Take Me to the River: BaaS Pushback Should Be a Moment for Banks and Regulators to Readjust Priorities
Generative AI is forcing banks to adapt over the next five years — and regulators should not get in the way, lest we want a broken banking system.
March ended with a tough week for Banking-as-a-Service (BaaS). Four of the remaining stalwart BaaS-friendly banks have received FDIC consent orders for various financial crime and consumer protection issues. This is on top of blowups or exits at Blue Ridge, Lineage, First Fed, MCB, Evolve and others, and many other banks effectively derisking their third-party/fintech relationships in various ways. This was telegraphed last summer, when the OCC, the Fed and the FDIC released interagency guidance on third-party risk (good analysis here).
The immediate impact of this guidance was for the larger, OCC-regulated banks to get spooked and derisk, because this interagency guidance was incrementally tougher on third-party relationships than the prior standalone OCC guidance. But the real news was that the Fed and FDIC, who between them oversee the state-chartered banks, were dramatically upgrading their standards in aligning with the OCC guidelines. That took a minute (in bank regulation time, that’s nine months) to flow through, but here we are.
This is not what the banking industry needs right now.
As I have covered elsewhere, the industry is facing an ongoing balance sheet crisis, due to a combination of underwater securities and deteriorating commercial real estate loans. These issues are most acute at the smaller banks. The industry will survive this crisis, but will soon run headlong into the next, more existential issue: the inevitable contraction in net interest margin (NIM) caused by AI agent-enabled rate shopping on both the deposit and lending sides of their business. The core business of banking is a relatively pure commodity predicated on opacity, latency and laziness, and all of those pillars are crumbling with the advent of open banking, real-time payments and generative AI.
Banks have three options
To prepare for the AI future, banks have three options. They can get big, get embedded or get creative.
Getting big is the easiest to understand. Banking is having its Ben Franklin moment: The banks must hang together or they will assuredly hang separately. If NIM is going to contract, say from 300 basis points to 150 basis points on average, then you need twice as many assets to generate the same dollars of profit (leaving aside for the moment that this will be terrible for ROEs). In that banking has a significant fixed cost component to it, which is going up as technology gets more central, these absolute dollars of contribution margin are essential to survival. A massive wave of M&A-driven consolidation has to start right now.
Another option is to embrace embedded finance (also covered elsewhere). Embedded finance is the idea that financial services are best distributed, underwritten and serviced in the context of sticky, data-rich software vs. as independent, discrete products. This embedding and bundling will greatly reduce the motivation and ability of customers to aggressively and frequently price shop. Banks play a critical role in embedded finance, as infrastructure providers (e.g., BaaS) and potentially as software-driven companies themselves (existence proof still pending).
Finally, banks can get creative; in other words, they can do the hard things. Banks have gotten themselves in this pickle by failing to innovate on the deposit side, and by gravitating to commodity, FICO score-based lending on the loan side. When all of your products look just like the other guy’s, and are underwritten using the exact same criteria, price shopping is quite easy. Banks need to create distinctive, relationship- and brand-based deposit products that leverage sophisticated loyalty systems and value propositions beyond interest rate. They need to create lending products that are differentiated by whom they will underwrite, at what size and in what timeframe. They need to create new loan types, or advance existing loan types, to create real customer value, like BNPL, MCA, factoring and other supply chain finance loans, etc. Right now, these loan types are considered exotic and are largely delivered by non-bank finance companies; they need to be mainstreamed by the banking industry.
And regulators have three priorities
As evidenced by their recent actions, our banking regulators are not currently well set up to assist the banking industry in making this transition. Regulators have three quite valid but often competing priorities: safety and soundness, financial crime prevention and consumer protection/competitiveness. We also have an alphabet soup of regulators who, as the blind men and an elephant parable goes, often can only feel their part of the elephant. I want to be very clear: This is not a critique of the regulators! I am highly sympathetic to their position, and certain that if I worked in any particular agency I would be acting just as they are. But their competing priorities produce a great deal of confusion and inertia for banks.
The safety and soundness, or prudential regulators, should want M&A, but are afraid of the fast money aspects of embedded finance and any creative lending or product development. These regulators are interested in passing what’s called “Basel III Endgame,” which perhaps appropriately seems like the name of a super boring Bond villain. Leaving aside whether this is necessary or not, it will significantly increase capital requirements for any novel or creative lending activities.
The financial crime regulators detest embedded finance, full stop, and want to go back to wet signatures in the branch (and potentially insist on customers arriving on horse and buggy). They are begrudgingly fine with M&A because they recognize that you need scale to have world-class systems, and if they can’t mandate the return of the 1950s, they need banks to have better tech. These were the folks riding herd on the BaaS industry last week, with their Oprah Winfrey-style willy-nilly distribution of consent orders.
The consumer protection and competitiveness regulators are good with embedded finance (and might even finally enforce Section 1033!) but despise M&A and consider creativity guilty until proven innocent, even if it’s much better for the customer than the alternatives.
Of course, most agencies have some responsibility for all three priorities, but they are not well-positioned to balance them in a nuanced way that considers the unintended consequences of each quite valid objective. Despite my irrepressible snarkiness, I am very sympathetic to the regulators — these issues are complex, and anything they do that appears to be friendly to the banking industry will be attacked, even though the health of our economy depends on the health of our banking system quite directly.
This is a case of talented, thoughtful people, each pursuing understandable and legitimate goals for the benefit of our society, whose insufficiently coordinated efforts, however, will be the end of our financial system as we know it. (Not investment advice, but you might want to buy some BTC, ETH and SOL, as we have.)
Moving past a cycle of sin and redemption
“Take Me to the River”, originally by Al Green but then taken to new heights by the Talking Heads, is clearly the best cultural meme related to Banking-as-a-Service, but frankly it’s also a confusing song.
It is prima facie about baptism, the cleansing of sin, the delivery of that inevitable consent order. That said, it’s also about the reality of vice, financial crime and questionable relationship choices. Perhaps the latter creates the need for the former, but the song reads more like a nuanced cycle of sin and redemption than a straightforward morality tale. The implication, surely, is that there will be more than one trip to the river.
The banking industry and its stalwart regulators are going to have to move beyond a Manichean model of sin and redemption. As banks fumble their way through the turbulent waters of the next five years, they will need to get big, get embedded or get creative — or some combination.
There will be some bad actors, but much more common will be those making simple mistakes, of strategy or execution. If the regulators, via a lack of coordination or a concerted effort to forestall change, inhibit scale, embedded finance and creativity, we may end up with a pure banking system, but not a healthy or effective one.
Related Insights
Selective State Space Models: Solving the Cost-Quality Tradeoff
As AI is increasingly used in production scenarios, costs are mounting. Are alternative architectures the solution?
Only Up From Here: 2024’s State of Fintech and the Hero’s Journey
We’re at the scary part of “the hero’s journey” in fintech, with a bottom for funding and a BaaS crisis, but we’re about to enter a new, exciting stage.