The Natural Rate of Unemployment: Idea, Rhetoric, and Empirical Trouble

The Seductive Curve and the Rude Question

For a long stretch of mid-twentieth-century policymaking, many economists and officials acted as if society faced a menu: accept a little more inflation, get a little less unemployment; accept a little more unemployment, buy a little price stability. Graphically, the menu was a Phillips curve: an empirical scatter or fitted line relating wage inflation or price inflation to some measure of labor-market tightness, often the unemployment rate.

Jargon note: The Phillips curve is named after A.W. Phillips’s work on UK wage behavior; in policy circles it became shorthand for a stable tradeoff between inflation and real activity.

In the late 1960s, Milton Friedman and Edmund Phelps—working separately—asked a question that sounds obvious only after you hear it: If policymakers try to exploit that tradeoff forever, why would workers and firms keep supplying the same “surprise inflation” that made the tradeoff work in the first place? Their answer, sharpened by later macroeconomics, became known as the natural rate hypothesis (sometimes linked to the concept of the Non-Accelerating Inflation Rate of Unemployment, or NAIRU in applied work).

This essay explains the hypothesis in plain language, separates the serious economic idea from the political slogans, and notes why modern empirical macro keeps revising “the” natural rate without ever quite owning it. We connect to Keynesian traditions via our multiplier realism piece, to monetarist monetary policy through k-percent rules, and to post-1970s inflation debates in stagflation and consensus.

What “Natural” Means—and What It Does Not

“Natural rate of unemployment” is a terrible label if you want public comprehension. People hear natural and think good, inevitable, or fair. In Friedman’s usage, “natural” was closer to equilibrium consistent with stable inflation expectations—not a moral optimum, not “full employment” in a humanitarian sense, and not a constant engraved in stone for all time.

Jargon note: Unemployment in these debates is usually the survey unemployment rate: people without jobs who are actively searching, divided by the labor force. It does not count “discouraged” workers unless survey definitions include them; it mismeasures underemployment; it says little about job quality.

The hypothesis’s core is about expectations. Imagine a central bank stimulates demand. Firms raise prices; workers initially misinterpret higher nominal wages as higher real wages; hiring increases; unemployment falls. This is the “movement along” a short-run Phillips curve if you like that language. But as workers learn prices are rising, they revise wage demands. The short-run tradeoff shifts. Try to hold unemployment permanently below the level consistent with stable inflation, and—you may—end up needing ever-accelerating inflation to keep fooling people, which is not a sustainable policy in a society with contracts, money balances, and political limits.

Hence the natural rate: the unemployment rate toward which the economy gravitates once expectations adjust, absent further surprises. Policy might push unemployment below that rate only by accelerating inflation—until expectations catch up again.

Friedman’s 1968 Address in Plain English

Friedman’s presidential address to the American Economic Association (1968) is the canonical statement many textbooks cite. Its rhetorical power came from timing: it predicted instability in naive Phillips-curve policymaking just before the 1970s made that instability visible.

Friedman distinguished nominal quantities (dollar wages, dollar prices) from real quantities (what wages buy, hours worked, consumption). He argued that labor supply and demand respond to real incentives, while policy often moves nominal demand first. The “surprise” channel is central: monetary policy can affect real outcomes when it is unanticipated relative to what wage-setters expected; systematic attempts to exploit the channel run into rational or adaptive learning.

Readers interested in how “expectations” became a technical battleground should see our Lucas critique primer, which asks when econometric relationships survive policy regime changes.

Phelps’s Microfoundations-Adjacent Story

Phelps’s work in the same era built search-and-information stories about labor markets: unemployment exists because matching workers and jobs takes time and because information is incomplete. Those micro stories help explain why “zero unemployment” is not a sensible benchmark without planning away frictions—but they do not by themselves prove a unique NAIRU stable across decades.

Still, the convergence of Friedman’s macro-language with Phelps-style labor-market frictions helped mainstream macro move away from treating unemployment as purely “demand deficiency” to be cured forever by stimulus. That move had ideological implications—sometimes used to argue against job guarantees or active labor programs—but the intellectual point can be separated: inflation expectations matter for whether demand stimulus buys durable real gains.

From Hypothesis to NAIRU: When Science Becomes a Number

Academic hypothesis became policy machinery when forecasters tried to estimate “the” NAIRU for central banks and finance ministries. Statistical offices produced time-varying NAIRU estimates that looked suspiciously like moving averages of actual unemployment: when unemployment rose for years, some models “discovered” a higher NAIRU, inviting the critique that policymakers were encoding pessimism as structural fact.

Jargon note: Hysteresis is the idea that long spells of high unemployment can raise equilibrium unemployment by eroding skills, networks, and attachment—so the “natural” rate might depend on history, not just demographics and search frictions.

If hysteresis matters, the clean separation between “cyclical” and “structural” unemployment blurs. European unemployment episodes in the 1980s and 1990s fed this debate: was high unemployment “natural” because labor markets were rigid, or did it become “natural” because macro policy accepted it long enough to scar the labor force? The hypothesis itself does not answer; it only frames why accelerating inflation might be the price of overheating.

Stagflation as a Political Fact, Not Only a Graph

The 1970s oil shocks complicated every simple Phillips story. If supply constraints raise prices and reduce activity, you can observe high inflation and high unemployment simultaneously—stagflation—without needing to deny expectations logic. But stagflation also damaged the credibility of policymakers who promised fine-tuned bliss. Monetarist narratives gained audience partly because they offered a nominal anchor story when supply shocks made demand management look intellectually exhausted.

