Unit 4 Inflation and unemployment

4.6 Expected inflation shifts the Phillips curve

In his presidential address to the American Economic Association in December 1967, Milton Friedman provided an explanation for why people’s expectations about inflation could change and why that would shift the Phillips curve. He referred to the recent experience in the US. Since 1966, unemployment had been steady, averaging 3.7%, but inflation had increased from 3.0% to 4.2%. He said that the only way unemployment could be kept as low as 3% was by allowing inflation to keep increasing to higher levels: ‘There is always a temporary trade-off between inflation and unemployment; there is no permanent trade-off,’ he claimed.¹

Supported by evidence from many countries from the late 1960s, he showed that if unemployment remains ‘too low’ the result will be not just higher inflation, but rising inflation as well. We will discuss what ‘too low’ means and how this is represented in the model by a shift in the Phillips curve.

expected inflation

The belief formed by wage-setters and price-setters about the level of inflation in the next period. See also: inflation.

Friedman claimed that at low unemployment, inflation keeps increasing so that the simple Phillips curve relationship like the one plotted by Bill Phillips is misleading. Why does inflation keep rising when unemployment is kept low? This is because of the way wage setters and price setters form their views about what will happen to inflation, which is called expected inflation.

We need to keep in mind that households care about wages and firms care about profits in real terms, and consider the following two points:

• People are forward-looking: They take actions now in anticipation of things they expect to happen. To stress this, economists say that ‘expectations matter’.
• People treat prices as messages: People treat changes in prices as messages about what will happen in the future, just as people treat a build-up of clouds as a prediction of rain.

The way that people expect inflation to behave will depend on how they make forward-looking decisions and how they interpret price messages.

Introducing expected inflation

We introduce the role of expected inflation (which we denote with the symbol \(\pi^E\)) by returning to the Phillips curve.

Suppose that the economy is in supply-side equilibrium at the intersection of the wage- and price-setting curves, with a real wage of 100 and unemployment at 6%. But everyone expects the inflation rate to be 3%. So there is no bargaining gap: each year both wages and prices will be raised by 3% and the real wage will remain at the intersection of the WS and PS curves.

In Figure 4.11, at the supply-side equilibrium with an unemployment rate of 6%, the inflation rate is 3% and not zero as in Figure 4.9. Once we introduce inflation expectations, the economy can be in supply-side equilibrium with rising prices—as long as they are rising at the rate expected.

If wage setters and price setters expect prices to rise by 3% per annum, and the level of aggregate demand is ‘normal’ and keeps unemployment at 6%, then the economy can remain at the supply-side equilibrium with inflation remaining constant at 3% per annum. Actual inflation is equal to expected inflation and the level of unemployment is constant at the equilibrium where the WS curve and PS curve intersect (point A).

inflation-stabilizing unemployment rate

The unemployment rate (at supply-side equilibrium) at which inflation is constant. Originally known as the ‘natural rate’ of unemployment. Also known as: structural unemployment rate, non-accelerating rate of unemployment (NAIRU). See also: equilibrium unemployment, structural unemployment.

At the supply-side equilibrium, where the WS and PS curves intersect, the equilibrium or structural unemployment rate is also the inflation-stabilizing unemployment rate.

Now consider a business cycle upswing. Aggregate demand rises, taking the economy to lower unemployment of 4% at point B. What will happen to inflation? Workers expect prices to rise by 3% and will require a nominal wage increase of 3% just to keep their real wage unchanged. But they require an additional 2% rise to give them an expected real wage rise on the wage-setting curve, so wages increase by 5%. With their costs rising by 5%, firms will increase prices by 5% to maintain their profit-maximizing markup. In the boom, inflation will be 5% with unemployment at 4%. This gives a Phillips curve like the one in Figure 4.9. The only difference is that inflation at supply-side equilibrium is 3% rather than zero.

