Economic history · When Money Stops Working
Eight steps, about twelve minutes, a question at each. You need no economics and no preparation, only a willingness to guess before you are told.
This part takes as read what caused it. In all three countries a state was spending money it had not raised, and the bank that issues the currency was making the rest.
The last part settled what Germany, Zimbabwe and Venezuela had in common, and this part is what that arithmetic does to the people living with it.
The money stops being what prices are quoted in. Traders, employers and landlords start setting prices off the day's dollar exchange rate, and in Zimbabwe in January 2009 the central bank wrote that into an instruction.
Then it stops being something a country's own figures can see. Zimbabwe's statistics office made its last monthly recording in July 2008; Venezuela's central bank stopped publishing in December 2015. The figures a reader finds today were estimated from exchange rates, and the percentages on offer cover different spans.
Then it stops being something a contract can be written in. Germany's whole war debt was worth 15.4 Pfennig in November 1923, and a lifetime's savings went the same way. And nearly eight million Venezuelans left.
What this module covers
Written for every level. Tap any underlined word for what it means, and open the boxes below for the economics behind each decision. If you already know the theory, skip both and the history reads straight through.
Step 1 of 8
In January 2009 Zimbabwe's told every licensed trader to quote goods in local money as well, on a formula built from the .
By then the local money had stopped being what people reckoned in. The Peterson Institute describes shops in an extreme inflation setting a price at whatever it would take to buy the thing in dollars.
The exchange rate in that formula was the inter-bank one, fixed at the mid-point and sent out to the market by the Bank, so the state was telling shops which number to price off.
Germany had reached the same place in 1923 without an order. The German museum record has shop opening hours following the times the day's exchange rates were announced, and a restaurant bill doubling during the meal.
Take a guess. You are not expected to know the answers, and being wrong is what makes a number stick.
In January 2009 Zimbabwe's central bank told traders how to work out a price in local money. What did the formula start from?
Six months before that order, in July 2008, Zimbabwe's statistics office made its last monthly recording of , four months before prices peaked.
Zimbabwe's money was losing value inside a day, and an index measures what shops charged over a whole month, so each survey described a month that had already gone.
The Federal Reserve Bank of Dallas records that last release at 2,600.2 per cent month on month, or more than 231 million per cent on a yearly reckoning. No official Zimbabwean figure followed it.
Venezuela's central bank did the same thing seven years later. An working paper records that the bank stopped publishing inflation and other indicators from December 2015.
The United Nations commission for the region was still recording it in 2024: no official information was available, its Venezuela note says, on the national accounts or the public finances.
Zimbabwe's own statistics office recorded inflation month by month until July 2008. What happened to prices after that date?
The figure most often quoted for Zimbabwe's worst month was worked out from the price of one company's shares.
Old Mutual was listed in Harare and in London at once, so one holding carried two prices at the same moment, one of them in a currency that still worked.
Steve Hanke and Nicholas Krus used that gap for August to November 2008, and their table names the basis in a column of its own: Implied Exchange Rate for Zimbabwe, Wholesale for Germany.
Zimbabwe's central bank names the same share. Its January 2009 statement records the Old Mutual price being used in the economy as a stand-in for the parallel market exchange rate.
So the most-quoted Zimbabwean figure is a currency price rather than a shop price. Nearly every number a reader meets about these episodes started as an exchange rate.
Zimbabwe's most-quoted peak figure was calculated from the price of Old Mutual shares, traded in Harare and in London. What did comparing the two prices give?
Zimbabwean prices doubled in about 24.7 hours at the worst of it, in mid-November 2008.
The figures below cover different spans, the Bank of England's a month and the Dallas one a year, so a doubling time can be compared and a percentage cannot.
The Bank of England gives around 80 billion per cent for a single month. The Federal Reserve Bank of Dallas gives 489 billion per cent for September 2008, which is the Fund's estimate on a yearly reckoning.
