Comprehensive List of Historical Hyperinflations

Two Cato scholars have produced an exhaustively researched list of all known hyperinflation episodes. Should be useful as a reference in the future.

Haven't read it all yet but it's incredibly annoying how their table is in no sorted order as far as I can tell. Not by date, country (Alphabetically ordered) or rate of inflation.

So sloppy and it makes me feel like it was unscientific!

[edit] Scratch that, I failed at reading for a second there! It's in order of equivalent daily inflation rate... I guess that's a sort of order but not one that makes much sense to me in category sorting....

[edit2]
The "chapter" is written in an incredibly horrible manner. Not scientific at all and more along the lines of a conversation. IMO, and from my background, very unprofessional. Or maybe this is common in the economics and mathematics areas of publication?

[edit3]

Additionally, in the case of the world’s first hyperinflation, that of France, there
were two separate currencies in circulation in 1796, the mandat and the assignant. But, in
reading the literature, you wouldn’t know it. Many experts incorrectly claim the assignant
as the most hyperinflating currency of this time (e.g. Cagan, 1987; Veigh, 1995;
Bernholz, 1995).

So, the world's first hyperinflation only just occurred in the 1700s after 1700+ years of fairly-well recorded currency in the relative modern age and millennia of poorly recorded currency before that? This just smacks of arrogance...

According to Cagan, an episode of hyperinflation starts when there is a month in
which the price level increases by at least 50%, and ends after at least one year when
rates do not exceed this threshold. However, even Cagan does not follow this definition
completely, stripping it of precision. Unlike Cagan, we chose to follow his definition.

So, no explanation of why Cagan didn't follow his definition? No discussion of that? Surely there was a reason otherwise there's no reason to follow anything he said if he was so unreliable in the first instance!

They want to let the data speak for themselves without giving a proper discussion about how the data was compiled or without showing that raw data.

I think this thing appears pretty amateurish, tbh...

Duoae wrote:

Haven't read it all yet but it's incredibly annoying how their table is in no sorted order as far as I can tell. Not by date, country (Alphabetically ordered)
The "chapter" is written in an incredibly horrible manner. Not scientific at all and more along the lines of a conversation. IMO, and from my background, very unprofessional. Or maybe this is common in the economics and mathematics areas of publication?

I don't know, but I despise papers that are written in difficult to penetrate jargon. What they are discussing here is not that complex and it needs to be readable and understandable to the general population.

So, the world's first hyperinflation only just occurred in the 1700s after 1700+ years of fairly-well recorded currency in the relative modern age and millennia of poorly recorded currency before that? This just smacks of arrogance...

They talk about that in the paper - they left out episodes for which there was no reliable, high-quality data, including North Korea. The French hyperinflation was unsual for the time because it was extensively documented.

They want to let the data speak for themselves without giving a proper discussion about how the data was compiled or without showing that raw data.

The footnotes clearly delineate their sources. The only compilation that is being done is assembling everything into the table and making sure it's right, along with the documented cases where they had to use alternate measures.

I'm at a loss to understand what exactly you would have them include without directly copying the data from their sources.

Aetius wrote:
Duoae wrote:

Haven't read it all yet but it's incredibly annoying how their table is in no sorted order as far as I can tell. Not by date, country (Alphabetically ordered)
The "chapter" is written in an incredibly horrible manner. Not scientific at all and more along the lines of a conversation. IMO, and from my background, very unprofessional. Or maybe this is common in the economics and mathematics areas of publication?

I don't know, but I despise papers that are written in difficult to penetrate jargon. What they are discussing here is not that complex and it needs to be readable and understandable to the general population.

I'm not talking about jargon. I'm talking about not having a conversational tone and having more of a paper/textbook style of reading. It's called being professional.

So, the world's first hyperinflation only just occurred in the 1700s after 1700+ years of fairly-well recorded currency in the relative modern age and millennia of poorly recorded currency before that? This just smacks of arrogance...

They talk about that in the paper - they left out episodes for which there was no reliable, high-quality data, including North Korea. The French hyperinflation was unsual for the time because it was extensively documented.

Then their command of the English language is lacking because it wasn't the world's first, it was their first reliable data source.

