Debanked by the Market

The crypto industry has been spreading a tale of federal bank regulators persecuting crypto and forcing banks to "debank" crypto companies. Like the grossly mischaracterized Operation Choke Point, the crypto debanking narrative is utter and self-serving bs. At best, the actual evidence shows the FDIC expressing very normal and reasonable risk management concerns—that is, the FDIC was just doing its job. There is zero evidence that the FDIC ever threatened or directed banks not to do business with crypto companies.

The simple truth is that crypto companies were debanked by the market, not regulators. Banking crypto poses a unique, correlated credit risk that should rightly concern any bank's risk committee. Crypto companies present a risk of correlated chargebacks that makes them all potential Fyre Festivals, so banks with prudent risk management practices determined that it was a value negative proposition to provide banking services to crypto companies. That's the invisible hand of the market at work, not the invisible hand of the Deep State.

Let's start with an empirical point:  everyone is ignorant here. No one knows how many crypto companies—or consumers—have been "debanked" and no one who has been debanked actually knows why because banks do not generally tell customers why they closed an account. So this whole debanking issue stands on a pedestal of (often self-serving) conjecture. 

Now for the main point: providing deposit services for crypto has a lot of unpriceable credit risk, so many well-managed banks reasonably avoid crypto firm customers on their own. No regulatory intervention is needed.

Banks provide three basic services: payments, deposits, and lending, but ultimately all three involve credit risk. Every crypto firm needs a payment processor--it needs some way to accept payments unless it's taking cash and checks. Payment processors (a/k/a "merchant acquirers") require that their customers maintain their deposit accounts with them. This is in part a risk-management tool. The result is that payment and deposit services are bundled. Now here's what's often not well understood: payment processing entails credit risk.

If a consumer claims that a transaction was unauthorized or that they did not receive the goods and services ordered, they can initiate a chargeback with their bank. When that happens, the consumer's bank has a right to claw the funds back from the merchant's bank. That leaves the merchant's bank holding the bag unless it can recover the funds from the merchant. If the merchant is bankrupt or absconds, the merchant's bank—the payment processor—is stuck with the loss.

Now most of the time this is not a big deal. Every merchant deals with a problem of onesies-twosies chargebacks. As long as chargeback levels are predictable, payment processors can price for them. Consumers never see this pricing, but merchants pay the acquirer bank a fee for every payment they receive. On credit and debit cards, it is the "swipe fee" (formally the "merchant discount fee") and a part of it gets remitted back to the issuer bank as the "interchange fee". Riskier merchants pay higher swipe fees. So a low risk operation (let's say Wal-Mart) is probably paying a merchant discount fee that is probably under 2%, while a high-risk merchant (say a gambling website) might be paying more like 10%-15% of every transaction as a processing fee. Chargeback risk is manageable when it is predictable; banks can price for predictable risk. 

Certain industries—e.g., adult websites, guns-and-ammo, gaming, bail bonds, and lawyers (!)—tend to have higher chargeback rates. These industries are often served by a subset of banks that specialize in high-risk payment processing. Crypto companies also fall into that category of having high chargeback rates, but the chargeback risk posed by crypto is fundamentally different.  Chargeback risk is manageable when it is predictable, but that requires that it be uncorrelated. Lots of happily married men might claim they did not authorize that purchase OnlyFans subscription, but they aren't all making that claim suddenly and at the same time. The chargeback level if high, but basically static and predictable.

Crypto doesn't work like that. Although there is probably always a relatively high baseline level of chargebacks for crypto, the chargebacks are also likely to be highly correlated whenever there is a market crash or allegations of fraud about a particular coin. If the market goes up, everyone is happy with their transactions, but when it falls, customers try to get out of their losses by claiming that the purchase was never authorized in the first place. Now Visa has a 120 day time limit for issuers to chargeback transactions to acquirers, so the acquirer has to worry about market movements in a completely unpredictable market for the next four months.  If say, Bitcoin crashes, chargebacks are going to soar at the very time when the risk of the crypto company customer going bankruptcy rises. So the scenario that the payment processor is facing is that it will be hit with a tidal wave of chargebacks and that it will not be able to recover them from the crypto company, but will instead just have an unsecured claim in the crypto company's bankruptcy.

That's the kind of risk that a bank's risk committee--and federal banking regulators—should be worried about. It is a materially different risk that generally exist for other types of high-risk merchants. The only similar situation is with troubled airlines and concert promoters. If the airline stops flying or the concert tour doesn't happen (because, say, it was for Michael Jackson, who has died), then there's going to be a tsunami of chargebacks. Not surprisingly, if an airline's finances start to decline, the processor will require a holdback of funds to cover potential chargebacks (and this can result in a self-fulfilling prophecy as the airline is deprived of liquidity). And for concert promoters, there might be a key person insurance policy requirement to protect the bank. But there isn't an obvious risk management move when dealing with crypto companies, particularly when banks had not had experience with it through multiple business cycles.

Once you understand that banking crypto exposes banks to an unpriceable type of credit risk, it becomes clear why crypto ventures were "debanked".  Banks reasonably recognized that the credit risk—not the reputational concerns—posed by providing services to crypto ventures outweighed the potential revenue. Yes, there might have been short-term gain from banking crypto companies (Silvergate), but it would be at the expense of existential risk (Silvergate again...). Unless treated as a loss-leader for other income streams, banking crypto is a value-negative proposition, so banks with prudent long-term risk management outlooks generally got out of it. That's market-driven debanking and cannot be blamed on the government. Bank risk committees don't need the FDIC to understand the risks of providing payment processing (much less actual lending) to crypto companies.

The implication here is that no amount of crypto-friendly bank regulation is going to change the fact that prudent risk management committees will eschew crypto companies. Instead, it will be the poorly managed operations—the Silvergates of the world—that chase the quick buck, and which will ultimately run into trouble because of it. (Indeed, the very fact that banks like Silvergate served the crypto industry indicates that regulators weren't pressing too hard--do we really think that Silvergate would have told the FDIC to take a hike?)

Coda. The other part of the "debanking" story—the one raised by some Democrats as well as Republicans—is that of closures of consumer accounts. It's a similar story:  consumers who repeatedly overdraw their accounts pose credit risk to banks. Yes, the bank can hit them with overdraft fees, but that doesn't do any good if the balance stays negative, and there are administrative costs of handling those accounts otherwise. But even if an account doesn't go negative, it's still a money loser for the bank unless (1) there balance is high enough to general sufficient net interest margin to cover costs, which is probably about a $5000 balance or so, (2) the account pays a lot of fees, or (3) the account is a loss-leader for other cross-sold products. If none of those three conditions are met, then the account relationship is value negative for the bank, and one would expect the account to be closed. The US doesn't have a basic banking requirement like, say, Canada, as a condition of the bank charter privilege, so account closures for consumers are going just be a market-driven process, just as they are for crypto firms. There's a policy argument for having a basic banking requirement, but until and unless we have that as a matter of law, banks should be making market-driven decisions.