« How Should Changes in Inequality Be Measured and Assessed? | Main | John Taylor: The Iraq Currency Plan Was a Big Success »

Monday, February 26, 2007

Inflation and Unemployment II

This post from earlier today noted that there is evidence that the relationship between inflation and measures of real activity such as the unemployment rate has weakened in the last twenty-five years. I want to return to this topic.

There are two types of causality we can examine with respect to the relationship between unemployment and inflation (there are others as well). The first is how a monetary (aggregate demand) shock that impacts inflation affects unemployment and other measures of real activity. It's possible that unemployment reacts less to a change in monetary policy than it did in the past. The second type of causality is how a shock to unemployment or output, perhaps from a supply-shock such as a change in productivity or a real cost shock, affects inflation in subsequent periods. The first is a demand driven change in the economy while the second originates on the supply side. This will address the first type - demand driven fluctuations of the kind that generate Phillips curve type movements in inflation and measures of real activity.

The causality here runs from monetary policy shocks to inflation to unemployment. I am going to use an expectations augmented Phillips curve framework to illustrate ideas. It's not the most up-to-date framework, but the ideas carry through to more modern theoretical structures and it has the advantage of being a simple and familiar structure.

The following figure shows two short-run Phillips curves, one for an expected inflation rate of π0 and one for an expected inflation rate of π1. Consider the PC labeled PC(πe = π0). This curve shows that if the inflation rate is equal to its expected value of π0, then the unemployment rate will be at the natural rate, U*. However, if the inflation rate is different than what is expected, then the economy will slide along the short-run Phillips curve and output will differ from its natural rate.


For example, suppose that agents expect an inflation rate of  π0, but the actual inflation rate turns out to be, unexpectedly, the lower rate of π1. Then instead of ending up at the natural rate rate, i.e. at point A, the economy will move to point B and the unemployment rate will increase. Thus, any unexpected inflation causes deviations from the natural rate of output.

If the lower inflation is permanent, over time expectations will adjust to reflect the lower actual inflation rate. That is, over time, the expected inflation rate will drop from  πe = π0  to  πe = π1 and the Philips curve will shift downward as shown in the diagram. As the Phillips curve shifts down, the economy will move from point B to point C as the unemployment rate returns to the natural rate.

Notice that the path of unemployment in response to an unexpected shock is from point A to B to C as unemployment first rises, then falls back to its initial level. But what if the change in policy was expected rather than unexpected? In this case, the fall in unemployment is less costly.

If the change in inflation is anticipated in advance, then the PC will shift as the inflation rate changes and the economy can move from point A to point C without as much (or with perfect flexibility without any) increase in unemployment. That is, with some degree of price and wage rigidity present, an unexpected change in inflation will move the economy from A to B to C, while an expected change will move more directly from A to C as shown, for example, by the dashed line in the diagram. Thus, importantly, anticipated changes lead to much less variation in unemployment than unanticipated changes.

This means that one explanation for increased stability (when causality is from policy to real activity) is that policy is easier to anticipate than in the past. Has this happened in the last twenty five years, i.e. is this a realistic assumption? It seems so for at least three reasons, the use of interest rate rules that have stabilized both actual and expected inflation making them more predictable, transparency about how the interest rule is implemented, and increased credibility.

Since the 1980s, the Fed has followed an interest rate rule and it has, at the same time, attempted to be much more transparent and communicative about its policy procedures. To the extent that this has allowed policy to be predicted more accurately, the economy should be more stable.

Increased credibility since the 1970s is also important. If the monetary authority says it is committed to a policy of lowering inflation and announces such intentions, but then allows inflation to continue to creep upward so that it loses credibility (as the Fed did in the 1970s), then there will be much more uncertainty about Fed policy and therefore much less predictability. All of these factors - committing to a rule, increasing transparency, and enhancing credibility allow policy to be predicted with less error and, according to the model above as well as more modern versions of it, the economy ought to stabilize. With a more stable economy, a given change in inflation will be associated with smaller changes in real activity.

    Posted by on Monday, February 26, 2007 at 05:46 PM in Economics, Inflation, Unemployment | Permalink  TrackBack (0)  Comments (6)



    TrackBack URL for this entry:

    Listed below are links to weblogs that reference Inflation and Unemployment II:


    Feed You can follow this conversation by subscribing to the comment feed for this post.