Productivity 1.4% and Unit Labor Costs 1.2% — Labor's Share of Output Hit a Record Low

Fundamentals · 2026-09-03

Nonfarm business productivity +1.4 percent annualized in the revised second quarter, unrevised from the preliminary, on output +1.7 and hours +0.3; unit labor costs revised down to +1.2 percent from 1.3, reflecting hourly compensation cut to 2.6 from 2.7; real hourly compensation fell 3.3 percent; the labor share of output came in at 52.8 percent, the lowest reading in a series that begins in the first quarter of 1947; manufacturing productivity revised up to +2.4 from 1.9 on output revised to +5.4 percent, the largest since Q2 2021, and manufacturing unit labor costs -0.3 percent, the first decline since Q2 2021; nonfinancial corporate unit profits +43.0 percent annualized, the highest since Q2 2021.

What Is This?

• What it is: The BLS Productivity and Costs report, released at 8:30am ET, second and revised estimate for the second quarter. Labor productivity is real output per hour worked. Unit labor costs are hourly compensation divided by productivity — BLS describes them explicitly as "an indicator of inflationary pressure on producers." The release also publishes the labor share, meaning the percentage of output that accrues to workers as compensation, and preliminary nonfinancial corporate figures including unit profits.
• Why it matters: It settles, with official data, the question of whether wages are driving inflation. It also arrives six days after Kevin Warsh asked at Jackson Hole whether artificial intelligence would produce "a significant, sustained rise in productivity across the economy. And if so, when?" This release is the first productivity data since he posed it.
• How to read it: Quarterly percent changes are seasonally adjusted annualized rates, so a 1.4 percent quarter is not a 1.4 percent year. Unit labor costs are the number that matters for inflation, and they are arithmetic: compensation growth minus productivity growth. Manufacturing output here is built from Federal Reserve industrial production and is not directly comparable to the business and nonfarm business output series, which come from GDP.

Want the full explainer? Economic News Events, Explained breaks down this release and every other one we cover, in plain English.

Summary

Nonfarm business labor productivity rose 1.4 percent at an annual rate in the revised second quarter, unchanged from the preliminary estimate, with output up 1.7 percent and hours worked up 0.3. Against the same quarter a year ago productivity is up 2.2 percent. Unit labor costs were revised down to 1.2 percent from a previously reported 1.3, reflecting a tenth of a point trimmed from hourly compensation, now 2.6 percent. Over the last four quarters unit labor costs rose just 1.4 percent.

Hold that last figure against headline PCE inflation running at 3.7 percent over twelve months. Unit labor costs are rising at well under half the rate of consumer prices. BLS's own framing calls this measure an indicator of inflationary pressure on producers, and by that measure labor is exerting almost none. It closes an argument that has been narrowing all week: Warsh conceded at Jackson Hole that wage growth "has not proven a reliable indicator of future inflation for a very long time," ADP showed base pay at 3.2 percent against 3.7 percent inflation yesterday, and the official productivity accounts now confirm it from the cost side. Whatever is producing 3.7 percent inflation, it is not the cost of labor.

The number that should lead every write-up of this release, and almost certainly will not, is one sentence in the third paragraph. The labor share — the percentage of output that accrues to workers in the form of compensation — was 52.8 percent in the second quarter, the lowest level in the series, which begins in the first quarter of 1947. That is a record low across seventy-nine years of data. Not a multi-year low, not a post-pandemic low. The lowest ever recorded.

It did not happen quietly. Real hourly compensation fell 3.3 percent at an annual rate in the quarter, revised from a preliminary -3.1, and is down 0.1 percent over the last four quarters. Nominal hourly compensation rose 2.6 percent while real compensation fell 3.3, which implies consumer prices rose at roughly a 6 percent annualized rate over the quarter — the energy shock quarter, with the Iran conflict running since late February. Workers produced 1.4 percent more per hour and were paid 3.3 percent less in real terms for it.

The other side of that ledger is in the nonfinancial corporate table. Unit profits increased at an annualized rate of 43.0 percent in the second quarter, the highest since the second quarter of 2021, and are up 17.8 percent over four quarters, the highest since the fourth quarter of 2021. Nonfinancial corporate productivity rose 2.2 percent with unit labor costs at -0.3 percent. Put the three facts in one line: output per hour rose, the cost of producing it fell, and the gain went to profit rather than pay. The labor share record and the profit figure are the same event described twice.

That finding does not sit alone. It is the fourth dataset this fortnight pointing at the same distribution. Michigan found sentiment declines concentrated among lower- and middle-income households and consumers with no stock holdings. ADP's own researchers said base pay growth for lower-paid workers is now slower than before the pandemic. Warsh told Jackson Hole that S&P 500 profits were up more than 20 percent with margins "quite elevated, relative to history" — and, in the same speech, that the most serious harm from policy error "is likely to befall those without financial assets." The productivity accounts have now put an official number on the gap he described.

