UK spine follow-ups: SPI donors funded, WAS coherence, residential split, Child Benefit, release-cut blockers (#1063)

[juaristi22]

Closes #1063. On main (after #1084). Receipts: experiments/1063-uk-spine-followups.md.

The eight items and the fold-ins

• Items 1, 2, 4 (c1): the SPI income band donors become a mass-conserving support channel. Each band seats max(120, ceil(published taxpayers / 250)) donors at equal weights, funded in proportion from the households of each donor's region; no created mass, graph policy declared, an identity-keyed seating (salt spi_income_band_donor_draw) the E8 receipt reruns from ids with the licensed tape.

• Item 3 (c2): mass_increasing_support leaves the coverage validator.

• Items 5, 6 (c3+c4): WAS wealth is coherent with tenure (no mortgage debt off mortgaged tenure, no main-residence value off owner tenures) and with its own totals (property and financial wealth derived from their components, identities asserted at cleaning); new gate uk_stage_was_wealth_coherence.

• Items 7, 8 (c5, c6, c5b): the residential-property flag is carried as weight by a new expansion stage, cgt_residential_split, after the incidence anchor: each liable gainer's solved probability splits the household into arms at p·w and (1 − p)·w (2^k arms for k gainers), so the residential count and gains are identities on HMRC Table 8a in every gain band at design weights, with no draw and no seed; calibration decides between the arms as it does for the incidence clones. The asset-type stage types the arms and fits Table 7 once per claimant status with an error-diffusion walk (seed 554 retired). The per-band walk of c5 is replaced because its gains error in the open band was bounded by one stake several times the band's expectation.

• Fold-ins: salary sacrifice predicted only for people who are paid (c7); Child Benefit claims and opt-outs redrawn after the SPI chain against HMRC's August 2025 statistics in a new child_benefit_take_up stage (c8).

• Release cut (c9): the certifier rehearsal's five blockers cleared at their source (candidate label, export surface read without ids and with the weight, 43 declared outputs allow-listed, incapacity_benefit_reported dropped at the release boundary, FRS 2024-25's TRAIN2 read, the LCFS donor floored at 0 with a receipt, three signed register entries). The plan-2 measure exclusions are renewed to 2026-11-03.

• Certifier: parts are verified, and the certificate signed, over the gate battery's canonical bytes; the logbook's Chronicle: one append-only trace for every build attempt at every rung #628 form the certifier used renders 1.0 as 1, so every real signed report failed to authenticate while the float-free fixtures agreed, a latent defect reached once the gates all passed. Regression test added.

• Identity receipts after the pensions PR: since microcosm#1069 c9 the salary_sacrifice stage lowers a converted record's pay in place, after the CGT support split and the incidence anchor banded on it, so the E8 recomputes banded 2,797 clone carriers on post-sacrifice pay and failed on the final build's spine. The stage now keeps a converted record's pre-conversion pay on salary_sacrifice_pre_conversion_pay (an internal carrier, dropped at the release boundary) and the identity tool reverses the conversion for every receipt of an earlier stage; the residential split walks the liable gainers in person_id order so its arms are bitwise identical for any row order. The certifier's non-negative and QRF tail gates read export-dropped declared outputs from the spine frame.

• Target registry: the SPI 2023-24 Table 3.6 employment, self-employment and private-pension amounts by total-income band are projected to 2025 with HMRC's band-specific Income Tax Liabilities growth, as the count, total-income and tax rows by band already were, instead of a flat national index. The flat average-earnings index raised the £15,000 to £20,000 band's employment amount 10.7% while HMRC projects the band's total income down 2.3%; the three lowest employment targets move from £18.7bn, £53.8bn and £190.4bn to £17.7bn, £47.5bn and £172.5bn, and the frame's design-weight gaps from 20 to 31% to 15 to 22%. The remaining gap is the frame's composition of the low bands, documented in the receipts note as a follow-up.

