Budget 2026 posts: cut process wording, recast work-in-progress notes as limitations

app/src/data/posts/articles/carers-allowance-taper.md

@@ -745,7 +745,7 @@ The table and chart leave out the £1 step at the limit itself. Without the tape
 
 ## Population estimates
 
-We also ran the measure on PolicyEngine UK's microsimulation model, which uses survey data on UK households; we projected the data to 2030-31 and 2031-32 on the OBR's October 2026 forecast. The estimate is static: carers do not change how much they work, and take-up of Carer's Allowance follows the assumptions in PolicyEngine's data. This is the run behind the figures for this measure on PolicyEngine's Autumn Budget dashboard, split here into its parts. We will update the 2030-31 population estimates after correcting how the taper interacts with overlapping benefits, such as the State Pension. The model applies half a year of the taper in 2030-31, because the taper starts at some point in that year. Like the dashboard, the run keeps PolicyEngine UK's current-law limit of £204. Unlike the worked examples, it does not yet count the PAYE on the allowance against Universal Credit earnings ([regulation 55(5)(b)](https://www.legislation.gov.uk/uksi/2013/376/regulation/55)); with that adjustment, our calculation gives a net cost of about £28 million in 2030-31 and £55 million in 2031-32.
+We also ran the measure on PolicyEngine UK's microsimulation model, which uses survey data on UK households; we projected the data to 2030-31 and 2031-32 on the OBR's October 2026 forecast. The estimate is static: carers do not change how much they work, and take-up of Carer's Allowance follows the assumptions in PolicyEngine's data. This is the run behind the figures for this measure on PolicyEngine's Autumn Budget dashboard, split here into its parts. The 2030-31 population estimates don't correctly reflect how the taper interacts with overlapping benefits, such as the State Pension. The model applies half a year of the taper in 2030-31, because the taper starts at some point in that year. Like the dashboard, the run keeps PolicyEngine UK's current-law limit of £204. Unlike the worked examples, it does not yet count the PAYE on the allowance against Universal Credit earnings ([regulation 55(5)(b)](https://www.legislation.gov.uk/uksi/2013/376/regulation/55)); with that adjustment, our calculation gives a net cost of about £28 million in 2030-31 and £55 million in 2031-32.
 
 | £ million, cost to the Exchequer |  2030-31 |  2031-32 |
 | -------------------------------- | -------: | -------: |
@@ -786,6 +786,6 @@ The estimate rests on fewer than 50 survey households, so it is imprecise, and w
 
 ### Methodology
 
-We used the `policyengine` Python package (version 6.11.1) with PolicyEngine UK (version 2.151.0). PolicyEngine UK does not yet include the taper, so we added it as a structural reform that changes Carer's Allowance entitlement (the `carers_allowance_pre_overlap` variable) above the earnings limit. The rules and values match the Carer's Allowance taper in the draft update to PolicyEngine's [Autumn Budget 2026 dashboard](https://policyengine.org/uk/autumn-budget-2026); the worked examples differ in two ways: they set the current-law limit to the government's rule, and they count the PAYE on the allowance against Universal Credit earnings (see the assumptions). The population estimates are the dashboard's own run.
+We used the `policyengine` Python package (version 6.11.1) with PolicyEngine UK (version 2.151.0). PolicyEngine UK 2.151.0 doesn't include the taper; in these calculations the taper changes Carer's Allowance entitlement (the `carers_allowance_pre_overlap` variable) above the earnings limit. The rules and values match the Carer's Allowance taper in PolicyEngine's [Autumn Budget 2026 dashboard](https://policyengine.org/uk/autumn-budget-2026); the worked examples differ in two ways: they set the current-law limit to the government's rule, and they count the PAYE on the allowance against Universal Credit earnings (see the assumptions). The population estimates are the dashboard's own run.
 
 Both sides of every calculation use the OBR's October 2026 forecast, as the dashboard does. In the worked examples, both sides also set PolicyEngine UK's current-law earnings limit to the government's rule and use our Universal Credit adjustment: for a person with employment income, we work out the income tax that Universal Credit deducts from earnings by treating Carer's Allowance as part of the earnings, so it includes the tax on the allowance. We calculated each household for gross pay from £0 (marginal rates) or £150 (other charts and tables) to £400 a week, in £1 steps for the other charts and tables. The charts of the allowance and of net income plot whole-pound points where the schedules change slope and a point every £20. The marginal tax rates are PolicyEngine UK's `marginal_tax_rate` variable, with its pay rise set to £52 a year (£1 a week) on both sides; the chart plots a point every £100 of annual pay and the points where the rates change, and leaves out the points whose £1 rise crosses one of those points, which would average two rates. The variable agrees with a direct difference of household net income to within 0.01 percentage points. PolicyEngine UK reduces all employees' pay by 0.16% in 2031-32, to reflect employers passing on the cost of the salary sacrifice cap, so we set pay 0.16% higher to give whole-pound gross pay.

