[juaristi22]

What and why

Budget 2026 raises every company car appropriate percentage by one point in 2030-31: Table 6.1 row 86, +£310m in 2030-31 and +£345m in 2031-32 (OBR split: income tax 226.2, NICs 92.4). The dashboard can't score it (PolicyEngine/uk-autumn-budget-dashboard-2026#316), because neither the engine nor the data had company cars.

• The engine. Model the company car and car fuel benefits, with Class 1A NICs policyengine-uk#2750 shipped in 2.165.0. It taxes the company car and car fuel benefits, and the Class 1A NICs on them, from nine person inputs that default to no car.
• The data (this PR). It locks 2.165.0 and adds a last stage, company_cars, that fills six of those inputs.

Without the stage, a file on 2.165.0 still scores the rise at £0.

Status: draft

1. The HMRC facts are in. The feed is re-pinned to 08962f1, the merge of Add HMRC benefits in kind statistics, July 2026, and keep one OBR EFO number per line and year chronicle#399 on top of UK facts for chronicle#395: Northern Ireland domestic rates, Welsh bus journeys, the Childcare Offer for Wales, VED by payer and CPI gas weights chronicle#396 (1,512,256 rows; the schema is unchanged). uk/hmrc_benefits_in_kind_facts.json vendors the 96 2024-25 rows the stage reads, and is registered in country_package.json.
   • The other 35 vendored resources, and the DESNZ constituency resource, carry the same values under the new identity.
   • The local surface is unchanged apart from its feed label and the DESNZ keys.
   • The coverage manifest and the H2 fixture are regenerated.
2. The OBR windows read both editions. Full eCPS input coverage: 158 required columns, 8 evidence-backed exclusions #399 keeps one OBR EFO number per line and year (OBR EFO March 2026: drop the years the October 2026 vintage also publishes (PolicyEngine/microcosm#1095) chronicle#397), so March 2026 now carries only FY2024-25 and October 2026 carries FY2025-26 on. The 19 OBR targets that window both years into calendar 2025 could no longer compile, since their window took both years from one series. A new selector key, source_table_by_opening_year, takes FY2024-25 from March's table and FY2025-26 from October's (María's ruling of 31 October).
   • Fifteen of the 19 move by under 1%: income tax −0.5%, UC +0.7%. The four larger moves are domestic rates (+3.2%), Pension Credit (+2.3%), Jobseeker's Allowance (−2.5%) and SDLT (−1.2%).
   • Three report-only diagnostics pinned March's FY2025-26 forecasts, which the feed no longer carries: obr.capital_gains_tax, obr.state_pension and the obr.ni total. They now pin October's outturns. None of them is in the fit.
   • obr.winter_fuel_allowance reads October's FY2025-26 (+0.02%).
   • The national membership also catches up with policyengine-uk 2.164.1's October-EFO indices, which Budget 2026 ISA data, PR B of #1418: ISA subscriptions per adult, the Lifetime ISA anchor, additive ISA income, policyengine-uk 2.164.2 #1421 locked after Budget 2026 slower data fixes, PR C of #1418: electricity at the ONS level, pension income by area #1420 had generated it: the eight SPI Table 3.10 self-employment bands move −0.55% and the thirteen dividend bands under 0.001%. The compile-parity receipt, which builds check against, already had these values.
   • The national references, membership and compile-parity receipt are regenerated from the 08962f1 national cut.
3. The build. A pair build on 2.165.0 against a control on the same base, with the results below. The machine order is María's.

Merge order

• Base. Main after Budget 2026 ISA data, PR B of #1418: ISA subscriptions per adult, the Lifetime ISA anchor, additive ISA income, policyengine-uk 2.164.2 #1421: it carries Budget 2026 ISA data, PR B of #1418: ISA subscriptions per adult, the Lifetime ISA anchor, additive ISA income, policyengine-uk 2.164.2 #1421's 2.164.2 lock, so this PR's lock step is 2.164.2 → 2.165.0.
• UK: take up #1095's actionable items on policyengine-uk 2.165.1 #1422 (draft) locks policyengine-uk 2.165.1. Whichever of the two lands second:
  • merges main and keeps the newer lock and stamps;
  • regenerates the shared surfaces: the gates.json digests, the coverage manifest, the H2 fixture, and the national references, membership and compile-parity receipt.
• Two overlaps with UK: take up #1095's actionable items on policyengine-uk 2.165.1 #1422.
  • It signs the nine company_car_* inputs at their default (UK_ENGINE_DEFAULT_INPUTS). Whichever lands second removes only the six this stage writes. Availability, private-use payments and the 6 April 2020 registration flag stay signed.
  • It also changes the calendar-year window code in ledger_targets.py and its tests, as this PR does for the table map. Whichever lands second expects a conflict there and keeps both changes. UK: take up #1095's actionable items on policyengine-uk 2.165.1 #1422 stays on Chronicle 53815b2 (María's call), so the table maps arrive with this PR.