When expected inflation is not zero, we can summarize the causal chain to inflation like this:

To work out the inflation rate:

\[\begin{align*} \text{inflation (%)} &\equiv \text{increase in prices (%)} \\ &= \text{increase in costs per unit of output (%)} \\ &= \text{increase in wages (%)}(\text{if wages are the only costs}) \\ &= \text{expected inflation (%) + bargaining gap (%)} \\ \pi_t &= \pi_t^E + \text{gap}_t \end{align*}\]

But Friedman pointed out that with continued low unemployment, inflation would continue rising rather than remain at 5% at point B. To understand why, we ask what happens next. The equation for inflation gives a hint—what will happen to expected inflation?

Expected inflation and the shifting Phillips curve

With low unemployment continuing, workers will be disappointed with the outcome, since they did not achieve their expected real wage. Why not? Workers expected a 2% real wage increase at B from their nominal pay rise of 5% (to give the real wage on the wage-setting curve), but they did not get this because firms raised their prices by 5%, thereby wiping out the real wage increase.

But the story does not end there. We know it is impossible for both parties to be satisfied with the outcome at low unemployment, because their claims add up to more than the size of the pie. What happens next depends on what workers expect inflation to be over the coming year. Higher expected inflation will shift the Phillips curve up.

We will model expected inflation by assuming that workers expect inflation over the year ahead to be equal to inflation over the last year—in this case, 5%. At the next wage-setting round, the human resources department has to take into account the fact that their employees expect prices to rise by 5%. So, in order to achieve a real wage increase of 2%, the size of the bargaining gap, the HR department sets a nominal wage increase of 7%.

In response, the marketing department will increase prices to compensate for the higher unit costs, and the rate of inflation continues to increase.

The table in Figure 4.13 summarizes the situation. We compare inflation over a three-year period with unemployment at two different levels: 6% and 4%.

The Expected inflation column of Figure 4.13 reflects forward-looking behaviour. Each year, expected inflation over the year ahead is equal to the previous year’s inflation. Then expected inflation together with the bargaining gap determines the actual inflation outcome, shown in the final column.

We can summarize the causal chain from the last period’s inflation rate to this period’s inflation rate like this:

To work out the inflation rate, we now have an extra step using \(\pi_t^E = \pi_{t-1}\):

\[\begin{align*} \text{inflation (%)} &\equiv \text{increase in prices (%)} \\ &= \text{increase in costs per unit of output (%)} \\ &= \text{increase in wages (%)}\text{(if wages are the only costs)} \\ &= \text{expected inflation (%)} + \text{bargaining gap (%)} \\ &= \text{last period's inflation} + \text{bargaining gap (%)} \\ \pi_t &= \pi_{t-1} + \text{gap}_t \end{align*}\]

The table in Figure 4.13 shows how a boom, in which unemployment falls below the equilibrium level (6%) to 4%, results in rising inflation. Figure 4.15 illustrates the same situation using the Phillips curve and WS–PS diagrams. Work through the diagrams below and then read the explanation that follows.

wage–price spiral

This occurs if an initial increase in wages in the economy is followed by an increase in the price level, which is followed by an increase in wages and so on. It can also begin with an initial increase in the price level.

At point A, with unemployment of 6% and inflation of 3%, inflation remains stable, year after year. A boom reduces unemployment to 4%. When inflation expectations in the year are 3%, there is a movement along the Phillips curve to B and inflation is 5%. In the next year, at the same low unemployment, inflation is expected to be 5% and the Phillips curve shifts up from the one through point B to the one through point C. This is called the wage–price spiral. It explains why, at low unemployment, inflation rises, not just in the year that unemployment fell, but year after year.

By plotting the path of inflation over time in Figure 4.16, we can determine the distinctive contributions of the bargaining gap and expected inflation to inflation. In this example, the bargaining gap opens up in year 1 because of the move to low unemployment. The assumption that unemployment remains below the inflation-stabilizing rate is reflected in the persistence of the bargaining gap. Inflation rises in every period because the previous period’s inflation feeds into expected inflation and therefore into wage and price inflation. Note that the real wage does not change, but remains on the price-setting curve.

To summarize, inflation will continue to rise period by period if the bargaining gap persists and the expected rate of inflation rises, shifting the Phillips curve up. We can model expected inflation over the year ahead by assuming it is equal to inflation over the past year.