Ending marginal rate tax relief on UK pension contributions

 
Much pre-Budget speculation has focused on ending marginal rate tax relief on pension contributions. One newspaper article even suggested that there was a four-out-of-five chance of this happening^[1], though that predated reports suggesting the idea was unlikely to proceed because of concerns about how public sector workers would be affected^[2].

An idea the Chancellor used to support

Rachel Reeves repeatedly advocated this, in various forms, before becoming responsible for Labour’s tax policy:

• In 2011, she wrote that “it cannot be right” that a £1,000 pension contribution costs a higher rate taxpayer less (in terms of post-tax income foregone) than it costs a basic rate taxpayer. This was “very inefficient” and “in urgent need of attention”. “Higher relief for those on lower incomes than those on higher incomes should be explored”^[3].
• In 2016, she advocated a flat 33% rate of tax relief for everyone, such that “For every £2 that savers put towards their pensions [from post-tax income], the Government would contribute another £1”. Tax relief “could be rebranded as a government-backed savings scheme”.^[4]
• In 2018, she was less specific, simply suggesting that higher rate reliefs could be “restricted”^[5].

However, she never spelled out how equivalent treatment would be applied to employer contributions.

Asked about her prior views during the election campaign, Ms Reeves replied: “ What we’re seeking at this election is a mandate to grow the economy. Not a mandate to tinker around with tax rates, that’s not what I’m about…”^[6]

Two taxes and a top-up that could be tinkered with

Ending marginal rate tax relief means moving to a system where at least some income saved in a pension is taxed both when it is earned and when it is paid in retirement (other than for the tax-free 25%, if that does not change), but topped up in between.

Top-ups per pound saved from post-tax income could be the same for everyone, or could vary – for example, they could be bigger on the first £X that someone contributed each year or could be bigger for specific groups, e.g., those with the lowest incomes.

A standalone top-up could be made more or less valuable almost at the flick of a switch and would be a tempting lever for Chancellors to reach for at every Budget. Savers would have to guess whether the top-up would soon get bigger or smaller when deciding whether to save now or later.

Because withdrawals would remain largely taxable, some pros and cons of the existing system would be preserved:

• Some tax revenue would be deferred until people receive their income in retirement, which helps mitigate the future public spending pressures that come with an ageing population^[7];
• The tax system would discourage people from withdrawing pension wealth rapidly (because spreading it across tax years would often lead to less tax being paid); but
• The end-to-end value of incentives for saving for retirement could not be known in advance, making them hard to communicate.

Could tax relief be rebranded as a “bonus”?

Curiously, it is sometimes suggested that making tax relief flat rate would allow it to be rebranded as a bonus/reward. For example:

• The Fabian Society recently recommend that “for the sake of transparency, tax relief should be rebadged as a pension ‘tax credit’, ‘match payment’ or ‘government top-up’. Workers could be promised, say, £1 of government money for every £3 of pension contribution from taxed earnings. This is the approach followed with Lifetime ISAs, Help to Save and Tax-Free Childcare”^[8].
• In 2015, the Association of British Insurers argued: “A key benefit of a single rate, expressed as Savers’ Bonus, is how simple it is for consumers to understand. At a rate of 33%, it could be presented as £1 from the Government for every £2 they contribute, making the benefits of saving into a pension much more visible than they are at present.”^[9]

Presenting things this way would overstate the fiscal incentives to save through a pension. A condition for getting the top-up would be agreeing to pay tax on the same income twice (or one and three-quarter times if 25% of withdrawals remain tax-free) – something the saver would not have to do if they saved outside a pension. That is not how Help-To-Save or the Lifetime ISA work. The top-up would not be a simple gift; much of it would be paid back like a loan. Would pension providers subject to the FCA’s consumer duty be able to reconcile this presentation with the requirement for communications to be “clear, fair and not misleading”^[10]?

For example, with a flat rate of relief at 25%, the Government would add one-third to contributions made from post-tax income, turning £75 contributed into £100. But if 20% tax applied to three quarters of a £100 withdrawal, the saver would be left with £85^[11]. So, the true end-to-end top-up for a person who had been a basic rate taxpayer when the contribution was made would be £10/£75 = 13.3% rather than 33.3%.

Nor could the Government highlight the true number. This cannot be known when the contribution is paid:

• If, when the saver comes to access their money, the basic rate of income tax is above/below 20%, the end-to-end top-up will be smaller/bigger.
• Where private pension income derived from a contribution falls within their personal allowance, the reward will be bigger. (It could be 33.3% in these cases, but that would only be true for the first bit of saving someone did and depends on future governments keeping the personal allowance above the full New State Pension.)
• The reward would become a penalty if a person who was a basic rate taxpayer when the contribution was made paid higher rate tax on the associated withdrawal: after paying 40% tax on a £100 withdrawal, they will have turned £75 into £60 . This could happen because they had income from other sources or if they drew down a DC pot rapidly.

Overseas experience

Proposals to move to a flat rate of relief in other jurisdictions have been limited in scope or have not progressed.

The Irish Government is replacing marginal rate tax relief with a flat top-up for employee contributions to its new automatic enrolment scheme^[12]. However: (a) this is not being extended to employees participating in existing occupational schemes^[13], and (b) employer contributions are not going to be treated as a taxable benefit in kind and will not attract a government top-up – in effect, they will still qualify for marginal rate relief.

