What Do We Know About the Labor Share and the Profit Share? Part III: Measures and Structural Factors

Summary

"What Do We Know About the Labor Share and the Profit Share? Part III: Measures and Structural Factors" is Levy Economics Institute of Bard College Working Paper No. 805 (May 2014), by Olivier Giovannoni, then assistant professor of economics at Bard College and a Levy Institute research scholar. It is the third of a three-part series (Parts I and II, working papers 803 and 804, cover theory and empirical evidence; this entry covers Part III, fetched and read in full — 44 pages — directly from the Levy Institute's publication server).

The paper's project is methodological: rather than take the labor share of income as a settled statistic, it builds one of the first national (whole-economy, not just nonfarm-business) US labor share series running back to 1929, cross-checks it against alternative datasets (BLS, OECD, AMECO), and decomposes the aggregate — sectoral shifts, gender composition, price-index choice, and top-income concentration — to test whether the labor share's apparent long-run stability survives disaggregation. It largely does not.

The Core Argument / Findings

The headline result. The paper's own framing: the labor share's near-constancy is "a bit of a miracle" (quoting Keynes) but simultaneously "a bit of a mirage" (quoting Solow) once decomposed. Combining NIPA data with the Piketty-Saez World Top Incomes Database, the paper finds "the labor share for the bottom 90%, 99% and 99.9% have decidedly fallen since the early 1980s, so much so that the labor share is lower today than at any other period since 1930." The fall "ranges between 8 and 18 points of the net national income, depending on which top income category is retained" — equivalent, per the abstract, to "a transfer of $1.8 trillion from labor to capital in 2012 alone," bringing "the US labor share to its 1920s level." The paper's best aggregate (undecomposed) labor-share estimate falls only about 5 points over the same period — an order of magnitude smaller, which is the paper's central point: aggregating across the whole income distribution masks a much larger internal redistribution.

Measurement choices matter, and the paper is explicit about the tradeoffs. It measures shares against net national income (GDP minus indirect taxes, capital consumption, and statistical discrepancy, plus net foreign income) rather than gross output, since depreciation and tax shares "did not change much over the past 80 years" and mostly add noise on a gross basis. For proprietors' ("mixed") income — neither unambiguously labor nor capital — it tests five apportionment methods: Johnson's (1954) fixed two-thirds-to-labor rule; Gollin's (2002) method of assigning proprietors the average employee wage (flagged as a likely understatement, "for self-employment covers professions such as doctors, consultants, lawyers and heads of business whose income is more likely to be above the market average than below"); Gomme and Rupert's (2004) variable-weight method; a cluster-analysis method; and a regression method (proprietors' income tracks compensation, coefficient 0.68, far more than property income, coefficient 0.07). All five "give very similar results" except Gollin's. Deflating the bottom-99% labor share by the CPI instead of the GDP deflator adds roughly 20% more purchasing-power decline since 1980 — "half of that drop is due to the rise of the top 1% incomes, and the other half to a higher deflator," since the CPI weights housing and gasoline more heavily. The paper also faults the official BLS series for excluding government, nonprofits, and owner-occupied housing — "approximately 20% of national income" — and for relying on partly unpublished data.

Structural/sectoral drivers. Comparing sectoral compensation shares 1977–2007 (EU-KLEMS), the paper finds "a large compensation share drop in manufacturing (-14 points) and a large gain in finance (+18 points)... both sectors have about the same average weight in value added over the sample, so gains of one canceled out the losses from the other, on average" — a major structural reallocation invisible in the aggregate. On top incomes, an additional decomposition asks what the "property share" looks like if the top 1% is treated as economic rent: "about 5% of the increase of the property share can be ascribed to different types of income, to which one must add 10-15% due to the rise of the top 1% and their capital gains," and "the reason for this increase is not to be found in profits (corporate or noncorporate), which have remained constant." The paper's own practical judgment: "knowing that the worsening position of labor is not so much due to corporate profits increasing but rather the top 1% increasing, could be of practical importance for tax design."

