Why the Employment Cost Index Is Not Average Hourly Pay
Learn how a fixed employment mix and benefit costs change what a labor-cost measure tells you.
Scenario
Average hourly earnings rise faster than the Employment Cost Index, and an automated summary says the two reports contradict each other. Before deciding that one is wrong, check whether they measure the same costs and use the same employment mix.
Evidence
BLS describes the ECI as a National Compensation Survey measure of changes in employer labor costs, designed to separate those changes from shifts among industry and occupation categories. Its total-compensation index includes wages and salaries plus employer benefit costs. Separate wage-and-salary and benefit indexes are also published.
The ECI uses employment weights held constant between periodic reweightings. By contrast, establishment-survey average hourly earnings divide aggregate payroll by aggregate paid hours. That average can change as the mix of higher- and lower-paid work changes, as well as when pay changes within establishments. It excludes employer benefit costs and is not a measure of total compensation.
Imagine two job groups paid $20 and $40 per hour, with unchanged pay rates. If more paid hours shift toward the $40 group, an economy-wide hourly average can rise without either group receiving a raise. Holding the job-category mix fixed helps separate this composition effect from changes in labor costs.
Why it matters
ATLAS4 analysis: use the ECI to investigate compensation-cost changes on its stated basis, and average hourly earnings to investigate payroll earnings per paid hour. Neither series means that every employee received its reported percentage increase. Compare compatible sectors and periods; do not equate a quarterly change with an annual change.
Definitions can also change. As checked October 25, 2026, BLS says the January 2027 release of December 2026 ECI data will introduce 2025 employment weights and remove workers’ compensation costs. That is an announced future change, not a change to apply early to the forthcoming September data.
Takeaway
A useful research record stores the series name, wage-versus-total-compensation coverage, sector, reference period, adjustment basis and methodology version. An AI-generated comparison should preserve those fields and explain differences in scope before calling two labor indicators inconsistent. These are recommended research practices, not claims about an already deployed ATLAS4 product feature.
Primary sources
BLS — ECI calculation and fixed weights
BLS — average hourly earnings concepts
BLS — announced January 2027 ECI changes
ATLAS4 Systems research is provided for informational and research purposes only and is not investment advice or an offer or solicitation to buy or sell securities or digital assets. INTERNAL_NEWS_INGESTION=UNVERIFIED; no ATLAS4 internal feed or asset-reaction data were used.