UK pensions after #1084: open rulings, residual gaps and the engine pieces

Status (16 October 2026)

This issue began as the plan for #1084 (work packages WP0 to WP7). #1084 merged on 2 October (fe4e38dfa) and delivered WP0 to WP5 and WP7's HBAI and reform-stability checks. Its measurements and rulings are in experiments/1069-uk-pensions-receipts.md, and the original plan is in this issue's edit history. The body now lists only what is still open.

Delivered since #1084

State Pension age by date of birth: Set State Pension age from date of birth, including the rise to 67 policyengine-uk#1899 (merged 3 October), in the lock since UK: port the remaining uk-data fixes tracked in #1095, engine items included #1121. uk_runtime/frs_take_up.py reads age_by_birth_date.
The triple lock computed from its statutory inputs, including the April 2027 rate: Compute the State Pension triple lock from its statutory inputs policyengine-uk#1939 (3 October). Generated-series reforms take effect: Read calendar-year series' bare years as calendar years, and make generated-series reforms take effect policyengine-uk#2404 (14 October). A reform= dictionary still edits parameters after processing, so reforms of the triple-lock inputs go through Scenario(parameter_changes=...).
Salary sacrifice is predicted only for people with pay: UK spine follow-ups: SPI donors funded, WAS coherence, residential split, Child Benefit, release-cut blockers (#1063) #1089 and UK: policyengine-uk 2.130.1 and the fourth batch of #1095 ports, with the release's weight year, identifiers and birth months fixed #1271.
The income-tax relief rate bands bind by tax-rate class (UK: port the #1095 fixes that need no new policyengine-uk release #1200). The all-rates total is a diagnostic, and the family's exclusions are retired.
Employer NI for employees over State Pension age: Charge employer NI on employees over state pension age policyengine-uk#1895 (30 September).

Open in microcosm

Rulings

Single ages (R3). WP2's 3% acceptance for single ages around State Pension age isn't met. Rule on the split (65, 66, 67–69), or bind the singles beside the five-year band.
SPI Table 3.8 contributions. Table 3.8 splits contributions by relief mechanism; the engine's columns split them by provider. Rule on binding the combined amounts against employee plus personal contributions.
Employer NICs relief at 15%. Confirm the restatement from 13.8%, which changes the rate alone.

Work

The State Pension type boundary. Re-measure the 70–74 boundary, which zig-zagged while state_pension_type keyed on age and sex. Set State Pension age from date of birth, including the rise to 67 policyengine-uk#1899 is in the lock, and UK: policyengine-uk 2.130.1 and the fourth batch of #1095 ports, with the release's weight year, identifiers and birth months fixed #1271 adds birth months to the spine.
Pensioner benefit coverage (from UK national calibration moves weight away from low-income households: reconcile the HMRC low-band rows, model UC take-up by award size, impute pensioner benefit receipt (pre-Budget) #1334). The AA, Pension Credit and pension-age HB totals start 20–32% below DWP and OBR before the solve (one Pension Credit age band 43%), and the solve meets them by reweighting. Align the rows' universe with the FRS (care-home recipients), allow at most a small draw for caseload growth, and do not draw onto sampled non-reporters. See the comment.
Pension Credit spending by component. Only OBR's total is bound; the Guarantee Credit and Savings Credit split is not.
The salary-sacrifice anchor. uk/salary_sacrifice_anchor.json is still vendored by hand (unchanged since 21 August). Vendor it from Chronicle; the values already match the feed. WS-C deferred adjudications and the reconciliation layer #736 lists the same item.
State Pension in the SPI bands and below the tax threshold. The £20–30k and £30–40k State Pension bands sit above target on the first national build after Draw SPI-channel incomes at the FRS twin's conditional rank (#840) #1256 (UK SPI channel after #840: residuals, post-merge measurements and stale pointers #1288). Non-taxpayer State Pension is short of what HMRC and DWP imply. UK SPI channel after #840: residuals, post-merge measurements and stale pointers #1288's item on State-Pension-only twins, drawn against a tape of mostly taxpayers, may explain part of both.
WP7's DWP path. Compare the baseline State Pension and Pension Credit path with DWP's forecast to FY2030-31, once the engine items below are in the lock.