They want to let the data speak for themselves without giving a proper discussion about how the data was compiled or without showing that raw data.

The footnotes clearly delineate their sources. The only compilation that is being done is assembling everything into the table and making sure it's right, along with the documented cases where they had to use alternate measures.

I'm at a loss to understand what exactly you would have them include without directly copying the data from their sources.

I guess graphs showing the 12 month periods for each event and what level of inflation each month in that period had/averaged is too much to ask for?
Looking at those hyperlinks they so thoughtfully provide as evidence of their data (hint: hyperlinking is now much more accepted than it was but generally you don't just link to the site and preferably you append the data you used to your paper because it may not be readily apparent on the site in question or it may be removed) they're not all that clear or useful. In fact, if I pick one at random (say, Bulgaria 1991, citation 33) it looks like there was a consumer price index spike in 1990, not 1991... and then again a huge one in 1996 - though this second episode is not listed in their table. I can't see where they got their data from as there are no tables and clicking on the links to the left of the chart does not take you back that far in time. Maybe I'm missing something?

But, ignoring that, let's go back to their definition of an event:

According to Cagan, an episode of hyperinflation starts when there is a month in
which the price level increases by at least 50%, and ends after at least one year when
rates do not exceed this threshold. However, even Cagan does not follow this definition
completely, stripping it of precision. Unlike Cagan, we chose to follow his definition

Then we go and look at the table in question and lo and behold a lot of their entries do not span a year's worth of time...

Duoae wrote:
According to Cagan, an episode of hyperinflation starts when there is a month in
which the price level increases by at least 50%, and ends after at least one year when
rates do not exceed this threshold. However, even Cagan does not follow this definition
completely, stripping it of precision. Unlike Cagan, we chose to follow his definition

Then we go and look at the table in question and lo and behold a lot of their entries do not span a year's worth of time...

And a number span more than a year. They are documenting the length of time hyperinflation was in effect, not that time + 12 months. The year's worth of time is just the amount of time the price level increase has to be below 50% in order for an event to be considered "over" - but the end of the event is the last month in which price levels rose over 50%, not twelve months after that. If you think about it, it's the only way to measure the events according to the definition.

Aetius wrote:

And a number span more than a year. They are documenting the length of time hyperinflation was in effect, not that time + 12 months. The year's worth of time is just the amount of time the price level increase has to be below 50% in order for an event to be considered "over" - but the end of the event is the last month in which price levels rose over 50%, not twelve months after that. If you think about it, it's the only way to measure the events according to the definition.

You're right, I misread that second bit.

Here is Cullen Roche's take on this report. I read that take before seeing this post.

http://pragcap.com/hyperinflation-st...

Anyhow, it was kind of nice to see a paper by the Cato Institute noting that hyperinflations tend to “arise under extreme conditions” as opposed to the usual “money printing” conclusion:

Hyperinflation is an economic malady that arises under extreme conditions: war,
political mismanagement, and the transition from a command to market-based economy –
to name a few.
In each of these circumstances, there are barriers to the recording and
publication of reliable inflation statistics. As we discovered over the course of our
investigation, overcoming these barriers was an arduous and painstaking process. In light
of this, it is little wonder that no one has been able to fully and accurately document
every case of hyperinflation.”

It’s nice to see that some people are beginning to connect the dots and make earnest attempts to understand why hyperinflation didn’t happen in the USA. Of course, if you don’t understand the basic operational realities of the monetary system then you likely can’t even begin to piece together the dynamics behind the recent lack of hyperinflation in the USA so if you’re bored on a Friday you might want to peruse this document first.

Oh, wow, this report just came out and it answers a ton of my questions. Thanks!

Washington Post has an article about this which just like me cites Cullen Roche.

http://www.washingtonpost.com/blogs/...

So what does any of this have to do with the Fed? One thing to observe, as Cullen Roche points out, that none of the most severe instances of hyperinflation appear to be triggered by a central bank simply trying to inject money into a basically intact economy in order to reduce the unemployment rate. “In fact,” he notes, “the monetary explosion is almost always the result of some other rare or extreme exogenous factor”— such as war or a collapse in productive capacity or a country ceding its monetary sovereignty.