The long shadow of the 2001 recession

Accompanied by neither a public health emergency nor a financial crisis, the 2001 recession is often overlooked. It lasted only eight months and was not particularly severe in terms of jobs lost, with payroll employment declining by less than 2 percent (compared with more than 6 percent during the Great Recession and nearly 15 percent during the pandemic recession).

But the 2001 recession was notable in one way: The labor market never fully recovered, according to one important measure. Among people ages 25 to 54 (the so-called prime-age population), the share who were employed—the employment-population ratio—was still below its pre-recession peak when the next recession began. In fact, the U.S. prime-age employment-population ratio is still below that level. In April 2000, 81.9 percent of prime-age people were employed. Just over 23 years later, that measure has reached its highest rate since the 2001 recession, but at 80.9 percent as of July 2023, it remains a full percentage point lower than that peak.

However, employment levels did not decline evenly across the entire population. Figure 1 plots changes in the employment-population ratio for different age groups, including people younger and older than prime age, with 2000 as the reference year. From 1994 to 2000, the employment-population ratios for all age groups were trending up consistently. After 2000, there is a notable divergence. People under 25 were 6 percentage points less likely to be employed in 2007 than they were in 2000, while prime-age people were 1 to 2 percentage points less likely to be working. Older workers saw no employment declines during the 2001 recession and ended up about 3 to 4 percentage points more likely to be employed in 2007 than they were in 2000.

Persistently lower employment-population ratios among younger workers following the 2001 recession could suggest that employers’ interest in hiring these workers was weaker, leaving more of them on the sidelines of the labor market. A larger pool of people out of work might make it more difficult for any particular young worker to increase their income by asking for a raise or finding a new job match, since employers have many other options among those workers waiting to be hired. Did income growth actually slow for younger workers after the recession? If so, lower employment rates could be a contributing factor.

National trends show young workers lagging

Statistics from the Opportunity & Inclusive Growth Institute’s Income Distributions and Dynamics in America (IDDA) project show how incomes for different age groups evolved following the 2001 recession. These statistics use tax data to create several measures of the annual growth in adjusted gross income (AGI) for groups defined by age and initial income. Putting this income data together with the employment-population ratio data will show whether groups that saw larger declines in employment following the recession also experienced lower income growth than other age groups.

Figure 2 shows how median annual growth in household AGI changed over time.¹ Specifically, it shows the median change in household AGI between the year shown on the horizontal axis and the preceding year, in inflation-adjusted dollars. In order to consider how changes in income growth relate to age, which is a characteristic of a person rather than a household, each person is assigned the income growth realized by their household.² People in households with income starting in the bottom 25 percent of the national income distribution are shown in Panel A, and people in households starting in the top 10 percent of the national income distribution are shown in Panel B.

All groups saw income growth decline in 2001 and remain slower for several years, but the decline in growth tended to be larger for younger people. Among people starting out in the bottom 25 percent of the national income distribution, those under age 25 had median annual income growth that was nearly $3,000 lower between 2006 and 2007 than it was between 1999 and 2000. This decline was about three times as large in dollar terms as the next largest decline experienced by the other age groups over this period.

People starting out in the top 10 percent of the national income distribution saw much larger declines in annual income growth in dollar terms than those at the bottom of the income distribution. The decline in income growth was steepest for the youngest (under 25) and oldest (over 65) workers, but income growth recovered more quickly for older workers.

Across states, growth slowed where employment fell

That both employment rates and income growth remained lower for younger workers in 2007 at the national level is consistent with the idea that there may be a relationship between declines in employment and declines in income growth. But doing this analysis at the national level provides only one test of this idea. Repeating the analysis for all 50 states provides 50 more tests of this hypothesis.

Figure 3 describes how changes in employment relate to changes in median income growth at the state level, focusing on people whose income is low (in the bottom 25 percent), a group whose employment and income tend to be more adversely affected by weak labor market conditions. Again, groups in which employment declined by more than other groups also tended to see median income growth decline by more, especially for younger people. Each point represents an age group within a state. Each point’s position on the horizontal axis corresponds to the change in the group’s employment-population ratio between 2000 and 2007, in percentage points. Each point’s position on the vertical axis corresponds to the change in median annual income growth (in dollars) for the group’s members over that same period.

For the youngest workers, shown in gold, states where employment declined more (those located further left on the horizontal axis) also tended to see larger declines in income growth (they are located lower on the vertical axis). The slope of the dashed gold regression line quantifies this relationship and suggests that each percentage point reduction in young workers’ employment-population ratio was associated with income growth being about $200 lower for the median person. Similar relationships are evident in other age groups, but their magnitudes are notably smaller, consistent with the national evidence in Figure 2.

In addition to the typical experience reflected by median income growth, Figure 3 also shows how changes in strong and weak growth experiences relate to changes in employment. Here, “strong” growth is annual growth that is greater than that experienced by 90 percent of people (that is, the 90th percentile of income growth), and “weak” growth is annual growth that is greater than that experienced by only 10 percent of people (the 10th percentile of income growth).

Figure 3 does not suggest that weak income growth becomes systematically slower in places with larger declines in employment, even for young workers. In other words, a weak year of income growth does not seem to have gotten much worse in places where employment fell more. In contrast, the plot of strong income growth in Figure 3 looks more like the median growth plot, suggesting that a strong year of income growth was slower in 2007 than it was in 2000 in places where employment declined by more, especially for younger workers.

Both national- and state-level evidence suggest that groups that saw larger declines in employment following the 2001 recession also saw typical income growth slow by more, especially younger workers with relatively low incomes. State-level evidence further suggests that, among people starting with relatively low incomes, strong income growth was also slower in groups that saw larger declines in employment.

Deterioration in what constitutes good income growth for low-income young people could inhibit upward mobility, making it more difficult for them to sustain higher incomes later in life. Though not necessarily causal, this is consistent with the idea that having more people out of work (“slack” in the labor market) can make it harder to improve one’s economic circumstances. Unfortunately for young workers, persistently weaker labor market conditions following the 2001 recession may have left them more vulnerable to the Great Recession, which research suggests also had more adverse long-term effects on the earnings of younger workers.

Endnotes

1 A household consists of all people living at a given address. In some cases, multiple 1040 tax returns may be filed from one household (e.g., if the household consists of roommates, unmarried partners, multi-generational families, etc.). The household AGI measure used here adds up AGI across all 1040 tax returns within each household.

2 When using a household-level income measure like AGI, falling employment could mechanically reduce income growth. Reducing the number of people contributing to a household’s income, which happened much more to households containing the youngest workers during and shortly after the 2001 recession, will reduce its income. But this mechanical effect on growth should be muted in later years and may even increase income growth as employment-population ratios recover. Considering median income growth, as in Figure 2, also limits the influence of movements in or out of employment because the median will not be affected much by changes in employment within households that otherwise would have had income growth below or much higher than the median. For these reasons, the slower growth among younger workers later in the recovery shown in Figure 2 more likely reflects other factors, such as differences in the labor market conditions they face, rather than mechanical effects of declining employment.