Measured so far (licensed local arms, nothing released)

• Arm C (the walk): all spine gates pass; residential gains gap 0.3% of target; calibration loss 0.00853 against main's 0.00836, ESS 5,217 (5,198), max-to-median weight 1,115 (1,079; bound 1,151), projection entrants 6,715 (6,711).
• Arm B (WAS): all spine gates pass including the coherence gate; the rule drops £127bn of £1,266bn off-tenure mortgage mass; renters with property wealth fall to 16.5% private and 4.6% social.
• Child Benefit trial: 86.7% of eligible children claimed for (HMRC 86.6%), 9.0% of claiming families opted out, 11.1m children in payment against 13.5m before.
• Final national build (two runs, 2fe7d8b and 64c460b; the second differs only by the carrier column): all spine gates, 7/7 seam gates, calibration loss 0.00713 over 1,090 targets (main's 2026-09-30 build 0.00836); evaluation against the enhanced-FRS incumbent passed (loss 0.00817 vs 0.327, 1,020 target wins vs 30); E7 and E8 pass with the licensed tape. The release-cut certifier reaches all 20 gates: all 20 pass with the three input-mass deferrals signed (ruled 2026-10-04, see the open calls), and the signed certification composes (shippable: true, candidate aa31bdf6…). The compile-parity receipt was re-signed for the band-aware Table 3.6 amounts and the stale adult_ema exclusion retired.

Tests

Stacked PRs get no CI here. At the stack head: 983 targeted tests, the engine-free lane (6,294 passed; the two failures were the extras-free-venv and importlib-mode environment checks), the engine-UK lane (137), the integration smoke build and the parity fixture tests pass; the c9 commit was checked on its touched files plus the engine-UK lane and the smoke build.

Open calls

• Child Benefit opt-outs are drawn from fully charged families first, then the taper; or one pool over every charged family.
• other_residential_property_value (WAS DVHseVal) excludes buy-to-let, which sits in the drawn remainder.
• The Table 8a rows stay fenced; promoting them to bound targets is the UK CGT: sub-exempt gains capped at exactly the annual exempt amount become 10.8m taxpayers under uprating; interpolate the remainder below the threshold and anchor incidence #970 leverage question after the final build.
• Two thin dwp/uc_payment_dist cells join the uk-data#452 measure exclusions (to 2026-11-26): SINGLE 28,800–30,000 (one supporting household) and COUPLE_NO_CHILDREN 22,800–24,000 (two, design weights 1,025 and 1,999 against 2,761), whose awards left their bands when the larger donor block reordered the rows the later positional draws read. Keying those draws by identity is a follow-up.

🤖 Generated with Claude Code

[vahid-ahmadi]

Automated review pass (Claude Code, high effort) — round 1 at ebff8412

Verdict: one blocker and one urgent split-out. The rest is design-sound. c9 allow-lists the five internal disability carriers on the export surface, and a test that was green on #1084 now fails. Separately, the plan-2 exclusion renewal sits behind unmerged #1084, so main's exclusions lapse on 3 October. The spine changes (c1, c3+c4, c5b, c8) read correctly. C2, the final build and the certifier run are still owed, so this stays draft.

All diffs are against #1084's head 004b5c9c, so they cover only this PR. I ran 13 targeted engine-free files: 356 passed, 1 failed (item 1). These are the new-stage files, plus stage health, terminal gates, battery bindings, diagnostics, source stages, release coverage, salary sacrifice and LCFS.

---

1. Blocking: c9 puts the internal disability carriers on the export allow-list.

• 48eb73bf adds person.attendance_allowance_reported, dla_sc_reported, dla_m_reported, pip_m_reported and pip_dl_reported to uk_export_surface's allowed_extra_columns (gates.json:926 and the four lines around it).
• Those are UK_INTERNAL_DISABILITY_REPORTED_COLUMNS (frs_disability.py:25). test_uk_source_stages.py::TestE3ManifestLockstep::test_internal_disability_carriers_stay_out_of_export_registers exists to keep them off the release.
• That test fails at this head. On 004b5c9c none of the five is listed, so the failure is new.
• R2's "every declared stage output the incumbent lacks is allow-listed" caught the internal carriers. Drop them from the allow-list, or drop them at the release boundary as incapacity_benefit_reported is. Stacked PRs get no CI here, so this would only have shown up after the retarget.

2. Urgent, outside the stack: ship the plan-2 renewal to main on its own.

• On main 5187fce2, slc.borrowers.plan_2_liable and plan_2_above_threshold expire on 2026-10-03 (calibration_measure_exclusions.json:362,371).
• c5368635 renews them to 2026-11-03, but only on this branch, which is stacked on UK pensions: resident State Pension, Pension Credit, contributions and salary-sacrifice relief (#1069) #1084, itself still a draft.
• Until a one-line PR to main lands, every UK build on main fails on those two exclusions from tomorrow. I'd cherry-pick c5368635 onto main today. It is self-contained, apart from the gate-digest re-pin it needs.

