Pathologies · Chapter 6

What happens when money dies?

A currency that dies, dies fiscally: a state that can pay only by printing. German wholesale prices rose about 3.8 billion times in the 15 months after August 1922; Venezuelan consumer prices rose about 126,000 times in the 24 months after December 2017, on the central bank’s later-published series. The losses do not land at random. They land on wages, on cash, and on the loans owed to you.

In this chapter

It arrives, when it arrives, as a news item about somewhere else. A banknote with too many zeros on it, held up to a camera. A queue outside a bank in a city you would struggle to place. A shopkeeper in a country you have never been to, repricing the shelves at lunchtime because the morning’s numbers have stopped being true. It looks like chaos, and it looks foreign, and the question underneath the watching is a private one that nobody asks out loud: could the money in my account do that, and if it did, what would actually happen to me? This chapter answers the second half, because the answer is more specific than most people expect. A currency’s death is not a random calamity that falls on everyone alike. It has a cause that repeats, a shape that repeats, and a distribution of losses that can be read off the legal record afterwards, name by name.

Start with what the dying looks like when it is counted rather than photographed.

The shape of the thing

Economists have a threshold for this. In 1956 Phillip Cagan proposed calling an episode a hyperinflation when prices rise by at least 50 percent in a month, and the line has stuck. Fifty percent a month is not ordinary inflation gone further. Sustained for a year it multiplies prices by about 130, and the episodes that pass the line do not stop there. Figure 6.1 puts three of them on one axis, each measured from the month it first crossed Cagan’s threshold, with prices shown as a multiple of what they were in that month on a scale where every step up the side is a factor of a thousand.

THREE CURRENCIES DYING, EACH TIMED FROM ITS OWN ONSET prices as a multiple of their level in the onset month; log scale ×1 ×103 ×106 ×109 ×1012 ×1015 ×1018 ×1021 ×1024 ×1027 official CPI ends here, July 2008 ZIMBABWE estimate, 14 Nov 2008 GERMANY Nov 1923 VENEZUELA Dec 2019 0 3 6 9 12 15 18 21 24 months since the episode first passed 50% inflation in a month (Cagan’s threshold) THREE DIFFERENT INDICES, NOT ONE SERIES Germany 1922–23: wholesale price index, Statistisches Reichsamt (via Sargent) Zimbabwe 2007–08: Reserve Bank CPI (solid), then Hanke & Kwok estimate (dashed) Venezuela 2017–19: consumer price index (INPC), Banco Central de Venezuela
Figure 6.1 Three currencies dying, each clock started at the month that episode first passed 50 percent inflation, and each line showing prices as a multiple of their level in that month. Every gridline is a thousandfold, which is the only way the three fit on one page: German wholesale prices multiplied about 3.8 billion times in 15 months, and Venezuela’s about 126,000 times in 24, which on this scale is the nearly flat line at the bottom. The three are not the same measurement and the legend names each one: a wholesale index for Germany, so it is not a cost-of-living figure; a consumer index for Venezuela, published by its central bank after a silence from early 2016 until 2019, when the missing years were filled in, so this is the backfilled series rather than what anyone could see at the time; and for Zimbabwe two different things joined at the open circle. Up to July 2008 the solid line is the Reserve Bank’s own reported inflation, chained from the monthly rates it published for these months. After that month the Reserve Bank published no further inflation figures, and the dashed line is not a measurement at all: it is Steve Hanke and Alex Kwok’s estimate, worked out from the price of a single share traded in both Harare and London, which is why it is drawn differently. Their estimate for 14 November 2008 is 79.6 billion percent in that month. Germany’s line stops in November 1923, the month prices stopped rising; Venezuela’s stops at 24 months by a convention chosen here, because that episode faded rather than ending at an event. Germany: Statistisches Reichsamt wholesale prices as printed in Thomas J. Sargent, “The Ends of Four Big Inflations,” Table G1, in R. E. Hall (ed.), Inflation: Causes and Effects, NBER/University of Chicago Press, 1982. Zimbabwe: Steve H. Hanke and Alex K. F. Kwok, “On the Measurement of Zimbabwe’s Hyperinflation,” Cato Journal 29(2), 2009, Table 1 (Reserve Bank of Zimbabwe rates March 2007 to July 2008; the authors’ own estimates thereafter). Venezuela: Banco Central de Venezuela, Índice Nacional de Precios al Consumidor, base December 2007, file vintage 2026-07-03. Threshold: Phillip Cagan, 1956. Retrieved 2026-07-17.

