By William R. Emmons, Lead Economist in Supervision
In an earlier post, I showed that all net employment gains in the last 20 years were in the 60-and-older age group. Among the 16-to-59 age group, employment was volatile but essentially unchanged during that time, on balance.
By examining data for these two broad age groups, I concluded that about 12 million fewer workers under 60 were employed at the end of 2020 than would have been the case if age-specific employment-to-population (E-P) ratios had not changed from their 2000 levels. The E-P ratio is the percentage of the population that is employed.
It turns out that population growth in the 16-59 age group during the 20-year period (the extensive margin of employment change) was almost exactly offset by decreased employment intensity, measured by the E-P ratio (the intensive margin of change). That is, although the number of 16- to 59-year-olds grew, this increase did not translate into a similar rise in employment because the percentage of those employed within that age group declined.
In this post, I use annual data to create an approximate longitudinal representation of five-year birth year cohorts (a “quasi-panel”).The exception was the 16-19 age range, which covers only four years. For example, I track people born in the years 1956-60 every five years between 1975 and 2020, when their ages approximated standard five-year age ranges, like 20-24 in 1980, 25-29 in 1985, etc. Creating a real panel would require tracking individual workers over time.
I show that the majority of the “missing” workers under 60 were, in fact, under 35 years old in 2020. Specifically, for those born after 1985, 5.4 million more men and 3.2 million more women would have been employed in 2020 if age- and gender-specific E-P ratios had not declined for those groups between 2000 and 2020. People between 35 and 59 years old in 2020, making up 55% of the 16-59 age group, represented only 33% of the missing workers. In fact, no workers were missing in the 55-59 group.
Low employment rates among young adults are concerning because research shows that scarring—an increased risk of future unemployment as well as diminished earnings prospects when finding a job—is more consequential for that group.
Using standard age ranges, the figures below show that E-P ratios declined between 2000 and 2020 for all young (ages 16-34) and middle-aged (ages 35-54) groups while they increased among older people (ages 55 and older). This pattern held for both men and women, as seen below.
The 16-24 age group had the largest decline in E-P ratio between 2000 and 2020, totaling 15.5 percentage points for men and 12.1 percentage points for women. Declines were 10.5 percentage points for men and 3.6 percentage points for women in the 25-34 age range. The largest increases were in the 65 and older groups, totaling 5.2 and 5.3 percentage points, respectively, for men and women.
Relatively weak attachment to employment early in life—whether voluntary or involuntary—may have long-lasting negative consequences for the affected individuals and the economy over time. Scarring that occurs among young adults today could persist, affecting their future employment prospects and reducing the economy’s potential growth. Therefore, it is natural to track cohorts of workers—groups of people born in specific time periods—as they traverse their life cycles.
The next two figures compare the E-P ratios of ten birth cohorts in particular age ranges, beginning with people born in the years 1956 through 1960. These ten groups constitute the entire population from the ages of 16 to 64 as of 2020. Men and women are shown separately because their E-P ratios are, in some cases, quite different. The figures below focus on the life-cycle stages before age 40, which are the most critical for potential long-term scarring from unemployment.
Looking first at men, a striking pattern within each age category (for example, ages 20-24) is a tendency for E-P ratios among more recent birth cohorts to decline as each subsequent birth cohort experiences that life-cycle stage. Seventy five percent of the cohort of men born in the years 1956 to 1960 were employed in 1980, when they were 20 to 24 years old. Twenty years later, in 2000, the 1976-80 cohort had an E-P ratio of 77% when they were 20 to 24. But in 2020, the ratio for the 1996-2000 cohort was only 61% when they were 20 to 24. The decline across recent cohorts is sharpest in the 16-19 age range, but it is evident in all age groups under 40.
Among women, the declines in E-P ratios over time are sharpest in the two youngest age ranges. Comparing the cohorts discussed above—1956-60, 1976-80 and 1996-2000—E-P ratios when these women were in the 20-24 age range were 62%, 68% and 58%, respectively. Thus, the E-P decline was 10 percentage points between women in the latter two cohorts from 2000 to 2020.
Decomposing changes in employment between 2000 and 2020 into extensive and intensive margins of change by birth cohort and gender, I find that two-thirds of missing employment can be attributed to people born after 1985.See my earlier article for details on decomposing employment changes into their extensive and intensive margins.
The table below shows that men and women born after 1985 constituted 45% of the population aged 16-59 in 2020, but the 8.6 million “missing workers” under age 35 made up 67% of the 12.7-million-worker shortfall.Using annual data and five-year birth cohorts to create a quasi-panel for this article, I estimate there were 12.7 million missing workers in the 16-59 age group in 2020. In my previous article, I used monthly not seasonally adjusted data and a single age range (16-59), finding 11.9 million fewer workers in December 2020 than if the group’s E-P ratio had not changed from December 2000.
|By Birth Year|
|Share of All "Missing" Workers Ages 16-59 in 2020 (Percentage)|
|Share of Population Ages 16-59 in 2020 (Percentage)|
|SOURCES: Bureau of Labor Statistics and author’s calculations.|
|NOTES: Actual employment for each birth cohort is for the year in which the age range corresponded exactly to the cohort’s range of birth years. Predicted employment is computed as the actual 2000 E-P ratio for each age and gender cohort multiplied by the group’s actual 2020 population.|
The most disproportionate effects were among men born after 1985. This group made up 22% of the population between 16 and 59 in 2020, but it accounted for 43% of the missing workers. Among women, the hardest hit by lower-than-predicted employment were those born after 1995; making up 10% of the population between 16 and 59, they accounted for 18% of these missing workers.
In sum, weakness in the job market in 2020 was experienced very differently across age groups and genders. Young men and women felt the greatest impact of lower employment during that period.