The short-run Phillips curve says that at any given point in time, there is a trade-off between unemployment and inflation. According to this view, policy makers have a choice: they can choose to accept the price of high inflation in order to achieve low unemployment. In fact, during the 1960s, many economists believed that this trade-off represented a real choice.

However, this view was greatly altered by the later recognition that expected inflation affects the short-run Phillips curve. In the short run, expectations often diverge from reality. In the long run, however, any consistent rate of inflation will be reflected in expectations. If inflation is consistently high, as it was in the 1970s, people will come to expect more of the same; if inflation is consistently low, as it has been in recent years, that, too, will become part of expectations.

So what does the trade-off between inflation and unemployment look like in the long run, when actual inflation is incorporated into expectations? Most macroeconomists believe that there is, in fact, no long-run trade-off. That is, it is not possible to achieve lower unemployment in the long run by accepting higher inflation. To see why, we need to introduce another concept: the long-run Phillips curve.

Figure 31-12 reproduces the two short-run Phillips curves from Figure 31-9, SRPC₀ and SRPC₂. It also adds an additional short-run Phillips curve, SRPC₄, representing a 4% expected rate of inflation. In a moment, we’ll explain the significance of the vertical long-run Phillips curve, LRPC.

Suppose that the economy has, in the past, had a 0% inflation rate. In that case, the current short-run Phillips curve will be SRPC₀, reflecting a 0% expected inflation rate. If the unemployment rate is 6%, the actual inflation rate will be 0%.

Also suppose that policy makers decide to trade off lower unemployment for a higher rate of inflation. They use monetary policy, fiscal policy, or both to drive the unemployment rate down to 4%. This puts the economy at point A on SRPC₀, leading to an actual inflation rate of 2%.

Over time, the public will come to expect a 2% inflation rate. This increase in inflationary expectations will shift the short-run Phillips curve upward to SRPC₂. Now, when the unemployment rate is 6%, the actual inflation rate will be 2%. Given this new short-run Phillips curve, policies adopted to keep the unemployment rate at 4% will lead to a 4% actual inflation rate—point B on SRPC₂—rather than point A with a 2% actual inflation rate.

Eventually, the 4% actual inflation rate gets built into expectations about the future inflation rate, and the short-run Phillips curve shifts upward yet again to SRPC₄. To keep the unemployment rate at 4% would now require accepting a 6% actual inflation rate, point C on SRPC₄, and so on. In short, a persistent attempt to trade off lower unemployment for higher inflation leads to accelerating inflation over time.

To avoid accelerating inflation over time, the unemployment rate must be high enough that the actual rate of inflation matches the expected rate of inflation. This is the situation at E₀ on SRPC₀: when the expected inflation rate is 0% and the unemployment rate is 6%, the actual inflation rate is 0%. It is also the situation at E₂ on SRPC₂: when the expected inflation rate is 2% and the unemployment rate is 6%, the actual inflation rate is 2%. And it is the situation at E₄ on SRPC₄: when the expected inflation rate is 4% and the unemployment rate is 6%, the actual inflation rate is 4%. As we’ll learn in Chapter 33, this relationship between accelerating inflation and the unemployment rate is known as the natural rate hypothesis.

The unemployment rate at which inflation does not change over time—6% in Figure 31-12—is known as the nonaccelerating inflation rate of unemployment, or NAIRU for short. Keeping the unemployment rate below the NAIRU leads to ever-accelerating inflation and cannot be maintained. Most macroeconomists believe that there is a NAIRU and that there is no long-run trade-off between unemployment and inflation.

We can now explain the significance of the vertical line LRPC. It is the long-run Phillips curve, the relationship between unemployment and inflation in the long run, after expectations of inflation have had time to adjust to experience. It is vertical because any unemployment rate below the NAIRU leads to ever-accelerating inflation. In other words, the long-run Phillips curve shows that there are limits to expansionary policies, because an unemployment rate below the NAIRU cannot be maintained in the long run. Moreover, there is a corresponding point we have not yet emphasized: any unemployment rate above the NAIRU leads to decelerating inflation.

Recall the concept of the natural rate of unemployment, the portion of the unemployment rate unaffected by the swings of the business cycle. Now we have introduced the concept of the NAIRU. How do these two concepts relate to each other?

The answer is that the NAIRU is another name for the natural rate. The level of unemployment the economy “needs” in order to avoid accelerating inflation is equal to the natural rate of unemployment.

In fact, economists estimate the natural rate of unemployment by looking for evidence about the NAIRU from the behavior of the inflation rate and the unemployment rate over the course of the business cycle. For example, the way major European countries learned, to their dismay, that their natural rates of unemployment were 9% or more was through unpleasant experience. In the late 1980s, and again in the late 1990s, European inflation began to accelerate as European unemployment rates, which had been above 9%, began to fall, approaching 8%.

In Figure 31-4 we cited Congressional Budget Office estimates of the U.S. natural rate of unemployment. The CBO has a model that predicts changes in the inflation rate based on the deviation of the actual unemployment rate from the natural rate. Given data on actual unemployment and inflation, this model can be used to deduce estimates of the natural rate—and that’s where the CBO numbers come from. As of the final three months of 2014, the CBO estimate of the U.S. natural rate was 5.5%.

Through experience, policy makers have found that bringing inflation down is a much harder task than increasing it. The reason is that once the public has come to expect continuing inflation, bringing inflation down is painful.

A persistent attempt to keep unemployment below the natural rate leads to accelerating inflation that becomes incorporated into expectations. To reduce inflationary expectations, policy makers need to run the process in reverse, adopting contractionary policies that keep the unemployment rate above the natural rate for an extended period of time. The process of bringing down inflation that has become embedded in expectations is known as disinflation, a concept we learned about in Chapter 23.

Disinflation can be very expensive. As the following Economics in Action documents, the U.S. retreat from high inflation at the beginning of the 1980s appears to have cost the equivalent of about 18% of a year’s real GDP, the equivalent of roughly $2.6 trillion today. The justification for paying these costs is that they lead to a permanent gain. Although the economy does not recover the short-term production losses caused by disinflation, it no longer suffers from the costs associated with persistently high inflation. In fact, the United States, Britain, and other wealthy countries that experienced inflation in the 1970s eventually decided that the benefit of bringing inflation down was worth the required suffering—the large reduction in real GDP in the short term.

Some economists argue that the costs of disinflation can be reduced if policy makers explicitly state their determination to reduce inflation. A clearly announced, credible policy of disinflation, they contend, can reduce expectations of future inflation and so shift the short-run Phillips curve downward. Some economists believe that the clear determination of the Federal Reserve to combat the inflation of the 1970s was credible enough that the costs of disinflation, huge though they were, were lower than they might otherwise have been.