Credit scores are about to acquire a big blind spot, and mortgage lenders worry they'll be on the hook for bad loans they made while relying on inflated scores.
On July 1, data on tax liens and civil judgments will be scrubbed from credit files under the National Consumer Assistance Plan. The plan, which came about as a
Right now, half of the tax lien and 96% of the civil judgment information in the credit bureaus files does not meet the identification criteria, according to LexisNexis Risk Solutions. That means scores will be less reliable for a significant swath of the population. About 11% of American consumers have a tax lien, a civil judgment or both, the vendor says. (That would total 28 million people, based on U.S. Census information.)
The data being removed from credit files can still be used during the underwriting process if obtained from other sources. But that means more work for lenders, which may have to buy back loans (or indemnify the Federal Housing Administration for losses) if borrowers they mistakenly approved default.
"The fact that isn't on the credit report isn't a showstopper, but it does in some ways increase rep and warrants risk to the lender," Bill Banfield, a vice president at Quicken Loans. "They now need to figure out how do they source that information, how do they officially bring it into the system and allow the underwriters to make decisions to properly underwrite the loan to the program requirements."
It's an additional cost that someone will have to bear and it is likely every lender will have some sort of operational change to accommodate it, Banfield said.
Nearly 700,000 U.S. consumers are projected to have a score increase of 40 or more points as a result of the change, according to Ethan Dornhelm, vice president for scores and analytics at FICO. This is compared to over 11 million consumers who would see score increases of 20 points or less.
Tax liens and judgments are considered derogatory items by the algorithm and are included in the calculation of payment history, which makes up 35% of the FICO score, Dornhelm said.
Consumers with these in their files are 5.5 times more likely to
Even a small increase in credit score could push consumers on the edge into a lower-risk lending tier, especially FHA applicants, said Mortgage Bankers Association President and CEO David Stevens.
Some with scores below 640 or 620 depending on the particular lender's guidelines could now pop over that line and get approved, Stevens said.
Public record data on liens and judgments can also be found during the title search process but only on the seller or a refinance borrower. There is no search on a buyer.
"The lien position has direct implications on a lender's rights and ability to recoup money should a borrower default and the property goes into foreclosure," said American Land Title Association spokesman Jeremy Yohe.
Fannie Mae, Freddie Mac and the FHA all said they are studying the changes for potential impacts on their portfolios and/or lender customers. Top lenders gave similar statements.
Wells Fargo Home Mortgage is "fully aware of the pending changes and we are evaluating the potential outcome,” said spokeswoman Vickee Adams. “We are working with our investors, the GSEs, to understand what impact the removal of tax liens and civil judgments from credit reports may have in possibly improving some customers' credit profiles."
Bank of America is "assessing the change and needs time to assess the data," said spokesman Terry Francisco. "While credit scores are important in underwriting a mortgage, equally important are documented ability to repay, the quality of the collateral, the amount of down payment and reserves."
The industry is looking for guidance from the secondary market around representations and warrants if loans go bad, Stevens said. When it comes to FHA, they are worried about enforcement actions
"I'm not sure from the creditor's standpoint that simply erasing something from a credit report that actually existed makes a better creditworthy borrower," Stevens said. "That just artificially raises the score. So how do you build it into your models?"
