Low Unemployment in 2023: Can It Last?
The United States has experienced unusually low unemployment in the past two years: The unemployment rate has been 4% or lower in every month since December 2021. It hit its lowest level (3.4%) since 1954 in January and April 2023. It then rose to 3.9% in October before falling to 3.7% in November. Low unemployment has been highly beneficial for millions of workers. But can it last? When unemployment has been low in the past, it has remained low for varying time periods before giving way to a recession that caused unemployment to rise.
Why Might Low Unemployment Be Followed by a Recession?
The recent combination of very low unemployment and high inflation is consistent with the economic concept of a nonaccelerating inflation rate of unemployment (NAIRU) toward which the unemployment rate will naturally gravitate over the course of an expansion. Unemployment below the NAIRU could be a sign that the economy is overheating, further fueling inflation and therefore complicating the Federal Reserve's efforts to restore low inflation. An overheating economy is, by definition, unsustainable, and the Fed is likely to respond to high inflation by raising interest rates, as it has done recently. A sharp rise in interest rates could itself cause a recession. Monetary tightening in such situations has historically resulted in recession more often than not. (For an in-depth discussion, see CRS In Focus IF12543, Has the Federal Reserve Achieved a Soft Landing in 2023?)
One difficulty with the NAIRU is it does not seem stable over time, although it is hard to tell because it is not directly observable. Thus, at any given time, low unemployment could be relatively close to or far from the NAIRU. (For an explanation of NAIRU and its relationship to inflation, see CRS Report R44663, Unemployment and Inflation: Implications for Policymaking.)
Has Low Unemployment Resulted in High Inflation and Recessions in the Past?
In the current expansion, unemployment first fell below 4% in December 2021 and has remained at or below 4% every month since. How unusual is this and how much longer can it last? To answer that, this In Focus looks at whether past periods of low unemployment were quickly followed by recession. The following analysis looks at expansions since the end of World War II featuring sustained periods of very low unemployment at or below 4%—comparable to the current period—and how long they lasted before the expansion ended and a recession began. Many economists estimate the NAIRU currently to be in the 4% range. But because it seems to change over time, the In Focus also looks at expansions featuring sustained periods of low unemployment at or below 5% to see whether similar patterns exist when the NAIRU was higher. Various cutoffs for low unemployment could potentially have been used, with similar results. Since World War II, there have been 12 recessions, and the unemployment rate rose significantly during all of them, peaking between 6.1% and 14.7%.
There are few very low unemployment episodes that lasted as long as the current one already has before unemployment started rising and the economy entered a recession. There were five previous very low unemployment episodes (see Figure 1), three of which featured a similar or shorter period of very low unemployment than the current expansion to date (expansions ending in 1948, 2001, and 2020). The other two cases, expansions ending in 1953 and 1969, maintained very low unemployment for longer. Very low unemployment was sustained for just a few months longer than the current expansion before ending in 1953. Thus, only the 1960s expansion featured a substantially longer period of very low unemployment—around four years—than the current one before a recession began in 1969. There were three low unemployment episodes, and all lasted less than three years before recessions beginning in 1957, 1973, and 2007 (not shown in Figure 1). (Note that the five expansions with unemployment below 4% maintained low unemployment for longer if a 5% threshold is used.) The unemployment rate never fell below 5% in the remaining four expansions before they ended.
Did the eight low and very low unemployment episodes result in high inflation? The four recessions starting in 1957, 1969, 1973, and 2007 were preceded by a combination of high inflation and low unemployment. The 1960s episode could be considered the exception that proves the rule that very low unemployment cannot be sustained for an extended period of time. It had the longest period of very low unemployment, but during that time, inflation steadily rose from 2% to 6%—consistent with an overheating economy with an unemployment rate below the NAIRU. But recessions in 1948 and 1953 followed periods of falling inflation, and 2001 and 2020 featured relatively low and steady inflation. Following consistently low inflation in the decades before the pandemic, economists had de-emphasized the inflation-unemployment relationship. The recent experience has raised questions of whether that relationship has reemerged.
