Given the increase in the unemployment rate since spring 2023, attention has turned to whether today’s softer labor market is still tight, back down to a more sustainable pace, or even showing signs of slack (meaning whether the labor market is showing signs that it could be stronger without putting unwelcome upward pressure in inflation). In this piece, we review a series of labor market indicators that provide insight into these questions. We find that multiple indicators reach the same conclusion: The labor market remains tight even though it has slowed.
However, relative to the strength in the labor market immediately prior to the pandemic, some measures suggest pockets of slack:
- White men between the ages of 25 and 54 as well as older white men are practically the only demographic groups to be participating in the labor force at lower rates than in the pre-pandemic period.
- Ten states have seen an increase in their unemployment rate of at least three-quarters of a percentage point.
- There have been recent upticks in transitions from unemployment into self-employment and in involuntary part-time work, both of which historically have occurred when the labor market is weak.
Indicator #1: Trends in the unemployment rate
The most common baseline indicator of slack is the national unemployment rate: the share of people in the labor force without a job and actively seeking work. Nonetheless, the national average unemployment rate is not the whole picture. Figure 1 shows the difference between the unemployment rate and the noncyclical rate of unemployment, that is, the Congressional Budget Office’s (CBO’s) projection of the hypothetical unemployment rate, which does not change with fluctuations in aggregate demand.
As Council of Economic Advisers Chair Jared Bernstein noted in his 2018 Hamilton Project policy proposal, since 1980, there have been more periods of slack than not. However, since 2017, the labor market has more often been hot than not. The unemployment rate rose sharply above the noncyclical rate in the midst of the COVID-19 recession, but the gap closed within a year-and-a-half. By comparison, it took about eight years following the Great Recession and 3 1/2 years following the 2001 recession to get to a healthy labor market by this definition.
While the pre-pandemic period (2019) and post-pandemic period (2022 to now) are characterized by notably low unemployment, the Q3 2024 unemployment rate (4.2 percent) is higher than its pre-pandemic rate (3.6 percent in Q3 2019). CBO’s noncyclical rate suggests that the unemployment rate was unsustainably low in both periods, but by less in the recent data.
Nonetheless, the national average unemployment rate is not the whole picture. Figure 2 is a data interactive that shows the difference in the Q3 unemployment rate for each year from 2019–23 to 2024, both overall and by state. A green arrow represents a reduction in the unemployment rate over that period, and a red arrow represents an increase. Between 2019 and 2024, about half of states saw declines on net in their unemployment rates, and some states experienced continued declines in the last year. While the national unemployment rate increased by 0.5 percentage points between Q3 2023 (3.7) to Q3 2024 (4.2), nine states saw their average unemployment rates decrease over this period. Currently, all but four states have unemployment rates below 5 percent; in fact, 37 states have unemployment rates below the Q3 2024 national unemployment rate.
That said, since Q3 2019, 10 states have experienced an increase in their unemployment rate of at least three-quarters of a percentage point; moreover, for six of those states, the increase was roughly 1.5 percentage points. For example, Nevada’s unemployment rate increased from 4.2 percent in Q3 2019 to 5.5 percent in Q3 2024. Our interpretation of the subnational picture is that there are a few geographic pockets of labor market slack.
Indicator #2: Who becomes self-employed?
Sources of inflows to self-employment have historically varied with the business cycle. Movement from unemployment to self-employment may also capture slack in the labor market if unemployed workers pivot from searching for wage and salary employment (hereafter “payroll employment”) to working for themselves due to poor job prospects. Figure 3 plots the number of people, in thousands, who transitioned into self-employment by their labor market status in the prior month—not in the labor force (NILF), unemployment, or payroll employment. We define those who are self-employed as those who report being self-employed as their primary job—this includes both unincorporated and incorporated self-employment, the latter of which we distinguish from working for an employer.
Generally, inflows to self-employment from NILF and payroll employment increase over time as the population expands, and those inflows have particularly increased during expansionary periods. In contrast, inflows from unemployment do not generally trend upward over time, but they do increase during periods of economic contraction.
