In a made-up town, 600 people have jobs. Its unemployment rate falls from 6.25% to 3.23%. Nobody finds work.
Twenty job seekers stop looking. They leave the unemployment count. The same 600 people still have jobs.
The headline looks better because people stopped searching, not because they found work. Reading the US jobs report means knowing who each number counts and who it leaves out.
One report, two surveys
The Employment Situation is the monthly US jobs report from the Bureau of Labor Statistics (BLS). Its two surveys ask different questions.
- The establishment survey counts nonfarm payrolls: paid civilian jobs at businesses and government agencies, excluding farm work and jobs in private households. It also measures private-sector earnings and hours.
- The household survey asks people about work and job searching. It supplies the unemployment and labor-force participation rates.
If you work at a shop and a warehouse, payrolls count two jobs. The household survey counts one employed person: you. Someone working for themselves can count as employed without being on an employer's payroll.
Two payroll jobs can belong to one person. The surveys measure different things, so their changes can disagree.
The payroll headline is the net change since the previous month, after jobs added and lost. It does not count everyone hired that month. The headline is seasonally adjusted to remove recurring patterns such as holiday hiring.
Who counts as unemployed
In US statistics, unemployed means without a job, available for work and actively looking during the four weeks ending with the survey week. Applying for a job counts; browsing listings alone does not.
The labor force combines employed and unemployed people. The unemployment rate tells you what share of that group is unemployed.
The labor-force participation rate divides the labor force by the eligible population: civilians age 16 and older living outside institutions such as prisons and nursing homes. The official term is civilian noninstitutional population.
Keep the town's eligible population at 1,000: 600 employed, 40 unemployed and 360 outside the labor force. Its labor force is 600 + 40 = 640.
Unemployment is 40 ÷ 640 × 100% = 6.25%. Participation is 640 ÷ 1,000 × 100% = 64%.
Twenty job seekers then go more than four weeks without looking. Only 20 remain unemployed, and the labor force shrinks to 600 + 20 = 620. Unemployment becomes 20 ÷ 620 × 100% ≈ 3.23%; participation becomes 620 ÷ 1,000 × 100% = 62%.
| Measure | Before | After |
|---|---|---|
| Employed | 600 | 600 |
| Unemployed | 40 | 20 |
| Outside labor force | 360 | 380 |
| Labor force | 640 | 620 |
| Unemployment rate | 6.25% | 3.23% |
| Participation rate | 64% | 62% |
The employed share stays at 60% in both bars. Only the boundary between job seekers and people outside the labor force moves.
A lower unemployment rate is not proof of more work. Participation helps explain where the job seekers went.
What strong hiring looks like
There is no single monthly payroll gain or unemployment rate that means healthy in every era. A growing labor force needs more new jobs just to keep unemployment steady. Compare several months and check whether gains spread across industries or depend on one pocket of hiring.
Unemployment can rise alongside hiring if more people enter the labor force than find work. Participation helps explain that result. Its direction also needs context: retirement can lower participation without signaling a shortage of jobs.
Average hourly earnings divides gross payroll by paid hours across private nonfarm jobs. It does not track the same workers' raises: the mix of jobs also matters.
Say two workers put in equal hours at $20 and $40 an hour. Their average is ($20 + $40) ÷ 2 = $30. If the $20 job disappears, the average becomes $40. Nobody gets a raise.
Hours matter to your paycheck too. At $20 an hour, 30 paid hours means $600 before tax; 25 hours means $500. You still count as employed in both cases. Wage growth also needs an inflation comparison to tell you what that pay buys.
Jobs, participation, pay and hours give you different clues about the economy. Together they also help explain changes in tax receipts and benefit spending.
Use the revised trend
BLS revises the previous two payroll months as more employers respond and seasonal factors are updated. The separate annual payroll benchmark revision checks job totals against fuller employment counts, mainly from unemployment-insurance tax records.
The revision changes what we know about last month, not what happened in it.
Take a fictional release. This month adds 180,000 jobs. Last month's 200,000 gain is revised to 150,000.
The two-month gain is 150,000 + 180,000 = 330,000. Keeping the old figure would give 200,000 + 180,000 = 380,000. Payrolls still grew, but the revised gain is 50,000 smaller.
Payroll growth also accelerated from the revised 150,000 to 180,000. Using the old 200,000 would make it look like growth slowed. The latest number needs the latest comparison.
For the market reaction, compare hiring with forecasts as well as with last month's revised gain. Stronger-than-expected hiring can support the sales outlook while pushing expectations for Fed policy toward higher rates.
Those forces can pull share prices in opposite directions. Their balance changes with the economy; interest rates affect stocks through more than one channel. A strong jobs headline has no fixed stock-market reaction.
Read jobs alongside output and spending to see the broader pattern. That is the next step in understanding the business cycle.
In short
- Payrolls count jobs; the household survey counts people.
- The unemployment rate divides unemployed people by the labor force, not by everyone.
- A falling unemployment rate can mean people stopped searching, while a rising rate can accompany hiring.
- Average pay can rise without anyone getting a raise; check hours, the mix of jobs and the revised trend too.
- A jobs surprise can change both expected company earnings and expected interest rates.