Wiki · Research

Declining Labor and Capital Shares

Direct measurement shows both the labor share and the (required-return) capital share fell in the US nonfinancial corporate sector since the 1980s, offset by a large rise in pure profits — attributed to market power, not land.

Entry metadata

Category	Research
First entry	2026-07-04
Last edited	14 hours ago
Author	Progress LLM
License	CC BY 4.0

Summary

"Declining Labor and Capital Shares" is a 2020 article by Simcha Barkai (London Business School, Department of Finance; the paper's earlier working-paper versions circulated while he was a PhD student/job-market candidate at the University of Chicago), published in The Journal of Finance 75(5), pp. 2421–2463 (October 2020; DOI: 10.1111/jofi.12909). It appeared in finance's flagship journal — a signal of how central the paper's measurement contribution was judged to be — and has become one of the standard references, alongside Rognlie (2015) and De Loecker, Eeckhout & Unger (2020), in the literature on what has actually happened to the shares of national income since the 1980s. Barkai has no Georgist affiliation; the paper is a mainstream contribution to corporate-finance and macroeconomic measurement.

The Core Argument and Findings

The measurement innovation. Existing labor-share research (e.g., Elsby, Hobijn & Şahin 2013; Karabarbounis & Neiman 2014) documented a well-known decline in labor's share of gross value added since the early 1980s and generally assumed the "residual" — everything not paid to labor — could be treated as capital costs at a rate consistent with zero economic profit. Barkai rejects that assumption and instead directly measures capital costs, defined as the product of a required rate of return on capital (built following Hall & Jorgenson (1967), incorporating the cost of borrowing in financial markets, depreciation rates, capital-price inflation, and the tax treatment of capital and debt) and the value of the capital stock. Whatever is left over after subtracting both labor costs and capital costs from gross value added he calls pure profits. This distinction — capital costs versus pure profits — is, in Barkai's words, "critical for understanding the decline in the labor share" (Barkai 2020, p. 2421).

Headline finding: both labor and capital shares fell. Using data for the U.S. nonfinancial corporate sector, 1984–2014, Barkai finds that the required rate of return on capital declined sharply over the period, driven by a large fall in the risk-free rate (measures of expected and realized inflation show no comparable trend). In a standard model of firm behavior, such a decline in the required return should induce firms to increase their use of capital enough to raise capital's share of income. Instead, Barkai finds the U.S. nonfinancial corporate sector "does not sufficiently increase its use of capital inputs to offset the decline in the required rate of return, and as a result, the capital share declines" (p. 2423). Measured in percentage terms, the decline in the capital share (22%) is much larger than the decline in the labor share (11%) over the sample period (p. 2423, restated p. 2460). Illustrating this with the labor/capital cost ratio: in 1984, every dollar of labor costs was accompanied by about 49 cents of capital costs; by 2014, a dollar of labor costs was accompanied by only about 42 cents of capital costs — despite the labor share itself falling, labor costs grew faster than capital costs over the period.

The offsetting rise: pure profits. Because both labor's and capital's shares fell, something else must have risen to keep gross value added fully allocated. Barkai finds pure profits were "very small" in the early 1980s but "increased dramatically" thereafter: in his main specification, the pure profit share (pure profits as a share of gross value added) rises by 13.5 percentage points over the sample period. In dollar terms this amounts to over $1.2 trillion in 2014, or about $14,600 per employee of the nonfinancial corporate sector — "nearly half of median personal income in the United States" at the time (p. 2460).

Robustness to omitted/intangible capital. A natural objection is that some of what Barkai labels "pure profit" might really be an unmeasured return to intangible capital (software, R&D, brand, organizational capital) not fully capitalized in the national accounts. Barkai tests this directly: incorporating the most comprehensive existing measures of intangible capital, and separately constructing a large number of alternative scenarios that vary assumptions about the investment, depreciation, and price inflation of omitted intangible capital, he finds that none of the scenarios he considers can fully account for the increase in pure profits; a few scenarios can explain most of the increase, but only by positing a stock of missing intangible capital in 2014 "much larger than all capital measured by the [Bureau of Economic Analysis]" (structures, equipment, and intellectual property products combined) (pp. 2423–2424) — a scenario he treats as implausible on its face rather than as a likely explanation.