The manufacturing revisions are large and they matter for a different reason. Manufacturing productivity was revised up to 2.4 percent from 1.9, on output revised up to 5.4 percent from 4.6 — the largest quarterly increase since the second quarter of 2021. Durable goods were revised harder still: productivity to 3.6 percent from 2.7, output to 8.9 percent from 7.3, hours to 5.1 from 4.5. And manufacturing unit labor costs came in at -0.3 percent, revised down from a flat 0.0, the first decline in manufacturing unit labor costs since the second quarter of 2021. Durable goods unit labor costs fell 2.2 percent.

This does not contradict Tuesday's ISM survey showing manufacturing losing momentum, or yesterday's ADP showing manufacturing shedding 17,000 jobs. It dates them. Q2 covers April through June; ISM and ADP cover August. What this release says is that the manufacturing boom was bigger than first reported in the spring — factories converted 2.9 percent more hours into 5.4 percent more output — and what August's data says is that it has since rolled over. The peak is now visible, and it was in the second quarter.

There is also a clean corroboration of something the ISM brief argued. Manufacturing unit labor costs fell 0.3 percent in the quarter while ISM's Prices Index has sat at 71.1 for twenty-three consecutive months with twenty-four commodities up in price against two down. Factory costs per unit of output are falling on the labor side and rising hard on the materials side. Manufacturing inflation is a commodities and energy story, not a wage story, and these two datasets now say so from opposite directions.

Finally, the AI question. Warsh asked six days ago whether AI would deliver a sustained productivity acceleration and when. The answer in this release is: not visibly, not yet. Productivity in the current business cycle, from the fourth quarter of 2019 through the second quarter of 2026, has grown at an annualized 2.1 percent — better than the previous cycle's 1.5 percent, but exactly equal to the long-term rate since 1947. Manufacturing is worse: 0.5 percent annualized in this cycle against a long-term rate of 2.1 percent since 1987. Capital goods imports may be up 46.9 percent year over year, but the productivity statistics have not yet registered anything the last seventy-nine years would call unusual.

The Internals

Revised second-quarter 2026, percent change from the previous quarter at an annual rate:

Sector · Labor productivity · Output · Hours worked · Hourly compensation · Real hourly compensation · Unit labor costs

Nonfarm business · 1.4 · 1.7 · 0.3 · 2.6 · -3.3 · 1.2

Business · 1.2 · 1.6 · 0.4 · 2.4 · -3.4 · 1.2

Manufacturing · 2.4 · 5.4 · 2.9 · 2.1 · -3.7 · -0.3

Durable manufacturing · 3.6 · 8.9 · 5.1 · 1.4 · -4.4 · -2.2

Nondurable manufacturing · 2.1 · 1.4 · -0.6 · 2.9 · -3.0 · 0.8

Nonfinancial corporate · 2.2 · 3.9 · 1.7 · 1.9 · -3.9 · -0.3

The same measures against the year-ago quarter:

Sector · Labor productivity · Output · Hours worked · Hourly compensation · Real hourly compensation · Unit labor costs

Nonfarm business · 2.2 · 2.5 · 0.2 · 3.7 · -0.1 · 1.4

Business · 2.1 · 2.5 · 0.4 · 3.5 · -0.3 · 1.4

Manufacturing · 1.1 · 1.6 · 0.5 · 4.5 · 0.7 · 3.4

Durable manufacturing · 2.5 · 3.6 · 1.0 · 5.6 · 1.7 · 3.0

Nondurable manufacturing · -0.2 · -0.6 · -0.4 · 2.2 · -1.6 · 2.4

Nonfinancial corporate · 3.1 · 3.8 · 0.7 · 3.5 · -0.3 · 0.4

What changed between the preliminary and the revised estimate for the second quarter:

Measure · Previously published · Revised · Change

Nonfarm productivity · 1.4 · 1.4 · Unrevised

Nonfarm output · 1.7 · 1.7 · Unrevised

Nonfarm hours worked · 0.3 · 0.3 · Unrevised

Nonfarm hourly compensation · 2.7 · 2.6 · Down 0.1

Nonfarm real hourly compensation · -3.1 · -3.3 · Down 0.2

Nonfarm unit labor costs · 1.3 · 1.2 · Down 0.1

Manufacturing productivity · 1.9 · 2.4 · Up 0.5

Manufacturing output · 4.6 · 5.4 · Up 0.8

Manufacturing hours worked · 2.6 · 2.9 · Up 0.3

Manufacturing unit labor costs · 0.0 · -0.3 · Down 0.3

Durable productivity · 2.7 · 3.6 · Up 0.9

Durable output · 7.3 · 8.9 · Up 1.6

Durable unit labor costs · -1.6 · -2.2 · Down 0.6

Nondurable productivity · 2.0 · 2.1 · Up 0.1

Nondurable unit labor costs · 1.1 · 0.8 · Down 0.3

Unit labor costs are arithmetic — compensation growth minus productivity growth — and every line checks:

