Happy Friday, Fintech Takers!

I have a pile of quick updates to share with you before we get to today’s essay.

1. You will notice that the “Read in Browser” button that is normally below the fold is not there anymore. That is because, starting today, we won’t be publishing most of my newsletters on the Fintech Takes website. The reason for this is simple: AI. The developers of large language models are rigorous harvesters of all human-written content on the internet (including my content) and, as a recent Cloudflare blog post makes clear, content publishers are getting the short end of the stick. So, for the time being, you’ll have to either read Fintech Takes in your inbox (as you are doing right now) or in the Library of the Fintech Takes Network (if you’re not yet a member, please apply to join!)
2. Speaking of the Fintech Takes Network, we are throwing open the doors next week. Our monthly virtual event — Fintech Office Hours — is normally only available to Plus members of the Network, but this month we are turning it into an Open House. All are welcome, so please register. These events are always fun and educational!
3. At the bottom of today’s email, you will see a section with the title “Input Requested”. If you have ever worked for a consumer lender that has relied on a credit decision engine and had a part in selecting, using, or maintaining that engine, I have some questions for you!
4. I am planning a very neat in-person event for fintech nerds in Washington D.C. later this year. I’m looking for 1-2 sponsors who would be interested in helping me bring it to life (and getting lots of face time with fintech operators, founders, and policy folks in D.C.) If that sounds like you, please reply to this email and I’ll share some additional details.
   

OK, phew! That’s enough of that. Let’s get into the fun stuff!

Martin Kleinbard, my fintech friend, long-time consumer lending operator, former CFPB staffer, and author of a terrific new research report, wrote a guest essay for Fintech Takes on what Bill Pulte has been up to over the last couple of months.

(Related — The "It's Just Math. Predictive Math." t-shirts are being made. It's happening. Let me know if you want me to share the Shopify link with you when they're ready!)

If you are interested in credit scoring, mortgage lending, cash flow underwriting, or the future of consumer lending, you will want to carve out time for it!

- Alex

---

Pulte Should Push For Cash Flow UnderwritingCopy anchor linkCopied

By Martin KleinbardCopy anchor linkCopied

Over the past decade, the Fair Isaac Corporation (better known as FICO) has experienced six single-day share price declines of 7.5 percent or greater. Four of those have come in the past eight weeks.

The root cause of all four recent dips was the Twitter/X account of Bill Pulte, Director of the Federal Housing Finance Agency, which oversees the government-sponsored enterprises that dictate the terms of the multi-trillion-dollar mortgage market.

In late May, he twice tweeted his disappointment with FICO’s recent steep price increases for soft pulls used for mortgage underwriting. Then, on July 8, he dropped the big bomb:

“Effective today, to increase competition to the Credit Score Ecosystem and consistent with President Trump’s landslide mandate to lower costs, Fannie and Freddie will ALLOW lenders to use Vantage 4.0 Score.”

And just like that, FICO’s run as the de facto sole permissible credit score in the conforming mortgage market was over, just days shy of exactly thirty years after it began.

The specific timing was likely a coincidence, but it provided a poetic bookend to a longstanding state-entrenched moat for Fair Isaac and its eponymous credit score. It also opened the door for a much-needed discussion about how to safely expand credit inclusiveness in mortgage underwriting.

Moses’ one discipleCopy anchor linkCopied

In July 1995, Freddie Mac, the little brother GSE to big sister Fannie Mae, published a letter to “CEOs and Credit Officers of all Freddie Mac Sellers and Servicers.” It directed mortgage lenders to supplement their underwriting policy with a wonky-sounding model score that was supposedly really good at predicting a borrower’s likelihood of default. Almost overnight, the FICO Score became a gatekeeper to the American Dream.

“To banks, mortgage bankers, thrifts and other firms with whom it does business,” the longtime real estate columnist Kenneth Harney wrote days after the letter went out, “Freddie Mac’s words about what it will and will not buy carry almost Moses-like force. And last week the words went out from Freddie: We strongly recommend that henceforth anyone who wants to sell loans to us use credit scores to evaluate consumers’ applications.”  