Changes

1. policyengine-uk 2.165.0.

   • The lock, the floors, the version mentions and the US worker identity's lock digest.
   • The uprating and bonus-sacrifice pins keep every value.
   • Concept coverage gains the nine company car inputs.
   • 2.164.3 also counts other_investment_income in market and HBAI income, which the spine fills (UK: data fixes for the Budget 2026 household measures #1418), so HBAI-based outputs move too.

2. The FRS job tab's company car answers. The root stage carries four columns per person from their car job with the lowest JOBTYPE (one car per person, as in the engine):

   • the CARVAL band, or "unknown";
   • FUELTYP, or "unknown";
   • EXPBEN03 (private fuel);
   • CARAMT where CARCON says the employee contributed.

   They are dropped at the release boundary.

3. The company_cars stage, the last in the roster, after isa_subscriptions, whose income cells read employment_income. Nothing reads the pay it rewrites. Every draw is keyed on identity.

   • Holders. FRS-channel people keep their car. An SPI-channel person keeps their twin's only where their drawn EPB is positive.
   • Fuel. From FUELTYP: a hybrid is an alternative fuel; a diesel meets RDE2. An unknown fuel draws at HMRC Table 2C's shares.
   • List price. Drawn within the CARVAL band from WAS round-8 company cars' list prices, grown by the engine's CPI index from 2021. The open and unknown bands draw by fuel class. A cell with fewer than 3 WAS households (the floor of 3) is drawn uniformly.
   • CO2. Zero for an electric car. For the others, HMRC Table 2A's shares among cars that emit CO2, hybrids first into 1 to 50 g/km, uniformly within each band; a car at 1 to 50 g/km takes a 50-mile electric range.
   • Defaults kept. Availability, payments for private use and the 2020 registration flag keep the engine's defaults.

4. SPI-channel pay. It is the tape's PAY + EPB + TAXTERM, so the engine's car and fuel benefits come out of employment_income (María's ruling of 30 October). They are capped at EPB, and at the pay the row's pension deductions leave: salary_sacrifice keeps employee contributions and salary sacrifice within pay and bonus sacrifice within half of it, and the carve-out keeps them there.

   • Private fuel goes first where its benefit alone exceeds the cap.
   • A car that still doesn't fit has its list price scaled down, with its contributions dropped.
   • Below £5,000 the car is dropped.

   The stage then has the engine price its output again: on every SPI-channel row, pay plus the engine's benefits must equal the pay the row brought in.

5. Targets. hmrc.spi.employment_income.amount_by_total_income_band and the local hmrc.employment_income.amount read employment_income_with_benefits_in_kind_annual, a new named term for employment income plus benefits_in_kind, which HMRC's employment income counts. The two count rows keep employment_income_annual.

6. The facts. hmrc_company_car_statistics.py reads Full eCPS input coverage: 158 required columns, 8 evidence-backed exclusions #399's 2024-25 rows by their source record ids, in persons and pounds:

   • a cell the feed lacks is refused, never read as zero;
   • a resource whose Table 2A or 2C total, or electric count, sits from Table 1 beyond each sheet's rounding is refused too.

7. Gates. uk_stage_company_cars_invariants is release-blocking and runs on the synthetic smoke too. It fails on:

   • a non-finite or negative value;
   • a holder without a price;
   • inputs on a person without a car;
   • an unknown fuel;
   • a CO2 figure missing for a car that emits it, or set for an electric car;
   • private fuel on an electric car;
   • contributions above the price;
   • an electric range outside 1 to 50 g/km;
   • an SPI benefit above EPB;
   • SPI-channel pay plus the engine's benefits, priced again on the output, more than 5p from the pay the row brought in (single precision).

