Wage growth: average hourly earnings, real wages and composition

Wage growth is how fast pay is rising. The most watched measure is average hourly earnings, published every month in the The jobs report: how to read nonfarm payrolls and the two surveys for all private-sector employees: $37.81 an hour in September 2026, 3.0% higher than a year earlier. Investors watch it because pay is both most of household income and a large share of business costs, so it sits at the junction of spending, profits and Inflation: what it is, how it’s measured and what it does to money.

A wage number on its own says little. What matters for living standards is pay after inflation, the real wage, and what matters for reading the data is whose pay is in the average. In April 2020 average hourly earnings jumped 8.1% from a year earlier without anyone getting a raise, because millions of low-paid jobs disappeared from the average.

This guide explains how the series is built, how to convert it into real wage growth, the composition effect that can distort it, and how it compares with the quarterly Employment Cost Index.

What average hourly earnings measures

Average hourly earnings come from the establishment survey (Current Employment Statistics), the same survey that produces payrolls. Employers report total payroll and total hours paid for their workers in the pay period that includes the 12th of the month. BLS divides total earnings by total hours. The headline covers all employees on private nonfarm payrolls; government workers are excluded. BLS also publishes a separate series for production and nonsupervisory employees.

Three features follow from that construction:

• It is gross pay before taxes and deductions, including overtime but not benefits such as health insurance or employer retirement contributions.
• It is an average of whoever is on payrolls that month. If the mix of workers shifts toward higher-paid jobs, the average rises even if no one's pay changes.
• It is nominal: dollars, not purchasing power.

Because it comes from a large sample, the series is fairly smooth month to month, but it is revised along with payrolls in the two following reports.

Real wage growth

To see whether pay is outrunning prices, deflate it by a price index, usually the Consumer Price Index (CPI).

Over longer periods compounding matters. From February 2020 ($28.54) to September 2026 ($37.81), average hourly earnings rose 32.5%, while CPI rose 29.7% (259.25 to 336.298). Real hourly pay over the whole stretch was up 1.325 ÷ 1.297 − 1 = +2.2% in six and a half years, despite large swings in between.

Wages against prices since 2008

The chart shows four distinct periods, using the year-over-year changes in the data.

• 2008–2019: slow wages, low inflation. Wage growth ranged from 1.6% to 3.7% and was between 2% and 3.5% in three months out of four, and real wage growth swung around zero: negative in 2008 and in 2011–2012 when CPI inflation picked up, positive in most later years. In December 2019 wages were up 3.0% against CPI of 2.3%.
• 2020: the composition spike. Wage growth jumped to 8.1% in April 2020 while CPI was up 0.3%. Twelve months later the base effect ran the other way: April 2021 wage growth was just 0.6%.
• 2021–2023: inflation outruns pay. Real wage growth was negative every month from April 2021 to April 2023. The low came in June 2022: wages +5.4%, CPI +9.0%, real −3.6%.
• 2023–2026: recovery, then a squeeze. Real wage growth turned positive in May 2023 and reached about +1.2% in December 2024 (wages +4.1%, CPI +2.9%). Wage growth then cooled steadily, to 3.0% by September 2026, while CPI inflation rose to 3.7%, putting real wages back below zero every month from April through September 2026.

The composition effect: April 2020

Average hourly earnings jumped from $28.73 in March 2020 to $30.04 in April 2020 in the current data, a 4.6% rise in one month. The first published figures, in the May 8, 2020 release, were $28.67 and $30.01, a rise of $1.34. BLS explained in that release that the increases largely reflected the substantial job loss among lower-paid workers. Over those two months payrolls fell by about 21.9 million, and the jobs lost were concentrated in lower-wage work.

The reverse happens in a recovery: as lower-paid workers are rehired, the average is pulled down, which is part of why wage growth fell to 0.6% a year after the spike. Any period with large, uneven job losses, such as a recession hitting low-wage industries hardest, can bias the average upward in the same way.

