Hyperinflation

Hyperinflation is an extreme, self-reinforcing loss of a currency's value, fast enough that holding the currency for any length of time is itself a loss and ordinary commercial life reorganizes around avoiding it. It sits at the far end of the same scale as ordinary inflation rather than being a different phenomenon, but the difference in degree is large enough to change behavior completely: households convert cash into goods or a stable foreign currency immediately, wages get renegotiated in weeks rather than years, and lenders either stop lending or price loans to cover the expected loss over the life of a very short credit period.

The word carries no official threshold. The Bureau of Labor Statistics publishes the Consumer Price Index and the Bureau of Economic Analysis publishes the price index for personal consumption expenditures, and what they publish is a rate of change rather than a classification of it. The Federal Reserve Bank of St. Louis put the position directly in a 2018 article on Venezuela: "Hyperinflation is an economic phenomenon that occurs when inflation increases very rapidly. Although there is not a precise number that defines hyperinflation, economists typically consider monthly inflation rates of above 50 percent as hyperinflation episodes."

Two reference points circulate, they come from different worlds, and they are nowhere near each other. Knowing that is more useful than memorizing either.

The economists' convention is monthly, and it is a convention. The 50 percent per month figure traces to a 1956 study of hyperinflations by Phillip Cagan, and it is a research threshold adopted so that episodes could be compared, not a rule any institution enforces. The St. Louis Fed article quoted above states the convention and, in the same sentence, states that no precise number defines the thing. Both halves are the point.

The accounting test is an indicator list, and it is far lower. IAS 29, the international accounting standard governing financial reporting in hyperinflationary economies, is where a company reporting in a collapsing currency has to make the call in practice. Its own text is careful. "This standard does not establish an absolute rate at which hyperinflation is deemed to arise," it says. "It is a matter of judgement when restatement of financial statements in accordance with this standard becomes necessary." What follows is a list of characteristics that "include, but are not limited to" five items, and only the last is numerical: "the cumulative inflation rate over three years is approaching, or exceeds, 100 %." That works out to roughly 26 percent a year sustained for three years. The research convention of 50 percent a month compounds to about 12,875 percent a year, so the two reference points differ by a factor of several hundred.

The gap between the two is not a contradiction. They answer different questions. The accounting standard asks when historical-cost financial statements stop being meaningful, which happens well before a currency collapses. The research convention asks when an episode belongs in a particular literature. Anyone quoting either as "the definition of hyperinflation" has stripped away the qualification the source itself supplies.

The four non-numerical indicators are the ones a person can actually observe. IAS 29's list begins with behavior rather than statistics: the general population "prefers to keep its wealth in non-monetary assets or in a relatively stable foreign currency," with local currency "immediately invested to maintain purchasing power"; it "regards monetary amounts not in terms of the local currency but in terms of a relatively stable foreign currency" and prices may be quoted in that currency; sales and purchases on credit "take place at prices that compensate for the expected loss of purchasing power during the credit period, even if the period is short"; and interest rates, wages and prices are linked to a price index. Those describe an economy that has stopped using its own money as a unit of account, which is what hyperinflation actually is. The percentage is downstream of that.

One caveat about the standard's own status. The International Accounting Standards Board lists an open project titled "High Inflation: Scope of IAS 29," so the wording quoted above could change. The current text is at ifrs.org.

A hypothetical showing why the monthly convention produces such extraordinary annual figures. Suppose prices rise 50 percent every month, the rate the research convention uses, and something costs $100 at the start.

• After one month it costs $150.00.
• After two months, $225.00.
• After three months, $337.50.
• After twelve months, about $12,974.63.

That is a compounding factor of roughly 130 over a single year, or an annual increase of about 12,875 percent. Twelve months of 50 percent is not "600 percent a year," which is the intuitive answer and is wrong by a factor of more than twenty. Compounding is the whole story at these rates, and it is why a hyperinflation that looks survivable month to month is not survivable across a year.

Set that against the accounting indicator. A cumulative rise of about 100 percent over three years, IAS 29's numerical characteristic, needs an annual rate of roughly 26 percent: 1.26 multiplied by itself three times is about 2.00, which is a doubling. Severe by any standard, disruptive to any household plan, and nowhere near the monthly convention.

What this means for a reader is mainly a matter of calibration. A period of high inflation is a serious event and a real planning problem, and the entry on inflation covers what it does to a long plan. It is not the same event as a currency ceasing to function, and the two are separated by orders of magnitude rather than by degrees. Claims that a developed economy is "on the verge of hyperinflation" are checkable against the indicators above, and the behavioral ones are usually easier to check than the arithmetic.

Why the concept is worth understanding

• It gives a reader a scale against which to judge alarming coverage of ordinary inflation, which is the commonest use for it in practice.
• The observable indicators are behavioral and public, so the claim can be tested rather than argued about.
• The accounting standard's refusal to set an absolute rate is itself informative about how contested the boundary is.

The traps in how the word gets used

• Neither circulating threshold is an official definition, and it is easy to meet one presented as though it were. One is a research convention and the other is an indicator in a list the standard says is not exhaustive.
• Because the two thresholds differ by orders of magnitude, a source quoting one without saying which can be off by an enormous factor.
• The word is frequently used rhetorically about economies with elevated but ordinary inflation, which drains it of meaning and makes real episodes harder to describe.
• No US statistical agency labels any reading, so there is nothing to appeal to when two commentators disagree about whether a threshold has been crossed.