In partnership with

Happy Monday, Fintech Takers!

I trust that you had a good weekend.

I did, although it remains ridiculously cold in MT.

However, none of us had as good a weekend as Kenya’s Sabastian Sawe, who became the first person to officially break the 2-hour barrier in the marathon. He finished yesterday’s marathon in London in an astonishing 1 hour, 59 minutes and 30 seconds.

Even more astonishing, the second-place finisher, Yomif Kejelcha of Ethiopia, also dipped under 2 hours by crossing the line in 1:59:41 in his first-ever marathon, while Jacob Kiplimo of Uganda broke the previous world-record time by seven seconds, finishing in 2:00:28.

Incredible.

We call it a “marathon” because of the famous run that Pheidippides (an Athenian running courier) made from Marathon to Athens (approximately 40 km / 25 miles) to deliver news of the Battle of Marathon, fought between Athens and Persia.

We don’t know how fast Pheidippides ran (probably not that fast, the legend is that he also fought in the battle and before that he had run 150 miles between Athens to Sparta to ask Sparta for help … he also died shortly after his run from Marathon). However, we do know that Spyridon Louis won the first Olympic marathon race (which was a slightly shorter distance back then) in 1896 with a time of 2 hours, 58 minutes, 50 seconds. Over the next century, humans shaved more than 55 minutes off that time, despite the race being standardized to a longer 42.195 km distance.

Until recently, it was thought that finishing a marathon in under two hours was physically impossible. Eliud Kipchoge, one of the greatest marathon runners of all time, spent his whole career trying to prove that wrong. He did it in 2019, but it didn’t officially count because he did it on a closed course with a rotating group of pace runners helping him.

And now, two guys (including one who was running his first ever marathon?!?) just sprinted (at a pace of about 4 minutes and 33 seconds per mile?!?) through the door that Kipchoge cracked open.

I’m not sure what the limits of human achievement are, but I am quite sure that we are nowhere close to reaching them. Totally... 1

- Claude Bot

#1: Keeping PACE

What happened?

Speaking of setting a fast pace, U.S. House Representatives Young Kim (Orange County) and Sam Liccardo (Silicon Valley) are trying to speed up payments for U.S. consumers and businesses. They just introduced the Payments Access and Consumer Efficiency (PACE) Act:

Legislation introduced in the U.S. House this week, known as the Payments Access and Consumer Efficiency Act (PACE Act), would establish a federal registration pathway for nonbank providers seeking direct access to the Federal Reserve’s core payment systems, including Fedwire, the FedNow Service and ACH.

The proposal lands as real-time capabilities and settlement certainty shift from competitive advantages to baseline expectations across the digital economy.

The PACE Act creates a new designation, “registered covered providers,” that would allow qualifying firms to apply for a “payments reserve account” at a Federal Reserve Bank, granting connectivity to Fedwire Funds Service, FedNow Service and FedACH Services, placing nonbank payment firms on more equal footing with insured depository institutions in terms of infrastructure access.

So what?

This isn’t exactly like the national fintech charter that the OCC tried to manifest between 2016 and 2018 or the national payments charter that the OCC floated, briefly, in 2020, but it’s similar.

The idea of carving out room for more novel entities (narrow banks, payment providers, etc.) to get direct access to the Federal Reserve’s payments infrastructure is not new. Novel, state-chartered banks and credit unions have been trying, with limited success, for more than a decade. More recently, the OCC has repurposed its national trust bank charter into a quasi-national payments charter, which various crypto and fintech companies have begun to pursue in order to get access to Fed master accounts. And the Fed itself has created a more limited version of its master account (a “skinny” master account … which Kiah Haslett and I discuss here) in order to open up wider access to its payments infrastructure, with Kraken (a Wyoming-chartered SPDI) getting the first one (which Kiah and I discuss here).

And now we have the PACE Act. It doesn’t have a companion bill in the Senate yet, but it does have enthusiastic support from the fintech and crypto lobbies. And, perhaps more importantly, it appears designed to avoid pissing off the Conference of State Bank Supervisors (CSBS), which is a trade association representing state banking regulators that vociferously opposed the OCC’s moves towards a national fintech charter and national payments charter (as I wrote about here).

Rather than preempting the states and their money transmitter licenses (MTLs) with a new limited-purpose national charter, the PACE Act would create a new class of “registered covered providers” that would be required to obtain MTLs from at least 40 states. They would also be required, for all outstanding payment obligations, to maintain reserves on a 1:1 basis, with reserves held in cash, deposits, short-term Treasuries or similarly liquid instruments (similar to a payment stablecoin issuer under GENIUS).

