Measurement of the labor share of national income is inherently ambiguous.
There is no single “correct” number because national accounts do not cleanly separate all income into pure labor versus pure capital categories. Economists and statistical agencies (BLS, BEA, Federal Reserve) therefore make assumptions, and different choices produce different levels and sometimes different trends.
Core Definition and Official Series
Labor share is generally compensation of employees (wages, salaries, and employer-paid benefits such as health insurance, pensions, and social insurance contributions) divided by a measure of total income or output.
- BLS official series (most commonly cited for the recent lows around 52.9–54%): Labor compensation as a share of current-dollar output in the nonfarm business sector. This covers roughly three-quarters of the economy. It excludes general government, private households, nonprofit institutions, the Armed Forces, and farms. BLS adjusts for proprietors by assuming they earn the same average hourly compensation as employees in the same sector (“labor approach”).
- BEA/NIPA measures: Often use compensation of employees as a share of national income, gross domestic income, or GDP. These can include or exclude more sectors.
- Corporate-sector measures: Compensation of employees in the corporate sector divided by corporate value added (minus certain taxes). These avoid many mixed-income problems and are preferred by some researchers for cleaner analysis of firm-level dynamics.
Recent headlines (Axios, Reuters, Fed analyses) referring to ~53–54% lows typically track variants of the BLS nonfarm business or similar gross measures.
Major Measurement Nuances and Adjustments
1. Proprietors’ / mixed / self-employment income: This is the largest and most debated issue. Proprietors (sole proprietors, partners) receive a single stream of income that mixes returns to their labor and their capital.
- Options: (a) treat all as capital (understates labor share); (b) assume proprietors earn the same average wage as employees and impute the rest as capital (“same wage” approach used by BLS); (c) assume the labor share is the same for proprietors as for the corporate sector (“same share”); (d) other imputations based on hours or assets.
- Impact: The choice affects both the level and part of the observed decline. Some studies attribute roughly one-third of the measured post-1980s/2000s decline to the particular imputation method plus the falling share of self-employment hours. Tax incentives (especially for S-corporations and pass-through entities) also encourage reclassification of labor income as capital income, which can further understate the true labor share.
2. Gross vs. net (depreciation): Most headline figures use
gross measures (before subtracting capital consumption/depreciation). Depreciation is a real cost of production that is not spendable income for anyone.
- Net labor share (labor compensation / net value added or net national income) is substantially higher—recent calculations put it around 64% when the gross figure is ~54%.
- The trend of decline usually remains under both, though the magnitude can differ depending on how depreciation rates have evolved (especially with faster obsolescence of IT and intangible capital).
3. Housing / owner-occupied housing: National accounts impute rental income to homeowners as if they were renting to themselves. Almost all of this is treated as capital income.
- Including housing lowers the measured labor share. Some argue it should be excluded because returns to owner-occupied housing largely accrue to middle-class households (workers) rather than pure “capitalists,” and because the imputation involves significant measurement error.
- Housing’s rising share of capital income has been an important contributor to the overall capital share in some decompositions (Rognlie and others).
4. Government sector: Government value added is mostly labor compensation (public employees). Including it raises the overall labor share. Capital income in the government sector is conceptually awkward (what is the “return” on public capital?). Many preferred measures focus on the private or nonfarm business sector partly for this reason.
5. Taxes on production and imports (less subsidies): These sit between labor and capital. Common practice is to allocate them proportionally or exclude them from the denominator so the shares of “factor income” sum closer to 100%. Different treatments change levels modestly.
6. Intellectual property products (IPP) and intangibles: BEA reclassifications (software, R&D, etc., treated as investment rather than intermediate inputs) increased measured value added and capital income without changing measured labor compensation. This mechanically lowered the labor share. Some researchers argue a sizable portion of the apparent decline is an accounting artifact from better capitalization of intangibles.
7. Other issues
- National vs. domestic income; inclusion of net foreign factor payments.
- “Factorless income” residual (the gap after assigning measured labor payments and imputed capital returns)—interpreted variously as pure economic profits/markups, unmeasured capital, or measurement error.
- Within-labor inequality: The aggregate labor share treats a CEO’s compensation the same as a production worker’s; it does not speak directly to distribution among workers.
- Sectoral composition: Much of the aggregate decline occurs within industries rather than solely from shifts toward low-labor-share sectors.
Practical Implications
- Different reasonable measures produce levels that can differ by 5–10+ percentage points.
- The direction of the long-term decline (especially since the early 2000s, with further post-COVID deterioration) is robust across most variants, though the exact size of the drop is sensitive to assumptions.
- Corporate-sector or net measures often show a clearer or cleaner decline than the broadest aggregates.
- Researchers such as Loukas Karabarbounis emphasize constructing multiple series (e.g., “Proprietors Same Share,” “Proprietors Same Wage,” pure corporate) and examining sensitivity. Justin Wolfers and others highlight the gross/net distinction and reclassification effects as important caveats to headline numbers.
In short, the recent lows are real under standard official definitions, but the precise number depends heavily on how mixed income, depreciation, housing, government, and intangibles are handled.
The structural forces (technology, globalization, market power, bargaining) that economists debate operate on top of these measurement choices.
When evaluating claims about “record lows,” it is useful to ask which exact series is being used and what adjustments (if any) have been applied.