app/src/data/posts/articles/fuel-duty-cut-extension.md

@@ -544,8 +544,8 @@ The Budget estimates that the government's fuel duty decisions in this Parliamen
 
 ### Methodology
 
-We used the `policyengine` Python package (version 6.11.1) with PolicyEngine UK (version 2.151.0). We ran `calculate_household` for each household with two parameter reforms to the fuel duty rate on petrol and diesel: one with the pre-Budget rates and one with the Budget rates, each keyed to the dates above. PolicyEngine UK keys its April 2027 rate at 59.75p, uprated by the March 2026 RPI forecast, so both scenarios replace it with 60.59p. PolicyEngine UK weights each rate by the days it applies in the tax year. The net income table and the marginal rates run on the OBR October 2026 forecast: we build each household as a PolicyEngine UK simulation with the dashboard's forecast scenario, applied before data load, and add the measure to the same scenario. The duty and VAT savings do not depend on the forecast.
+We used the `policyengine` Python package (version 6.11.1) with PolicyEngine UK (version 2.151.0). We ran `calculate_household` for each household with two parameter reforms to the fuel duty rate on petrol and diesel: one with the pre-Budget rates and one with the Budget rates, each keyed to the dates above. PolicyEngine UK keys its April 2027 rate at 59.75p, uprated by the March 2026 RPI forecast, so both scenarios replace it with 60.59p. PolicyEngine UK weights each rate by the days it applies in the tax year. With and without the measure, the net income table and the marginal rates use the dashboard's OBR October 2026 forecast. The duty and VAT savings do not depend on the forecast.
 
-PolicyEngine UK holds spending on fuel fixed and does not pass a duty change into VAT, so we add 20% of the duty change for the VAT on the pump price. In those simulations, the VAT step lowers the household's `vat` by that amount, so the change flows through `vat_change` into `household_net_income`, as the dashboard's VAT modifier does. We use the rates, dates and VAT treatment of the fuel duty measure in PolicyEngine's [Autumn Budget 2026 dashboard](https://policyengine.org/uk/autumn-budget-2026). The saving is linear in litres, so the chart plots its value every 100 litres; we compute each point from the rules and check it against PolicyEngine UK.
+PolicyEngine UK holds spending on fuel fixed and does not pass a duty change into VAT, so we add 20% of the duty change for the VAT on the pump price. In the net income table and the marginal rates, the VAT step lowers the household's `vat` by that amount, so the change flows through `vat_change` into `household_net_income`, as the dashboard's VAT modifier does. We use the rates, dates and VAT treatment of the fuel duty measure in PolicyEngine's [Autumn Budget 2026 dashboard](https://policyengine.org/uk/autumn-budget-2026). The saving is linear in litres, so the chart plots its value every 100 litres.
 
-The population results come from a run of the dashboard's measure on PolicyEngine's certified UK household data. We reproduced them in a separate run with the same package, data and forecast: the cost is £301 million in 2026-27 and £521 million in 2027-28, of which VAT is 17%, and the average gain is £10 and £18 a household.
+The population results come from a run of the dashboard's measure on PolicyEngine's certified UK household data: the cost is £301 million in 2026-27 and £521 million in 2027-28, of which VAT is 17%, and the average gain is £10 and £18 a household.

app/src/data/posts/articles/pension-tax-allowance.md

@@ -1003,4 +1003,4 @@ The Treasury's costing (Budget 2026, Table 6.1, row 52) follows a National Accou
 