Two Congressional Research Service reports give 471 billion per cent for that same September and 500 million per cent or more for the collapse. Zimbabwe's own office, counting to July 2008, gave 231 million.
At the peak in mid-November 2008, roughly how often did Zimbabwean prices double?
Venezuela's National Assembly counted the country's inflation for 2018. Its central bank had given up publishing figures in December 2015.
The Assembly's finance committee collected shop prices in Caracas, which is thinner work than a statistics office running a national survey, and it covered less ground.
An International Monetary Fund working paper takes the Assembly's figures and reports inflation of 488,865 per cent year on year at September 2018.
The United Nations commission for the region gives 130,060.2 per cent for the whole of 2018, the highest annual inflation rate in the region's history. One month against one year: the two are about four times apart.
Venezuela's central bank had given up publishing prices. Where did the inflation figures for 2018 come from instead?
The Reserve Bank of Zimbabwe called the 2008 surge in prices self-fulfilling inflation , and that phrase is the mechanism.
The European Central Bank sets the loop out. Prices rise, so wage demands rise to match; those wages raise costs; costs raise prices again; and the belief has made itself true.
Holders of the money also spend it faster, to be rid of something worth less by the hour. The same money then does more buying in a week than it once did in a month.
The Peterson Institute puts the same thing in the seller's hands: in an extreme inflation, pricing stops looking back at what things cost last week and starts looking at the exchange rate.
So a belief moved Zimbabwe's prices in 2008 while nothing physical had changed. A belief can also change on a date, which is what makes an ending quick.
A trader is sure that next week costs more than this week. On the European Central Bank's account, what does that belief do?
Germany's whole war debt, 154 billion Mark, was worth 15.4 Pfennig on the day the new currency arrived in November 1923.
Marks were what the debt was written in, and by November 1923 a mark bought almost nothing, so the borrower was clear and whoever was owed had nothing.
The New York Fed went through the accounts of about 700 German firms and found their debts, set against their assets, down by more than half between 1919 and 1924. Bankruptcies fell as prices rose.
The other side is in the German museum record: the savings of generations destroyed, fixed interest worth almost nothing, and war that had counted as the safest holding in the country gone.
The Congressional Research Service reports the same in Venezuela. Hyperinflation there reduced the value of what households had saved, while the state stayed in default on what it owed abroad.
Three of these were wiped out by the price rise. Which one came out ahead?
Nearly eight million Venezuelans were living outside the country in December 2024, against 26.7 million still inside it.
That is close to a quarter of the people born there. The count is the United Nations agencies', reported by the Congressional Research Service in September 2025, and it moves.
The Service also puts the economy's fall at roughly three-quarters between 2014 and 2021, on the Fund's figures. It is among the deepest falls any country has had outside a war.
A later report of its own gives more than 80 per cent over the longer span from 2013 to 2020. The economy in 2024 was still under half the size it was in 2013.
A quarter of a country leaving is what a price does when it stops working. Those people were still abroad in December 2024, years after the prices stopped moving.
Nearly eight million Venezuelans were living abroad by December 2024. Roughly what share of the country's people is that?
Module 5 of 7 in When Money Stops Working
What Zimbabwe told its traders to set a price from, January 2009
the inter-bank exchange rateZimbabwe's last official monthly recording, against the peak
July 2008, against mid-November 2008What the most-quoted Zimbabwean peak figure was calculated from
one company's share price in two citiesHow long Zimbabwean prices took to double at the peak
24.7 hoursVenezuela's 2018 , on two bodies' figures
488,865% year on year at September, 130,060.2% for the yearWhat the Reserve Bank of Zimbabwe called the 2008 surge
self-fulfilling inflation expectationsGermany's war debt at the currency change, November 1923
154 billion Mark, worth 15.4 PfennigVenezuelans living abroad in December 2024, and the economy they left
nearly eight million, from a fall of roughly three-quartersYou met six terms in this module
, , , , ,
Prices in somebody else's money, figures that stopped before the worst of it, and a quarter of a country gone. How it ends is next.
