Intermediatemacroeconomics
The Phillips Curve: Inflation vs Unemployment
Overview
Few diagrams have shaped policy more than the Phillips curve — the inverse relationship between unemployment and inflation. Governments once thought it offered a menu of choices; then the 1970s stagflation seemed to destroy it; then it reappeared, flatter and expectations-driven. This lesson traces the curve's three lives — original, expectations-augmented, New Keynesian — and explains what it can and cannot tell policymakers today.
Learning objectives
- Describe the original Phillips curve and its policy interpretation.
- Explain the expectations-augmented curve and the long-run vertical result.
- Define the NAIRU and explain its role.
- Discuss why the curve has flattened and what that implies.
The original discovery
In 1958, A. W. Phillips plotted a century of British data and found a stable inverse curve: low unemployment came with fast wage growth, high unemployment with slow or falling wages. Paul Samuelson and Robert Solow translated it into an inflation-unemployment 'menu' for the US: policymakers could choose their preferred point — a little more inflation for a little less unemployment. In the 1960s, governments believed they had found a reliable exploit in the economy's machinery.
Pick your point on the curve: accept 4% inflation to get 4% unemployment, or 2% inflation at 6% unemployment. The tradeoff looked stable — until it wasn't.
Friedman and Phelps: expectations strike back
Milton Friedman (1968) and Edmund Phelps argued the menu was an illusion. Workers bargain over real wages: if they expect 5% inflation, they demand 5% higher nominal pay upfront, neutralizing any stimulus. Only surprise inflation can temporarily push unemployment below its 'natural rate.' The long-run Phillips curve is therefore vertical at the natural rate (NAIRU): any sustained attempt to hold unemployment lower just accelerates inflation without limit. The 1970s stagflation — high inflation and high unemployment together — was the devastating confirmation.
Push unemployment below NAIRU and expectations ratchet up each round: 5% inflation expected leads to 5% built into wages, requiring 7% to surprise again — the wage-price spiral logic.
The NAIRU: the anchor of modern policy
The Non-Accelerating Inflation Rate of Unemployment (NAIRU) is the unemployment rate consistent with stable inflation — determined by labor-market structure: demographics, skills mismatch, hiring frictions, labor laws, union power. It is unobservable and drifts over time (estimates for the US fell from ~6% in the 1990s toward ~4% recently). Central banks estimate it constantly: unemployment below NAIRU signals coming inflation; above it signals slack. Getting the estimate wrong means mistimed policy — the perennial central-banking anxiety.
Frictional and structural unemployment always exist — people between jobs, skills mismatches. The NAIRU is the efficient-ish floor, not a policy failure.
The short run vs the long run
The modern synthesis: a short-run Phillips curve slopes downward — demand stimulus can temporarily lower unemployment at the cost of higher inflation — but it shifts with expectations, and the long-run curve is vertical at the NAIRU. Credible central banks flatten even the short-run curve: anchored expectations mean demand swings move output with little inflation. This expectations-dependence is why central bankers obsess over 'anchoring' — credibility literally reshapes the tradeoff they face.
The flattening puzzle
Since the 1990s the Phillips curve has gone strikingly flat: unemployment fell to historic lows with barely a ripple in inflation — until 2021–22, when it roared back. Explanations compete: globalization disciplining prices, better-anchored expectations, mismeasured slack (gig work, underemployment), or nonlinearities — flat when inflation is low, steep when it surges. The 2021 inflation burst reminded everyone the curve is not dead, merely dormant at low inflation. Humility about its shape is now standard.
What the curve means for policy today
Policymakers use the framework daily: estimate the output gap and NAIRU, forecast inflation from expected demand pressure, and set rates to keep the economy near — not below — the natural rate. Fiscal authorities watch it too: stimulus into a below-NAIRU economy mostly buys inflation. The Phillips curve no longer offers a exploitable menu, but as a diagnostic — where is demand relative to capacity, and are expectations anchored? — it remains indispensable.
Key points
- Original curve (1958): stable inflation-unemployment tradeoff — the 1960s 'menu' view.
- Expectations-augmented: only surprise inflation moves unemployment; long-run curve vertical at NAIRU.
- NAIRU = unemployment consistent with stable inflation; unobservable, drifting, structural.
- Short-run tradeoff exists but shifts with expectations; credible policy flattens it.
- The curve flattened after the 1990s but revived in 2021–22 — shape depends on context.
Common misconceptions
✓ Truth: That was the 1960s menu view, refuted by stagflation. The long-run curve is vertical: permanent below-NAIRU unemployment buys accelerating inflation, not stable higher inflation.
✓ Truth: It disproved the stable-tradeoff version, not the expectations-augmented one — which predicts exactly that supply shocks and de-anchored expectations can raise both simultaneously.
✓ Truth: Flat at low inflation is consistent with steep at high inflation (nonlinearity). 2021–22 showed the tradeoff reawakening once shocks hit and expectations stirred.
Practice questions
Q1. The expectations-augmented Phillips curve predicts that a permanent attempt to hold unemployment below the NAIRU will cause:
- Permanently lower unemployment with stable inflation
- Accelerating inflation with no lasting unemployment gain
- Deflation
- A permanently flatter curve
Each round of surprise inflation gets built into expectations, requiring ever-higher inflation for the same surprise. Long-run unemployment returns to NAIRU; only inflation is higher — the accelerationist result.
Q2. Stagflation in the 1970s is best explained in this framework as:
- A movement along a stable Phillips curve
- Adverse supply shocks plus de-anchored expectations shifting the short-run curve outward
- Excessive fiscal surpluses
- A sudden fall in the NAIRU to zero
Oil shocks raised inflation at every unemployment level while expectations de-anchored — the short-run curve shifted out. The stable-menu view could not produce high inflation with high unemployment.
Q3. A flatter Phillips curve implies that:
- Small demand changes cause large inflation swings
- Demand fluctuations move unemployment with relatively little inflation change
- The NAIRU is zero
- Expectations are completely unanchored
Flatness means the inflation cost of demand variation is small — consistent with well-anchored expectations. (Caveat: flatness at low inflation may not hold when inflation is high.)
Click a question to reveal the answer and explanation.
Frequently asked questions
What is the difference between the short-run and long-run Phillips curve?
Short-run: downward sloping — demand shifts move unemployment and inflation oppositely, given fixed expectations. Long-run: vertical at the NAIRU — once expectations fully adjust, no permanent tradeoff remains.
What causes the NAIRU to change?
Structural labor-market factors: demographics, skill mismatches, labor regulation, unemployment benefit generosity, union density, hiring technology. It drifts over decades, which is why real-time estimates are so uncertain.
What is a supply shock's effect on the Phillips curve?
An adverse supply shock (oil spike) shifts the short-run curve outward — higher inflation at every unemployment rate. Policymakers then face a cruel choice: fight inflation with unemployment, or accommodate inflation to protect jobs.
Why do central bankers talk about 'anchored expectations' constantly?
Anchored expectations flatten the short-run tradeoff and keep the long-run curve vertical at low inflation — making stabilization nearly painless. De-anchored expectations steepen everything and force harsher tightening.