How to Trade NFP (Non-Farm Payrolls): The Macro Trader’s Guide

Most retail traders read one number on NFP day: the headline jobs print. They take a position based on whether it beat or missed, and an hour later they cannot understand why the dollar reversed. The institutional desks read five numbers, in a specific order, and the headline is the least important of them. Wages move the Fed. Revisions move the trend. Participation moves the curve. The headline moves only the algorithms, and only for ninety seconds.

By Ken Chigbo · Founder, KenMacro · 18+ years in markets, London trading floor and institutional FX

This guide is reviewed and refreshed periodically to reflect the current cycle. The framework itself is timeless.

In one sentence: NFP is not a jobs print, it is a labour-market complex of five numbers, and the durable trade lives in average hourly earnings and prior-month revisions, not in the headline.

What NFP Actually Is

Non-Farm Payrolls is a monthly report from the Bureau of Labor Statistics measuring the change in employment across the US economy excluding farm workers, government employees in some categories, private household employees, and non-profit employees. The "non-farm" qualifier reflects the volatility of agricultural employment, which the BLS strips out to produce a cleaner read on the rest of the labour market.

The report is built from two surveys conducted in parallel. The Establishment Survey covers roughly 122,000 businesses and government agencies; it produces the headline payrolls number. The Household Survey samples about 60,000 households; it produces the unemployment rate, participation rate, and labour-force flows. The two surveys often disagree because they measure different populations through different methodologies, that disagreement is itself a signal.

NFP is released on the first Friday of every month at 08:30 ET, with the data referring to the previous calendar month. The report is published online by the Bureau of Labor Statistics at bls.gov, distributed to financial newswires under embargo, and released to markets simultaneously. Algorithmic trading systems read the JSON release and fire orders within milliseconds of the print.

What makes NFP exceptional is its position in the data calendar. It is the first major US data print of every month. CPI follows roughly a week later. Retail sales, PPI, and industrial production print throughout the month. NFP sets the tone, every subsequent release is read against the labour-market read NFP just produced. This is why NFP day is consistently one of the highest-volatility days in any given month for DXY, gold, the Treasury curve, and equity index futures.

The Five Numbers in the NFP Report

The institutional read decomposes NFP into five numbers, each with its own market sensitivity. Reading them in order is the difference between trading the headline and trading the actual labour market.

1. Headline non-farm payrolls. The change in employment from the prior month, in thousands. Consensus expectations are published before the release. The headline is what algorithms react to in the first 90 seconds.

2. Average hourly earnings (AHE). Both month-over-month and year-over-year. This is the wage-growth measure the Fed cares about most. A hot AHE print signals continuing wage pressure and forces the Fed to delay cuts; a soft AHE print signals labour-market cooling and supports cuts.

3. Unemployment rate. From the Household Survey. Includes only people actively looking for work. Can fall for the right reasons (more hiring) or the wrong reasons (people leaving the labour force).

4. Participation rate. The percentage of the working-age population in the labour force. Read alongside the unemployment rate to determine whether changes are real or compositional.

5. Revisions. The BLS revises the prior two months' headline figures with each new release. Revisions are routine and often substantial, a single revision can be larger than the new month's print. Markets sometimes ignore revisions in the first reaction; sophisticated desks read them first.

Two more readings often matter, the average workweek (a measure of labour-market tightness; declining hours signal weakening demand) and the U-6 unemployment rate (a broader measure including discouraged workers and part-time workers wanting full-time work). U-6 is the cleanest read of underutilisation; if it is rising while U-3 is flat, labour-market slack is increasing despite the headline.

The Headline: Why It Does Not Move Durable Trends

The headline non-farm payrolls number is the most-watched, least-durable signal in the NFP report. Algorithms react to it within ninety seconds. The dollar moves, gold moves, the Treasury curve moves. And then, by 09:00 ET, half of those moves reverse.

That reversal happens because the headline is noisy. Monthly payrolls bounce around a long-run trend of roughly 150,000 to 200,000 jobs per month in the current cycle. Standard deviation on the print is large. A 50,000 beat or miss against consensus is statistically unremarkable; a 100,000 beat or miss can be a one-sigma event that mean-reverts in subsequent months.

Institutional desks know this. They use the first ninety seconds of algorithmic reaction to pre-position into wages and revisions, which print at the same time but require human reading to interpret. By 09:00 ET, the wages number has been digested. By 09:30, revisions have been factored in. By the time the New York equity open arrives at 09:30 ET, the durable direction for the day has been set, and it is rarely the direction the headline implied.

