A data-driven analysis of what labor productivity growth actually measures, why the current readings matter more than they appear, the unresolved debate about whether a structural acceleration is underway, and what the answer means for inflation, wages, hiring, and the broader economic outlook.
Nonfarm business sector labor productivity grew at a 4.9 percent annualized rate in Q3 2025, the strongest reading since Q3 2023. For the full year 2025, average annual productivity growth came in at 2.1 percent, broadly in line with the long-run post-war average of 2.2 percent going back to 1947 and above the 1.5 percent rate of the previous business cycle from Q4 2007 through Q4 2019.
Those numbers receive a fraction of the coverage devoted to monthly payroll reports, inflation readings, or yield curve movements. That imbalance is analytically difficult to justify. Productivity growth is the mechanism through which wages can rise without generating inflation, through which the economy can expand without tightening, and through which the same level of labor input produces more output over time. It is the variable that most directly determines whether the current economic expansion is sustainable and whether the labor market can continue to operate close to full employment without reigniting the inflation that the Federal Reserve has spent four years trying to contain.
This article explains what productivity data actually measures, places the current readings in historical context, examines the unresolved question of whether a structural acceleration is underway, and draws out what the answer means for wages, inflation, hiring budgets, and the Fed's policy path.
Labor productivity, as measured by the Bureau of Labor Statistics, is calculated by dividing an index of real output by an index of hours worked. The resulting figure — output per hour — captures how much economic value is produced per unit of labor input. An increase in productivity means the economy is generating more output for the same number of hours worked, which is the fundamental mechanism of economic growth beyond simple additions of workers or working hours.
The BLS publishes productivity data quarterly with approximately a six-week lag, and annual average figures are revised regularly as source data from the Bureau of Economic Analysis and the Federal Reserve are updated. The primary series covers the nonfarm business sector, which excludes government, nonprofit organizations, and private households. A separate series covers the manufacturing sector. Neither series receives the immediate market-moving attention that monthly payroll releases or CPI prints attract.
That asymmetry in attention reflects a genuine feature of how the data is used in real time: productivity is a quarterly series released with a lag, subject to substantial revision, and its implications play out over periods of quarters to years rather than days to weeks. But the practical consequence of underattending to it is that one of the most consequential signals about the medium-term health of the economy is systematically underweighted in the analysis that informs business and policy decisions.
The three variables that productivity connects most directly to each other are wages, inflation, and employment. Understanding those connections is the first analytical step before examining what the current data shows.
The relationship between productivity, wages, and inflation is not metaphorical. It is arithmetic.
Unit labor costs — the labor cost per unit of output — are calculated as hourly compensation divided by labor productivity. When productivity rises faster than compensation, unit labor costs fall. When compensation rises faster than productivity, unit labor costs increase. Because unit labor costs are among the most important inputs to price-setting decisions for goods-producing and service businesses, the relationship between productivity and wages is a direct determinant of inflationary pressure.
As the Federal Reserve Bank of Cleveland has articulated, the slowdown in trend productivity growth in the early 1970s is widely believed to have contributed to inflationary pressures during that decade. Policymakers were slow to recognize that the economy could not grow as rapidly without generating inflation as the higher-productivity era had allowed. The mirror image is the late 1990s, when Fed Chair Alan Greenspan held off tightening monetary policy despite strong GDP growth specifically because he judged that rising productivity growth allowed the economy to expand faster than the old rules of thumb would have indicated.
This is not ancient history. It is directly relevant to the Fed's current policy path. If productivity growth is sustainably above its post-financial-crisis trend, the Fed can allow the economy to run somewhat hotter without reigniting inflation. If the recent readings are cyclical rather than structural, the same wage growth that current productivity figures make look benign would reveal itself as inflationary once productivity normalizes lower.
Understanding whether the current productivity readings are signal or noise requires placing them in the long sweep of post-war US productivity history. That history is not a smooth trend. It is a sequence of regime shifts that have mattered enormously for economic conditions at each transition.
The quarter-century following World War II was defined by exceptionally high productivity growth. The United States enjoyed roughly 3 percent growth in output per man-hour from 1947 through 1967. This was the era of mass electrification, the expansion of the interstate highway system, the consolidation of mass production manufacturing, and the maturation of a generation of technologies — refrigeration, air conditioning, synthetic materials — that had been developed earlier but required years to achieve broad economic impact.
The productivity gains of this era were large enough and durable enough to support rising real wages and price stability simultaneously. The labor share of income remained high, unemployment was structurally low, and the distributional gains from productivity were broadly shared.
