The algorithm as foil is no mere abstraction to Kempfer: “I actually have a loan right now that’s not getting approved,” she said. “The credit score is OK – it’s not horrible; we’ve got 5% down. By all accounts, it should be approved. The only reason it’s not is we have down payment assistance and a thin credit file. These people are trying to keep their debt down; they don’t make a lot of money; and they have to buy a condo, which is a higher risk. So these people, in my mind, are getting redlined out, and they’re being discriminated against through an algorithm – your income is this, your LTV is too high, you’re buying a condo instead of a single-family home and you don’t have a lot of credit. So too bad – you’re not getting a house.”
And it’s not the first time she’s encountered the pitfalls of an automated underwriting system dampening the dream of homeownership with the cold dispassion of an algorithm-based assessment. “I had one almost identical to this, and it also was a Black borrower,” Kempfer noted. While the other loan she described remains in limbo, the earlier one did have a happy ending after considerable effort. “I had to fight tooth and nail to use the 12-month rental history,” she said. “No-one knew how to do it, so it took moving heaven and earth to get someone to even try to do it.”
Kempfer said it will be education that will save the day in the future. “I’ve talked to some of the top lenders – I won’t name them – and I’ve told them this needs to be in your job aid,” Kempfer said. “People need to have this top of mind.”
But in the end, the future is still fraught with questions: “Where is the disconnect between Freddie and Fannie and whomever is making these programs available and the lender? How do we educate underwriters and the account executives as well as mortgage professionals so that, at the end of the day, we are getting more people in homes?”
Comments are closed, but trackbacks and pingbacks are open.