The US Employment Situation brings together payroll jobs, unemployment and earnings from different surveys. Reading them together helps you understand a headline surprise and its revisions. This article explains the units, the release calendar and the execution conditions to check around the event.

What the NFP report actually is and who publishes it

NFP is the headline nonfarm payroll-job estimate in the monthly US Employment Situation report, published by the Bureau of Labor Statistics. Check the official schedule: releases commonly fall on Friday at 8:30 a.m. Eastern Time, but exceptions occur. Convert the time for the actual date, including differences in daylight-saving transitions.

The establishment survey (CES) estimates nonfarm payroll jobs in private businesses and government. It excludes farm workers, private-household workers and the unincorporated self-employed, among other groups. A person working for two employers can count twice; the household survey (CPS) counts that person once. Payroll jobs are not full-time equivalents. CES refers to the pay period including the 12th; CPS generally refers to the week containing the 12th, not conditions at publication time.

Three extra payroll jobs, no new people

Invented firms within nonfarm employment have ten unique people, each initially working for one employer. Three of these same people take another job with a different employer.

Scroll sideways to see the full diagram

Additional payroll jobs held by the same peopleTen unique people initially hold ten payroll jobs. A, B and C retain their first job and take a second with another employer. There are still ten people, but thirteen payroll jobs.MORE JOBS, THE SAME PEOPLE3 of 10 people take an additional jobPeopleExisting jobExtra jobAjobjobBjobjobCjobjobThe other 7 people: one payroll job eachUnique people: 10 → 10, change 0Payroll jobs: 10 → 13, change +3A payroll job need not be full-time.

Firms and people are invented. Payroll coverage and reference period are unchanged; all jobs meet the model's criteria. Nobody leaves, no new person enters and coverage does not change. We do not reproduce BLS sample estimation or working hours. BLS: CES Frequently Asked Questions — Employment Concepts confirms that CES covers payroll jobs, including full- and part-time work; someone working for multiple employers is counted by each.

The diagram shows A, B and C with two payroll jobs each. The other seven people have one each.

Name the unit when interpreting NFP: payroll jobs, rather than unique people or full-time equivalents. In our model jobs change by +3 and people by 0. We calculate neither unemployment nor an actual household-survey result.

The three numbers the market watches

The report contains several measures with different units: payroll-job changes in thousands, unemployment as a percentage of the labour force, and private-sector average hourly earnings in dollars per hour or percentage changes. They are not three interchangeable measures of the same population.

Read the change in payroll jobs, the CPS unemployment rate and CES average hourly earnings with their definitions and revisions. There is no universal healthy payroll threshold. CES reports earnings rather than wage rates: gross payroll divided by total hours. It is not median pay, total employment cost or a direct inflation measure.

A fictional composition example has two groups of 50 people, each working one paid hour at USD 20 or USD 10. Total payroll is USD 1,500 for 100 hours: USD 15 per hour. In the next period only the USD 20 group remains at the same rate, giving USD 1,000 for 50 hours: USD 20 per hour. The average rises 33.33%, total payroll falls 33.33% and 50 jobs disappear, without a pay rise for anyone. This illustrates composition, not the complete CES estimator or an inflation forecast.

The unemployment denominator can also change. In a separate fictional model, 190 million employed and 10 million unemployed give 10 / 200 × 100% = 5%. If two million unemployed leave the labour force while employment stays unchanged, 8 / 198 × 100% ≈ 4.04%. The rate falls about 0.96 percentage point without an extra employed person. Inspect participation and employment alongside the percentage.

Why this single release moves the dollar

Labour-market information may change expectations about the Fed’s employment and inflation objectives. It does not force a rate decision or a dollar direction. A stronger or weaker report must be interpreted alongside its components, revisions, earlier expectations and other data.

Suppose, hypothetically, consensus is one hundred and eighty thousand jobs and the release is one hundred and eighty-five thousand. The five-thousand difference alone does not establish a negligible market reaction: other components, revisions and expectations also matter. There is no universal tens-of-thousands threshold for a currency move.

Keep revisions separate from the current surprise. In a fictional report, current payroll growth is +180 thousand against +150 thousand forecast: +30 thousand. Earlier increments change from +170 to +130 thousand and from +150 to +120 thousand: revisions −40 and −30, totalling −70 thousand. The new three-month mean is (120 + 130 + 180) / 3 ≈ 143.3 thousand. Unemployment at 4.2% against 4.1% forecast surprises by +0.1 percentage point; earnings growth at 0.2% against 0.3% surprises by −0.1 point. Do not add these percentages to thousands of jobs, or convert 30 minus 70 into a dollar signal.

CES revises estimates twice before annual benchmarking can change their history again. Save what was known at each decision. A later revision is not a second independent release of current-month employment, and using it to judge an earlier signal introduces later information.

What the first minute after the release looks like

Specify the event time and observation window when describing a post-release move. Do not assume every NFP event produces the month’s largest movement or a particular cross-asset path. Review actual quotes and execution conditions for the instrument you observe.

Order execution depends on the broker, instrument and market conditions. Spreads and fills can differ from a calm-period observation, and an ordinary stop loss is not a guarantee of the selected price. A price gap or rapid reversal can produce a different fill or, under some execution arrangements, a rejected request. Check the agreement and platform rules; no fixed spread or slippage distance applies to every NFP event.

How NFP fits into the Fed cycle

Interpret the report against the current policy setting and previous expectations. A jobs gain does not always delay a cut, and a weak reading does not require one. See how Fed decisions relate to the dollar for the broader conditional comparison.

The weekly report on new claims for unemployment benefits measures a different population and reference period. It is not a weekly NFP or CPS unemployment count. Compare the definitions before combining it with Employment Situation in fundamental analysis; frequency alone does not rank the size of a market response.

What to do before the next NFP

  1. Check the next release’s exact date and time. Verify the BLS schedule and your economic calendar. If you want a five-minute reminder, convert the announced time into your own zone for that date before setting it.
  2. Decide your event-exposure rule in advance. Your plan may specify observation, reduced exposure or no new orders during the release. Twelve months of experience is not a tested boundary, and liquidity or spreads are not guaranteed to return to normal at 15:00 or thirty minutes after publication. Check actual execution conditions before any later order.
  3. Record a reproducible observation. Save the forecast provider and capture time, first actual values, old and revised previous values, and when you received the contents. Choose price-observation windows in advance and record instrument, quote source, spread and actual fills. Separate overlapping events. A journal documents observations; it does not establish a profitable rule or prove causation.