How Did the 1970s Stagflation Cause a Revolution in Economic Theory?

In 1958, New Zealand economist A. W. Phillips published a paper documenting a stable statistical relationship in British data: when unemployment was low, wage inflation was high; when unemployment was high, wage inflation was low. The relationship was christened the "Phillips curve" and was quickly generalized into a policy tool—a menu of trade-offs that policymakers could navigate, choosing more inflation for less unemployment or vice versa. The Phillips curve became the analytical centerpiece of 1960s Keynesian demand management. In the 1970s, the curve moved, then broke entirely. Stagflation—both unemployment and inflation high simultaneously—was not merely a policy failure; it was a falsification of the theoretical prediction that inflation and unemployment could not rise together. The response was a revolution in macroeconomic theory that replaced the simple Phillips curve with a more sophisticated framework incorporating inflation expectations, natural rate dynamics, and the importance of monetary credibility.

Quick definition: The Phillips curve breakdown refers to the 1970s empirical failure of the assumed stable trade-off between inflation and unemployment, as stagflation produced simultaneously high inflation and high unemployment—validating the Friedman-Phelps critique that the short-run Phillips curve shifted with inflation expectations, that there was a "natural rate" of unemployment that monetary policy could not permanently reduce, and that sustained inflation policy would ultimately produce only higher inflation without lower unemployment.

Key takeaways

• The original Phillips curve (1958) documented a stable empirical relationship between unemployment and wage inflation in UK data; it was generalized into a policy tool by Samuelson and Solow (1960), suggesting policymakers could choose along a stable inflation-unemployment trade-off.
• Milton Friedman (1968) and Edmund Phelps (1968) independently predicted that the short-run Phillips curve was not stable: once workers and firms adjusted their expectations to a new inflation rate, the curve would shift upward, requiring ever-higher inflation to maintain any given unemployment level.
• The 1970s stagflation empirically confirmed the Friedman-Phelps critique: the Phillips curve shifted upward repeatedly; inflation and unemployment rose together, contradicting the stable trade-off assumption.
• The natural rate hypothesis—that there is a long-run "natural" unemployment rate that monetary policy cannot reduce without continuously increasing inflation—became the foundation of modern macroeconomics.
• The New Keynesian synthesis that emerged from the theoretical revolution incorporated rational expectations, credibility, and the natural rate into a more sophisticated framework that has guided central bank practice since.
• Modern central bank practice—inflation targeting, forward guidance, credibility management—is directly derived from the theoretical lessons of the 1970s Phillips curve breakdown.

The original Phillips curve

A. W. Phillips's 1958 paper examined UK wage and unemployment data from 1861 to 1957 and documented a consistent negative relationship: lower unemployment was associated with faster wage growth; higher unemployment was associated with slower wage growth or wage cuts. The relationship appeared stable across nearly a century of data.

Paul Samuelson and Robert Solow (both eventually Nobel laureates) adapted the Phillips curve to US data in 1960 and reframed it in policy terms: the curve provided a menu of trade-offs between unemployment and price inflation (using price inflation rather than wage inflation) that policymakers could navigate. The Kennedy and Johnson administrations explicitly used the Phillips curve framework in their policy decisions—accepting modestly higher inflation to reduce unemployment in the early 1960s.

The empirical relationship seemed to hold through the early 1960s. Unemployment fell from approximately 7 percent in 1961 to approximately 4 percent by 1966; inflation rose from approximately 1 percent to approximately 3 percent. The data points moved along the predicted curve.

The relationship began deteriorating in the late 1960s as Vietnam War spending pushed unemployment below 4 percent while inflation rose above the curve's prediction. By the early 1970s, the data no longer fit the 1960s curve at all; inflation was higher at every unemployment rate than the curve predicted.

The Friedman-Phelps critique

Milton Friedman's 1968 presidential address to the American Economic Association and Edmund Phelps's 1968 paper—published independently but making similar arguments—provided the theoretical explanation for why the Phillips curve was unstable:

The expectational Phillips curve: The actual unemployment-inflation relationship holds only in the short run, when workers and firms are surprised by inflation. If workers expect 0 percent inflation, a 3 percent inflation rate is a surprise that makes firms willing to hire more workers (their real labor costs are lower than workers expected) and makes workers willing to accept lower real wages (they don't realize their real wages are falling). Unemployment can be reduced below the natural rate temporarily—but only through inflation surprise.