3. Should-fix: the WAS tenure rule drops £127bn of mortgages but keeps the property they're secured on.

• mortgage_debt is HMortGR8, all of the household's mortgages (was_wealth.py:251; the comment at :712 says so).
• The rule zeroes it off a mortgaged tenure, so outright owners' and renters' buy-to-let and second-home mortgages go: £127bn of £1,266bn, all of it OthMortR8, per the receipts.
• The property those loans are secured on stays, in other_residential_property_value and the drawn remainder. So property and net wealth rise by £127bn on exactly the landlord households the WAS undercount already hits.
• Mortgaged-tenure households also keep their other-property mortgages inside mortgage_debt. So the column means "all mortgages" on one tenure and "main-residence mortgage only" on the others.
• Suggested fix: apply the tenure rule to the main-residence mortgage (TotMortR8), which you already read for diagnostics. Carry the other-property mortgage (OthMortR8_sum) unrestricted, as its own column or added back to mortgage_debt. Then the coherence gate tests the thing that is actually structural.

4. Should-fix: the residential split's band identities hold at design weights only, and nothing reports them after calibration.

• cgt_residential_split makes Table 8a's count and gains identities in every gain band before calibration.
• Calibration binds only the two national totals: hmrc.cgt.residential_property_taxpayers and _gains in uk_population_targets.json. The by-band rows stay fenced, as the receipts note says.
• The arms differ only in capital_gains_residential_property, and main and residential rates are both 18/24 from 2025-26. So no other target distinguishes the arms, and the post-calibration split within each household is set by the solver's update rule, not by data.
• That's fine as a design, and it's the incidence-clone precedent. But C2 should report the by-band residential count and gains on the calibrated weights next to the design-weight identities, so the "identity in every band" claim isn't read as a release property.
• Also worth recording in C2: the row-level minimum weight and the positive-weight median with the arms in place. A 0.04-household arm next to a 20.76 one is fine on the fold, but it is new for any row-level reader.

5. Question: Child Benefit children in payment are about 5% below HMRC.

• The vendored August 2025 statistics give 12.73m children registered and 1.00m opted out (all ages), which implies 11.73m in payment.
• The trial reads 11.1m in payment, while the claim rate (86.7% against 86.6%) and the opted-out family share (9.0% against 684,635 / 7,552,330 = 9.07%) both match.
• So the gap is in the eligible-child base or in the age mix of opt-outs, not in the rates. HMRC's opt-outs are concentrated at ages 10–16 (63,975 to 94,905 a year against 9,770 at age 0).
• Is the gap in the frame's eligible children, or in opt-out families having more children than the frame's charged families? One line in the receipts would settle it.

6. Nit: the stage notes overstate what calibration binds. cgt_residential_split.py:23 and the source_stages.json / sources.yaml notes say calibration "binds" Table 8a. It binds the two national totals only (see item 4). Reword it to "binds the national residential taxpayers and gains".

---

Your open calls

• Child Benefit opt-out pool: I'd use one pool over every charged family, with selection probability proportional to the charge as a share of the benefit (1 for fully charged, tapering below). That keeps fully charged families most likely, as now, without making taper families structurally ineligible. Opting out has a real cost for them (a Self Assessment return) and HMRC doesn't split opt-outs by income. Either way, check the age profile of opt-outs against the vendored by-age rows, since that's where item 5's gap may sit.
• other_residential_property_value excludes buy-to-let (DVBltVal): fine as long as the buy-to-let value lands somewhere in property_wealth. Item 3's fix matters more here, because buy-to-let is exactly where the dropped mortgages sit.
• Table 8a rows fenced: agree. Leave promotion to UK CGT: sub-exempt gains capped at exactly the annual exempt amount become 10.8m taxpayers under uprating; interpolate the remainder below the threshold and anchor incidence #970 after the final build, with item 4's post-calibration readings as the evidence for it.

