Expectations and the Phillips Curve: Adaptive vs. Rational

In the late 1960s, economists thought they had cracked the code on managing the economy. The Phillips Curve suggested a simple trade-off: accept a bit more inflation, and you could push unemployment down. Governments could seemingly dial in their preferred mix of these two variables like adjusting a thermostat. Then came the 1970s, and everything fell apart. Inflation soared while unemployment simultaneously climbed-a phenomenon called stagflation that the original Phillips Curve said was impossible. The missing ingredient? People’s expectations about future inflation.

Why expectations became the game changer 🔗

Think about negotiating your salary. If you expect prices to rise by 5% next year, you’re going to push for at least a 5% raise just to maintain your purchasing power. This intuitive logic-first formalized by economists Milton Friedman and Edmund Phelps in the late 1960s-fundamentally transformed how we understand the relationship between inflation and unemployment. They argued that workers and employers don’t suffer from “money illusion.” Rather, they care about real wages-what their paychecks can actually buy after accounting for inflation.

When the government tries to boost employment through expansionary policy, it initially works. Firms see higher demand and hire more workers. Unemployment falls. But here’s the catch: workers eventually notice that prices are rising faster than their wages. They start demanding higher pay to catch up, and firms pass these costs onto consumers through higher prices. The economy ends up back where it started in terms of unemployment, but with a permanently higher inflation rate. This insight explains why stagflation plagued advanced economies during the 1970s, confounding policymakers who had relied on the simple Phillips Curve framework.

Adaptive expectations and the short-run trade-off 🔗

Under the adaptive expectations framework, workers form beliefs about future inflation by looking backward at historical data. If inflation was 3% last year, people expect roughly 3% inflation this year. This backward-looking approach creates a short-run Phillips Curve where expansionary policy can temporarily reduce unemployment below its natural rate.

Imagine an economy where inflation has been stable at 2% for years, and unemployment sits at 5%-the natural rate. The government decides to stimulate the economy through increased spending or easier monetary policy. Demand surges, firms increase production, and unemployment drops to 4%. Workers see their nominal wages rising and supply more labor, not immediately realizing that prices are climbing even faster. In this short window, the government has seemingly achieved its goal of lower unemployment.

But this victory is temporary. As workers gradually adjust their inflation expectations upward, they demand higher wages to protect their purchasing power. The short-run Phillips Curve shifts outward. At each level of unemployment, there’s now higher inflation than before. To maintain unemployment at 4%, the government would need to accept even higher inflation-say 5% or 6%. The trade-off has worsened, creating a painful dilemma for policymakers.

The mechanics of adaptive adjustment 🔗

The speed of this adjustment matters enormously. If workers quickly update their expectations when they observe rising prices, the short-run period where policy can affect unemployment becomes very brief indeed. The formula for adaptive expectations is straightforward: expected inflation this period equals actual inflation from last period. This means each round of inflation feeds into future expectations, creating a self-reinforcing cycle.

Consider the American experience in the early 1970s. Between 1968 and 1970, unemployment rose from 3.6% to 4.9% while inflation increased from 4.7% to 5.6%. Expected inflation in the Michigan survey climbed from 3.8% to 4.9% during this same period. This data strongly supports the expectation-augmented Phillips Curve framework-even as the economy weakened, inflation kept rising because expectations had become unanchored.

The long-run vertical Phillips Curve 🔗

When workers fully adapt their expectations to match actual inflation, something remarkable happens: the Phillips Curve becomes vertical at the natural rate of unemployment. This means that in the long run, there’s no trade-off between inflation and unemployment at all. The economy gravitates toward a single unemployment rate-sometimes called the Non-Accelerating Inflation Rate of Unemployment or NAIRU-regardless of the inflation rate.

Why does this occur? Because once everyone expects 5% inflation, all wages and prices adjust by 5%. Real wages remain constant, employment stays at the natural rate, but nominal values are rising across the board. The government can choose any inflation rate it wants, but it cannot permanently reduce unemployment below the natural rate without ever-accelerating inflation. As Friedman and Phelps predicted, this is precisely what happened in the 1970s: when average inflation jumped from 2.5% in the 1960s to 7% in the 1970s, unemployment didn’t fall-it actually rose from about 4% to over 6%.

