How Fannie Mae, Freddie Mac and non-QM digital data use compares

Nontraditional lenders making loans in the non-qualified mortgage market (which lies outside current regulatory indicators of an ability-to-repay) originally started to turn to things like digital bank data as an alternative or supplement to traditional credit qualifications.

The same in recent years has been true to an extent for the GSEs, but their borrowers are more likely to be a lower income buyer with a thin-credit file than a wealthier self-employed business owner with irregular income.

To be sure, some overlap may occur given the regulatory framework no longer categorizes Fannie and Freddie loans automatically as QM, but generally the two remain separate markets.

Although consideration of some nontraditional forms of income may be more longstanding in non QM lending, the potential efficiency gains may be larger at the GSEs, because their underwriting is more standardized and less piecemeal.

“There will always be an aspect of manual underwriting for Fannie and Freddie, but there is much more manual activity in non QM,” said Knochel. “Investors in the non QM space have different appetites. Some will say they’re comfortable with six months of bank statements. For others, it’s more or less.”

Fragmentation has been a challenge for the technology, said Elan Amir, CEO of MeasureOne, a platform that provides access to consumer-permissioned data, like information from bank accounts, and is working to consolidate the processing in more efficient ways.

“Right now, there are only solutions that verify income and employment data alone, so looking to scale would require lenders to work with multiple vendors and in the long run could make digital underwriting as tedious, complex, and fractured as the existing manual process,” Amir said.

Also, while both GSE and non QM market segments have been increasingly comfortable with using digitized bank data to verify income and even validate employment in some cases, some caution remains around relying on them as more of a primary source of underwriting information.

“Improvements on those tools are constantly evolving to be able to capture income a little more accurately for users, but it really takes a lot of in-depth tech to understand what the data being read is and how you parse that out,” said Josh Hager, head of mortgage operations at Button Finance. “Sometimes even direct deposits may not accurately reflect the employer’s name.”

A closed-end home equity specialist like Button rarely has to size up income and employment unless a borrower needs to, for example, count on the cash involved to close a primary mortgage; so providers of these loans may not yet have the kind of large-scale efficiency incentives to use the technology.

However, the GSEs have reported broad efficiency gains. Digital verifications of assets and incomes were reportedly shaving 15 days off loan processing cycles and reducing costs by as much as 30% as early as 2020, according to Freddie Mac. It’s further automated the process since then, extending the use of digital bank data to things like new rent-based underwriting.