What the Jobs Report Measures, and Why Its Two Surveys Disagree

The monthly US employment report is the most closely watched economic release in the world, and it is really two reports published together. They use different methods, measure different populations, and regularly disagree.

Coverage tends to pick whichever survey supports the cleaner narrative. Reading both is not much harder and is considerably more informative.

The establishment survey

The first survey asks businesses how many people are on their payrolls. It produces the headline payrolls figure and the average hourly earnings data.

It samples a large number of employers, which makes it statistically precise on the level of employment. It counts jobs, not people — someone holding two jobs is counted twice. It excludes the self-employed and agricultural workers, which is a growing gap as independent work expands.

The household survey

The second survey asks households about their own employment. It produces the unemployment rate, the participation rate, and the count of people who are employed.

It counts people rather than jobs, and includes the self-employed. Its sample is much smaller, so it is noisier month to month — which matters, because it is the source of the unemployment rate that generates most headlines.

Why they diverge

The surveys can point in opposite directions for entirely mundane reasons: growth concentrated in self-employment shows up in one and not the other; people taking second jobs inflate payrolls without changing the number employed; sampling error in the smaller survey produces month-to-month swings that mean nothing.

Sustained divergence over several months is worth attention. A single month of it usually is not.

What the unemployment rate excludes

The unemployment rate counts people without work who are actively looking. Someone who stops looking leaves the labour force entirely and stops being counted as unemployed.

This produces a well-known perversity: the rate can fall because people gave up. It can rise because conditions improved enough that discouraged workers resumed searching. The rate alone does not distinguish these, which is why the participation rate — the share of the working-age population in the labour force — should be read alongside it.

Revisions are the norm

Initial payrolls estimates are revised twice as more responses arrive, and the whole series is periodically benchmarked against tax records. Revisions are routine and sometimes large enough to change the interpretation of a month entirely.

This has a practical implication for anyone reading the release: the first print is a provisional estimate with meaningful error, and confident conclusions drawn from a single unrevised month are built on sand. Markets react to it anyway, which is a fact about markets rather than about the data.

The figures worth reading

Beyond the headline, a few series carry more signal than they get credit for.

Seasonal adjustment

Employment follows strong seasonal patterns — retail hiring before the holidays, construction slowing in winter, teaching staff leaving payrolls in summer. Almost every figure reported is seasonally adjusted to strip these out.

The adjustment is estimated from historical patterns, which works well when patterns are stable and less well when behaviour shifts. A retail sector that hires earlier than it used to will show a distorted adjusted figure until the model catches up.

For most months this is a technicality. After a period of unusual disruption it becomes a live source of error, and it is worth knowing that an unexpected number can reflect the adjustment rather than the underlying hiring.

Wages, and the productivity qualifier

Average hourly earnings is watched as an inflation signal, on the logic that rising labour costs feed into prices. The reasoning is sound but incomplete.

The measure is affected by composition. If job losses fall disproportionately on lower-paid roles, average wages rise without anyone receiving an increase. The reverse happens when hiring is concentrated in lower-paid sectors.

More importantly, wage growth is only inflationary to the extent it exceeds productivity growth. Workers producing more per hour can be paid more without raising unit costs. Commentary treating any wage growth as an inflation risk skips this, and in doing so treats a rise in living standards as a problem to be managed.

The other labour market releases

The monthly report is one input among several, and the others fill gaps it leaves.

The quits rate deserves particular attention. People leave jobs voluntarily when they are confident of finding another, so it functions as a measure of worker bargaining power, and it tends to turn before the headline figures do.

Why it moves markets

Employment data drives expectations about the Federal Reserve, and Fed expectations drive the price of money, which prices everything else. A strong report implies rates staying higher for longer; a weak one implies cuts.

That is the transmission channel, and it explains why bond markets sometimes fall on good economic news. They are not pricing the economy. They are pricing the policy response to it.