Relation to the Georgist Case

This paper does not address land directly and makes no claim about land's share of the capital-share rise — a distinct empirical question the wiki covers elsewhere (see Most of the modern rise in the capital share is land, not capital, which relies on different sources such as Rognlie and Knoll-Schularick-Steger). Giovannoni's contribution is a complementary decomposition of the labor side of the same national-accounts identity: how much of the "stable aggregate labor share" story is a composition effect driven by rising top incomes booked as labor compensation, and by a sectoral reallocation of compensation away from manufacturing into finance. The finance finding bears directly on the wiki's FIRE sector analysis: an 18-point rise in finance's compensation share against a 14-point fall in manufacturing's is the kind of sectoral evidence the FIRE-critique literature (Hudson and others) points to — though the paper does not itself characterize financial-sector compensation as rent extraction, and attributing that framing to it would be an overreach. [VERIFY: whether the financial-sector compensation gains documented here are substantially rent (e.g., from land-collateralized lending) versus a return to scarce financial skill is a separate, contested question this paper does not adjudicate.] The top-1%/property-share finding is more directly rent-adjacent — the paper's own framing question is "what would the property share be if one treats the top 1% of incomes as economic rents?" — but it is offered as an illustrative accounting exercise, not an argument that top incomes are economic rent in the technical sense used elsewhere on this wiki (see Economic Rent); the paper does not test whether that income reflects genuine scarcity rent, monopoly power, or return to skill.

Nuances and Limits

• A working paper, not peer-reviewed — checked 2026-07-18, remains unpublished. The paper circulated simultaneously as Levy Economics Institute Working Paper No. 805 and University of Texas Inequality Project (UTIP) Working Paper No. 66 (dated March 3, 2014, confirmed from the UTIP PDF's own title page) — the same "Levy/UTIP" dual release Giovannoni used for this series generally. Searched IDEAS/RePEc (which lists no "published in" field for wp_805), SSRN, Google Scholar (Giovannoni's own author profile), general web search, and the Levy Institute's own publications page; found no evidence Part III (or the companion Parts I/II, working papers 803-804) was ever published in a peer-reviewed journal. Giovannoni's subsequent Levy work (e.g., "Income Distribution Macroeconomics," Levy WP 807) also remains working-paper tier. Absence of a Scholar/RePEc/SSRN listing is not proof no journal version exists anywhere, but no evidence of one was found across any of these channels. This paper should continue to be cited and weighted as an unpublished working paper — useful for methodological transparency and an unusually long (1929–2012) US series, but below peer-reviewed work in the wiki's source hierarchy.
• US-focused. International comparisons (Mexico, Canada, Japan, UK, France, Germany) show the US and Canada notably more stable than European labor shares 1960–2012, but the detailed top-income and sectoral decompositions are US-only; the paper does not explain why the US/Canadian series are unusually stable.
• Data end in 2012–13. More recent trends are not covered and would need checking against newer data before citing this paper for present-day magnitudes.
• The "top 1% as rent" framing is illustrative, not a defended identification. It is one alternative accounting lens, not an argument that top-1% income is rent in the technical sense; citing this paper to assert a specific share of the capital-share rise "is rent" would overstate what it shows.
• No causal mechanism tested. The paper documents composition effects but does not adjudicate between market-power, financialization, technology, or bargaining-power explanations for them.

Bears On

• Concept: FIRE Sector — the sectoral decomposition (manufacturing compensation share -14 points, finance +18 points, 1977–2007) is independent national-accounts evidence that compensation has shifted substantially into finance, though the paper itself does not characterize the shift as rent extraction.
• Concept: Economic Rent — the paper's "top 1% as economic rent" decomposition is a citable instance of testing that framing quantitatively, illustrating how contested the rent/non-rent line becomes once one moves past land to personal top incomes.
• Problem: Most of the modern rise in the capital share is land, not capital — this paper does not test the land-vs-capital decomposition and should not be cited as supporting or challenging that claim; it is complementary context, a different compensation-side decomposition of the same broad shift.
• Research: Barkai, "Declining Labor and Capital Shares" — both papers find pure profit or top-income concentration, rather than an aggregate capital-share rise per se, explains much of labor's loss, via different sectors, periods, and methods; complementary rather than directly corroborating on specifics.

See Also

• FIRE Sector
• Economic Rent
• Most of the modern rise in the capital share is land, not capital
• Barkai, "Declining Labor and Capital Shares"

Sources

1. Olivier Giovannoni (2014), "What Do We Know About the Labor Share and the Profit Share? Part III: Measures and Structural Factors," Levy Economics Institute of Bard College Working Paper No. 805, May 2014. PDF — used for all findings, the apportionment-method comparison, the sectoral and top-income decompositions, and all direct quotations; fetched and read in full (all 44 pages) this session.
2. Same paper, also released as University of Texas Inequality Project (UTIP) Working Paper No. 66, dated March 3, 2014. PDF — confirms dual working-paper release, no peer-reviewed venue.
3. Olivier Giovannoni, Google Scholar author profile — checked to confirm no later peer-reviewed publication of this series. scholar.google.com/citations?user=h2y9TJEAAAAJ