Narrow banks — unbundling money from credit

The fusion, and the obvious fix

Recall the deepest of the three root causes from Module 01. Conventional banks fuse two functions that need not be joined: issuing the money people use, and financing loans. They do it because the bundle is profitable — the spread on lending, magnified by leverage, is the business model. But the fusion has a price that everyone pays: because your deposit is a claim on a bank that has lent the money out, holding everyday digital money makes you an unsecured creditor of a leveraged lender, and the whole structure is fragile because callable deposits fund illiquid loans.

Once you state the problem as a fusion, the answer suggests itself. If the trouble is that money and credit are bundled, unbundle them. Build an institution that does the money part — holding deposits, making payments — and simply does not do the credit part. A bank that takes your money and does not lend it out cannot lose it to bad loans, and cannot be unable to return it on demand, because it never tied the money up in anything illiquid. Your deposit would stop being a claim on a risky loan book and become what most people wrongly assume it already is: safe money, fully there, available on demand.

That institution has a name — the narrow bank — and the simplicity of the idea is exactly the point. It attacks one of banking's most fundamental problems not with a clever new technology but with a different answer to the question "what is a bank for?" It is a structural innovation in the precise sense of Module 01: it changes what the institution is, not what tools it uses. Hold that in mind, because it is why this module leads the innovation track.

What a narrow bank actually is

Precisely defined, a narrow bank takes deposits and holds them entirely in the safest, most liquid assets available — ideally balances at the central bank itself (reserves), or failing that, short-term government securities. It makes no loans, takes no credit risk, and runs no maturity mismatch. Its assets are as safe and liquid as money gets, and they match its deposits one-for-one. Compare the two balance sheets, per $100 of deposits:

The contrast is the whole lesson. The conventional bank holds $6 of liquid reserves against $88 of deposits it owes on demand, with the rest tied up in loans — the recipe for the credit risk and the run risk of the classic track. The narrow bank holds $100 of reserves against $100 of deposits. Every dollar deposited is sitting, fully and liquidly, at the central bank. Two consequences follow immediately and are worth stating sharply:

• It cannot suffer a run. A run happens when withdrawals exceed liquid assets (Module 04 of the classic track). A narrow bank's assets are entirely liquid and exactly equal to its deposits, so it can always pay everyone, instantly, in full. The coordination trap that topples conventional banks simply cannot form — there is no reason to be first in line when everyone can be paid.
• It carries no credit risk for the depositor. The deposit is backed by central-bank money, the safest asset in the system, not by a portfolio of loans that might default. The depositor is no longer an involuntary creditor of a leveraged lender. The fusion is undone.

This is run-proof, credit-risk-free money, available to the public, in digital form — precisely the "missing cell" the classic track's Module 03 identified, the combination that conventional banking cannot offer. And it requires no blockchain, no app, no new technology of any kind. It requires only the decision not to lend.

A very old idea

If the narrow bank sounds like a fintech-era novelty, it is not — and this matters enormously for the course's thesis. The idea is nearly a century old, predating computers entirely, which is the clearest possible proof that the most powerful financial innovations are often structural and legal rather than technological.

The roots reach back to the goldsmith warehouse of Module 01: an institution that simply stored your money and charged a fee, taking no risk with it. The idea became a serious reform proposal after the bank failures of the Great Depression, when economists at the University of Chicago advanced what became known as the Chicago Plan — a call for 100% reserve banking, under which deposit-taking institutions would hold full reserves against all deposits and could not create money by lending. The economist Irving Fisher championed it in his 1935 book 100% Money, arguing it would end bank runs and tame the boom-bust credit cycle in one stroke. Henry Simons and, later, Milton Friedman lent the proposal support across the ideological spectrum.

The idea never went away; it returns after every major crisis, because every crisis re-teaches the lesson that fractional-reserve banking is fragile. After 2008, economists at the International Monetary Fund revisited the Chicago Plan in detail, and the broader debate over "sovereign money" and full-reserve banking flared again. Narrow banking is less a new invention than a recurring answer that the financial system keeps declining to adopt — which raises the obvious question this module builds toward: if the idea is this old, this simple, and this safe, why does it almost never happen?

Modern incarnations

Although no pure narrow bank operates at scale for ordinary depositors, several modern instruments are narrow-bank-like — they approximate the structure in different ways, each revealing something about the idea's promise and its obstacles.

Two of these deserve a moment. Money-market funds are the closest thing most people can already use to a narrow bank — they hold safe short-term assets rather than lending — but they revealed the catch in 2008, when one prominent fund holding Lehman debt "broke the buck" and triggered a run on the whole sector. A narrow vehicle is only as safe as the assets it actually holds; "short-term and safe" is not the same as "central-bank reserves."

The most striking modern incarnation is the fully-reserved stablecoin. Strip away the cryptocurrency packaging and a stablecoin that holds one dollar of cash and short Treasury bills for every token it issues is, economically, a narrow bank: it takes in money, holds it in the safest assets, issues a claim that trades as money, and does not lend. The technology is new, but the structure is the ninety-year-old Chicago Plan idea wearing a digital coat. This is a recurring pattern you will see across the track — a genuinely old structural or legal innovation is what does the real work, and the new technology is the delivery vehicle, not the breakthrough.

Why the idea keeps getting blocked

Here is the puzzle at the center of the module. Narrow banking is old, simple, and demonstrably safer for depositors. Yet no pure narrow bank serves the public at scale, and the one that tried (TNB) was kept out. Why? The reasons are partly principled and partly self-interested, and an honest account includes both.

