Bryan’s question on testing for inflation regimes drew interesting comments. Some notes:
1. The idea that the rate of inflation and its variability are correlated is not original with me. I know the observation was made earlier in the literature, but I don’t have a specific citation at my fingertips.
2. I actually think there are three inflation regimes. Low, stable inflation; high, variable inflation; and hyperinflation. The latter is often defined as inflation at an annual rate of 100 percent or more. As I read the literature, hyperinflation is a fiscal phenomenon. That is, when a government needs to run deficits and cannot borrow, it runs the printing presses and inflation gets out of control. The only way to stop the hyperinflation is to get control over the deficit.
3. The empirical example I had in mind is the United States. I think of low inflation as under 3.5 percent and high inflation as over 5.5 percent, with the 3.5 to 5.5 percent range the region where the Fed may be “toggling” from one regime to another.
From 1952 through 1967, the December-December percentage change in the Consumer Price Index ranged from -0.7 (1954) to 3.5 (1966). Low and stable.
From 1968 through 1972, we were “toggling” into the high-inflation regime. The sequence was 4.7, 6.2, 5.6, 3.3, 3.4, with the latter two years temporarily and artificially suppressed by wage-price controls.
From 1973 through 1981, the range was 4.9 percent (1976) to 13.3 percent (1979). Those were the high and variable years.
From 1982 through 1986, we were “toggling” back to a low-inflation regime, with a sequence of inflation rates of 3.8, 3.8, 3.9, 3.8, 1.1. Then, we started to toggle up, with a sequence of , 4.4, 4.4, 4.6, 6.1. But at that point the Fed applied the brakes, and from 1991 through 2006 the maximum was 3.4 and the minimum was 1.6.
We have shown that we can maintain an inflation rate between 0 and 3.5 percent for long periods–the low and stable periods from 1952 through 1967 and from 1991 through 2006. We have shown that we can maintain an inflation rate above 4.5 percent for a long period, from 1973 through 1981. That was the high and variable period (which perhaps should include 1969-1972 if one factors out the wage-price controls).
The only question is how to interpret the early and late 1980’s, when inflation was mostly between 3.5 and 5.0 percent. My interpretation is that the intention was to toggle down into the low and stable regime. However, in the late 1980’s we started toggling up again, which point the Fed decided to quash inflation before it got out hand. Another interpretation would be that this period shows that we can remain in the 3.5 percent to 5.0 percent range for a long time, perhaps indefinitely. If we could indeed maintain the 1980’s pattern at will, then I would have to concede that my claim that the Fed must choose between two regimes is a false one. (Recall that the third, hyperinflation regime, is a fiscal phenomenon rather than a monetary phenomenon.)
READER COMMENTS
jsalvati
Aug 29 2010 at 11:09am
I should first say that I am not sympathetic to your macroeconomics. However, I do recommend putting a page/post together that links to all of your Macro Doubtbook that will get updated when you make additions and that is easy to find (thought a permalink or through google). When you have a long series of related posts introducing new material, it’s fairly frequent that people want to go back and reread things (if they’re interested anyway), so you want to make it easy for people to do so.
jsalvati
Aug 29 2010 at 11:42am
heh, I just noticed the link to the left!
Kevin Dick
Aug 29 2010 at 2:16pm
So this is an interesting observations, but have you run any statistical analysis to see if it’s significant? This looks like it could be a simple random walk.
Andy Harless
Aug 29 2010 at 3:06pm
The 1980’s look pretty stable to me, especially if you use core inflation data. The outliers in 1986 and 1990 are the result of one-time (and partly temporary) shifts in the volatile price of oil that can be explained by specific historical events. (The OPEC collapse in 1986 was arguably endogenous, but it would be hard to argue the same about the invasion of Kuwait.) We certainly did not have the impression at the time (in 1986 or 1990) that the monetary policy regime was shifting. It was only after the subsequent recession that that regime appeared to have shifted. And I think it is accurate hindsight to say that the Fed was engaging in opportunistic disinflation, where each recession effectively becomes a regime change. During the period from 1982 to 1990, we were in an intermediate regime, which, as best I can tell, worked out just fine.
ajb
Aug 30 2010 at 8:59am
Even if your theory is correct — and there are many reasons to believe it may be flawed — why shouldn’t the Fed work harder to stay near the 3% end of the inflation range in the midst of an historically awful recession? That’s the essence of Sumnerism and the core of the claim that tight monetary policy has dramatically exacerbated the problems of the Great Recalculation both here and in Europe and Japan.
Steve Roth
Aug 30 2010 at 1:37pm
It really bothers me when you (and others) use single-country data series to try to prove something like this. It’s just story-telling, the very antithesis of rigorous analysis.
Comments are closed.