Economy · Analysis
How to read a jobs report without being misled
Employment data arrives as several surveys measuring different things. The headline number is the least stable of them.
This is analysis. It contains the interpretation of the Wallcrest Economics Desk.
The short answer
- Household and establishment surveys can disagree for months at a time.
- Revisions are routine and often larger than the market-moving surprise.
- Participation and hours worked describe slack better than the unemployment rate alone.
A monthly jobs release is not one measurement. It combines a survey of employers, which counts payroll positions, with a survey of households, which counts employed people. A person with two jobs appears twice in one and once in the other; a self-employed contractor may not appear in the payroll count at all.
Revisions are the norm
Initial estimates are based on partial responses and are updated as more arrive, then benchmarked against administrative records. A first print that surprises by a modest amount can be revised away entirely, which is why single-month reactions are frequently reversed.
Where the slack actually shows
- Participation rate: people who left the labour force do not count as unemployed.
- Average hours: employers cut hours before they cut heads.
- Underemployment measures: part-time work for economic reasons.
- Quits and vacancies: how confident workers and firms feel about switching.
Wages are the inflation link
For policy, the question is whether pay growth is consistent with the inflation target once productivity is accounted for. Composition effects complicate this: if low-paid jobs disappear, average wages rise without anyone getting a raise.
Sources
- Employment situation technical note — U.S. Bureau of Labor Statistics
- Labour market statistics — Eurostat
Spotted an error? Tell our corrections desk.