The monthly jobs report measures U.S. employment in two ways: employer payroll jobs and the employment status of people age 16 and older.
It is not one head count, but the Bureau of Labor Statistics' Employment Situation release built from two sample surveys. One survey produces the headline change in nonfarm payrolls, along with earnings and hours. The other produces the unemployment rate and estimates of employment, unemployment and labor-force participation.
Table of Contents
- One report, two surveys
- Payroll jobs are not the same as workers
- What the unemployment rate means
- Why the two surveys can appear to conflict
- How to read the first estimate
One report, two surveys
The establishment survey collects payroll information from about 119,000 businesses and government agencies representing roughly 622,000 worksites. It measures nonfarm wage-and-salary jobs, average hours and earnings. The household survey covers about 60,000 eligible households.
It classifies people in the civilian noninstitutional population age 16 and older by employment status. The BLS Employment Situation technical note explains that the monthly release combines these separate surveys. Because the surveys measure different units, their results answer different questions. Payroll data show how the number of covered jobs changed; household data show how many people were employed or unemployed under official definitions.
Payroll jobs are not the same as workers
The headline payroll figure counts jobs, not unique people. A person with two covered payroll jobs appears twice, according to the BLS guide to Employment Situation data. Losing one of those jobs would reduce payroll employment even if that person remained employed elsewhere. Payroll employment also excludes several kinds of work.
The household survey includes self-employed people, agricultural and private-household workers, unpaid family workers, and people on unpaid leave. The payroll survey generally does not include those groups, as the BLS comparison of payroll and household employment details. That difference matters when interpreting a headline. An increase in self-employment could appear in household employment without increasing nonfarm payrolls. Conversely, one worker taking a second payroll job could raise the payroll count without increasing the number of employed people.
What the unemployment rate means
The official unemployment rate comes from the household survey. A person qualifies as unemployed only when all three conditions apply: the person had no job, was available for work and actively sought work during the previous four weeks. There is an exception for people awaiting recall from a job.
📨 Get Free News Stories Alerts
Free · No spam · Unsubscribe anytime
Someone without work who does not meet the active-search requirement is not included in the official unemployed count. The unemployment rate therefore does not answer every question about financial hardship or interest in working. It measures the share meeting a specific labor-market definition, not everyone who lacks a paycheck or wants a different job.
Why the two surveys can appear to conflict
The surveys differ in whom they cover, whether they count people or jobs, and how they define employment. Their reference periods are also close but not identical. The household survey generally covers the calendar week containing the 12th of the month. Payroll data cover each employer's pay period containing the 12th.
Those pay periods can span different dates. Sampling uncertainty adds another source of divergence. BLS says a monthly payroll change of about 122,000 is statistically significant, compared with about 676,000 for household employment. A modest one-month increase or decrease may therefore reflect sampling variation rather than a clear underlying shift.
How to read the first estimate
The initial payroll number is preliminary. BLS revises it in each of the next two monthly releases, then annually re-anchors payroll estimates using near-complete unemployment-insurance tax records, according to its payroll survey FAQ.
A careful reading should include these checks: If two headlines appear inconsistent, first check whether one cites payroll jobs and the other cites household employment. They may be reporting different estimates rather than contradicting each other.
- Identify whether the number describes payroll jobs, employed people, unemployment, earnings or hours.
- Confirm whether the figure counts jobs or unique workers.
- Treat the first payroll estimate as an early reading rather than a final count.
- Check the next two releases for revisions to the month in question.
- Be cautious about drawing broad conclusions from a small one-month movement.
You Might Also Like
- Siemens to Add 1,400 Jobs in Jackson County, Georgia, With $185 Million Investment
- Texas Freezes Data Center Grid Approvals Pending State Audit
- How the Federal Reserve Sets Interest Rates