Warehousing Promises

Earlier this year, a prominent fintech infrastructure startup precipitated a massive collapse of a portion of the fintech ecosystem because of growing errors in its ledger over about 18 months. In the wake of this failure, many parties are looking to understand how to prevent another failure of this kind in the future.

Let’s take a closer look at how fintech intermediaries like Synapse facilitate non-banks to offer banking like accounts, and dig into why one solution should be to think about these services as warehousing promises.

What is a promise?

In a financial context, a promise—sometimes called an IOU (”I Owe You”)—is an assurance that one party will give the other party cash. For example, suppose you loan a friend $10 for lunch. In exchange, your friend offers you a verbal promise that they will pay you back the $10 next week.

Promises are a fundamental building block of our financial system. You find them everywhere from bank accounts to loans of various stripes—any time that cash is exchanged for future consideration.

In the mid-1800s in the United States essentially anyone or any business could issues promises. But outside of informal arrangements with friends, such promises are risky, and allowing any entity to make them brought about a great deal of chaos and confusion for everyone involved. To reign in this chaos and stabilize the American banking system, the National Banking Act of 1863 mandated that only banks with a federal charter could make a promise.

Making a promise

Banks are in the business of making promises. That is, they aim to make a profit by creating promises. Suppose you deposit $100 into a bank account. You hand the bank cash, and the bank hands you a promise, specifically a promise to return your $100 as cash on demand.

Sounds reasonable, and risk-free. But just holding onto your money is not a great business model. Banks in turn make money by investing the cash you gave them into financial products that earn interest—perhaps they loan your $100 out as part of a home mortgage for example. Of course any investment that earns interest does so because it has some degree of risk attached to it. Which means that the promise the bank made to you is itself mired in at least some risk.

It’s for this reason that banks and the promises they make are extremely tightly regulated. It’s also why the FDIC and similar institutions in other countries exist: Events like a bank run can lead to insolvency because a large percentage of the funds will be tied up in illiquid investments—the bank can’t just demand that mortgage holders immediately pay their loan off, to continue the example. These governmental financial institutions exist to prevent fraudulent behavior, ensure that banks are properly managing their exposure to risk, and ultimately to protect the individuals and businesses who have deposited funds with those banks.

When a promise is not a promise

Fintechs are in a gray area in the sense that they are in the business of making things that look and act like promises. This is what we mean when we say that fintechs are not banks. Fintechs are not chartered to make promises to customers in the same way. Instead, they act as a broker between customers and banks. Neobanks look like a bank, and appear to make promises to depositing customers in the same way as a bank does, but the reality is a bit more confusing.

Imagine this simplified scenario. You deposit $100 into an account at Neobank. Neobank in turn deposits that $100 into a partner bank (a chartered bank, such as Evolve or Lineage). The bank issues a promise to Neobank, and Neobank issues a promise to you on the basis of the bank’s promise. FIntechs issue promises backed by bank-issued promises.

Of course the real world is more complicated. There might be intermediaries who manage deposits and promises on behalf of neobanks and banks, such as Synapse did. The deposited funds are almost certainly aggregated across customers, and then placed into multiple accounts at different banks. So oftentimes fintechs issue promises backed by other fintech-issued promises, which in theory are ultimately backed by bank-issued promises.

This is why maintaining a careful ledger is so important, and why Synapse’s failure was so devastating—because without an accurate ledger, it was impossible to know which funds belong to which beneficiaries. (The observant will have noticed the close analogy to money laundering—the only difference is that money laundering is designed to deliberately obfuscate.) As the errors in Synapse’s ledgers compounded, the promises they were issuing to their fintech clients were gradually not backed by actual bank promises, because there was no clear way to connect the promises they issued to the promises issued to them by banks.

Warehousing promises

But in a post-Synapse world, customers, legislators, and regulators are increasingly asking for a fundamental change in the role of fintechs as intermediaries that offers better protection to end customers. Fundamentally, fintechs should not be in the business of making promises, but of warehousing them.

Here is a simplified version of what this looks like: Again, imagine you deposit $100 at Neobank. Neobank hands this money to a partner bank, and tells the bank this money belongs to you. The bank then issues a promise not to Neobank, but to you, which Neobank then manages on your behalf. Neobank has not issued a promise of its own. This seems like a small conceptual change, but it will require a fundamentally different—and (more importantly) improved ledgering model that prevents a situation where a fintech’s promises become unmoored from the underlying bank’s promises.

What does warehousing mean? Rather than tracking the movement of fungible funds—whose movement does not provide an easy-to-audit connection between the funds and their owner, fintechs should inventory uniquely identified promises on behalf of their customers and their partner banks. This way of tracking is much more explicit about who the funds belong to. The job of the fintech is to keep the ledger that allows them to query their warehouse of promises to tell users (or clients) precisely how much they have in their accounts, and to tell banks precisely which beneficiaries own which funds.

There is a complication, in that promises are not fungible in the same way that money is, which means we need a new set of semantics that enable the system of records of fintech companies to properly orchestrate the promise factory chain, with perfect end-to-end traceability. For example, a fintech might be storing your funds in Bank A, but might help you perform an FX instant transfer by leveraging funds in Bank B. Managing these kinds of transactions will require a semantics that allows transmuting promises from Bank A to you into promises from Bank B to you.

We’ll explore how the semantics of a fintech ledger will differ from a bank ledger in our next article.

Conclusion

Building and maintaining a ledger in an error-free way is difficult work, moreso than many early fintechs realized. The only semantics available to manage fintech ledgers is modeled on the movement of fungible value from one account to another. But because fintechs act as intermediaries, this model requires fintechs to issue promises backed by bank-issued promises, which we now understand to be deeply problematic, and difficult to track. Instead, fintechs should think of themselves as warehouses of promises, managing promises between banks and clients on their behalf. But because promises are non-fungible, warehousing will require a new ledger semantics, but one that will ultimately make auditing transactions easier, and that provides greater protection for customers and other fintechs in the value chain.