• Disintermediation — the serious objection. This is the real argument, and it must be taken seriously. The economy needs credit. Conventional banks fund their lending with deposits. If safe narrow-bank money were widely available, depositors might move their money out of conventional banks and into narrow ones — and conventional banks, losing their cheap deposit funding, would have less to lend. Safe money for individuals could mean less credit for the economy. The fusion that creates the problem is also what funds the lending the economy runs on, so unbundling it has a real cost.
• Monetary-policy plumbing. A narrow bank holding vast public deposits as central-bank reserves would change how the central bank operates and how it controls interest rates. Central banks worry, with some justification, about large untested shifts in the plumbing they use to run monetary policy.
• The master-account moat. As Module 06 of the classic track showed, a narrow bank needs an account at the central bank to hold reserves — and that gate can be, and was, kept shut. Here the anti-competitive structure of the classic track does direct work: the safer alternative is blocked not because it is unsafe but because it threatens the existing arrangement and depends on access the incumbents help control.
• Incumbent interest. Conventional banks have every reason to oppose a competitor offering safer money that could drain their cheap deposits, and they have the lobbying power to shape the rules on entry and access. Not all the resistance is high-minded.

The trade-off: safety versus the credit channel

The disintermediation objection deserves its own section, because confronting it honestly is what separates a serious treatment of narrow banking from a naive one. The objection says: safe money is good, but the economy needs maturity transformation — someone has to fund the long-term loans that build houses and factories — and conventional banks do that by lending out deposits. Strip the deposits away into narrow banks and you have made money safe at the cost of the credit the economy depends on.

The narrow-banking reply is not to deny the need for credit but to relocate it — to say that lending should be funded by people who knowingly take risk, not by depositors who think they hold safe money. In this vision, the system splits cleanly in two: narrow banks provide safe money and payments, holding only the safest assets; and separate lending institutions fund loans with money from investors who understand they are taking credit risk and could lose it — equity holders, bondholders, fund investors — rather than from depositors who believe their money is safe. The economist Laurence Kotlikoff pushed this to its logical end with "limited purpose banking," in which all lending is done through clearly-labelled investment funds and no leveraged deposit-taking institution exists at all.

Seen this way, the deep question narrow banking forces is not "safety or credit?" but something sharper: should the economy's lending be funded on the backs of money-users who do not know they are bearing the risk? The conventional system says yes, implicitly, and patches the resulting fragility with the safety net. The narrow-banking vision says no — make money genuinely safe, and fund credit transparently with risk capital from people who have chosen to bear it. Whether that transition is achievable without choking off credit during the change is the genuine open question, and reasonable people disagree. What is not in doubt is that the trade-off is real and the choice is a choice, not a law of nature.

The idea around the world

Narrow banking and its full-reserve cousins surface in different forms across countries, and the comparative picture shows both the idea's reach and the consistency of its rejection.

• Switzerland put it to a vote. In June 2018, Swiss citizens voted on the Vollgeld ("sovereign money") initiative, which would have required that only the central bank create money and that commercial-bank deposits be fully backed — a national referendum on something close to the Chicago Plan. Voters rejected it decisively, roughly three to one, persuaded in part by warnings about disrupting the credit system. It remains the clearest case of a country directly deciding whether to unbundle money from credit — and choosing the status quo.
• Postal savings as quasi-narrow money. As Module 01 noted, systems like Japan Post Bank historically held enormous public deposits while investing conservatively, mostly in government securities rather than commercial lending — a partial, state-backed approximation of narrow money that gave citizens a very safe place to keep funds.
• Money-market funds everywhere. The narrow-vehicle structure exists globally in money-market funds, and so does its weakness — regulators in the US, the EU, and elsewhere have repeatedly reformed them after runs, precisely because "safe-but-not-quite-narrow" vehicles can still break.
• The stablecoin frontier. Fully-reserved stablecoins are pushing the narrow-bank structure into digital money worldwide, often from outside the traditional banking system entirely — which is why regulators are now scrambling to decide whether to treat them as narrow banks, as something new, or to bring them inside the perimeter.

Assessment — and the seam to public money

Narrow banking is the cleanest answer to the deepest root cause, and it earns its place at the head of the innovation track for three reasons. It attacks the fusion of money and credit directly, by simply not performing it. It delivers run-proof, credit-risk-free digital money to the public — the missing cell conventional banking cannot fill. And it does so as a pure structural innovation, requiring no new technology, which makes it the sharpest demonstration of the course's thesis that the most powerful financial innovations are often not technological at all.

Yet it keeps not happening, and the reasons divide into a genuine trade-off and a defended moat. The genuine trade-off is disintermediation: a world of fully safe money forces an honest reckoning with how the economy's credit gets funded, and the transition is not costless. The defended moat is the master-account gate and the incumbent interest that kept TNB out and helped sink Vollgeld. The honest verdict is that narrow banking is safer for depositors and genuinely disruptive to the credit system, and that its repeated rejection reflects both a real concern and an entrenched interest — exactly the kind of mixed judgment Module 01 asked you to be ready to make.

Notice the recurring obstacle: every private path to narrow money runs into the trust-gate — a narrow bank needs the central bank's reserves and its account, which the central bank controls and can withhold. That points directly at the next module. If the cleanest safe money is central-bank money, and the obstacle is access to the central bank, then the most direct version of the idea is for the central bank itself to issue safe digital money to the public — bypassing the gate by removing the need for a private intermediary at all. That is a central-bank digital currency: the public version of the narrow bank, and the subject of Module 03.

Test your understanding

Six questions on narrow banking — what it is, why it is old, its modern forms, and the trade-off at its heart. The questions test the reasoning rather than recall of any single date.