Confidence in a currency can drain for years. When it finally goes, it goes all at once. Weimar, the 1933 gold order, and 1971 are three versions of what happens next. Each one ended with a bill, and with a particular kind of person paying it.
Every currency is a promise that the money you hold today will buy roughly the same tomorrow. Confidence in that promise holds until it snaps. The reason is simple game theory. The moment enough people suspect the promise will be broken, the winning move is to get out before everyone else, and everyone knows everyone else is doing the same maths. Governments know it too. So the break arrives as a done deed, not a proposal you get to debate. It lands overnight or over a weekend, with the terms already changed. A decree, a banking holiday [banks ordered shut so no one can take money out], a “temporary” suspension.
And underneath, it’s the same move, time after time. When the promise can’t be kept, the issuer rewrites what the promise means rather than defaulting openly. It keeps the upside of the rewrite and leaves the loss with whoever was still holding the old promise. Weimar, 1933, and 1971 are that one move at three intensities.
Why the state ends up cornered
The three episodes look like three separate accidents until you see what they share. In our system, money is loaned into existence [new money is created when banks lend, so most money is someone’s debt], so the whole structure needs prices and incomes to keep rising. The debts are fixed in money terms, and falling prices make them heavier. When losses appear, they get passed upward instead of taken. The firm is rescued by the bank, the bank by the state, and the state’s own promise, the currency, is the last balance sheet [the list of what it owns and owes] in the chain. Suppressing every small failure works like suppressing every small forest fire. The fuel builds, and the fire you finally get burns the money itself.
By the time confidence breaks, the loss already exists. The wealth was spent or misallocated years earlier. The break doesn’t create that loss. It decides who carries it.
Weimar, 1921 to 1923: the promise is destroyed outright
Germany came out of the war owing enormous sums. The cruellest part was that reparations were set in gold marks and foreign currency. Printing paper marks couldn’t shrink that debt, because the debt wasn’t in paper marks. The state printed anyway, to pay wages and obligations, to buy the foreign currency it owed, and eventually to pay striking workers in the Ruhr after France and Belgium occupied it. Each round of printing pushed prices up faster. People learned to spend marks the hour they received them, because holding money was the losing move. That’s a crack-up boom [a late-stage rush out of money into anything real, as confidence in the unit erodes]. By November 1923 the unit was dead.
Almost every promise written in marks died together, from savings accounts and pensions to insurance policies and the war bonds people had bought patriotically. That included the state’s own internal war debt, quietly wiped. Printing is a default on everything denominated in your own unit, your citizens’ savings included. Meanwhile, anyone who owned real things with mark debts against them, factories, land, buildings, walked out the other side with the greater part of those debts gone. Then came the reset [replacing a failing currency with a fresh promise under new terms], a new unit called the Rentenmark. The gold-denominated external debt was restructured by negotiation under the Dawes Plan in 1924. The savers were never made whole. And the damage went further than money, because a wrecked middle class went looking for someone to blame and a strongman to fix it, and got both.
1933: the promise is rewritten mid-contract
The American version came from the opposite direction. Prices were falling, and not because anyone had got better at making things. Dollars were claims on gold at $20.67 an ounce. After 1929, prices and wages fell hard while debts stayed fixed, so every debt took a bigger bite of every shrinking income. That’s debt deflation [falling prices making fixed debts heavier until defaults cascade]. The government wanted to create money and push prices back up. The gold peg blocked it, because printing more claims against the same gold invites a run on the vaults.
In April 1933 the state ordered citizens to hand in their monetary gold at the old price of $20.67 and made refusing a criminal offence. Congress also cancelled the gold clauses in contracts, the clauses savers and lenders had written in to protect themselves from this exact manoeuvre. Gold came in, out of bank vaults and out of household drawers. Then, in 1934, with that gold in official hands, the government repriced it at $35 an ounce. The dollar was worth roughly 40% less in gold, and the profit went to the Treasury, not to the people who had complied. The Supreme Court let the cancelled clauses stand in 1935, a year after the profit was banked. Britain, by the way, had run its own milder version in 1931, simply leaving the gold standard and letting the pound fall by about a quarter.
Gold didn’t fail as a scarce thing. Its purchasing power was fine. Gold failed as a system, because using it for a whole economy meant vaults, custodians, and paper claims, and whoever controls the custody can change the rules in an afternoon. The seizure was cheap because the metal was already centralised.
1971: the promise is withdrawn from the last claimants
Bretton Woods rebuilt the world’s money in 1944 with the dollar at the hub. Other currencies pegged to the dollar, and the dollar to gold at $35. But after 1933 only foreign governments held that conversion right. Citizens were already out. Through the 1960s the US created dollars for the Vietnam War and programmes at home, until dollar claims abroad were several times the gold available to honour them. Foreign governments did the arithmetic and began redeeming, France most famously. On 15 August 1971, rather than let the vault drain, Nixon suspended convertibility [the right to swap the paper claim for the underlying gold]. It was announced as temporary. It was never restored.
Every dollar on earth, including the reserves of other nations’ central banks, became a promise with nothing behind it. The packaging fits the pattern. The same speech imposed a 90-day freeze on wages and prices and a surcharge on imports, all framed as defending citizens against speculators. The loss was then spread across a decade. The inflation of the 1970s repaid the world’s dollar holders in units that bought less each year. That’s an implicit default [reducing what a debt is really worth by devaluing the money, rather than missing a payment]. Much of that loss was pushed onto foreigners, who held the promise but had no vote on its terms.