[1] When Zimbabwe's statistics office last measured inflation, and what an inflation did to German company debts: Federal Reserve System: Federal Reserve Bank of Dallas, Globalization and Monetary Policy Institute 2011 Annual Report, Hyperinflation in Zimbabwe; and Federal Reserve Bank of New York, Liberty Street Economics, Inflating Away the Debt, 13 July 2023. The Dallas report's chart note is the load-bearing sentence: "Zimbabwe's last official month-to-month recording of inflation by the country's Central Statistics Office, July 2008, although estimates are much higher. The official annual rate recorded for July 2008 is 231 million percent, and the International Monetary Fund estimated the annual inflation rate for September 2008 at 489 billion percent." The body adds that at the last release the month-over-month rate "had reached 2,600.2 percent". So the 231 million per cent is Zimbabwe's own office on a yearly reckoning of a monthly figure, and the 489 billion per cent is the Fund's estimate for a later month, reported here rather than measured here. The report also gives the Z$100 trillion note as the largest denomination, US$1 for Z$4 million at the official rate on 31 December 2008 "although parallel black-market rates were much greater", and credits Hanke (2008) rather than an official index for the currency losing "more than 99.9 percent of its value". The New York piece digitises about 700 German joint-stock firms: "leverage fell by over 50 percent between the start of the inflation in 1919 and the aftermath of the inflation in 1924", and "Bankruptcies consistently declined with rising inflation and remained at historically low levels". It is a research post carrying the standing note that the views are the authors' own.
[2] What Zimbabwe's central bank told traders to price off, and what it said about its own statistics office: Reserve Bank of Zimbabwe, Monetary Policy Statement, January 2009; and the 2008 Annual Report. Paragraph 5.18 of the January 2009 statement, under Adoption of Dual Pricing Arrangements: "All traders shall therefore in addition to selling their goods and services in foreign currency, adopt a dual pricing framework where goods will also be quoted in local currency." Paragraph 5.19: "the pricing formulae to be implemented shall be based on the inter-bank market determined exchange rate which shall be fixed at the mid-rate level and communicated to the market by the Reserve Bank of Zimbabwe on a regular and/or as appropriate." Paragraph 4.15, on the share price: "The fungibility of the Old Mutual shares resulted in the share price of the counter being used as a proxy for the parallel market exchange rate. Consequently, the increase in share price due to the asset price bubble translated to a surge in the prices of goods and services in the economy." Paragraph 3.17 asks that the Central Statistical Office "be adequately capacitated so as to enable them to carry out substantive surveys", and records that Government "has directed the CSO, effective this month, to begin tracking developments in price indices in foreign currency terms". The 2008 Annual Report, paragraph 1.6, gives that year's stock market surge as "a result of self-fulfilling inflation expectations, unbridled speculation and fraudulent activities".
[3] The table that states its own basis, and where the doubling time comes from: Cato Institute, Steve H. Hanke and Nicholas Krus, World Hyperinflations, Working Paper, August 2012. The only table among these sources that declares its own basis on its face. Its columns are LOCATION, START DATE, END DATE, MONTH WITH HIGHEST INFLATION RATE, HIGHEST MONTHLY INFLATION RATE, EQUIVALENT DAILY INFLATION RATE, TIME REQUIRED FOR PRICES TO DOUBLE, CURRENCY and TYPE OF PRICE INDEX. Against Zimbabwe, March 2007 to mid-November 2008, that last column reads "Implied Exchange Rate" and the doubling time is 24.7 hours; against Germany, August 1922 to December 1923, it reads "Wholesale" and the doubling time is 3.70 days. The Zimbabwe footnote: "The authors calculated Zimbabwe's inflation rate, from August to November 2008, using changes in the price of the stock, Old Mutual, which was traded both on the Harare and London stock exchanges. The stock prices yielded an implied exchange rate for Zimbabwe dollars, under purchasing power parity." The paper's own heading The Arduous Data Compilation Process is where it describes official statistics being unavailable or unusable. cato.org returns 403 to this project's fetcher and the paper was read in full through the cloud container instead; it is signed work by two authors and the module names them rather than the institution.