Blocked on policyengine-uk

Pension Credit FY2025-26 rates and the Savings Credit projection: Correct 2025-26 Pension Credit rates and the Savings Credit projection policyengine-uk#2083 (open). Until it lands, Pension Credit calibrates at the engine's current values.
Additional State Pension and protected payments uprated by CPI: Uprate additional State Pension and protected payments by CPI, not the flat rate policyengine-uk#1941, with PR Uprate additional State Pension and protected payments by September CPI policyengine-uk#2336 (both open).
Winter Fuel Payment recovery above £35,000: Model Winter Fuel Payment high-income recovery separately from entitlement policyengine-uk#1760 (open). DWP's recipient count stays a diagnostic until it lands.
No policyengine-uk issue or PR found yet (check before filing):
- the basic-to-new State Pension reclassification double count;
- a guard against datasets that carry a state_pension column;
- the income-tax easement switch for pensioners whose only income is the State Pension (original decision 10);
- Pension Age Disability Payment, so Attendance Allowance in Scotland can bind.
The original decisions 9 (the Pension Credit guarantee after 2026) and 10 (the easement switch) belong to the engine side.

Multi-year runs: UK static aging: ONS 2024-based projection reader and weights-only annual projections to 2050-51 #1075 and UK State Pension projection bundle: entitlement sidecar, abroad dataset, manifest and scorecard for PolicyEngine UK #1076 (microcosm-dynamics#501).
Pension wealth: UK source packages: ONS WAS Round 8 household wealth distribution (total wealth, components, percentile aggregates) chronicle#205.
One permanent signed reference exclusion in uk/target_reference_signed_exclusions.json names this issue. It is provenance with no expiry, and needs no action when this issue closes.

Done when

The three rulings are recorded.
The open microcosm items are bound, vendored or re-measured, with receipts.
The engine items are in the lock, and the WP7 DWP path has been run.

Let synthetic smoke builds record the Pension Credit take-up gate wit…

Lower the SPI channel's pension-age prior share to 0.2 (#1069, ruled …

Uprate the SPI State Pension amount rows by their count rows' recipie…

Receipts: review round 1, the σ_p ruling and the State Pension amount…

Merge pull request #1084 from PolicyEngine/uk-pensions-1069

#1084 is merged (fe4e38d). It delivered WP0–WP5 and WP7's HBAI and reform-stability checks; the measurements and rulings are in experiments/1069-uk-pensions-receipts.md. This issue stays open for what's left, most of which is blocked on policyengine-uk.

Blocked on policyengine-uk

Set State Pension age from date of birth, including the rise to 67 policyengine-uk#1899 (open PR: State Pension age by date of birth, including the rise to 67). It blocks:
- c0, microcosm's engine bump with the State Pension age migration. uk_runtime/frs_take_up.py reads gov.dwp.state_pension.age.{male,female}, which #1899 deletes.
- The State Pension type boundary. Until then state_pension_type keys on age and sex. The 70–74 boundary zig-zags, and single ages 69–73 are off by 10–20%.
- Two items of the engine follow-up below: the basic-to-new reclassification double count, and Additional State Pension uprated by CPI rather than the triple lock.
The engine follow-up PR (WP6), to be filed after #1899. Four of its items are drafted locally; nothing is filed yet. The items:
- Reforms of triple-lock inputs propagate.
- The Pension Credit FY2025-26 rates and the guarantee's uprating rule. Until then Pension Credit calibrates at the engine's current FY2025-26 values, which WP3 wanted corrected first.
- The April 2027 rate.
- A guard against datasets that carry a state_pension column.
- The two #1899-dependent items above, and the income-tax easement switch (decision 10).
Model Winter Fuel Payment high-income recovery separately from entitlement policyengine-uk#1760 (Winter Fuel Payment recovery above £35,000). Until it lands, DWP's 11.0m recipients stay a diagnostic.
Pension Age Disability Payment isn't modelled in the engine, and there is no engine issue yet. So Attendance Allowance in Scotland isn't bound.
WP7's DWP path, the baseline State Pension and Pension Credit against DWP's forecast to FY2030-31, runs after the engine follow-up.

Open in microcosm, not blocked on the engine

Single ages. WP2's 3% acceptance isn't met: 65 is −17%, 64 −6%, and 62 and 63 are +15% and +12%. A ruling is owed on the R3 split (65, 66, 67–69).
SPI Table 3.8 contributions are out of the fit, because Table 3.8 splits by relief mechanism and the engine's columns by provider. A ruling is owed on binding the combined amounts against employee plus personal contributions.
The salary-sacrifice amount. £34.3bn is sacrificed against the £24.6bn implied, because the forest predicts sacrifice for people with no pay. The fix is UK spine follow-ups: SPI donors funded, WAS coherence, residential split, Child Benefit, release-cut blockers (#1063) #1089's c7 (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).
The income-tax relief rate bands wait for a band-test fix, and their exclusions expire 2026-11-25. Bands or total needs deciding before it lands, so nothing double-binds.
The employer NICs relief is restated from 13.8% to 15% by the rate only; that ruling is to confirm.
The Pension Credit Guarantee/Savings spending split isn't bound; only OBR's total is.
The salary-sacrifice anchor isn't vendored from Chronicle yet; the values already match the feed.
Remaining gaps. The SPI £20–30k State Pension band is +13.1% (8% too many recipients). Non-taxpayer State Pension is short of what HMRC and DWP imply.