8. Measured, not bound. No HMRC company car fact is a calibration target: the stage draws to Tables 2A and 2C, so binding them would be circular (María's ruling of 28 October on the ISA facts). The calibrated file is reported against Tables 1, 3A and 4.

Results

PENDING the pair build. Every figure is checked for disclosure before it goes in.

• Holders against HMRC's 920,000 (2024-25 provisional), and by income range against Table 3A.
• Car and fuel taxable value, income tax and Class 1A against Table 1 (£3,070m, £1,090m, £420m).
• The list price base, electric and not (the receipt's list_price.weighted_total_gbp). Two published comparators frame it: HMRC's 2024-25 taxable value implies roughly £27bn to £37bn, and the OBR's NICs line about £60bn.
• The static 2030-31 cost of COMPANY_CAR_TAX_RISE_2030 against Table 6.1 row 86 (+£310m). Expect it well under £310m, from the size of the base (docs/uk-company-cars-316.md, "What to expect of the 2030-31 score"). The file holds the build year's fleet: fuel and CO2 are frozen, prices grow by CPI, and the weights are not projected. A file that matches Table 1 exactly should still fall about 30 to 45% short.
• Gate and fit movement:
  • the re-pointed SPI rows, obr.income_tax, and obr.ni_employer (Class 1A);
  • HBAI income. 2.164.3 adds other_investment_income to market_income, household_market_income and hbai_household_net_income, and 2.164.4 to capital_income. Those feed stage predictors, so the draws below can move for households with other investment income:
    • was_wealth, which materializes household_net_income and capital_income;
    • was_lisa (household_net_income);
    • lcfs_consumption and etb_services (hbai_household_net_income).

Tests

• New:

  • test_uk_company_cars.py, engine-free and engine (22 tests);
  • test_uk_company_cars_stage_health.py;
  • the carrier tests in test_uk_frs_spine.py;
  • the named term in test_uk_measure_simulation.py.

• The engine tests on 2.165.0 show that:

  • the engine reads the stage's inputs;
  • SPI pay plus the engine's benefits is the tape's total to the penny;
  • the 2030-31 rise reaches the holders alone.

• The engine-free suite, run locally in chunks on the re-pin, failed only on the national surface this PR then regenerated, and on three local-only checks:

  • the uk-extra-absent test, since this environment installs the extra;
  • a live release-pointer test;
  • the orrery check, whose Node package isn't installed in this checkout.

  After the regeneration, every file that failed passes, as do the touched surfaces' tests.

• The 16 feed-gated files, one process each against the pinned feed: 250 passed. CI skips them.

• The national compile-parity regeneration on the full pinned artifact, with its release gate, passes.

• The H2 parity test passes on the regenerated fixture.

• tools/run_integration_tests.sh was not run locally; CI's integration-uk runs it.

Disclosure

The FRS and WAS are UK Data Service data (SN 9563, SN 7215). The stage's receipt carries weighted aggregates, rounded counts and support bands only. The HMRC figures quoted here are published.

Not in this PR

• Class 1A is not passed through to households (Pass Class 1A NICs through the employer NI incidence policyengine-uk#2752). The 2030-31 rise is scored statically, comparable with the OBR's costing.
• Vans, fuel provided for part of the year, optional remuneration arrangements, and the rest of the "not modelled" list in policyengine-uk#2750.
• Whether to anchor holders to Table 3A by income range, if the FRS count falls short of HMRC's. That is María's call after the results.
• The dashboard measure: policyengine.py certifies a build on ≥2.165.0, then the dashboard adopts the release and adds the measure.

🤖 Generated with Claude Code

[vahid-ahmadi]

Review of 21e3623 (draft, so an early look as a comment). First round; there are no earlier comments.

The stage writes what policyengine-uk 2.165.0 expects. I checked the nine inputs at the 2.165.0 tag. The six written ones have the right units, periods and enum names (GBP stocks uprated by CPI, g/km, miles, CarFuelType names, bool). The draws are keyed on identity, there is one car per person by lowest JOBTYPE, household weights pass through, and no company car fact is bound. The stage is the last spine stage, so calibration still runs after it: it adds no weights, but the solve then moves holders, which fits "measured, not bound". There are four shoulds and one nit.

Findings

1. should: on a frozen fleet, expect the static 2030-31 score well under £310m. The gap comes from the size of the base, not the EV share.

