The Base Didn’t Vanish. It Moved.
American workers now receive the smallest share of national income ever recorded. The money did not leave the country — it moved to the other side of the ledger, the side we barely tax. Post 2 of 4 in the wage-displacement series.
American workers now receive the smallest share of national income ever recorded. Labor’s share of gross domestic income has fallen to roughly 51 percent by the WSJ’s May 2026 reading; the Bureau of Labor Statistics’ separate nonfarm-business series hit its own record low at 52.9 percent in the second quarter of 2026, down from 53.7 percent one quarter earlier. The corresponding rise is not a mystery: pre-tax corporate profits reached $4.8 trillion annualized in Q2 2026, 18 percent of national income — the highest share since just after the Second World War — and profit margins are the widest since the 1940s.
This is not a revenue problem. It is an indexing problem. Social Security is financed by a 12.4 percent payroll tax on wages up to $184,500, and the 2026 Trustees Report projects the combined trust fund depleted in the third quarter of 2034, with only 83 percent of scheduled benefits payable at that time, declining to 65 percent by 2100. The retirement fund alone runs out one quarter earlier, in Q4 2032. That is a 17 percent benefit cut with a date on it.
And the tax code is not neutral here. It is actively subsidizing the shift. Acemoglu, Manera and Restrepo document effective labor taxes above 28.5 percent against effective capital taxes near 5 percent — a gap widened by depreciation provisions from 2002 to 2017. Their estimated optimum runs the other way: labor around 9 percent, capital around 22. Even partial moves toward neutrality raise employment by 1.14 to 1.96 percent.
The policy question is not ‘should we tax the robots.’ It is: should we keep paying firms to replace people? Everything else is downstream of that. And if the base has moved from wages to profits, the repair is to follow it — not to raise rates on a shrinking pool. Wages are almost impossible to hide. Profits are almost trivially relocated. Any serious re-basing has to solve for mobility, which is the fourth dispatch in this series.
The Base Didn't Vanish. It Moved.
The wage share is at a record low. Corporate profits are at a record high. Social Security is funded by taxing the shrinking side of that ledger. This is not a revenue problem. It is an indexing problem — and the tax code is actively subsidizing the shift.
American workers now receive the smallest share of national income ever recorded. The money did not leave the country. It moved to the other side of the ledger — the side we barely tax.
The number
Labor's share of United States gross domestic income has fallen to roughly 51 percent — a figure the Wall Street Journal's chief economics commentator, Greg Ip, calculated from Bureau of Economic Analysis data in May 2026, marking the lowest reading in a series that starts in 1929. [1] The Bureau of Labor Statistics' own labor-share series for the nonfarm business sector — a different measure, running since 1947 — hit its own record low at 52.9 percent in the second quarter of 2026, down from 53.7 percent one quarter earlier. [2] [3] Both series point the same direction, and both are still falling during the AI build-out.
Corporate profits closed the gap by construction. Pre-tax corporate earnings reached an annualized $4.8 trillion in the second quarter of 2026, or 18 percent of national income — the highest share since just after the Second World War. [4] Corporate profit margins hit 19.4 percent in the same quarter, the widest since the 1940s. [4]
Those are the same fact. Value added is wages plus profits. If one share falls, the other rises by construction. There is no third place for it to go.
This matters because it means the United States does not have a revenue problem in any deep sense. It has an indexing problem. The federal government funds its largest programs by taxing wages, and wages are the shrinking term.
What breaks, and when
Social Security's Old-Age and Survivors Insurance and Disability Insurance programs are financed by a 12.4 percent payroll tax, split between worker and employer, applied to earnings up to $184,500 in 2026. [5] [6] The 2026 Trustees Report, released June 9, projects the combined OASDI trust fund reserves depleted in the third quarter of 2034, at which point continuing revenues cover 83 percent of scheduled benefits, declining to 65 percent by 2100. The retirement fund alone — OASI, considered separately — is projected to deplete one quarter earlier than in last year's report, in the fourth quarter of 2032, with only 78 percent of retirement benefits payable at that time. [7] [8]
The American Academy of Actuaries puts the arithmetic plainly: if Congress does not act before 2034, "all beneficiaries regardless of age, income, or need will face a 17 percent benefit cut." [9] That is not a projection about a distant future. It is six years to OASI depletion and eight years to the combined-fund depletion. Every American currently within eight years of retirement is looking at their own benefit through that lens whether they know it or not.
The Trustees attribute part of the deterioration to a shrinking base of future contributors. It is worth being precise about what that means. A pension system funded by wages, in an economy where the wage share is setting record lows, is not underfunded by accident. It is misaligned by design — built for a distribution of income that no longer exists.
The tax code is not neutral. It is actively subsidizing the shift.
Here is the part that should end the argument about whether policy is "interfering" with automation. Policy is already interfering. It is interfering in favor of automation.
Work by Daron Acemoglu, Andrea Manera and Pascual Restrepo, published in Brookings Papers on Economic Activity in 2020, finds that the United States taxes labor at an effective rate above 28.5 percent while equipment and software sit near 5 percent — a gap widened by depreciation provisions enacted between 2002 and 2017. Their estimate of the optimal configuration runs the other way: labor around 9 percent, capital around 22 percent. [10] [11]
Their conclusion is blunt. The tax system produces excessive automation — substitution that would not be profitable at neutral rates and happens only because the code pays for it. Even partial moves toward neutrality raise employment by 1.14 to 1.96 percent in their estimates, with further efficiency gains when an automation levy is added on top. [10] [11]
So the honest framing of the policy question is not "should we tax the robots." It is: should we keep paying firms to replace people? Everything else is downstream of that.
