Researcher warns of risks with using alternative data in lending:
Traditional credit scoring is based on a person's demonstrated ability to take on debt and pay it off. But with the dawn of larger data pools and access to more sophisticated modeling programs, lenders and credit agencies are taking more nonfinancial factors into rating creditworthiness, particularly those without an extensive credit history. This group tends to include vulnerable populations who are often more susceptible to predatory lending practices.
The problem is the systems developing these alternative scores can be like a black box, according to University of Georgia financial regulation researcher Lindsay Sain Jones. With the pool of personal data available growing, Jones argues that it's time to take a second look at how the American credit scoring system works and is regulated.
[...] In their recent paper, Jones and her co-author argue further regulation of financial reporting entities — both large credit bureaus and new data collectors — is needed in the same way gas, electric and water providers regulated their services. They argue participation in the credit system has become as necessary as having a phone or electricity.
[...] Jones and her co-author are also concerned that much of the lifestyle-related data points lenders correlate with creditworthiness can connect to race, gender, age, socioeconomic status, a person's ZIP code or where they attended college. Successfully challenging this kind of disparate impact under the ECOA [Ed: Equal Credit Opportunity Act] is nearly impossible.
One agency pulled information on how often people pay for gas at the pump versus paying inside the store. People who paid at the pump were deemed more creditworthy.
"There are all kinds of factors that can be correlated with creditworthiness, but that doesn't mean they should be used," Jones said.
When they factor in the web sites that people visit, do you suppose SN would be an asset or liability towards creditworthiness?
[ed note: See also Black Mirror, Season 3 Episode 1, "Nosedive". - fnord]
Janine S. Hiller and Lindsay Sain Jones, Who's Keeping Score?: Oversight of Changing Consumer Credit Infrastructure [open], Am. Bus. Law J., 2022
(Score: 0) by Anonymous Coward on Friday May 20 2022, @02:44AM (1 child)
Umm... "Creditworthiness" has very little to do with one's ability to do math, especially if it just involves saving a few cents here and there. I've known quite a few math and hard science (i.e., math competent) instructors who lived paycheck-to-paycheck because they couldn't manage money well.
When push comes to shove, what creditors are interested in is, "If we give you credit, will you be able to pay it back in a reasonable fashion? Or will you go on a 'financial bender' and dig yourself into bankruptcy and not pay us our money back??"
Lots of people who are bad at managing money tend to deal in cash, because they can't spend more than they have. Dealing primarily in cash doesn't demonstrate anything to creditors. In fact, using cash often means you are UNABLE to get credit, perhaps due to previous bad use of credit. Which is probably why people who pay for gas with cash generally correlates with worse credit risks.
What does demonstrate ability to manage credit is USING CREDIT. Then responsibly paying it back.
You don't need to have a billion credit cards or use them at the gas pump to have a great credit score. You don't even really need to use credit cards very often at all (as long as they're not so inactive that they get canceled). Take out a loan and pay installments back regularly. Or pay off your credit cards every month and get some points (which in many cases can easily be advantageous over "cash price"). You build up credit trustworthiness by actually USING CREDIT (surprise!) and demonstrating you can do that in a responsible manner.
If gas is a losing proposition for the use of a credit card for you (or in your state or whatever), then don't do it. But in general (maybe not for you, but on average), I can definitely understand how the type of person who pays cash always at the gas station is more likely to be the type of person who doesn't have good credit to begin with -- i.e., someone who may not have the option to use a credit card.
(Score: 4, Interesting) by PiMuNu on Friday May 20 2022, @02:28PM
This is an interesting case where naive statistical models break down.
The "trustworthiness" is binodal - there are some people who don't use credit cards because they always have money in the bank and simply don't need credit at all; there are some people who don't use credit cards because they can't cope with regular payments. And then there is the morass of folks between.
Sucks if the "good at finance" crowd get hammered when they finally want a mortgage or new bank account.
Chuck out your gaussians! A case for multivariate and non linear analysis (AKA Artificial Stupidity).