How are frozen tax thresholds reshaping who pays personal taxes?

The thresholds at which income tax and higher rates of income tax begin to apply have not been increased in cash terms since April 2021, and for National Insurance contributions (NICs) the same has been true since at least April 2023.1 While, by default, thresholds are uprated annually by CPI inflation, under current policy they are due to remain frozen until April 2028. This represents a substantial real-terms reduction in tax thresholds, increasing personal tax liabilities for a given level of real earnings. Put another way, a worker whose earnings only just keep up with inflation is paying a larger fraction of those earnings in tax because of the freezes. Taken as a whole, these real-terms threshold reductions represent one of the most significant changes to the personal tax system in recent decades – fundamentally reshaping who pays tax and at what rates. In this comment, we summarise the scale of the freezes’ impact so far and highlight the implications of extending the freezes should the Chancellor choose to do so at the forthcoming Budget.

In fiscal terms, the impact of threshold freezes has already been substantial. Freezes to the thresholds at which the basic (20%) and higher (40%) rates of income tax begin to apply are alone expected to raise £39 billion a year in 2029–30 (roughly similar to the amount of revenue that would be raised by increasing all rates of income tax by 3½ percentage points). Similarly, the thresholds at which employee and self-employed NICs begin to apply have been frozen in cash terms since July 2022 and April 2023 respectively (having been significantly raised shortly before that). Meanwhile, the threshold at which the top rate of NICs begins to apply has been frozen since April 2021; together these NICs threshold freezes are set to raise £3 billion a year in 2029–30. The threshold at which employer NICs begin to apply has also been frozen in cash terms since April 2022, but was then subject to a large one-off reduction (i.e. a tax rise) in April 2025. Despite these already substantial changes, more revenue could be raised from threshold changes. If the Chancellor chose to keep all currently frozen NICs and income tax thresholds frozen for two further years (until April 2030), for example, this would raise a further £8.3 billion a year by 2029–30.2  

The threshold freezes mean that, all else equal, anyone paying income tax or NICs will see their taxes increase. An additional impact of frozen thresholds is that a substantial minority of taxpayers will face higher marginal tax rates (for example, by going from being a basic-rate to a higher-rate income tax payer). This matters because the marginal tax rate is an important component of the incentives individuals face when making decisions over whether to increase or reduce how much they work.

Figure 1 shows how the shares of adults that pay income tax and higher rates of tax have changed over time and as a result of the freezes. A two-year extension to the freezes would leave the total number of income tax payers at 42.1 million (73% of people aged 16 or over) – 960,000 higher than under current plans and 5.1 million higher than if there had been no freezes at all. The impact on the number of people subject to income tax at higher rates (i.e. marginal rates of 40% or above), meanwhile, is even more striking. An extension would bring the total number of higher-rate taxpayers to 10.1 million (18% of people aged 16 or over), an increase of 790,000 compared with current policy and a 4.8 million increase relative to if there had been no freezes at all.3

One implication of the substantial increases in both the number of income tax payers and the number of people subject to higher rates of income tax is that the composition of both groups has changed substantially.

First, higher rates of tax are reaching significantly further down the earnings distribution. In 2019–20, the last time the higher-rate threshold was increased by more than inflation, it stood just below the 90th percentile of employee earnings. As Figure 2 shows, however, under a two-year extension to the freezes, the higher-rate threshold would be expected to reach the 75th percentile of employee earnings by 2029–30, meaning one in four employees would be paid more than the higher-rate threshold that applies outside Scotland and so generally be subject to a 40% or higher rate of income tax.4 

Second, at the other end of the earnings distribution, more minimum wage workers are being brought into income tax – driven by both the tax freezes and substantial minimum wage increases. In 2015–16, just before the minimum wage started rising rapidly, a minimum wage worker would have needed to work 31 hours a week for a year to pay income tax.5  Should the Chancellor choose to continue the tax freezes for another two years, we estimate that figure would fall to just 18 hours by 2029–30 – the lowest level since the minimum wage was introduced in 1999. In other words, increasingly, even part-time minimum wage workers can expect to pay at least some income tax on their earnings. One consequence of this is to reduce how much of a minimum wage rise goes to workers, with more of the pay rise being recouped by the exchequer in the form of tax. A freeze extension would result in a full-time minimum wage worker paying £137 per year more in tax relative to current policy and £759 more than if there had been no freezes in the first place.6  

