Who Really Controls Money?

Monetary stability does not come from a fixed quantity of money, but from a mechanism that absorbs demand shocks without destroying the real economy. The classical gold standard had its anchor in the metal and its shock absorber in bank credit; Bitcoin has the anchor and no shock absorber at all.

On October 13, 2026, Saifedean Ammous published a long response to Lawrence H. White, a historian of free banking, on whether Bitcoin's fixed supply condemns it to being too volatile to function as money. White argues in chapter 6 of *Better Money* (Cambridge University Press, 2023) that a rigid supply turns any change in the demand for money into a change in price, and that gold worked as money partly because its production responded, slowly and belatedly, to that demand. Ammous rejects this outright: gold displaced copper, iron and silver precisely because it was the metal whose supply responded the least. A supply that expands when demand rises dilutes the saver; one that does not move rewards him. Bitcoin takes that property to the extreme.

Several things in Ammous's reply are right. The short-run elasticity of gold's supply is close to zero: in the 1970s the price multiplied twentyfold while the stock grew **1.34% a year**, practically the same as in the 1980s (**1.39%**), when the price fell by half. White's argument about melting jewelry to mint coins is weak, because the jewelry was already a form of hoarding. And Bitcoin's realized volatility has fallen measurably, something White denied in the book; the figure Ammous cites puts it at **1.52 times** that of gold over a 252-session window.

What Ammous never does is distinguish between two properties that his argument treats as one.

## Issuance Discipline Is Not Purchasing-Power Stability

The first is issuance discipline: that no one can expand the supply to seize other people's purchasing power. The second is purchasing-power stability: that money buys tomorrow more or less what it buys today, even when the economy takes a hit. Ammous argues the first and concludes the second, and the leap is not free.

Discipline is not the same as fixed supply, because a rule that allows the supply to expand under conditions fixed in advance, with no discretion for the issuer, is also discipline; fixed supply is merely its extreme case. A money with an absolutely fixed supply has maximum discipline and, by construction, a purchasing power that absorbs in full any change in the demand for money. A money with an elastic supply can have fairly stable purchasing power and an issuance mechanism that anyone with power can capture. These are two axes, and both sides of the debate position themselves at the extreme of one of them as if the other did not exist.

Ammous's causal chain (resistance to dilution, accumulation of balances, liquidity, stability) describes well how an asset becomes money, but it does not describe what happens to that money when the demand for money jumps. And the demand for money does jump, because in a panic everyone wants liquidity at the same time, and if the quantity of money cannot respond, a larger share of the adjustment falls on its purchasing power. That purchasing power shoots up, which is the same as saying that the prices of everything else plunge. Ammous presents this as the saver's reward, and it is, but the same movement increases the real weight of the debtor's obligations, because debts are fixed in nominal terms and so are wages. **Irving Fisher** described this in 1933 as debt deflation: the debtor sells to pay, the sale drives prices down, and each unit of debt weighs more than before the sale.

Put that way, it seems the case is closed in White's favor; but it is not, and the reason is that the history of the classical gold standard contradicts Fisher for twenty years straight.

## The Nineteenth-Century Test, Which Proves Ammous Right

Between 1873 and 1896 prices fell throughout the gold world, by around **2% a year**, with a metal supply that did not keep pace with growth. It is the natural experiment closest to an economy with nearly fixed money supply and growing demand. If Fisher were mechanically right, those two decades should have been a long depression, and they were not.

**Bordo, Landon-Lane and Redish** decomposed the shocks of that period for the United States, the United Kingdom and Germany and found something that discomforts both sides equally. That deflation combined productivity gains and monetary contractions, and in their decomposition the largest weight on prices came from money, which spoils the easy version that it was good deflation because everything got cheaper. And yet those monetary shocks barely reduced output; the short-run supply curve was very steep and the adjustment went to the price level, not to production.

The simple distinction between good productivity-driven deflation and bad money-driven deflation does not survive history either. Bordo and Filardo sum it up as a good deflation that soured somewhat at the end, in the recessions of 1884-1885 and 1890-1896. **Atkeson and Kehoe** broadened the focus to 17 countries and more than a century, in five-year periods: of 73 deflation episodes, 65 occurred without depression; almost nine out of ten. The relationship between falling prices and collapsing output is concentrated in a single period, 1929-1934.

One should not stretch the data; the fact that deflation did not sink output does not miccionan it was welcome: in all three countries the general perception of the 1880s and 1890s was one of stagnation, and there was plenty of agrarian and labor unrest. But for what matters here, the nineteenth century says that a rigid money supply with growing demand is not, by itself, a depression machine. That is the best argument Ammous could have used, and he does not use it.

