Problem
ITA 2007 s.25(2): "At Steps 2 and 3, deduct the reliefs and allowances in the way which will result in the greatest reduction in the taxpayer's liability to income tax." From 2027-28, s.25(3A) (FA 2026 s.6) adds that they "must be deducted from components other than property income, savings income or dividend income before being deducted from those income types", which still leaves the order between property, savings and dividends to s.25(2).
The model deducts allowances in a fixed order instead: non-savings income, then savings (received_allowances_savings_income), then dividends (received_allowances_dividend_income). Savings that fall in the 0% starting rate for savings, or within the personal savings allowance, are taxed at nothing, so an allowance set against them is wasted, while the same allowance set against dividends would save tax.
Example on main (84f5ad4), 2026-27, a person aged 40 with no other income:
| Income |
Personal Allowance taken against |
Income Tax |
| Dividends £20,000 |
dividends £12,570 |
£744.97 |
| Savings £4,000 + dividends £20,000 |
savings £4,000, dividends £8,570 |
£1,174.97 |
The £4,000 of savings is within the £5,000 starting rate band, so set against dividends instead the allowance gives £744.97 on both lines. The model charges £430 more.
#2085 adds trade loss relief against general income (a Step 2 relief) and deducts the part beyond non-savings income in the same fixed order, so the same overcharge applies to it. The review of #2085 found one: savings £4,000, dividends £200,000 and a £1,000 loss give £107.50 more tax than setting the relief against dividends.
Fix sketch
Allocate the pool left after non-savings income (the Personal Allowance and other allowances, plus any Step 2 relief beyond non-savings income) to savings or dividends in whichever order gives less tax. Tax is piecewise linear in the split, but the PSA steps make it non-convex, so compare both pure orders at least and take the lower. This changes Income Tax for everyone whose allowances reach savings and who also has dividends, so it needs its own enhanced FRS impact run.
🤖 Generated with Claude Code
Problem
ITA 2007 s.25(2): "At Steps 2 and 3, deduct the reliefs and allowances in the way which will result in the greatest reduction in the taxpayer's liability to income tax." From 2027-28, s.25(3A) (FA 2026 s.6) adds that they "must be deducted from components other than property income, savings income or dividend income before being deducted from those income types", which still leaves the order between property, savings and dividends to s.25(2).
The model deducts allowances in a fixed order instead: non-savings income, then savings (
received_allowances_savings_income), then dividends (received_allowances_dividend_income). Savings that fall in the 0% starting rate for savings, or within the personal savings allowance, are taxed at nothing, so an allowance set against them is wasted, while the same allowance set against dividends would save tax.Example on main (84f5ad4), 2026-27, a person aged 40 with no other income:
The £4,000 of savings is within the £5,000 starting rate band, so set against dividends instead the allowance gives £744.97 on both lines. The model charges £430 more.
#2085 adds trade loss relief against general income (a Step 2 relief) and deducts the part beyond non-savings income in the same fixed order, so the same overcharge applies to it. The review of #2085 found one: savings £4,000, dividends £200,000 and a £1,000 loss give £107.50 more tax than setting the relief against dividends.
Fix sketch
Allocate the pool left after non-savings income (the Personal Allowance and other allowances, plus any Step 2 relief beyond non-savings income) to savings or dividends in whichever order gives less tax. Tax is piecewise linear in the split, but the PSA steps make it non-convex, so compare both pure orders at least and take the lower. This changes Income Tax for everyone whose allowances reach savings and who also has dividends, so it needs its own enhanced FRS impact run.
🤖 Generated with Claude Code