[PavelMakarchuk]

Summary

Adds a UK blog post that explains the Budget 2026 fuel duty measure step by step at household level: the 5p cut kept to 1 July 2027 and the planned RPI rise deferred to the same date. The main Autumn Budget dashboard does not show the measure by date or by litre.

Base: this PR is stacked on #1457 (pavel/pension-tax-allowance-post), because the post uses the line chart type that #1457 adds. The branch merges #1457's latest head (5b6eb05). Retarget it to main after #1457 merges.

Post

• Rules from the Budget 2026 document, HM Treasury's policy costings and the OBR EFO October 2026, paragraph 3.19, in our own words.
• Pre-Budget path: 52.95p to 31 December 2026, 55.95p from 1 January 2027 (SI 2026/164 as amended by SI 2026/555), 57.95p from 1 March 2027 (HODA 1979 s.6), and 60.59p from 1 April 2027. Budget path: 52.95p to 30 June 2027, then 60.59p from 1 July 2027. The paths are the same from then on.
• Assumption, stated in the post: no source gives the pre-Budget April 2027 rate. We assume the RPI rise that the Budget moves to July was due in April at the same size (60.59p), as the dashboard does. With policyengine-uk's own 59.75p, the 2027-28 saving for 1,000 litres is £19.17 instead of £21.54; the post gives both.
• Per litre, with 20% VAT on the duty: 3.6p less in January and February 2027, 6.0p in March, 9.2p from April to June.
• Worked examples (no names) at 500, 1,000 and 2,000 litres a year. At 1,000 litres: £33.71 in total, £12.17 in 2026-27 and £21.54 in 2027-28. The post notes that PolicyEngine's tax years (from 6 April) differ from the Treasury's fiscal years (from 1 April).
• Other programmes: the post explains why none interact: means tests use income, not spending; September 2026 and September 2027 duty rates are the same with and without the Budget, so benefit uprating does not change; marginal rates do not change. Business fuel use by the self-employed is noted as out of scope.
• Net income and other benefits: full household_net_income for a single earner, a couple with two children on Universal Credit, and a pensioner on Housing Benefit, Council Tax Reduction and Pension Credit. Net income rises by exactly the duty and VAT saving; no benefit, income tax or NI amount changes. Effective marginal rates are the same before and after, so there is no MTR chart.
• Two chart blocks: a bar chart of the rate by period, before and after the Budget, and a line chart of the saving by litres (a point every 100 litres).
• Cites Table 6.1 row 3 (−£575m in 2026-27, −£780m in 2027-28, then 0).
• Population results: from the dashboard's preview run of fuel_duty_rise_deferral (Vahid, commit 0d9dccb2, generated 28 Oct 14:30 UTC, ab2026-local-preview manifest; the four CSVs match the receipt's SHA-256 hashes), the same source Autumn Budget 2026 property income tax explainer #1460 uses. −£301m and −£521m (static, household fuel, incl. VAT) beside Table 6.1, with the scope, timing, behaviour and VAT differences from María's and Vahid's Create Donate page #290 notes; average gain £10.33 and £17.77 per household; 61% and 73% of people affected; two bar charts by income decile (£ and % of income); poverty unchanged because HBAI income does not deduct taxes on spending. Max asked for this section on Add blog post: the Pension Tax Allowance in the Autumn Budget 2026 #1457. Replace these figures when the dashboard publishes its final run.
• Cover: a PolicyEngine title card of the two rate paths (no photo, no credit needed).

Method

pe.uk.calculate_household (policyengine 6.11.1, policyengine-uk 2.151.0) with two dated parameter reforms on gov.hmrc.fuel_duty.petrol_and_diesel, one per path. The engine day-weights the rates over each tax year (fiscal_year_blend). VAT is 20% of the duty change, as in pump_vat_on_fuel_duty_deferral in PolicyEngine/uk-autumn-budget-dashboard-2026#290; for the net income table a structural reform lowers vat by that amount, so it flows through vat_change into household_net_income, as the dashboard's apply_vat_base_change does. Rates and dates match fuel_duty_rise_deferral there (FUEL_DUTY_FROZEN_UNTIL = "2027-07-01"). The chart values agree with PolicyEngine UK to within £0.01 at every 100 litres from 0 to 2,500, in both tax years.