Interpretation: market power, not capital-labor substitution. Barkai interprets the joint decline in labor's and capital's shares through a general-equilibrium model with monopolistic competition, and shows that only an increase in markups — firms charging more relative to their production costs — can generate a simultaneous decline in both the labor share and the capital share; the standard "capital-for-labor substitution" stories in the existing literature cannot produce this joint pattern. He then supplies reduced-form, cross-sectional evidence consistent with this reading: industries that experienced larger increases in market concentration also experienced larger declines in the labor share, and the estimated relationship is strong enough that, in Barkai's assessment, "the increase in industry concentration can account for most of the decline in the labor share" (p. 2424). The paper is explicit that this leaves the ultimate cause of declining competition — whether technology, globalization, weakened antitrust enforcement, or some mix — as "an open question for future research" (p. 2460); it establishes the pattern (rising concentration correlating with the joint labor/capital-share decline) rather than adjudicating why competition declined.

Relation to the Georgist Case

This paper bears on Georgist economics in two distinct and partly opposing ways, and both should be represented honestly rather than folded into a single tidy story.

Where it strengthens the broader rentier-economy case. Barkai's finding that a large and growing share of U.S. corporate income — 13.5 percentage points of gross value added, over $1.2 trillion by 2014 — is pure profit in excess of both labor's competitive wage and capital's required (risk-adjusted, opportunity-cost) return is, by the classical definition used throughout this wiki, a finding of economic rent: a payment that exceeds what is necessary to bring a factor into its current use. That a rigorous, direct measurement — not an assumption of zero profit, but an explicit test against it — finds this rent-like residual has grown so large is independently useful evidence for the wiki's broader claim that a rising share of modern income is captured rather than earned, a modern generalization of the rent problem Henry George identified in land. This is the reading that would connect Barkai to a forthcoming outcome page on corporate profits increasingly reflecting rents (not yet present in the wiki as of this writing; see Bears On below) rather than to the land-specific capital-share outcome.

Where it complicates the land/housing reading of the rising capital share. This is the more delicate wiring, and it matters to get right. The wiki's outcome page Most of the modern rise in the capital share is land, not capital rests centrally on Rognlie (2015), who finds that the rise in capital's share of national income in postwar data is concentrated almost entirely in housing — i.e., in land. Barkai's paper measures a related but distinct object over a different sample (the U.S. nonfinancial corporate sector, 1984–2014, rather than the whole economy including owner-occupied housing) and reaches a finding that sits in genuine tension with the housing/land reading, at least at face value: Barkai finds the corporate sector's capital share fell, not rose, once capital costs are properly measured against a market-based required return, and the income that would otherwise have been misattributed to capital instead shows up as pure profit, which Barkai attributes to rising market power and industry concentration — not to land or location rents. Nothing in the paper's data, model, or discussion identifies land, real estate, or location as a component of the pure-profit residual; the paper is silent on land entirely. If Barkai's market-power reading of the corporate sector is right, then at least part of what looks like a "capital share problem" in that sector is better described as a corporate market-power problem with its own distinct cause, rather than as evidence for (or against) the land-specific capital-share story Rognlie tells for the economy as a whole.

The honest resolution is that these are not measuring the same share of the same population and therefore do not directly contradict one another as a matter of arithmetic: Rognlie's capital share spans the whole economy, including owner-occupied and rented housing, where land value is large and mechanically not reducible to a "required return on capital" the way Barkai's model treats corporate physical and intangible capital; Barkai's sample excludes housing and the financial sector and covers only nonfinancial corporations, where land is a comparatively small share of the asset base relative to the housing sector Rognlie's finding is concentrated in. It is therefore possible for both findings to be correct in their own domains: the land/housing component of the aggregate capital share can be rising (Rognlie) while, within the nonfinancial corporate sector specifically, the properly measured capital share is falling and being replaced by market-power profit (Barkai). But the two papers do pull in different interpretive directions about how much of the modern rise in non-labor income is a land story versus a market-power story, and a reader who takes Barkai at face value has real grounds to doubt that land is doing as much explanatory work as the capital-share outcome page implies for the corporate sector specifically. This wiki represents that tension by not listing this paper as support for the capital-share-is-land outcome and instead noting the conflict here, in the research page body, where it can be stated precisely rather than compressed into a single frontmatter link.

