“Hot” or not? Learning the right lessons from the labor market after COVID

How should policymakers interpret signals from the labor market during periods of high inflation? This is a crucial question about which much can be learned from the inflation surge of 2021–2023. As the effects of that period continue to play out, it increasingly looks like key indicators from that time were not telling the story that many thought they were.

A common narrative is that the labor market in the post-pandemic period was “hot” and was a proximate cause of inflation. The supposedly feverish temperature of the labor market was measured by low unemployment, high vacancies, and high quit rates of employed workers (Figure 1). As reported widely, the “V/U” ratio—the ratio of vacant jobs to unemployed workers—hit a historic high in early 2022 (Panel C).¹

This narrative’s lesson for the future is that addressing high inflation requires policymakers to first cool the labor market. Two recent papers challenge this view. They argue that these measures used to diagnose the labor market do not point to causes of inflation. Rather, these indicators are symptoms of broader economic and social phenomena—including inflation itself.

A hot labor market seems incompatible with falling wages

A key aspect of both papers is an additional data point that has become more salient as time passes: real wages. After a brief increase in 2020, real wages declined during the primary inflationary period from April 2021 to September 2023 (Figure 2). The average real wage fell more than 5 percent below trend.

Falling real wages were experienced by workers across the income distribution. Real wages of low-income workers did not catch up to their pre-pandemic trend until the end of 2024, while real wages for high-income workers remain nearly 4 percent below trend. If a hot labor market was driving inflation, why did wages lag prices and, from the perspective of 2025, never really catch up?

New macroeconomic models from economists including Erik Hurst and Gianluca Violante offer strong explanations for falling real wages alongside heightened measures of labor market tightness and worker mobility. Contrary to the prevailing narrative around “hot labor market” data, both economists’ narratives suggest a cool labor market in terms of its pressure on prices. After reviewing their findings, I ask Hurst and Violante for alternative, real-time measures that might be used to diagnose whether a hot labor market is putting pressure on prices.

High inflation drives employed workers to search for new jobs

In “A Theory of How Workers Keep Up with Inflation,” Hurst of the University of Chicago Booth School of Business (with Hassan Afrouzi, Andres Blanco, Andres Drenik) finds that after a burst of inflation, the labor market can generate data that looks hot but is cool in the sense of declining real wages.

The paper’s key mechanism is as follows. When inflation takes off, firms can sell their goods for higher prices but may be slow to increase their workers’ nominal wages. This is great for firms, but bad for workers who face stagnant wages and higher prices. Unable to get a raise at their current job, employed workers increase their effort to find higher-paying new jobs. This surge in worker mobility incentivizes other firms to post vacancies, which are mostly filled by employed workers.

Quits increase, vacancies (V) increase, and the unemployment rate (U) remains flat, causing a sharp increase in the V/U ratio. The labor market looks hot by this measure, but real wages are falling.

The model used to study this mechanism is consistent with empirical facts documented over the last five years:

1. On-the-job wage raises are indeed infrequent and small. A pre-pandemic study found that one-third of job-stayers receive no adjustment in their wages in a typical year. Of those who get raises, more than 60 percent receive raises of less than 3 percent.
2. To keep up with inflation, many workers did have to move to a new job. Hurst and his co-authors find wage increases of job-changers were roughly double those of job-stayers between 2021 and 2023.
3. Layoff rates fell and profits jumped as firms paid workers pre-pandemic wages but sold goods at higher prices.
4. Hiring rates out of nonemployment were relatively flat. Employed workers scrambled to keep up with inflation, while nonemployed people’s benefits were indexed (for example, Social Security recipients received an 11.9 percent cumulative cost-of-living adjustment over 2021 and 2022).

The authors conclude by showing that many past periods of high V/U were also associated with high inflation but flat unemployment rates. These periods coincided with negative supply shocks in the 1950s, mid-1970s, and late 1970s. Like 2021 to 2023, these were weak labor markets from the perspective of their pressure on prices. “Policymakers should be cautious about viewing the rise in the V/U rate as a sign of a hot labor market during inflationary periods without holistically looking at other labor market indicators,” the authors note.

