Recent Posts

Wednesday, March 22, 2023

Narrow Banking: A Bad Solution To A Non-Existent Problem

Narrow banking is a concept for a bank that holds 100% reserves against deposits. It attracts people who are deeply concerned about the symbolic content of “money” on both the left (e.g. Positive Money) and the free market right (the Chicago Plan). Devotees of narrow banking are happy to talk your ear off about how their plans work, so I leave finding out more as an exercise as a reader. I just want to focus on the core principle: they want banks to not take risks lending deposits, so that “money” remains “money”: a numeric entry that corresponds in a 1:1 fashion to a claim on a “monetary asset,” like a gold coin or claims on specific gold coins, and not a messy credit relation.

Note: Once again, this article is an unedited draft from my banking primer manuscript. This explains some unusual wording used here, since I want to be able to paste in text into the full text without having editorial bombs that are hard to spot later.

As the title of this section suggests, I am not a fan of the concept. As I will explain, creating a narrow bank within modern economies is largely a solved problem. You will need capital and to be able to hire competent personnel, but assembling the bank thereafter is just tying together existing software packages and personnel together into an institution. The reason why nobody does this is simple: there is no market for such an institution. Sure, you might be able to assemble a cult following among narrow banking bugs in a large country, but it is going to be flop among the broad public.

What I Mean By A “Solved Problem”

When I write “creating a narrow bank is a solved problem,” I have a somewhat narrow definition in mind. What has been solved is the ability to create the product being sold, but not the problem of selling it profitably. The following engineering analogies hopefully illustrate the point.

  1. Setting up a construction firm is a solved problem. You need to hire architects, civil engineers, and workers in the trades, and administrators (although some administration can be outsourced). All the equipment you need to buy is standard industrial equipment, and the people you hire will have been trained at other firms and/or the educational system. So long as everyone is somewhat competent, you should expect to be able to build a building. However, the running a construction firm profitably is not solved — you need to find clients, and competently manage expenses.

  2. Setting up an electric car company (now) is not a solved problem (assuming that you are not stealing an existing design, which is hard to pull off legally). Although you will be hiring trained engineers and have some parts available at existing suppliers, you still need to design a car as well as some key components. Even if you assemble your team, there is no guarantee that you will be able to produce a viable working electric car.

Solved Problems in Finance

Setting up a financial firm is normally closer to electric car company example, as you need some distinctive “hook” (e.g., proprietary trading algorithms) to get people to invest with you. But if we put aside the proprietary algorithms, investment management can be closer to the construction company example.

For example, before the 2008 Financial Crisis, there were a lot of quantitative credit funds that consisted of two traders, a trading terminal, a spreadsheet with Gaussian copula models, and some sales people. The rest of the administration could be outsourced. (Given the track record of such funds in that crisis, that model for hedge funds is now frowned upon, but the risk appetite cycle might turn back.) The spreadsheet was the only proprietary technology that was needed, everything else was software and training that was available already. However, if the fund was a naïve index fund, you did not even need the Gaussian copula models. Nobody is going to create a hedge fund to do a naïve index portfolio, but an existing financial firm could bolt on an index fund structured this way. (As an aside, most index funds do require some proprietary strategies to deal with the thorny problem of transaction costs, but some fixed income indices are very simple and clients do not reallocate funds rapidly, and a naïve matching of the index works.)

Creating A Narrow Bank

Let us imagine that we want to set up a narrow bank that is largely internet-based. It does not have bank branches, and possibly does not deal with cash or cheques (“checks”). It offers deposits and term deposits that are 100% covered by central government liabilities (reserves, government bonds and bills, and repo transactions on government paper). In a country with a somewhat modern payments system (e.g., not the United States), not covering cheques and cash might not be too limiting for a “internet savings bank” (which exist).

As a disclaimer, I have not set up a bank previously, but my guess is that most components are largely solved.

  1. Administration would be similar to any other bank.

  2. Risk management could use off-the-shelf software, and you can hire risk managers familiar with that software. They will need to develop internal reports built on top of that software, which they would have done at the old job.

  3. Portfolio managers could be hired from treasury, money market, and government bond trading desks. Their trading objectives and style would be different, but they would be working within the same frameworks.

  4. The interface to the wholesale payments system is country-specific. Whether off-the-shelf software exists might depend upon how many banks there are in the country (creating a market for independent software). At worst, you need to hire programmers to build an interface to a known protocol, replicating existing banks’ software.

  5. The website or apps facing clients would be proprietary, but programmers could use off-the-shelf components.

  6. Cheque processing and the management of cash implies a larger infrastructure, and would only be appealing for a large internet bank (or a bank with a single branch).

Other than the last point, none of the above represent insurmountable problems for an entity with fairly deep pockets. Furthermore, an existing bank could bolt on a “narrow bank” subsidiary and re-use most of its technology and know-how (if regulators let them, which is unclear).

So why not do it?

No Market

The problem for narrow banking is that the client base is tiny. It consists of the following.

  1. Ideologues who want to bank with a narrow bank on principle.

  2. People and firms with deposits above the deposit insurance limit who are too dim to sweep deposits into other money market instruments.

