A friend of mine has spent years metal detecting around this part of upstate New York, and occasionally he'll show me something that came out of a field we've probably driven past hundreds of times.
Some of the finds are ordinary enough: old copper coins, buttons, bits of hardware and tokens from businesses nobody remembers. Others reach surprisingly far back. One of my favorites is a Spanish silver two reales from the eighteenth century, roughly the diameter of a modern quarter.
It's difficult to hold an old coin without wondering what it once bought.
That's also where things become complicated, because translating money across centuries isn't as simple as converting one number into another. The coin contains a measurable amount of silver, but that doesn't tell us exactly what the silver meant to the person who carried it. Wages were different. Goods were produced differently. Some things that are cheap today were enormously expensive, while things we now spend considerable money on didn't exist at all.
Even so, the temptation to find a conversion is strong.
We want to know what the coin was really worth.
I've become increasingly suspicious of the question, or at least of the assumption hiding inside it. It suggests that somewhere underneath all the changing prices there must be a stable measure that will let us compare 1776 with 2026.
Dollars won't do it. A dollar today clearly doesn't buy what a dollar bought generations ago.
Silver won't quite do it either.
Neither will gold.
That doesn't make any of those measures useless. It means we're trying to measure a moving economy from inside the economy.
For long periods, gold and silver played a much more direct monetary role than they do in ordinary American commerce today. Spanish silver circulated widely in early America, and the monetary system that developed in the United States was strongly connected to existing silver coinage. Later monetary arrangements changed repeatedly, eventually leading to the Bretton Woods system, under which foreign governments and central banks could exchange dollars for gold at an official price.
That arrangement ended in 1971 when the United States stopped converting foreign official dollar holdings into gold.
It's tempting to draw a straight line there: before 1971 money represented something real, and afterward it became an abstraction.
History is messier than that.
Metallic monetary systems had inflation, deflation, financial crises, changing exchange ratios, shortages of coin and political arguments about what counted as money. Gold and silver themselves changed value relative to the things people needed.
What changed wasn't that money stopped having value.
The mechanism by which the monetary unit was defined changed.
That distinction matters because the dollar still works remarkably well for one of the things I ask money to do. If I buy groceries this afternoon, I don't need to calculate the price in ounces of silver, hours of labor or bushels of wheat. Dollars give everyone involved a common unit in which to quote prices and settle the transaction.
The trouble appears when I ask the same unit to carry information across long periods of time.
Suppose someone tells me that a house sold for $20,000 many decades ago.
I know the number.
I still don't know what it meant.
How much did the buyer earn? What interest rate did they pay? How large was the house? What did land cost in that particular place? How expensive were food, transportation and energy? How many years of work did the purchase represent?
The nominal price is real information, but it isn't enough information.
Economists have a term for one version of the confusion that follows: money illusion, our tendency to respond to nominal amounts without fully accounting for changes in purchasing power. A raise in the number printed on a paycheck, for example, doesn't necessarily mean someone can buy more with the paycheck.
Once you notice this, historical prices bec