[MaxGhenis]

This brings docs/parity-ledger.md up to date with what has merged since its last check at ad42ecc0 (18 October): the work rulings d1802 and d1803 started, the three approved lane designs, and U2's ratification. It changes no code, gate or run.

Verified at 2195258c (first commit)

• Forward earnings: S1 is B2's successor under d1803. It is specified and tested on invented data (#555), with P2 in reserve (#561, draft). Neither has registered.
• Family maximum: G v2 merged (#530). The row now lists the five configurations it computes, the two it refuses by name and its 29 worked examples. G part 2 (#563) is in review.
• Retirement earnings test: R3 now computes a reentitled spouse's pre-FRA months (#564). R2 part (b) is in Track B G part 2 and R2 part (b): deductions before reduction; SSA's operational saving clause #563.
• Insured status: #562, a draft in review.
• Critical path: steps 1, 2, 3 and 5 updated to match.
• Rulings: adds d1724, d1802 and d1803; d1338 now points to G part 2.
• Progress scorecard (docs/progress/scorecard.json, which the live progress page fetches from master): the forward-projection row now says S1 is built and tested on invented data, not registered.

Verified at f8ffc7d9 (commit 4200e1af)

• Twelve rows the merged designs cover now cite the design files and name their first milestones and what each waits on:
  • from docs/design/mortality_groups_financing.md (#558, d1922): mortality differences, education, results by group, national costs, and trust funds;
  • from docs/design/income_and_resources.md (#559, d1923): covered earnings, SSI, and wealth and pensions;
  • from docs/design/open_population.md (#557, d1925): births, immigration and emigration, alignment, and projections beyond 2030.
• Poverty row (U2): the second structural pass completed under d1724 after U2 births: count out seeds outside the birth law's support (d1724) #552, and U2 §20 steps 4–6: record the step-4 pass, materialize the final parameter block and rulings record, ratify by merge #556's merge ratified the specification. Next are the evidence record, a fresh forecast and the Candidate 2 design: latent-permanent conditioned chained QRF (research memo) #42 registration.
• Known gaps: keeps only what no issue, PR or approved design plans: moving the engine into Microcosm, and entrants' later disability and benefit receipt. (Gate 3's cells were listed here until review r1 found them drafted; see below.)
• Rulings: adds d1922, d1923 and d1925.
• Scorecard: the wealth row names the approved design.
• Header: records the second verification.
• docs/progress.md is regenerated with scripts/build_progress_page.py.

Review r1's corrections (commit b64705e2)

The independent review r1 (GPT-6.1 Sol) re-verified every row at f8ffc7d9 and asked for eight changes. Each was checked against the code before it was applied:

• Benefit rules run in Axiom: Axiom's AIME agreement for 7,486 people is against the statutory-count reference; Track A's fixed 35 years gives a different AIME for 251 of them, all born before 1929 (scripts/track_c_aime_agreement.py).
• Results by group: MINT's beneficiary-type classification is built (group_breakdowns/common.py). Only rows that need projected household income or poverty status are not computed.
• Scorecard, DYNASIM output comparisons: exercises 1–4 are quantitative comparisons with DYNASIM3, not "not yet".
• Projections beyond 2030 and the known gaps: gate 3's cells, tolerance formula and pass rule are drafted in docs/plans/microcosm-start-registration-draft.md. Posting waits on d947.
• Scorecard, gate m6: it gates 2015–2019 flows and the 2016 and 2018 earnings, with 2020–2022 as a diagnostic. The holdout also informed the gated surface and the forward earnings law's structure.
• Disability insurance entitlement: conversion at full retirement age has no draw.
• Behavioral responses: exercise 3's claiming alternatives cap at age 70, and one moves only claims at 65 or later.
• Scorecard, gate w1: its 44 gated cells are 43 cross-sectional joints plus one compression swap. SSA margins are report-only.

docs/progress.md is regenerated.

Delta review r2

The delta review r2 (GPT-6.1 Sol) returned APPROVE at b64705e2. It found all eight corrections resolved and accurate against f8ffc7d9, and requested no replacement text. It also checked that the commit changes only the three documents, that the committed generator reproduces docs/progress.md byte for byte, that every ledger reference resolves, and that neither the body nor the diff discusses data-use terms. The PR merges on its gates.

Later merges (#560, #568, #576 and #578, and #567 once it lands) go in the next ledger update; this one stays verified at the commits above.

axiom: n/a: microcosm-dynamics documentation, not a PolicyEngine country model

🤖 Generated with Claude Code

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
social-security-model	Ready	Preview	Oct 24, 2026 4:19pm UTC

[MaxGhenis]

Closing and reopening only to rerun CI on current master (4e20c6e4, after #582). The head is unchanged.

[MaxGhenis]

The dynamics coordinator merged this at its reviewed head b64705e2e12457e8bcf4383ae42ef2e6f7b8e0b2 (merge commit a190d7d55dbd8a248495996f9702d72f258f1278), with --merge --match-head-commit.

Gates.

• gh pr checks exited 0 on the latest run. It reran on current master after Registry: compare a pull request's own commits from its fork point #582, since the earlier green run predated Track B: held-out data registry for every Track B and open-population gate #560's history test.
• The PR was MERGEABLE and CLEAN, and no review requested changes. It was marked ready just before the merge; the test workflow doesn't run on that event.
• It changes only docs and progress data (docs/parity-ledger.md, docs/progress.md, docs/progress/scorecard.json), so tier counts are unaffected. The progress builder reads only docs/progress/*.json, which OP0.1: source, exposure, access and power audit for the demographic gates #567 did not touch.

Independent review. Two GPT-6.1 Sol rounds:

• Round 1 (20261024-115152-ledger-569-review-r1) re-verified every row at f8ffc7d9 and asked for eight changes. All eight are applied in b64705e2.
• Round 2 (20261024-121926-ledger-569-review-r2): APPROVE at b64705e2, with every round-1 finding resolved.

Not yet covered. The ledger is verified at f8ffc7d9. A ledger delta will cover what merged after that: #560, #568, #576, #578, #582, #567 and #577.