• The rise is one point in every band, zero-emission and the maximum included. So the extra taxable value is about 1% of total list price (less contributions), plus 1% of the fuel multiplier for each fuel-benefit holder. The frozen EV share barely moves the rise. It moves the baseline level instead (EVs go from 3% to 9% by 2029-30).
• The base the rise acts on is the build year's fleet, carried forward to 2030-31. Fuel type and CO2 are not uprated, list price and contributions rise by CPI, and the weights are not projected.
• Rough HMRC-implied base for 2024-25: Table 1 car taxable value £3,070m over Table 2A's band-weighted 2024-25 appropriate percentages gives about £27bn to £37bn of list price (depending on PHEV range and EV price premium). With CPI to 2030-31 and 40,000 fuel holders, that is about £330m to £440m of extra taxable value. At HMRC's 2024-25 ratios (income tax 35.5% of taxable value, Class 1A 15%), that is roughly £170m to £225m.
• The OBR's NICs line (£92.4m / 15%) implies about £616m of extra taxable value, roughly a £60bn base. Its income tax line gives the same 36.7% ratio. So even a file that hits HMRC's 920,000 holders and £3,070m exactly should come in around 30 to 45% short. Recipients rose from 840,000 to 920,000 in one year, and HMRC flags under-reporting of payrolled cars.
• Matching Table 1's taxable value does not pin this base: a file can hit £3,070m with a different EV mix and different prices. Suggest the receipt add weighted total list price (by electric and not), so Results can compare it with the HMRC-implied range and read the £310m gap as base size.

2. should: the SPI pay-identity check in the release-blocking gate can't fail.
company_cars.py:983-987 sets new_pay = pay - carve.benefit, then measures |new_pay + carve.benefit - pay|, so max_abs_pay_plus_benefit_error_gbp is zero by construction. That means uk_stage_company_cars_invariants (stage_health.py:4954-4960) never tests the claim in its notes. The real risk is the engine's benefit on the written inputs differing from the carved amount. The price scaling assumes the defaults for availability and private-use payments, and that contributions were dropped, so a later engine change would break it without the gate noticing. Suggest one more materialize of company_car_benefit + car_fuel_benefit on the output frame, comparing each SPI-channel row with carved.

3. should: one CI failure is not on the body's expected-red list.
engine-uk fails on tests/engine/uk/test_uk_annual_income_measures.py:156: set(receipt["computed_measures"]) == set(UK_ANNUAL_INCOME_MEASURES). The test never computes the new employment_income_with_benefits_in_kind_annual. This is independent of chronicle#399. Computing it there, and asserting it equals employment_income_annual when nobody has a car, would fix it. The job stops at this first failure ("1 failed" and nothing passed), so CI has not run the new engine tests in test_uk_company_cars.py yet.

4. should: the lock step pulls in more than the body lists.
2.164.2 → 2.165.0 includes 2.164.3 (other_investment_income in market_income, household_market_income and hbai_household_net_income) and 2.164.4 (also in capital_income). The body mentions only HBAI outputs. These engine values are stage predictors:

• was_wealth materializes household_net_income and capital_income (was_wealth.py:63-89);
• was_lisa reads household_net_income (was_lisa.py:164);
• lcfs_consumption and etb_services use hbai_household_net_income (lcfs_consumption.py:650, etb_services.py:62).

So the wealth, LISA, consumption and services draws can move for households with other investment income. Suggest adding these to Results' "gate and fit movement" list for the pair build. 2.165.0 also adds Class 1A to ni_employer, which the body already covers.

5. nit: unknown-fuel draws never produce a hybrid.
company_cars.py:588 draws only ELECTRIC, PETROL or diesel, and :920 flags hybrids from the FRS code alone. So "petrol/diesel including hybrids" from Table 2C never feeds the hybrids-first split into 1 to 50 g/km. The effect is small, but worth a sentence in the doc.

Checked and fine: the three defaults (availability 1, no private-use payments, the 2020 flag) only matter in 2020-22 or slightly raise each car's taxable value for part-year holders; the WAS donor (cars not vans, household weight, floor of 3); CARVAL/FUELTYP mapping and the open band; CO2 bands within Table 2A's bounds; the SPI carve-out arithmetic (fuel first, exact price scaling, £5,000 floor); the re-pointed SPI targets, since taxable_employment_income now includes benefits_in_kind, so total-income bands agree.