Symptoms of the rise of remote work

In “Job Amenity Shocks and Labor Reallocation,” Gianluca Violante of Princeton University and his co-authors (Sadhika Bagga, Lukas Mann, and Aysegul Sahin) model another mechanism by which a labor market can generate data that looks hot but is actually a cooling influence on prices.

Their model incorporates the scenario of a surge in working from home, as we experienced after the pandemic. Before COVID, many jobs in the U.S. economy were not viewed as compatible with remote work. Some workers might have preferred working from home, but the option was not even on the table.

The pandemic changed this. With small changes to technology, many jobs were revealed to be compatible with remote work, generating a mismatch in the economy between work-from-home jobs and those who wanted them. Here a important similarity with Hurst’s story emerges, as employed workers stepped up their search for jobs that aligned with their preference for remote work. This drove a sharp increase in vacancies from the jobs that these workers left behind, and fresh job postings as employers developed new work-from-home positions. Meanwhile, real wages stagnated as workers traded off wages against the amenity value of working from home.

Like Hurst’s paper, Violante’s exercise aligns with several salient facts about the U.S. labor market:

1. The willingness to work from home is high. In surveys of workers, around 50 percent reported being prepared to take a pay cut in exchange for working from home.
2. Many jobs were revealed to be compatible with working from home. In surveys of human resources managers, nearly 40 percent of jobs have task content compatible with working from home.
3. The model predicts that sectors that have lower compatibility with working from home should have seen large increases in vacancies as workers quit, while struggling to fill these vacancies. This was exactly what we saw in the data. Hospitality, construction, and durable goods manufacturing sectors experienced among the largest increases in vacancies and the largest decline in the rates at which they were filled.

While not explicitly modeling inflation, the authors note that this type of job reallocation tempers wage growth. Labor reallocation is high, vacancies are high, and quits are high. This can lead to the perception of a tight labor market. Yet wage growth is tempered by workers accepting lower wages and higher amenity values in work-from-home positions. This provides a clean narrative for the persistent decline in real wages beyond the pandemic, especially among high-income earners.

A key question for policymakers

When faced with an increasing V/U ratio, what additional information could policymakers use to assess whether the labor market is putting upward pressure on inflation? I put this question to both authors. According to Hurst, policymakers can look at wage changes of job-changers and flows from unemployment to employment. “Models of hot labor markets almost always predict that wages of job-changers should be increasing in real terms,” Hurst said. “These are the new-hire wages that are directly putting pressure on firms’ marginal costs and hence prices.” In the recent inflationary episode, wage changes of job-changers did not keep up with inflation (Figure 3).

Violante responded that it is “crucial to understand whether the surge in labor market tightness is accompanied by a reallocation of employment toward jobs that are either more productive or, as in our paper, more attractive to workers than their original ones. In both scenarios, such reallocation helps mitigate inflationary pressures.” To gauge this, Violante points to a measure proposed by Shigeru Fujita, Giuseppe Moscarini, and Fabien Postel-Vinay that captures mobility of employed workers across jobs relative to flows into the labor force from unemployment. In the recent episode, unemployment-to-employment flows fell relative to employment-to-employment flows (Figure 4). Violante emphasizes that such measures can be complemented with “real-time microdata on the labor market, in particular the origin and destination of workers swapping jobs.”

Not as hot as we thought

While these papers provide different perspectives on the U.S. labor market in the early 2020s, they come to common conclusions. The spike in vacancies and quits relative to unemployment arose due to an increase in labor mobility. Workers went looking for new jobs that either helped them keep up with inflation or aligned with their preference for remote work. Both imply a labor market that looked hot by standard metrics—in particular rising vacancies relative to unemployment. Yet rather than stoking inflation, both theories are consistent with a glaring data point in the post-pandemic labor market: Wages failed to keep up with prices.

Endnote

1 In a 2024 article, we argued that the officially reported vacancy rate was, for other reasons, already overstating the tightness in the labor market (“Fewer openings, harder to get hired: U.S. labor market likely softer than it appears,” Sept. 2024).