Loading field map
Why does inflation persist even when unemployment is high? The answer, macroeconomists have come to agree, depends on how people form expectations about future prices. The history of thinking about expectations and inflation is a story of successive refinements in how economists model the way households, firms, and financial markets anticipate the future—and what those models imply for the power of monetary and fiscal policy.
In the original Keynesian framework that dominated from the 1930s through the 1960s, expectations played a surprisingly passive role. The Phillips curve, which described a stable inverse relationship between unemployment and wage inflation, was treated as a reliable menu for policymakers. If the government wanted lower unemployment, it could accept higher inflation—and vice versa. This trade-off rested on the assumption that workers and firms formed expectations by looking at past inflation (static or extrapolative expectations) and did not adjust their behavior when policy changed. The framework gave policymakers considerable confidence: they could exploit the Phillips curve to manage aggregate demand without worrying that expectations would shift and undermine the relationship. But this confidence was a blind spot. The Keynesian model offered no mechanism for how expectations might update when inflation became persistent, leaving it vulnerable to the critique that would come from monetarists.
Milton Friedman and Edmund Phelps independently challenged the Keynesian Phillips curve in the late 1960s by introducing the natural-rate hypothesis. They argued that there is no long-run trade-off between inflation and unemployment. In the short run, an unanticipated increase in inflation can lower unemployment if workers mistake nominal wage rises for real wage gains. But once workers adapt their expectations to the new inflation rate—using adaptive expectations, where the expected inflation rate is a weighted average of past inflation—unemployment returns to its natural rate. The result was the expectations-augmented Phillips curve: inflation depends on expected inflation and the unemployment gap. Monetarism replaced the Keynesian static-expectations assumption with a backward-looking learning rule. This shift had a sharp policy implication: systematic attempts to keep unemployment below the natural rate would produce accelerating inflation, not lower unemployment. The framework narrowed the scope for activist demand management, but it still allowed short-run policy effects as long as expectations were slow to adjust.
New Classical Macroeconomics, developed by Robert Lucas, Thomas Sargent, and others in the 1970s, took the next logical step. If agents are rational, they argued, expectations should be formed using all available information, including knowledge of the policy regime itself. Adaptive expectations, which rely only on past data, are systematically wrong when policy changes. Rational expectations meant that anticipated monetary policy cannot affect real output or employment—only unanticipated surprises can. This led to the policy-ineffectiveness proposition: systematic stabilization policy is useless. The Lucas critique went further: any estimated relationship (like the Phillips curve) that does not account for how expectations respond to policy changes will break down when policy shifts. New Classical macroeconomics replaced adaptive expectations with rational expectations and, in doing so, narrowed the scope for policy even more than Monetarism had. The framework treated the economy as continuously clearing, with fluctuations driven by real shocks rather than demand management.
New Keynesian Macroeconomics emerged in the 1980s as a response to the New Classical challenge. It accepted rational expectations as the benchmark for modeling expectations but rejected the assumption of continuous market clearing. Instead, it introduced nominal rigidities—sticky prices and wages—that prevent the economy from adjusting instantly to shocks. In the New Keynesian Phillips curve, current inflation depends on expected future inflation and the output gap, but because prices are sticky, monetary policy can affect real activity even when it is anticipated. The framework absorbed rational expectations from New Classical macroeconomics but added frictions that restored a role for stabilization policy. The key difference: New Classical models implied that only unanticipated money matters; New Keynesian models showed that credible, anticipated policy can stabilize the economy by managing expectations about future inflation. This opened the door for a new kind of policy regime.
Inflation Targeting, adopted first by New Zealand in 1990 and later by dozens of central banks, translated New Keynesian insights into an operational regime. The framework commits the central bank to a publicly announced numerical inflation target and holds it accountable for achieving that target over a medium-term horizon. By making the target transparent and credible, Inflation Targeting aims to anchor private-sector expectations of future inflation. When expectations are anchored, the central bank can respond to supply shocks without triggering a wage-price spiral, and disinflation becomes less costly because the public believes the commitment. Inflation Targeting coexists with New Keynesian macroeconomics as the dominant academic model for monetary policy analysis. It replaced earlier Monetarist-style money-supply targeting by focusing directly on inflation rather than intermediate targets. The framework's distinctive contribution is not a new theory of expectation formation but a practical mechanism for shaping expectations through institutional design.
Behavioral Macroeconomics, which gained momentum after 2000, challenges the rational-expectations assumption that underpins both New Classical and New Keynesian models. Drawing on psychology and experimental evidence, behavioral models assume that agents have bounded rationality: they process information slowly, rely on heuristics, and update expectations in heterogeneous ways. Two influential strands are sticky-information models (Mankiw and Reis, 2002), where firms update their plans infrequently, and heterogeneous-expectations models (Branch and Evans, 2006), where agents switch between forecasting rules depending on past performance. These models can generate persistent inflation dynamics and hump-shaped responses to shocks that standard rational-expectations models struggle to replicate. Behavioral Macroeconomics does not aim to replace New Keynesian macroeconomics entirely; rather, it modifies the expectations block within New Keynesian frameworks, often coexisting as a refinement. The approach remains an active frontier, with ongoing debates about whether bounded rationality or sticky prices better explains observed inflation persistence.
Today, New Keynesian Macroeconomics and Inflation Targeting are the leading frameworks in academic research and central-bank practice. There is broad agreement that the long-run Phillips curve is vertical—that is, inflation has no permanent effect on unemployment—and that the credibility of monetary policy matters for anchoring expectations. The main disagreements center on the microfoundations of expectation formation. Rational-expectations models remain the workhorse for policy analysis because they are tractable and discipline the model with consistency. But behavioral and heterogeneous-expectations models are gaining ground, especially for explaining episodes like the Great Inflation of the 1970s and the low-inflation puzzle of the 2010s. A key open question is the size of the sacrifice ratio—how much output must be lost to reduce inflation—which depends critically on how quickly expectations adjust. Another dispute concerns the role of forward guidance: rational-expectations models predict powerful effects, while behavioral models suggest that households and firms may not fully understand or believe central-bank announcements.
The journey from static expectations to adaptive, rational, and finally bounded expectations reflects a progressive deepening of macroeconomists' understanding of how beliefs shape inflation outcomes. Each framework refined the treatment of expectation formation and, in doing so, altered the perceived scope for policy. Keynesian economics gave policymakers a trade-off that vanished when expectations adapted. Monetarism and New Classical macroeconomics progressively narrowed that scope by making expectations more forward-looking. New Keynesian macroeconomics restored a role for policy by combining rational expectations with sticky prices, and Inflation Targeting provided the institutional tools to anchor those expectations. Behavioral macroeconomics now pushes the frontier by questioning whether rational expectations are a realistic description of how people actually think. The subfield remains vibrant because the central question—how expectations are formed and how they affect inflation—is both theoretically deep and practically urgent.