The net income table and the marginal rate check run on the dashboard's OBR October 2026 forecast (forecast/__init__.py and obr_october_2026.json at 069203c4, applied as a scenario before data load, as with_forecast does). The 2027-28 values are the same on the March 2026 forecast; from 2028-29 the forecasts differ, but the post's figures stop at 2027-28.

Consistency with other work: the rates and dates match PolicyEngine/policyengine-uk#2715 (Max, draft: 52.95p to 30 June 2027, then the RPI step on 1 July 2027 that used to fall on 1 April) and the dashboard's 60.59p (Vahid, d442ef77). The dashboard's own population cost (−£301m and −£521m against the Treasury's −£575m and −£780m) is a gap under investigation on #290; the post cites only the Treasury.

An independent audit recomputed every number and reran the model; it found no numeric errors. Its wording points (the April 2027 assumption, tax year vs fiscal year, HODA citation, net income definition, stated inputs, other fuels, rural relief pass-through) are fixed.

Checks

• Website tests for components, content and lib pass locally (349 tests). tsc --noEmit is clean. No TypeScript files change. Prettier is clean on the article and posts.json.
• I checked the post, the cover and both charts on a local dev server.

To review

• Post date: 2026-10-28 19:00.
• The April 2027 rate assumption.
• Litres are fixed (no demand response) and pass-through to pump prices is full; the post states both.
• Red diesel, the HGV road tax holiday and mileage rates are out of scope.

🤖 Generated with Claude Code

[vercel]

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

Project	Deployment	Actions	Updated
policyengine-calculator-next	Ready	Preview	Oct 28, 2026 5:06pm UTC
policyengine-website	Ready	Preview	Oct 28, 2026 5:06pm UTC

[MaxGhenis]

@PavelMakarchuk, I've checked e3a4bee; the post is the same at 6df8990. Every point from my 16:11 comment is fixed. There's one wording change left, plus a correction of my own.

One change: the pensioner doesn't get Pension Credit

The appendix says the pensioner "claims Pension Credit and Housing Benefit". On the October scenario they get no Pension Credit. Their income for it is £12,944, above the £12,761 guarantee, and the table rightly lists none.

Suggested wording: "...pays £7,800 in rent and £1,600 in council tax, and gets Housing Benefit and Council Tax Reduction; their income is above the Pension Credit guarantee, so they get no Pension Credit."

Correction to my 15:45 comment

I said the Budget's words are "35 million individual motorists – and businesses across the UK", so the 35 million counts motorists only. The Budget uses both forms:

• The opening summary has "35 million individual motorists – and businesses across the UK".
• The fuel duty paragraph has "the 35 million motorists and businesses across the UK".

Your current wording follows the first and is fine. But the Budget doesn't settle whether businesses are inside the 35 million, so "counts motorists only" was too strong.

Checked at e3a4bee

• Net income table. It matches the dashboard's October scenario:
  • single adult £24,333 → £24,355;
  • couple £38,045 → £38,067;
  • pensioner £20,002 → £20,024, with Housing Benefit £7,682, Council Tax Reduction £1,564 and Winter Fuel Payment £200, on £13,036.40 of State Pension.
• Gini (0.01%, 0.02%). These now read as relative changes.
• Uprating line and "on our assumption".
• The new 275-day row (1 July 2027 to 31 March 2028; 2028 is a leap year), and RPI from 1 April 2028.
• The £230 baseline. It matches the Budget's footnote: "the 5p cut being reversed in April 2025, and annual RPI uprating from April 2025 onwards".
• "Other fuels and rebated fuels, where practical". It matches the costings, page 13.
• Deciles. They are ranked by equivalised HBAI net income before housing costs, weighted by people (household_income_decile), with shares from each decile's total net income (calculators.py).
• The OBR sentence. "The OBR scores the Treasury's figure as fuel duty receipts, with no VAT line" matches Table 1.1.
• The Table 6.1 link. Table 6.1 is in the Budget document's HTML, in chapter 7.

With the Pension Credit wording fixed, I have nothing else on the fuel duty content. Whether the population section goes out today is for Max and Vahid. #1459 merges after #1457, which it is stacked on.

