How Some Algorithm Lending Programs Discriminate Against Minorities
SCOTT SIMON, HOST:
Racial profiling by banks against African-Americans, Latinos and other minorities who are trying to qualify for the lowest priced home or auto loans has been a fact for decades. But shouldn't the algorithms of computer-driven lending programs be blind to ethnicity? Not quite, according to a new study. Washington Post personal finance columnist Michelle Singletary has been looking into this and joins us. Michelle, thanks so much for being with us.
MICHELLE SINGLETARY: Oh, it's my pleasure.
SIMON: Now, there's a new study from UC Berkeley that says these computer-driven programs somehow manage to discriminate.
SINGLETARY: That is correct. They found that both online and face-to-face lenders charge higher interest rates to African-American and Latino borrowers. And in this study, they were looking at, OK, face to face, lenders can't discriminate based on, you know, race. But you'd think with the machines, they don't know what color I am. Well, they figured out that some of the data that was being input into the algorithms may in fact create some discriminatory situations.
SIMON: I mean, how does that work out? How does that happen?
SINGLETARY: So what they were figuring is that minority borrowers don't tend to shop around as much for a whole bunch of reasons, but they don't. And so the machine was thinking, OK, so these folks are not going to shop around. We can price them higher because they're not going to know that they're paying more than they need to pay. And they found that that was happening more often with African-American and Latino borrowers. And this was again holding cost and for credit history and other factors that would make people credit worthy.
SIMON: So, I mean, not to split hairs, are they discriminating because somebody is African-American or Latino or some other minority or because in the judgment of the algorithm, they're betraying (ph) certain traits that leads them to conclude that?
SINGLETARY: It's unintentional, we hope, but it's discrimination nonetheless. And, again, the intent might not have been there, but let's just say you are offering higher rates to people who live in a certain zip code, and we know that the people who mostly live in this zip code are African-American, well, that's not right.
SIMON: What should lenders do, or legislators, for that matter?
SINGLETARY: Well, certainly legislators should look at the input of this and make sure that the result is not discriminatory. From a consumer's point of view, if you know that this is going on - and hopefully you know now because you've been reading my column and you know about this Berkeley study, which is fascinating - now you know that if you are good with your credit, you can get the best rates out there. And that means shopping around. One study showed that if you shop around with one or two lenders or three, you can save thousands of dollars on your mortgage. And if you know that you've got a good credit score, that puts you in the driver's seat.
SIMON: But if a lender says, well, you're welcome to shop around, that's not very friendly, is it? That's not solving the problem.
SINGLETARY: You know, it's not, and it sometimes intimidates people or sometimes, especially with minority borrowers because, let's just be honest here, there was a lot of red money and a lot of discrimination, and many people may not think that they're worthy of a better rate. And so you tend to go with what you know, but knowing that there's much more competition out there, don't sell yourself short when it comes to mortgages and lending. You know, shopping around can save you a lot of money. And it's really worth feeling uncomfortable by asking other people what they would offer if it's going to save you some money.
SIMON: Michelle Singletary writes The Washington Post's nationally syndicated personal finance column, "The Color Of Money." Thanks so much for being with us, Michelle.
SINGLETARY: You're welcome.
NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.