This is why an apparently "hot" NFP, say 250,000 jobs against a 180,000 consensus, can produce a dollar rally that fades by lunch if wages came in soft and revisions were large negatives. The headline reaction is reversed by the deeper read.

The trader who fires on the headline gets the algorithmic move and stops out on the reversal. The trader who waits ninety seconds for wages, two minutes for revisions, and trades the resolved direction at 09:00 ET captures the durable trend.

Average Hourly Earnings: The Real Driver

Average hourly earnings is the wage-growth measure the Fed watches most closely. It is reported as a percentage change month-over-month and year-over-year. The year-over-year number is the more important of the two because it strips out monthly noise; a single hot month can be statistical, but a sustained year-over-year trend is signal.

The mechanism through which AHE moves markets is direct. The Fed's mandate is price stability; service-sector inflation excluding shelter, sometimes called supercore, comes from labour costs because services are mostly delivered by people. If wages are growing at 4 to 5% year-over-year, services inflation cannot sustainably return to 2%. The Fed therefore cannot cut while AHE is hot.

Historically, the Fed has tolerated AHE growth around 3.0, 3.5% as compatible with 2% inflation given normal productivity growth. AHE above 4% has typically forced rate hikes; AHE below 3% has typically allowed cuts. The current cycle's threshold may differ, productivity has been volatile post-pandemic, but the directional rule holds: hot wages delay cuts; soft wages enable cuts.

The market reaction is the same in both directions. Hot AHE: yields rise across the curve (most at the front), DXY rises, gold falls, long-duration tech sells off. Soft AHE: yields fall, DXY softens, gold rallies, long-duration tech rips. The reaction is roughly proportional to the surprise versus consensus, but with an asymmetry, soft prints often produce larger reactions than hot prints because the market expects wage stickiness, so a downside surprise is more informative.

Watch for the composition. AHE growth driven by low-wage workers seeing larger raises (catch-up wage growth) is less alarming than AHE growth driven by mid-wage workers, the latter signals broader inflation pressure. Bureau of Labor Statistics releases sectoral AHE detail; it is not in the headline release but worth reading on hot prints to confirm whether the growth is broad-based.

Unemployment Versus Participation

The unemployment rate looks simple, the percentage of people actively looking for work who do not have a job. It is misleading in isolation. The participation rate is the second variable that determines whether unemployment is moving for healthy or unhealthy reasons.

Consider two scenarios. Scenario A: unemployment falls from 4.0% to 3.8%, participation rises from 62.5% to 62.8%. People are entering the labour force and finding jobs; the labour market is genuinely tightening. Bullish for growth, hawkish for the Fed, bullish for the dollar. Scenario B: unemployment falls from 4.0% to 3.8%, participation falls from 62.5% to 62.3%. People are leaving the labour force, which mechanically reduces unemployment. The labour market is weakening; the headline read is the opposite of the truth. Bearish for growth, dovish for the Fed, bearish for the dollar, even though the headline unemployment rate fell.

Institutional desks therefore always pair the two numbers. A falling unemployment rate with rising participation is durable; a falling rate with falling participation is a statistical artefact that subsequent months will reverse.

Powell reads them this way explicitly. In dovish cycles, he has repeatedly emphasised the participation rate as evidence of a "warm but not overheating" labour market. In hawkish cycles, he has highlighted the unemployment-participation mix to signal that the labour market is tighter than the headline suggests. Listening for the language at FOMC reveals which read the committee is using.

Equally, the same logic applies to the U-6 unemployment rate. If U-3 (the headline) is flat but U-6 (the broader measure) is rising, slack is increasing despite the headline. The reverse, U-3 falling while U-6 also falls, is the cleanest tightening signal.

Revisions: The Quiet Signal

Every NFP release revises the prior two months' headline figures based on additional sample data the BLS receives after initial publication. Revisions are routine. They are often large. And they are routinely under-traded by retail because they print simultaneously with the new month's headline and are easy to miss.

Crucially, a 50,000 downward revision to last month combined with a 50,000 beat on the new month is a wash on a six-month average, the labour market is exactly where the previous read suggested. A 75,000 downward revision combined with a 25,000 miss on the new month is a 100,000-job downward swing in two months, the labour market is meaningfully weaker than the previous read suggested, and the Fed's read of the cycle has just changed.