Of the decisions listed above, 9 (the Pension Credit guarantee after 2026) and 10 (the easement switch) belong to the engine follow-up; the plan ruled the rest.

Hold the salary-sacrifice income-tax and employee-NIC reliefs on the …

Bind the salary-sacrifice relief bands by income tax rate class (#1095)

Tie pension contributions to the pay they come from

New work item: pensioner benefit coverage, moved here from #1334 (which closes with #1344 and #1347 merged).

Why it is here

#1334's fix 3 imputed Attendance Allowance to non-reporting pensioners with a care or disability need, and drew pension-age Housing Benefit claims among entitled non-reporters. María dropped it on 22 October. The FRS 2024-25 takes both from administrative data for about 95% of respondents (AA through DWP's General Matching Service, HB through SHBE). So a sampled pensioner who reports neither almost certainly receives neither, and imputing receipt added income where there is none, and moved pensioner poverty away from HBAI. The shortfall against DWP's caseloads is coverage: care-home residents (in DWP's counts, not in the FRS sample), caseload growth since the survey year, and weighting.

Where it stands (#1344 arm at `e399d21ef`, policyengine-uk 2.147.2, 2025; #1338's certified build is the same before the solve)

Row	Before the solve	After
`obr.attendance_allowance`	−28.1%	+0.1%
`dwp.attendance_allowance.recipients_england`	−32.3%	+0.1%
`dwp.attendance_allowance.recipients_wales`	−20.4%	0.0%
`obr.pension_credit`	−26.1%	0.0%
`dwp.pension_credit.benefit_units`	−24.1%	−0.1%
Pension Credit by age band (65–69 to 90+)	−6.4% to −42.7%	within 0.2%
`dwp.hb.households_pension_age`	−21.0%	0.0%
`dwp.hb.amount_pension_age`	−21.8%	−0.1%

The solve meets these rows by reweighting. It up-weights reporters and cuts poor pensioners who receive nothing. In #1334's replica re-solves, the benefit targets as a whole accounted for about 2.9pp of the 4.5pp fall in pensioner absolute BHC poverty between a demographics-only solve and the full solve.

Pensioner poverty in 2024 on the same arm, against HBAI FYE25: relative BHC 12.7% (HBAI 15.9%, −3.2pp) and relative AHC 12.2% (13.9%, −1.7pp). The absolute rates moved up slightly under #1344 (BHC 11.4% → 11.8%, AHC 11.4% → 11.6%), so #1344's fall in the relative rates comes from the poverty line. The remaining gap is on the income and coverage side, not a result of #1344's weighting (#1334 comment).

The work

Align each row's universe with the FRS. Estimate the care-home recipients in DWP's AA and Pension Credit counts (and the HB equivalent, on UK spine: housing benefit £2.7bn against OBR £12.1bn — pensioner and supported-accommodation HB missing after the UC migration #867), then net them out of the targets or declare the reconciliation on them.
At most a small draw for caseload growth since the survey year, on a stated rate. No draw onto sampled non-reporters.
Pension Credit. pension_credit_take_up.py runs on the whole frame after the SPI income chain, and the caseload is still 24% short before the solve. Check whether FRS Pension Credit receipt is matched like AA and HB before changing the draw. Measure how much the SPI copies carry, and check the State-Pension-only boundary in the SPI income draw (UK SPI channel after #840: residuals, post-merge measurements and stale pointers #1288).
PIP and DLA for pensioners. The same check.
Pension-age Housing Benefit is tracked with the rest of HB on UK spine: housing benefit £2.7bn against OBR £12.1bn — pensioner and supported-accommodation HB missing after the UC migration #867, which now has the same lesson.

Acceptance (from #1334's fix 3, restated for coverage)

At design weights, the AA, Pension Credit and pension-age HB caseloads are within ±10% of DWP's, on a universe that the FRS samples.
Between prior and final weights, the mass of pensioners who are poor and receive none of these benefits moves by less than 10%. It fell to ×0.64 when UK national calibration moves weight away from low-income households: reconcile the HMRC low-band rows, model UC take-up by award size, impute pensioner benefit receipt (pre-Budget) #1334 was written.
Pensioner absolute and relative BHC poverty are reported at design and final weights, against HBAI.

