Every month, the U.S. Bureau of Labor Statistics releases “The Employment Situation,” widely known on Wall Street as the monthly jobs report. Market participants frequently watch two headline numbers: the change in total nonfarm payroll employment and the civilian unemployment rate. Yet behind this single press release sit two entirely independent surveys with distinct sample sizes, counting rules, and collection methodologies: the Establishment Survey and the Household Survey.
Because the two surveys measure different slices of the labor market, they frequently appear to send conflicting signals. Nonfarm payrolls can surge by hundreds of thousands of jobs in a month while the unemployment rate simultaneously ticks higher. Understanding how both surveys operate is essential for investors seeking to decipher whether hiring momentum is expanding or softening. For broader orientation on economic indicators and market mechanics, visit our learning center.
The Two Surveys Behind the Jobs Report
The monthly employment report combines findings from two separate statistical operations conducted under federal guidelines:
- The Establishment Survey (Current Employment Statistics, or CES): Administered directly by the BLS, this survey collects payroll records from approximately 119,000 businesses and government agencies representing roughly 629,000 individual worksites. It produces the headline “nonfarm payrolls” number, along with data on average hourly earnings and weekly hours worked.
- The Household Survey (Current Population Survey, or CPS): As noted by the U.S. Census Bureau, the survey is sponsored jointly by the U.S. Census Bureau and the U.S. Bureau of Labor Statistics (BLS), is the primary source of labor force statistics for the population of the United States. Conducted by Census field representatives across a probability sample of approximately 60,000 eligible households, it determines how many individuals are employed, unemployed, or outside the labor force.
According to the official Bureau of Labor Statistics Employment Situation Technical Note, “The household survey provides information on the labor force, employment, and unemployment that appears in the ‘A’ tables, marked HOUSEHOLD DATA.” Meanwhile, the establishment survey generates the nonfarm payroll figures found in the “B” tables. When timely reports emerge—such as our recent analysis of how September jobs added just 29,000 positions—traders react to both sets of tables simultaneously.
Structural Differences: What Each Survey Measures
The primary source of divergence between payroll changes and the unemployment rate lies in their structural definitions. The establishment survey counts jobs, whereas the household survey counts people.
As the BLS technical note explicitly highlights: “In the establishment survey, employees working at more than one job and thus appearing on more than one payroll are counted separately for each appearance.” If an individual works one full-time position at a logistics warehouse and a second part-time weekend job at a retail store, the establishment survey records two nonfarm payroll jobs. In contrast, the household survey interviews the person and counts them as exactly one employed individual.
Furthermore, the household survey includes agricultural workers, unincorporated self-employed individuals, unpaid family workers, and private household workers among the employed. All of these categories are excluded from nonfarm business payrolls. Age restrictions also differ: the household survey restricts its labor force universe to civilian noninstitutional individuals aged 16 and older, whereas payroll records measure jobs regardless of employee age.
| Feature | Establishment Survey (CES) | Household Survey (CPS) |
|---|---|---|
| Primary Metric | Nonfarm payrolls, hours, earnings | Unemployment rate, labor force participation |
| Sample Size | ~119,000 businesses (~629,000 worksites) | ~60,000 eligible households |
| What Is Counted | Jobs on commercial and government payrolls | Persons (employed, unemployed, not in labor force) |
| Multiple Jobholders | Counted on each payroll separately | Counted once as employed |
| Self-Employed & Farms | Excluded from nonfarm payrolls | Included in household employment |
| Reference Period | Pay period including the 12th of the month | Calendar week containing the 12th of the month |
Why Payrolls and the Unemployment Rate Diverge: A Worked Example
Consider an illustrative scenario where nonfarm payrolls expand by 250,000 in a given month, yet the reported unemployment rate rises from 4.0% to 4.2%. Many observers mistakenly assume one of the numbers must be incorrect. In reality, mathematical definitions and labor force dynamics fully explain the divergence.
The unemployment rate is defined as:
Unemployment Rate = (Unemployed Persons / Civilian Labor Force) × 100
To be classified as unemployed in the household survey, a person must not have had a job during the survey week, must have been available to work, and must have actively sought employment during the preceding four weeks. If an individual gives up searching, they exit the labor force entirely and are neither employed nor unemployed.