Checked and correct

• c1, mass and seating:
  • Donors are funded from each donor's region stratum. Every stratum's and the total household mass are conserved, and the stage refuses if funding would empty an incumbent row (spi_band_donors.py:694-702).
  • Seating is an Efraimidis–Spirakis draw without replacement on identity-keyed uniforms, salted per band, and households already taken are excluded across bands. So no source household seats twice.
• c1, identity keys: donors carry their own support clone index (2), so geography identity keys stay unique.
• c5b, identity and the fold:
  • Arms take id + j × multiplier and carry cgt_residential_clone_index. The geography identity kernel reads both and enforces an int64 index (graph_population.py:84-85, 402).
  • The family fold now folds residential arms onto their household as well as support copies (diagnostics.py:492), so the weight-ratio and ESS gates read the folded weights.
  • k > 3 liable gainers is refused rather than grown.
• c8, rates: the claim rate and opted-out family share match the vendored HMRC rows (item 5 aside).
• c3+c4: the coherence gate and the derived-total identities are asserted at cleaning.
• Tests: all the new-stage test files pass.

[juaristi22]

Thanks, all six addressed in the latest commit; the branch is now on main (#1084 merged) with the same twelve commits plus this one.

1. Carriers (blocker). The five internal disability carriers are off the allow-list and leave at the release boundary with incapacity_benefit_reported (UK_RELEASE_EXPORT_DROPPED_COLUMNS); test_internal_disability_carriers_stay_out_of_export_registers passes, and the terminal-gates test now asserts the boundary drops exactly the reviewed exclusions plus the carriers.
2. Plan-2 renewal. María's call: it stays on this PR, which she expects to merge today or tomorrow, and no dataset is built without it.
3. Mortgages. Taken as suggested, with one difference: the other-property mortgage is HMortGR8 less TotMortR8 on the donor rather than OthMortR8_sum, so the identity mortgage_debt = main_residence_mortgage + other_mortgage holds on every donor row and no mass leaves (the £15bn of HMortGR8 that is neither TotMortR8 nor OthMortR8_sum stays). Only the main-residence mortgage is tenure-stratified; the other mortgages are drawn on every tenure, conditioned on the property the chain already placed, and mortgage_debt is the derived sum. Eight segments (the first seven roots unchanged), the coherence gate holds main_residence_mortgage_off_mortgaged_tenure_rows at zero and mortgage_debt_violation_rows at zero, and the receipt reports the other-mortgage mass held off the mortgaged tenure on each side instead of a dropped mass.
4. By-band identities after calibration. scripts/residential_readout.py on C2's calibrated weights: the two bound national totals hold (count 202,628 vs 202,630; gains £12.246bn vs £12.242bn) and within them the solver moves mass between bands (£250–500k 3,680 → 4,256 taxpayers, £500k–1m 2,004 → 2,173, £1–2m 730 → 552, £2–5m 427 → 274, open band 134 → 139; under £250k within 3%). Row level with the arms in place: smallest positive calibrated weight 0.012 (an arm, 0.050; smallest design arm 0.144), positive-weight median 26.6 (arms 11.8), no row at zero. Both are in the receipts note, with the by-band identities now described as a design-weight property.
5. Children in payment. The 11.1m is at design weights on the 1 October spine, where the eligible-child base is 14.04m; the rates match HMRC and opted-out families carry 1.82 children each, so claimed children are 12.17m and 1.09m of them opted out. On the 30 September build's calibrated weights the base is 14.79m, which at the same rates gives 11.7m against HMRC's 11.73m. The gap is the design-weight child base, not the rates or the opt-out ages.
6. Wording. Done in the docstring; the manifest notes already named the two rows.

On the open calls: the opt-out pool and the Table 8a promotion are María's; the buy-to-let value sits in the drawn remainder of property_wealth, and with item 3 its mortgage now sits beside it.

One more thing the stack arm found: dwp/uc_payment_dist/SINGLE_annual_payment_28_800_to_30_000 has a single supporting household on every main-based spine (a disabled single council tenant, design weight 760 against 1,654). On the stack its award moved £678 lower, out of the band, when the larger donor block reshuffled the rows every later non-identity-keyed draw reads; the cell joins the measure exclusions beside its 27,600–28,800 sibling of the same uk-data#452 class, for María's signature.

🤖 Generated with Claude Code

[vahid-ahmadi]

Automated review pass (Claude Code, high effort) — round 2 at fd13f71e

Verdict: one blocking item, and it's a one-line fix. CI's engine-free job fails on both Python versions because this PR adds two UC payment-band exclusions and a test still pins the old count. Every round-1 item is closed. Main is now past the plan-2 expiry, so this should merge (or the renewal should go to main on its own) as soon as the pin is fixed.

1. Blocking: the engine-free failure is this PR's

test_uk_uc_family.py::test_full_uc_payment_registry_matches_independent_relationship_and_band_cases fails with assert 81 == 83 (run 37081918992, on fd13f71e). I reproduced it locally at this head.

The cause is 747313ae and 55832501, which add two dwp/uc_payment_dist exclusions:

• SINGLE_annual_payment_28_800_to_30_000;
• COUPLE_NO_CHILDREN_annual_payment_22_800_to_24_000.

The test pins 83 active and 17 excluded at test_uk_uc_family.py:355-356, with a comment dated to the 2026-09-21 A16 retirements. The fix is 81 and 19, with the comment naming the two new cells.

Your reply mentions only the SINGLE cell. The COUPLE_NO_CHILDREN cell (two supporting households, 1,025 and 1,999 against 2,761) should be in the body and the receipts as well.

2. Plan-2 exclusions: main is past expiry

On origin/main (c1611fbf), both slc.borrowers.plan_2_liable and plan_2_above_threshold still carry expires_on: 2026-10-03.

apply_uk_calibration_measure_exclusions (measure_simulation.py:714-724) raises a ValueError once the evaluation date leaves approved_on..expires_on. So from today, 2026-10-04, any UK build from main that compiles the national register refuses at that step.

The renewal to 2026-11-03 (05d6c9bf) exists only on this branch. María's call to keep it here is fine, provided this merges today. If item 1 holds it up, the renewal commit alone should go to main.

Round-1 items

Item	Status	Evidence
1 Internal disability carriers on the export allow-list	Closed	All five, plus incapacity_benefit_reported and the new salary_sacrifice_pre_conversion_pay, are in UK_RELEASE_EXPORT_DROPPED_COLUMNS (national_frame.py:466-480). test_internal_disability_carriers_stay_out_of_export_registers passes locally.
2 Plan-2 renewal on main	Open, now urgent	See 2 above.
3 £127bn of other-property mortgages dropped	Closed in code; one receipt line stale	Other mortgage is HMortGR8 − TotMortR8 via _donor_remainder (was_wealth.py:489), which refuses a donor row whose main-residence mortgage exceeds the total. So mortgage_debt = main + other holds on every donor row and no mass leaves. Only main_residence_mortgage is stratified (was_wealth.py:89). However, Arm B in the receipts (1063-uk-spine-followups.md:53-55) still says the rule "drops … £127bn". Your reply says the receipt now reports the other-mortgage mass held off the mortgaged tenure instead, but I couldn't find that figure in the note. Please mark the Arm B line as superseded and add the held mass.
4 Table 8a bands after calibration	Closed as reported	The readout is in the receipts (lines 76-84). The national totals hold and the bands move: £1–2m 730 → 552 and £2–5m 427 → 274. Stating the by-band identities as a design-weight property is honest. Promoting the bands is María's call. One gap: scripts/residential_readout.py is cited but isn't in the tree. Please commit it, or say it lives with the licensed data.
5 Children in payment, 11.1m against ~11.73m	Closed	The explanation is in the receipts (lines 141-146): a design-weight base of 14.04m against a calibrated 14.79m gives 11.7m at the same rates.
6 "Binds" wording	Closed	The split's docstring and notes now name the two national rows.

New since round 1

7. Question: the certifier's last failure. The receipts say the rerun passes 19 of 20, with only uk_input_mass_parity failing, on access_fund +535%, jsa_income_reported +1,058% and working_tax_credit_reported +747%.
   • The analysis that this predates the stack is convincing for JSA and tax credits.
   • On access_fund, one household carries 98% of the mass. Its £1,035.62 a week has FRS period code 5, and the spine annualises it to £54,037.
   • If period code 5 is a calendar month, the weekly figure may itself be a mis-weeklyised annual or termly amount. That would make it a data or period-code fix rather than an outlier fence.
   • Worth checking before María rules. Either way, the PR is titled "release-cut blockers" and the cut still fails one part, so the body should say so.
8. Nit: thin cells flipping on draw order. Both new UC exclusions are cells with one or two supporting households, whose awards left the band because later draws that aren't keyed by identity read their rows in a new order. The exclusions are signed and consistent with the 2026-08-26 siblings (uk-data#452, expiry 2026-11-26). But each upstream stage change can now cost a thin cell. Keying the remaining draws by identity would stop that churn, so it's worth a line in UK spine follow-ups after #1045: SPI band donors as a mass-conserving support channel, WAS tenure coherence, the residential draw's bound and Table 7 composition, then a national build #1063's follow-ups.
9. Note: weight-shape margin. The folded max-to-median is 1,028, against main's 609 and the 1,151 bound. That passes, but it's closer to the bound than main has been. The receipts already flag it, so no change is needed.