🤖 Generated with Claude Code

Declare the age rows' overlap with all taxpayers, and review is_SP_age (

Re-pin the UK spine graph fixture for the vendored age rows (#1069)

Answer the review of #1414: SPI universe evidence, frozen-share note,…

Answer the round-2 review of #1414: README scope, implied range, prov…

Answer the delta review of #1414: England off its regions, structured…

Merge pull request #1414 from PolicyEngine/claude/uk-pensioner-taxpay…

State Pension in the SPI bands and below the tax threshold: measured on #1414 (merged as a172afa)

The shortfall of non-taxpaying State Pension recipients was mostly target error. The SPI 2023-24 State Pension band rows reached calendar 2025 by HMRC's projected growth in all taxpayers in each band. That is 4.8% in total from 2023-24, while HMRC projects taxpayers of State Pension age 10.4% higher (ITL July 2026, Table 2.1).

#1414 makes two changes:

It binds HMRC's resident taxpayers of State Pension age, and those aged 65 and over, over the twelve region-tier areas (Table 2.2).
It moves every SPI State Pension band by that level's factor, ×1.0835. The band shares stay at 2023-24.

Two local national builds measure the effect. They share the spine, policyengine-uk 2.162.8 and feed 272c987, and are built with --no-staging --no-publish-orrery: no release.

	main `7d45c0c`	#1414 at `8925039`
Taxpaying State Pension recipients (HMRC-implied 8.65m–8.81m)	8,839,850	8,743,572
Non-taxpaying recipients (HMRC-implied 3.58m–3.74m)	3,530,991	3,628,068
All recipients (DWP + DfC target 12,383,022)	12,370,841	12,371,640
Taxpayers of State Pension age, 12 areas (target 8,841,750)	8,932,401, not bound	8,833,654
SPI State Pension count, £20–30k	+17.9%	+4.1%
SPI State Pension amount, £20–30k	+20.0%	+6.1%
SPI State Pension count, £30–40k	+0.8%	−0.1%
SPI State Pension amount, £30–40k	+2.2%	+1.5%
SPI State Pension count, 12 bands, mean absolute error	1.96%	0.50%

The HMRC-implied range starts from SPI 2023-24's 7.98m taxpayers with a State Pension, which is 97.79% of ITL Table 2.1's 8.16m taxpayers of State Pension age. Holding that share, it is applied to the calendar-2025 window of two levels:

the twelve areas' residents, which gives 8,646,711 taxpaying recipients;
the United Kingdom row, which gives 8,806,360.

Non-taxpaying recipients are the 12,383,022 recipients minus each end. Both of the #1414 build's counts fall inside the range, and both of main's fall outside it (hmrc_implied_recipients.json in the evidence).

PR #182's sole-income State Pension allowance (uk-autumn-budget-dashboard-2026) was run through policyengine.py 6.11.1 on each build:

	main	#1414
2027-28	1,234,867 people, −£84m	1,326,038, −£93m
2031-32	1,731,415 people, −£330m	1,841,789, −£361m

The receipts and scripts are in docs/evidence/uk-pensioner-taxpayers-1069/.

Still open on this item

The £20–30k State Pension band is still 4.1% above its count target and 6.1% above its amount target.
The band shares are a 2023-24 snapshot. Pilot (not merged): bind the SPI pension band rows at 2023-24 (option 1) #1413 binds each SPI band row at its own year, which removes the band projection altogether.
SPI private pension amounts at £15–20k got worse, from −7.8% to −11.1%. They still move by Table 2.5's all-taxpayer band growth, which is the same projection problem Pilot (not merged): bind the SPI pension band rows at 2023-24 (option 1) #1413 pilots.
UK SPI channel after #840: residuals, post-merge measurements and stale pointers #1288's hypothesis about State-Pension-only twins was not tested here.

Answer the independent review and Vahid's review of #1420

Navigation Menu

Repository navigation

UK pensions after #1084: open rulings, residual gaps and the engine pieces #1069

Status (16 October 2026)

Open in microcosm

Blocked on policyengine-uk

Related

Done when

Why it is here

Where it stands (#1344 arm at `e399d21ef`, policyengine-uk 2.147.2, 2025; #1338's certified build is the same before the solve)

The work

Acceptance (from #1334's fix 3, restated for coverage)

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

UK pensions after #1084: open rulings, residual gaps and the engine pieces #1069

Description

Status (16 October 2026)

Open in microcosm

Blocked on policyengine-uk

Related

Done when

Activity

juaristi22 commented on Oct 2, 2026

juaristi22 commented on Oct 23, 2026

Why it is here

Where it stands (#1344 arm at e399d21ef, policyengine-uk 2.147.2, 2025; #1338's certified build is the same before the solve)

The work

Acceptance (from #1334's fix 3, restated for coverage)

MaxGhenis commented on Oct 28, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Where it stands (#1344 arm at `e399d21ef`, policyengine-uk 2.147.2, 2025; #1338's certified build is the same before the solve)