The U.S. economy grew in 2025, but the gains came increasingly from deploying more resources rather than using them more cleverly. The latest BLS annual productivity report reveals a two-year deceleration in total factor productivity, the economy’s deepest measure of efficiency. Beneath the headline, a surge in software and R&D investment is quietly reshaping how American businesses grow — even as traditional productivity gains moderate. Assistance from Claude AI.
Source: U.S. Bureau of Labor Statistics Report: Total Factor Productivity — 2025 Annual Released: March 19, 2026 Reference Period: Full Year 2025 (Preliminary)
The Headline Numbers
The Bureau of Labor Statistics reported on March 19, 2026 that private nonfarm business sector total factor productivity (TFP) grew 0.8 percent in 2025 — a meaningful step down from 1.5 percent in 2024 and part of a two-consecutive-year deceleration that deserves close attention.
Real output in the private nonfarm business sector rose 2.6 percent, down from 3.0 percent in 2024 and continuing the same downward trajectory. This number tells us the economy kept expanding — production genuinely increased — but the pace of that expansion is cooling.
Labor productivity — the more familiar measure of output per hour worked — rose 2.2 percent in 2025, compared to 3.0 percent in 2024. That 0.8-percentage-point drop was the second-largest single-year decline in this current business cycle. Worker output is still improving, but the rate of improvement is slowing.
Combined inputs (capital plus labor together) rose 1.7 percent, actually rebounding slightly from 1.4 percent in 2024. That rebound is worth noting: it tells us businesses were pumping more resources into production in 2025, yet still getting less output-per-dollar than the year before — the definition of an efficiency slowdown.
Capital input alone grew 2.7 percent, maintaining the same steady pace it has held for the past four years. The economy is not pulling back on investment; it is investing just as heavily as before, but the productivity payoff from that investment is arriving more slowly.
No consensus analyst forecast is published for TFP in the same way that jobs or inflation figures are; this is a less-watched but arguably more consequential annual data release with no conventional “beat/miss” benchmark.
What This Actually Means
Think of the economy like a restaurant kitchen. You can produce more meals in two ways: hire more cooks and buy more stoves, or reorganize the workflow so the same crew turns tables faster. Total factor productivity measures only the second kind of gain — the smarter-not-just-bigger improvements.
In 2025, the American economy kept growing, but an increasing share of that growth came from the first method: hiring slightly more, and investing more capital. The share coming from pure efficiency improvements — new techniques, better software, smarter processes — shrank.
This matters because efficiency-driven growth is the good kind of growth. It raises living standards without requiring proportionally more resources, energy, or labor. If an economy can only grow by consuming more inputs, that growth eventually runs into physical limits. If it grows by becoming more productive with the same inputs, that potential is much more open-ended.
The 0.8 percent TFP figure in 2025 is not alarming on its own — it is positive, and it remains above the long, disappointing stretch of 0.6 percent annual TFP growth that defined the 2007–2019 business cycle. But the back-to-back deceleration from 3.5 percent in 2021 down to 0.8 percent in 2025 is worth watching, particularly as businesses are making enormous investments in AI and automation that have not yet fully shown up in the productivity numbers.
Key Internals and Nuance
The TFP contribution to labor productivity nearly halved year-over-year. In 2024, TFP accounted for 1.5 percentage points of that year’s 3.0 percent labor productivity growth — the single largest driver. In 2025, TFP’s contribution fell to just 0.8 percentage points out of 2.2 percent total labor productivity growth. Capital intensity (the capital-per-worker ratio) contributed 0.9 percentage points, and labor composition (the shifting skill and education mix of workers) contributed 0.4 points. The efficiency engine did not stall, but it downshifted significantly.
Intellectual property is now the dominant form of capital investment. Within the 2.7 percent growth in capital input, intellectual property products — software, research and development, and artistic originals — accounted for more than half of that growth, contributing 1.6 percentage points. In 2019, intellectual property contributed 1.4 points out of a larger 3.0 percent capital input total, representing less than half. This structural shift means businesses are increasingly investing in ideas and code rather than machines and buildings, which makes intuitive sense given the tech and AI investment wave of recent years.
Software and R&D within IP are both accelerating. Software capital grew at roughly 9.6 percent in 2024 (the most recent year for detailed capital data), and R&D grew at 6.3 percent. These are the fastest-growing components of the entire capital stock by a significant margin. The puzzle is that this investment surge has not translated into proportionally higher TFP — suggesting either that the gains are still building in a pipeline, or that measurement challenges are obscuring true productivity in digital-intensive industries.
Equipment investment is quietly fading. Physical equipment — machines, vehicles, computers — has gone from contributing 1.1 percentage points to capital growth in 2019 to just 0.5 percentage points in 2024. Within equipment, information processing equipment (computers, communications gear) is still the largest contributor but has decelerated. Traditional physical machinery is playing a shrinking role in how the economy grows.
Hours worked recovered, barely. Hours worked in the private nonfarm business sector fell 0.1 percent in 2024 — an unusual contraction — and then rebounded to a 0.4 percent increase in 2025. This is a modest recovery, not a surge. It suggests the labor market added modestly to economic capacity in 2025 but was not the primary driver of output growth.
Important data caveat: The 2025 TFP figures are explicitly preliminary. The BLS uses only 6 aggregated asset categories for the most recent year, compared to the full 90-asset breakdown used for all prior years. This means the capital input estimate for 2025 is an approximation, and the headline 0.8 percent TFP figure could be revised — in either direction — when complete data become available.