Fairness requires acknowledging heterodox and post-Keynesian critiques: focusing on NAIRU can distract from markup pricing, conflict inflation between firms and workers, indexation, and global commodity regimes. The natural rate hypothesis is not a theory of all inflation; it is a warning about expectations when demand stimulus is the main tool.

Empirical Trouble: Flat Phillips Curves, Anchored Expectations, and Puzzles

Later decades brought empirical puzzles. In many countries, inflation looked less responsive to unemployment (“flattened Phillips curve”) while expectations seemed anchored by inflation targeting. Some episodes showed unemployment falling without explosive inflation, challenging crude NAIRU calendars. After the global financial crisis, inflation in many rich countries stayed muted even as unemployment fell, pushing macroeconomists toward global slack, inflation expectations, markups, demography, and digital prices as complementary explanations.

This is not a funeral for the natural rate concept; it is a reminder that one number rarely summarizes labor-market equilibrium across structural change. Readers of Adam Smith on division of labor and Schumpeter on creative destruction already suspect that “the” structure of employment shifts with technology and institutions—why wouldn’t the inflation-unemployment mapping shift too?

Matching Frictions, the Beveridge Curve, and Why “Labor Shortages” Aren’t NAIRU

Labor economists sometimes translate NAIRU intuitions into matching language: vacancies and unemployed workers coexist because search takes time. The Beveridge curve plots vacancies against unemployment; outward shifts can signal worse matching efficiency (skills mismatch, geographic mismatch, sectoral change). During COVID-19, many countries saw unusual vacancy–unemployment combinations, and commentators rushed to pronounce a “new NAIRU.” Some moves were temporary—health risk, childcare disruptions, early retirements—rather than eternal structural ceilings.

Jargon note: The Beveridge curve is a relationship between job openings and unemployment; it helps diagnose whether hiring is harder because the economy is hot or because matching technology worsened.

This matters for readers because shortages talk in the business press often smuggles in a moral claim (“people don’t want to work”) when the data problem is composition, wages, and sectoral reallocation. The natural rate hypothesis is about inflation expectations in macro equilibrium; it is not a license to ignore bargaining power, safety, or migration constraints.

Ethics and Rhetoric: When “Structural” Blames Workers

Because NAIRU estimates influence interest-rate decisions and austerity debates, they carry distributive stakes. If a model says NAIRU is high, policy may tolerate slack; if NAIRU is low, policy may chase overheating. When those estimates move opaquely, democracy has a communication problem.

A reader’s shield is conceptual: separate (a) the claim that expectations limit a naive inflation–unemployment tradeoff from (b) the claim that today’s unemployment is mostly “voluntary” or that labor markets are perfectly competitive. The first can be true without the second.

Cross-Country Snapshots: The Same Idea, Different Institutions

The natural-rate idea traveled, but it did not land identically. In Germany, central-bank independence and a cultural memory of hyperinflation made inflation aversion a public norm; in that setting, “don’t overheat” sounded like prudence, not cruelty. In Spain or Italy in the 1980s and 1990s, debates about convergence to European monetary union mixed NAIRU talk with labor-market reforms—sometimes fairly, sometimes as a blanket justification for deregulation that shifted risk onto workers.

In Latin American stabilization episodes, the relevant “natural” concept often intertwined with exchange-rate regimes and dollarization debates rather than with a single domestic Phillips curve. When inflation is triple digits, the Friedman–Phelps point is almost trivial: expectations are not anchored; “surprise inflation” is the weather, not the exception. Stabilization then requires coordination—fiscal, monetary, sometimes political—in ways a one-line NAIRU estimate cannot replace, as readers of our Washington Consensus reconstruction will recognize.

These comparisons matter because NAIRU rhetoric can sound universal while quietly importing US labor-market institutions—at-will employment, weak unions, employer-based health benefits—into places where wage-setting works differently. A hypothesis about expectations is not automatically a handbook for copying American institutions.

How This Connects to Modern Central Banking

Modern inflation targeting is, in part, a institutional answer to the natural rate hypothesis: publish an inflation goal, communicate a reaction function, build credibility so expectations do not need to be “surprised” into alignment. The Fed and peers still debate maximum employment using range language rather than a single eternal NAIRU—an admission that precision is fiction. Our essay from monetarism to inflation targeting traces how that framework absorbed (and partly replaced) older monetarist money-growth targeting.

For a broader institutions lens—rules, power, growth—see Acemoglu and Robinson–style arguments; for financial instability, see Minsky. None replaces the natural-rate lesson; they surround it.

If you take one practical lesson into policy reading groups, make it this: whenever a speaker cites NAIRU to end an argument, ask which model produced the number, what data updated it last year, and what would falsify it. Science-minded macroeconomists ask those questions routinely; political actors often do not. The hypothesis’s enduring value is not a sacred unemployment constant—it is a discipline about nominal surprises, learning, and the limits of using inflation as a perpetual engine for real output.

Undergraduate textbooks sometimes illustrate the natural rate with long-run vertical Phillips curves at “the” NAIRU. Treat that diagram as a pedagogical cartoon: it encodes the expectations logic cleanly, but it can mislead if you forget that immigration policy, childcare policy, training systems, and antitrust all move the labor-market frictions that sit underneath the cartoon. Good economics keeps the cartoon and the institutions in the same frame.