Verdict

Sound design. Items 2 and 3 are small code fixes. Item 1 is about setting expectations before the pair build, and item 4 is about what that build should report. Not ready until the #399 vendoring, the regenerations and the build, as the body says.

What I ran

Fresh clone at 21e3623 and uv sync --all-packages --locked (no UK engine extra), deleted afterwards; one process at a time.

• Engine-free pytest: test_uk_company_cars.py and test_uk_company_cars_stage_health.py gave 24 passed. With test_uk_frs_spine.py, test_uk_measure_simulation.py, test_uk_source_stages.py, test_uk_terminal_gates.py, the ISA, local-target, private-school and stochastic tests, test_engine_core_pins.py, test_country_spec.py and microcosm-data's test_contract.py, 848 passed and 4 failed. All 4 are KeyError: 'hmrc_benefits_in_kind_facts.json', as expected.
• I read the 2.165.0 variables, parameters and the company_car_tax_rise_2030 scenario at the tag, and the 2.164.2...2.165.0 compare.
• CI: engine-free fails on the missing facts and integration-uk on the same resource (expected), engine-uk as in item 3; lint, wheels and engine-us pass.

Not verified: the engine tests (no UK engine locally, no build or Modal), any FRS or WAS figure, and the score itself. Item 1's figures are rough arithmetic from published HMRC and OBR numbers.

🤖 Generated with Claude Code

[juaristi22]

Thanks, Vahid. All five items are addressed. Item 2 led to one more fix, and to a disclosure pass on the stage's receipt.

1. The 2030-31 expectation and the list price base.

• The receipt now records list_price.weighted_total_gbp for electric cars and for the others, each with its support band.
• docs/uk-company-cars-316.md has a new section, "What to expect of the 2030-31 score". It says why the rise acts on list price, why the frozen electric share moves the baseline but not the rise, and why the static cost should come in well under £310m: the OBR's NICs line implies a base of about £60bn, against roughly half of that from HMRC's 2024-25 taxable value. It credits your arithmetic for the 30 to 45% range.
• The body's Results now compare the receipt's base with that range, so the gap reads as base size.

2. The pay identity check couldn't fail.

• The stage now prices its output again on the engine. On every SPI-channel row, pay plus the engine's company_car_benefit and car_fuel_benefit on the written inputs must equal the pay the row brought into the stage.
• The receipt records the largest miss as spi_carve_out.max_abs_pay_identity_error_gbp, rounded up to the penny. uk_stage_company_cars_invariants allows 5p, for the engine's single precision (maximum_pay_identity_error_gbp: 0.05). The battery and certification digests are re-pinned for that gates.json change.
• A new test runs the stage with an engine stub whose car benefit ignores the list price. The check reads a miss above £1,000, so it fails as it should.
• While doing this I found one more thing to fix. salary_sacrifice keeps employee contributions and salary sacrifice within pay, and bonus sacrifice within half of it, and its comment asks any later pay rewrite to hold those bounds. The carve-out didn't. Its budget now also leaves the pay those deductions need, and the receipt counts the rows where that binds (limited_by_pension_deductions).
• The disclosure pass covers every weighted figure in the receipt:
  • each one is withheld when a cell it rests on has fewer distinct source households than the floor, so support copies count once;
  • shares that sum to one are withheld together;
  • invariant counts are zero exact and banded otherwise.

3. The engine-uk failure. test_uk_annual_income_measures.py now computes employment_income_with_benefits_in_kind_annual and checks that it equals employment_income_annual when nobody has a car. Locally on 2.165.0, that file gives 4 passed and the company cars engine tests 3 passed.

4. The 2.164.3 and 2.164.4 predictor moves. They are in Results' gate and fit movement list:

• was_wealth (household_net_income, capital_income);
• was_lisa (household_net_income);
• lcfs_consumption and etb_services (hbai_household_net_income).

5. Hybrids. The doc now says that an unknown fuel never draws a hybrid, since Table 2C counts hybrids inside petrol and diesel, and that the effect is small.

What I ran:

• engine-free company cars and stage-health tests: 28 passed;
• the engine tests above;
• the contract, gate battery pin, country spec, release certification and terminal gate tests: 523 passed.

The chronicle#399 re-pin and the BiK vendoring are pushed alongside. So is the fix for the 19 OBR windows that #397 split across the March and October EFOs: they take FY2024-25 from March and FY2025-26 from October (María's ruling of 31 October). The body's Status says what each moved.

