The Economist has an article on central bank digital currencies (CBDCs):

If citizens can convert bank deposits into central-bank money with a simple swipe, the technology “has the potential to be run-accelerant,” said Lael Brainard, a Federal Reserve governor, in 2019. This could pull deposits out of the banking system and onto the central bank’s balance-sheet, disintermediating the banks. . . .

If CBDCs proved popular, they could suck all deposits out of the banking system. In America this would stretch the central bank’s balance-sheet from $8trn to a whopping $21.5trn. Who, then, would provide the $15trn of loans that banks now extend to the American economy?Perhaps a central bank could simply pass the funds back to the banks by lending at its policy interest rate. But it is hard to see the idea of the Federal Reserve extending trillion-dollar loans to the likes of JPMorgan Chase or Bank of America as being politically uncontroversial.

I’m no expert on digital currencies, so I encourage commenters to correct me.  But I suspect that the fundamental problem here is not digital currencies, rather it is digital currencies combined with zero interest rates.

Let’s step back from digital currency for a moment, and think about dollar bills.  Why don’t people hold all their wealth in the form of ultra-liquid cash?  There are two reasons.  First, alternative investments might offer a higher rate of return.  Second, even if alternative safe investments pay zero interest (as is the case today), alternative investments are less risky than currency notes, because large cash holdings can be stolen, lost in a fire, or attract the attention of government money laundering investigators.

Imagine there were digital currencies in the 1990s, when safe investments such as T-bills and CDs paid 5% interest.  In that case, the demand for zero interest digital currency would be relatively modest, as alternative investments would pay a much higher rate of return.   Today, risk-free interest rates are zero, but people continue to hold much more wealth in bank deposits than in currency, because physical cash is inconvenient.

But what if interest rates fell to zero and a perfectly safe form of cash was offered by the central bank—a digital currency?  Might savers prefer digital cash to zero interest bank accounts?  That seems possible, and that’s presumably the scenario the Economist worries about in the quotation above.

This hypothetical relates to a fundamental problem with our entire financial system; we have socialized risk in such a way that savers are shielded from the consequences that result when their funds are misused.  Imagine the following (highly simplified) data for the US:

1. Treasury bonds = $20 trillion

2. FDIC insured bank accounts = $20 trillion

3. Bank loans = $20 trillion

Of course these numbers are not exact, but they demonstrate a basic problem with our financial system.  In this example, the economy has $20 trillion in safe assets and $20 trillion in risky assets.  But from the perspective of savers, it looks like we have $40 trillion in safe assets–the T-bonds plus the FDIC-insured bank accounts.  While banks do take risks with depositor funds, the depositors are shield from those risks by FDIC.  The bank accounts are privately safe but socially risky.

Now suppose that a digital currency is offered and savers move all $20 trillion from bank deposits to digital currency.  On the central bank balance sheet the digital currency “liabilities” will be “backed” by $20 trillion in T-bonds.  In other words, the central bank will have to buy up $20 trillion in T-bonds to match the growth in the liability side of its balance sheet as people hold more digital currency.  The $40 trillion in safe assets falls to $20 trillion.  Unless . . .

Presumably, the people who sold the $20 trillion in T-bonds to the Fed will need another place to put their funds.  And since those “other places” will be more risky, they will have to offer a higher rate of interest to savers.  Suppose that in order to keep money in the banking system, savers demand 1% higher interest on bank accounts than on digital currency (which pays zero interest by assumption.)  In that case the financial system will look like the following:

1. Digital currency (backed by Treasury bonds) = $20 trillion

2. FDIC insured bank accounts paying 1% interest = $20 trillion

3. Bank loans charging an extra 1% interest = $20 trillion

So as long as there are enough safe assets like T-bonds (or mortgage backed securities backed by the Treasury) to back the newly issued digital currency, the move to digital currency doesn’t actually destroy the banking system.  But it does make banking more costly (and thus might shrink the system.)  Because digital currencies are an attractive substitute for bank deposits, banks must pay depositors a higher interest rate, and those extra costs get passed along to borrowers.  Recall the complaints from the real estate industry that reforming the GSEs might cause mortgage rates to rise by a few basis points—that’s the type of issue we are looking at here.

At least I think that’s the issue; please correct me if I am wrong.  And if there are not enough safe assets to back digital currency, then the central bank might have to fund private banks, the nightmare scenario involving JP Morgan discussed in the quotation above.

One advantage of the second system (with $20 trillion in digital currency) is that it might be a bit easier to abolish government deposit insurance.  People would have a safe place to put their saving, which would be much more convenient than holding Treasury bonds.  I say “a bit” easier, as the banking industry would continue to strongly oppose the abolition of FDIC.

At a still deeper level this example exposes a quirk in our monetary system that most economists probably don’t think about.  Prior to 1913, all base money was cash (currency and coins).  Everyone could hold any type of base money they wanted to hold.  There was no law saying that only blue-eyed people could hold $10 bills.  This all changed after 1913.  A new type of base money was created—deposits at the Fed.  But only banks (and a few other institutions such as GSEs) were allowed to hold deposits at the Fed.  This may or may not have been a good idea, but it’s a very strange system.  Digital currencies threaten to upend that system.  Holding central bank digital currency is essentially like holding a deposit at the Fed.

It seems to me that if we issue a central bank digital currency, we’d want to have a system where negative interest rates are possible in times of stress, to prevent a massive flow of funds from the banking system into digital currencies.  I don’t know if that’s technically feasible.  No doubt the technological revolution in money will raise many other difficult questions.

PS.  I wonder if younger economists are familiar with Eugene Fama’s brilliant 1980 article entitled “Banking in the Theory of Finance“, which indirectly relates to this post.  I strongly recommend this article; it’s the one where Fama suggests replacing base money with a new unit of account, reserve balances that entitle one to own a spaceship.