Why Is U.S. Inflation Still High?
The continued tightness of the labor market remains the primary driver of persistently high inflation, according to our econometric model (Benigno and Eggertsson, 2023).
We initially tested an early version of this model at the end of 2022 and finalized it by February 2023, when inflation was still significantly elevated. At that time, our model effectively captured the inflationary surge, offering a compelling explanation for its escalation. More remarkably, it also provides strong support for the subsequent disinflationary period, demonstrating its robustness across different inflationary phases.
As illustrated in the graph above, the sharp rise in inflation—measured using core CPI at quarter-to-quarter annualized rates—began in the first quarter of 2021. This was largely driven by supply shocks (red bars) and labor market tightness (gold bars). Notably, the model’s residual (gray bars) is sizable in that quarter, possibly reflecting the limitations of traditional supply indicators in fully capturing the supply-chain disruptions that occurred at the time.
From then until the first quarter of 2022, inflation continued to rise, primarily fueled by labor market tightness, though supply shocks also played a role. This was the period when the Federal Reserve initiated its rate-hiking cycle, which continued until July 2023. Interestingly, as the Fed began tightening monetary policy, inflation expectations had already started to become unanchored, further contributing to inflationary pressures—as indicated by the blue bars. However, these expectations quickly stabilized following the Fed’s actions.
The disinflationary process eventually took hold, initially driven by falling energy prices and the easing of supply constraints. Over time, loosening labor market conditions played an increasingly important role. However, by the second quarter of 2024, the impact of declining energy prices had largely faded.
The third quarter of 2024 appeared to be the point at which inflation nearly converged to the Fed’s target. However, in the final quarter of the year, inflation ticked back up. What changed? The labor market had initially loosened, bringing the vacancy-to-unemployment ratio down to 1.08. But by the last quarter of the year, it had tightened again, rising to 1.12. This suggests that labor market conditions remain a key factor in explaining inflation’s persistence.
Background on the model
The econometric model of inflation we present in Benigno and Eggertsson (2023) is a simple yet structured framework grounded in our theoretical approach to understanding the post-pandemic inflationary surge in the U.S.
At the core of our approach is a Phillips curve framework, but with a key innovation: instead of using the unemployment rate or the output gap as a measure of economic slack, we use the vacancy-to-unemployed ratio. This choice aligns with recent studies that emphasize the importance of this variable in driving inflation dynamics (see references in our paper).
The novel aspect of our model is the introduction of non-linearity in the Phillips curve, specifically a piecewise linear function that becomes significantly steeper when the vacancy rate exceeds the unemployment rate. We identify this threshold—the point where vacancies equal unemployment—as the Beveridge threshold, representing a balanced labor market.
Like any Phillips curve model, proper identification is crucial. To address this, we control for multiple proxies of supply shocks and inflation expectations. Our benchmark specification captures supply shocks using a principal component of non-core price variations and import prices, while inflation expectations are proxied by 2-year Cleveland Fed inflation expectations. Further details on the econometric specificaiton and estimates are provided in the paper.
References
Benigno, Pierpaolo and Gauti Eggertsson (2023).“It is Baaack. The Surge in Inflation in the 2020s and the Return of the Non-Linear Phillips Curve,” NBER Working Paper No. 31197. (April 2023.)