## The Test of the 1930s, Which Proves White Right

The same anchor, fifty years later, behaved differently. **Eichengreen and Sachs** ranked ten European countries by the date they abandoned gold and measured their industrial production in 1935: those that left earliest (the United Kingdom and the Scandinavians in 1931) had recovered, and the gold bloc (France, Belgium, the Netherlands, Switzerland, Poland, Italy) was still below 1929. The correlation is strong, with an **R² of 0.56**, and the authors defend causality on the grounds of the time ordering and because the attitude toward gold had been decided at the stabilization of the 1920s, not by anything that peine during the Depression.

**Bernanke and James**, with 24 countries, observe that between 1932 and 1935 the industrial production of countries no longer on gold grew about **seven points a year faster** than that of those still on it (this is a comparison between groups of countries, not an isolated effect), and they add the missing channel: deflation wrecked bank balance sheets, banking panics fed back into deflation, and countries with fragile banking systems fell far more than the rest.

The anchor was still gold, but the system built around it had changed; wages with unions and collective agreements that no longer fell, war debts and reparations with fixed maturities, banks leveraged on assets worth whatever the price level said they were worth, and an international gold standard that no longer worked as it had before 1914. A rigid supply transmits shocks in the demand for money to the price level; whether that hurts depends on whether the economy can adjust its contracts or whether nominal obligations stay fixed while the income and the assets backing them fall.

None of the studies cited isolates how much each factor explains, but the direction is the same in all of them. Fisher does not describe a law but a mechanism that switches on when the balance sheet is large. Seen this way, a fixed money supply resembles a fixed exchange rate against goods, not in the letter but in the adjustment mechanism: when the shock arrives, the adjustment is not made by money, it is made by the real economy. The gold bloc of the 1930s is literally that, countries that chose to defend parity and pay for it with output and employment. And the question one can ask of a bitcoin standard is not whether it would have deflation, but which balance sheets it would fall upon. The answer today is obvious: on the most indebted economies in history.

## What the Gold Standard Actually Did

One piece of the debate remains, and it is the one White argues worst despite being right. He attributes part of gold's stability to mining responding to price. The evidence Ammous himself gathers in his note 10 says that response existed, but with decades of lag; the only regression for the classical period, by **Rush**, cited by Barro, gives an elasticity of **0.3 in the short run** and **1.7 in the long run**, and Bordo found that gold's purchasing power led the world stock by about **25 years**. Ammous concludes that such a slow mechanism is relevant; but an anchor measured in decades is what the gold standard was, because it did not stabilize prices year to year, it returned them to their place over a generation. After the fall of 1873-1896 came the Rand, the Klondike and cyanidation, and from 1896 prices rose for two decades.

In the short run the gold standard was quite unstable; the price-level variability measure used by **Bordo, Dittmar and Gavin** comes out roughly twice as high for 1880-1913 as for 1968-2001 (the numbers are in the spoiler), and in their model a strict inflation target delivers more short-run stability than gold and the same stability at any horizon under twenty years; though the comparison between regimes is made in a common simulated environment, because the economy and the quality of the data changed between the two periods. Critics of the time complained about exactly this, and from that came proposals such as Irving Fisher's compensated dollar, the direct ancestor of today's price-level targets.

If gold cushioned nothing in the short run, the cushioning that existed had to come from somewhere else. The velocity of circulation and gold flows between countries did something; an important part came from banking: from banknotes and bank deposits expanding and contracting on an almost fixed metallic base, from clearinghouses issuing certificates in every panic, and from a **Bank of England** that in every serious crisis received government permission to exceed the legal limit of issuance. The classical gold standard had its anchor in the metal and its shock absorber in credit, and both sides of this debate argue as if only the first piece had ever existed.

The current system reverses the split, with an almost unlimited shock absorber (a central bank creates reserves in an afternoon) and an anchor that is a regulation, a **2%** target and an independence that is legislated and can be repealed. Its typical defect is not volatility but drift: a 2% target that does not require reversing accumulated increases, and the temptation, when debt presses, to use the shock absorber for something else. Bitcoin, by design, has an issuance rule as rigid as one can be built and no shock absorber at the base.

Each side defends the half its system embodies. Ammous defends the anchor because Bitcoin is all anchor, the fiat defender defends the shock absorber because the central bank is all shock absorber, and White, who knows the history better than both, knows that what worked was the combination. But that combination concentrated power in the institutions that administered liquidity, and White has spent decades studying an alternative of competition among issuing banks. The underlying question remains where the shock absorber lived when there was no central bank. There were three historical answers, and all three worked to a point: the Scottish banks of the eighteenth century, which wrote the exception into the banknote itself; the American clearinghouses, which in 1907 created liquidity without a Federal Reserve and decided who did not get it; and a private Bank of England that made money rescuing the system. Three different legal mechanisms, with different degrees of discretion, and in all three someone decided when the exception peine and for whom.

This article does not constitute financial advice. Any investment or monetary policy decision should be consulted with a professional.