🤖 Generated with Claude Code

[juaristi22]

Independent review of #1430 at e67e76c29

This is an independent adversarial review. Four reviewers each took one area:

• the stage;
• facts and windows;
• a cross-cutting pass;
• conventions.

A verifier then reproduced each finding on synthetic data or published tables, and a last pass looked for gaps. Vahid's round isn't repeated. Nothing here uses licensed data. Paths are relative to packages/microcosm-build/src/microcosm/build/ unless they start with tools/ or docs/.

Verdict: two things to fix before the pair build (1 and 2), and two that will skew how its Results read (3 and 4).

Fix before the build

1. obr.ni_employer now counts Class 1A; OBR's line doesn't.

• The row binds ni_employer (uk/uk_population_targets.json:2111). In policyengine-uk 2.165.0 that variable adds ni_class_1a, which is the rate × benefits_in_kind.
• OBR's line is Class 1 employer NICs only. Class 1A sits in "Other NIC" (October EFO table 4.4, row 21, footnote).
• Once this stage fills the car inputs, the measure gains Class 1A on them, about £420m in 2024-25 on HMRC Table 1, and the solve moves weight to absorb it.
• Fix: bind ni_class_1_employer. The notes at :677 and :34829 still say no household carries Class 1A.

2. The carve-out breaks the identity-stability tool.

• tools/verify_uk_identity_stability.py:2555 requires that pay plus salary sacrifice equal the pre-conversion pay exactly, on every converted record.
• company_cars lowers SPI pay after salary_sacrifice has run, and keeps no carrier for the carved amount. salary_sacrifice converts SPI rows too (salary_sacrifice.py:658).
• So a converted SPI person with a car makes --check e4, e5 and e8 raise. At full scale that overlap is near-certain.
• e8 also re-bands CGT on pay the CGT stages never saw.

What the Results will measure

3. SPI copies keep a car only on an independent EPB draw (uk_runtime/company_cars.py:991).

• A copy keeps its twin's car only where its drawn EPB is positive. EPB is drawn on identity-keyed uniforms that are never conditioned on the car (spi_rank_transport.py:123; spi_income.py:168-175).
• The SPI channel holds half of the prior working-age mass (sources.yaml:1415-1416). So before calibration, holders and list price come to about (0.5 + 0.5p) of the FRS level, where p is the share of copies that draw a positive EPB.
• No car target restores this.
• The receipt gives only a rounded row count (spi_dropped_without_epb). Meanwhile the doc's "What to expect" puts the whole shortfall against £310m down to base size.
• Suggestion: record a weighted figure for the dropped holders, or condition the keep on the twin; either would make this measurable.

4. Assumptions documented as "baseline-only" drive the 2030-31 base, through the price scaling (:918-925).

• A scaled SPI car's price is the room ÷ the appropriate percentage at 2024-25 rates, and the 2030-31 rise is 1% of price.
• On a £2,000 room and a £45k car, the scaled price is:
  • £25,000 at 8% (the fixed 50-mile range);
  • £14,286 at 14% (under 30 miles);
  • £45,000, unscaled, for an electric car at 2%;
  • £8,000 for an RDE2 diesel at 25%;
  • £6,897 at 29% for a diesel without RDE2.
• That contradicts sources.yaml:6360 ("matters only to 2027-28") and docs/uk-company-cars-316.md:40-41 and :86-87.

5. The CO2 receipt predates the carve that skews the fleet (:1009-1017, :1147).

• The carve drops or scales cars in inverse proportion to their appropriate percentage. So electric and low-CO2 cars survive, and high-CO2 cars drop first.
• On synthetic data, the kept 1-50 g/km share rises from 0.39 to 0.50, and the electric share from 0.25 to 0.38.
• The receipt reports the pre-carve shares, so "matches Table 2A" isn't true of the output.

6. The list-price base is read at pre-calibration weights (:1134).

• list_price.weighted_total_gbp is at the stage's weights, which carry the SPI channel's prior 50%. Calibration moves holders after this point, and nothing measures list price on the calibrated file.
• Yet the doc (:96-99) and the body plan to read it as the score's base.

