Band savings and dividends above property income

[MaxGhenis]

Fixes #2351

axiom: ukpga/2007/3/16 encoded-correct (rulespec-uk c65fd7df: uk/statutes/ukpga/2007/3/16.yaml, and uk/statutes/income_tax/individual/savings_dividend_oracle_pipeline.yaml stacks savings and dividends on all non-savings income, with no separate property slice)

What changes

ITA 2007 s. 16 treats savings and dividend income as the highest part of total income, with dividends above savings. Property income is non-savings income, so it fills the bands below them, as pay does. That holds in 2026-27 and, with property income's own rates, from 2027-28.

Since #1408, earned_taxable_income subtracts taxable_property_income, because property_income_tax taxes property income separately. But these variables stack savings or dividends on earned_taxable_income:

• basic_rate_savings_income_pre_starter, basic_rate_savings_income, higher_rate_savings_income, add_rate_savings_income;
• the starting-rate taper in savings_starter_rate_income;
• dividend_income_tax.

So savings and dividends were banded as if property income were not there.

• New taxable_property_income_after_allowances: taxable property income less the allowances it absorbs. This is the slice property_income_tax already computed inline; it now reads this variable, with no change in behaviour.

• New non_savings_taxable_income: earned_taxable_income plus property income after allowances. With no Marriage Allowance it equals non_savings_income_after_allowances (a property test checks this). earned_taxable_income also subtracts Marriage Allowance, which the law gives as a tax reduction (ITA 2007 s. 55B). That is a separate open defect (Marriage Allowance is booked on the recipient as the transferor's unused allowance, not the fixed s.55B transfer #1947, PR Model Marriage Allowance as the statutory s.55B transfer and Step 6 reduction #1963); YAML Case 7 pins the current behaviour.

• The six variables above stack savings and dividends on non_savings_taxable_income.

• taxed_income, which sets the marginal rate of the annual allowance charge (personal_pension_contributions_tax), adds non_savings_taxable_income too. On main, £10,000 of excess contributions with £60,000 of property income was charged at 20% instead of 40%.

• New non_savings_income_after_allowances: adjusted net income less savings, dividends and allowances, floored at nil. Both taxable_property_income_after_allowances and property_income_tax read it, so there is one copy of the expression.

• From 2027-28 the order is statutory:

  • s. 25(3A) sets allowances against income other than property, savings and dividends first;
  • s. 16A places property income immediately below savings.

  Both are cited.

• Unchanged: earned_taxable_income and the earned-income bands, which property income must stay out of, because property_income_tax taxes it.

Worked example, 2026-27: £40,000 of property income and £20,000 of dividends. Non-savings taxable income is £27,430. The £500 dividend allowance and £9,770 fill the basic-rate band at 10.75% (£1,050.28), and £9,730 is at 35.75% (£3,478.48): £4,528.75. main charges all £19,500 at 10.75%: £2,096.25. The same income as pay gives £4,528.75 on main too.

Impact (real run)

Measured on enhanced_frs_2023_24.h5 (local policyengine-uk-data storage, 6 October build) for 2026. This branch at 2de2210 (whose model code is the head's, after a clean merge of main) and main's code (c6d15c8) each ran a full simulation of the same data (impact_compare.py).

	main	This PR	Change
Income Tax	£349,674.5m	£349,705.3m	+£30.7m
of which savings income tax	£356.7m	£382.6m	+£25.9m
of which dividend income tax	£1,613.7m	£1,618.5m	+£4.8m
Household net income	£1,645,439.5m	£1,645,408.7m	−£30.7m

The £30.73m Income Tax change is £25.90m on savings plus £4.81m on dividends, plus about £0.02m from the annual allowance charge (taxed_income).

• Who changes: 1,856 records (101k people weighted), all with property income.
• Increases: 99k people pay more, by up to £9,873.52 a person.
• Float noise, not decreases: 397 records (2k people weighted) show "decreases" of £0.0078 to £0.125. Each is one or two float32 steps at a tax level of £93k to £1.28m.
• Not included: the starting rate for savings still tapers from the personal allowance plus £5,000 rather than from the start of taxable non-savings income (Starting rate for savings: band is not tapered until other income exceeds £30,140; should taper from the Personal Allowance (£12,570) #1799). Fixing that would raise tax further for landlords with £0-£17,570 of taxable non-savings income.

