Still Skeptical of FinTech
Andy Kessler interviews Mike Cagney, founder of SoFi, which, Kessler reports, found its initial success lending money to Stanford graduate students:
Digging for good credit risks worked at Stanford, and Mr. Cagney soon realized that he could apply the same concept elsewhere. Instead of relying on notoriously inaccurate backward-looking FICO scores, SoFi is “forward-looking.” That means asking basic questions—“Do you make more money than you spend?”—and calibrating where applicants went to college, how long they’ve been employed, how stable their income is likely to be over time. [Emph. added]
FICO scores are “notoriously inaccurate”? This is news to me. Rather, I thought the problem with FICO scores last cycle wasn’t that they didn’t work, but rather that too many banks lent too much money to borrowers that had low ones. For that matter, credit scores are just one of many inputs lenders take into account in the underwriting process. Anyway, if Cagney really believes he’s developed superior underwriting metrics, such as where an applicant went to college, one wonders why the FICO people haven’t thought of them already, too, and folded them into their algorithms.
I’ve long been skeptical of the bank-disintermediating potential that FinTech companies are said to have. But I’ve also seen enough other industries become semi-obsolete via similar-looking innovations that I’ll keep my mind somewhat open. Still, even Kessler seems to have some doubts. It’s hard to lose money lending to Stanford grad students, after all:
But isn’t SoFi cherry-picking loans? Absolutely. Why can’t banks do this? Because if you use depositor money for loans, as all banks do, you fall under the jurisdiction of the Federal Deposit Insurance Corp. and the Community Reinvestment Act, which bans discriminatory credit practices against low-income areas, known as redlining. In exchange, banks use leverage, but that’s courting trouble.
SoFi doesn’t take deposits, so it’s FDIC-free. . . .
SoFi and its ilk may not fall under the regulatory oversight of the FDIC and OCC, but as Kessler notes later on, they are under the thumb of the CFPB, which hasn’t been shy about, say, using “disparate impact” as evidence of racial discrimination. Which is to say, the agency isn’t especially subtle in its approach. So soon enough, the SoFis of the world may have their fair share of regulatory and compliance—and, once the cycle turns, credit—headaches. Maybe I shouldn’t be keeping an open mind after all.
What do you think? Let me know!
4 Responses to “Still Skeptical of FinTech”
Totally spot on. Yes, technologies and industries evolve. And there’s always a new set of smart people popping up to tell bankers how backwards they are. Fintech companies have a vested interest in creating widespread panic about being “left behind.” But never ask a barber if you need a haircut.
The strong point in this case is that they have set out a contemporary framework for ” eligibility”, a politically incorrect concept that disappeared from comercial Banks in the early 2000s.
FICO scores are as scammed as ratings by the agencies that are still around after 2008 for no economic justification. Try correcting reported delinquencies on outsourced, unsolicited merchant credit cards!
I would encourage these initiatives from the non-Banks since the TBTF contingent is hunkered behind abusive ATM and card delinquency fees..
I think you have this one ” just right ” Tom.
To be fair, it’s not that FICO scores are inaccurate. It’s that they may not be representative of the TRUE credit risk of various borrowers. Especially in areas where there is significant collateral against the loan (think auto, mortgage, etc.). The reason for this is very simple. If we both have a 750 FICO and one of us puts $10k down on a car loan and the LTV is 62% and the other puts no money down and the LTV is say, 105%, they present uniquely different risks to a particular bank. This isn’t news to anyone. That’s the whole reason a vast majority of larger banks invest in what’s know as internal score or custom score. This allows them to understand the unique buying habits of their customer base and product set to get a better judge of the true credit risk of the individual. It’s great when we have companies looking at unique elements to underwrite credit. Show me how this data holds up over say 2-3, 3-4 credit cycles and I’m all in…..These tend to be very product specific and your not inventing a new way to look at credit. FICO isn’t perfect for all products, all the time and they are always looking for additional ways to make the score more predictable. At the end of the day, you make it more predictable by customizing it to the unique characteristics of the product.
Comments are closed.