Then came the second half of the move. It’s why 1971 never looked like Weimar. With the gold link gone, the US made sure the world’s oil stayed priced and settled in dollars. Any country that needed oil still needed dollars. The petrodollar arrangement extended the dollar’s life without ever restoring the promise.
The same move, seen once
Strip the period detail away and every episode runs the same sequence. Promises are made that the money coming in can’t cover. The gap is hidden with money creation for as long as the rules allow. Where the rules block that, as the gold peg did in 1933, the pressure comes out as falling prices and defaults instead, and the rule itself becomes the thing that has to give. Losses are passed upward until the currency is the only balance sheet left. Then, instead of open default, the money itself is redefined, by printing until the unit dies and a new one takes its place, by seizing and repricing the collateral, or by severing the link entirely. The redefinition arrives suddenly, because announcing it in advance causes the very run it’s meant to prevent. It comes wrapped in emergency powers and a story about necessity, usually with an enemy attached, speculators, hoarders, foreigners. Afterwards, new rules restore calm, and the cycle begins again under the new promise.
This isn’t ancient history either. Malawi’s kwacha was devalued by roughly 44% overnight in 2023, sprung on people the same day. A national supermarket chain shut for a day just to relabel the goods. Nobody got a 44% pay rise to match. On the long-run chart of a managed currency [a currency whose exchange rate the state sets] the fall comes in cliffs, not a slope. And every cliff is the same move, smaller.
So, who absorbs the loss?
Whoever is still holding the promise when the terms change, and can’t exit.
That means savers in cash and bank deposits, pensioners, bondholders, and people who live on wages, because wages adjust slowly, so even if your pay packet is untouched, what it buys falls. It means the rule-followers, the Americans who handed in gold at $20.67, the lenders whose protective contract clauses were cancelled. It means the people furthest from the source of new money, because prices reach them before the money does, while those closest to the source, governments, banks, leveraged asset owners, get the money before prices adjust. And when the money is a reserve currency [the currency widely held by central banks and used for global trade], it means foreigners, because reserve status puts much of the promise in foreign hands, so that’s where much of the loss is assigned.
The indebted don’t absorb it, so long as the debt is written in the money being rewritten. The biggest such debtor is the state itself, and its obligations in its own unit shrink or vanish. Owners of real assets carried on borrowed money come through enriched. So inflation and resets are a transfer with a direction, from holders of promises to issuers of promises.
And there’s a second loss the ledger doesn’t show. Trust itself gets consumed. The zeros on the banknotes were the cheap part of Weimar. What cost more was a society that stopped believing the rules were fair and reached for scapegoats and a strongman. When the money breaks, division and control rise together.
The four best arguments against me
“Modern institutions would prevent all this now.”
I’d answer with 1933. It happened inside a strong constitutional order, with courts and a free press, and it went through anyway. The contracts were still cancelled, and the courts still upheld it. 1971 ran through that same order too, where a single speech imposed a 90-day freeze on wages and prices. When the survival of the money system is at stake, law bends to money, not money to law. Institutions are only as strong as the incentives pressing on them.
“Devaluation is a legitimate adjustment tool, not theft.”
It’s a mass pay cut and a levy on savings, imposed in a day, without consent. Describe it for what it is and put it to a vote. Would it pass? The mechanism only works because nobody is asked.
“Gold worked for centuries, so surely the metal wasn’t the problem.”
Agreed, and that’s the point. Scarcity never failed. Custody failed. Any money that has to be warehoused and used through claims ends up governed by whoever holds the warehouse.
“1971 wasn’t really a break, since the dollar is still here.”
The promise broke completely. What survived was demand for the unit, re-anchored to oil and to the lack of anywhere else to go. Spreading a loss thinly across the world and across a decade doesn’t cancel it. Delay isn’t repair.
Why this history carries the thesis
The loss always lands on whoever holds the promise and can’t exit. So every version of the move depended on people not being able to get out. Sometimes the exits were locked, by gold in vaults, banking holidays, capital controls [restrictions on moving money across borders], and cancelled clauses. Sometimes what people held had no exit in the first place, like a pension, an insurance policy, or a war bond. And today’s system is the post-1971 arrangement running at full stretch, roughly 185 trillion dollars of new debt in the two decades before the pandemic to buy about 46 trillion of growth. The debt is now too large for the sharp Volcker-style cure [the early-1980s fix, when the US central bank raised rates hard enough to crush inflation], so the loss is being assigned quietly instead, through financial repression [holding the return on savings below inflation so that debts melt gradually at savers’ expense]. That is the drain, not the break. It is the same implicit default as the 1970s, running ahead of the rewrite this time instead of after it. If markets push back, the central bank can cap borrowing costs directly, buying bonds to hold rates where it needs them.
For the first time, both kinds of trap have a structural counter. A money you can hold yourself, without a vault, and carry across a border in memory, is not a claim on anyone. There is no custodian to hand it to and no issuer who can rewrite what it means. That removes the choke point every previous version of the move relied on. The move can still be tried. What’s changed is that you no longer have to be holding the promise when the terms change.