The Establishment Survey
The establishment survey, formally called the Current Employment Statistics Survey, gathers data from approximately 122,000 nonfarm businesses and government agencies for some 666,000 work sites and about one-third of all payroll workers. The survey is based on the weekly pay period that includes the 12th day of the month
Anyone on the payroll of a surveyed business during that reference week, including part-time workers and those on paid leave, is included in the count used to produce an estimate of total U.S. nonfarm payrolls
Farm workers are not included because of agriculture’s seasonal nature; the sector’s reliance on self-employment, unpaid family work, and undocumented workers; and its partial exemption from unemployment insurance requirements, since those records are used to compile the survey sample. The payroll data also does not include self-employed workers.
The establishment survey provides estimates for nonfarm payrolls, average weekly hours worked, and average hourly and weekly earnings nationwide as well as by state and metropolitan area, and by industry. The report also tallies the hours worked and earnings of production and nonsupervisory employees.
Survey results are adjusted based on the “birth/death model”—for businesses rather than people. The BLS uses it to estimate monthly payroll changes resulting from the openings and closings of businesses beyond its survey sample.
The numbers are seasonally adjusted to filter out fluctuations like the annual surge in retail sector hiring ahead of the holiday shopping season and the wintertime slowdowns in construction, though the BLS also provides the data without the seasonal adjustments.
Nonfarm payroll totals are adjusted in each of the two monthly jobs reports following the initial release to incorporate additional survey responses and the latest seasonal adjustment factors. They’re also subject to annual revisions benchmarking them to updated counts from unemployment insurance tax records
The Household Survey
The household survey is based on monthly interviews of 60,000 households conducted for the BLS by the U.S. Census Bureau. Survey participants are asked about their employment status during the week including the 12th day of the month.
4
The most prominent product of the household survey is the official, or U-3, unemployment rate, calculated as a percentage of the unemployed actively seeking work relative to the labor force, or the sum of the employed and the unemployed. To be officially counted as unemployed, the survey respondent has to have been available for work in the reference week and made specific efforts to find work during the four prior weeks, unless awaiting an expected recall from a layoff.