🤖 Generated with Claude Code

[MaxGhenis]

@PavelMakarchuk, Max wants the explainer posts merged today so the dashboard's Policies tab can link them. Is it OK to merge this post at 42a6de0?

• Content. Every must-fix from my reviews is in at 42a6de0, including the Pension Credit line. I have nothing else on the fuel duty content.
• What may still change before the merge, with no change to the post's text:
  • a merge of origin/main into this branch, for the posts.json conflict that comes once other posts land;
  • the base moving from pavel/pension-tax-allowance-post to main. This PR stacks on Add blog post: the Pension Tax Allowance in the Autumn Budget 2026 #1457, which stacks on Add a line chart type for blog posts, and valueLabels for bar charts #1462 (line chart), and the post uses the line chart and valueLabels. So those two go in first, and this one follows.
• Figures. I'll check the population figures against Create Donate page #290's final regeneration, Vahid's freeze, before merging. If any figure moves at whole pounds, I'll update it here as a separate commit and tell you exactly what changed.

A "yes" here is enough. I'll squash-merge on it, after #1457, once CI is green.

🤖 Generated with Claude Code

[MaxGhenis]

@PavelMakarchuk, a correction to my 16:29 comment. Your OK is still needed, but it isn't enough on its own. Merging this post also needs Max's yes, for this PR and for whether the costing section goes out before the dashboard approves fuel duty. The coordinator will put both to him after the webinar.

In the meantime, once #1457 is in, I'll merge origin/main into this branch and check the figures against #290. I won't merge until Max has said yes.

🤖 Generated with Claude Code

[MaxGhenis]

@PavelMakarchuk, I pushed one commit to this branch: be1b1cf, on top of your merge of main (da89d29). It changes only the two decile charts' data values in fuel-duty-cut-extension.md, nothing else.

• What the charts held. Values from the 0d9dccb2 preview, which predates the October detailed-table inputs.
• What they hold now. The dashboard's own DistributionalImpactCalculator and WinnersLosersCalculator on fuel_duty_rise_deferral at Create Donate page #290's 9b98370e, with the bundle's certified data and the October forecast. The values are stored to 3 decimal places for pounds and 5 for shares. That way the chart's toFixed rounds once: 19.495 stored as 19.5 would have shown £20.
• Shown values that change:
  • decile 2, 2026-27: £7 → £8;
  • decile 7, 2027-28: £20 → £19;
  • decile 10's 2027-28 share: 0.015% → 0.014%. That now matches the text, which already said 0.014%.
• Unchanged. Every other bar, and every figure in the text. The same run gives −£301m and −£521m, average gains of £10 and £18, Gini −0.01% and −0.02%, and no change in poverty.

When Vahid freezes #290, I'll check that nothing feeding the fuel measure has changed since 9b98370e, and rerun if anything has.

🤖 Generated with Claude Code

[MaxGhenis]

@PavelMakarchuk, Max has said yes to merging this post today, with the costing section as it is, once you OK it.

• Head: 0ddbe9b, your merge of main (which brought in Add blog post: the Autumn Budget 2026 energy bill measures #1461). The post itself is unchanged since be1b1cf.
• CI: every check passes at 0ddbe9b. The diff against main is the post, its cover image and the posts.json entry.

If you're happy, reply here with "OK to merge". I'll then approve and squash-merge at the head you OK. If you'd rather change something first, push it and say so; I'll recheck and merge that head instead.

The final dashboard regeneration hasn't been published yet. When it is, I'll check the post's figures against it, and fix anything that moves in a follow-up PR.

🤖 Generated with Claude Code

[PavelMakarchuk]

OK to merge at 0ddbe9b, given Max's yes.

[MaxGhenis]

Approving at 0ddbe9b on Max's yes (relayed by the coordinator: "Yes, once Pavel OKs") and Pavel's OK (6476100182). Checked: every microsimulation figure in the post against the dashboard's final fuel duty receipt (test link revision 28, artifact e077f8d4, built at #290's freeze head 4de166f0; no displayed figure differs), whole-pound rounding, the household examples on the October scenario, and CI (all checks pass at this head). The post is unchanged since be1b1cf; 0ddbe9b only merges main.

🤖 Generated with Claude Code