How many jobs were lost and then regained during the US pandemic recession and recovery (Feb 2020–2023) in BLS data?
Executive summary
The Bureau of Labor Statistics’ payroll series shows an unprecedented collapse of pandemic">U.S. nonfarm payrolls after the February 2020 peak — a combined decline of 22.4 million jobs in March and April 2020 — followed by a rebound that returned employment to and slightly above its prepandemic level by late 2022, meaning the BLS-conducted payroll count of jobs lost was effectively fully regained during the recovery [1] [2]. Household-survey measures and subgroup outcomes paint a more nuanced picture: official counts recovered by 2022–23, but some workers and measures of underutilization lagged [3] [4].
1. The scale of the initial job loss: an historic two-month collapse
After peak employment in February 2020, nonfarm payroll employment plunged by a combined 22.4 million jobs in March and April 2020 — a 15 percent decline that dwarfed recent recessions and constitutes the clearest BLS measure of how many payroll jobs were lost at the nadir [1].
2. The path of the recovery: returning to — and then exceeding — the February 2020 payroll level
The employer (payroll) series shows recovery progressed rapidly through 2021 and 2022 and, by November 2022, total nonfarm employment had risen to 100.7 percent of its February 2020 level, indicating the payroll jobs lost had been fully recovered and marginally exceeded in absolute terms by that date [2]. The BLS and other analysts treat that November 2022 milestone as the point at which payroll employment reclaimed the prepandemic peak [2].
3. How to translate those percentages into “how many jobs were lost and regained”
Using the BLS headline figures: the immediate job loss from the February 2020 peak was about 22.4 million payroll positions in March–April 2020 [1]; by November 2022 payroll employment reached 100.7 percent of the February 2020 total, meaning those roughly 22.4 million lost payroll jobs had been regained and the total payroll count was modestly higher than the prerecession peak [2]. Some BLS products and press summaries also note large annual losses earlier in 2020 and strong gains thereafter, but the 22.4 million drop and the return to 100.7 percent are the core BLS benchmarks for “lost and then regained” 2023/data-on-display/employment-gains.htm" target="blank" rel="noopener noreferrer">[5] [2] [1].
4. Important caveats: survey differences, revisions, and unequal recovery
BLS publishes two primary measures — the employer payroll (CES) and the household survey (CPS) — which differ in coverage (for example, CPS includes self-employed and agricultural workers) and occasionally diverge; BLS cautioned that 2023 census population adjustments affect comparability of household-series levels to earlier years [3]. The agency also identified categorization errors early in the pandemic that likely understated unemployment initially, and it flags persistent labor underutilization and elevated part‑time-for‑economic‑reasons counts even as headline payroll totals recovered [4] [6]. Moreover, industry- and demographic-specific recoveries varied: leisure and hospitality saw very large swings and only gradually recouped losses, and some groups were slower to return to work [6] [7].
5. The broader reading: full payroll recovery but lingering frictions
By late 2022 and through 2023 the BLS and policy analysts judged that payroll employment had returned to prepandemic levels and that, from a jobs-and-growth standpoint, the economy was largely healed by end of 2023 — a conclusion echoed in policy research noting regained aggregate employment and continued job-openings strength even as the labor market cooled in 2023 [3] [8]. That aggregate restoration, however, sits beside continued questions about distributional scars, survey comparability, and the composition of jobs regained [9] [4].