FRED: Business Sector Labor Productivity (OPHNFB)
- Rows
- 318
- Columns
- 2
- Data range
- 1947-01-01 to 2026-04-01
- Auto-refresh
- Re-checks its source every 14 days
- Last refreshed with new data
- 2026-10-10
Quarterly output per hour in the nonfarm business sector. Rising productivity relative to employment is a key indicator of software/AI-driven labor substitution over time.
AI analysis
Dataset Analysis: FRED Business Sector Labor Productivity (OPHNFB)
1. Dataset Overview & Research Value
This dataset captures quarterly nonfarm business sector output per hour — a foundational macroeconomic indicator published by the Federal Reserve Bank of St. Louis (FRED), one of the most authoritative and methodologically consistent economic data sources available. The source URL (fred.stlouisfed.org) confirms this is a direct API pull from FRED, lending high credibility and suggesting the data is regularly refreshed through early 2026. Spanning nearly eight decades (1947–2026), the dataset is exceptionally well-suited for long-horizon correlation studies — particularly those examining how technology waves (mainframes in the 1960s, PCs in the 1980s, internet in the 1990s, and AI/software automation from the 2010s onward) have structurally shifted labor productivity. The "software influence" framing is analytically compelling: productivity gains that outpace employment growth are a classic empirical signature of labor-substituting automation.
---
2. Data Quality Observations
Data quality here is exemplary for a time-series dataset of this length. Across both columns and all 317 rows, there are zero null cells, which is rare for a dataset covering 79 years of quarterly observations and suggests consistent measurement methodology or reliable backfill by the BLS/FRED pipeline. The OPHNFB column contains 316 distinct values against 317 rows, implying at most one duplicate value (not a duplicate row, but a repeated index reading) — a negligible concern that likely reflects coincidental rounding at two different time points rather than a data integrity issue. No type mismatches are indicated, and the observationdate column has 317 distinct entries perfectly matching the row count, confirming no duplicate timestamps. This dataset is essentially ready for analysis without preprocessing.
---
3. Key Column Distributions & Statistical Patterns
The sole metric column, OPHNFB, tells a rich story through its distribution statistics. The range of 22.065 to 119.437 reflects nominal index growth (not percentage change), so the absolute spread is expected and meaningful. The mean of 61.27 versus a median of 55.52 reveals a moderate positive skew (0.464), consistent with accelerating productivity gains in later decades pulling the distribution rightward — the series is not symmetric around a central tendency but rather shows compounding upward drift. The interquartile range (Q1=38.93, Q3=85.32) spans 46.4 index points, suggesting the middle 50% of observations covers a substantial portion of the productivity journey. The standard deviation of 27.47 is large relative to the median, further confirming a strong secular trend rather than a stationary process — standard correlation methods should be applied to first-differenced or growth-rate transformed values to avoid spurious correlations driven by shared trending behavior.
---
4. Recommended Join Key for Cross-Dataset Correlation
The observationdate column is the clear and only viable join key, and it is well-structured: 317 distinct quarterly timestamps with no nulls or gaps apparent. For joining, dates should be standardized to quarter-start format (e.g., YYYY-Q1 = January 1) since FRED uses the first day of each quarter. Any partner dataset should be aggregated or filtered to quarterly frequency before joining to avoid many-to-one mismatches. For datasets with monthly frequency (e.g., CPI, employment), a quarterly average or end-of-quarter snapshot transformation is recommended prior to the join.
---
5. Suggested Complementary Datasets for Correlation Discovery
Several dataset pairings would unlock high-value correlation analyses:
- FRED Total Nonfarm Payrolls (PAYEMS) — directly tests the labor substitution hypothesis; divergence between productivity growth and employment growth is the core AI/automation signal - FRED Real GDP (GDPC1) — decomposes how much of output growth is attributable to productivity vs. labor input expansion - BLS Software & IT Sector Employment (CES5051) or IT Investment (from BEA Fixed Assets) — correlates software capital deepening with productivity inflections, particularly post-1995 and post-2015 - FRED Federal Funds Rate or 10-Year Treasury (DGS10) — examines whether productivity cycles align with monetary policy regimes - Patent grant data (USPTO) or R&D expenditure (FRED: Y694RC1Q027SBEA) — tests whether innovation input leads productivity output with a measurable lag, useful for Granger causality analysis
Columns
- observation_date (date)
- OPHNFB (decimal)