On September 17th, the industry’s top quant minds gathered to discuss the state of the credit market and how lenders can use alternative data to widen the credit box. View the #NEXTONLINE webinar, presented in partnership with RiskSpan.
Unique as it is, the COVID-19 crisis has amplified a familiar response to mortgage lending risk. The mortgage market is going very well, said Laurie Goodman, director, Housing Finance Policy Center, Urban Institute, at the #NEXTONLINE webinar: Using Alternative Data to Widen the Credit Box, but a look at the mortgage conditions beyond the rates, prices and home sales shows “it’s less pretty.”
The Mortgage Bankers Association Credit Availability Index shows there are “huge decreases in credit availability, Goodman said. After loosening up before the great recession, starting early in 2011 and up to 2019, credit tightening has continued. If credit was tightening for years, she said, Covid-19 has intensified it.
Currently the decrease in credit availability overall is down 33%, at about 19% in the government backed loan market, 23% in the conforming market, and 58% in the jumbo market which does not enjoy any government guarantees. “Even within the government market there has been a very appreciative tightening of the credit,” for both purchase and refinance.
For example, it is getting harder and harder to qualify for borrowers with FICO scores less than 700 and debt-to-income ratios greater than 40. In January 2019 44% of Ginnie Mae purchase mortgages were in that category, by Jan. 2020 the number fell to 38%, and now is 36%; compared to 39% in Jan. 2019; 13% in Jan. 2020 and 5% now, for refinancing.
Data show the credit box has tightened for Fannie Mae and FHA borrowers and across the board for all other loan types due to a combination of reasons including the Coronavirus Aid, Relief & Economic Security (CARES) Act forbearance relief measures.
What’s behind the credit squeeze?
The credit tightening is occurring for three reasons, according to Goodman.
Initially the agencies would not buy loans that were in forbearance, she recalled. That requirement changed. The borrower should request forbearance, than the agencies would buy the loan, but the agencies also attached a penalty of 5-7 points to it. For example, for Federal Housing Administration (FHA) loans the servicer would have to absorb 20% of the eventual loss if the loan misses two payments in the first two years.
Servicers reacted to this requirement by tightening the credit box, “to make sure that if they make the loan in good faith will be able to sell it without paying a penalty,” she said. “We estimate that approximately 255,000 borrowers have lost access due to credit tightening, which means 1% fewer purchase loans and 5% fewer refi loans that otherwise would have been.”
Furthermore, according to Goodman, this requirement has produced “a negligible amount of revenue to the FHA and the GSE’s, but it has had a significant impact on the credit box.”
Another factor that affected credit tightening is the high cost of servicing of delinquent loans. MBA data suggests it cost $150,030 a year to service a performing loan and $190,069 a non-performing loan, with FHA loans being three times more expensive to service than GSE loans.
Lastly, advancing “weighs heavy on servicers, especially in the government loan market,” she explained, since if the homeowner misses a payment, the servicer is obligated to advance the payment on these loans to investors. As a result, forbearance places a burden on servicers requiring them to advance the principle and interest payments, taxes, insurance, the GSE guarantee fees, the FHA premium and Ginnie Mae fees.
On agency loans, servicers have to advance payments for four months, which forces the servicer to buy the loan from the pool, or continue to advance, “which is a very heavy burden” on them, she said. Consequently, due to liquidity constrains, especially non-bank servicers are particularly reluctant to lend to more risky, and more costly borrowers.
“Forbearance was very, very, very well intentioned, but I think there have been some implications for the servicers that were not completely thought through in the congressional mandate. Obviously, Fannie and Freddie have done what they could to try to alleviate this, and I think that Ginnie and FHA have gone as far as they can,” Goodman said.
“Certainly, penalties for purchasing loans in forbearance at the time of sale, which is very easily corrected by administrative actions, is probably the biggest reason we had a tightening of the credit box; and that is very easily fixable from an administrative point of view.”
Behind the CARES Act’s consequences
“There’s a lot of disruption in the market,” agreed Amy Crews Cutts, president, AC Cutts and Associates, the CARES Act and its requirement to the Credit Bureaus to suppress credit reporting is one of them.
It is a positive pro-consumer first order effect, with a negative second order effect for servicers, she said. “The sad thing is that when Congress changed the Fair Credit Reporting Act …there was no new reporting code introduced.” The credit reporting agencies had to improvise.
“A mortgage borrower has the option to apply for two consecutive six month deferrals, but all federally-backed student loans are automatically deferred until September 30, 2020, Crews Cutts said, which means 91 percent of all student loans are now deferred. Unlike with mortgages, the CARES Act requires not just that the account status is frozen at the level the account was in when the forbearance was granted, but that deferred delinquent federal student loans be reported to credit bureaus as current during this period. Eight million student loan accounts that were delinquent are now reported as current. So at least until October their credit scores will positively reflect this suppressed status. And because there is no CARES Act reporting code for these accommodations, the lender can’t distinguish the reason for the accommodation and has to look beyond credit scores for signs of stress.”
“People keep calling it a recovery as if we are on a rocket ship here and going into a full recovery of the economy, and I want to caution everybody,” she said. “We have bounced off the bottom, and it was a terrible bottom, I think in 2008 it unfolded fast…but it cannot be compared to the pandemic and how fast things happened in March. Since then we have regained half of the payroll jobs that have been lost, so we are not back to normal by a long stretch.”
Who is being affected the most?
So far, it has been great for homeowners, because if they do not have a job and are facing hardship they have forbearance, but it is not so great for renters, said Goodman. As a rule, homeowners are a more affluent group and tend to have a lower unemployment rate than renters, who tend to be employed in the five more vulnerable industries, such as hospitality and entertainment.
According to the MBA, of the total number of homeowners 4.6% of fannie and Freddie borrowers, 9.1% of Ginnie Mae borrowers and about 10.7% of other borrowers are in forbearance. “The MBA has been pleasantly surprised” by the strength of the housing market, she said, homeowners have been much less affected by unemployment, and those who have their jobs are in good shape, which resulted in incredible strength in the market.
The MBA’s weekly Mortgage Purchase Application Index shows that after a dip in March, the index has recovered and even increased compared to previous years, “reflecting the strength of the first-time buyer market taking advantage of the generational lows in interest rates Interest rates have never been this low before.” Even though home prices have risen very rapidly, they can afford to pay more for their dream home.
This crisis certainly did not affect everyone the same, said Janet Jozwik, managing director, RiskSpan, Inc., states with higher forbearance such as New York and Nevada have 14% of loans in forbearance, compared to 4% in many Western states. Forbearance data tracking is important at this time in determining origination risk levels at the state and county level, she said, and Covid-19 has provided additional, more granular data that is updated daily or weekly, as opposed to monthly.
The pandemic predictions tap on roughly six months of data, which is a limited, so the ability to target all datasets more specifically is key to predicting risk drivers, said Jozwik. Using GPS data, for instance, tracking mobile phone data, even Covid-19 data helps. Finding the right risk modeling techniques that use all the granular data, including behavior estimates, to refine the analysis and locate the population of acceptable risk loans, help lenders to make informed decisions and rethink pricing parameters.
Can we make more loans with less risk if we start incorporating more data into the analysis, she said. “Those 255,000 loans that Goodman mentioned, that could have been originated for instance, for an originator, that is money left on the table.