The productivity growth rate fell sharply in the early 1970s and remained depressed for more than twenty years. Multifactor productivity in the overall economy grew at approximately 0.3 percent annually from 1973 to 1994, compared with 2.2 percent over the preceding 25 years. In manufacturing, it fell from 1.8 percent in 1949 to 1973 to 0.8 percent in 1973 to 1992.
The causes remain debated. The OPEC oil shock and regulatory burden were early explanations that largely ran their course without restoring growth. The influx of less experienced workers from the baby boom generation, the shift toward services, and the mismeasurement of service sector output have all been proposed. What is clear is that policymakers who assumed the economy could continue growing at the old pace without generating inflation were wrong, and the resulting stagflation of the 1970s was partly a consequence of that misdiagnosis of the economy's productive capacity.
From 1995 onward, productivity growth roughly doubled from its post-1973 pace. Between 1995 and 2003, the rate more than doubled to levels comparable to the pre-1973 era. The acceleration is widely attributed to the production and use of information technology: computers, semiconductors, software, and eventually the Internet. The US invested heavily in IT capital throughout this period, with investment in information-processing equipment and software growing at approximately 20 percent per year from 1995 through 1999.
The IT boom vindicated the argument that technology adoption, with appropriate lags, eventually shows up in aggregate productivity statistics. But it also illustrated the Solow paradox — named for Robert Solow's observation that you can see the computer age everywhere but in the productivity statistics — which notes that new general-purpose technologies take years or decades before their productivity effects become visible in aggregate data. Computers were widespread in US offices by the 1980s, but the productivity payoff did not register until the mid-1990s.
After 2004, the IT-driven productivity acceleration faded. Productivity growth returned approximately to its 1973 to 1995 pace for the decade leading into the Great Recession and remained subdued through the slow recovery of 2009 to 2019. The business cycle from Q4 2007 through Q4 2019 produced an average annualized productivity growth rate of 1.5 percent, below the long-run post-war average and well below the IT boom rates.
Since Q4 2019 — the starting point of the current business cycle in BLS analysis — nonfarm business productivity has grown at an annualized rate of 2.1 percent. That is higher than the previous cycle rate, approximately in line with the long-run average, and has been associated with quarterly readings that reached 4.9 percent in Q3 2025.
The BLS data for 2025 presents a specific picture that is worth examining carefully. Quarterly annualized productivity growth ran at 4.1 percent in Q2 2025, then accelerated to 4.9 percent in Q3 2025, the strongest reading since Q3 2023. Q4 2025 then showed a deceleration to 1.8 percent, with output expanding 2.6 percent and hours worked declining slightly.
The year-over-year change — which smooths the quarterly volatility — showed nonfarm business productivity rising 2.5 percent from Q4 2024 to Q4 2025, and the annual average for 2025 came in at 2.1 percent, moderating from a 3.0 percent increase recorded for 2024.
Unit labor costs in the nonfarm business sector increased 4.4 percent in Q4 2025, reflecting a 6.3 percent increase in hourly compensation and only 1.8 percent productivity growth in that quarter. That combination produced the highest quarterly unit labor cost increase since Q3 2022. For the full year, unit labor costs increased 2.4 percent over the last four quarters — a reading that is consistent with the Fed's inflation target if it is sustained, but which bears watching if productivity decelerates and compensation growth does not follow.
In the nonfinancial corporate sector — a cleaner measure for understanding business profitability — productivity grew at 4.5 percent in Q3 2025 and unit labor costs increased only 1.2 percent, with unit profits rising 6.6 percent. That configuration describes a productive economy with controlled cost pressures and healthy corporate margins, which is the positive scenario for sustained expansion.
Manufacturing sector productivity tells a different story. Manufacturing productivity decreased 2.5 percent in Q4 2025, producing unit labor cost increases of 9.1 percent in that quarter — the largest since Q3 2022. For the current business cycle as a whole, manufacturing productivity has grown at only 0.4 percent annualized, far below the 2.1 percent rate for the broader nonfarm business sector.
This divergence is analytically important. The nonfarm business productivity acceleration is driven primarily by services, technology, and finance — sectors where AI tools, workflow automation, and digital transformation are most directly applicable. Manufacturing productivity, by contrast, has been held back by structural challenges including tariff-induced supply chain disruption, the capital expenditure required for reshoring, and the slower pace at which physical production processes respond to digital productivity tools.