Sometimes expansions end because of an economic shock, not overheating. For example, the recession in 2020 was caused by the onset of the pandemic, not the very low unemployment preceding it. Some of the recessions that were preceded by a period of low unemployment and high inflation may have had more proximate causes than the general overheating, such as supply shocks or distress in financial markets. Nonetheless, overheating may have made the economy more vulnerable to such occurrences.
Do Other Indicators Suggest Unemployment Is Unsustainably Low?
The unemployment rate is just one measure of the state of the labor market, and a look at other indicators can provide more insight into whether unemployment is unsustainably low. Anecdotal evidence and other labor market indicators point to significant labor market tightness that may be hard to sustain. For example, the ratio of unemployed workers to job openings has been less than one since May 2021, indicating there is more than one job opening for every person seeking work (see Figure 2).
|
Figure 2. Unemployed Persons Per Job Opening January 2001 to October 2023 |
 |
|
Source: Bureau of Labor Statistics, Job Openings and Labor Turnover Survey. |
In 2020, labor shortages initially developed because of the extreme sectoral disruptions caused by the pandemic. Industries that shed jobs at historic rates due to shutdowns then had trouble restoring employment levels during reopenings. These shortages were alleviated (but not eliminated in all cases) when economic and health conditions normalized. Meanwhile, unemployment became very low and stayed low despite lackluster economic growth throughout 2022 and the first half of 2023. The combination of low growth and low unemployment is unusual and suggests that low unemployment may not be solely a cyclical phenomenon. Some observers have hypothesized that unemployment has remained low despite the slowdown in economic growth because employers, previously hamstrung by pandemic labor shortages, are now hoarding workers they do not need. Although labor shortages seem to have alleviated somewhat in recent months, the low unemployment-to-job-opening ratio suggests that the unemployment rate is still unsustainably low.
On the other hand, the employment-population ratio is not particularly high right now—it was higher during the previous expansion and continually from 1986 to 2008. This suggests that if more people could be drawn back into the labor market, it could alleviate labor market tightness without higher unemployment. (Only a person who is seeking a job and does not have one is classified as unemployed; people who are not seeking jobs are classified as not in the labor force.) Given the low ratio of unemployed workers to job openings, an influx of workers to the labor force could potentially result in available job openings being filled (i.e., higher employment) rather than increased unemployment. Since unemployment first fell below 4% in December 2021, the employment-population ratio has increased by one percentage point, and the economy has added roughly 7.3 million jobs.
A similar metric to the employment-population ratio, the labor force participation rate (LFPR), measures the size of the labor force relative to the population. It is unclear whether an increase in the size of the workforce large enough to ease labor market tightness is feasible based on a demographic breakdown of the LFPR. There is more room for growth in some demographic groups than in others. Part of the decline in LFPR is due to the aging population, and older workers who leave the labor market are least likely to reenter. Prime-age female participation has already fully rebounded since the pandemic. But younger workers and prime-age men have seen a long-term decline in labor force participation that could potentially be reversed. It is unclear whether a strong labor market could induce higher participation among groups that are dragging down overall participation or whether policy changes would be needed.
Another indicator that sheds light on whether unemployment is unsustainably low is the NAIRU itself. Even though the NAIRU is not directly observable, it is possible that it has trended downward in recent years, making a lower unemployment rate more sustainable than previously. One well-understood driver of the NAIRU is the age profile of the population—young workers consistently have higher unemployment rates than older workers—so an economy with relatively more old workers would be expected to have a lower NAIRU than one with more young workers. Because of the aging of the population and a decline in the participation rate of young workers—trends that pre-date the pandemic—the share of the labor force under age 25 has recently been the lowest since the data series began in 1948. Nonetheless, most current estimates of the NAIRU are not below 4%.