 ### Methodology
 
-We used the `policyengine` Python package (version 6.11.1) with PolicyEngine UK (version 2.151.0). Every calculation, with and without the PTA, uses the OBR's October 2026 Economic and Fiscal Outlook forecast, as PolicyEngine's Autumn Budget 2026 dashboard does; PolicyEngine UK 2.151.0 ships the March 2026 forecast. PolicyEngine UK does not yet include the PTA, so we added it as a structural reform that reduces the `income_tax` variable and the income tax that Universal Credit deducts from earnings. The rules and values match the PTA modifier in the draft update to PolicyEngine's Autumn Budget 2026 dashboard, including its Universal Credit adjustment; the one difference is the Pension Credit guarantee (see the assumptions). Income tax, the PTA and net income are piecewise linear in total income, so the charts plot their breakpoints and a point every £100. These values agree with PolicyEngine UK to the penny at every £2 of total income, and the effective marginal tax rates agree to within 0.04 percentage points away from the breakpoints.
+We used the `policyengine` Python package (version 6.11.1) with PolicyEngine UK (version 2.151.0). Every calculation, with and without the PTA, uses the OBR's October 2026 Economic and Fiscal Outlook forecast, as PolicyEngine's Autumn Budget 2026 dashboard does; PolicyEngine UK 2.151.0 ships the March 2026 forecast. PolicyEngine UK 2.151.0 doesn't include the PTA; in these calculations the PTA reduces the `income_tax` variable and the income tax that Universal Credit deducts from earnings. The rules and values match the PTA modifier in PolicyEngine's Autumn Budget 2026 dashboard, including its Universal Credit adjustment; the one difference is the Pension Credit guarantee (see the assumptions). Income tax, the PTA and net income are piecewise linear in total income, so the charts plot their breakpoints and a point every £100. These values agree with PolicyEngine UK to the penny at every £2 of total income, and the effective marginal tax rates agree to within 0.04 percentage points away from the breakpoints.

app/src/data/posts/articles/uk-autumn-budget-2026-briefing.md

@@ -29,7 +29,7 @@ Sources: [Budget 2026](https://www.gov.uk/government/publications/budget-2026-do
 
 The Budget makes other changes that reach households, which our dashboard does not model:
 
-- **Savings income tax.** From 6 April 2027 the rates on savings income also rise 2 points, to 24%, 44% and 49% across the UK. We give the official costing [below](#savings-income); our own estimate follows with a data update.
+- **Savings income tax.** From 6 April 2027 the rates on savings income also rise 2 points, to 24%, 44% and 49% across the UK. We give the official costing [below](#savings-income) but no national estimate of our own.
 - **The Triple Lock.** From April 2030 the State Pension rises each year by the highest of 2.5%, inflation, or the amount that keeps its value relative to earnings at its April 2029 level. The government estimates this reduces State Pension spending by £15 billion a year by the late 2030s. We analyse it [separately](https://www.policyengine.org/uk/research/uk-triple-lock-budget-2026).
 - **The National Living Wage** rises 3.9% to £13.21 an hour on 1 April 2027, as the Low Pay Commission recommended.
 - **The Bank Corporation Tax Surcharge** rises from 3% to 6% from 1 January 2027 to 31 March 2028, to help fund the energy support.
@@ -55,7 +55,7 @@ Source: the OBR's [detailed forecast table 1.1](https://obr.uk/efo/economic-and-
 
 ## The combined effect
 
-We run the seven measures together as one package, in a single simulation for each year. We leave out the savings income rate rise until our estimate is ready.
+We run the seven measures together as one package, in a single simulation for each year. The package leaves out the savings income rate rise.
 
 ### Cost to the public finances
 
@@ -192,7 +192,7 @@ Our estimates differ from the official costings for four main reasons: timing (w
 
 Official: the OBR's [detailed forecast table 1.1](https://obr.uk/efo/economic-and-fiscal-outlook-october-2026/), measures 1, 2, 3, 52, 61, 15 and 7, the line or lines that match what we model, under the table's own names. Table 6.1 total: HM Treasury, [Table 6.1](https://www.gov.uk/government/publications/supporting-documents-for-budget-2026), rows 1, 2, 3, 52, 61, 15 and 7. Its totals also include the OBR table's other lines for each measure that aren't zero: Other AME (current) for zero VAT; PSCE in RDEL, Scottish BGA (current) and Scottish self-financed expenditure for the Pension Tax Allowance and the trading allowance; Onshore CT, Scottish BGA (current) and Scottish self-financed expenditure for property income tax; and Scottish BGA (current) for the Carer's Allowance taper.
 
-Why each differs. The reasons come from the [dashboard's notes](https://www.policyengine.org/uk/budgets/autumn-2026?tab=dashboard&sub=finances) on its comparison with the official costings, from the OBR's October 2026 Economic and fiscal outlook and from the Budget documents. The notes explain the comparisons and also draw on further runs of our model, so some of their figures aren't in Table 6 or in our published results:
+Why each differs. The reasons come from the [dashboard's notes](https://www.policyengine.org/uk/budgets/autumn-2026?tab=dashboard&sub=finances) on its comparison with the official costings, from the OBR's October 2026 Economic and fiscal outlook and from the Budget documents. The notes explain the comparisons, and some of their figures appear in neither Table 6 nor our published results:
 
