In total, 80% of respondents favor manual underwriting over automating as the way to restore confidence. However, manual underwriting is fraught with error as well. In actuality both the human and the computer need to work together to restore confidence. I’m not sure the human can do it alone, but I don’t think the computer can either.
“Loan quality and automation go hand in hand but when you get down to the core of each loan you find a fundamental element: the automated underwriting findings,” said Kevin Marconi, chief operating officer at United Fidelity Funding. “Anyone in the originating or production process now relies on the algorithms of a Web-based computer to let them know if their loan complies with agency standards.
“But when it comes to agency standards (Fannie, Freddie, FHA and USDA) we now find that they are behind the times. In reality agency standards are trumped by aggregator guidelines and overlays. Therefore we find ourselves underwriting to aggregator guidelines not just agency guidelines. This invalidates the agency findings and forces those in the origination and production channels to rely on the old reliable manual underwrite. Very few lenders, including many of the aggregators themselves, have an automated underwriting engine which incorporates the agency guidelines and lender overlays in to the algorithms to derive a ‘true’ approval: subject to validation. More likely the standard LP, DO/DU, GUS or FHA Scorecard is run and left to the discretion of the originator and underwriter to apply a tangled web of overlays to the loan scenario. This leaves borrowers and originators guessing until the very end.”
Bruce Backer, president of LoanSifter agrees that you can’t successfully underwrite a loan without some kind of automation. “Better underwriting standards and automating what has traditionally been a paper-driven process are both critical to loan quality. However, when we talk about perfecting loan quality, this clearly cannot happen without automation. Human involvement will remain a constant in our industry, but even the most dedicated is susceptible to occasional errors simply due to the sheer complexity of the mortgage transaction. On the other hand, automation—which can provide lenders with instant access to accurate pricing and automatically match each file with the applicable disclosures, guidelines and state regulations—can do much more, at far less cost.”
But can automation truly improve the process? “When does approve mean approved in a paper world? In conclusion to perfect loan quality, automated underwriting systems, which reflect true market standards need to be developed and enhanced to once again take the subjective ness out of underwriting once again,” concluded Marconi. “We need to be able to rely on the validity of our AUS findings.”
