Bryan Caplan  

Testing Arnold's Dual Inflation Regimes Hypothesis

PRINT
Bond Bubble Watch... Austrian Resurgence...
Arnold has a novel theory of dual inflation regimes:
I don't think that the Fed can fine tune the economy. I think that, at most, it can toggle between two regimes: a regime where inflation is high and variable; and a regime where inflation is low and stable.
This seems testable by optical least squares.  If we get data on a bunch of countries over the last 60 years, put mean inflation on the x-axis, and inflation variance on the y-axis, there will be two easily visible clusters.  I'd like to run the test myself, but I can't easily find the data (or better yet, a pre-made graph).  Help!

P.S. If you want to allow individual countries to shift regime, one easy place to start is to break each country's results down by decade, yielding N*6 total decade-long "regimes."  If you make the time periods much shorter Arnold probably needs a new label for his hypothesis...


Comments and Sharing





TRACKBACKS (1 to date)
TrackBack URL: http://econlog.econlib.org/mt/mt-tb.cgi/3912
The author at Eli Dourado in a related article titled Are There Two Inflation Regimes? writes:
    Arnold Kling tentatively postulates that central banks can, at most, select between a low-stable inflation regime and a high-variable inflation regime. Bryan Caplan proposes a quick test, which I hereby supply. Below are some scatterplots of inflation ... [Tracked on August 28, 2010 3:05 PM]
COMMENTS (12 to date)
Eric Rall writes:

The IMF World Economic Outlook database has data for most countries post-1980.


You can use the report generator to build reports for various country groups, for example

John Hall writes:

The monetary policy can change over time, so that doesn't seem like a good way to look at the data.

Another way to do it would be to do a VAR-type regression for inflation and growth, but instead do it as a multivariate markov regime-switching equation. For each country you can test whether multiple regimes are present.

Eric Rall writes:

Here's a graph based on the "Advanced Economies" country set. X-axis is inflation rate in %, and the Y-axis is the % change year-over-year in the inflation rate.

jsalvatier writes:

Wasn't it obvious that Arnolds theory was nonsense? There are countries that have had between 0% and 6% inflation in the recent past, I think singapore was the most dramatic. I got most of my data from here: http://www.indexmundi.com/australia/inflation_rate_%28consumer_prices%29.html

Unfortunately I don't have my excel sheet with me.

Bryan Caplan writes:

Eric, I'm having trouble making sense of the graph you linked to. It almost seems like it's showing inflation versus average inflation. Is there another page with an explanation?

Eric Rall writes:

I misinterpreted the data when I made the graph. On referring back to my source, I'm actually graphing consumer prices vs. inflation. I'd thought I was graphing inflation vs the change in the inflation rate.

I'd cited my source for the data in an earlier comment that seems to still be stuck in moderation. It's from the IMF's World Economic Outlook database (available in the Data and Statistics tab of the website), looking at the Advanced Countries dataset, from 1980-2009, and selecting the "Inflation, average consumer prices
Index, 2000=100" and "Inflation, average consumer prices Percent change".

In order to generate the graph you want, I would need to take only the percent change dataset and generate the variance by hand.

Eric Rall writes:

Here's a new-and-improved graph, in which the X axis is annual inflation (%), and the Y axis is the absolute value of the difference between this year's inflation and last year's inflation.

And a close-up of the lower-left corner, in which most of the data clusters.

Steve Roth writes:

Wouldn't it be better, for the y axis, to use a volatility measure like coefficient of variation for, say, rolling 5-year periods?

I fond that useful here:

http://www.asymptosis.com/ahh-for-european-stability.html

Yancey Ward writes:

What are the definitions of the following terms:

"High Inflation", "Low Inflation", "Variable Inflation", and "Stable Inflation"?

Peter writes:

Yancey, I'll take a stab at defining them:

Low Inflation: Inflation that averaged over a sufficient period (probably one business cycle), does not exceed average GDP growth. This means that average wealth is not destroyed by inflation.

High Inflation: Inflation that does not meet the standard of low inflation.

Stable inflation: Inflation where the rate of change (first derivative) of the inflation rate is less than the first derivative of the change of GDP* over a sufficient period, probably one business cycle.

Variable inflation: Inflation that does not meet the standard of stable inflation.

I use a relative standard to GDP as opposed to an absolute standard, because the variation of growth matters. If the US had 6% inflation right now, that would be very high, as we only have ±2% growth. If China growing at 9% has 6% inflation, it is much less of a problem for them.

*Technical note, the standard should actually be the absolute value of the change of inflation being lower than the absolute value of change in GDP, because otherwise downturns get mathematically silly.

Doc Merlin writes:

'Inflation that averaged over a sufficient period (probably one business cycle), does not exceed average GDP growth. This means that average wealth is not destroyed by inflation.'

Average wealth isn't necessarily affected by inflation. It depends if there are alternative effective currencies and other value stores. You need to stop thinking of inflation in the the macroeconomic sense. It is complete garbage. A better way to think of it is in competition between currencies and other sources of store-of-value.

David B. Collum writes:

I will repeat a notion sent in an email this AM: I think any inflation model that fails to include an acceleration of the replacement cycle (accelerated depreciation) has missed some significant inflation. While Michael Boskin would aggressively adjust for the improvements of cell phones over land lines, I would correct for the fact that the former lasts about a year whereas the latter lasted for decades. Our disposable society is highly inflationary. What used to be household assets (appliances) are really liabilities demanding constant replacement. They can't even be repaired.

Just watching from the cheap seats...

Comments for this entry have been closed
Return to top