The Phillips Curve: Inflation vs Unemployment

The original observation

In 1958 the economist A.W. Phillips plotted UK wage growth against unemployment from 1861 to 1957 and noticed a clear negative relationship: when unemployment was low, wages (and prices) rose faster. Policymakers in the 1960s took this as a menu — pick the combination of unemployment and inflation you want and pursue policies that get you there.

That worked for about a decade. Then in the 1970s, inflation rose sharply while unemployment was high. The simple Phillips trade-off seemed to have broken. What went wrong?

Expectations

The fix is the expectations-augmented Phillips curve:

$$\pi = \pi^e - \beta (u - u_n)$$

In words: inflation $\pi$ equals what people expect inflation to be ($\pi^e$) plus a short-run kick from the unemployment gap $(u - u_n)$. If unemployment is below the natural rate, firms compete for scarce workers, wages rise faster than expected, prices rise faster than expected, $\pi$ runs hot.

But over time, $\pi^e$ catches up. People are not stupid; if inflation has been running at 6% for three years, they start expecting 6%. The whole Phillips curve shifts up. To keep unemployment below the natural rate you’d need ever-faster surprise inflation — and once people figure that out, you can’t surprise them anymore.

That’s why in the long run the Phillips curve is vertical at $u_n$. There is no permanent trade-off; the only sustainable unemployment rate is the natural one.

Play with it

Expected inflation πᵉ (%): 2.00Natural rate uₙ (%): 5.00Slope β: 0.50Actual unemployment u (%): 5.00

Today's inflation π = 2.00%

Pin a low unemployment rate. Slowly raise the expected-inflation slider $\pi^e$ and watch the curve drift up. Today’s actual inflation moves with it. That’s the textbook story of the 1970s — sustained low unemployment unanchored expectations, and inflation kept ratcheting up.

The 2022 episode

When US inflation hit 9% in mid-2022, the question on every policymaker’s mind was: have expectations come unanchored, the way they did in the 1970s? Survey measures of $\pi^e$ rose somewhat but stayed well below 9%. Bond markets’ implied $\pi^e$ stayed near 2.5%. The Fed bet that expectations were still anchored and tightened aggressively to bring actual inflation back down — which it did, without unemployment spiking.

That’s the textbook soft landing. It worked partly because the Phillips-curve framework gives policymakers a clear test: watch expectations, not just realized inflation.

A modest takeaway

In the short run there’s a real trade-off. Stimulating an economy that’s already at full employment pushes inflation up; cooling an overheated economy raises unemployment. In the long run, neither policy can buy permanently lower unemployment — only better matching of workers to jobs (training, labor-market reform) moves the natural rate itself.