The Employment Cost Index

BLS publishes a second, slower measure designed to avoid that problem. The Employment Cost Index (ECI) is quarterly and measures the change in employers' hourly labor costs using a fixed basket of jobs, so it is free of the effects of workers moving between occupations and industries. It covers wages and salaries and, in a separate measure, the cost of benefits. The September 2026 quarter was published in October 2026. The ECI is not charted on this page, but when the two series disagree around a recession or a large shift in hiring, the ECI is the cleaner read on what employers are actually paying for a given job.

How to read wage data

• Use the 12-month change for trend, the monthly change with care. A 0.2% or 0.4% monthly print is the difference of a few cents; annualizing it multiplies the noise.
• Check hours and payrolls in the same report. If total hours or low-wage industries fell sharply, expect a composition boost to the average.
• Compare with inflation, and with productivity. Wage growth above inflation plus productivity growth raises labor cost per unit of output, which can feed into prices; wage growth near that sum is broadly consistent with steady inflation.
• Why markets care. Faster wage growth can mean stronger consumer spending and also stickier service-sector inflation, which matters for how long the Fed holds rates high. Weaker wage growth can signal a cooling labor market. As with payrolls, bond yields usually react more than stocks.

Pitfalls and limitations

• Composition. The average changes with the job mix, as April 2020 showed.
• No benefits. Average hourly earnings misses health insurance and retirement contributions, a meaningful share of total compensation.
• Private sector only. Government workers are excluded from the headline.
• Calendar quirks. Pay periods and the number of weeks in a month can shift the monthly change by a few cents.
• Deflator choice. Using CPI or the PCE price index gives different real wage paths; see PCE inflation: the Fed’s preferred gauge vs the CPI for why the two indexes differ.
• Data gaps. The missing October 2025 CPI means any real-wage calculation spanning that month must match dates, not row counts.

Wage measures compared

How Finabase does it

Finabase does not publish wage, inflation or any other macro data, and has no economic calendar. The chart on this page is ours, drawn from the public BLS series on FRED as downloaded on October 30, 2026.

Where the app helps is the market side of a wage surprise. The Treasury page shows yields from 1 month to 30 years, where a hot or soft wage print shows up quickly in the 2-year yield. The Fed topic and economy topic in Discover gather web coverage of each jobs report and what it implies for policy.

Frequently asked questions

What is a normal rate of wage growth?

There is no official target. In the data from 2008 through 2019, the year-over-year change in average hourly earnings ranged from 1.6% to 3.7%, and was between 2% and 3.5% in about three months out of four. Whether a given pace is consistent with steady inflation depends on productivity growth.

Why did wages jump in April 2020?

Because the average changed composition, not because pay rose. Average hourly earnings rose $1.34 in the first published figures for April 2020, and BLS attributed the increase largely to the substantial job loss among lower-paid workers, which removed low wages from the average.

What are real wages?

Real wages are wages adjusted for inflation: what an hour of pay buys. Real wage growth is approximately nominal wage growth minus the inflation rate. In September 2026, wages up 3.0% and CPI up 3.7% meant real hourly earnings fell about 0.7% over the year.

What is the difference between average hourly earnings and the ECI?

Average hourly earnings is monthly and reflects whoever is employed, so a shift in the job mix moves it. The Employment Cost Index is quarterly, holds the mix of jobs fixed, and also measures benefit costs.

Summary

Average hourly earnings is total private payroll divided by total hours, published monthly with the jobs report. Because it averages whoever is employed, it can rise when low-paid jobs disappear, as it did by 8.1% year over year in April 2020. Deflating it by CPI gives real wage growth, which was negative from April 2021 to April 2023 (−3.6% at the June 2022 low) and again from April to September 2026 (−0.7% in September). The quarterly Employment Cost Index holds the job mix fixed and is the cleaner gauge of labor costs when composition shifts.