UK Fiscal Drag 2026-27 — Frozen Thresholds and Their Real Cost
Original analysis: who crosses into higher-rate, additional-rate and the £100k Personal Allowance taper in 2026-27, thresholds frozen to April 2031.
Frozen vs CPI-indexed counterfactual (2026-27)
Threshold
Frozen 2026-27
CPI-indexed 2026-27
Gap
Personal Allowance
£12,570
£16,150
£3,580
Higher-rate threshold (above PA)
£50,270
£64,570
£14,300
Additional-rate threshold
£125,140
£236,600
£111,460
Personal Allowance taper start
£100,000
£157,730
£57,730
CPI-indexed counterfactual = each threshold's nominal value in the tax year it was last set, × cumulative ONS CPIH (series L522) growth from that month to April 2026. PA/HRT run from April 2021 (+28.4%); ART/taper run from April 2010, when they were introduced (+57.7%). Rounded to nearest £10.
OBR estimates (Nov 2022 EFO — original freeze to April 2028)
Metric
Value
New taxpayers created by the PA and higher-rate threshold freeze to 2027-28
3.2 million
New higher-rate taxpayers created by the same freeze
2.6 million
Yield of the PA and higher-rate threshold freezes in 2027-28 (vs CPI-indexed thresholds)
£26 billion a year
Including the primary and secondary NICs threshold freezes
£35 billion a year
All figures from the OBR November 2022 Economic and Fiscal Outlook (paragraph 38 and footnote 24), which estimated the effect of the freezes as extended to 2027-28 at Autumn Statement 2022. This freeze was assumed to end in April 2028 — it did not; see the update below.
Update: freeze extended to April 2031 (Autumn Budget 2025)
Before the original freeze reached its April 2028 end-date, Autumn Budget 2025 extended it a further three years to April 2031, for income tax PA/HRT/ART and the equivalent NIC thresholds. The figures below are the OBR's estimate of the incremental effect of that extension, not a re-statement of the cumulative 2022 figures above.
Metric
Value
Extra people paying income tax by 2029-30 (vs Mar 2025 forecast)
~780,000
Extra higher-rate taxpayers by 2029-30 (vs Mar 2025 forecast)
~920,000
Extra additional-rate taxpayers by 2029-30 (vs Mar 2025 forecast)
A 1p basic-rate rise is highly visible — politicians own it, the press reports it, and voters react. A threshold freeze is invisible — tens of billions arrive gradually as wages grow, and no individual paycheck shows a "tax rise" line. Most affected taxpayers will not realise they have been pulled into a higher band until their first April pay rise crosses £50,270 and the change appears in their next tax code (1257L → BR for the over-band portion).
The freeze interacts particularly sharply with the £100,000 Personal Allowance taper. Workers with bonus-driven income (consultants, sales, financial services) crossing into the £100,000-£125,140 band hit a 60% effective marginal rate that didn't exist for them at any prior salary level — see the companion 100k tax trap analysis.
Fiscal drag is the silent tax rise that happens when wages grow with inflation but tax thresholds stay frozen. More of each worker's pay enters higher tax bands, so the average effective tax rate climbs even though Parliament hasn't voted to raise rates. The Office for Budget Responsibility tracks it explicitly in every EFO.
A 1p rate rise is visible — politicians own it and voters react. A freeze is invisible — it raises the same revenue over a few years but never appears in a Budget headline. The OBR's November 2022 estimate was that freezing the Personal Allowance and higher-rate threshold to 2027-28 would raise £26 billion a year by 2027-28 relative to CPI-indexed thresholds, or £35 billion including the matching National Insurance threshold freezes. Since the freeze was extended to April 2031 at Autumn Budget 2025, the OBR's November 2025 EFO puts the freeze's annual revenue at over £55 billion by 2030-31 — with no equivalent rate-rise headline at any point.
Who is most affected?
Workers with rising wages who cross thresholds. The biggest cohort is the people pulled into higher-rate (40%) tax, which the OBR put at 2.6 million by 2027-28 in its November 2022 EFO — typically dual-income households, public-sector pay-band climbers, and anyone whose nominal wage grew with inflation. The sharpest individual marginal-rate jumps fall on those crossing the £100,000 taper and the additional-rate (45%) threshold, which was cut from £150,000 to £125,140 in April 2023; the OBR's March 2026 Economic and Fiscal Outlook detailed forecast tables (receipts, Table 3.18) put the number brought into the 45% band by that cut at around 537,000 in 2027-28.
Where do the CPI-indexed counterfactual figures come from?
Apply ONS CPIH (series L522) growth from each threshold's last uprated month through to April 2026. The Personal Allowance and higher-rate threshold were both last uprated for 2021-22: CPIH rose from 110.4 (April 2021) to 141.8 (April 2026), +28.4%. The additional-rate threshold and the £100,000 Personal Allowance taper start have not risen since they were introduced in 2010-11: CPIH rose from 89.9 (April 2010) to 141.8 (April 2026), +57.7%. We round to the nearest £10 in the table.
Did fiscal drag end in April 2028 as originally planned?
No — it was extended before it ever arrived. The freeze was originally set to end in April 2028 (thresholds would then rise with CPI), but Autumn Budget 2025 pushed the freeze a further three years to April 2031, before any unfreezing took effect. The OBR's current (Nov 2025) baseline assumes thresholds rise with CPI from April 2031. Watch for any future Budget statement that pushes the freeze further out again — that is the de-facto tax rise to track.
Is the National Insurance threshold also frozen?
Yes. The Class 1 Primary Threshold and Upper Earnings Limit are aligned with income tax PA / HRT respectively, so they share the same freeze through April 2031 (also extended from April 2028 by Autumn Budget 2025). Class 4 thresholds for the self-employed are similarly frozen at £12,570 / £50,270.
Can I cite this analysis?
Yes. Published under CC-BY 4.0 with attribution. Underlying inputs are all from gov.uk, ONS, and the OBR Nov 2022 EFO.
Re-publishable under CC-BY 4.0. Please link back to this URL so corrections propagate.
