Fewer openings, harder to get hired: U.S. labor market likely softer than it appears

Monetary policymakers currently face two important questions: (1) Has the labor market made a “soft landing” around its pre-pandemic level, or is it likely to contract further? (2) Do we need to hunt for structural changes (examples might include immigration, automation, or working from home) to understand a labor market that is fundamentally different from the past? Or are we simply moving in a natural, expected way through the business cycle?

The raw data on U.S. job vacancies present a puzzle on both counts. In the official data, vacancies soared during the pandemic and—though much lower—still remain at levels that suggest the labor market is moderately tight. Yet the unemployment rate is unequivocally on the rise, from a historic low of 3.4 percent in 2023 to 4.3 percent in July 2024. At the same time, we have seen no sign of mass layoffs.

Where are we in the business cycle, and is there something deeper going on?

In a December 2023 article, we used an adjusted vacancy rate to clarify our view of the U.S. labor market. When we update this work to include the latest data, today’s mixed signals recede to reveal the labor market may be accumulating slack, in a “plain vanilla” fashion that is consistent with standard theory and not dependent on structural shifts. In the context of the Fed’s tighter monetary policy since 2022, open jobs have become less abundant, leading to a natural decline in job-finding. This development is the primary factor causing the unemployment rate to increase.

From this vantage, the labor market has softened considerably. With 12-month personal consumption expenditures (PCE) inflation at 2.5 percent, still somewhat above the Fed’s official target, policymakers are confronting a trade-off between the Fed’s “dual mandate” goals of maximum employment and low and stable inflation. As Minneapolis Fed President Neel Kashkari put it in a recent interview, “the balance of risks has shifted.”

U.S. labor market not as tight as it may appear

In our prior article, we argued that the long-run upward trend in the published data on job openings—from the U.S. Bureau of Labor Statistics’ Job Openings and Labor Turnover Survey (JOLTS)—poses a challenge for assessing labor market tightness. The relationship between the vacancy rate (job openings per U.S. worker) and other labor market series—the quit rate, job-finding rate, and hiring rate—has become unstable since the Great Recession (possibly because of measurement issues). To gain a clearer view, we created an adjusted vacancy rate to correct for this deviation, retroactive to 2010, roughly when it began.¹

Figure 1 shows how the unadjusted and adjusted series paint different pictures of labor market tightness (the ratio of job openings to active job seekers, which we define as those who are unemployed but not on temporary layoff) through July 2024. The unadjusted series shows a labor market that is about as tight as the average for 2019. However, the adjusted vacancy rate shows a labor market that is not only slacker than the official data but has now fallen well below its pre-pandemic level.

Put in terms of levels, the divergence is especially stark: The unadjusted series suggests that right now each job seeker faces around 1.4 job openings. The adjusted series suggests the opposite: Each job faces 1.5 job seekers—much worse odds for unemployed workers.

The dimmer view provided by the adjusted measure of labor market tightness conforms with other labor market indicators. Most importantly, the job-finding rate (Figure 2, the fraction of unemployed workers who become employed each month) has also fallen below 2019 levels. 

Decline in the job-finding rate mostly responsible for rising unemployment

The unemployment rate is determined by both flows into unemployment (layoffs and people joining the ranks of job seekers) and flows out of unemployment (job seekers findings jobs). Intuitively, it is reasonable to associate a rising unemployment rate with workers losing jobs. However, a key insight from academic work on the U.S. labor market is that the large changes in the job-finding rate that occur over time (the outflows) are generally more important to the level of unemployment than inflows. That is, layoffs are historically the smaller factor, and this insight appears to also apply in today’s labor market.

Figure 3a presents data on the unemployment rate and the job-finding rate from 1967 to 2024. The job-finding rate typically explains more than 80 percent of changes in unemployment. Figure 3b shows the historic pattern also holds since July 2022.

Focusing on the job-finding rate takes us back to basics, cutting through alternative explanations for the recent increase in unemployment. There have been notable inflows into the ranks of unemployed workers, including a rise in immigration since the pandemic and a wave of tech sector layoffs in 2023. But the obvious explanation for the persistent rise in the unemployment rate seems more likely to be the factor that affects everybody in the pool: a persistent decline in the demand for labor and, thus, the job-finding rate.

This development has coincided with the correction in market tightness revealed by our adjusted tightness measure and other signs of a weakening labor market, including a shift in the composition of separations toward layoffs rather than quits.² The Federal Reserve Bank of New York tracks workers’ expectations of their ability to find a job within three months, if needed. This measure has been below pre-pandemic levels since mid-2023.

Updating the Beveridge curve for a new view of the soft landing

Putting the focus on job vacancies leads naturally to the “Beveridge curve,” which plots the relationship each month between the unemployment rate and the vacancy rate.

The Beveridge curve with official, unadjusted vacancy data shows a dramatic journey during and after the pandemic (Figure 4). Although the ride looks largely over, as of July 2024 the curve still shows the labor market at an extreme end of the pre-pandemic experience. In the language of the “soft landing,” this implies we might still have room to coast to a gentle touchdown.

However, we argued back in December 2023 that the U.S. economy was already near this spot, at the peak of the pre-pandemic Beveridge curve, when using the adjusted vacancy series. Updating that Beveridge curve with the latest adjusted labor market data shows the economy continuing to move around the familiar territory of the stable pre-pandemic curve (Figure 5).

This adjusted Beveridge curve suggests the U.S. labor market has well and fully landed. Job openings have grown relatively scarce. We should not be surprised that workers are reporting that jobs are harder to get and the unemployment rate is rising.

Going forward, if the labor market continues to soften, we would expect to observe small declines in measured, unadjusted vacancies—along with increases in unemployment. Suppose over the next year the official, unadjusted vacancy rate decreased by only 0.30 percentage points. Once we adjust for the upward trend in vacancies, this translates into a (larger) 0.50 percentage point decline in adjusted vacancies (the same as experienced over the past 12 months). Our estimates suggest this would lead to a significant 1.4 percentage point increase in unemployment. Adjusting for the long-term trend in vacancies shows that JOLTS vacancies need to be increasing for us to infer a stable labor market.

As is always the case, many idiosyncrasies differentiate the current position of the U.S. economy relative to past similar points in the business cycle. Yet by correcting for seeming aberrations in job vacancy data, we see a standard view of the current U.S. labor market. In this case, at least, the principle of “Occam’s razor” rings true: The simpler explanation appears likely to be the best one.

Endnotes

1 While other series fluctuated around similar levels, the vacancy rate has fluctuated while moving steadily upward. If job openings really were persistently increasing, we would have seen hiring rates increase as well, but we have not. Briefly, the adjustment is conducted as follows: (1) Between 2001 and 2009, we take data on the quit rate (quits relative to employment) and the vacancy rate (openings relative to employment) and fit a constant linear relationship in logs; (2) Between 2010 and 2019, we use this relationship and the log quit rate to construct a counterfactual log vacancy rate; (3) Between 2010 and 2019, we fit a linear trend to departures of the log vacancy rate from the counterfactual log vacancy rate; and (4) Between 2010 and 2024 we remove this linear trend from the log vacancy rate data. Whether the quit rate, job-finding rate, or other stationary labor market series is used as the basis for the counterfactual vacancy rate doesn’t matter for the results.

2 Research by Kathrin Ellieroth and Amanda Michaud suggests that layoffs may be higher than measured in official statistics once transitions out of the labor force are taken into account. This presents additional evidence that the labor market may be weakening.