Money is one of those surprisingly common domains where practice has generally outstripped theory, but it’s not the only one. There are actually a lot of them: The first caveman to throw a rock at an animal he was hunting was, in some sense, solving a calculus problem in his head even if he would have been befuddled by the concept of a meter-per-second-squared. Evolution gave us sweat glands even though the atomic theory-related reason sweat cools us down would take a lot longer to discover. And we've been trading goods and services for tokens that represent some kind of value for a long time without asking what it all means.

The best way to think about money is that it's a yardstick that emerges as a Schelling Point. It's possible to have fruitful, positive-sum exchanges through barter, and it's tempting to imagine a barter economy slowly developing into a monetary one—there might be some category of product that's easy to store, easy to divide, hard to consume, and naturally scarce, etc. But that implicitly assumes modern norms of work, social relationships, and exchange. If your social system has overlapping, reciprocal obligations, and mostly consists of repeat interactions with the same set of people over an entire lifetime, in a system of technological stasis where there's almost no capital to accumulate, there just aren't many times when someone wants to exchange X chickens for Y cows and settles on some intermediate quantity Z—either there isn't an exchange at all, or the exchange that happens is part of some social ritual that's only loosely tied to its underlying economics.

That model works, but it doesn't scale for two reasons: first, it means that getting access to goods and services entails being an established part of some community, so travel or solo immigration are a challenge. More importantly, early states have a hard time assessing and collecting taxes if they're paid in kind, especially if most of the people they're taxing live at a subsistence level, and a minority who don't spend much of their surplus on 1) ritualistically re-establishing their high status, often by giving away or destroying valuable goods, and 2) acquiring substitutes for this legitimacy in the form of various means of committing and withstanding violence.

Monetary transactions make more sense in a world where there's some abstract force providing security; the fact that money is more convenient to carry around than a farm also means that in a monetary economy, it's a much better idea to rob people. And this introduces a feedback loop that keeps making money more abstract, and more disconnected from its origins.

Some early monetary systems used assorted rare objects—cowrie shells, gold, silver.¹ But over time, precious metal-based economies tended to develop a two-tier system, with internal commerce conducted in units of currency that were backed by precious metal reserves, and global trade settled in those reserves themselves. This creates an admirable balancing mechanism—when a country issues too much paper currency, there's an incentive to redeem it for gold, and the only way to offset that is to get a positive trade balance, i.e. to weaken the economy to the point that people can't afford to buy many imported goods and can't afford not to work hard in export-focused industries. This leads to brutal recessions, but short ones.

What this eventually means is that the literal precious metal backing matters less than managing the interface between paper currency and gold. And that creates the necessary financial infrastructure to cut the tether entirely, and switch to a purely fiat system. It's an awkward shift for someone who's used to using currency redeemable for precious metals, largely because currency in the first half of the twentieth century often says something to the effect that the physical bill in question is redeemable for five dollars, or that it represents five dollars in silver payable to the bearer on demand. (Both images via this page.) But functionally, this $5 bill was not worth $5 because it was "redeemable," or "payable on demand." It was just $5. Still, it was a shock to break the silver connection in 1968 or the gold connection in 1971.

There are two things that made those bills continue to be valuable currency rather than a neat-looking design drawn on a cotton-and-linen canvas. First, they remained legal tender for debts. And second, the IRS would periodically request a certain amount of them, and it behooved the recipient of that request to comply.

Another explanation for the end of backing currencies with precious metals is that countries accumulated too much debt, and could no longer make good on their obligations. But it's equally accurate to say that as the world accumulated more debt, the redeemability of a given note become superfluous: if you have a taxable income or a mortgage, if you're a company with accounts payable and bank loans, if you're a foreign entity that exports goods to the US and intends to import some later on—you need dollars, lots of dollars, and you just don't care what they're backed by.

This creates another interesting dynamic, though. It means that in practice, spending power is a function of the amount of credit in the economy, and only incidentally the amount of hard currency in circulation. That creates two kinds of novel risks: first, excessive credit creation, where the increase in credit leads to higher spending, which makes dubious loans pay off better, at least for a while. The second problem is when this stops working, and the level of credit contracts too fast. If you go back and look at asset prices in late 2008 and early 2009, you'll see plenty of screaming bargains—talk to anyone who was investing back then and they'll say they were perfectly aware of how cheap everything was, but there wasn't any liquidity available to take advantage of it. 

And that's where the money-printing comes in: central banks can create more credit, and have some clumsier tools for ensuring that this credit actually gets used. It leads to some ominous graphs of the money supply during crises, but that's mostly a statistical artifact—if most of what counts as money in practice isn't directly tracked by monetary aggregates, those aggregates just represent how much additional more-legible money has been injected to keep things even. It does lead to problems, but roughly symmetrically: after 2008, the US could have handled larger government deficits financed by central banks (our evidence for this is that as deficit spending ramped up in the late 2010s, inflation didn't take off). But that doesn't mean there's an unlimited capacity for spending and money-creation, as the 2021 inflationary episode demonstrates.

Money is as slippery a concept as ever. It may always be the case that our theories lag practice, and that the intuitive approaches work even if there isn't much intellectual rigor behind them. In the end, money is simply a way for strangers to transact with one another—if there's some token that both sides treat as representing fair value for a transaction, it's as money-like as it's ever going to need to be.

Read More in The Diff

The Diff is a newsletter about finance and technology, and money is both a platform for financial technology and a financial technology itself. It comes up a lot, in many contexts:

• One of the functions of modern central banks is to turn real estate into a more liquid asset—in short, to monetize land ($).

• We’ve looked at the post-Bretton Woods system, which has lasted longer than Bretton Woods itself ($). (It’s striking that the ad hoc compromise made in an emergency turned out to be more stable than Bretton Woods’ clean-slate redesign, which was designed to be implemented after a crisis.)

• Modern payment platforms like Stripe create a sort of solid-state economics.

• The dollar’s dominance has a distorting effect on US policy and the global economy ($), but it’s hard to imagine the next-best option being any better.

1. An interesting contrast here is that in a small, culturally homogeneous group, you can use something abundant and easy to counterfeit as currency, because your peers will all call you out on it you end up being mysteriously rich. In a more complex economy, with more external trade, you have to worry about counterfeiting both among the locals and among people who trade with them. So classical Sparta, a mostly agrarian and mostly autarkic society with strong norms against enjoying material wealth could use iron for currency—their norms meant that even if you did forge some fake money, there wasn't much for you to buy. The Athenians, a trade-based empire with colonies and vassals, used silver. To counterfeit it, you have to find some or buy some, which means you're either creating some value or exchanging previously-created value.

Join the discussion!