Every year or two, I realize I don’t actually understand what banks do and how they work. This happens reliably enough that I keep getting less confident that I, or anyone else, will ever really truly understand what a bank does.
But I’ve recently settled on a model that’s both novel and useful, that neatly synthesizes the two broad schools of thought. We can think of those schools as Hard Money and Bagehot.
- The Hard Money Theory is popular with a motley crew of Bitcoiners, Austrian economists, Ayn Rand’s imaginary pirates, and, oh yeah, 99% of people who ever lived before the twentieth century. The hard money theory holds that the supply of money should be as fixed as possible, and that economic growth should lead to deflation. Hard money is good at preventing inflation — inflationary episodes like Spain’s gold and silver imports from their colonies produced price increases on the order of 1% to 1.5% per year, hardly what we’d consider hyperinflation. However, hard money makes economic crises very severe, because when economic activity slows, liquidity leaves the system and it’s extremely hard to service debts.
- The Bagehot Model holds that central banks should be lenders of last resort. At first, this model was applied to countries with a gold standard, but that’s ultimately a Martingale Bet: if a central bank commits to providing other banks with as much capital as they need, eventually they’ll need more capital than the central bank has reserves, at which point the bank must print more money.
Most countries have defaulted to a Bagehotian model today, because available cash is a limiting factor for growth whenever real GDP per capita is rising. Hard Money was a simple default to choose when economic growth was slow and institutions were weak. If you gave governments the ability to produce arbitrary amounts of money, they’d, well, produce arbitrary amounts of money, and spend them on negative-sum activities like wars. One of the Western world’s first experiences with paper money was the Mississippi Bubble, which essentially financed a series of wars after the fact, and then went under.
When productivity growth and population growth are stagnant — or, equivalently, when we’re at a Malthusian level where population grows exactly as fast as a country’s ability to support people at a subsistence level — then money creation gets wasted. But when real GDP per capita growth is positive, that money finds its way to productive uses. In late 18th century England, there were lots of opportunities to make money and increase GDP per capita by building steam engines to pump water out of mines, or by building factories, so money creation led to investment. As the industrial revolution really got into gear, this became increasingly true.
The Second World War was a catalyst for finally putting old monetary ideas out to pasture. It was the first war in which it made sense to reverse the cliche: not “men and matériel” but “matériel and men.” The Axis’ major strategic blunders (bombing Pearl Harbor and invading the USSR) were both gambles to secure oil supplies. Governments on both sides took extreme financial measures: rationing and severe financial repression among the allies, total war for the Axis (Nazi Germany spent 75% of its GDP on the military in 1944). In the first world war, governments had been stretched to their limits to mobilize populations, but in the second they stretched to their limits to mobilize stuff. The US, for example, had 1,700 military aircraft at the start of 1939, and built around 300,000 by the end of the war.
It’s impossible to finance this level of mobilization by conventional means; it’s certainly not possible to raise taxes fast enough to finance a war effort without running a deficit. So, by the end of the war, every country, including the victors, was in a fiscal situation that would be untenable under a gold-based system.
1945 wasn’t exactly when we switched, but it was the point of no return; from then on, the combination of high war debts and ample investment opportunities meant that flexible money was the way to go.
The Bagehot model enjoys overwhelming support right now, but it doesn’t line up with simple intuitions about how banks work. The Hard Money model fits what a small child thinks a bank is: it’s a building where you store cash. The Bagehot Model lets us do lots of more interesting things with our banking system: since there’s a lender of last resort, banks don’t have to worry quite so much about literal money shortages, but the lender of last resort has to calibrate the hypothetical terms of those loans in order to encourage the right level of growth.
In Econ 101, you might learn that the way a money works is that governments print money, then banks accepts deposits and lend them out. That’s true in an accounting sense, but not in a real sense. What really happens is that banks create money: if a bank has $200 in deposits and lends $100 of it to you, you deposit that money in a bank; now, the system has another $100 in deposits.
This seems not only fake, but slightly more fake than the model where governments create money. A government, at least, can compel you to use its currency for some transactions. They create money out of nothing, but a bank creates money out of even less than that.
What’s going on?