The natural rate of unemployment: There is a long-run equilibrium unemployment rate—the "natural rate"—determined by structural factors: job matching dynamics, search frictions, skill mismatches, unemployment insurance generosity. This natural rate is not a policy variable; it cannot be permanently reduced by monetary policy.

The long-run Phillips curve is vertical: In the long run, when expectations have fully adjusted, there is no trade-off between inflation and unemployment. At any inflation rate, the economy gravitates toward the natural rate of unemployment. A higher inflation rate does not purchase permanently lower unemployment—it only temporarily reduces unemployment while the inflation is higher than expected.

The accelerationist implication: If policymakers try to maintain unemployment below the natural rate through monetary stimulus, they must keep inflation above expected inflation indefinitely—which requires continuously accelerating inflation. Any attempt to hold unemployment at 3 percent when the natural rate is 5 percent would eventually require hyperinflation. The only way to avoid accelerating inflation is to allow unemployment to return to the natural rate.

The rational expectations revolution

The Friedman-Phelps critique assumed adaptive expectations—workers and firms adjust their expectations based on past experience (last year's inflation plus some adjustment). Robert Lucas and Thomas Sargent pushed further in the early 1970s: what if expectations were rational rather than adaptive? What if workers and firms used all available information efficiently to forecast future inflation?

The rational expectations hypothesis (Lucas 1972, Sargent 1975) had powerful implications:

Policy ineffectiveness in the long run: If economic actors form expectations rationally using all available information, systematic monetary policy that is predictable will be incorporated into expectations immediately—before it affects employment. Only unanticipated monetary policy can have real effects on unemployment; anticipated policy affects only prices.

The credibility revolution: If expectations are forward-looking and information-efficient, then a central bank's announcements matter—not just its actions. A credible commitment to low inflation, if believed by all economic actors, would lower inflationary expectations immediately, reducing the output cost of disinflation. The credibility of the commitment was the policy variable.

Time inconsistency: Finn Kydland and Edward Prescott (1977 paper) showed that monetary policy faces a time consistency problem: the optimal policy announced in advance (credible commitment to low inflation) differs from the optimal policy in the moment (inflate to reduce unemployment). This creates incentives for central banks to renege on announced low-inflation policies, which rational economic actors will anticipate—leading to higher inflation even without explicit policy failure. The solution is institutional commitment devices that make reneging costly.

The New Keynesian synthesis

The theoretical revolution produced by the 1970s Phillips curve breakdown ultimately generated a "New Keynesian" synthesis that incorporated the key insights while preserving the role for monetary policy:

Sticky prices: Firms don't adjust prices instantly; they set prices based on expectations of future costs and demand. This creates a short-run trade-off between inflation and output that monetary policy can influence—but only temporarily and only if not fully anticipated.

Expectations-augmented Phillips curve: The modern version incorporates expected inflation explicitly: actual inflation equals expected inflation plus a term reflecting the output gap (actual vs. potential output) plus supply shock terms. Managing inflation requires managing expectations; credibility determines how costly disinflation is.

Inflation targeting: The practical policy implication of the New Keynesian synthesis is that central banks should announce explicit inflation targets, build credibility through consistent policy, use forward guidance (communicating future policy intentions) to manage expectations, and treat their credibility as a strategic asset.

The sacrifice ratio: The output cost of disinflation—the "sacrifice ratio"—is lower when central bank credibility is higher. A credible disinflation announcement shifts expectations immediately; less actual unemployment is required to demonstrate commitment. Volcker's high sacrifice ratio (10.8 percent unemployment) reflected the low credibility inherited from Burns-era accommodation; a credible central bank could achieve disinflation with lower unemployment costs.

Modern central banking as product of the 1970s revolution

Every major element of modern central bank practice traces to the theoretical revolution that the 1970s stagflation motivated:

Inflation targeting: Adopted by New Zealand in 1990, the UK in 1992, and explicitly by the Federal Reserve in 2012—the commitment to a specific inflation rate (typically 2 percent) provides the expectational anchor that the natural rate hypothesis shows is essential.

Central bank independence: The time inconsistency problem identified by Kydland and Prescott explains why political control of monetary policy produces inflation bias; independent central banks with credibility as their primary asset are less subject to the inflation temptation. The Fed's quasi-independence was reinforced by the 1970s experience.