This is weird! And it leaves me with many questions, including:

• What does CSBS think of this proposal? As far as I can tell, they haven’t commented publicly yet. Perhaps they don’t want to dignify the bill with a response, given that it’s very early in its lifecycle and may not have the requisite support in Congress to go anywhere. If it did gain traction, is the requirement to hold at least 40 MTLs sufficient to mollify CSBS? Might CSBS want to get onboard with this idea, given that the alternative (which they despise) is the OCC’s national trust charter landgrab?

• And speaking of national trust banks, how would a theoretical PACE Act-regulated entity differ from a payments-focused national trust bank? Both would be primarily regulated by the OCC and be subject to strict requirements on the custody of funds and reserves. If the PACE Act were to pass, would it shunt the national trust bank charter back to the corner where it had been collecting dust?

• How would this bill — if it were passed and signed into law — impact the Fed’s strategy on master accounts? Skinny master accounts aren’t as popular with the fintech and crypto industries as you might expect because of their limitations (no FedACH, no daylight overdrafts, etc.) Many of them want “fat” master accounts and the PACE Act would seem to give them a different path to get there, given that “registered covered providers” under PACE would be considered less risky, under the Fed’s evaluation criteria, than novel state-chartered institutions like Kraken.

#2: FICO is Trying to Price Itself Out of a Tough Situation

What happened?

FICO had a tough week:

Shares of Fair Isaac Corp., known as FICO, sank on Wednesday, after the US government took steps that officials said were aimed at cutting costs for credit scores and making homeownership accessible to a wider swath of buyers.

The Federal Housing Administration and mortgage-finance giants Fannie Mae and Freddie Mac are implementing the first new credit score models for mortgages in decades, according to a news release from the Federal Housing Finance Agency. The models include factors like rental and utility payment history.

[FHFA Director Bill] Pulte said Fannie and Freddie will use a VantageScore 4.0 mortgage origination score and continue to work toward using a new, more advanced FICO score known as a “10T” that employs credit information and rental payment history to offer lenders a “more nuanced” look at borrower behavior, according to FICO. Lenders can also continue to use the Classic FICO scores the industry has long relied on.

So what?

This is the culmination of the plan that was set in motion with the passage of the Credit Score Competition Act in 2018, which required the FHFA to create a process for validating newer mortgage credit score models.

After 2018, VantageScore, a competitor to FICO that is owned by the big three credit bureaus, took steps to have its latest score (VantageScore 4.0) validated and approved by the FHFA.

During that same time, FICO took steps to have its latest score (FICO 10T) validated and approved by the FHFA as well, while also raising the price of FICO Classic for mortgage from roughly $0.50 per score in 2018 to $4.95 per score in late 2024.

It’s important to note that the credit bureaus also benefited enormously from these price hikes, as they resold FICO Classic at a markup that roughly doubled the effective price for lenders (e.g., the $4.95 price was actually about $10).

Last year, after Bill Pulte accelerated the endgame for the Credit Score Competition Act, FICO launched a direct license program, allowing mortgage lenders to access FICO Classic, without going through the bureaus, for $10 per score or $4.95 per score plus a $33 fee for each closed loan. Additionally, FICO announced that it would be directly offering FICO 10T for mortgage lenders for free if they continue buying FICO Classic or standalone for $0.99 per score plus a $65 fee for each closed loan.

This is tricky!

It anchors the price for FICO Classic at familiar levels for mortgage lenders: $10 if they want to buy it for a flat fee or $5 if they want to buy it and pay the backend performance fee. Additionally, the $0.99 per file fee for FICO 10T matches the aggressive per-score price that the credit bureaus are charging mortgage lenders for VantageScore 4.0, but, again, only if lenders are willing to pay FICO the backend performance fee.

What’s wild about this pricing model is that if mortgage lenders’ pull-to-close rate (the ratio of scores they pull to loans that they close and fund) is greater than roughly 15%, then the performance model for FICO Classic is a worse deal than the $10 flat-fee model. This is true for FICO 10T as well ($0.99 per score + $65 per closed loan averages out to more than $10 per score with a high enough pull-to-close ratio), except there is no $10 flat equivalent for FICO 10T. The company only offers 10T under its performance model (or for free when bundled with FICO Classic). This means that for many mortgage lenders, FICO is either incentivizing them to stay on FICO Classic or to pay more for their mortgage credit score than they were paying before!

Either Bill Pulte doesn’t understand this distinction, or he does and he is pressuring FICO to lower its price for 10T. Here’s Bloomberg again:

FICO said in a statement that it applauds the “initiative to get FICO Score 10T into the market,” adding that it “incorporates rental and utility payment history, enabling more consumers to qualify for mortgages.” That score is available for free with the purchase of FICO’s “classic” score; via FICO’s “Mortgage Direct License Program,” it costs 99 cents and a funding fee of $65, FICO said.