[4] When Venezuela's central bank stopped publishing, and who counted after it: International Monetary Fund, Working Paper WP/18/266, Jose Luis Saboin Garcia: The Modern Hyperinflation Cycle, December 2018. A working paper by one author, carrying the Fund's standing note that the views are the author's and not the Fund's. "At the time of writing these lines (September 2018) year-on-year inflation has reached 488,865 percent." Its footnote 2 says where that came from and why: "According to the figures of the National Assembly, because the Central Bank stopped publishing these and other indicators since December 2015." That figure is a year-on-year rate at one month in 2018 and is not comparable with the United Nations commission's rate for the calendar year, which is why the module prints both with their spans attached and picks neither.
[5] Venezuela's inflation for 2018, and what was still unavailable six years later: United Nations Economic Commission for Latin America and the Caribbean, Economic Survey of Latin America and the Caribbean: Bolivarian Republic of Venezuela country notes, 2019 and 2024 editions. The 2019 note: "In 2018 the Bolivarian Republic of Venezuela recorded an annual inflation rate of 130,060.2%, the highest in the region's history; and, up to April 2019, it had suffered 17 months of hyperinflation." It gives the composition too, with housing rent at 486,684.5% and food and non-alcoholic beverages at 143,786.9%, which the module does not print. The 2024 note opens with footnote 1: "At the time of writing, no official information was available on national accounts, public finances, the labour market or the balance of payments." The commission's annual figure and the Fund working paper's September 2018 year-on-year figure are about four times apart and are not the same measurement; both are printed.
[6] How far Venezuela's economy fell, and how many people left: Congressional Research Service: R44841, Venezuela: Background and U.S. Relations, 6 December 2022; IF10230, Venezuela: Political Crisis and U.S. Policy, updated 30 September 2025; IF11268 and R44633 on Zimbabwe. Read at everycrsreport.com. R44841: "Venezuela's economy contracted by roughly 75% from 2014 through 2021 ... according to the International Monetary Fund (IMF)", so the Service is reporting the Fund's figure rather than measuring it; "hyperinflation reduced the value of household savings"; and "As of October 12, 2022, U.N. agencies estimated that more than 7.1 million Venezuelans have fled the country." IF10230 carries the later count: "UN agencies estimated there were some 7.9 million Venezuelan refugees and migrants globally in December 2024. As of May 2025, some 6.9 million of those resided in Latin American and Caribbean countries", and "roughly 73.2% of the population of 26.7 million lived in poverty in 2024". THE SAME REPORT GIVES A DIFFERENT CONTRACTION OVER A DIFFERENT SPAN: "Venezuela's gross domestic product (GDP) contracted by more than 80% from 2013 to 2020", with the economy in 2024 "still less than half the size it was in 2013". Both are printed and neither is picked. On Zimbabwe, IF11268 gives an annual rate "of 471 billion percent in September 2008" and R44633 gives "hyperinflation of 500 million percent or more" at the start of the unity government, and elsewhere "489 billion percent per annum or higher" citing the Dallas report -- three figures in two reports from one body.
[7] How far Zimbabwe's prices rose in one month, in a central bank's explainer for the public: Bank of England, explainer: How have prices changed over time?, 16 December 2021. "In extreme cases, high and volatile inflation can cause an economy to collapse. Zimbabwe is a good example. It experienced this between 2007 and 2009 when the price level increased by around 80 billion per cent in a single month." And: "As a result, people simply refused to use Zimbabwean banknotes and the economy ground to a halt." The Bank's figure is monthly and is the same order as the Cato table's 7.96 x 10 to the tenth per cent for the month to mid-November 2008; the Dallas and Congressional Research Service billions are annual rates for September 2008. That is the whole of the difference between them and it is why the module gives a doubling time instead.