Sector · Hourly compensation · Less productivity · Equals unit labor costs

Nonfarm business · 2.6 · 1.4 · 1.2

Manufacturing · 2.1 · 2.4 · -0.3

Durable manufacturing · 1.4 · 3.6 · -2.2

Nondurable manufacturing · 2.9 · 2.1 · 0.8

Nonfinancial corporate · 1.9 · 2.2 · -0.3

First-quarter 2026 revisions, for completeness:

Measure · Previously published · Revised

Nonfarm productivity · 0.8 · 0.8, unrevised

Nonfarm unit labor costs · 1.3 · 1.3, unrevised

Manufacturing productivity · 1.9 · 2.2

Manufacturing unit labor costs · 3.5 · 3.3

Durable productivity · 4.6 · 4.9

Durable unit labor costs · 2.9 · 2.6

Nonfinancial corporate productivity · 1.4 · 1.4, unrevised

Where The Output Went

The distributional numbers, which is what this release is really about this quarter:

Measure · Reading · Context BLS attached

Labor share of output · 52.8 percent · The lowest level in the series, which begins in the first quarter of 1947

Real hourly compensation, nonfarm · -3.3 percent annualized · Revised down from a preliminary -3.1

Real hourly compensation, four quarters · -0.1 percent · Effectively flat in real terms over a year

Nonfinancial corporate unit profits · +43.0 percent annualized · The highest rate since the second quarter of 2021

Unit profits, four quarters · +17.8 percent · The highest rate since the fourth quarter of 2021

Nonfinancial corporate unit labor costs · -0.3 percent · Costs per unit fell while profits per unit surged

Nominal hourly compensation rose 2.6 percent while real hourly compensation fell 3.3, which implies consumer prices ran at roughly a 6 percent annualized rate in the quarter. That is the arithmetic gap between the two figures BLS published, not a separate estimate.

Productivity growth across business cycles, which is the frame for the AI question:

Period · Nonfarm productivity, annualized · Output · Hours worked

Current cycle, Q4 2019 to Q2 2026 · 2.1 percent · 2.5 percent · 0.4 percent

Previous cycle, Q4 2007 to Q4 2019 · 1.5 percent · Not stated in this release · Not stated in this release

Long-term, since Q1 1947 · 2.1 percent · Not stated in this release · Not stated in this release

The current cycle beats the previous one and exactly matches the seventy-nine-year average. It does not yet show an acceleration.

Manufacturing tells a weaker version of the same story:

Period · Manufacturing productivity, annualized · Output · Hours worked

Current cycle, Q4 2019 to Q2 2026 · 0.5 percent · 0.2 percent · -0.3 percent

Previous cycle, Q4 2007 to Q4 2019 · 0.1 percent · Not stated in this release · Not stated in this release

Long-term, since Q1 1987 · 2.1 percent · Not stated in this release · Not stated in this release

Manufacturing productivity in this cycle is running at roughly a quarter of its long-term rate.

Two records and two firsts in one release, worth listing plainly:

Item · Reading · Standing

Labor share · 52.8 percent · Lowest since the series began in 1947

Manufacturing output · +5.4 percent · Largest quarterly increase since Q2 2021

Manufacturing unit labor costs · -0.3 percent · First decline since Q2 2021

Nonfinancial corporate unit profits · +43.0 percent · Highest since Q2 2021

One methodological caution BLS flags directly: the concepts, sources and methods behind the manufacturing output series differ from those behind the business and nonfarm business series, so the two output measures are not directly comparable. Manufacturing output here is built on Federal Reserve industrial production indexes; nonfarm business output comes out of GDP.

Impact on USD

• Mixed, lean bearish — unit labor costs at 1.2 percent annualized and 1.4 percent over four quarters remove the last credible wage-inflation argument for tightening.
• Against 3.7 percent headline PCE, labor costs are rising at well under half the rate of consumer prices, which narrows what a rate hike would actually fix.
• Real hourly compensation down 3.3 percent in the quarter and flat over four quarters is a consumption headwind, and consumption is 70 percent of the economy.
• Offsetting: solid productivity at 1.4 percent quarterly and 2.2 percent year over year is genuinely good news for potential growth, and a higher productivity trend argues for a higher neutral rate.
• Second-order to tomorrow's payroll report, which after this week's stack of ADP, JOLTS and ISM carries more weight than any figure in this release.