Now observe the numerical mechanics in our worked model:
- Starting baseline: Assume a civilian labor force of 168,000,000 individuals, comprising 161,280,000 employed workers and 6,720,000 unemployed job seekers. The initial unemployment rate equals (6,720,000 / 168,000,000) × 100 = 4.00%.
- Subsequent month changes: Robust hiring optimism encourages 500,000 previously discouraged individuals to resume looking for work, expanding the labor force to 168,500,000. Over that same period, household employment grows by 150,000 (reaching 161,430,000).
- New outcome: Because 500,000 entered the labor pool but only 150,000 secured positions, the number of officially unemployed persons rises by 350,000 to 7,070,000. The new unemployment rate becomes (7,070,000 / 168,500,000) × 100 = 4.20%.
Simultaneously, if existing workers take on 100,000 secondary part-time payroll jobs and businesses create 150,000 net new positions, the establishment survey reports a robust gain of 250,000 nonfarm payrolls. Both metrics accurately captured their respective domains: business hiring expanded, but labor force entrants outpaced net absorption.
Alternative Measures of Underutilization: U-1 to U-6
Headline unemployment is technically labeled U-3. However, economists recognize that U-3 omits workers who desire full-time work but cannot find it, or those who have ceased searching due to frustration. To provide a broader perspective, the BLS computes six measures of labor underutilization (U-1 through U-6):
- U-1: Persons unemployed 15 weeks or longer, as a percentage of the civilian labor force.
- U-2: Job losers and persons who completed temporary jobs, as a percentage of the civilian labor force.
- U-3: Official civilian unemployment rate (total unemployed as a percentage of the civilian labor force).
- U-4: U-3 plus “discouraged workers”—individuals who want a job and are available to work, but have ceased looking because they believe no jobs are available.
- U-5: U-4 plus all other “marginally attached workers” who desire work and have looked within the past 12 months, but not within the past 4 weeks.
- U-6: U-5 plus total employed part-time for economic reasons (often called underemployed workers who seek full-time hours but work part-time due to slack business conditions).
Because U-6 encompasses underemployed individuals and marginally attached job seekers, it provides a comprehensive assessment of labor slack. When U-6 rises while U-3 holds steady, it often reveals that businesses are cutting employee hours rather than initiating outright layoffs.
Birth-Death Adjustments and Benchmark Revisions
When reviewing establishment survey payroll numbers, investors must account for statistical modeling techniques. Because sample surveys cannot immediately detect every new company formed or business closed each month, the BLS employs a net birth-death model to estimate hiring from newly opened and shuttered enterprises.
During economic turning points, the birth-death model can introduce lag into initial estimates. To preserve long-term accuracy, the BLS conducts an annual benchmark revision that recalibrates monthly payroll survey samples against comprehensive administrative tax records from the Quarterly Census of Employment and Wages (QCEW), which covers approximately 95% of all U.S. jobs. Understanding these revisions prevents market participants from overinterpreting first-run preliminary prints. Just as inflation data requires deep dive analysis—such as our walkthrough on how to read the monthly CPI report—labor market data requires evaluating both surveys in tandem.
What to Watch in Future Releases
When tracking upcoming labor releases, follow this structured evaluation sequence:
- Examine the household employment change alongside payrolls: Compare the net payroll gain with the civilian employment count to detect discrepancies driven by multi-job holding or self-employment shifts.
- Inspect labor force participation: Check whether changes in the unemployment rate were caused by hiring gains or by workers entering or leaving the labor pool.
- Review U-6 and wage growth: Track part-time for economic reasons and average hourly earnings to gauge underlying wage pressure and labor market tightness.
Sources
- Bureau of Labor Statistics Employment Situation Technical Note — Detailed collection methodologies, sampling confidence intervals, and conceptual definitions for the CES and CPS surveys.
- U.S. Census Bureau Current Population Survey (CPS) — Survey structure, household sampling framework, and labor force statistical standards.
Disclosure: This article is for informational purposes only and is not investment advice.