Interchange Arbitrage as a business model

Marqeta & Chime both depend on IC arbitrage, but there is a downside

Key insights in this post

• The way Networks and Regulators set interchange creates arbitrage opportunities to shift volume between card types or between cards and non-card methods

• For some Fintechs, such arbitrage is key to their revenue models, for example:

  • Marqeta exploits both exempt debit IC and virtual cards

    • It provides Cash App with exempt debit interchange by using an exempt BIN-sponsor bank

    • It shares Virtual Card interchange with other Fintechs by intermediating between a consumer card IC rate and the ~2.5% virtual card IC rate

  • Chime exploits both exempt debit interchange and the gap between credit card IC and debit card IC

    • ~70% of Chime revenue is from exempt debit IC which Chime gets by splitting its volumes over two exempt BIN-sponsor banks

    • Chime’s Credit Builder secured credit card offers no Credit but manages to earn high credit card interchange

  • These are just two prominent examples, but many more Fintechs and incumbent payments companies capture similar interchange arbitrage

• These arbitrage opportunities are beginning to come under pressure

  • Some exempt BIN-sponsors are under FDIC consent orders to improve BSA compliance, which will be expensive relative to BIN-sponsor revenues

  • The merchant settlement may loosen or eliminate the Honor All Cards rule which would allow merchants to steer away from high interchange services

Introduction

Economics is about incentives, and payments are no exception. The way networks set interchange incents providers to substitute one payment method for another, where the target method earns higher revenue or has a lower cost.

Merchants optimize against these differences:

• Starbucks gets consumers to load a stored value account via card and then scan their app for purchase. That saves the fixed component of interchange X-1 times where X is the number of purchases against each value load

• Merchants work to change their MCC (Merchant Category Code) to one with a lower IC rate. This works where the merchant is selling a wide variety of merchandise

• I worked with a merchant that left each acquisition with their own legal entity. Had they consolidated under a single legal entity, their volume would have qualified for a lower interchange “tier”. Instead of actually consolidating, they negotiated with the networks to treat them as a single tier despite the legal structure. This saved a small fortune.

Merchant litigation against the networks is another front in this battle. We are now waiting for a final decision on a 10-year old lawsuit.

The government created a different set of incentives under the Durbin Amendment. Durbin capped Debit interchange (IC) for larger banks, creating an advantage for smaller, “exempt” banks; V/MC set Exempt interchange at a premium to Regulated IC. I have done projects for smaller banks on how to grow revenue while keeping assets under $10B to retain Exempt status. Durbin created an incentive to stay small.

Fintechs are masters at interchange arbitrage – at some, it accounts for the bulk of revenue. This post will discuss two prominent Fintechs and how they do this: Marqeta and Chime. Of course, “there ain’t no such thing as a free lunch” (TANSTAAFL), so this arbitrage has costs as well. I will note both the upside and the downside.

Marqeta

Marqeta is an innovative company with impressive growth, so the commentary below is not a knock on their business. Marqeta just illustrates the points I want to make. I could have done the same thing with many Fintechs.

Marqeta practices interchange arbitrage for both debit cards and virtual cards:

Debit: Exempt versus regulated debit interchange

Marqeta has been in the news recently as they blamed at least part of their Q3 results on throughput issues at their BIN-sponsor — Sutton bank. Sutton has $1.5B in assets, yet they sponsor the Cash App Card with 24M actives. This earned them ~$44M YTD (Q3) in fees while Net Interest Income was only ~$68M. Unfortunately for Marqeta, Sutton now faces an FDIC consent order to improve Bank Secrecy Act (BSA) compliance. Dealing with this no doubt created those growth bottlenecks. Many exempt BIN-sponsors are under similar FDIC scrutiny.

Why use such a small BIN Sponsor? Because the only way the Cash App Card can earn Exempt interchange is by using a BIN Sponsor with less than $10B in assets. Exempt interchange varies, but a typical transaction might earn 45¢; in contrast, the same transaction at a regulated bank earns the Durbin cap of ~25¢.