Impact on US Indices (ES / NQ / YM)

• Bullish — unit profits up 43.0 percent annualized with unit labor costs at -0.3 percent in the nonfinancial corporate sector is as clean a margin signal as this dataset produces.
• The labor share at a record low is, mechanically, the profit share at a record high — the corporate sector is capturing more of each unit of output than at any point since 1947.
• Manufacturing revised up hard: output +5.4 percent, the best since Q2 2021, with unit labor costs falling for the first time in five years. Industrials had a better Q2 than first reported.
• The catch is timing. This is April to June. ISM and ADP say August was materially weaker, so treat the manufacturing strength as a peak that has already passed rather than a run rate.
• The consumer-facing caveat is the same one that has run all week: real pay is falling, and the households feeling it are the ones without equity exposure.

Impact on Gold

• Mixed, lean bearish — softer unit labor costs weaken the inflation-persistence case that has been supporting the metal.
• If labor costs are not driving inflation, then the inflation that exists is a commodity and energy story, which is a narrower and more reversible base for a debasement trade.
• Offsetting: real hourly compensation falling 3.3 percent, with the labor share at a record low, is the kind of distributional stress that supports a longer-run bid.
• Conditional: tomorrow's payroll report matters more, and the September 11 inflation print more still.

What To Watch

• August employment situation — Friday, September 4, 8:30am ET. FactSet consensus is +65,000. ADP said +38,000 private, JOLTS implied -18,000 for July, and claims this morning were benign but trending up.
• The labor share. At 52.8 percent it is at a seventy-nine-year low. Whether it stabilizes or keeps falling is the most consequential number in this dataset and it is reported once a quarter.
• August CPI — Friday, September 11. With unit labor costs at 1.4 percent over four quarters, a hot print would be unambiguously about commodities rather than wages.
• FOMC decision and dot plot — Tuesday and Wednesday, September 15-16. Three officials dissented for a hike in July. This release removes the wage-inflation leg of that argument entirely.
• Preliminary third-quarter Productivity and Costs — Thursday, November 5, 8:30am ET. It will show whether the Q2 manufacturing peak was a peak, and whether unit profits held near 43 percent.
• The AI productivity question. Warsh asked when it would show up. The cycle is running at the long-term 2.1 percent rate. A break above that is the thing to watch for.

TLDR

Productivity and Costs, Second Quarter 2026, Revised (released September 3):

• Nonfarm business productivity: +1.4 percent annualized, unrevised from preliminary; +2.2 percent over four quarters
• Unit labor costs: +1.2 percent, revised down from 1.3, on hourly compensation trimmed to 2.6 from 2.7; +1.4 percent over four quarters
• The labor share of output was 52.8 percent — the lowest level in a series that begins in the first quarter of 1947
• Real hourly compensation fell 3.3 percent in the quarter, revised down from -3.1, and is -0.1 percent over four quarters
• Nonfinancial corporate unit profits rose 43.0 percent annualized, the highest since Q2 2021, and 17.8 percent over four quarters, the highest since Q4 2021
• Manufacturing productivity revised up to 2.4 percent from 1.9, on output revised to +5.4 percent, the largest since Q2 2021
• Manufacturing unit labor costs fell 0.3 percent — the first decline since Q2 2021; durable goods unit labor costs fell 2.2 percent
• Durable manufacturing output revised up to +8.9 percent from 7.3, with hours up 5.1 percent
• Current-cycle productivity growth is 2.1 percent annualized, above the previous cycle's 1.5 but exactly equal to the long-term rate since 1947
• Manufacturing productivity in this cycle runs at 0.5 percent against a 2.1 percent long-term rate
• Unit labor costs at 1.4 percent over four quarters against 3.7 percent headline PCE — labor is not the source of inflation

The revisions are small and the sentence nobody will quote is the whole story: the labor share of output came in at 52.8 percent, the lowest in the seventy-nine years BLS has measured it. Real hourly compensation fell 3.3 percent in the quarter while nonfinancial corporate unit profits rose 43.0 percent annualized, the highest since 2021. Output per hour went up, the cost of producing it went down, and the difference went to profit rather than pay. That is the same distribution Michigan found in sentiment, ADP found in base pay, and Warsh described from the podium at Jackson Hole. It also closes the wage-inflation debate: unit labor costs are rising 1.4 percent over four quarters against 3.7 percent consumer inflation, so whatever is driving prices, it is not labor. Manufacturing was revised up hard — output +5.4 percent, the best since Q2 2021, with unit costs falling for the first time in five years — but that is April to June, and ISM and ADP say August has already rolled over. Payrolls September 4, CPI September 11, FOMC September 15-16, next productivity report November 5.

_For informational purposes only. Not investment advice._