Sustained patterns of negative revisions are particularly important. The BLS sometimes revises down for several months in a row late in an economic cycle as administrative records catch up to slowing actual hiring. When you see two or three consecutive months of meaningful downward revisions, the cycle is rolling over, even if the headlines are still positive.

The reverse pattern is rarer but happens during early-cycle rebounds. Sustained positive revisions late in a slowdown signal that hiring is reaccelerating before the headlines pick it up. This is when the early dollar bid and steepener trades work.

Track the rolling three-month average of payrolls instead of the headline. The three-month average smooths revisions and noise; a meaningful change in the three-month average is a meaningful change in the labour market, regardless of any single month's print.

Why NFP Matters to the Fed

The Fed has a dual mandate set by Congress and codified in the Federal Reserve Act: price stability and maximum employment. NFP feeds both reads, the unemployment and participation data inform the employment side; AHE informs the wage-cost component of the price-stability side.

In a cycle where inflation is the active concern, the Fed reads NFP through the wage lens. Hot AHE means delayed cuts. Cool AHE means cuts on the table. The unemployment rate is secondary because mild labour-market loosening is exactly what the Fed wants, the cooling that precedes lower wage growth. Powell has said as much directly in multiple press conferences.

In a cycle where growth is the active concern, the Fed reads NFP through the unemployment lens. Rising unemployment, falling participation, declining hours, these are the recession-leading signals. AHE becomes secondary because labour-market weakness will eventually pull wages lower regardless. The Fed shifts from delaying cuts to delivering them, even before inflation is fully extinguished.

The transition between regimes is the point of maximum NFP-reading complexity. A single hot AHE print combined with a soft unemployment print is contradictory; the market often spends the morning chopping while desks figure out which signal the Fed is prioritising. Powell's most recent press conference is the best guide; if he emphasised wages, AHE leads. If he emphasised employment, unemployment leads.

For the full FOMC framework that integrates NFP into the rate-path read, see how to trade FOMC.

NFP Transmission Across Forex, Gold, and Bonds

Hot NFP, strong headline, hot AHE, low unemployment, produces a clean transmission chain. Treasury yields rise (most at the 2-year, less at the 10-year, curve flattening). DXY rises across G10. Gold falls on real yields. Long-duration tech sells off. Cyclical sectors (financials, industrials) outperform tech. Emerging markets weaken on dollar strength.

Soft NFP, weak headline, soft AHE, rising unemployment, produces the opposite. Yields fall, curve steepens. DXY softens. Gold rallies. Long-duration tech rallies. Financials lag.

Mixed NFP, strong headline but soft wages, or weak headline but hot wages, produces choppy reactions. The day's direction is usually set by the wage component, not the headline. Pre-position into the wages reaction, not the headline.

The DXY trade is the cleanest expression. NFP is one of the highest single-day realised volatility events in DXY's calendar, comparable in magnitude to FOMC days. Average DXY range on NFP day is 80 to 120 basis points; data-surprise days can extend to 200 basis points. Sizing must respect the realised volatility.

For the full DXY framework around macro releases, see how to trade DXY.

Gold's reaction is more nuanced because it depends on whether the move is real-yield driven or risk-sentiment driven. Hot NFP that lifts both nominal yields and breakevens leaves real yields flat, gold can hold ground. Hot NFP that lifts nominals while breakevens fall (the market reading the strength as the Fed staying restrictive) drops real yields hard, and gold cracks. The breakeven move is the tell.

How to Trade NFP: The KenMacro Framework

The framework is built around the order in which the report should be read and the order in which trades should be sized. Discipline is the differentiator.

Step 1, pre-print preparation (24 hours before). Pull the consensus on headline, AHE, unemployment, and prior-month revisions. Pull the OIS curve to identify what the market has priced for the next FOMC. Pre-write three scenarios, hot, in-line, soft, and the expected reaction in DXY, gold, the 2-year, and tech for each. Identify your two primary instruments for the trade.

Step 2, pre-print positioning (one hour before). Reduce or flatten existing positions in instruments correlated to NFP. The opportunity cost of carrying directional risk into the print is rarely worth it. Set alerts at the levels you have pre-defined for each scenario.

Step 3, the print (08:30 ET). Read the headline. Read AHE month-over-month and year-over-year. Read unemployment. Read participation. Read revisions. Form a one-sentence verdict in your head: "hot wages, soft headline, weak revisions" or "in-line headline, hot wages, healthy participation". The verdict tells you the durable direction.