Cash App has gradually insourced some of the activities that Marqeta previously provided. In fact, Marqeta changed its financial reporting to reflect that Cash App does so much in-house. They used to report Cash App interchange as “Net Revenue” and then subtract all direct expenses as “Cost of Revenue” to derive “Gross Profit”. But now, Marqeta effectively reports just Gross Profit for Cash App revenue.

With that change, Net Revenue dropped by more than 50%, Cost of Revenue dropped by ~75% but Gross Profit dropped by only 18%. Quick math shows that, the Cash App contract had roughly a 10% gross margin. Cash App was 79% of net revenue before but is now under 50% of net revenue. Volumes are still comingled with other clients.

The challenge Marqeta faces is a common story in payments:

1. Any payments company processing large clients would experience a similar challenge: having big customers helps with scale, but squeezes margins. We used to call this the “winners curse” at JPM. This is not unique to Marqeta or Fintechs in general. If TSYS ever disclosed their economics for Capital One or Bank of America we would likely see the same phenomenon; likely the same for Fiserv with Synchrony and Citi Retail. Most of the large credit card issuers pay almost nothing to the networks; And Amazon & Walmart have the lowest acquiring, interchange & network fees.

2. Marqeta uses an exempt BIN-sponsor because their biggest clients need exempt interchange – and exempt banks are sub-scale by definition. The small BIN-sponsor then became a growth constraint as they need to scale compliance capabilities to keep pace with client growth. Regulators have been scrutinizing BIN-sponsors in general, and other small sponsors are in the same boat. Even Wells retreated from BIN-sponsorship due to the compliance risk

3. Cash App could insource so much functionality because debit card processing in general is a commodity. Marqeta bills itself as “the modern card issuer” and that is a true statement, but debit cards are hemmed in by the legacy ecosystem: network rules, Reg E, the Durbin amendment, the ISO 8583 messaging standard, the EMV standard and other constraints. It is hard to sustainably differentiate.

Based on “Cash App Actives” of 24M for Q3 24 and Marqeta Q3 Net Revenue from Cash App of $60M, Marqeta earns ~$2.50 per active card per quarter or $0.83 per month. Assuming 35 transactions per active card per month we get 2.4¢ per transaction. These are not windfalls. Even if the Cash App cards are less active, the net revenue is modest.

Ironically, Exempt interchange isn’t always as high as advertised. The Card networks discount Exempt interchange to get routing volume. The biggest merchants get the biggest discounts, and debit spend is concentrated in those big merchants. So, the net lift for Cash App from using an exempt BIN-Sponsor is not as high as one might expect.

Virtual card interchange versus consumer card interchange

Marqeta also disclosed that, after Cash App, their next 3 biggest clients account for a collective 39% of Net Revenue. That leaves only 14% of Net Revenue in their long tail. I am guessing, but I suspect the top 3 include POS lenders like Affirm & Klarna and delivery companies like Door Dash and Uber Eats. Those companies receive payment from a consumer credit or debit card, but pay merchants using a virtual card.

Virtual cards (VCs) are single-use payment cards that carry a high interchange rate of ~2.5%. They are virtual because no plastic is ever produced — it is just a 16-digit number. VCs have a thousand uses, but have become a key monetization method for many Fintechs. Because of the networks’ “Honor all Cards” rule, merchants cannot turn them down despite their high cost.

Affirm offers one example of how this works; when a consumer is in a physical store they can tell the Affirm app how much they want to borrow to finance a purchase. Affirm pushes a single-use VC into Apple Pay for that amount. The consumer then taps their mobile at the point of sale just like any other Apple Pay transaction. Marqeta received ~2.5% interchange on the transaction, most of which it shares with Affirm. When paying the installments to Affirm, the consumer uses a debit card, usually at only 25¢ cost per transaction.

Since these are high-ticket transactions the gap between consumer interchange and Virtual Card interchange is material. The consumer card could have IC from 1.4%-1.8% while the virtual card IC is at ~2.5%.

When integrated directly with a merchant, Affirm reports an average 2.5% subsidy from merchants for a longer-term, interest-bearing loan. Since the VCs are largely used at non-integrated merchants, Affirm gets roughly the same subsidy, but without a merchant agreement. On Pay-in-X loans (nee Pay-in-4) they report a 5% subsidy, so Affirm gets half as much revenue on an interest free, short-term loan from an un-integrated merchant – but it is likely useful as a retention vehicle for active users and Pay-in-X loans tend to be low ticket — so the actual cost is modest.

