Friday, March 02, 2007

Is the Natural Rate of Unemployment Dead?

Warning: This post is Macro related, and I am going to go a bit deeper than introductory level here. It is not "testable"

The WSJ ran an interesting article a few days ago on the Natural Rate of Unemployment, which those of you who took Macro will remember as being structural + frictional unemployment (the type of unemployment the Fed cannot eliminate with their policy tools. The idea being is that there is no getting around the natural rate of unemployment. Economist's old (old) idea was that inflation and unemployment were inversely related. If inflation rose, real wages would fall and businesses would hire more workers, lowering the inflation rate. The argument works in reverse as well. However, work done in the late 1960's by two Nobel Prize winners (Edmund Phelps and Milton Friedman) suggested that this relationship is not permanent because eventually workers would realize that their real wages were falling, negotiate higher wages and lead to a decrease in employment back to the natural rate. This view has, for the most part, been consensus thinking in macroeconomics since then since the theory ended up predicting the great inflation of the 1970s.

Recently, however, higher employment levels have not led to an increase in inflation.

But Fed officials have rethought that notion. They believe it takes a far bigger change in unemployment to affect inflation today than it did 25 years ago. Now, when inflation fluctuates, they are far more likely to blame temporary factors, such as changes in oil prices or rents, than a change in the jobless rate.

What is my opinion on this?

The natural rate of unemployment is related to potential GDP ( the vertical Long Run Aggregate Supply Curve). Thus any demand side stimulus will not have any impact on output, but rather be fed directly through to prices. This theory is likely true over a 1-2 year time span.

In the short run, however, there may exist a short run aggregate supply curve that is upwards sloping. If we accept this hypothesis for a moment (and I recognize that economists get into fist fights over the existence of this), then a boost in demand side stimulus will increase both prices and output.

What I think the article is trying to argue is that the SRAS supply curve has gotten flatter (more Keynesian) in recent years. Thus a boost in demand side stimulus has led to more of an increase in output than prices, in the short run.

However, the ONLY reason the SRAS supply curve has become more Keynesian is because the Fed has focused (and been more successful) on containing inflation. Workers are not as sensitive to upside blips in CPI anymore because they do not expect them to be long lasting. Their inflation expectation is unchanged. Workers are confident that inflation, on average will be around 2-2.5%, thanks to the Fed.

An attempt to exploit this Keynesian SRAS curve will result in the curve reverting back. Worker expectations will be for higher inflation because they are no longer confident in the Fed's ability to hold inflation at around 2%. Therefore, I agree with Edmund Phelps that

Mr. Phelps says the new thinking on the Phillips Curve doesn't change the implications of his Nobel-winning work. If the Fed never responded to higher inflation, consumers and businesses eventually would begin to expect higher inflation, and "then the game is up."

Economists need to keep the Lucas Critique of Macro Economics in the back of their minds. Just because a relationship is appearing in the data does not mean that the relationship is fundamental. If we cannot explain this relationship using economic theory, then we should, at the very least, be highly suspicious about our ability to exploit it. Remember the lessons of the 1970's.

Note: I grant that I was not alive in the 1970's. I was, however, very good at history in high school and college.

No comments: