The labor share. Is value really moving from labor to capital? The data isn’t on anyone’s side yet.

📊 Full opportunity report: The labor share. Is value really moving from labor to capital? The data isn’t on anyone’s side yet. on ThorstenMeyerAI.com — validation score, market gap, and execution plan.

TL;DR

The data on labor’s share of income remains stable over 70 years, but early signs suggest AI may be reallocating value at the margins. The overall impact remains uncertain, with implications for ownership policies.

Recent data shows that the overall share of income going to labor in the US has remained within a narrow range over the past 70 years, despite rapid technological change. However, early signals suggest that AI may be beginning to shift value at the margins, especially among entry-level, routine jobs, raising questions about the future of labor’s share.

The US labor share of income has fluctuated between approximately 57% and 64% from the 1950s to 2023, remaining relatively stable through waves of automation, computers, and the internet. A Stanford study analyzing millions of payroll records found a roughly 13% decline in employment for 22-to-25-year-olds in AI-exposed occupations since late 2022, after controlling for firm shocks. These younger workers in routine, entry-level jobs have experienced displacement, while older workers in the same roles have remained stable or grown, indicating a shift at the margin. Experts argue that the aggregate labor share’s stability does not negate these early signals, which are concentrated and predicted by economic theory. The debate centers on whether these marginal shifts will eventually impact the overall share, with some analysts emphasizing the importance of the early, localized signs, and others pointing to the long-term stability of the aggregate data.

The Labor Share — Thorsten Meyer AI

● DISPATCH / JUNE 2026

THORSTEN MEYER AI · POST-LABOR · § 02

POST-LABOR · 02

EVIDENCE / SHARE

Essay · The Empirical Floor Under The Stake · 2026-06-07

The ownership case rests on a premise. This dispatch tests it — and holds my own argument to the standard I hold everyone else’s.

The skeptic’s strongest chart: the US labor share has stayed within a 57-64% band from the 1950s to 2023, through industrial machinery, computers, and the internet. The other side’s strongest number: a Stanford study found a ~13% relative employment decline for 22-25-year-olds in the most AI-exposed jobs since late 2022 — while older workers held steady. The aggregate is stable; the margin is moving. The structural argument: the premise under the ownership case is true at the margin and not yet true in the aggregate — genuinely unresolved, because a durable share-shift is confirmable only in retrospect. Which means the ownership case rests not on a proven aggregate shift but on a marginal one that may or may not become aggregate — and that uncertainty is the strongest argument for a no-regrets response.

Thorsten Meyer ThorstenMeyerAI.com Munich · Iffeldorf Essay · ~5,000 words Post-Labor · 02 · weekend read

57-64%

US labor share band · 1950s-2023 ·
the skeptic’s strongest chart

−13%

Relative employment, 22-25-yr-olds
in AI-exposed jobs since 2022 (Stanford)

238 regions

EU areas where AI patenting tracks
declining labor share (Minniti et al.)

not yet

Knowable · a share-shift is
confirmable only in retrospect

THE LABOR SHARE· IS VALUE REALLY MOVING FROM LABOR TO CAPITAL· THE AGGREGATE IS STABLE · THE MARGIN IS MOVING· 57-64% BAND FOR 70 YEARS · THE SKEPTIC’S CHART· −13% ENTRY-LEVEL IN AI-EXPOSED JOBS · THE SIGNAL· AUTOMATION → DECLINE · AUGMENTATION → STABLE· THREE QUESTIONS · JOBS · WAGES · SHARE OF VALUE· THE OWNERSHIP CASE NEEDS ONLY THE THIRD· THE BARGAINING-POWER CHANNEL · A DRIFT, NOT AN EVENT· NBER · ENTRY-LEVEL DECLINE MAY BE INTEREST RATES, NOT AI· EXPOSURE IS NOT DISPLACEMENT· CONFIRMABLE ONLY IN RETROSPECT · NOT YET KNOWABLE· THE UNCERTAINTY IS THE CASE FOR A NO-REGRETS RESPONSE· THE LABOR SHARE· IS VALUE REALLY MOVING FROM LABOR TO CAPITAL· THE AGGREGATE IS STABLE · THE MARGIN IS MOVING· 57-64% BAND FOR 70 YEARS · THE SKEPTIC’S CHART· −13% ENTRY-LEVEL IN AI-EXPOSED JOBS · THE SIGNAL· AUTOMATION → DECLINE · AUGMENTATION → STABLE· THREE QUESTIONS · JOBS · WAGES · SHARE OF VALUE· THE OWNERSHIP CASE NEEDS ONLY THE THIRD· THE BARGAINING-POWER CHANNEL · A DRIFT, NOT AN EVENT· NBER · ENTRY-LEVEL DECLINE MAY BE INTEREST RATES, NOT AI· EXPOSURE IS NOT DISPLACEMENT· CONFIRMABLE ONLY IN RETROSPECT · NOT YET KNOWABLE· THE UNCERTAINTY IS THE CASE FOR A NO-REGRETS RESPONSE·

FIG. 01 — THE STABLE AGGREGATE · THE SKEPTIC’S STRONGEST CHART

Seventy years of enormous technological change — and labor’s slice stayed in its band

If labor’s share survived every prior wave, why would AI break it?

64%

57%

1950s

2023

stable

The US labor share fluctuated within roughly 57-64% across industrial machinery, the computer, and the internet — each, in its moment, the technology that was going to break the work-income link. The economy keeps inventing new labor-side work as fast as the old is automated. As of early 2026, the aggregate data is on the skeptic’s side: the share is stable, employment is stable, wages are not falling. Any honest ownership argument has to begin by conceding this.

FIG. 02 — THE MOVING MARGIN · WHERE THE SIGNAL ACTUALLY APPEARS

The aggregate is a sum — and sums can be flat while components move oppositely

The displacement appears exactly where the theory predicts: entry-level, AI-automated work

22-25, AI-exposed jobs

−13%

Relative employment decline since late 2022 — controlling for firm shocks (Stanford / Brynjolfsson)

Older workers, same jobs

steady

Held steady or grew — experience and tacit knowledge as a buffer against displacement

AI automates (code, customer chat) → entry-level hiring declines

AI augments (problem-solving, accuracy) → employment holds or rises

The signal tracks the mechanism — displacement appears where AI substitutes rather than complements, which is evidence it’s causal, not coincidental. And the European data shows the share-shift itself: across 238 regions in 21 countries, higher AI-patenting intensity tracks more pronounced declines in labor’s share of income (Minniti et al.) — AI as a capital-biased technology.

FIG. 03 — THE THREE QUESTIONS · WHAT “LABOR SHARE” ACTUALLY MEANS

Much of the disagreement dissolves once you separate three questions

They have different answers — and the ownership case depends on only one

Question oneDo jobs disappear?

Mostly not, yet

Question twoDo wages fall?

Mostly not, yet

Question three — the real oneDoes labor’s share of the value fall?

Unresolved

A worker can keep their job and their wage while the share of output going to wages (versus profits) declines — that’s the capital-share rise, and it’s compatible with full employment. The skeptic’s strongest evidence answers questions one and two; the ownership case concedes those and asks the third — harder to measure, slower to appear, visible mainly in retrospect. The debate talks past itself because each side is answering a different question.

FIG. 04 — THE BARGAINING-POWER CHANNEL · HOW THE SHARE MOVES WITHOUT JOBS VANISHING

If the share can fall while jobs and wages hold, there has to be a mechanism

AI shifts leverage from labor to capital even when it doesn’t eliminate the job

What we look for

A layoff (an event)

Visible, datable, easy to count. The thing the aggregate employment data tracks — and it’s stable.

What’s actually happening

A drift (erosion)

AI as a credible partial substitute weakens leverage; the automated learning curve breaks the entry-level deal. Value shifts to capital gradually — as wages growing slower than productivity.

AI doesn’t have to replace a worker to weaken their position; it only has to be a credible partial substitute. The “deal” of junior work — rote labor for mentorship — breaks when AI does the rote labor, and the career ladder loses its bottom rung. A bargaining-power shift is a slow drift, invisible in real time and obvious in retrospect — which is why the aggregate hasn’t “moved” yet even if the mechanism is already operating.

FIG. 05 — THE VERDICT · WHAT THE DATA CAN AND CANNOT SUPPORT

Narrower than either camp would like — and the narrowness is the point

The skeptic’s case is serious: the entry-level decline may be interest rates, not AI (NBER)

What the data supports

What it does NOT support

A real, concentrated, mechanism-consistent marginal signal — entry-level displacement where AI automates, EU regional share declines.

An aggregate share-shift, or a confident forecast that the margin becomes the aggregate. The band holds; the confounds are real.

Reasonable belief the marginal shift is real and AI-related.

Anyone claiming the shift is proven or certainly coming reads more than the data holds.

The verdict is not “yes” and not “no” but “not yet knowable” — and that’s not a dodge; it’s the accurate epistemic state. A share-shift is confirmable only after it has happened, so waiting for proof means waiting until it’s irreversible.

The empirical ambiguity that weakens a confident displacement narrative is precisely what strengthens the case for a response that doesn’t require the narrative to be confident. You don’t need the premise proven to justify a no-regrets response. You only need it plausible — and the marginal evidence makes it more than plausible.

Thorsten Meyer · The Labor Share · Post-Labor 02

Source dossier

ITIF (May 2026) · AI is not going to reduce labor’s share — the seventy-year band: US labor share ~57-64% from the 1950s to 2023 despite industrial machinery, computers, the internet; workers reallocate rather than disappear; the tax base erodes only if labor share falls dramatically and persistently — “highly unlikely” · [itif.org](https://itif.org/publications/2026/05/14/ai-not-going-reduce-labors-share-of-income-or-destroy-tax-base/)
Brynjolfsson, Chandar & Chen (Stanford) · Early employment effects of generative AI — ~13% relative employment decline for 22-25-year-olds in the most AI-exposed jobs since late 2022, controlling for firm shocks; older workers steady or growing; automation → decline, augmentation → stable; BofA — recent-grad unemployment exceeding overall for the first time · [cnbc.com](https://www.cnbc.com/2025/08/28/generative-ai-reshapes-us-job-market-stanford-study-shows-entry-level-young-workers.html)
Minniti et al. / The Economy · When AI captures the surplus — 238 regions across 21 European countries: higher AI-patenting intensity tracks more pronounced declines in labor’s income share, especially industrial areas; AI as capital-biased, reallocating returns to capital; diminished worker bargaining power even when employment persists · [economy.ac](https://economy.ac/review/2026/03/202603288663)
International Center for Law & Economics · empirical-evidence review — no economywide displacement through 2024-25; 35.9% AI adoption by Dec 2025 with small positive wage effects (Hartley et al.); early adjustment via task reallocation and within-firm productivity; concentrated entry-level effects in exposed occupations · [laweconcenter.org](https://laweconcenter.org/resources/ai-productivity-and-labor-markets-a-review-of-the-empirical-evidence/)
Stanford Review · The class of 2026 — not because of AI — the NBER paper (25,000 workers, 7,000 workplaces): zero effect on earnings/hours, entry-level decline replicated but shown NOT driven by firms adopting AI; the interest-rate confound (Fed near-zero 2020-22 fueling speculative hiring that reversed) · [stanfordreview.org](https://stanfordreview.org/the-class-of-2026-is-struggling-to-find-jobs-and-its-not-because-of-ai/)
Citadel Securities · The 2026 Global Intelligence Crisis — little evidence of AI disruption in labor-market data today; share of working-age adults using GenAI for work looks stable, not non-linear; high substitution elasticity could collapse labor share in theory, but policy would offset · [citadelsecurities.com](https://www.citadelsecurities.com/news-and-insights/2026-global-intelligence-crisis/)
Rezi · Crisis of entry-level labor — the “learning curve” automating; the entry-level deal (rote labor for mentorship) breaking; UK tech graduate roles −46% in 2024 (projected −53% more by 2026); healthcare/government/hospitality absorbing 2024-25 job growth · [rezi.ai](https://www.rezi.ai/posts/entry-level-jobs-and-ai-2026-report)
ALM Corp · AI job displacement statistics — Dallas Fed — AI augmenting not replacing in exposed jobs; Anthropic’s usage-based measure — high-usage AI occupations seeing modestly slower hiring (modest but accelerating); Goldman 6-7% long-term (~11M); NBER ~3.9% at high exposure + high automatability · [almcorp.com](https://almcorp.com/blog/ai-job-displacement-statistics/)
Frontiers in Human Dynamics · Observed displacement 2020-25 — PRISMA review: AI-driven displacement observable at the margin (lower-skill, routine cognitive) before the core employment relationship · [frontiersin.org](https://www.frontiersin.org/journals/human-dynamics/articles/10.3389/fhumd.2026.1815037/full)
The stake · Thorsten Meyer · Post-Labor 01 · the ownership case this dispatch’s empirical test underwrites — the no-regrets response to exactly the unresolved share-shift the data shows

Colophon

Set in Source Serif 4 (display, italic accent), EB Garamond (body), IBM Plex Sans (UI labels), IBM Plex Mono (mastheads, ticker, stamps). Paper-cool gray-cream #e6e7e4.

Chromatic register: labor-rose dominant (the loss/cautionary register, fitting the contested premise), structural-slate for the stable aggregate and the skeptic’s case, empirical-clay for the three-questions disaggregation, synthesis-deep for the not-yet-knowable verdict and the no-regrets close.

Key frame:

LABOR SHARE · TESTING THE PREMISE 57-64% BAND · 70 YEARS · STABLE −13% ENTRY-LEVEL · THE MARGIN MOVES THREE QUESTIONS · ONLY THE THIRD MATTERS A DRIFT, NOT AN EVENT NBER · MAYBE INTEREST RATES, NOT AI NOT YET KNOWABLE CONFIRMABLE ONLY IN RETROSPECT THE UNCERTAINTY IS THE CASE

Implications for Ownership and Future Policy

This divergence between stable long-term data and early displacement signals matters because it influences policy debates on whether to promote broad-based ownership of capital. If value is beginning to shift from labor to capital at the margins, policies encouraging ownership could be justified. However, the lack of definitive evidence at the aggregate level means that immediate sweeping actions may be premature. Recognizing the uncertainty helps policymakers design responses that are resilient to future developments, rather than reacting to unconfirmed shifts. The debate underscores the importance of understanding whether AI’s impact will be a temporary, marginal phenomenon or a durable, structural change that redefines income distribution.

Amazon

AI automation labor displacement books

As an affiliate, we earn on qualifying purchases.

Historical Stability Versus Emerging Signals

Over the past seven decades, the US labor share of income has remained within a narrow band, despite multiple technological revolutions. The 1950s to 2023 saw automation, digitalization, and internet proliferation, yet the overall share of income going to labor has not significantly declined. Recent research, including a Stanford study, indicates that while the aggregate share remains stable, early signs of displacement are emerging at the margin, particularly among young, entry-level workers in AI-affected sectors. These signals align with economic theories predicting that new technologies initially displace routine tasks before affecting broader income distribution. The debate is whether these marginal effects will accumulate into a long-term shift or remain localized and temporary.

“The premise that value is moving from labor to capital is true at the margin but not yet in the aggregate, and the evidence remains unresolved.”
— Thorsten Meyer

Amazon

entry-level routine job automation tools

As an affiliate, we earn on qualifying purchases.

Unresolved Questions About Long-Term Impact

It remains unclear whether the early, marginal displacement signals will coalesce into a sustained, aggregate shift in the labor share. The data currently shows stability at the macro level, but the signals at the margins are real and predicted by economic models. Whether these will accumulate into a broader redistribution of income from labor to capital is still an open question, only answerable after more time passes and additional data becomes available.

Amazon

AI impact on workforce training courses

As an affiliate, we earn on qualifying purchases.

Monitoring Displacement Trends and Policy Responses

Researchers will continue tracking employment and income share data, especially among entry-level and AI-affected sectors. Policymakers are advised to consider responses that are robust to uncertainty, such as policies promoting broad-based ownership, which can mitigate potential future shifts regardless of whether the aggregate labor share eventually declines. The next steps include refining measurement techniques and expanding data collection to better understand the evolving impact of AI on income distribution.

Amazon

As an affiliate, we earn on qualifying purchases.

Key Questions

Currently, the overall labor share has remained stable over the past 70 years, despite early signals of displacement at the margins.

What are the early signs that AI might be shifting value from labor to capital?

Recent studies show a decline in employment among young workers in AI-exposed roles, especially in routine, entry-level jobs, indicating displacement at the margin.

The aggregate data shows stability, but early signals are localized and may not yet reflect a long-term trend. The shift, if it occurs, can only be confirmed after it has happened over time.

What policy measures are recommended given the current uncertainty?

Policies promoting broad-based ownership of capital are advised, as they are resilient to whether the long-term shift occurs or not.

Only after more time passes and additional data confirms a sustained decline can this be definitively known.

Source: ThorstenMeyerAI.com

This content is for general information only and is not financial, tax or legal advice. Consult a qualified professional for decisions about your money.

The labor share. Is value really moving from labor to capital? The data isn’t on anyone’s side yet.

Up next

DojoClaw: The Engine Behind the Fleet

Author

leftbrainmarketing Team