# The Shareholder Commonwealth — working notes

Third paper in the sequence. ADVOCACY, declared as such: the case that World III
(from Disarmed Countermovement Part VI) is the best feasible world, and the road.

## Thesis
Best on three grounds: republican freedom (Pettit non-domination / eyeball test —
Caretaker State = benevolent master; Douglass "slavery, not its incidents"; Belloc's
Servile State), Rawlsian justice (Rawls HIMSELF: welfare-state capitalism fails
justice as fairness, "permits a small class to have a near monopoly of the means of
production"; POD = "widespread ownership of assets... at the beginning of each
period" — verified verbatim p.139 JF:AR), and stability (only world where the
concession function stays permanently bound — leverage constitutionalised: a share
is the instrument that makes a claim survive the claimant's usefulness).
Achievable by RATCHET: seed (subsidies→equity; Intel precedent) → dividend fast
(Alaska rule, anti-Meidner) → entrench (constitutionalise corpus AND formula) →
broaden (auto-accounts, NEST ~90%; baby bonds) → extend to rents (limit case base).

## Original claim
THE HEDGE PROPERTY: equity-funded universal endowment is the only social insurance
whose payout is positively correlated with the disaster it insures (wage collapse =
equity boom). Negative-beta social insurance. Not found stated in the literature —
flagged in disclosure as the paper's own claim.

## Key verified facts (all URLs verified 2026-08-02, agents' dossiers)
- Verasight Jun 2026: 69% support AI firms transferring 50% of stock to public fund
  (64% even when attributed to Sanders). CNBC Jul 2026: only 19% support "government
  owning shares of companies" (up from 13% Oct 2025). THE 50-POINT FRAMING WEDGE.
- Sanders American AI Sovereign Wealth Fund Act (introduced ~Jun 18 2026): 50% stock
  levy on AI firms >$200M receipts; ~$7T fund; 5% draw; ~$1,000/person/yr initial;
  no-sale rule. 200 economists incl. 16 Nobels warning statement Jul 13 2026.
- US stakes: Intel 9.9%/$8.9B (CHIPS conversion, Aug 2025); US Steel golden share
  (Jun 2025); MP Materials ~15% DoD; Nvidia/AMD 15% China revenue share (legally
  contested); EO 14196 (Feb 2025) SWF — no fund yet; Canada Strong Fund C$25B 2026.
- OpenAI Foundation ~$130B (~26% at $500B — 26% UNVERIFIED at source), Oct 28 2025,
  extracted via CA/DE AG leverage. Anthropic LTBT: governance-only Class T.
- Norway: ~$2T, ~1.5% of world listed equity, votes 11,000+ AGMs/yr, publishes
  intentions 5 days ahead; 3% rule since 2017 (4% 2001).
- Alaska: 1976 amendment 75,588–38,518 (~2:1); dividend since 1982; Harstad 2017:
  64% prefer income tax over ending PFD (29% in 1984); 90% prefer universal;
  corpus constitutional, dividend statutory → cut post-2016 (Wielechowski v. State
  2017 upheld veto). LESSON: constitutionalise the formula too.
- Meidner: 1976 proposal 20% of profits in new shares; enacted 1984 diluted (5 funds,
  SEK 17.5B cap, 8% vote cap); 4 Oct 1983 march 80-100k largest business demo in
  Swedish history; abolished 1991-94, assets → research foundations; zero visible
  worker benefit ever. Failure = class-bound + accumulation-first + statutory.
- Australia super: 1992 at 3% → 12% Jul 2025; A$4.3-4.4T; lesson: individual
  accounts mirror wage inequality → need both legs.
- New Mexico: 2022 referendum ~70-30 raised LGPF distribution; universal free
  childcare funded from fund returns 2025.
- Homestead: 1.6M claims, 270M acres (~10% of US land). GI Bill: 7.8M educated,
  49% of 1947 college admissions, 2.4M home loans by 1952.
- Paine 1797 verified verbatim (SSA archive): £15 at 21, £10/yr at 50; "not charity
  but a right"; ground-rent logic. Jefferson 1785 Fontainebleau: "legislators cannot
  invent too many devices for subdividing property."
- Objections bank: Konczal why-not-tax; Clifford Chance £300B (transfer-not-
  destruction reply; 55% foreign shareholders); Weyl "dangerous ideology" +
  Prewitt "peasant class" (CNBC 2021); Rand Paul "step toward socialism";
  Jones-Marinescu no employment effect; 1MDB >$4.5B; PIF $2B Kushner;
  Madrian-Shea/NEST ~86-90% default power.

## Attack log
- Rawls p.139 says "widespread ownership of ASSETS and human capital" (not
  "productive assets") — quote used correctly.
- OpenAI Foundation 26% share UNVERIFIED at OpenAI source; use "$130B" (verified).
- Meidner march size: 80-100k (sources vary; "75k" also cited) — used range.
- Verasight Q wording ("required to transfer") likely inflates support — caveated
  in disclosure ("polling before costs attached overstates").
- Fund arithmetic is mechanical/nominal, no GE — stated in fig caption + App A;
  year-25 transform numbers explicitly labeled absurd-by-construction.
- Conflict flagged twice: AI execs proposing levies on themselves; this paper
  drafted with Claude (Anthropic) assistance — in disclosure box.

## Files
- build_paper_v1.py → core (19pp) + full PDFs
- fund_arithmetic.py → fund_arithmetic_result.json
- generate_figures.py → fig1_lineage, fig2_arithmetic, fig3_politics

## Error log
- fig1 v1: linear time axis piled 1944-2026 labels; v2 staggered offsets still
  collided; v3 evenly-spaced index axis with year labels — fixed.

## Draft 1.1 — dividend math expanded (same day)
- §11.1 The dividend ladder (Fig 4): conversion factor $6.5T fund = $1,000/adult/yr
  at 4% draw. Green rungs reachable from present equity base; red rungs
  (wage replacement, $273-520T) reachable ONLY in the transform state — which is
  the hedge property read off a bar chart. Anchors: poverty ~$15.6k, median
  personal ~$42k, median household ~$80k (approx, labeled).
- §11.2 Milestones (Fig 5, log scale): ratchet crosses $1,000 in yr 6-8 in ALL
  states (political payoff arrives regardless — anti-Meidner timing point:
  "Sweden's funds died in year seven having paid nobody"); $5,000 floor yr 13/17/22;
  Altman $13,500 yr ~19-24 except fizzle.
- §11.3 Funding stack table: ~$1T/yr total. Share levy $600B ≫ token tax $25B
  (25:1) — stock-vs-flow argument in miniature; stack diversifies the politics.
- Appendix A: draw×return sensitivity table ($1,132-$3,584 year-10 span);
  draw rate framed as a values choice (legacy vs constituency); year-by-year
  central path to $95T/yr25. All numbers from fund_arithmetic.py extensions →
  fund_arithmetic_ext.json.
- fig4 anchor labels collided with bottom bar → moved to top, staggered.