While aggregate inflows have generally increased over time, outflows have been frequently greater than inflows to self-employment since 2000, with the highest number of people exiting self-employment to NILF (not shown). The result is that the share of the labor force that is self-employed has generally declined over time.
Despite that long downward trend, the share of the labor force that was self-employed rose from 10.5 percent in late 2019 to a peak of 11.3 percent in summer 2021 and remained elevated until early 2023. That occurred because of a sharp increase in inflows to self-employment from those with payroll jobs and those out of the labor force. Those inflows more than offset a fall in inflows from unemployment, suggesting that strong labor demand was pulling the unemployed into wage and salary employment. However, inflows to self-employment from unemployment have increased by about 40,000 people relative to 2023. In September 2024, transitions into self-employment from unemployment reached 235,000 people—higher than any level in 2018 or 2019. A growing number of unemployed people opting into self-employment may indicate labor market slack.
Indicator #3: Newly employed workers
To the degree the pool of available workers now, more so than in the past, comprises both the unemployed and those out of the labor force, that suggests that we need to rethink how well the unemployment rate captures labor market slack. Figure 4 shows, for newly employed workers, the labor force status in the previous month (unemployed, not in the labor force and wanting but not actively searching for a job, and otherwise not in the labor force). By this measure, we see signs that there is more labor market slack than suggested by the unemployment rate.
Except in the immediate aftermath of the COVID-19 recession, over the time period shown, most newly employed people were previously not in the labor force. In fact, and perhaps surprisingly, the typical newly employed person was not in the labor force the previous month and did not indicate that they wanted a job. Relative inflows to employment among those who were out of the labor force the previous month but who indicated that they wanted a job have remained relatively flat over time: This share remained around 15 percent in 2024, consistent with historical trends since 2000.
In contrast, the share of the newly employed who were previously NILF and had not indicated that they wanted a job has averaged roughly 60 percent since late 2021, consistent with its average from 2016 to 2019 but higher than its average in previous years. From 2022 to 2024, 72 percent of the newly employed had previously been out of the labor force, a statistically significant increase from 64 percent in the early 2000s. That said, the share ticked down in August 2024. If this reduction continues, it may be a sign that those out of the labor force see fewer job opportunities, and that the increase in slack is actually greater than the rise in the unemployment rate suggests.
Indicator #4: The employment-to-population ratio
Figure 5 shows the employment-to-population ratio (EPOP; lime green) and the potential EPOP (dark green). To measure potential employment, we multiply the inverse of the noncyclical rate of unemployment (the potential employment rate) by the potential labor force. EPOP is a useful measure for assessing slack because it relates employment levels to the size of the population. However, given the downward trend in the potential EPOP since the turn of the century due to population aging, it can be difficult to make sense of EPOP levels over time without making this comparison. We find that EPOP continues to surpass potential EPOP, indicating that the labor market is tight.
In Q3 2024, EPOP stood at 60.1 percent, down only marginally from the year prior and on par with lauded pre-pandemic levels, even as CBO’s estimate of potential EPOP has declined. Furthermore, in June 2023, prime-age EPOP (among those ages 25 to 54) reached its highest peak since 2001, and it remains at that peak as of September 2024.
Indicator #5: Underemployment
Another dimension of slack is underemployment: the gap between a worker’s desired hours and the hours they actually work. Figures 6a (all workers) and 6b (prime-age workers) show the share of workers who are part-time, overall and by whether the part-time hours are voluntary or involuntary. The composition of part-time work suggests a softening labor market.
For more than three decades before the COVID-19 pandemic, trends in part-time work have been driven by changes in the involuntary part-time workforce. While voluntary part-time employment accounts for most part-time work, its trend is notably flat, even among the prime-age population. Involuntary part-time work, on the other hand, has tended to follow the business cycle. It is a new development that recent growth in overall part-time work reflects an increase in the share of workers working voluntarily part-time. That development reflects the combination of typical cyclical patterns and an unusual decline in voluntary part-time work during the height of the pandemic and its gradual recovery to pre-pandemic levels since.
