HMRC Overcharged Millions of Pensioners for Decade Due to Income Calculation Error

The UK tax authority HMRC has been systematically overcharging pensioners for more than a decade due to a persistent error in how state pension income was calculated, affecting up to 8.7 million people. The error resulted from HMRC using 52 weeks of the current tax year's state pension rate rather than the correct 51 weeks of the current rate plus one week of the previous year's lower rate, as its own guidance prescribes. This miscalculation compounded annually as the state pension rose under the triple lock mechanism, leaving pensioners paying an average of £5 extra per year, though some higher-rate taxpayers faced overcharges of £3.60 to £4.

The scale of the overcharging is substantial, with estimates suggesting more than £350 million may have been incorrectly taken over the period, and as much as £43.5 million extracted in a single recent year. The problem was first identified by affected pensioners as far back as 2016, but HMRC failed to address it publicly or issue refunds. Former HMRC chief executive Jim Harra, who stepped down in April 2026, was made aware of the issue two years prior yet the department did not launch a formal investigation until a victim contacted the chief executive directly, triggering action only recently.

The error affected both self-assessment taxpayers and those paying through Pay As You Earn (PAYE), as HMRC used income data provided by the Department for Work and Pensions rather than its own records. The impact widened each year due to the state pension's guaranteed annual increases. Despite acknowledging the problem, HMRC has not yet notified affected individuals or provided automatic refunds, though the department stated it is "working at pace" to fix the issue by summer 2026. Critics have highlighted the significant delay between initial identification and public disclosure, as well as HMRC's characterization of the "impact as small" for individuals despite the cumulative effect of the overcharges.