The BLS also publishes total factor productivity data, which measures output growth not explained by growth in either labor or capital. Total factor productivity increased 0.8 percent in the private nonfarm business sector in 2025, following a 1.5 percent increase in 2024. The deceleration in TFP is worth noting: the headline labor productivity figures held up partly because hours worked declined while output expanded, meaning the productivity number reflects both genuine efficiency gains and a reduction in labor input rather than purely the former.
The most important analytical question about the current productivity data is whether it represents a structural shift to a higher trend growth rate or a cyclical fluctuation that will revert to the post-2004 pace. The evidence is genuinely mixed, and the range of credible estimates is wide.
The Federal Reserve Bank of Cleveland published a regime-shift analysis in early 2025 using a statistical model designed to detect changes in trend productivity growth. Their model, updated with the most recent data, estimated approximately a 40 percent probability that the economy has shifted into a high-growth productivity regime. That is a substantial increase compared to pre-pandemic estimates but, as the Cleveland Fed noted, still less than even odds. The model requires several more quarters of data before a more unambiguous conclusion is possible.
The Atlanta Fed CFO Survey — drawing on responses from nearly 750 corporate executives — found that implied revenue-based labor productivity gains in 2025 were positive and varied by sector. High-skill services and finance showed the largest effects at roughly 0.8 percent, while low-skill services, manufacturing, and construction saw about 0.4 percent. Critically, these effects were expected to roughly double in 2026, with high-skill services and finance potentially exceeding 2 percent. If those expectations are realized, the aggregate productivity signal would strengthen materially.
The productivity gains documented at the task level from AI adoption in specific domains are large: approximately 30 percent improvements in coding productivity and customer service operations, substantial quality and speed gains for management consultants, and measurable efficiency improvements in professional writing, legal analysis, and call center operations. These micro-level gains are real, well-documented, and growing. The question is whether they are large and diffuse enough to show up in aggregate productivity statistics.
Goldman Sachs published a research note in early 2026 finding no meaningful relationship between AI and productivity at the economy-wide level. While management teams quantifying AI-driven productivity impacts on specific tasks reported a median gain of around 30 percent in targeted functions, Goldman found that aggregate labor market impacts remained negligible and that there was no discernible effect on major economic indicators including the unemployment rate, layoffs, or measured productivity outside of a few specific use cases.
The historical analogy is cautionary. As the Cleveland Fed explicitly noted, the information technology revolution widely accepted to have begun in the 1970s coincided with a productivity slowdown that lasted more than 20 years, with higher productivity growth only emerging in the late 1990s. If the pattern holds, AI investments being made today may produce their aggregate productivity payoff in the mid-2030s rather than in the current data.
Daron Acemoglu's analysis concluded that aggregate total factor productivity gains from AI over the next ten years were unlikely to exceed 0.66 percent in total, implying only a marginal increase in annual TFP growth. That estimate stands at the low end of the range but is not outside the mainstream of credible academic research.
The mismeasurement problem is also genuinely difficult. As documented in the history of the productivity paradox, standard national accounts methodology can either overstate or understate productivity growth when the nature of output is changing rapidly. If AI is improving the quality of output without proportionally raising its measured value — better code, faster analysis, higher-quality writing — the productivity gains will not appear in the data in proportion to their real economic benefit.
The honest analytical position is that the available data provides tentative but not conclusive support for an above-trend productivity regime. The Cleveland Fed's 40 percent probability estimate is the most rigorous formal assessment available, and it is neither dismissible nor decisive. Several more quarters of data above the long-run average would shift the probability materially; a reversion to the post-2004 pace would reduce it.
What the data does clearly establish is that productivity growth since Q4 2019 has been running above the previous business cycle rate of 1.5 percent, with the current annualized rate of 2.1 percent more consistent with the long-run post-war average than with the secular stagnation narrative that dominated economic analysis through much of the 2010s. That is not a minor change in the macroeconomic backdrop.
[H2] What the Productivity Signal Means for Inflation, Wages, and the Fed
The policy implications of productivity growth are direct and consequential, and the current readings have specific implications for each of the major economic actors monitoring the data.
Higher productivity growth creates space for wage growth without proportionate price increases. PIMCO's analysis framed this precisely: productivity gains that do not accrue to workers in the form of compensation tend to be disinflationary, as they reduce unit labor input costs, which eventually feed into price dynamics. Mathematically, nominal wage inflation equals price inflation plus productivity gains plus changes in the labor share of income.
