The Foundry · Complete Guides
The Complete Guide to Reading the Jobs Report
Titan Macro Desk • 4 July 2026 • 8 minute read
Two days ago, US payrolls printed 57,000 against a forecast of 114,000. The market rallied 566 points in two hours, then gave every point back before the close. Both moves were driven by the same number. If that seems impossible, this guide is for you, because the jobs report is not one number and it never was. It is a bundle of measurements that regularly contradict each other, and the traders who get hurt by it are almost always the ones reacting to the headline while the market trades the internals.
What the Report Actually Is
On the first Friday of most months (or Thursday when a holiday intervenes, as it did this week), the US Bureau of Labor Statistics publishes the Employment Situation report. It is really two surveys stapled together, and the staple is where most of the confusion lives.
The establishment survey asks roughly 120,000 businesses how many people were on their payroll. This produces the headline: non-farm payrolls, the famous NFP number. It measures jobs, not people. One person holding two jobs counts twice.
The household survey asks about 60,000 households whether their members are working or looking for work. This produces the unemployment rate. It measures people, not jobs, and it uses a completely different definition of employed.
Because they are different surveys of different things with different sample sizes, they can and do disagree. This week was a textbook case: payrolls halved against forecast while the unemployment rate fell from 4.3% to 4.2%. Weak hiring and falling unemployment in the same release. Neither number is wrong. They are answering different questions, and the tension between their answers is often the single most informative thing in the report.
The Six Numbers That Matter, In Order
| Number | What it tells you | Common trap |
|---|---|---|
| Headline NFP | Net jobs added versus forecast | Trading it before reading anything else |
| Revisions | Whether the last two months were better or worse than first reported | Ignoring a big revision that changes the trend entirely |
| Unemployment rate | Share of the labour force without work | Forgetting it falls when discouraged people stop looking |
| Average hourly earnings | Wage growth, the inflation channel | Reading hot wages as good news for stocks in an inflation fight |
| Participation rate | Share of adults in the labour force at all | Missing that it explains “impossible” unemployment moves |
| Diffusion index | How many industries are adding jobs versus cutting | Missing that one sector (often government or healthcare) is carrying the whole print |
The order matters. A weak headline with positive revisions and broad diffusion is a very different animal from a weak headline with negative revisions and three sectors doing all the hiring. The first is noise. The second is a trend change announcing itself.
Why the Same Number Can Be Bullish and Bearish
Markets do not price the economy. They price the path of interest rates, corporate earnings, and the balance between them. A jobs number feeds both, in opposite directions, and which channel dominates depends entirely on what the market fears most that month.
The rates channel: weak jobs means a cooling economy, which means the central bank can cut rates sooner. Lower rates raise the present value of future earnings, which is mechanically bullish for equities, especially long-duration growth stocks. This is the famous “bad news is good news” regime.
The earnings channel: weak jobs also means fewer employed consumers, softer demand, and eventually weaker corporate revenue. If hiring is falling because businesses see trouble coming, today’s rate-cut celebration is borrowing against tomorrow’s earnings recession. This is “bad news is just bad news”.
Thursday’s session ran both channels in sequence, in plain sight. Act one: 57K triggered rate-cut repricing and a 566 point rally in NAS100. Act two: the market did the earnings arithmetic, asked who buys the products if nobody is hiring, and unwound the entire move into the close. The number did not change between 14:30 and 20:00. The dominant channel did. Watching which act wins the close, rather than which wins the first hour, tells you which regime you are actually in.
The Cross-Asset Confirmation Check
Equities are the least honest first responder to a jobs number, because they are fighting the two-channel battle above. The cleaner reads come from the assets with only one channel:
- Short-dated Treasury yields reprice rate expectations directly. If two-year yields plunge on a weak print and stay down, the rate-cut interpretation has bond market backing. If they bounce back within the session, the bond market is calling the equity rally a tourist trade.
- Gold benefits from lower rates and from fear. When both channels of a weak jobs number point its way, it does not need to pick an act. It won Thursday outright, up 1.78% while equities round-tripped, precisely because it was the only major asset with no internal contradiction to resolve.
- The dollar weakens when cut odds rise. A weak-jobs day where the dollar barely moves is a market that does not believe the cuts are coming.
- Consumer discretionary versus staples is the earnings channel made visible. If discretionary underperforms staples after a weak print, the market is trading recession, whatever the index level says.
The discipline is simple: never grade a jobs reaction from the equity tape alone. Three confirming assets make a regime. One index in the first hour makes a trap.
Reading This Week’s Report With the Framework
Apply the checklist to Thursday’s release and the confusion resolves into structure. Headline: 57K versus 114K, a genuine shock, half of consensus. Context: the second weak employment signal in two days, after ADP missed at 98K, which is what upgrades a miss into a trend. The household survey contradiction: unemployment fell to 4.2%, which softens the recession case and explains why the sell-off, when it came, was orderly rather than panicked. Cross-asset: gold confirmed, bitcoin decoupled upward, crude kept falling on its own demand worry, and volatility was crushed 10% into a three day weekend, which is the one reading that should make a careful trader uncomfortable, because it prices certainty nobody has yet earned.
That is why the week’s closing posture was reduced exposure into Monday’s reopen. Not because the framework knows which act wins, but because the report left the argument genuinely unresolved, the market is priced as if it is resolved, and the gap between those two states is where weekend risk lives. The next instalments arrive fast: the central bank’s June meeting minutes on Wednesday, then weekly jobless claims on Thursday, which after a 57K print stop being routine and become the referee.
The Checklist
1. Headline versus forecast, and versus the three month average, not just last month.
2. Revisions first, always. They change trends silently.
3. Does the household survey agree? If not, which story does participation support?
4. Wages: hot wages plus weak hiring is stagflation risk, the worst quadrant.
5. Diffusion: is hiring broad or is one sector carrying the print?
6. First hour reaction versus closing reaction. The close is the vote that counts.
7. Confirm across two-year yields, gold, the dollar, and discretionary versus staples.
8. Ask what is now priced. The number matters less than the gap between the number and the positioning that met it.
The jobs report rewards the reader who treats it as a document and punishes the trader who treats it as a signal. It takes eight minutes to work through the internals properly. Thursday’s two-act session was worth 566 NAS100 points in each direction inside four hours. Few eight minute investments in markets pay better.
Titan Macro Desk. This is analysis and education, not financial advice. Markets carry risk. Always manage your position size and do your own research.