Headlines say worker productivity surged 3.6% in the third quarter. The Bureau of Labor Statistics release behind them says that is true, and also that workers' share of output fell to 53.0%, the lowest in a series that starts in 1947. Real hourly pay is up just 1.0% over the past year.

What did the BLS actually report?

In its Productivity and Costs release for the third quarter, issued Thursday, BLS said nonfarm business productivity "increased 3.6 percent." Output rose 3.7% and hours worked rose 0.1%. In plain terms, the economy produced about 3.7% more while the country put in almost no extra hours.

Reuters reporter Lucia Mutikani wrote that economists had forecast 2.8%, so the number beat expectations. She also reported that unit labor costs fell 0.7%, the first drop since the second quarter of 2025.

Why the 3.6% needs a footnote

BLS reports quarterly changes at an annualized rate, and productivity swings a lot from one quarter to the next. The same release puts the gain over the past four quarters at 2.1%, and the second quarter was revised up to 2.1% as well. Reuters quotes Stephen Stanley, chief U.S. economist at Santander U.S. Capital Markets, saying "AI has yet to make a noticeable contribution," and notes the gains have run steadily for several years.

BLS also says the figures reflect the 2026 annual update of the national accounts, which revised output and productivity back to 2021. Older published numbers will not line up with these.

Where did the workers' share go?

BLS describes labor share as "the fraction of economic output that accrues to workers as compensation in exchange for their labor." A record low means output grew faster than pay did. The release's year-over-year column shows the gap: productivity up 2.1%, hourly compensation up 4.5%, and real hourly compensation up 1.0% once consumer prices are counted. That is a gap of about 1.1 points between output per hour and real pay per hour, by my arithmetic.

There is one better data point for workers. Real hourly compensation rose 1.1% in the third quarter after a revised 1.9% decline in the second. One good quarter does not undo a bad one.

Graduates of the Portsmouth Naval Shipyard trades apprenticeship and worker skills progression program, Class of 2025. U.S. Navy photo by Hanna Lanoie via Wikimedia Commons, public domain. Cropped; illustrative, not tied to the BLS figures.

What the hiring numbers say about the output

Extra output with almost no extra hours matches a slow-hiring economy. BLS's September jobs report showed payrolls up only 29,000, against a prior 12-month average of 45,000. The October report is due Friday morning. The Labor Department's weekly claims report shows initial jobless claims at 189,000 for the week ending Oct. 31, so layoffs are not driving this.

Factories did not share in the surge. The release has manufacturing productivity up only 0.3% for the quarter and 0.4% over the year, while manufacturing unit labor costs rose 4.0% over the year.

What this does not prove

The release does not say why labor's share fell, and it does not credit or blame any policy. It does not say AI caused the productivity gain. What it does show is a stronger-producing economy where pay per hour, adjusted for prices, is lagging output per hour. For a reader watching a paycheck against the grocery bill, that is the number to carry past the headline.

Header photo: banners reading "American Workers First" on the Frances Perkins Department of Labor Building in Washington, D.C., photographed March 18, 2026. Per the Commons description, the banners show Presidents Donald Trump and Theodore Roosevelt. Photo: PatrykTBrown via Wikimedia Commons, CC0. Cropped.