At the current business cycle pace of 2.1 percent productivity growth, wages can rise at a rate of approximately 4 percent annually without generating more than 2 percent inflation — assuming the labor share of income remains stable. The Q4 2025 data, which showed 6.3 percent hourly compensation growth against only 1.8 percent productivity growth in that quarter, pushed unit labor costs to 4.4 percent. If that quarterly imbalance persists into 2026, it would represent a genuine inflationary signal. If it reflects a one-quarter mismatch between compensation timing and the productivity cycle, the full-year average of 2.4 percent unit labor cost growth is consistent with the Fed's target.
The Federal Reserve's ability to allow the economy to run at strong growth and low unemployment without tightening monetary policy is directly tied to its assessment of trend productivity growth. The late 1990s precedent — when Greenspan held off rate increases despite strong growth because he judged that rising productivity was sustainably expanding the non-inflationary growth rate — is the positive analogy. The 1970s precedent — when the Fed allowed the economy to overheat because it misjudged the productivity slowdown as temporary — is the cautionary one.
Charles Schwab's 2026 fixed income outlook anticipated the Fed lowering the federal funds rate to the 3.0 to 3.5 percent range over the next year, citing labor market weakness as the primary driver. If that cutting cycle proceeds while productivity growth remains above trend, the combination would be supportive of non-inflationary expansion. If productivity decelerates simultaneously with the cutting cycle, the inflation risks rise.
PIMCO's analysis identified what it described as a K-shaped economy, in which large capital-intensive firms aggressively deploying AI are pulling ahead while more workers fall behind. As of Q3 2025, US productivity grew roughly 2 percent from a year earlier — above post-pandemic averages — but the labor share of income had fallen, meaning workers were not capturing the full value of their productivity gains.
The labor share was 54.4 percent in Q4 2025. A declining labor share means that nominal wages have not kept pace with productivity growth plus inflation. Workers are generating more output but capturing proportionally less of its value, with the difference accruing to capital. As PIMCO noted, a lower labor share has important implications for aggregate demand: productivity gains that primarily flow to capital rather than labor boost corporate profits and asset prices but provide less direct support to consumer spending, which constitutes approximately 70 percent of US GDP.
For organizations making hiring and budget decisions, the productivity reading matters in a specific and practical way that is often overlooked. Productivity growth means more output from the same labor input. At the aggregate level, that implies that a given growth target can be achieved with fewer additional hires than a lower-productivity baseline would require.
If the economy is growing at 2 percent annually with 2.1 percent productivity growth, gross hiring needs are significantly lower than if it is growing at 2 percent with 1.5 percent productivity growth. The difference — roughly 0.6 percentage points of productivity — compounds over time into materially different labor market dynamics.
The BLS current business cycle data captures this precisely. From Q4 2019, nonfarm business output has grown at 2.6 percent annualized while hours worked have grown at only 0.4 percent. The gap — which is the productivity growth — means the economy has been generating more output without proportionate labor input expansion. That dynamic directly explains the combination of strong GDP growth and tepid hiring that has characterized much of the period since 2022, and which the Bankrate economic survey described as one of the strangest economic moments in recent memory.
For organizations evaluating whether to add headcount, the productivity signal has a specific implication: if above-trend productivity growth is structural, the threshold for adding a marginal full-time employee rises. The same growth objective can be achieved with more technology investment and fewer incremental workers, which is the organizational analog of the macroeconomic pattern in the aggregate data. This calculus is increasingly explicit in sectors where AI adoption is furthest advanced, and the CFO survey data showing expected productivity doubling in high-skill services through 2026 suggests it will become more explicit across more sectors over the next several years.
Measure | Current Reading | Historical Context |
Nonfarm business productivity, 2025 annual average | 2.1% | Long-run post-war average: 2.2%; Previous cycle (2007–2019): 1.5% |
Current business cycle annualized rate (Q4 2019–Q4 2025) | 2.1% | Above previous cycle; approximately at long-run average |
Q3 2025 quarterly rate (annualized) | 4.9% | Strongest since Q3 2023 |
Q4 2025 quarterly rate (annualized) | 1.8% | Deceleration; unit labor costs rose 4.4% |
Unit labor costs, full year 2025 | +2.4% year-over-year | Consistent with 2% inflation target if sustained |
Labor share, Q4 2025 | 54.4% | Declining; productivity gains not fully accruing to workers |
Manufacturing productivity, 2025 annual | Below nonfarm business | Structural divergence; tariff disruption significant |
Cleveland Fed probability of high-growth regime | ~40% | Tentative support; not conclusive |
Several data points will determine whether the current productivity readings represent the beginning of a structural acceleration or a cyclical fluctuation that will revert.