The first thing the figure shows is that these episodes are not each other’s cousins in magnitude. Venezuela’s currency lost enough value in two years to multiply prices by more than a hundred thousand, which is a catastrophe by any standard a household applies, and on this page it is the flat line. Germany went four and a half orders of magnitude further in less time. Zimbabwe’s estimated terminal multiple runs to twenty-seven digits. The second thing, and the more useful one, is the shape they share: a long, comparatively gentle climb, and then a near-vertical final stretch in which the great majority of the total destruction happens in the last few months. German prices multiplied by about 400 over the first eleven months after onset, and then by about ten million over the four months that followed. That is the signature of a process that feeds on itself.

Why a currency dies

Money is created in the ordinary course of business, mostly by commercial banks lending it into existence, which is the subject of the history volume’s chapter on where money actually comes from. Hyperinflation is not that mechanism running fast. It is a different mechanism, and its cause is fiscal rather than monetary in origin: a government whose spending commitments exceed what it can raise in taxes and what anyone will lend it, and which therefore covers the difference with newly issued money because that is the only creditor it has left. Germany after 1918 was carrying reparations, occupation, and a tax system that had not been rebuilt for any of it. Zimbabwe in the 2000s had an economy shrinking year after year and a state that had moved its bills to the central bank. Venezuela after 2014 had an oil revenue that fell and a spending structure that did not. In each the printing was the symptom of an arithmetic that had already failed, which matters because it tells you what the disease is not. A central bank that creates money to hit a 2 percent inflation target, or that buys assets to stop a deflation, is doing something the record distinguishes sharply from this; why the target is 2 percent and not zero, and the record of the large-scale asset purchases that followed 2008, are treated in the history volume’s chapter on why a little inflation is the target.

What turns a fiscal hole into the vertical part of the curve is that the public works out what is happening. Money is held on the understanding that it will still be worth roughly this much tomorrow, and the moment that understanding goes, everyone tries to hold it for a shorter time. Wages are spent the day they arrive; sellers price in tomorrow’s expected fall rather than today’s cost; anyone with a choice moves into goods, or foreign notes, or anything that is not the currency. That flight is itself inflationary, because the same quantity of money chasing the same goods, changing hands more times a day, does the work of a larger quantity. So the state must print faster to buy the same real resources with money that is worth less by the week, and the faster it prints the faster the flight. Bresciani-Turroni, writing up the German case from inside the wreckage, records how far the compression went: it became the custom to pay an advance on Tuesday and the balance on Friday, and later, in his words, “some firms used to pay wages three times a week, or even daily,” because by then it made a difference to a workman whether he was paid in the morning or the afternoon. What Germany’s currency was still doing at that point was serving as a way to move value across a few hours. It had stopped being a way to hold it at all.

Money that cannot store value across a week has effectively resigned from one of its jobs, and the other two follow. It stops being a unit of account, because contracts and prices get written in dollars or in goods; and it eventually stops being a medium of exchange, which is where these episodes end. Zimbabwe’s did not end with a clever policy. Hanke and Kwok record that at the November peak people simply refused to use the Zimbabwe dollar, and the hyperinflation came to an abrupt halt. Germany’s ended with a new unit, the Rentenmark, declared on 15 October 1923 to be worth a trillion of the old paper marks. The declaration is not the same thing as the stabilization, and it is worth keeping them apart: Sargent’s reading of the same tables in Figure 6.1 is that prices stopped rising and the mark stopped depreciating in late November, more than a month after the decree, once the fiscal position behind the new unit was credible.

Who pays, and who does not

The wheelbarrow photograph shows the chaos and not the direction of it. A hyperinflation is not a lottery in which everyone loses. It is a transfer, and it runs in a direction that can be established from the legal record. The German case is the one documented thoroughly enough to itemize, and Figure 6.2 does that: every row names the decree, law, or statistical record that establishes it.