Readers familiar with the demise of Silicon Valley Bank in March 2023 would note that there were a lot of firms and rich people with very large uninsured deposits. Am I saying that all those uninsured depositors were stupid? No, because those deposits seem to largely explained by Silicon Valley Bank implicitly and explicitly pushing those clients into those deposits. In some cases, holding the firm’s cash as deposits at Silicon Valley Bank was required in debt covenants. In others, it appears that the deposit holding was used as a cross-subsidy to other transactions. The capitalist system is populated by people, and cross-subsidies happen. However, cross-subsidies are not want you want when liquidating a firm, so most traditional firms are more averse to them than Silicon Valley and its major depositors.

The Product Stinks

The narrow bank has to pay a lot of expensive salaries and software licenses to keep the business running. The only revenue are fees, and the spread of interest paid on government debt over what it paid on the deposit accounts. If the deposit rate is 0%, that spread might be attractive in a high interest rate environment. However, the New Keynesians at central banks can take away the high interest rate environment at any time.

How the narrow bank makes a profit is uncertain, but it certain involves hefty bank fees and/or paying a decent negative spread versus the risk-free curve.

Government Bill Funds Exist

By not attaching the savings product to the payments system, mutual fund companies can offer Treasury Bill funds that pass through interest on Treasury bills with a fee that is typically under ten basis points (0.1%). These funds tend to be loss leaders for fund complexes: they are there to provide a product to keep client asset allocations within the same fund family, but they are not expected to be a profit centre.

For anyone who is not a narrow banking ideologue, this is where you stick your extra cash if you are risk averse. Narrow banking is the bad solution to risk aversion, Treasury bill funds are the good one.

Incompatible With Industrial Capitalism

We can now see why banks generally do not bolt on “full reserve” subsidiaries — they can cover the client base with a government bond fund issued by a broker subsidiary (courtesy of the demise of the pillar system). The terms associated with fully reserved deposits would insult clients when compared to existing products.

Banks only offer deposit services because they are profitable — a diversified deposit base is sticky, and provides cheap funding for their lending operations. And lending operations are the core purpose of banks. Much as air pollution is the breath of industrial capitalism, credit flows are its blood.

Fans of narrow banking believe that stability of the value of money is the paramount interest of actors in the economy. In reality, that is only true during the worst of a financial crisis. When animal spirits are running rampant, portfolio allocations drift towards riskier money market credit products. In a crisis, those products collapse. Bank deposits’ ambiguous status between “money” and “credit” allows them to weather the storm, and provide stability for the broader financial system.

Deposit insurance eliminates the worry about credit risk for the vast majority of the population. The handful of people and firms who run deposit balances above the insured limits have the resources to deal with credit risks. Meanwhile, the risks of uninsured deposits is minimal in most cases given their priority in a restructuring, under the assumption that regulators make a slight effort to regulate banks (ahem). The only real credit risk to deposits in well-managed banks are in systemic meltdowns — which have much larger costs than just what happens to uninsured deposits.

Systemic Risk — Probably Worse

Although fans of narrow banking love to argue that narrow banking reduces systemic risk, my expectation is that it makes it worse. Since narrow banks cannot offer loans, we still need a risky bank/non-bank system.

The risky system will offer money market funds, which will out-yield narrow bank deposits, and are integrated into financial market trading. In the absence of crises, these risky products will attract more an more inflows. (“Stability is destabilising.”)

And when things go wrong in the real economy, credit losses will blow up those credit products. If people flee to narrow banks, there is no way of recirculating those flows back into risky assets. This is different than traditional banks, which participate in risk asset markets.

The usual response from narrow bank fans is along the lines that this is exactly what is supposed to happen — financial markets are risky, and should not be bailed out by central banks. However, I am exceedingly unconvinced that politicians are going to step aside and watch large employers fail due to an inability to roll over their borrowing, or getting their liquid assets impaired.

Concluding Remarks

There is a dispersed ideological constituency for the creation of narrow banks. These efforts never gain much ground, since everybody else realised that government bill funds and deposit insurance deal with the alleged problem.

References and Further Reading

This section was largely the result of my experience in working in finance. As should be obvious, I am not interested in the details of the suggested narrow banking plans, rather the more basic implementation question. I also question the theoretical claims about the advantages of the system of proponents, but that discussion rapidly goes out of the scope of this text.

  • On the libertarian right, “the Chicago Plan” is the usual framework. Jaromir Benes and Michael Kumhof wrote “The Chicago Plan Revisited” which discusses the idea. URL:

  • On the progressive side, the “Positive Money” movement covers this. URL:

  • My comments about stability are a very high level summary of Hyman P. Minsky’s dictum that “stability is destabilizing.” His analysis of the evolution of American financial system in a destabilising fashion away from a highly stable system that came out of World War II provides examples of the principle in action. Although Stabilizing an Unstable Economy is his most accessible book, Can “It” Happen Again?: Essays on Instability and Finance (Routledge, 2015) is a better source for this particular topic.

Email subscription: Go to 

(c) Brian Romanchuk 2023

No comments:

Post a Comment

Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.