Stage behaviour

7. Fuel-first ordering drops cars that would fit without fuel (:910-928).

• With EPB of £9,000, an appropriate percentage of 29%, a £35k car and an £8,062 fuel benefit, nothing is kept. Without the fuel flag, the car is kept at £31,034.
• This follows the 30 October ruling's wording literally, but the result isn't monotone.
• Related: an unknown-fuel holder who reported private fuel draws electric at Table 2C's share (about 51%) and so loses that fuel (:1018).

8. Count rows drop SPI rows whose pay the carve takes to zero.

• The employment-income counts filter on employment_income_annual != 0 (uk/uk_population_targets.json:3696, :31014), while the amounts read the new term.
• An SPI row whose pay is all EPB can end with pay 0 and its car kept: carve_spi_benefits(epb=2000, pay=2000, retained=0, car=6000) does this. Such a row leaves the counts but stays in the amounts.
• Suggestion: filter the counts on the new term.

9. The band-1 fallback draws prices from £0 (:197, :735).

• With no band-1 WAS donors, prices are drawn uniformly on (0, £10,000]. About half fall below £5,000, the least price the carve treats as plausible.
• A CARAMT capped at a near-zero price then zeroes the benefit.

10. Carving last leaves upstream pay rules unrestored (:188-192).

• These were all drawn on pre-carve pay:
  • the bonus-sacrifice candidacy (pay ≥ UEL, salary_sacrifice.py:286);
  • UC take-up awards;
  • the Carer's Allowance earnings test;
  • the gainful self-employment flag.
• So "nothing reads the pay it rewrites" holds only for later stages.
• Suggestion: split the car part out of EPB inside the SPI income stage. That keeps these rules consistent and also clears item 2.

11. SPI-channel rows that aren't tape rows lose the twin's car (:538-543, :992).

• Undrawn copies and dependants aged 16 to 19 keep their twin's pay with EPB 0, so no_epb drops the car.
• spi_income documents person_is_spi_income_recipient for exactly this distinction, and property_components uses it.

Coverage and texts

12. company_cars has no coverage-manifest family.

• Of the 15 stages that declare record_mass_conservation_receipt, it is the only one without a family_coverage entry. tools/build_uk_release_input_coverage_manifest.py:931 lists isa_subscriptions but not this stage.
• So release_input_coverage.py never checks its mass record, nor requires it in the plan.

13. Small items:

• the annual measure scales pay by employment_income_annual_share but not the car inputs, which puts a full-year benefit on part-year FRS pay (small);
• the BIK fact fence's count is popped but never compared;
• company_cars.py:196-216 re-declares FRS constants from frs_spine.py:957-963 (one of them unused), against docs/shared-constants.md;
• _STAGE_HELPER_MODULES["company_cars"] omits uk_engine_adapter and salary_sacrifice;
• the stage notes omit the pension-deduction cap and the re-pricing check;
• graph.py:586 still says one engine materialization.

Cross-PR with #1422

• Council tax nation rows. UK: take up #1095's actionable items on policyengine-uk 2.165.1 #1422's three OBR council tax nation rows need this PR's source_table_by_opening_year map, or the merged tree refuses. This is posted on UK: take up #1095's actionable items on policyengine-uk 2.165.1 #1422.
• Company car defaults. The guard at test_uk_terminal_gates.py:512 can't see this stage's outputs. So UK: take up #1095's actionable items on policyengine-uk 2.165.1 #1422's nine company car inputs signed at default would merge silently with the six this stage writes; this is also posted on UK: take up #1095's actionable items on policyengine-uk 2.165.1 #1422. The other three keep a reason that becomes false ("every list price is zero"), and the approximation they make (full-year availability, no private-use payments) isn't stated.

Checked and found true

• The OBR windows. The weights are 3/12 and 9/12, and income tax reproduces exactly. Editions match by row header for all 19 rows, the refusals work, and the value moves are as stated.
• The BIK facts. There are 96 rows, and the tolerances match Chronicle's stated rounding. The re-pin changes only feed identity on the other 35 resources, and only the keys of the DESNZ resource.
• The stage's mechanics:
  • the engine inputs' names, units and periods match 2.165.0;
  • the price scaling is exact;
  • the float32 re-pricing error stays under 2.3p;
  • the draws are deterministic;
  • the carriers are dropped at the release boundary.
• Fixtures. The H2 fixture CSVs and the test frames are synthetic.

Disclosure-control items went to the author separately.

🤖 Generated with Claude Code