A major credit union serving military members and veterans rejected more than half of its Black conventional mortgage applicants
A major credit union serving military members and veterans rejected more than half of its Black conventional mortgage applicants
I noticed the article indicated that the analysis didn’t include credit score, available cash deposit, and relationship history with the lender. Honestly, it is usually credit score and cash on hand that gets you approved to receive a loan, debt to income ratio to determine maximum loan amount, and property value to pass underwriting requirements.
CNN’s response is basically, we don’t have the credit scores because it isn’t in the public data, and sort of glosses over it from there.
Now, I’m all for rooting out implicit and systematic bias, especially with systems that have profound effects on life trajectory, and I think credit scores are particularly prone to disadvantage minorities, but I feel like CNN really could have done better here. I know they asked the bank for credit score data, but the bank can’t release that information to a third party like that. I would have liked to see CNN being in some more expertise on the subject.
I’m contrast both NPR and NYT have done much more thorough investigative journalism on these topics. NPR in particular did a great podcast on how black home owners were routinely having their homes appraised for markedly less than white peers, and NYT covered discrepancies in underwriting outcomes even when credit history was the same.
I think CNN saw the easy article against this particular bank and took it to fill column space.
“Notably, Navy Federal approved a slightly higher percentage of applications from White borrowers making less than $62,000 a year than it did of Black borrowers making $140,000 or more.”
I don’t know man. Like, I agree that data is missing, but the data they did pull is very damning. Like how is there a higher approval % of white applicants making less than half the yearly salary of black applicants?
That’s a good point, though the whole thing is weird. I went back and reread the article and this jumped out at me as well:
It feels like they have some biased criteria in their approval process, but would really need the data to show it. They originate a higher percentage of conventional loans to black borrowers, but possibly as an artifact of having more apply.
One possible reason for the discrepancy with approval rates and income could also be cost of living areas. It would make sense that higher salaries also correlate to more expensive areas.
It would be nice if they shared their statistical results that they talk about being embedded in their processes to prevent bias so we could see what they are keying off of.
The poorer people are likely applying for smaller loans so the income to loan ratio could be better.
The fuck? I don’t have relationships with blood sucking financial parasites. Who here is playing golf on the weekend with bank CEOs?
Credit unions in particular still give some weight to how long you have been a member. Banks don’t so much, and even with credit unions it only goes so far.
The article is discussing the Navy credit union.
I bet there is zero science backing this up. Just someone some moron in a suit thought would sound smart.
Nope. There is zero evidence that a long-standing, mutually-beneficial financial relationship works out.
FFS, of course working with people over time is beneficial. Of course, both parties view one another more favorably, and act on that favoritism. You’re like some kind of financial incel.
that’s not what’s meant. they mean, how long you’ve had an account with them, whether you have multiple accounts or loans with the institution, if you’ve been late in paying or carried very minimal balances or have a history of harassing customer service to the point customer service felt the need to record it.
it’s your relationship with the institution, not the ceo, and whether you’ve been a good customer or not really.
Been with my credit union for 15-years or so. We have a solid relationship. If I need a thing, bet they stand up for me. No one has as clue what the other looks like, but here we are.
Had 3 car loans, couple of personal loans, saving and checking accounts, and a credit card with them. They helped me with financial needs, I repaid as agreed. It’s been a mutually beneficial relationship.
If your take on business relationships is built on schmoozing, you’re gonna fail every time.
Also, how does that compare with other financial institutions?
It’s a well known fact that black people are more likely to live in poverty or to at least be raised in it, that’s not very good to build a credit score!
Also, mortgages are approved by computers, not humans, the computer doesn’t know your skin colour, it knows your financial situation.
It’s the widest but what’s the rate for the other institutions? Is 29% the worst but the best is only a 4% difference or is it a 20% difference? I’m sure you know exactly what I’m saying, I just find it frustrating that it’s always the best or the worst that gets talked about when it’s the whole portrait that’s important.
Yeah. The concern I have with these things is when a constellation of metrics is assembled for the purpose of decision making, each data element on its own innocuous seeming, but together they manage to encourage training bias. Which is part of the concern around machine learning models being used to make real world decisions.
In this case there may well be a structural issue with the navy credit union, but I don’t think CNN showed that as well as they think.