The Unemployment-Inflation Trade-off Revisited: The Phillips Curve in COVID Times
We estimate the natural rate of unemployment, often referred to as u*, in the United States using data on labor market flows, short-term and long-term inflation expectations and a forward-looking New-Keynesian Phillips curve for the 1960-2021 period. The natural rate of unemployment was at around 4.5% before the onset of the pandemic and increased to 5.9% by the end of 2021. This pronounced rise was primarily informed by strong wage growth rather than changes in inflation expectations. Despite the rise in the natural rate of unemployment, the secular trend of unemployment continued to fall and stands at around 4.2% reflecting ongoing secular developments which have been pushing down the unemployment rate over the last 30 years. Our model forecasts strong wage growth to moderate only sluggishly continuing to put upward pressure on inflation in the medium-run. We project underlying inflation to remain 0.5 percentage points above its long-run trend by the end of 2023 even if long-run inflation expectations remain well anchored.
Given the importance of wage growth for the inflation outlook, we examine detailed micro data on job-filling rates, posted wages for vacant positions, and workers' reservation wages. In particular, we construct a composition-bias free measure of wage growth at the employer-job level using Burning Glass Technologies data and document strong wage growth for both teleworkable and non-teleworkable jobs. Moreover, we find that workers' reservation wages increased substantially after the pandemic. Our empirical analysis suggests that the strong wage growth is likely not a one-time adjustment of additional compensation for jobs that pose health risks to workers but rather reflects a tight labor market accompanied with a changing work-leisure trade-off.
This paper was initially presented at the January 3, 2021 ASSA session on “Measuring the Unemployment Gap.” We thank Emmanuel Saez and Pascal Michaillat for organizing the session and the discussant, Regis Barnichon. Jin Yan and Charles Smith provided excellent research assistance. The views expressed in this paper are those of the authors and do not necessarily represent those of the Federal Reserve Bank of New York, Federal Reserve Bank of Dallas, the Federal Reserve System, or the National Bureau of Economic Research.