“A lot of people have called non-QM ‘storied loans’, because there’s a story behind the borrower, but it’s hard to write code to know a borrower’s story.
“There’s technology that can read bank statements, but it’s harder when it comes to making a decision on whether or not a deposit is OK, and it’s a challenge to automate the entire process,” he said.
The problem was compounded because the focus of technology vendors “has always been on streamlining the agency process”.
He said: “It makes sense because it makes up 99% of the business, and if you’re a technology firm you’re not going to spend a lot of time and resources on non-QM, which to date makes a really small piece of their business. But those are just some of the challenges we face.”
He highlighted how the agency world had found it relatively easy to build the technology, since the process started with automation, thanks to the development by Fannie Mae and Freddie Mac of automated underwriting systems, Desktop Underwriter (DU) and Loan Product Advisor (LP).