Checked and correct

• Determinism fixes: the salary-sacrifice pre-conversion pay carrier and the person-order walk for the residential split, from the E8 diagnosis, both look right. The second build's spine is bitwise identical in every shared column.
• Tests: the targeted suites for the WAS wealth, residential split, Child Benefit, salary sacrifice, terminal gates, battery bindings and source stages tests all pass locally. The only failure is item 1.
• Merge base: the branch is on main after UK pensions: resident State Pension, Pension Credit, contributions and salary-sacrifice relief (#1069) #1084 merged, with 21 commits.

[vahid-ahmadi]

Automated review pass (Claude Code, high effort) — round 3 at 5720b80d

Verdict: nothing outstanding in the code; I'd approve once engine-free is green at this head. I wrote this round against 427bec73; 5720b80d landed while I was writing it and closes the one remaining test failure, so the first row records both. The access-fund record is the one thing I'd still rather fix than defer.

Item	Status	Evidence
Engine-free CI	Fixed at 5720b80d (CI still running at post).	At 427bec73 the UC pins were fixed (81 and 19, test_uk_uc_family.py:360-361), but engine-free failed on test_uk_national_graph.py::test_national_gates_evaluate_the_seam_scope_and_record_the_evidence (assert ['uk_target_fit'] == []): the fixture reviewed exclusions at 2026-09-29 while 747313a signs the SE £12,570–15,000 deferral as approved 2026-10-02, so the gate called it premature. 5720b80d replaces the five hard-coded dates with _review_date(), the latest approved_on across the five registers, and asserts it precedes the earliest expiry. That fixes it without pinning a date the next approval would break, and an expired entry still fails.
New in 5720b80d: certification signatures	Correct, and worth calling out.	_verify_part_signature and _sign_certification now canonicalise with the battery's canonical_report_bytes instead of the logbook's canonical_json_bytes. The two differ on floats (1.0 vs 1, 1e-05 vs 0.00001), so the old verifier would have refused every real battery report while passing float-free synthetic fixtures. I found no other certification verifier still using the logbook form.
access_fund input-mass parity	Closed as a signed deferral; the period fix is owed.	It is now signed: approved_by juaristi22, dated 2026-10-04, expiring 2027-01-04, with the diagnosis. One record carries 98% of the candidate's access-fund mass (£1.01bn against £160m), and calibration takes it and its two CGT copies to the 10× weight bound. So the release would ship £1bn of access fund that is almost certainly wrong. The fix could be small: in the spine, treat this record's access fund as annual (or cap it at the scheme maximum). That would be better than publishing with the deferral. If it stays deferred, the release notes should name it.
New deferrals (jsa_income_reported, working_tax_credit_reported)	Accepted, with a note.	Both are signed, dated, expire 2027-04-04, and give their reasons: the SPI channel's stage-2 QRF imputes these legacy leaves onto taxpayer rows, and the 30 September main candidate was already at +428% and +421%. Since WTC closed in April 2025, the follow-up she lists (the SPI channel's legacy leaves) should land before these expire.
Readout tool	Closed.	tools/uk_residential_readout.py is committed, and --help runs. The by-band numbers need licensed weights, so I didn't reproduce them.
Receipts' £127bn wording	Closed.	Arm B's line is marked superseded (experiments/1063-uk-spine-followups.md:51-60). It now gives the held mass on the final spine: the donor holds £127.3bn off the mortgaged tenure (£111.6bn of it other-property mortgages), and the recipient frame holds £273.1bn on 1,386 households, with zero identity violations.
Plan-2 exclusions	Renewed here; main is still past expiry.	This branch has slc.borrowers.plan_2_liable and plan_2_above_threshold at 2026-11-03. On origin/main (c1611fbf, 4 Oct) both still say 2026-10-03, and no separate main PR exists. So UK builds from main fail at the exclusions step until this merges. So this PR should merge as soon as CI is green, or the renewal goes to main on its own.