Tests

• New YAML (gov/hmrc/income_tax/property_income_band_base.yaml), seven cases:

  • dividends above £40,000 of property income: £4,528.75;
  • savings above it: £5,846;
  • pay and property income together;
  • a Scottish taxpayer, whose dividends use UK bands;
  • 2027-28, with property income at the 22% property rate and dividends still above it;
  • the annual allowance charge at 40% with property income (Case 6);
  • Marriage Allowance pinned (Case 7).

  Cases 1-6 fail on main.

• Changed YAML (integration/income.yaml, "Property income uses post-allowance rate bands"): it expected nil savings tax for two people with property income and savings interest, which encoded this bug. With savings above the property income:

  • person 1, a higher-rate taxpayer with a £500 allowance, pays (£2,900 − £500) × 40% = £960;
  • person 2, a basic-rate taxpayer with £1,000, pays (£1,525 − £1,000) × 20% = £105.

• Property tests (test_property_income_band_base_properties.py, Hypothesis):

  • replacing property income with pay equal to its taxable amount leaves savings and dividend tax unchanged, in England and Scotland;
  • with no Marriage Allowance, non_savings_taxable_income = adjusted net income less savings, dividends and allowances, floored at nil;
  • more property income never lowers savings or dividend tax (a sanity check: main passes it too);
  • with £40,000 of property income, some of £20,000 of savings pays 40% (main charged £2,900, below the 20%-only £3,900).

  The first and last fail on main. The first moves the taxable amount, because the £1,000 property allowance relieves small property income in full; Hypothesis found that case (£312).

• Run locally at 2de2210:

  • the full policy YAML suite, 3,015 passed;
  • the property tests, 4 passed.

• Review: a GPT-6.1 Sol review approved 1a80b21. Its should-fix and nit findings are addressed in 2de2210:

  • taxed_income;
  • the Marriage Allowance claim and pin;
  • the Starting rate for savings: band is not tapered until other income exceeds £30,140; should taper from the Personal Allowance (£12,570) #1799 comment;
  • the 2027-28 citations;
  • one base variable;
  • a strict test;
  • the decreases explained.

Not changed

The savings starting rate still tapers from the personal allowance plus £5,000 rather than from the start of taxable non-savings income (#1799). This PR changes only the income that taper reads.

🤖 Generated with Claude Code

[MaxGhenis]

CI run cancelled to free runners (UK hub, 2026-10-11 23:30Z). The PolicyEngine org is at GitHub's limit of 60 concurrent jobs, with about 360 queued. Fleet ops asked the hub to let the Budget-critical #2305 take runners first. This PR's queued run had not started, so I cancelled it. Nothing is wrong with the PR.

Please don't re-run it or push just to retrigger CI. The hub re-runs this run itself once #2305 has runners. The record is ~/reviews/uk-hub/ci-yield-2026-10-11.json; see notice 17 in ~/reviews/uk-hub/notices.md.

[MaxGhenis]

Hand-off to the UK hub (session local_66e4b1fd, PolicyBench UK round 2).

• Head: c1067e325e10b8a1621cf7dddf6c9997ffa6dad0 (includes main at ed7b42f95).
• Review: GPT-6.1 Sol, two rounds, then a delta review of the main merge; APPROVE at this head: ~/reviews/policybench-uk-round2/reviews/pe-uk-merge-delta-2352.md (round 2: pe-uk-2352-review-r2.md).
• CI: not yet run at this head. The run was queued, and the hub cancelled it to free runners (notice 17, ~/reviews/uk-hub/ci-yield-2026-10-11.json); the hub re-runs it. No CI run has completed on this PR at any head: each was superseded by a push or cancelled while queued. Lint and Test are therefore the open gate.
• Local tests (~/reviews/policybench-uk-round2/fixes/evidence/), at the approved head before the main merge, which changed none of this PR's files (pr2352_approved_net.diff and pr2352_merged_net.diff are identical): 3,015 YAML cases and 4 property tests (pr2352_r2_full_yaml.log, pr2352_r2_py.log).
• PolicyBench check at this head: all 2,800 round 2 references recomputed on this head (audit/attribution/refs_prop_final.csv); 2 move, the same cells and values as at the reviewed head. With all five fix heads merged onto main ed7b42f95, 96 move, also unchanged.
• Impact (enhanced FRS, 2026, in the PR body): Income Tax +£30.73m, 101k people, all with property income.
• Depends on: nothing. It and Set the Personal Savings Allowance on rest-of-UK bands for Scottish taxpayers #2348 both touch savings tax, in different files; they merge cleanly in either order.
• Decision: none. A bug fix to ITA 2007 s. 16.
• Follow-ups: none owed. Axiom parity: encoded correctly already (uk/statutes/ukpga/2007/3/16.yaml).