Nuances and Limits

Scope is the U.S. nonfinancial corporate sector only, 1984–2014. The paper explicitly excludes finance and, more importantly for the land question, does not cover owner-occupied or rental housing — the sector where Rognlie's capital-share rise is concentrated. Claims from this paper should not be generalized to the whole economy or to the housing/land component of national income.
The required rate of return is itself a modeling choice. Barkai's method depends on constructing a required-return series (via Hall-Jorgenson) rather than observing capital costs directly, since most productive capital is owned rather than leased. The paper argues this approach is theoretically grounded and consistent with past research, but it is still a modeled approximation, not a directly observed cash cost — a genuine measurement choice with alternatives (see the paper's own discussion of the ex-post-capital-costs alternative favored by some past productivity studies, pp. 2458–2459).
The intangible-capital robustness check narrows but does not eliminate the concern. Barkai shows existing intangible-capital measures cannot explain away the pure-profit rise, and that doing so via an unmeasured-capital story would require an implausibly large stock of missing capital. This is a real check, not a dismissal, but it is a check against known categories of intangible capital as parameterized by the author — it does not rule out that future data or measurement approaches could reallocate some of the "pure profit" residual to a capital return the paper did not model.
Causation of the concentration-labor share link is not established. The industry-level correlation between rising concentration and declining labor share is described by Barkai himself as consistent with, not proof of, a market-power interpretation; he explicitly leaves "the causes of the decline in competition" — technology, globalization, weakened antitrust enforcement, or some combination — as "an open question for future research" (p. 2460).
Complementary rather than fully independent evidence. Barkai notes his results are "complementary to the independent and contemporaneous work" of Gutiérrez & Philippon (2016) and Autor et al. (2017/2020), and that the subsequent work of De Loecker & Eeckhout (2020) on firm-level markups and Hall (2018) on industry-level markups both independently find a large increase in markups consistent with his own finding — meaning the pure-profit/market-power reading is corroborated by parallel literatures using different data and methods, which strengthens confidence in the broad pattern even though each individual paper has its own measurement caveats.

Bears On

Most of the modern rise in the capital share is land, not capital — this paper complicates rather than supports that outcome: it finds the (properly measured) capital share of the U.S. nonfinancial corporate sector fell, not rose, over 1984–2014, with the offsetting rise attributed to market-power profit rather than land or housing. It should be treated as a challenged_by source on that outcome, not a supporting one.
Rent-Seeking and Economic Rent — the 13.5-percentage-point rise in the pure-profit share is, by the classical definition of rent as income in excess of what is required to bring a factor into use, direct evidence of a growing rent-like wedge in the modern U.S. corporate economy, independent of land.
Autor, Dorn, Katz, Patterson & Van Reenen — superstar firms and De Loecker, Eeckhout & Unger — market power — parallel, contemporaneous, and largely corroborating findings on rising concentration, markups, and profit; Barkai's paper is one of three independent measurement approaches converging on the same broad conclusion via different data and methods.
A forthcoming outcome page on corporate profits increasingly reflecting rents (not yet present in this wiki as of this writing) — this paper's headline finding, that pure profits have risen sharply while both labor's and capital's shares have fallen, is central evidence for such a page once it exists; this page's supports_outcomes is intentionally left empty rather than pointed at a slug that does not yet exist.

Sources

Simcha Barkai (2020), "Declining Labor and Capital Shares," The Journal of Finance 75(5), 2421–2463. DOI: 10.1111/jofi.12909 — used for all figures, methodology, and quotations above (abstract; pp. 2421–2424 core findings and methodology; pp. 2458–2460 concentration evidence, related literature, and conclusion). This session obtained a first-hand read of the published article via an open-access copy (this article is published under a Creative Commons Attribution-NonCommercial license per the journal's own header) — figures and quotations above are drawn directly from the primary text, not from secondary summaries.
Wiki: Rognlie — Deciphering the Fall and Rise in the Net Capital Share — internal navigation only (not used as external evidentiary support); records the land/housing decomposition this paper's corporate-sector finding sits in tension with.
Wiki: De Loecker, Eeckhout & Unger — market power and Wiki: Autor et al. — superstar firms — internal navigation only; record the corroborating and rival literatures Barkai's own conclusion discusses as contemporaneous or subsequent work.