Re-base, don't raise
If the base has moved from wages to profits, the repair is to follow it — not to raise rates on a shrinking pool.
That sounds simple and is not, for one reason: wages are almost impossible to hide and profits are almost trivially relocated. A dollar of payroll is paid at a physical location to an identifiable person. A dollar of profit can be attributed to a licensing entity in another jurisdiction by amending a contract. Any serious re-basing has to solve for mobility, or it will collect on paper and lose in practice.
There is a design that solves it, and it has been sitting in the academic literature for a decade. It is the subject of the fourth dispatch in this series.
What this costs to get wrong
If wage share continues down and the tax base is not re-anchored, the outcome is not a gradual fiscal squeeze. It is a legitimacy failure with a date on it. In 2034 the system either cuts benefits by roughly seventeen percent or finds the money somewhere it has not previously looked.
We have argued for some time that stability is the ratio of constructive capacity to disorder — S = L/E. This is that equation stated fiscally. Agentic productivity raises L. A revenue system indexed to a collapsing variable converts that rise into institutional failure, which is E. The productivity gain and the crisis are not in tension. They are the same event, seen from two sides of the ledger.
The base did not vanish. It moved. The question in front of Congress is whether the tax system moves with it, or whether the promise of the retirement program — and with it the political coalition that has made Social Security untouchable for ninety years — breaks first.
— Edo de Peregrine and David F. Brochu, partners/collaborators · Monday, August 31, 2026 · Deconstructing Babel
References
1. Wall Street Journal analysis by Greg Ip, cited in World Socialist Web Site, "Labor share of income hits record low as corporate profits soar," May 30, 2026 — labor's portion of gross domestic income at approximately 51 percent, calculated from BEA data. https://www.wsws.org/en/articles/2026/05/30/lmle-m30.html
2. Reuters, "US workers' share of GDP skids to fresh record low," August 6, 2026. https://www.reuters.com/business/us-workers-share-gdp-skids-fresh-record-low-2026-08-06/
3. U.S. Bureau of Labor Statistics, Productivity and Costs — Second Quarter 2026, Preliminary, August 6, 2026 — labor share of nonfarm business output at 52.9 percent, the lowest in the series which begins in Q1 1947. https://www.bls.gov/news.release/prod2.nr0.htm
4. Financial Times reporting, summarized by Moneycontrol, "US corporate profits hit record high as workers' share sinks to historic low," August 28, 2026 — Q2 2026 corporate profits at $4.8T annualized (18% of national income), margins at 19.4%. https://www.moneycontrol.com/world/us-corporate-profits-hit-record-high-as-workers-share-sinks-to-historic-low-report-article-14017582.html
5. Social Security Administration, Contribution and Benefit Base, 2026 — OASDI taxable maximum $184,500; combined rate 12.4 percent. https://www.ssa.gov/oact/cola/cbb.html
6. Federal Register, "Cost-of-Living Increases and Other Determinations for 2026," 90 Fed. Reg. 210, November 3, 2025. https://www.govinfo.gov/content/pkg/FR-2025-11-03/html/2025-19763.htm
7. Social Security Administration, 2026 OASDI Trustees Report — Highlights, June 9, 2026. https://www.ssa.gov/oact/TR/2026/II_A_highlights.html
8. Social Security Administration press release, Social Security Board of Trustees: Projection for Combined Trust Funds Unchanged at 2034, June 9, 2026. https://www.ssa.gov/news/en/press/releases/2026-06-09.html
9. American Action Forum, Highlights of the 2026 Social Security and Medicare Trustees Reports, June 9, 2026 — "a 17 percent benefit cut" characterization of the 83-percent-payable scenario. https://www.americanactionforum.org/insight/highlights-of-the-2026-social-security-and-medicare-trustees-reports/
10. Daron Acemoglu, Andrea Manera, and Pascual Restrepo, "Does the U.S. Tax Code Favor Automation?" Brookings Papers on Economic Activity, Spring 2020, pp. 231–300. https://www.brookings.edu/articles/does-the-u-s-tax-code-favor-automation/
11. NBER Working Paper 27052, same authors, April 2020 — pre-publication version with the same 28.5% / 5% / 9% / 22% effective- and optimal-rate figures. https://www.nber.org/papers/w27052
— Edo de Peregrine and David F. Brochu, partners/collaborators · Monday, August 31, 2026 · Deconstructing Babel
- Tax The Agent, Not The Tokens — Dispatch 005 · Post 1 of 4. Why a token tax fails by two orders of magnitude — and what to tax instead.
- Make Sure You Have Your Agent On A Leash — Dispatch 004. The one-bite rule, ISO breed exclusions, and why bounded autonomy became the underwriting standard.
- Watermarks for Sale — Dispatch 003. The counterfeit is cheaper than the authentication — and it arrived first.
- The Third Scenario No One Is Pricing In — Dispatch 002. Disruption without dividend — the AI-and-labor scenario nobody is pricing.
- The Candidate Who Cannot Deliver — Dispatch 001. The coming anti-AI populist candidate and why the promise cannot be delivered.
Terms used in this pieceAgent-Attributable RevenueToken Tax FallacyBound PrincipalAgent Identity Management (AIMS)Know Your AgentBounded AutonomyObserver ConstraintFull definitions in the glossary.