Reading a Jobs Report: The Three Numbers Traders Watch

On the first Friday of most months, at 8:30 AM Eastern, the Bureau of Labor Statistics releases the Employment Situation report — and markets jerk. Bond yields jump and equity futures swing, all from a survey-based snapshot of one month’s labor market. What are traders actually reading in those tables?

Why one report is really two surveys

The Employment Situation is built from two separate surveys that frequently disagree at the edges.

The establishment survey (Current Employment Statistics, or CES) polls roughly 142,000 businesses and government agencies covering about 689,000 worksites. It asks employers who was on payroll during the pay period including the 12th of the month. It counts positions, so a person with two jobs is counted twice — and it excludes farm workers, the self-employed, and private household workers. What it gives you: the headline payrolls number, with detailed industry breakdowns.

The household survey (Current Population Survey, or CPS) interviews about 60,000 households, conducted by the Census Bureau for the BLS. It counts people — a worker with two jobs appears once — and captures the self-employed and farm workers. What it gives you: the unemployment rate, with detailed demographic breakdowns.

The two surveys diverge for legitimate methodological reasons — and each owns one headline: payrolls come from the establishment survey, the unemployment rate from the household survey. At turning points, watch both.

The adjustments inside the sausage

Raw monthly data is lumpy — holiday retail hiring, winter construction pauses — so BLS seasonally adjusts the series. The subtler adjustment is the birth-death model: new firms take roughly a year to reach the survey’s sampling frame, and dead firms simply stop responding, so BLS models the net employment change from business births and deaths using past Quarterly Census of Employment and Wages (QCEW) data — the near-universe count of payroll jobs from state unemployment-insurance records.

One detail trips people up every cycle: the net birth-death figures are not seasonally adjusted, and they are applied to the not-seasonally-adjusted estimates before the headline number is adjusted. Critics who “discover” hundreds of thousands of model-added jobs are usually mixing adjusted and unadjusted numbers — the BLS publishes the model components openly. The agency’s own caveat: the model assumes historical patterns continue, so it is weakest at economic turning points.

Revisions: the monthly kind and the annual kind

The prior two months’ payroll figures are revised with every release as late reports arrive — and revisions are frequently larger than the headline surprise. Professionals read the three-month average, not the single print.

Then, once a year, BLS re-anchors the whole series to the QCEW full population counts for March of the prior year — the benchmark revision. A preliminary revision is announced in late summer (for March 2026 it was −79,000, released August 28, 2026); the final revision lands with the January release the following February. Over the last decade the absolute average revision has been just 0.2% of total nonfarm employment — small as a share of 150+ million jobs, but enough to redraw the narrative of a year.

Number 1: Nonfarm payrolls (the headline)

This is the net number of jobs added or lost in the month, excluding farm workers. Consensus expectations are set beforehand, and the surprise — actual versus expected — is what moves markets. A big beat suggests a hot economy, but it also means the Fed is less likely to cut rates, which is why strong jobs data sometimes sells off equities. Context decides the reaction: in a slowdown, good news is good news; in an inflation scare, good news is bad news.

Number 2: The unemployment rate

The share of the labor force actively seeking work, from the household survey. It moves slowly — labor markets turn like cargo ships — which makes sustained changes powerful signals. The unemployment rate rises around the start of every US recession; the Sahm rule, named for economist Claudia Sahm, formalizes the pattern: when the three-month average of unemployment rises half a percentage point above its 12-month low, the economy has historically been entering a downturn. Its track record is unusually clean, though Sahm herself has warned against relying on any single indicator.

But the rate has quirks: it can fall for the wrong reason (people giving up and leaving the labor force) and rise for the right one (optimistic workers re-entering the job hunt). Always check the labor force participation rate alongside it.

Number 3: Average hourly earnings

Wage growth is the Fed’s favorite page of the report. Wages rising faster than productivity feed directly into service-sector inflation — the cost-push channel. Hot wages keep rate-cut hopes in check; cooling wages let the Fed breathe. Of the three numbers, this one moves bond yields most directly.