Sustained quarterly readings of above 2 percent annualized nonfarm business productivity across multiple consecutive quarters, combined with continued sector-level evidence of AI-driven efficiency gains in services and finance, would increase the Cleveland Fed model's probability of a high-growth regime toward and potentially beyond 50 percent. That would represent the kind of evidence that would change Fed policy frameworks and wage growth tolerance materially.
A reversion of annual average productivity growth toward the 1.4 to 1.5 percent range of the previous business cycle, accompanied by continued compensation growth at 4 to 5 percent, would produce unit labor cost pressures that are inconsistent with the Fed's 2 percent target and would require either tighter monetary policy or an acceptance of persistent above-target inflation. That scenario replicates the misdiagnosis risk of the early 1970s.
The manufacturing productivity divergence warrants specific tracking. A recovery in manufacturing productivity toward the nonfarm business rate would indicate that the tariff-induced supply chain disruption is temporary and that the sector is achieving the efficiency gains from reshoring investment. A persistent gap would indicate that the productivity benefits of the current expansion are structurally concentrated in services, with manufacturing operating in a different regime.
The labor share of income — which the BLS reports quarterly as part of the productivity release — is a leading indicator of whether productivity gains are translating into wage growth or corporate profit expansion. A sustained decline in labor share alongside rising productivity would signal the K-shaped dynamic that PIMCO identified, with implications for aggregate demand sustainability that eventually return as a constraint on the expansion.
Productivity is the most consequential economic variable that receives the least proportionate attention relative to its significance. The current readings — 2.1 percent average growth for 2025, 2.1 percent annualized over the current business cycle, a 40 percent probability of a structural high-growth regime from the Cleveland Fed's regime-shift model — are not dramatic. They are precisely in the range that matters most for whether the current expansion is sustainable and whether the labor market can continue to operate close to full employment without reigniting inflation.
The historical context is clear: every major shift in the economic backdrop since 1947 — the post-war golden age, the 1970s stagflation, the 1990s boom, the 2000s stagnation — had productivity dynamics at its core. Policymakers who correctly identified the productivity regime and adjusted accordingly made better decisions. Those who misdiagnosed it — assuming the 1970s productivity slowdown was temporary, or failing to recognize the 1990s acceleration early enough — made predictable errors with lasting consequences.
The current question is whether AI-driven efficiency gains are beginning to produce the aggregate productivity acceleration that micro-level task experiments suggest is coming, or whether the historical pattern of long lags between technology adoption and aggregate productivity measurement means the payoff is still years away. The data does not yet answer that question definitively. But watching for the answer in the BLS quarterly releases, in unit labor cost trends, and in the sector-level evidence of productivity gains diffusing beyond software and finance into the broader service economy is among the most important things an economically informed observer can be doing in 2026.
All data cited in this article is drawn from primary sources including:
US Bureau of Labor Statistics Productivity and Costs release: Fourth Quarter and Annual Averages 2025, Revised (March 24, 2026), US Bureau of Labor Statistics Productivity and Costs release: Third Quarter 2025, US Bureau of Labor Statistics Productivity Home Page and FRED series OPHNFB and PRS85006092, Federal Reserve Bank of Cleveland Economic Commentary: Cline, Kahn, and Rich, Is High Productivity Growth Returning? (2025), Federal Reserve Bank of San Francisco Economic Letter: Fernald, The Recent Rise and Fall of Rapid Productivity Growth (2015), Federal Reserve Bank of Atlanta Business Inflation Expectations CFO Survey on AI productivity (2026), US Bureau of Labor Statistics Monthly Labor Review: The US Productivity Slowdown (2021), Congressional Research Service: Productivity Growth Trends and Policy Issues (September 2025), PIMCO Macro Signposts: Why US Productivity Gains No Longer Reach Workers (February 2026), Goldman Sachs research analysis: No meaningful relationship between AI and productivity at the economy-wide level (2026 earnings season analysis), Fortune: Goldman Earnings AI Anxiety (March 2026), International Center for Law and Economics: AI, Productivity, and Labor Markets: A Review of the Empirical Evidence (February 2026), European Central Bank: AI and the Euro Area Economy (March 2026), Acemoglu productivity modeling research (2025), OECD AI productivity estimates, NBER Macroeconomics Annual: Fernald, Productivity and Potential Output before, during, and after the Great Recession, Charles Schwab 2026 Fixed Income Outlook, and Bankrate Economic Indicator Survey (2026).