Policy implications of the vertical curve 🔗

The vertical long-run Phillips Curve carries a sobering message for policymakers: monetary and fiscal stimulus cannot permanently reduce unemployment. Attempts to do so only generate inflation without lasting employment gains. This realization transformed central banking. Instead of trying to exploit an inflation-unemployment trade-off, modern central banks focus on anchoring inflation expectations at a low, stable level while letting unemployment settle at its natural rate, which is determined by real factors like labor market flexibility, education levels, and technology.

Rational expectations and policy ineffectiveness 🔗

Adaptive expectations assumes people learn slowly, looking backward at past inflation. But what if people are smarter than that? The rational expectations hypothesis, championed by Robert Lucas and Thomas Sargent, suggests that forward-looking agents use all available information-not just past inflation-to form expectations. This includes understanding how economic policy works and anticipating government actions.

Under rational expectations, policy becomes even less effective. Suppose the government announces an expansionary monetary policy. Under adaptive expectations, workers would be fooled initially because they haven’t yet observed higher inflation. But with rational expectations, workers immediately incorporate the expected impact of the policy into their wage negotiations. They know expansionary policy causes inflation, so they demand higher wages right away. The result? Prices and wages both jump immediately, with no temporary reduction in unemployment-not even in the short run.

This “policy ineffectiveness proposition” sparked intense debate. Early new classical theories assumed that with rational expectations and flexible prices, the Phillips Curve should be very steep and deviations from the natural rate very short-lived. Only unexpected policy changes could move unemployment, and even then, only briefly.

The nuanced reality 🔗

In practice, the strong version of rational expectations proved too extreme. Even its proponents eventually conceded that wages and prices are somewhat “sticky”-they don’t adjust instantly to new information. Labor contracts last months or years. Prices in catalogs and menus change infrequently. This stickiness means that even anticipated policies can have short-run real effects, though these effects are smaller and shorter-lived than under adaptive expectations.

Modern central banks, including the Federal Reserve and European Central Bank, operate under the assumption that expectations matter tremendously. They communicate extensively about their inflation targets and policy intentions precisely because they want to anchor expectations. When people trust that the central bank will keep inflation around 2%, they build that assumption into wage negotiations and pricing decisions, making the central bank’s job easier. Transparency and credibility have become essential policy tools.

Implications for modern macroeconomic policy 🔗

The expectations revolution in macroeconomics fundamentally changed how we think about policy. The message is clear: governments cannot fool people systematically. Any attempt to permanently exploit a trade-off between inflation and unemployment will fail, leaving only higher inflation and no lasting reduction in unemployment. The original Phillips Curve, which suggested a stable menu of policy choices, was dangerously misleading.

Today’s consensus view incorporates both adaptive and rational elements. People form expectations based partly on past experience and partly on forward-looking information. Central banks can influence unemployment temporarily through unexpected policy changes or when expectations are not fully anchored. But sustainable economic performance requires maintaining price stability and allowing the labor market to find its natural equilibrium. The painful lessons of the 1970s-when unemployment and inflation both reached double digits in many countries-serve as a reminder of what happens when policymakers ignore the role of expectations.

Understanding whether expectations are adaptive or rational has profound consequences. It affects how long monetary stimulus influences the real economy, how quickly inflation responds to policy changes, and whether announced policy frameworks have credibility. These aren’t merely theoretical curiosities-they determine whether central banks can effectively stabilize the economy or whether their efforts will be neutralized by savvy economic agents who see through the policy game.

What do you think? In today’s era of explicit inflation targeting and extensive central bank communication, do you believe people form expectations adaptively or rationally? How might the rise of real-time economic data and social media affect how quickly expectations adjust to policy changes?

References

1. https://en.wikipedia.org/wiki/Stagflation
2. https://www.econlib.org/library/Enc/PhillipsCurve.html
3. https://en.wikipedia.org/wiki/Phillips_curve
4. https://policonomics.com/expectations-augmented-phillips-curve/