In Delivery, the consumer pays the delivery company using a personal card, while the delivery company pays the merchant using a VC at ~2.5% IC. This is a particular windfall if the consumer pays with debit, but there is still a gap on most credit cards.

Chime

As I discussed in my post, “Neobanks are not the future”, Chime likely gets 70%+ of its revenue from exempt debit interchange. So, it is in the same boat as Cash App. However, Chime splits volumes over two exempt BIN-sponsors so that neither one risks exceeding the $10B asset cap.

Chime also figured out a way to arbitrage between Credit Card interchange and Debit card interchange. They launched the “Credit Builder” secured credit card which is more flexible than traditional secured cards. Instead of posting a fixed security deposit to back a revolving line, the customer moves money from their Chime deposit account to their Credit Builder account and can then spend up to that amount only – no credit line is available. In fact, they advertise 0% APR because Chime won’t approve any transaction that exceeds the unspent security deposit.

Chime reports transaction activity to the credit bureaus and claims this can increase credit scores by “30 points on average” after 8 months. Neat trick when you are not actually providing credit! Effectively this is just a debit card in credit card clothing – and at Credit Card interchange. It is hard to tell what IC category these transactions qualify for, but they earn at least 50bp more IC than exempt debit.

Chime does not report adoption rates. The card doesn’t pay rewards, so the full interchange lift goes to Chime. Since it also doesn’t provide credit or a real grace period, the prospect of a credit score lift is the key value proposition. In return for that, the consumer needs to split liquidity across their Chime deposit account and their Credit Builder account, and manage that balance to fund spending. Given Chime clients are low-balance, that complexity may be daunting.

The “merchant settlement” could disrupt the status quo

Needless to say, merchants do not like Virtual Cards, or “high cost debit”. They have to accept them under the networks’ “honor all cards” rule, but they don’t have to like it. Honor All Cards is at risk from the pending litigation settlement between the networks and the merchants. The judge in that case rejected a prior settlement because it didn’t address the needs of all merchant classes. The judge sent it back to the parties to negotiate an improved deal, but made a particular point of criticizing Honor All Cards.

If Honor All Cards is retired, merchants are likely to take aim at Virtual Cards – and some have already started. The big mobile providers have changed their bill pay terms to discount for ACH or to surcharge card usage. This is likely because the new wave of bill pay Fintechs push virtual cards into the telcos online while taking payment themselves via ACH or Debit (similar to what Marqeta facilitates for online lenders and delivery companies)

Other verticals subject to virtual card substitution are likely to follow suit. While most surcharging is paid by the consumer directly, here the surcharge would be borne by the fintech intermediary. If the intermediary passes it through to the consumer, demand will decline, but if they absorb that cost, the transaction becomes unprofitable. If that happens, the lenders, delivery companies, bill pay companies and other Fintechs relying on IC arbitrage may need to find a new revenue model.

The most obvious resolution is more partnerships between fintechs and retailers as Affirm already has online. In the case of delivery companies it could be that supermarkets and restaurants cut a deal with a single company rather than accept all of them. The Fintech will get lower revenue per transaction but many more transactions. It general, power will shift from the Fintechs to the merchants, but they need each other so deals will be done.

But not every Fintech adds value to merchants like lenders and delivery companies. Bill Pay is an example. As we saw, the big telcos have already changed their payment terms to avoid high interchange. This will no doubt spread to other verticals where simple bill pay is at issue. Of course, if the settlement keeps Honor All Cards as is, none of this applies.

Conclusion

Many Fintechs exploit price differences between payment methods. In particular, they take advantage of exempt Debit interchange and virtual cards. Exempt interchange requires Fintechs to use sub-scale, exempt BIN-sponsors that have recently faced regulatory scrutiny. Furthermore, These models rely on the card networks’ “honor all cards” rule that requires merchants to accept any card payment, regardless of cost. The merchant settlement may weaken Honor All Cards, putting pressure on these revenue models. 2025 will be an interesting year for all.