We are currently witnessing an internet-based transformation of many major institutions, which would have seemed like science fiction not so long ago (and still might to many currently living). The future of banking is one such area- a transformation from retail bank buildings to a global computer system that makes decisions about how to move money by the use of algorithms rather than humans. A bank that requires almost no people or real estate to function. No more than a headquarters, with some managers and some developers, who may be working remotely, will be sufficient. Although purely remote employees and a purely virtual service is entirely plausible.

The customer of this algorithm bank will rely exclusively on the web interface of the bank for all interactions. For most users of banking services this is pretty much already the case anyway. Even today the beginnings of the future are starting to appear before us.

Additionally, this new algorithm bank will be operating in a different environment than retail bank locations have enjoyed for the past few decades. Universally people will be perfectly capable of securely keeping their money with nothing more than a phone, so the “secure vault” functionality of the bank becomes superfluous and unnecessary. Further, people will be able to securely transfer anywhere they wish at any time for a negligible fee, so the “send/receive” functionality of conventional banks is also superfluous.

The Return of Real Interest

Therefore I make the strong prediction that there will be a return to the canonical trait of a bank deposit; interest payments. Banks have offered virtually zero interest for many years, because the other services they provide, as well as just inertia, has been sufficient to convince people to stay with them. By contrast the “algo bank” will be left with no alternative but to offer a competitive interest rate to persuade customers to deposit money in their bank as opposed to hold their own money. Or select a competitor bank that offers a better interest rate. A bank of this type that does not offer enticing interest to its depositors will fail, defeated by competitors who do offer enough interest to encourage customers to choose them over the competition.

Due to the low transaction costs and high speed of transfers, the algo bank must expect their customers to be highly fickle as to which bank is offering the best interest rate at any given time. This is an unavoidable consequence of ultra-fast and high-efficiency transactions. A bank that is offering the highest interest rate will likely see a sudden influx of depositors, who will also swiftly leave the moment a competitor is offering a higher rate.

I think that this ‘fickle customer’ issue will be considered untenable by many financial institutions, at least for the near future. Aggressive players might be willing to live with the uncertainty, but most will not.

I believe the most likely solution banks will apply to solve this problem is offering “products” that involve lockups for variable periods of time. Specifically, suppose Bank of America were to offer a “product” that involves a 7 day lockup for a specific interest rate. The bank may offer a range of such products of variable lengths and interest rates, attempting to incentivize depositors to select a longer deposit period at a higher percentage rate. Perhaps this may take the form of 1 day, 3 day, 14 day, 30 day, 90 day, 180 day, and even 365 day options to choose from. Given this approach the bank will likely assign the best rates to the longest terms, making their availability of deposits more predictable in exchange for a beneficial interest rate.

The interest rate that the bank can offer for a specific product will depend most directly on the rate of return that the bank expects from whatever it plans to do with the money deposited. The bank’s obvious idea is to collect more income than it is required to pay out to its depositors in interest. There are any number of ways that banks can collect income by lending, investment, trading, etc. that I won’t speculate about here. But to do so they require capital which is obtained from customers. I believe the bank of the future is likely to rely heavily on quantitative or algorithmic approaches that are designed by people but executed by computer, but the specific methods behind the scenes are not important to the discussion of the consumer-facing algo bank concept.

One way this might play out, is the banks providing software for the end user, such as a BofA app. But I think it is more likely that we will see third-party software that is bank-agnostic and which connects to the API’s of many banks. Much like how you can buy Android apps that aggregate social media accounts from various different services. And this app will select from offerings of the various banks which it has API’s for, according to the user’s preference. Such as the user selecting “highest APY” as a possible example. Even a standardized API could be created which would allow for the existence of a “micro bank” with an arbitrarily small size.

Micro Banks & Regulations

From a practical, and a technical perspective, there is no reason why an algo bank could not be a “micro bank” of an infinitesimal size, indeed an arbitrarily small size. The barrier here is regulatory, and the regulatory concerns regarding banking are very substantial, for good reason. But bank regulations were simply not designed for this scenario.

The approach that makes sense to adopt here is to change the basic method and principle of the style of regulation, from a retroactive punishment as a response to a transgression by people, to a proactive examination of the code and processes that will be used by the algo bank.

A financial regulator will likely find it extremely challenging to investigate and hold accountable an endless myriad of tiny entities simultaneously executing all over the world in real time. The amount of data flowing in will be unmanageable. No human can keep up with that level of complexity after the fact. This retroactive approach is doomed to failure- and I am calling it now, we are going to see the technologically illiterate cry Armageddon rather than actually attempt to work the problem in any serious way.

However, the fact that the execution is code, provides the very tools that makes it entirely possible to verify its execution BEFORE it is deployed, and thus reach informed conclusions about broad patterns before they happen, even if that code is being executed billions of times all over the world.

I propose as a solution to this potential problem, that regulators require “algo banks” to register with them, likely with a fee, which could be very small. Further, each “product” that an algo bank will offer must also be registered, also with a fee. The regulator will then inspect the code associated with the product, rather than attempt to investigate its execution after the fact, which may not be humanly possible.

At first this will be a difficult project. Initially there will be a tremendous variety of wild experimentation which will require careful examination and mistakes will be made. But eventually we will see successful, well-tested forms emerge that can be effectively copied and largely reduplicated. Forms that will become familiar. Forms that may become so familiar that it may be adequate for approval simply because they are the exact same forms that someone else has already used with success for years.

Large and aged institutions tend to be slow to change. So I am not optimistic about the speed at which this shift in methodology will occur. But eventually they will be left with no practical alternative but to change their approach from a retroactive one, to a proactive one, simply because the retroactive effort is so laborious and ineffective compared to a relatively small pre-emptive one. An ounce of prevention is worth a pound of cure, as they say.