Trend Context
Zoom out to see the full picture. TFP growth in the current business cycle (2019–2025) is averaging 1.0 percent per year, which genuinely outpaces the prior cycle (2007–2019) average of 0.6 percent. So the current era is meaningfully better than the post-financial-crisis decade, when productivity growth was described by many economists as a “mystery” and a crisis in slow motion.
But within this cycle, the trend has been unmistakably downward since the 2021 reopening surge. TFP peaked at approximately 3.5 percent in 2021, driven by the extraordinary efficiency gains of businesses adapting to COVID constraints and the rapid reallocation of labor to more productive uses. It then went negative in 2022 as those one-time gains faded and supply chains normalized. It recovered to 1.0 percent in 2023 and reached 1.5 percent in 2024 before dropping back to 0.8 percent in 2025.
That 2024 peak now looks, in retrospect, like a partial recovery from 2022’s dip rather than a new upward trajectory. The 2025 deceleration extends the pattern and raises a legitimate question: is the post-pandemic productivity boost fully exhausted? Or is there a second wave building as AI investments mature?
The long-run comparison is instructive. During the dot-com buildout of the 1990s (1990–2000), TFP averaged 0.9 percent annually. During the tech diffusion era of 2000–2007, it peaked at 1.3 percent. The 2007–2019 cycle was the low point at 0.6 percent. The current cycle at 1.0 percent sits between the dot-com era and the pre-crisis peak — which is actually a decent outcome given everything that has happened since 2020.
What Economists and Analysts Are Saying
Note: This report was released March 19, 2026. The following represents the expected range of analytical interpretation based on established economic frameworks rather than quotes from specific post-release commentary.
The optimistic view holds that the current deceleration is a normal mean reversion after the extraordinary 2021 post-pandemic productivity surge, and that the massive wave of AI and software investment now underway will generate a productivity dividend over the coming years — just as it took roughly a decade for the internet to show up in measurable productivity gains after its commercial rollout. Economists like Erik Brynjolfsson at Stanford have argued that AI’s productivity effects are real but will arrive in measured data with a substantial lag.
The pessimistic view argues that we have been here before: enormous investments in transformative technology repeatedly fail to materialize as broad productivity gains on any reasonable timeline. The “productivity paradox” — first described by Robert Solow with computers — has not been resolved even after decades of digital investment, and there is no guarantee AI will be different.
A structural concern focuses on the composition of where productivity is and isn’t improving. TFP gains tend to be concentrated in a small number of highly productive industries (tech, finance, professional services), while large employment sectors like healthcare, education, and retail continue to show minimal productivity growth. Average TFP figures can mask this bifurcation.
Watch for motivated framing: advocates for deregulation will cite any TFP slowdown as evidence that regulatory burden is strangling innovation; advocates for public investment will cite it as evidence that private R&D needs to be supplemented by government spending. Both are ideological priors looking for data support rather than conclusions the TFP data itself can definitively establish.
Policy Implications
For the Federal Reserve, productivity data matters because it shapes the “neutral rate” of interest — the rate at which monetary policy is neither stimulating nor restricting. Higher trend productivity growth generally supports a higher neutral rate, because a more efficient economy can sustain faster non-inflationary growth. A decelerating TFP, if sustained, would push the other direction — suggesting the Fed has less room to hold rates elevated without dampening growth. However, one year of deceleration is not a trend shift; the Fed will not react to this report in isolation.
For Congress and budget debates, productivity is the arithmetic foundation of long-run fiscal sustainability. The Congressional Budget Office’s long-run budget projections are built on assumptions about trend productivity growth. If TFP growth softens toward 0.6–0.8 percent (the post-financial-crisis norm) rather than the 1.0–1.5 percent of recent years, the revenue projections that underpin current fiscal policy become meaningfully more optimistic than reality. This creates quiet pressure for either spending cuts or higher revenues over time.
For AI and technology policy, the report highlights the growing IP investment wave while simultaneously showing TFP deceleration — a combination that will fuel the ongoing debate about whether AI is productively deployed or whether much of the investment is going into speculative capacity rather than efficiency-improving applications. Policymakers weighing AI regulation, antitrust, and R&D tax credits will find ammunition on both sides of the debate in this data.
What to Watch Next
The quarterly Productivity and Costs release (next due June 2026) will provide the first Q1 2026 labor productivity reading and will be the earliest signal of whether the 2025 TFP deceleration is continuing into 2026 or stabilizing. This is the same series the annual TFP report draws on for its preliminary output and hours estimates.
The BEA’s GDP release for Q1 2026 will provide the underlying output data that feeds into all productivity calculations. A significant deviation from the 2025 trend in either direction will either reinforce or complicate the TFP story.
BLS industry-level TFP data (available at bls.gov/productivity) will allow analysts to see where the efficiency gains are concentrated and which sectors are dragging the aggregate figure down. The headline number masks enormous variation across industries; the sector breakdown is where the real analytical action is.
Bottom Line
The U.S. economy kept growing in 2025, but it did so by working harder rather than smarter — deploying more capital and labor rather than extracting more from what it already has. The measure of pure economic efficiency declined for the second year in a row, from 1.5 percent in 2024 to 0.8 percent in 2025, though the current business cycle as a whole remains more productive than the disappointing post-2008 decade. The one structural bright spot is a surging investment in software and R&D that now dominates capital spending — but whether that investment will eventually unlock a new wave of productivity gains, or simply be the latest chapter in technology’s long history of overpromising and underdelivering on efficiency, remains the central open question in American economic policy.
Data source: Bureau of Labor Statistics, USDL-26-0504, “Total Factor Productivity — 2025,” released March 19, 2026. 2025 figures are preliminary estimates subject to revision.