Forward guidance: The rational expectations revolution showed that announced future policy affects expectations immediately. Central banks now use explicit forward guidance—statements about future interest rate paths—as a policy tool supplementing the current interest rate.

Quantitative easing credibility challenges: Post-2008 quantitative easing—large-scale asset purchases—raised questions about whether the Fed could maintain credibility in a zero-lower-bound environment; the theoretical framework of credibility and expectations helped structure those debates.

Real-world examples

The European Central Bank's post-2012 evolution illustrates the ongoing influence of 1970s theoretical lessons. ECB President Mario Draghi's 2012 "whatever it takes" statement—a credibility commitment that reversed the eurozone sovereign debt crisis without a single euro being spent—demonstrated the power of credible central bank communication that the rational expectations revolution had identified as a policy tool. The commitment to act credibly changed expectations; changed expectations resolved the crisis.

The Fed's 2021-2022 "transitory" inflation misjudgment—and the subsequent aggressive rate increases to restore credibility—similarly reflected the theoretical framework. When the Fed acknowledged that inflation was not transitory and began aggressive tightening, the explicit goal was to restore the credibility that the 2021 misjudgment had damaged—a credibility-management framework directly derived from 1970s theory.

Common mistakes

Treating the Phillips curve as completely useless. The original Phillips curve captured a real short-run phenomenon that the expectations-augmented version preserves: in the short run, when inflation surprises, there is a trade-off. The critique is that this trade-off is not stable and cannot be exploited systematically. The curve exists; it just shifts.

Treating the natural rate as a precise, known number. The natural rate of unemployment is a theoretical construct that is not directly observable. Estimates for the US have ranged from 4 to 6 percent over different periods. The Fed's inability to precisely identify the natural rate in real time contributed to the 1960s-70s policy errors; acknowledging uncertainty about the natural rate is built into modern policy frameworks.

Confusing Keynesian economics with the original Phillips curve. Modern Keynesian and New Keynesian economics incorporates the Friedman-Phelps critique—expectations matter, the natural rate exists, credibility is central. Characterizing "Keynesian economics" as the pre-1970s simple Phillips curve model misrepresents how the framework has evolved.

FAQ

Was Milton Friedman entirely right about the Phillips curve?

Friedman was right about the non-stability of the simple Phillips curve and the role of expectations. He was more controversial about the superiority of monetary rules over discretion—the subsequent New Keynesian synthesis preserved a role for discretionary monetary policy (particularly in response to demand shocks) while acknowledging that systematic inflation bias requires institutional constraints. The full rational expectations implications (systematic policy is ineffective) have not been entirely borne out by empirical evidence; wages and prices do exhibit some stickiness that allows monetary policy real effects.

Has the Phillips curve fully recovered its predictive power after the 1970s?

The expectations-augmented Phillips curve has had periods of better and worse predictive performance. During the "Great Moderation" (1985-2007), when inflation was well-anchored at low levels, the Phillips curve's output gap term was often small, making the curve difficult to identify empirically. In the 2021-2023 episode, the Phillips curve dynamic—tight labor markets contributing to inflation—seemed more visible, though supply-shock effects complicated interpretation. The relationship exists but is neither as simple as the original Phillips curve nor as cleanly identifiable as its critics suggest.

Related concepts

• Stagflation and the 1970s
• Supply Shock Economics
• The Volcker Shock
• The 1973-74 Bear Market
• Lessons from the Oil Shock

Summary

The 1970s stagflation produced the most significant revolution in macroeconomic theory since Keynes's General Theory—the Phillips curve breakdown demonstrating that the assumed stable inflation-unemployment trade-off was not stable and could not be systematically exploited. Friedman and Phelps had predicted in 1968 that expectations-adjustment would shift the Phillips curve; the 1970s empirically confirmed this prediction. The theoretical response—rational expectations, time inconsistency, the natural rate hypothesis, the expectations-augmented Phillips curve—produced the New Keynesian synthesis that governs modern central banking. Inflation targeting, central bank independence, forward guidance, and credibility management are direct institutional responses to the theoretical lessons of the 1970s. The most practical implication: central banks that allow inflation expectations to become unanchored face higher sacrifice ratios when they eventually tighten—more unemployment per point of inflation reduction. Volcker's high sacrifice ratio illustrated this principle; subsequent credibility management has been aimed at maintaining the anchored expectations that reduce future sacrifice ratios.

Next

→ The 1974 Recession: Depth and Recovery