During the press conference, [Wolfe Research analyst Scott] Wurtzel said, “Pulte appeared to frame this as FICO lowering the $10 per-score royalty to 99 cents,” instead of offering 10T through a direct model at 99 cents plus the $65 fee.

Either way, not great for FICO!

#3: The Informed Minority

What happened?

A fascinating new academic paper on prediction markets was released:

Prediction markets are remarkably accurate, yet the source of this accuracy remains poorly understood. The conventional view attributes it to crowd wisdom, whereby prices aggregate information from a large and diverse pool of participants. We show instead that accuracy is driven by a small minority of informed traders. Using the universe of transactions from a large prediction market platform, we identify these traders and show that they, around 3% of all accounts, generate the bulk of price discovery. Their trades predict future prices and final outcomes, make prices more accurate throughout a market's lifespan, and react to news the moment it arrives. The remaining majority does not produce accuracy; rather, it funds it. Their trades generate most of the volume, but little of the information, and their losses flow as profits to the informed minority. Prediction market accuracy thus reflects the wisdom of an informed minority, not the wisdom of the crowd.

So what?

The value of prediction markets, we are repeatedly told, is that they are a highly accurate tool for predicting the future; civilization-scale truth-telling machines, as the venture capital class likes to say.

The predictions generated by prediction markets can be used to hedge specific economic risks (this is their purpose, as derivatives, under the Commodity Exchange Act), and, more broadly, to help people make better decisions in their day-to-day lives (Polymarket CEO Shayne Copland has described the use of war markets on Polymarket by people living in war zones and deciding whether or not to sleep near a bomb shelter as an “undeniable value proposition”).

That’s the output.

The input — the source of the prediction markets’ insights — is, we’re told, the “wisdom of crowds.”

This framing is important because it feels democratic. Everyone has some notion of what might happen in the future and if everyone trades on their convictions about the future with everyone else, the end result will be an accurate aggregate-level view of future events, expressed probabilistically, and a set of winners and losers that is fairly distributed.

However, the core insight from this paper, which analyzed transaction data from Polymarket, is that knowledge about the future is not distributed evenly. It is, in fact, highly asymmetrical.

This makes intuitive sense. Insiders know more about the industry or organization or event that they are inside of than even the smartest outside observers do. A market that efficiently generates accurate predictions about the future needs to incentivize insiders to bet on what they know. Shayne Coplan made this exact point in an interview with Axios’ Dan Primack, saying, “I think what’s cool about Polymarket is that it creates this financial incentive for people to go and divulge the information to the market.”

The challenge for the prediction markets is ensuring that the financial incentive for this “informed minority” is large enough to guarantee their participation. In order to do that, prediction markets — which, as they constantly remind us, are platforms for peer-to-peer trading — need to attract enough dumb, speculative money to take the other sides of these bets with the informed minority.

This is the root of my problem with prediction markets. In order to deliver the value they claim to provide, they need to attract the informed minority. In order to attract the informed minority, they first need to attract the uninformed majority. And in order to attract the uninformed majority, they need to offer bets on activities that a large number of people enjoy betting on recreationally, in the places where recreational gamblers already congregate, which is why sports betting is now available through Robinhood and Coinbase.

This may be an “undeniable value proposition” for the market overall, but it certainly doesn’t feel fair, in any sense of the word. And structurally, I’m not sure it ever can.

#1: Online Gambling Is Breaking Containment (by Charles Fain Lehman, The Dispatch) 📚

This is a smart and broadly-framed take that touches on some important long-term questions. Questions that I will be hashing out in more detail in future Fintech Takes podcast episodes.

Stay tuned!

#2: The $100B+ Missing Connectivity Layer (by Jason Mikula, Fintech Business Weekly) 📚

Come for Jason’s reporting on a batshit-crazy story involving the CEO of failed fintech infrastructure company Synapse. Stay for a fun guest essay by Knot Co-founder and CEO Rory O’Reilly.

There are a TON of interesting questions being asked in the Fintech Takes Network. I’ll share one question, sourced from the Network, each week. However, if you’d like to join the conversation, please apply to join the Fintech Takes Network.

What are the best arguments against disparate impact as a mechanism for monitoring and correcting for fair lending violations?

I got some helpful pushback on this subject after publishing last week’s newsletter and I’d love to dig in even more.

What are the problems with disparate impact? What would be a better way to regulate fair lending?

If you have any thoughts on this question, reply to this email or DM me in the Fintech Takes Network!

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

— Alex