[8] What a price is quoted in when the money has stopped working: Peterson Institute for International Economics, Gonzalo Huertas, Policy Brief 19-13: Hyperinflation in Venezuela: A Stabilization Handbook, September 2019. "Finally, once levels of inflation become extreme, the public's price-setting behavior undergoes a final shift. Nominal signals are so distorted that much of the pricing in the economy is done using foreign currency as a reference; goods may come to be denominated in domestic prices at whatever rate it would take to buy them in US dollars. In effect, money loses its role as a unit of account." And, on why fixing a rate acts quickly: "Price-setting behavior during episodes of extreme inflation tends to be less backward-looking and more exchange rate-based." The brief's own inflation levels are not quoted here.
[9] The loop by which an expectation about tomorrow raises a price today: European Central Bank, Price stability: why is it important for you?, April 2009. On the German episode: "As many people lost their savings, this led to a substantial loss in wealth for broad segments of the population. The realisation that price levels were constantly rising sparked a vicious circle. People naturally asked for higher wages, anticipating higher price levels in the future. These expectations became a reality, since higher wages translated into higher costs of production, which again meant higher prices. In the same vein, people started to pass on their money -- which lost its value -- by spending faster and faster." The module takes the ORDER of that explanation and writes its own sentences to it, which is what tools/step_skeleton.md asks for; the wording above is the Bank's.
[10] What Germany's war debt was worth on the day the currency changed: Deutsche Bundesbank, Inflation - lessons learnt from history, 15 October 2012. "As galloping inflation became hyperinflation in 1923, the currency could no longer function as a general means of payment or store of value. ... Inflation had almost completely devalued all financial assets and liabilities that had been in Mark. The state benefitted most from this because the sum of all German war debt, valued at 154 billion Mark, dropped to just 15.4 Pfennig on the day the Rentenmark was introduced." THE DEUTSCHES HISTORISCHES MUSEUM GIVES THE SAME FACT WITH DIFFERENT FIGURES, 164 billion Mark and 16,4 Pfennige, and neither institution notes the other. The module prints the Bundesbank's pair and says the museum's are ten billion Mark higher. Neither is corrected.
[11] What a German shop's opening hours followed in 1923, and what was left of people's savings: Deutsches Historisches Museum, LeMO: Die Inflation (Stand 14 September 2014). Under the heading „Die sozialen und politischen Folgen der Inflation“: „Ersparnisse wurden völlig entwertet, Spargelder von Generationen vernichtet. Feste Erträge oder Zinsen waren praktisch wertlos.“ -- savings were completely devalued, the savings of generations destroyed, and fixed returns or interest practically worthless. On the shops: „Ladenöffnungszeiten richteten sich nach den Bekanntgabeterminen für aktuelle Wechselkurse. In Restaurants konnte sich die Zeche während der Mahlzeit verdoppeln.“ -- shop opening hours followed the times at which the current exchange rates were announced, and in restaurants the bill could double during the meal. On the other side of it: „Gemäß dem Grundsatz 'Mark = Mark' konnten Kredite, die in höherwertigem Geld aufgenommen worden waren, mit entwertetem Geld zurückgezahlt werden. Schulden lösten sich in nichts auf.“ -- under the principle that a mark is a mark, loans taken out in better money could be repaid in devalued money, and debts dissolved into nothing. The page opens by dating the whole thing to 1914 and names the war bonds, „als mündelsicher angesehenen Kriegsanleihen“, as the holding whose complete devaluation cost the republic most in trust. Its figure for the state's war debt is „164 Milliarden Mark“ worth „16,4 Pfennige“ at the changeover of 15 November 1923, against the Bundesbank's 154 billion and 15.4.
Country data from the World Bank (CC BY 4.0) and the UNDP Human Development Report (CC BY 3.0 IGO)
Games · Learn · Atlas · Privacy