Average hourly earnings or the Employment Cost Index: two ways to measure pay
Average hourly earnings and the Employment Cost Index both track pay, but they answer different questions, and the difference is mostly about who is in the average.

Two Bureau of Labor Statistics (BLS) measures often get called “wage growth,” yet they can point in different directions in the same period. Average hourly earnings come with the monthly jobs report. The Employment Cost Index (ECI) comes out quarterly. They are built to answer different questions, and knowing which question each one answers explains most of the gap between them.
This guide explains the two measures. It does not forecast pay and does not say what wage growth should be.
What average hourly earnings measure
Average hourly earnings come from the Current Employment Statistics (CES) survey, which BLS describes as producing industry estimates of nonfarm employment, hours and earnings of workers on payrolls. BLS reports the survey covers roughly 119,000 employers and about 622,000 worksites (BLS, CES, checked November 18, 2026).
At the most basic level, BLS says the all-employee figure is payroll divided by hours. Higher-level figures average the component cells weighted by aggregate hours (BLS Handbook of Methods, CES calculation, checked November 18, 2026). In plain terms, it is total pay divided by total hours worked, across whoever is on payrolls that month.
That is what makes it timely and also what makes it sensitive to who is working. BLS says its Real Earnings averages are not the earnings of “typical” jobs or workers, have no adjustments for occupational, age or schooling differences, and are influenced by fluctuations and varying trends in employment in high-wage versus low-wage industries, as well as by changes in wage rates (BLS, Real Earnings technical note, checked November 18, 2026).
What the Employment Cost Index measures
BLS says the ECI measures the change in the hourly labor cost to employers over time, and that it uses a fixed “basket” of labor to produce a pure cost change (BLS, ECI, checked November 18, 2026). The technical note describes it as free from the influence of employment shifts among occupations and industries. It uses fixed employment weights, which BLS updates periodically (BLS, ECI technical note, checked November 18, 2026).
The ECI covers both wages and salaries and the cost of benefits, so it also reports a total compensation figure. Wages and salaries include incentive pay such as commissions and production bonuses; benefits cover paid leave, supplemental pay such as overtime and shift differentials, insurance, retirement and savings, and legally required benefits such as Social Security and Medicare (BLS, ECI questions and answers, checked November 18, 2026).
For the third quarter of 2026, BLS reported that civilian compensation costs rose 3.4 percent over the year, wages and salaries rose 3.1 percent and benefit costs rose 3.9 percent. The next release, covering December 2026, is scheduled for January 29, 2027 (BLS, ECI, checked November 18, 2026).

Why mix matters: an invented example
The easiest way to see the difference is to invent a tiny labor market. These figures are made up and describe no real economy.
Imagine 100 low-wage jobs paying $20 an hour and 100 high-wage jobs paying $40, so the average is $30.00. A year later, every worker who keeps a job gets a 3 percent raise, so the pay rates are $20.60 and $41.20.
Case 1: low-wage jobs disappear. Suppose 20 of the low-wage jobs are lost. Total hourly pay is (80 × $20.60) + (100 × $41.20) = $1,648 + $4,120 = $5,768 across 180 workers, an average of $32.04. That is 6.8 percent higher than $30.00, even though no one’s rate rose more than 3 percent.
Case 2: low-wage jobs are added. Suppose instead 20 low-wage jobs are added. Total pay is (120 × $20.60) + (100 × $41.20) = $2,472 + $4,120 = $6,592 across 220 workers, an average of $29.96, a decline of about 0.1 percent despite 3 percent raises for everyone.
A fixed-weight measure holds the mix constant, so in both cases it would show the 3 percent change in the price of labor. Neither number is wrong. They answer different questions: what the average paycheck-hour looked like, and what the same work cost.
What each measure can and cannot tell you
Average hourly earnings are useful as a quick read on the average pay of whoever is employed. They can move because raises were larger, because the job mix shifted, or both. The monthly jobs report is covered in the guide on how to read the monthly jobs report.
The ECI is better suited to the question “is the cost of an hour of the same labor rising faster or slower?” It is quarterly, so it is slower, and it includes benefits, which average hourly earnings do not. BLS says its ECI estimates are final when published, except that seasonally adjusted three-month series are revised each March for the previous five years (BLS, ECI questions and answers, checked November 18, 2026). Monthly CES figures, by contrast, are revised after first publication, and the latest months in the jobs report are labeled preliminary.
Neither one is your own raise. A person’s pay change depends on their employer, job and hours, and the national measures are averages. To test whether a raise kept pace with prices, see is a raise keeping up with prices?.
How to compare two wage figures sensibly
Check four things before comparing two wage-growth numbers. First, which measure it is, and what population it covers. Second, whether it includes benefits. Third, the time span: monthly changes, three-month changes and twelve-month changes are different things. Fourth, whether the figure is adjusted for prices, which turns nominal pay into real pay, a distinction covered in nominal versus real.
If two sources report different wage growth, a gap is not automatically a contradiction. It is often the mix effect shown above, a different time span or the inclusion of benefits.
Common questions
Which measure is better?
Neither is better in general. Average hourly earnings are timelier and show what employed people were paid on average. The ECI isolates the cost of labor from shifts in who works where. The right one depends on the question.
Why did average hourly earnings jump when the economy was weak?
One possibility, shown in the invented example, is that lower-paid jobs were lost, which lifts the average without any raises. That is a mechanism, not a claim about any particular month, and the BLS release itself is the place to check.
Does the ECI include benefits?
Yes. BLS reports separate wage and salary and benefit series, plus total compensation.
Sources checked November 18, 2026: BLS, Employment Cost Index; BLS, ECI technical note; BLS, ECI questions and answers; BLS, Current Employment Statistics; BLS Handbook of Methods, CES calculation; BLS, Real Earnings technical note.