One of the common themes of history is that we reinvent the forms of institutions but preserve the function. In higher education, Title IX is a patchy, gradual reimplementation of the in loco parentis standard. In US politics, gerrymandering effectively cancelled out the 17th amendment; now Senators are elected by popular vote, while House members are appointed by state legislators by way of congressional districts that all look like a Rorschach tests or poorly-copied Kama Sutra diagrams. Literally the oldest story in the Western canon, the Iliad, opens with everyone complaining that they’re stuck fighting a war in the Middle East and it’s way harder than they thought.
And banks are just reinventing the private mint. When money is backed by physical specie because it’s literally made of specie, governments don’t actually have to control the money supply. Nature does it for them, by putting valuable metals underground where they’re hard to find. So it’s fine for people who own silver mines to dig up silver, weigh it, stamp it with a local ruler’s face, maybe mill the edges (so nobody is tempted to clip bits of metal off the sides of the coin in order to take a rake on a transaction), and then exchange it for goods and services.
Some of these entrepreneurs achieved the great distinction of making a product so popular it became a generic term: like Google, Kleenex, and Xerox, “Dollar” was originally a brand name — an abbreviation for Joachimsthaler, a silver join minted in Joachimsthal. (It’s a weird linguistic coincidence “dollar” as a word for money and “Neanderthal” as a word for cave man are both derived from the same root, “thal,” which just means “valley.”)
It’s not risky for someone to run a private mint under a hard money regime; if they dilute their coins, the coins lose value in the market, and that’s directly reflected in exchange rates. And diluting the metal content of a coin over time is even worse, since it means all of your coins, old and new, are less trustworthy, so you irritate all of your older customers.
If silver is money you can spend, a mint is just in the business of finding it, weighing it, and slicing it into convenient increments. They’re not so much “making money” as they are making money legible.
Turning back to modern banks: they don’t just make loans at random. They make loans when there’s a reasonable prospect of repayment. An incremental loan, then, represents incremental certainty about the future. Since our collective wealth is the value of all the goods and services we’ve created plus the present value of all the future goods and services we’re likely to create, this money-creation is an elegant way to ensure that the supply of money always matches the total amount of wealth the world has.
Maturity Transformation, Seigniorage, and Bad Debt
The strongest Hard Money critique of the Bagehotian paradigm is the problem of maturity transformation. It goes like this: a bank owns lots of loans that come due in the future. The bank also has deposits, many of which are demand deposits that can be converted into cash right now. If a bank transforms demand deposits in the liability column into long-term debt on the asset side, they’re basically making a promise they can’t keep. They simply don’t have the means to pay back their debt, and must constantly roll it over.
In the bank-as-mint model, that’s not so pathological, because those debts can be paid in the aggregate even if they can’t be paid by that specific bank. It’s not fraudulent to use expectations about the future to shift wealth to the present, it’s just analogous to the way that planes make New York closer to London than it was a hundred years ago, and Instagram makes them closer still. The system can face hiccups if a bank runs into a funding crisis, but sufficient money-printing can keep the system running.
So long as there’s a Bagehotian central bank to step in and help roll over debts, the system works.
At least, it works as long as the loans are good.
What happens if a bank makes a poor assessment of a borrower’s prospects? The bank loses money. They expect to do this, some of the time, which is why your credit card has a higher interest rate than a T-Bill. But if a bank systematically underestimates the probability of default or the value of recovery, it’s doing something else: it’s minting coins whose face value is higher than the value of the metal. Bad loans are just a form of seigniorage.
The upshot of the bank-as-mint model is that we should be less skeptical of growth in the financial industry and growth in total loans outstanding, but much, much stricter in apportioning liability to shareholders and employees of banks that make bad decisions. We don’t have to roll things back to the days of Joachimsthalers and gold-backed money, when counterfeiters stopped briefly in jail before visiting Tyburn, but we can feel good about clawing back comp and cramming down shareholders in the next crisis.
It might slow economic growth down a bit in the short term if we adopted a more punitive attitude to collapsed financial institutions, but that, too, is a reflection of reality rather than a policy decision. If banks create currency, then excess lending debases currency, and debased currency leads to cycles of bubble and collapse. If banks tend to create money based on more certainty than is warranted by the facts, we’re just trading economic growth in the short term for crises later on, and while it’s hard to call a bubble at the time, it’s comparatively easy to set up a legal standard that puts the onus of prevention on the people best positioned to do so.