Third, for the first time since its introduction, the full new state pension is set to exceed the personal allowance in 2027–28. Since the state pension is taxable, this will require many more pensioners to pay income tax. Currently, the full new state pension is £11,973, meaning that pensioners with no other taxable income do not incur income tax. In 2022–23, just under half of those on the full new state pension were taxpayers.7  By 2027–28, that figure will be 100%. Unless the government grants an exemption, pensioners with low incomes will be required to begin paying tax directly to HMRC (note that taxation of the state pension is not currently handled through the PAYE system),8  creating an additional administrative burden for millions of people. What is more, it is possible that many single pensioners who are in receipt only of the full state pension could become eligible for a small amount of pension credit (a means-tested benefit) because the tax payment will push their after-tax income below the pension credit threshold. Many other entitlements, including the free TV licence, are passported from pension credit, so this could both impose material administrative costs and be fiscally costly. It will also make future increases to the state pension less costly overall for the government and less beneficial to pensioners, since a larger fraction of state pension increases will be returned to the exchequer via higher tax. If the Chancellor chooses to extend the freezes, therefore, everyone in receipt of the full new state pension can expect to experience an increase in taxes.

Finally, since most means-tested benefits are by default uprated in line with inflation, individuals in households that receive means-tested benefits are increasingly likely to also pay income taxes. The number of taxpayers in families entitled to universal credit (UC) would rise to 3.1 million if the freezes are extended, 690,000 more than if there had been no freezes and 110,000 more than under current policy. This matters because these workers face particularly weak work incentives. Families in receipt of UC see 55p of benefit withdrawn for every £1 increase in the family’s after-tax earnings, meaning that workers on UC who do not also pay income tax or NICs keep 45p for every pound they earn. However, those who do pay income tax and NICs typically keep at most 32p from an extra £1 earned.

The flip side of this final effect is that since the effective marginal tax rates of UC claimants brought into income tax as a result of freezes are high, UC claimants who are subject to income tax are partially protected from the tax increases that threshold freezes would otherwise entail: because UC entitlement depends on earnings after deducting taxes, every £1 increase in tax means a £1 decrease in after-tax earnings, and so results in a 55p increase in UC. Even without this effect, a two-year extension to threshold freezes would be progressive over much of the household income distribution – largely because people who earn less than the personal allowance are unaffected by the freezes. As illustrated in Figure 3, however, the UC offset would make such a measure modestly more progressive by sheltering lower-income households from some of the tax increase. 

Conclusion

Since 2021, successive Chancellors have raised revenue by extending the freezes to personal tax thresholds – allowing inflation to erode their real value over time. At last year’s Budget (despite speculation to the contrary), Ms Reeves chose not to follow suit. This year, with the public finances tight, the Chancellor may choose to raise £8.3 billion by extending the freezes for two more years.

Reducing the real value of tax thresholds is not an unreasonable way to raise revenue. If the Chancellor does choose such a course, however, she should do so because she favours the model of personal taxation that such a policy entails – one in which the lowest incomes (including those of pensioners with no private pension income, minimum wage workers, and workers receiving universal credit) are increasingly subject to tax and where a rising fraction of employees are subject to higher rates of tax – not because she believes such a change would be less likely to be noticed by taxpayers.

What is more, if the Chancellor does wish to reduce tax thresholds, extending the cash freezes is a poor way of achieving such an aim. Freezes mean that the real value of thresholds, and the amount of revenue raised, are determined by the (uncertain) path of future inflation. Rather than specifying nominal thresholds and letting the vagaries of inflation determine their real value, Ms Reeves would be far better advised to do the reverse: she should specify the real value she thinks thresholds should be set to, and adjust nominal thresholds in light of what happens to prices.