The Phillips curve shifts when something changes the rate of inflation that occurs at any given level of unemployment. The most powerful shift factor is a change in inflation expectations: when workers and businesses expect higher inflation, the entire curve moves upward. But expectations are only one of several forces. Supply shocks, productivity changes, globalization, and structural shifts in the labor market all push the curve to new positions.
Inflation Expectations Are the Primary Shift Factor
The Phillips curve plots the relationship between unemployment and inflation. At any point along it, a lower unemployment rate corresponds to higher inflation, and vice versa. But the curve sits at a particular height on the chart, and that height is set mainly by what workers and firms expect inflation to be.
Here’s the mechanism. Suppose the economy is at its natural rate of unemployment (around 4.4% in the Congressional Budget Office’s current projections for the U.S.) and everyone expects 3% inflation. Workers will demand at least a 3% raise just to keep their purchasing power flat. Firms, facing those higher labor costs, raise prices by 3% to protect their margins. The result: inflation runs at 3% and nobody is surprised. The economy can sit at this point indefinitely.
Now imagine expected inflation rises to 5%. Workers start demanding 5% raises as a baseline, firms pass those costs along, and the entire curve shifts upward. Every unemployment rate on the curve now corresponds to a higher inflation rate than before. The formula that captures this is straightforward: actual inflation equals expected inflation plus whatever extra pressure comes from the labor market being tighter or looser than normal (economists call that extra pressure the “bargaining gap”). When expected inflation rises, actual inflation rises at every level of unemployment, which is exactly what an upward shift looks like.
This is why central bank credibility matters so much. When a central bank has a strong track record of keeping inflation near its target, people treat inflation blips as temporary. Their expectations stay “anchored,” often approximated by a constant value like 2%. The Phillips curve stays put. But if monetary policy is lax and inflation drifts away from target for extended periods, people start basing their expectations on whatever inflation was last period. Each burst of inflation feeds into expectations, which feeds into the next period’s inflation, and the curve keeps shifting upward. That self-reinforcing cycle is precisely what made the 1970s so painful.
Supply Shocks Push the Curve Outward
A supply shock raises the cost of producing goods for reasons that have nothing to do with how tight the labor market is. The 1970s oil crises are the textbook example. U.S. consumer price inflation hit 6.4% in 1970 and kept climbing through the decade, even as unemployment rose. Policymakers at the time treated inflation as a pure “cost-push” process: prices were being driven up by energy costs, and weaker demand couldn’t offset it. The Phillips curve appeared to have shifted outward, delivering both higher inflation and higher unemployment simultaneously, a combination the original stable curve said shouldn’t happen.
The post-pandemic period provided a modern replay. Research from the IMF found that during the 2021-2023 recovery, the Phillips curve shifted outward by roughly 130 basis points for core inflation, meaning that the same level of economic activity produced notably higher inflation than it would have before COVID. The curve also became much steeper, with its slope increasing by 170 to 270 percent compared to the pre-pandemic period. The main culprit was supply chain disruptions that hit the goods sector especially hard. When demand surged during the recovery but manufacturing capacity and shipping logistics couldn’t keep up, firms in the goods sector raised prices far more aggressively than historical patterns would have predicted. The goods sector drove most of the shift, while services behaved more normally.
This matters because it shows that supply-side constraints don’t just cause a one-time price increase. When those constraints are widespread enough, they change the entire inflation-unemployment tradeoff for as long as they persist.
Productivity Growth Shifts the Curve Favorably
When labor productivity accelerates, each worker produces more output per hour. In theory, this lets firms absorb higher wages without raising prices, which should improve the inflation-unemployment tradeoff. In practice, the shift happens because of a timing mismatch: workers’ wage expectations adjust slowly to changes in productivity growth.
Research from Johns Hopkins economist Laurence Ball explains the mechanism. When productivity growth speeds up, firms can produce more cheaply, which directly reduces price inflation for any given rate of wage growth. But workers, whose expectations are shaped by recent experience, don’t immediately demand larger raises to match their higher productivity. The gap between actual productivity growth and workers’ wage aspirations acts as a disinflationary force, shifting the Phillips curve downward. This helps explain why the late 1990s saw falling unemployment alongside low inflation: productivity was booming, but wage demands hadn’t caught up yet.
The reverse is equally important. When productivity growth slows down, as it did in the early 1970s, workers continue expecting real wage gains based on the old, faster trend. Their aspirations exceed what productivity can support, firms face rising unit costs, and the Phillips curve shifts upward. Ball’s research provides a unified explanation for both episodes: the productivity slowdown of the 1970s worsened the inflation-unemployment tradeoff, and the productivity acceleration of the late 1990s improved it.
Globalization Flattens the Curve
Globalization doesn’t just shift the Phillips curve up or down. It changes its slope, which affects how much inflation responds to changes in unemployment. A study of 35 countries found that greater integration into the global economy systematically weakens the link between domestic economic activity and inflation. The output gap (how far the economy is from its potential) still pushes inflation up, but the effect is considerably smaller in more open, globalized economies.
The reason is competition. When domestic firms compete with producers around the world, they have less ability to raise prices during a boom because customers can switch to imports. Global supply chains also mean that domestic cost pressures get diluted across many countries. As more nations participate in global value chains, global output gaps start mattering more for domestic inflation, while domestic slack matters less. This flattening intensified in recent decades and hit particularly open economies the hardest.
A flatter Phillips curve has a practical consequence: central banks need to push unemployment much further below its natural rate to generate meaningful inflation, and recessions do less to bring inflation down. This helps explain why many advanced economies saw persistently low inflation in the 2010s despite falling unemployment.
Structural Labor Market Changes Shift the Natural Rate
The Phillips curve is anchored around a natural rate of unemployment, sometimes called the NAIRU (the rate at which inflation neither accelerates nor decelerates). When the natural rate itself changes, the curve shifts horizontally. If the natural rate falls from 6% to 4%, the entire curve slides to the left, meaning the economy can sustain lower unemployment without triggering accelerating inflation.
Several structural forces move the natural rate. Automation and technological change can make certain skills obsolete, raising structural unemployment as displaced workers struggle to find new roles. Demographic shifts matter too: a younger workforce with less experience tends to have higher frictional unemployment than a prime-age workforce. Changes in labor market institutions, such as the decline of union power or the rise of gig work, alter how wages are set and how quickly workers match with jobs. According to the Bureau of Labor Statistics, these structural factors produce “long-lasting, often permanent changes” to the economy’s equilibrium unemployment level. If a structural shift raises the natural rate, the curve shifts rightward, and what previously looked like a normal level of unemployment now sits below equilibrium, generating inflationary pressure.
Multiple Forces Often Act at Once
In the real world, the Phillips curve rarely shifts for a single reason. The 1970s combined rising oil prices (a supply shock) with a productivity slowdown and unanchored inflation expectations, all pushing the curve in the same unfavorable direction. The late 1990s combined accelerating productivity with deepening globalization and well-anchored expectations, all pulling it favorably. The post-pandemic period layered supply chain disruptions on top of a rapid demand recovery and a brief unmooring of inflation expectations.
Understanding which force is dominant at any given time is what makes inflation forecasting so difficult. The IMF’s post-pandemic research noted that policy institutions failed to anticipate the inflation surge partly because their models didn’t account for the sectoral imbalance between goods and services. When the goods sector hit a supply wall while services were still recovering, standard Phillips curve models that treat the economy as a single sector badly underestimated the inflationary impact. The lesson is that the Phillips curve is not a fixed relationship. It moves, and knowing what moves it is essential to understanding why inflation behaves the way it does.