Locally at 427bec73: test_uk_uc_family, test_uk_weighted_integrity, test_uk_battery_bindings and the shared test_contract all pass. engine-uk, integration-uk and lint are green in CI; engine-us was still running.

[juaristi22]

Thanks. Round 2 addressed at 5720b80. CI's engine-free lane had two failures, not one: the UC registry pins you found, and test_uk_national_graph.py's seam-scope test, which the path-tiered lane only selected once round 2's diff widened. That test reviewed the real target-fit register at a fixed 2026-09-29, so the South East deferral approved 2026-10-02 was refused as not yet in force; it now reviews at the latest approval across the committed registers, inside every live window (an expired entry still fails, as the real review would).

1. UC registry pins (blocker). test_uk_uc_family.py now pins 81 active / 19 excluded, with the comment naming the two cells. The COUPLE_NO_CHILDREN 22,800–24,000 cell is in the receipts (Stack 2) and now in the body: two supporting households (design weights 1,025 and 1,999 against 2,761) whose awards left the band when the reworked WAS chain reordered the rows the later positional draws read.
2. Plan-2 renewal. María's call stands: it merges with this PR, which she is merging as soon as CI is green; the next national build is from main after the merge.
3. Arm B receipt. Marked superseded, with the final build's figures: the donor holds £127.3bn of mortgage debt off the mortgaged tenure (£111.6bn of it other-property mortgages) and the recipient frame £273.1bn on 1,386 households, with zero main-residence mortgage off the mortgaged tenure and zero mortgage_debt identity violations.
4. Readout script. Committed as tools/uk_residential_readout.py; the receipts cite it there.
5. The certifier's last part and access_fund. You were right to ask about the period code. The FRS 2024-25 variable listing gives ACCSSPD 5 as "Calendar month", and £1,035.62 a week is exactly £4,500 a calendar month (1,035.616438 × 365/7 ÷ 12). A £4,500 access fund a month is not a plausible award; £4,500 a year sits beside the next FRS values (£4,803, £4,654, £3,602), so the record is an annual or termly award under the wrong period code, which makes it a data fix in the spine's period handling rather than an outlier fence. The household's weight is the calibration's: on the 30 September main build the same record sat at design weight because its incidence clone drew £2,355 of gains (under the exempt amount, so the anchor returned the clone's mass); on this build the Table 3 redraw's positional stream, shifted by the 1,920 inserted donor households, gives the clone £3,776, a liable South East gainer with a residential arm, and the calibration takes the household and both copies to the 10x bound. María's ruling (2026-10-04): the deferrals needed to build and publish are signed, each with its diagnosis, in input_mass_reviewed_exclusions.json: access_fund to 2027-01-04 (the period rule is owed before it expires), jsa_income_reported and working_tax_credit_reported to 2027-04-04 (the SPI channel's legacy-benefit leaves; the 30 September main candidate sat at +428% and +421% on the same columns). With the deferrals signed, the certifier rerun from 5720b80 on the second final build's artifacts passes all 20 gates and composes a signed certification (shippable: true; spine, calibration-seam and release-cut parts, 31 entries) for candidate aa31bdf6…. The rerun from the signed head also found a latent certifier defect, reached for the first time because no certification had got past the gates before: the gate battery signs its reports over json.dumps(sort_keys, compact) while the certifier verified the parts, and signed the certificate, over the logbook's Chronicle: one append-only trace for every build attempt at every rung #628 form, which renders 1.0 as 1, so every real spine report failed "signature does not authenticate" while the float-free fixtures agreed. The certifier now verifies and signs over the battery's canonical bytes (canonical_report_bytes, the form the data contract's verifier already uses), with a regression test that puts such floats in a part.
6. Thin cells and positional draws. Agreed; it is in the receipts' follow-ups (keying the Table 3 redraw's within-band draws and the other positional draws by identity), beside the SPI channel's legacy leaves and running salary_sacrifice before the capital-gains stages, which would retire the pre-conversion pay carrier.
7. Weight shape. Noted; no change.

🤖 Generated with Claude Code