Nonfarm business labour productivity (TTM)
Trailing twelve-month nonfarm business output per hour, year-on-year change.
Reading
The single statistic that, if it actually shows up in the aggregate data on a multi-year basis, materially eases the diagnosis. The honest counterweight: the bull case requires this number to deliver.
Thresholds
- watch3 — sustained above 3% — would represent a genuine AI productivity dividend
Context
Why this matters
Total nonfarm-business labour productivity is the single statistic that, if it actually shows up in the aggregate data on a multi-year basis, materially eases every other signal in the diagnosis. Faster productivity growth means real GDP grows faster, which expands the denominator of debt/GDP, eases the worker-to-beneficiary arithmetic in real terms, and supports the corporate margin assumption underwriting current equity valuations. The bull case for the entire diagnosis sits on this number delivering.
Who watches this
- BLS Productivity & Costs team — publishes the canonical series
- Larry Summers — has framed productivity as the single most important macro variable for the 2020s
- Tyler Cowen (GMU) — The Great Stagnation and successor work places productivity at the centre of the diagnostic question
- Erik Brynjolfsson (Stanford) — academic work on whether IT and AI investment translates into measured productivity (the "productivity J-curve")
- Jan Hatzius / Joseph Briggs (Goldman Sachs) — Goldman's pro-AI productivity case; have published the most-cited bull-side estimates
Recent history
Productivity readings since 2022 have been volatile — quarterly prints have at times accelerated into 3%+ year-on-year ranges before reverting. The relevant question is multi-year average, not single quarters. Through 2024–25 the trailing average has remained inside the post-2008 band of roughly 1–2%, despite the AI-narrative framing.
What would change my read
Sustained productivity growth above 3% for 3+ years would be a genuine regime shift — the kind that historically has accompanied resolution of similar fiscal-arithmetic problems (postwar US, post-Bretton-Woods adjustment). This is the bull case's load-bearing requirement; everything else hinges on whether the number actually arrives.