Fannie would follow suit a few months later, cementing FICO’s hold on not only the mortgage market but all other consumer credit markets that took their creditworthiness cues from the GSEs. By 1999, bankers had already confirmed FICO’s status as the “universal grade” and “common currency” that guided both direct lending and securitizations. It was an honor that Fair Isaac would keep all to itself until this past week.

“Anything to help the consumer”Copy anchor linkCopied

Pulte’s July 8 announcement confirmed that there is now a second common mortgage currency alongside FICO: VantageScore 4.0, the latest model iteration from the credit scoring company jointly owned by the Big Three credit bureaus.

Technically, FICO lost its “only show in town” title back in 2022, when the FHFA announced that Fannie and Freddie would eventually phase in Vantage. The mandate was supposed to go into effect in the fourth quarter of 2025, so from a nuts-and-bolts implementation perspective, Pulte’s “effective today” tweet only moved up the timeline by a few months.

But there’s a lot more to this story than nuts and bolts. Pulte’s subsequent tweets from that same day implied a desire not just to increase market competition and resilience, but to expand the pool of eligible homebuyers. 

“We will be incentivizing lenders, who USE both Vantage 4.0 and FICO, with better pricing,” he wrote a few hours after the initial pronouncement. “[A]nything to help the consumer.”

Reading tweet tea leaves can be a tricky business, but I’m pretty sure I know what Pulte is looking to accomplish here. He knows that his agency’s 2022 bulletin only required lenders to pull and deliver VantageScores alongside FICO Scores, without any mandated change in credit policy. He wants to use every lever at his disposal to push lenders to actually underwrite applicants based on both scores. 

More specifically, I think he wants lenders to put more weight on the higher of the two scores, especially for applicants with a higher VantageScore. If consumers get two cracks at a “yes” and two cracks at a prime or super-prime interest rate, homeownership rates will go up and borrowing costs will go down. 

“If you [the mortgage lender] use Vantage and not just FICO,” he followed up, “for the betterment of the American people and the consumer, you should get better pricing. It’s just math. Predictive math.”

Less than meets the eyeCopy anchor linkCopied

There are reasons to be skeptical of this plan, but there’s also clearly something alluring about giving consumers as many opportunities as possible to become homeowners. In a country that loves second chances, this would literally be a federally-mandated second chance. 

The unfortunate reality is that, for the vast majority of potential borrowers, it’ll be closer to two draws of the same card from the deck.

Fair Isaac and Vantage may be competitors, but their proprietary model scores are highly correlated because their underlying data sets are nearly identical. Recent press releases from Vantage that tease “$1 Trillion in Incremental Mortgage Activity” likely grossly overstate its additive value. The benefits touted in its newest, FHFA-approved model version (VantageScore 4.0) — trended data, rent reporting — are mostly mirrored by FICO’s own newly FHFA-approved version (FICO Score 10 T).

That last point about furnishing rental payments — which Pulte referenced on several occasions in recent tweets — is worth a deeper dive. It sounds like a big deal, but there’s no proof that just this change will move the needle for most would-be mortgage borrowers. 

For starters, the recent Urban Institute study that Vantage itself pointed to as evidence of rent reporting’s positive scoring impacts only found material effects on moving consumers from unscored to subprime or near-prime. The study, which consisted of a mere 141 consumers in the treatment group, did “not find that rent reporting has a statistically significant impact on the likelihood of having a ‘prime’ credit score or better” — the score range that most borrowers would need for a conforming, modestly priced mortgage. 

In other words, rent reporting on its own may help you get a high-APR personal loan or credit card, but it probably won’t help you get a home with affordable monthly payments.

Even if we assume the rosier credit inclusiveness conclusions from that minuscule sample size hold for the broader population, they would only apply to renters whose landlords agree to furnish the repayment data to the credit bureaus. The most widely cited data point for the share of renters who are receiving that benefit — also courtesy of the Urban Institute — is just five percent.

The real dealCopy anchor linkCopied

If Pulte is really serious about giving applicants as many bites at the mortgage apple as possible, he’ll take a long look at a new data source that is much less correlated with FICO and inherently geared toward giving the fairest shake to the “millions of forgotten Americans” (his words) who don’t heavily avail themselves of the mainstream credit markets. 