The share of people who worked involuntarily part-time has ticked up in 2024 (from 2.5 percent to 2.8 percent); however, these 2023 and 2024 year-to-date average levels remain the lowest since 2001. This follows a spike in involuntary part-time work in April 2020, during which about 8 percent of all workers and prime-age workers were involuntary part-time, hovering near the voluntary part-time levels for both groups. Since then, the share of involuntary part-time workers has fallen over 66 percent from that peak for both prime-age workers and all workers. Still, the uptick in 2024 may suggest softness in the labor market and merits watching.
Indicator #6: Contributions to labor force participation
Finally, figure 7 shows how different demographic groups’ changing propensities to participate in the labor force between the pre-pandemic period (January–September 2019) and today (January–September 2024) contribute to the net change in labor force participation after holding demographic changes constant. Over this period, the labor force participation rate declined by 0.6 percentage points, but the contribution of an increasing propensity to participate (+0.6 percentage points) counteracted the drag of demography-related changes (–1.2 percentage points).
Consistent with our prior research from 2023, white men aged 25−54, 55−64, and 65+ persist as the groups whose propensities to participate are lagging (together a drag of -0.3 percentage points on the total labor force participation rate, or 73 percent of the total drag). White men are the only prime-age gender race/ethnicity group with a lower propensity to participate today than in the pre-pandemic period. In contrast, prime-age women of all races lead the groups who have increased their propensity to participate since 2019, with a particularly large increase in participation among prime-age white women.
One group watched closely during the pandemic was people aged 55–64; that group saw a sharp drop in participation early on. Although observers worried that early retirements among this group could mean a smaller labor force for some time, participation among that group has largely rebounded. White men aged 55−64 are the only group within that age band to be participating less now than in 2019: All other men and women aged 55−64 are now participating at higher rates, holding other demographics constant.
Conclusion
In this piece, we assess the state of the labor market by looking at key indicators—employment, unemployment, labor force participation, part-time work, and churn into employment and self-employment. Particularly when making comparisons to the 2019 labor market, the indicators point to some discrete pockets of slack: White men 25 and older today have a lower propensity to participate than they did in 2019, and a fifth of states have unemployment rates today that are at least three-quarters of a percentage point higher than in 2019. In addition, involuntary part-time work and inflows into self-employment from unemployment have very recently ticked up. However, in general, we find that the labor market today remains tight and some further slowing is still in train as it returns to a more sustainable pace.
That said, developments in the labor market over the next year or so may be determined more by policy changes than by business cycle trends. For example, significant reductions in immigration flows could lead to reduced aggregate demand and thus reduced labor demand. The implementation of large, across-the-board tariffs would also reduce labor demand, particularly in the manufacturing sector. Such policy changes put at risk the soft landing, which otherwise appears well within reach.
Acknowledgments
The authors thank Chloe East and Brad Hershbein for expert and helpful comments and Victoria Yan and Jimmy Zheng for material contributions to earlier versions of this work.
Data references
Congressional Budget Office (CBO). 2024. “Historical Data and Economic Projections.” Congressional Budget Office, Washington, DC.
U.S. Bureau of Labor Statistics (BLS). 1988–2024. “Current Population Survey.” Bureau of Labor Statistics, U.S. Department of Labor, Washington, DC.
U.S. Bureau of Labor Statistics (BLS). 1990–2024. “Current Population Survey.” Bureau of Labor Statistics, U.S. Department of Labor, Washington, DC.
U.S. Bureau of Labor Statistics (BLS). 1994–2024. “Current Population Survey.” Bureau of Labor Statistics, U.S. Department of Labor, Washington, DC.
U.S. Bureau of Labor Statistics (BLS). 2000–24. “Current Population Survey.” Bureau of Labor Statistics, U.S. Department of Labor, Washington, DC.
U.S. Bureau of Labor Statistics (BLS). 2019–24a. “Current Employment Statistics.” Bureau of Labor Statistics, U.S. Department of Labor, Washington, DC.
U.S. Bureau of Labor Statistics (BLS). 2019–24b. “Local Area Unemployment Statistics.” Bureau of Labor Statistics, U.S. Department of Labor, Washington, DC.
