What Hourly Pay Measures, and Why the Average Can Mislead
The pay figure in each jobs report is an average, and an average can rise when nobody gets a raise. Its fastest growth since 1981 came in the spring of 2020, when millions of low-paid workers lost their jobs.
October 29, 2026 · updated October 30, 2026
In one sentence
The almanac's hourly pay line is the average hourly earnings of production and nonsupervisory workers, the roughly four in five private-sector employees who are not managers, from each month's jobs report, shown as the change from a year earlier.
New articles and charts arrive in one short email on Monday mornings. Get the weekly digest
The analogy
The average height of a room. If the shortest people walk out, the average rises, and nobody grew. That is what happened in April 2020: growth in average hourly pay jumped from 3.6% to 7.8% in a single month, largely because millions of low-paid workers, many of them in restaurants, shops and hotels, lost their jobs and stopped being counted, as the Bureau of Labor Statistics noted at the time. The room was taller because it was emptier. When those jobs came back in 2021, the same arithmetic ran in reverse, and the average slowed while restaurants were raising pay to hire.
How it reaches your kitchen table
Your raise: against prices. A raise is only as good as what it buys. In September 2026 hourly pay was up 3.4% from a year earlier and prices 3.7%, so an hour of work bought about 0.3% less than it did a year before. (The Bureau of Labor Statistics' own calculation for September 2026, which uses the price index for wage earners, puts the drop at 0.6%.) That gap went below zero in September. The almanac draws it as "real wage growth" on the hourly pay chart; real vs. nominal explains why it is the number that matters.
Your next job: through the switch. People who change jobs usually get bigger raises than people who stay, and the gap is a running measure of how hard employers are competing. In September, by the Atlanta Fed's count, people who had switched were getting raises of 4.7% a year and people who stayed 3.4%. In mid-2022 switchers were getting 8.6%, the most since the series began in 1997. The gap between the two closed entirely from mid-2024 to mid-2025, when moving paid no more than staying, and has since reopened to about its 2019 size. What hasn't come back is the moving itself: about a fifth fewer workers quit each month than in 2019.
Prices and rates: through the Fed. The Fed watches pay for signs that inflation is settling into wages. The test Jerome Powell described as chair in 2021 was real pay rising persistently faster than productivity, which pushes businesses to raise prices. In 2026 real pay is falling, so the question is whether the oil shock will spread into pay demands, not whether it already has.
What moves it
How tight the job market is, first: when quits are high and openings plentiful, employers raise pay to keep people. Inflation, second, because pay chases prices with a lag of a year or so. Productivity, over the long run, because it is what lets pay rise without prices following. Minimum-wage increases at the bottom, now mostly set by states. And the mix of who is working, which is the part that fools people.
How to read this month's number
Against inflation. The headline pay figure is nominal. Subtract inflation, or turn on the inflation line on the chart below, before deciding whether it is good news.
Against the other measures. Three others check the average, each by removing a different distortion. The Employment Cost Index prices a fixed mix of jobs, so a change in who is working can't move it; in the year to September 2026 it showed private-sector wages and salaries up 3.1% (CIU2020000000000I), the same pace as in the year to June, and down 0.6% after inflation. The Atlanta Fed's tracker follows the same people a year apart and takes the median of their raises; in September that was 3.9%. It usually runs above the average because it counts only people who held a job in both years; the average is also pulled down by new workers starting at lower pay and by higher-paid older workers retiring. Median weekly earnings, from the household survey, is the middle full-time worker's paycheck rather than the average: $1,232 in the third quarter, up 1.5% from a year earlier while prices rose 3.5%, a real decline of about 2%. When all four agree, believe them. When the average jumps and the others don't, suspect the mix of jobs, not a raise.
Against its own history. The 2015–19 average was 2.7%. The 2022 peak was 7.0%, the fastest since 1982 apart from the 2020 composition spike; in the early 1980s pay growth topped 9% while inflation ran in double digits. At 3.3% in October 2026, the reading is lower than about 63% of months since 1965 and not far above the pre-pandemic norm.
Where the analogy breaks
People in a room don't grow, and workers do get raises, often in the same months that the mix of jobs is shifting, so the average blends the two and can't tell you which did the work. The room also has a door policy. The series leaves out managers and the self-employed, and it counts jobs, not people, so someone with two part-time jobs counts twice. The tracker fixes the mix problem by following the same individuals, but only those who had a job a year ago and still have one, so it misses exactly the people who lost work or just found it.
What people argue about
Whether workers are better off than before the pandemic. One side: average hourly pay for these workers is up about 39% from 2019 and prices about 31%, a real gain of about 5%, and the lowest-paid gained most in 2020–23, when switching jobs was easiest and employers competed for them. The other: the gain came in bursts and is now eroding, real pay has fallen since April, and some of what people can't cut back on rose faster than prices overall: rent is up about 36% over the same years, electricity about 44%. The wage-earner path takes this up with the chart.
Whether low immigration raises pay. One side: fewer new workers means employers have to compete harder for the ones already here, and pay should rise, especially in construction, farming and food processing. The other: the National Academies' review of the research found immigration's effect on native-born wages very small over a decade or more, with any losses falling mostly on earlier immigrants and workers without a high school diploma; and the 2025–26 slowdown in immigration came with slower hiring, not faster pay growth, as hourly pay growth eased from 3.7% in February 2026 to 3.3% in September. Both sides expect any effect to show up first in particular industries, not in the national average.
Whether wages are an inflation risk now. One side: unemployment is still low, and if the oil shock feeds into pay demands, 2026 could repeat the 1970s spiral, when wages and prices chased each other for a decade. The other: pay growth has been slowing all year, about one worker in ten belongs to a union, and research by Ben Bernanke and Olivier Blanchard found that most of the 2021–23 surge came from prices rather than wages, though tight job markets had a longer-lasting effect. Supply vs. demand has the larger version of this argument.
See also
- Real vs. nominal: the raise that didn't feel like one.
- The jobs report: where this number comes from each month.
- You work for a wage: pay, prices and your leverage, in order.
- The almanac's hourly pay with every overlay.
Sources
- FRED: average hourly earnings of production and nonsupervisory employees (AHETPI), the almanac's series; CPI-U (CPIAUCSL), rent of primary residence (CUSR0000SEHA) and electricity (CUSR0000SEHF01), for the changes since 2019 (2019 average to September 2026).
- BLS, The Employment Situation, September 2026: who the series covers (81.6% of private jobs); Real Earnings, September 2026.
- BLS, The Employment Situation, April 2020: the bureau's note that the jump in average pay reflected job losses among lower-paid workers.
- BLS, Employment Cost Index, September 2026 and its fixed-weight method; Atlanta Fed Wage Growth Tracker, with job switchers and stayers, and FRED: quits rate (JTSQUR); BLS, Usual Weekly Earnings, third quarter 2026.
- Federal Reserve Board, "What Do Various Wage Measures Tell Us about Underlying Wage Growth?" (January 2025), on how the measures differ; San Francisco Fed, "What's Up with Wage Growth?" (2016), on how new workers and retirements pull the average down.
- Federal Reserve, press conference transcript, December 15, 2021: Powell on real wages and productivity.
- Autor, Dube and McGrew, "The Unexpected Compression", on the low-paid's gains after the pandemic; National Academies, The Economic and Fiscal Consequences of Immigration (2016); Bernanke and Blanchard, "What Caused the US Pandemic-Era Inflation?" (2025); BLS, union membership rate 10.0% in 2025.