WHO LOST, WHO GAINED: GERMANY, AND THE DOCUMENT THAT SAYS SO every row names the decree, law or record that documents it THOSE WHOSE CLAIMS WERE WIPED MORTGAGE AND DEBENTURE CREDITORS revalued at about 15% of their original gold value; the rate for mortgages later raised to 25%, but not demandable until 1932 Decree of 14 February 1924; Law of 16 July 1925 HOLDERS OF GOVERNMENT WAR LOANS 1,000 old marks converted into 25 new, or 2.5%; those who bought before July 1920 got a 30-year lottery from 1926, worth 12.5% Law of 16 July 1925 SAVINGS-BANK DEPOSITORS, PRUSSIA deposits revalued at between 17% (Berlin) and 29% (Upper Silesia) of the original value of the deposit the decree that finally settled the matter WAGE EARNERS miners’ real wages, as a share of pre-war, 1923: 47.7% in January, 86.2% in March, 47.6% in July, 81.2% in October Statistical Bureau of the Reich, reported by Bresciani-Turroni THOSE WHO STEPPED OVER THE WRECKAGE THE STATE the sum of all German war debt, valued at 154 billion Mark, dropped to just 15.4 Pfennig the day the Rentenmark came in Deutsche Bundesbank DEBTORS AND OWNERS OF REAL ASSETS “it became one of the rules of good management to contract as many debts as possible: debts which were repaid later with depreciated currency” Bresciani-Turroni; no balance sheets retrieved, so no figures here
Figure 6.2 The German hyperinflation as a transfer, itemized from the record that documents it. Each row names its instrument, because the losses were not merely suffered; most of them were eventually adjudicated, and the adjudication is where the size of the loss is written down. The creditors who had lent against property recovered about 15 percent of the gold value they had lent, raised on paper to 25 percent for mortgages by the law of 1925, though that law also postponed payment to 1932, so the higher figure without the delay overstates what anyone got. People who had lent to their own government by buying war loans drew the lowest rate on the plate: 25 new marks for a thousand old ones. Wages tell a different story from the others, because they were renegotiated constantly and so whipsawed rather than simply fell. On the other side the state extinguished a war debt the Bundesbank puts at 154 billion Mark for a final settlement of 15.4 Pfennig, and the borrowers who had bought real things with borrowed money repaid in currency that no longer meant anything. No source retrieved for this plate computes a total sum transferred from one group to the other, so no such total appears here. Costantino Bresciani-Turroni, The Economics of Inflation (Allen & Unwin, 1937 English edition), for the decrees of 14 February 1924 and 16 July 1925, the Prussian savings revaluation, the Stinnes record, and the miners’ real-wage series he attributes to the Statistical Bureau of the Reich. War-debt figure: Deutsche Bundesbank. Retrieved 2026-07-17.

Read the ledger as a single sentence and it says this: the losses fall on everyone whose wealth was a promise denominated in the currency, and the gains fall on everyone whose obligations were. A savings account, a pension, a life-assurance policy, a government bond, a mortgage you hold as the lender, the wage you have earned but not yet spent: each is a claim on a fixed number of units, and a hyperinflation does not confiscate them so much as evaporate them, which requires no legislation and generates no confiscation notice. Meanwhile a debt is also a claim on a fixed number of units, seen from the other end, and so the farmer with a mortgage, the industrialist who borrowed to buy the factory, and above all the state with a war to pay for, all watch their obligations shrink to nothing at the same rate. This is why the state is so badly placed to stop it. The government running the printing press is also the economy’s largest debtor, and the inflation that is destroying its citizens’ savings is simultaneously the only instrument available for retiring debts it cannot otherwise pay.

What sorts the losses is worth naming, because it explains why the experience feels arbitrary from the inside. The sorting principle is whether a claim can be repriced. Cash cannot be, and it loses value for exactly as long as the episode runs. The pension, the insurance policy and the bond are promises written in fixed units of the currency, and they cannot be renegotiated either, which is why the ledger settles them at revaluation rates set years later. Wages can be renegotiated, raggedly, and the figure shows what that looked like: a whipsaw rather than a floor. What survives is what is not denominated in the currency at all: land, machinery, a house, foreign notes, and a debt owed by you rather than to you. A country that has been through this once does not forget which of those categories it was in, and that memory outlives the episode by generations.

Printing money always causes hyperinflation.

