"Fintech" Borrowers More Likely to Default?

Figure Files for a Charter, BAML Gets "No Action", Beam Blows Up

Hey all, Jason here.

Well, it’s finally over, and Joe Biden will become President on January 20, 2021. Maybe some VC dollars need to flow to ‘votetech’ (and ‘pollingtech’, for that matter).

Biden has said little specifically about fintech, but two safe bets are: a beefed up CFPB, and that policy making will be viewed through the prism of coronavirus and its economic impacts. I wrote about this in more detail a couple weeks ago.

Despite being glued to the NBC News livestream since Tuesday, I’ve managed to put together this week’s issue. Here goes.

Don’t Subscribe Yet?

Deep Dive: “Fintech” vs. Traditional Lender Analysis

This paper from Harvard Business School’s Marco Di Maggo and Georgia State University’s Vincent Yao compares the performance of similar personal loan borrowers from “fintech” vs. “traditional” lenders.

The paper is full of interesting data, and while I concur with much of the analysis, there are some problematic arguments. Part of the problem is in the frame itself - eg, the distinction between “fintech” and “traditional” lender is rapidly losing significance.

The paper defines fintech lenders as:

“…those who operate exclusively online and do not have a brick and mortar presence, do not accept deposits, and are not regulated by the Federal Reserve or the Office of the Comptroller of the Currency (OCC). We also require them to be recently founded.”

And names the largest fintech lenders included in its data set as: Lending Club, GreenSky, SoFI, Avant, LoanDepot, Upstart, and CashCall. This is logical enough, but I question if these attributes are responsible for the differences in borrower behavior and outcomes described.

Further, as “fintechs” increasingly seek to acquire charters (Lending Club/Radius, Jiko, SoFI, Varo, Figure), this delineation between “fintech” and “traditional” will no longer hold.

The full paper is 73 pages, so let me break down the most interesting parts (with some critiques) to save you the time.

Fintech vs. Traditional Lenders: Differences in Borrower Behavior and Outcomes

As part of establishing its definition of “fintech”, the paper argues that

Fintech lenders are “less regulated” than traditional lenders, which may allow them to adopt laxer lending standards

This oft-repeated notion is just not accurate. Lending-related regulation (ECOA, FCRA, TILA, FDCPA, UDAAP) applies equally to fintech as to traditional lenders.

While the fintech lenders in the study aren’t directly overseen by federal regulators (FRB, OCC, or FDIC), most (if not all) originate loans via partner banks (usually WebBank or Cross River Bank) that have state and federal regulators.

Further, it’s unclear to me how having or not having a direct prudential federal regulator would influence a lender’s underwriting standards.

Borrowers from fintech lenders are more likely to default than similar borrowers from traditional lenders. This is driven in part by an increased likelihood of fintech borrowers total indebtedness and revolving balances increasing post-origination vs. traditional (including an increased likelihood financing a vehicle). The result is that fintech borrowers are more likely to become over extended and default.

This consumer use pattern - using an unsecured personal loan to consolidate outstanding revolving debt, but then re-accruing revolving debt - is not surprising to me, and is well known to anyone who has worked in the industry.

An interesting quirk of credit scoring is that users who consolidate revolving debt with a personal loan will see a jump in their credit score (as revolving utilization drops), even as their overall debt load remains the same or even increases (depending on the interest rate and term of loan, monthly payments may drop).

“…even when borrowers use the personal loan to consolidate their credit card debt, we find that their revolving balance increases again a few months after origination.”

That jump in score may enable borrowers to qualify for new credit (eg, car loans) or for revolving line increases, encouraging them to take on more debt. As overall debt load increases, borrowers become unable to keep up with their payments and begin to go delinquent.

While credit score initially increases post-origination, as borrowers re-accrue debt and delinquencies increase, credit score declines. Chart images from “Fintech Borrowers: Lax-Screening or Cream-Skimming?”

The aspect of this that surprised me is that, controlling for credit quality, “fintech” borrowers are more prone to this behavior and thus more likely to default than a borrower from a traditional lender.

The authors argue that this is because fintech lenders experience adverse selection and are more likely to rely on bureau data, and thus extend credit to someone who would be rejected by a traditional lender:

“The evidence on the underlying screening technology and the results on the borrowers’ delinquency suggest that fintech lenders are likely to be exposed to adverse selection. Specifically, due to the lack of soft information available to them, fintech lenders might end up giving credit to borrowers that would have been rejected by a traditional lender who might be able to detect weaker financial management skills.”

An element driving adverse selection may be the perceived ease and simplicity of fintech lenders vs. traditional:

“…the evidence points out that the increased ease and speed with which borrowers can have access to credit is particularly appealing to certain households who tend to use these funds, in conjunction with other forms of credit, to sustain their consumption, which ultimately makes them more financially vulnerable.”

However, the author’s own analysis of fintech lender pricing suggests to me that their underwriting compared to traditional lenders is no mistake, but rather part of an intentional customer acquisition strategy:

“…higher defaults are likely not to translate to lower profits for the lenders as fintech lenders seem to be better at pricing than traditional institutions.”

This, combined with my experience in the industry, suggests to me that fintech lenders intentionally targeted customers less likely to be approved by traditional lenders and priced this risk accordingly; not that fintech lenders faced adverse selection or a ‘soft information’ deficit that led to them mistakenly approving these borrowers.

Follow-Ups: New Developments on Prior Stories

Small-Dollar Lending: BAML Gets “No-Action”

CFPB issues “no-action letter” for Bank of America’s announced small-dollar loan, Balance Assist. Such a letter increases certainty that Bank of America won’t face a supervisory or enforcement action related to the credit product, and follows interagency guidance from regulators earlier this year encouraging more banks to offer small-dollar lending products.

Startups & User Trust: Beam Blows Up

I previously wrote about how fintechs benefit from user trust in the established financial and regulatory system. Beam, a startup that had been promising as high as 7% APY in an FDIC-insured savings account, is the latest and maybe most extreme case-in-point.

With users reporting that they are unable to withdraw money and an investigation launched by the FTC, there are more questions than answers. Dwolla, the app’s ACH processor, has cut ties; Beam’s banking partner, Huntington National, says it doesn’t have the money; and the company handling Beam’s sweep account, R&T, says the funds are “held in a demand deposit account at one, well-capitalized FDIC member bank.”

While users should eventually get their money back, this story is a cautionary tale, to both consumers and banks considering partnering with startups.

Charter Watch: SoFI Gets Conditional OCC Approval; Figure Files

The last big ‘charter watch’ story was Revolut (and just a rumor at that).

Now SoFI has won conditional approval from the OCC, and late-breaking news that Figure (also founded by Mike Cagney) has announced it has applied for a national bank charter (the announcement, interestingly, makes no mention of deposit products).

This is particularly interesting, given the challenges of VC-backed startups qualifying for a national charter without their investors facing bank holding company issues.

SoFI is the furthest along in diversifying from its original product (student loan refinance for prime borrowers) to looking, well, like a multi-product bank, offering mortgages, investments, and a checking-like account.

The real question to me becomes, what distinguishes a “fintech” like SoFI from existing online-only banks like Ally or a developing one like Marcus by Goldman Sachs?

And, can fintechs justify “tech” valuations when they so closely resemble publicly traded peers with much lower multiples?