 - **Household Energy Payment (scope).** Ours falls 25% below the official £2,627 million because the official figure also covers two things we don't model: the devolved governments' share of the funding (about £425 million, our estimate using the Barnett formula) and £100 million for councils in England to target support at the households that rising energy bills hit hardest. Without them, about £2,100 million remains for the payment in England, and ours is 94% of that. At £100 each, the up to 20 million households the Budget names take up to £2,000 million of it; neither the OBR nor the Budget itemises the rest.
 - **Zero VAT on electricity (coverage).** Ours exceeds the OBR's VAT line by £106 million (14%) in 2026-27 and by £57 million (19%) in 2027-28. The official costing also covers charities, residential care and some businesses, and counts households switching from gas to electricity; both make the official figure larger, so neither explains the gap. The two start from different spending: we calculate ours from the Department for Energy Security and Net Zero's domestic electricity volumes and the average prices households pay, then project that spending using CPI; the official costing starts from ONS consumer spending, which is 10% to 12% lower.
@@ -215,23 +215,21 @@ From 6 April 2027 the income tax rates on savings income rise 2 points, on top o
 
 Official: the OBR's [detailed forecast table 1.1](https://obr.uk/efo/economic-and-fiscal-outlook-october-2026/), measure 62, the line that matches what we model. Table 6.1's total also includes the OBR table's Onshore CT line for the measure.
 
-The official costing includes some savers moving their savings into ISAs and other tax-advantaged products. We hold our estimate back from this release while we update the savings income in our data; it follows with that update.
+The official costing includes some savers moving their savings into ISAs and other tax-advantaged products. Our national estimates leave out the rate rise.
 
-## What we're improving
+## Limitations
 
-We plan to update these results later this week. The update will:
+These results:
 
-- add poverty and people-affected figures;
-- add national figures for the savings income tax rise, once we have improved our savings income data (the calculator already includes it);
-- apply the trading allowance to miscellaneous income as well as trading income, which will raise our estimate of its cost;
-- correct how the Carer's Allowance taper interacts with overlapping benefits, such as the State Pension, in 2030-31;
-- extend the fuel duty cut to road fuel gases (LPG and natural gas), as the government's costing does;
-- give the Pension Tax Allowance to everyone who reaches State Pension age during the tax year.
+- leave out poverty figures and the dashboard's separate people-affected figures;
+- leave out national figures for the savings income tax rise (the calculator includes it);
+- apply the trading allowance to trading income but not to miscellaneous income (including it would raise our estimate of the allowance's cost);
+- don't correctly reflect how the Carer's Allowance taper interacts with overlapping benefits, such as the State Pension, in 2030-31;
+- apply the fuel duty cut to petrol and diesel but not to road fuel gases (LPG and natural gas), which the government's costing covers;
+- don't give the Pension Tax Allowance to everyone who reaches State Pension age during the tax year;
+- project household electricity spending with general inflation rather than retail energy prices.
 
-We are also working on:
-
-- projecting household electricity spending with retail energy prices rather than general inflation;
-- adding the Budget's measures to PolicyEngine UK itself, so anyone using our open-source model can run them.
+PolicyEngine UK 2.151.0, our open-source model, doesn't include the Budget's measures; the dashboard runs them.
 
 ## Methodology
 
@@ -243,6 +241,6 @@ Our estimates come from PolicyEngine's [Autumn Budget 2026 dashboard](https://ww
 - **The package.** The combined results come from one simulation of the seven measures together, which models each exactly as in its own run.
 - **Small samples.** The dashboard withholds any result that rests on fewer than three survey households. For the package in 2030-31 and 2031-32, it withholds the results for two income deciles and the average change across all households, the same cells it withholds for the Carer's Allowance taper on its own.
 
-Every PolicyEngine figure in this briefing comes from the dashboard's published results (manifest revision 17), and we read each from files that match the checksums in their receipt; the exceptions are the further figures we quote from the dashboard's notes on the comparison with the official costings, which come from further runs of our model. Every official figure comes from the OBR's October 2026 Economic and fiscal outlook or the Budget documents we link above.
+Every PolicyEngine figure in this briefing comes from the dashboard's published results (manifest revision 17), except the further figures we quote from the dashboard's notes on the comparison with the official costings, which the published results don't include. Every official figure comes from the OBR's October 2026 Economic and fiscal outlook or the Budget documents we link above.
 
 _PolicyEngine's UK Budget analysis is funded by the [Nuffield Foundation](https://www.nuffieldfoundation.org/). The views expressed are ours and not necessarily the Foundation's._