Phillips curve

For a while, economists thought they had found a stable menu: pick your inflation, accept the matching unemployment. The Phillips curve, and its breakdown, taught a harder lesson.

The Phillips curve describes an observed inverse relationship between inflation and unemployment: historically, lower unemployment tended to come with higher inflation, and higher unemployment with lower inflation. It suggested a trade-off that policymakers might exploit, and its later failure reshaped macroeconomics.

The apparent trade-off

The original relationship, drawn from decades of data, implied that an economy running hot, with low unemployment, would see wages and prices rise faster, while a slack economy with high unemployment would see inflation ease. This seemed to offer governments a menu: accept a little more inflation to buy lower unemployment, or the reverse. For a time, demand management was conducted as if this trade-off were stable and reliable.

The breakdown

Then came stagflation, high inflation and high unemployment together, which the simple Phillips curve said should not happen. Economists, led by Milton Friedman and Edmund Phelps, had anticipated why: once people come to expect inflation, they build it into wage demands and prices, so the trade-off vanishes. Trying to keep unemployment below its natural rate simply produces ever-accelerating inflation, not a stable lower unemployment. The curve, it turned out, shifted with expectations and held only in the short run.

What survives

The modern view keeps a short-run Phillips curve, where unexpected demand changes do move inflation and unemployment in opposite directions, but denies any stable long-run trade-off. In the long run the economy gravitates to a natural rate of unemployment set by supply-side factors, regardless of inflation. The role of expectations, which the early version ignored, is now central, and the credibility of central banks in anchoring those expectations is treated as crucial.

The Phillips curve is a case study in the limits of empirical relationships in economics: a pattern that held until policymakers tried to exploit it, at which point behaviour adjusted and the pattern broke. Its history is a standing caution against treating a historical correlation as a stable trade-off a government can rely on.