The headline month-to-month payroll estimates are produced by the BLS via the Institution Survey, a part of the Present Employment Statistics (CES) program. The survey collects responses from roughly 119,000 companies and authorities businesses, protecting about 622,000 particular person worksites.
As a result of these figures are derived from a pattern reasonably than a full rely of employment, they’re topic to statistical estimation and revision. The closest approximation to a complete rely of jobs is the Quarterly Census of Employment and Wages (QCEW), which compiles administrative data from unemployment insurance coverage filings and covers roughly 97% of US employment. Because of this, the QCEW serves because the benchmark to which nonfarm payrolls are periodically revised.
Annually, the BLS benchmarks the CES knowledge to the QCEW. On this course of, the payroll stage for the earlier March is in contrast with the QCEW estimate, and the distinction is distributed evenly throughout the prior 12 months, reasonably than making use of it all of sudden (a linear “wedging” adjustment). The result’s that the extent of nonfarm payrolls is introduced into alignment with QCEW knowledge for March of the given benchmark 12 months.
In recent times, these benchmark revisions have been comparatively massive and persistently adverse. Over the previous three years alone, changes have diminished beforehand reported payroll employment by a mixed 1.75 million jobs.
