgage analytics at 1010DATA. “Those are a lot of toxic assets.”
Another area of MBS credit risk is an unknown level of oversupply. “The real question is what is the excess level of housing in the system,” said Samir Shah, head of ABS and MBS at MF Global, adding that current data modeling platforms have yet to pin down shadow housing inventories. The current foreclosure pipeline, however, should give an indication of when loan delinquencies will peak, said Dapeng Hu, managing director at Blackrock, adding that in the subprime market, 50-60% of loans are currently delinquent. “It takes time to clean out the pipeline, not factoring in loan modifications.”
If that happens in early ‘08 or ‘09, there will be a peak of defaults a year later, he added. In the meantime, data providers are looking for new ways to provide investors with more predictive loan metrics, which include using consumer data to model a borrower’s behavior. In the past, the market took note only after a loan had gone bad. One way data metric providers are addressing this now is by trying to utilize more granular data. One disadvantage is a lack of information on borrower’s current income. “You can determine their credit history, whether they have student loans or have recently moved to a new location," Hu said. “Income is the final missing piece.” The other issue is that unemployment is not a uniform data point. "In California, we know that it is construction workers and the transport sector that has been hard hit," Hu said.
But, again, there is a dearth of data on higher income workers. James Yao, v.p. of capital markets at Transunion, said his group is looking at the available data as an actuary exercise. “We don’t have pay stubs, but we have a model based on their credit profile," said Yao. “We provide an estimate of income that is accurate between $28,000 to $30,000," he said, adding that incomes greater than $200,000 prove less accurate.
One bright spot, however, has been a leveling off this year of subprime home foreclosures. One reason has been the government’s $8,000 tax credit for first-time home buyers, analysts said. They would like to see the program extended to the Alt-A and other troubled areas of market.