Step 4, wait ninety seconds. The algorithmic reaction to the headline will overshoot in the wrong direction roughly half the time. Patience pays. The 09:00 ET pause is when the durable direction sets.

Step 5, trade the verdict. Enter your two primary instruments at the pre-defined levels. Halve normal size. Set wider stops to respect realised volatility. The trade should hold through the New York equity open and into the London close.

Step 6, post-trade review. Read the BLS sectoral detail later in the day. Note which sectors led the headline change. Note whether revisions matched your read. Refine the framework for next month.

NFP Trade Example: How the Sequence Plays Out

Consider a hypothetical NFP Friday. Consensus: headline 180,000, AHE 0.3% MoM / 4.0% YoY, unemployment 4.0%, prior month revised flat. Pre-print, DXY at 105.50, gold at 2,400, US 2-year at 4.50%.

08:30 ET, release. Headline 165,000 (slight miss). AHE 0.5% MoM, 4.3% YoY (hot). Unemployment 3.9% (small drop). Participation up 0.1pt. Prior month revised down 25,000.

First 90 seconds, algorithms read the headline miss and sell dollars. DXY drops 25 basis points. Gold rallies $8. The 2-year yield drops 4bp. The first move reflects only the headline.

09:00 ET, desks digest the wages number. AHE printed substantially above consensus, with year-over-year accelerating from 4.0% to 4.3%. The 2-year reverses, yield rises 12bp from the open. DXY reverses, bid extends through 105.65, through 105.80. Gold reverses, drops back to flat, then $5 below open. The durable direction has set.

09:30 ET, equity open. Tech opens softer; financials open firm. The 2-year continues to bid. DXY pushes 105.95.

10:00 ET, revisions get a second look. Prior month revised down 25,000; combined with the 165,000 print, the two-month run-rate has slipped to 145,000. The directional dollar bid stays intact through the lunch session.

13:00 ET, late-morning chop, but DXY holds 105.85. Gold holds $2,392. The 2-year holds 4.62%.

16:00 ET, New York close. DXY closes 106.00 (+50 basis points on the day). Gold closes 2,388 (-$12). 2-year closes 4.63% (+13bp).

The trade was visible by 09:00 ET in the wages number, despite the headline miss. The trader who fired short-DXY at 08:30:01 stopped out by 09:15. The trader who waited ninety seconds, read AHE, and went long DXY at 09:00 ET captured 35 basis points by the close.

Common Mistakes Traders Make on NFP Day

Trading the headline

The headline reverses too often. Wait for AHE.

Ignoring revisions

A 50,000 downward revision combined with a 50,000 beat is a wash. Read the rolling three-month average; one number is noise.

Reading unemployment without participation

Also, a falling unemployment rate driven by falling participation is bearish, not bullish. The combination is the signal.

Holding through the print

Existing directional positions in DXY, gold, equities, or yield-sensitive instruments will be hit by NFP regardless of whether the trade thesis depends on it. Reduce or flatten.

Sizing for normal volatility

NFP day realised volatility is 1.5, 2x average. Stops blown on NFP are usually a sizing problem, not a directional one.

Confusing AHE direction with rate-cut direction

Hot AHE delays cuts (hawkish); soft AHE enables cuts (dovish). It is not "hot economy = cut" or "weak jobs = cut". The Fed is reading wages, not jobs.

Trading every NFP

Some prints are too in-line to be tradeable. The framework should produce a "no trade" verdict when all five numbers come in close to consensus. Forcing a trade on a non-event print is how P&L gets wasted between the few real prints that matter.

Ignoring CPI's lead-in. NFP is the first major data print of the month; CPI follows roughly a week later. The market often pre-positions for CPI based on what NFP signalled. Reading the two together produces better trades than reading either alone. For the CPI framework, see how to trade CPI.

Final Takeaway: NFP Is a Wage Print, Not a Jobs Print

The headline is for algorithms. The wages are for the Fed. The revisions are for the trend. The participation is for context. Read all five numbers in sequence; trade the verdict, not the headline.

NFP is one of the highest-volatility events in any monthly calendar. The discipline of waiting ninety seconds for the wages number, reading revisions before sizing, and respecting realised volatility on stops is what separates traders who use NFP as a controlled income event from traders who use it as a coin flip.

Frequently Asked Questions: How to Trade NFP

Sources: Bureau of Labor Statistics Employment Situation Report methodology; observed institutional desk practice across multiple cycles. Scenarios are analytical frames, not forecasts.