Beyond the headline trio: the U-1 through U-6 range

The official unemployment rate is U-3 — just one of six measures BLS publishes each month:

• U-1: people unemployed 15 weeks or longer
• U-2: job losers and people who completed temporary jobs
• U-3: the official rate — all unemployed as a share of the labor force
• U-4: U-3 plus discouraged workers (those who quit looking for job-market reasons)
• U-5: U-4 plus all other marginally attached workers
• U-6: U-5 plus people working part-time for economic reasons — who want full-time work but can’t find it

U-6 is the widest measure of labor underutilization: when a recovery is driven by part-time work, U-3 falls while U-6 stays elevated — a weaker labor market than the headline suggests.

Participation and the prime-age ratio

The labor force participation rate — the labor force as a share of the civilian noninstitutional population — has been drifting down for decades, mostly because the population is aging. That secular drag makes the headline rate misleading over long horizons. The cleaner gauge is the prime-age (25–54) employment-population ratio: the share of the core working-age population with a job. It strips out the two big demographic distortions — young people staying in school longer and older workers retiring — and it isn’t affected by people entering or leaving the labor force.

Wages beyond hourly earnings: the Employment Cost Index

Average hourly earnings has a composition problem: if hiring shifts toward lower-paying industries, average wages can fall even as nobody’s pay was cut. The Employment Cost Index (ECI) fixes this by tracking the cost of employing a fixed basket of occupations — like a CPI for labor. It covers wages and salaries plus employer-paid benefits, and BLS publishes it quarterly. The Fed watches the ECI closely — former chairs have called it indispensable — so the quarterly release can move rate expectations on its own. If hourly earnings are hot but the ECI is cooling, the wage scare may be a composition effect, not genuine pressure.

How the Fed reads it

The Fed’s dual mandate is stable prices and maximum employment, and the jobs report speaks to both: strong payrolls, falling unemployment, and accelerating wages argue for holding rates higher; the reverse argues for cuts. Read the report as a Fed input first, an economic scorecard second.

A worked walkthrough: expectations vs. surprise

Take a hypothetical release. Consensus expects +175,000 jobs; the print is +95,000 — a miss. But unemployment holds at 4.2% while average hourly earnings rise 0.4% on the month against 0.3% expected. The payrolls miss is dovish (slower hiring); the hot wage print is hawkish (persistent inflation pressure). Bonds tend to rally on the jobs miss, but rate-cut bets only extend so far — and equities get the worst of both: weaker growth and higher-for-longer rates. That mixed signal is why professionals never read the headline alone. The day’s trading is the market’s net vote on all three numbers at once.

How to read it like a pro

1. Compare to expectations, not to zero — markets price the consensus.

2. Read all three together: jobs strong + wages hot + unemployment low = hawkish; the reverse = dovish.

3. Check the three-month payroll average and the revisions, not the single print.

4. Cross-check the headline rate with U-6 and the prime-age ratio — they catch what the headline misses.

5. Read the BLS tables themselves: Table A (household) and Table B (establishment) carry the detail the press release summarizes.

6. Remember: one noisy month matters less than the trend over quarters, even if the day’s trading says otherwise.

The bottom line

The jobs report is really three reports — growth (payrolls), slack (unemployment), inflation pressure (wages) — built on two surveys, smoothed by seasonal adjustment, and corrected by revisions. Markets react to the mix, not the headline. Learn to read the mix, and the 8:30 AM chaos starts looking orderly. Pair it with our recession checklist and the GDP explainer for the full cycle picture.

Sources: U.S. Bureau of Labor Statistics (Employment Situation FAQ, CES handbook and benchmark pages, alternative measures of labor underutilization); FRED (Federal Reserve Bank of St. Louis); Bloomberg Law analysis of the CES/CPS surveys; Investopedia.

This article is educational and is not investment advice.