Oversimplified Moderate confidence

The claim has a true kernel, which is why the ruling is not that its reverse holds. Every episode in this chapter was financed by issuance. A state that covers its bills with newly created money, at the scale and duration these three did, produced a hyperinflation in all three cases, and anyone who answers that money creation is simply unrelated to the value of money is contradicted by Figure 6.1. That is what keeps this ruling off backwards. The word doing the damage is “always.” Rich-country central banks created money on a great scale after 2008 and again in 2020, and consumer prices in those countries did nothing resembling the lines in this chapter; the argument that followed 2008, treated in the history volume’s chapter on why the inflation target is a low positive number rather than zero, was that inflation was running too low, not that it was getting away. The claim collapses a distinction the cases themselves insist on. In each of the three episodes here the printing was downstream of a fiscal position that had already failed, a government that could not tax what it spent and could no longer borrow, so the press was the last creditor available. Printing on those terms is a symptom of that arithmetic, and the arithmetic, not the press, is the thing to look for. The claim takes a mechanism that is present in every case and states it as a law that holds in every case, which the last two decades in the advanced economies did not bear out.

Sources
  • The episodes and their measurements: Germany, Statistisches Reichsamt wholesale prices as printed in Thomas J. Sargent, “The Ends of Four Big Inflations,” Table G1 (NBER/University of Chicago Press, 1982) — a wholesale index, not a cost-of-living one. Venezuela, Banco Central de Venezuela, Índice Nacional de Precios al Consumidor, base December 2007 (the series published after the 20162019 silence and backfilled; file vintage 2026-07-03). Zimbabwe, Steve H. Hanke and Alex K. F. Kwok, Cato Journal 29(2), 2009, Table 1: the Reserve Bank of Zimbabwe reported rates for March 2007 to July 2008 and the authors calculated the rates after that. No official series exists past July 2008; the 79.6 billion percent for 14 November 2008 is Hanke and Kwok’s estimate, derived from the Old Mutual share price in Harare and London, and is not an official figure.
  • The threshold: Phillip Cagan, “The Monetary Dynamics of Hyperinflation,” in M. Friedman (ed.), Studies in the Quantity Theory of Money (1956) — at least 50 percent a month, as quoted and applied by Hanke and Kwok.
  • Money creation that did not do this: the creation mechanics are treated at History, Chapter 2, and why the target is a low positive number at History, Chapter 3. The rich-country price record since 2008 is the published consumer-price series of those countries’ own statistical agencies, and nothing in it approaches the threshold that defines this chapter’s episodes. What an exchange rate does in ordinary times is at Countries, Chapter 8.
  • Confidence is moderate under the rubric, which scores the weaker of evidence directness and construct match. Evidence directness is strong for the episodes: three price records, each pinned to its issuer, with the one estimated segment marked as an estimate. The binding, weaker leg is construct match. “Printing money,” as the claim uses it, is a single folk category that runs together two things this chapter’s evidence holds apart: a treasury financing its deficit directly with issuance, which is what the three cases are, and a central bank buying assets under an inflation target, which is what the counter-cases are. No source cited here counts how often the first happens without producing hyperinflation, so the “always” is falsified rather than measured, by counter-cases from a different construct than the one the claim probably intends. The direction is not in doubt, since fiscal financing at scale and hyperinflation travel together in all three episodes here, which keeps the ruling firm and off low; the looseness of the claim’s own category is what keeps it off high.

What a dead currency actually costs

The answer to the question the news item provokes runs through the ledger rather than the wheelbarrow. Your currency will not die because a central banker was careless or because too much money was created in the abstract. It will die if your state reaches a point where it can pay only by printing, and the warning sign is fiscal rather than monetary: a government that cannot raise what it spends and can no longer find a lender. If it happens, the arithmetic of who is hurt is knowable in advance. Everything you are owed in the currency goes. Everything you owe goes with it, which is why a hyperinflation can leave a mortgaged household better off in net worth and a prudent saver with nothing. And the state, which is the largest borrower in the room and the one holding the press, is on the winning side of the ledger throughout.

The three episodes here also share something the figure cannot show, which is that in each case the currency was the last thing to fail rather than the first. The money died because the state had already run out of ways to pay, and the state had run out because of a lost war, a collapsed economy, a fallen oil revenue. Money is the instrument through which those failures were transmitted to households, and the instrument is what everyone remembers, because it is the part that touched them. Whether that pattern, in which a failure somewhere else reaches people through the claims they hold, is confined to money is a question with a harder case attached: how can famine strike a country that has food in its granaries? A later chapter in this volume takes up that question.