I’m talking, of course, about cash flow underwriting, which allows lenders to instantaneously and programmatically evaluate an applicant based on the daily inflows and outflows of their bank account. 

No less an authority than Larry Rosenberger, one of the original creators of FICO, the score, and a longtime CEO of FICO, the company, recently referred to cash flow data as “gold.” It’s set up to find creditworthiness nuggets out of exactly the types of behaviors that Pulte was referencing that don’t usually make their way to a traditional credit report, such as payments for rent, utilities, and small-dollar loans. And unlike FICO and Vantage, which rely on the better angels of landlords’ and electrical companies’ nature to furnish payment information in a timely and accurate manner to the credit bureaus, cash flow scorers have access to that raw payment information every month.

So what’s the next step? I wouldn’t expect or want Pulte, Priscilla Almodovar (CEO of Fannie Mae), or Mike Hutchins (CEO of Freddie Mac) to offer a blanket “yes” to a cash flow score to sit alongside FICO or Vantage, sight unseen. 

Their ideal response would be, “Prove it.” 

That’s exactly what their predecessors did to get us to this point. Freddie’s initial 1995 “Moses” letter referenced a study of “how hundreds of thousands of loans performed over several years” to determine FICO’s suitability for mortgage underwriting. The FHFA’s decision to bring in Vantage was the result of a Joint Credit Score Solicitation that began in 2020. 

Cash flow model scores should have to follow the same set of rules. They’ll need to earn their place at the table by demonstrating best-in-class default predictiveness that can compete head-to-head with FICO and Vantage and add orthogonal value to expand desirable credit opportunities. In doing so, they’ll have to adhere to the requirements for model predictiveness and safety and soundness as laid out in the Economic Growth, Regulatory Relief, and Consumer Protection Act in 2018 and subsequent FHFA regulations. 

If Pulte wants responsible mortgage access expansion to be part of his legacy, he needs to get the ball rolling now by directing Fannie and Freddie to commission a cash-flow-specific Credit Score Solicitation. They can more or less reuse the blueprint from the 2020 solicitation, with just a few additional requirements on the underlying cash flow data sources. All the evaluative parts can be left virtually unchanged. The same methodologies and metrics that determine the efficacy of a credit-bureau-based model apply to a cash-flow-based model. Ditto for the legal framework, where all relevant requirements are technology-agnostic. The big “lift” will be on behalf of the score providers, not the FHFA or the GSEs.

Make no mistake about it: if Pulte builds it, they will come. Several high-quality companies are already producing highly predictive, regulatorily compliant cash flow scores, and I’m sure would love the opportunity to compete for FHFA approval. 

Think about what that potential future state of the mortgage market would look like. Three battle-tested scores, including one from a completely new data source, would give Americans more chances at that prime mortgage rate than two scores from the same data source. The additional competition would also help to bring down underwriting costs and increase systemic resilience against legacy score degradation.

It’s just math.

---

MORE QUESTIONS TO PONDER TOGETHER

Big news for the endlessly curious (yes, you): I’m collecting your fintech questions on a rolling basis. 

What’s keeping you up at night? What great mysteries in financial services beg to be unraveled? Think of it this way, if a stranger is a friend you just haven't met yet, your question is a Fintech Takes conversation waiting to happen. 

One that could headline a Friday newsletter or be answered in an upcoming Fintech Office Hours event.

Drop your question here, whenever inspiration strikes!

---

INPUT REQUESTED

I’m working with Dilly Labs and the wise and powerful Tom Johnson on a little research project and I am looking for folks who work at a consumer lending company (bank, credit union, non-bank lenders, etc.) and who have experience buying/implementing/working with credit decision engines to fill out a quick survey.

Credit Decision Engine Vendor Survey

I promise it won’t take long! And it will be extremely helpful! So …

(Editor’s Note — If you work at a technology company that sells a credit decision engine, feel free to pass the link to the survey on to your clients. That’s perfectly fine. However, if you attempt to fill out the survey yourself, a terrible curse will befall you and your company. This isn’t a joke. The curse is real.)

---

Thanks for the read! Let me know what you thought by replying back to this email.

— Alex