Ahead of the 2020 US Presidential election, President Biden’s campaign proposed a flat rate of relief for 401k contributions^[14]. This has not come to fruition .

The “easy bit”: applying flat rate relief to DC pensions

Currently:

• Both basic rate taxpayers and higher rate taxpayers can pay pre-tax income directly into their pension^[15]. A £100 pension contribution can therefore be financed with £100 of pre-tax income.
• The cost in terms of post-tax income will be £80 for basic rate taxpayers and £60 for most higher rate taxpayers^[16]. Not making employees pay tax upfront on the value of employer contributions delivers an equivalent income tax treatment.

A flat rate of relief would seek to change this principle so that basic rate and higher rate taxpayers give up the same amount of post-tax income for every pound credited to their pension pot.

If the flat rate of relief were set between the basic rate and the higher rate, savers currently receiving only basic rate relief would be net winners and savers currently receiving only higher rate relief would be net losers. This is not the same as saying that current basic rate taxpayers would always gain and only current higher rate taxpayers would lose: some basic rate taxpayers effectively receive higher rate relief – either because their own pension contributions reduce their taxable income below the higher rate threshold or because employer contributions would take them over this threshold if taxed as a benefit-in-kind.

Policymakers seriously considering this sort of change would need to consider detailed examples. That is partly to know the scale of gains and losses (especially losses, if they think these will be noticed more).

But it is also because the implementation method chosen would affect how gains and losses are experienced, pending adjustments to pay packages – through pension balances or through take-home pay. The table below shows this, for employees with gross salaries of £25,000 and £75,000, who both currently pay 2% of salary as an employee pension contribution and benefit from an 8% employer contribution. It assumes a 25% flat rate of relief.

Under the first approach, all gains and losses are experienced through the pension input. Under the second, the higher rate taxpayer suffers a reduction in take-home pay (but sees a bigger pension input than now). Under the third, pension inputs are unaffected; the basic rate taxpayer gains in take-home pay while the higher rate taxpayer loses.

Some people’s tax status can change during the year – they might get a pay rise or be out of work for part of the year or have overpaid tax if a new employer applied an emergency tax code. Where this means that the wrong tax/top-up was applied, correction mechanisms could be necessary.

More radically, employer contributions could be replaced with salary supplements used to finance higher employee contributions. In that case:

• If all schemes switched to the relief at source method of administering tax relief, the top-up that schemes claim under RAS could be increased to match the flat rate.
• If there were no separately identifiable employer contributions, NI relief would cease to exist. Some or all of the resulting revenue could be channelled into providing a bigger government top-up to contributions paid from post-tax income. But loss of employee NI relief would entail a bigger first-order effect on take-home pay. The salary supplements employers offered in lieu of pension contributions would be likely to include an adjustment for employer NI costs.
• The status of contracts entitling employees to employer contributions would have to be resolved, and legislation (for example, setting out minimum contributions required under automatic enrolment) would have to change.

Integration with other features of the tax/benefit system

Currently, diverting pay into a pension can increase some people’s entitlement to means-tested benefits such as Universal Credit (which more than a quarter of working-age families are projected to receive when it is fully rolled out)^[18]. The Government would need to be clear about how it wants employee and employer contributions to affect entitlement and to ensure that the new legislation achieves this, noting that changes here could sometimes make more difference than a change in the headline rate of relief.

Unlike with Universal Credit, the current rules for Child Benefit withdrawal (on incomes in the £60k-£80k range) count taxable benefits-in-kind as income. Making employer pension contributions taxable could reduce some families’ entitlements unless the legislation and/or thresholds were changed.

Taxable benefits also affect entitlement to various subsidised childcare initiatives, which stops once one parent’s income exceeds £100,000 (meaning that some people can significantly increase their family’s income net of childcare costs by working fewer hours); if not implemented carefully, pension tax changes could push this cliff-edge further down the income scale, potentially affecting labour supply in some public services.

The “hard bit” – applying flat rate to DB

Although active membership of private sector defined benefit schemes is around 700,000 and falling^[19], DB provision remains a key part of remuneration for around five million public sector employees.

Recent attempts to reduce the cost of pensions tax relief have adversely affected public services: the Tapered Annual Allowance meant some doctors would effectively pay to work overtime (a problem that became particularly acute just before the pandemic). Ending higher rate relief would affect far more people, so the Government would want to tread carefully.

The central challenge with applying flat rate relief to DB is assigning employer contributions (which reflect the cost of providing benefits to the whole membership) to individuals, which is necessary to calculate tax liabilities.

Option 1: tax contributions

The simplest approach would be to multiply the employer’s contribution rate by the member’s pensionable pay and say that this is the value of employer-financed accrual for that individual.

These contribution rates are high. In the main unfunded public sector schemes, the contributions covering current service cost are currently^[20]:

• 18.6% of pensionable pay in the NHS Pension scheme^[21]
• 18.0% of pensionable pay in the Teachers’ Pension Scheme^[22]
• 43.8% of pensionable pay in the Armed Forces Pension Scheme^[23] (which is non-contributory for members)
• 23.6% of pensionable pay in the Civil Service Pension Scheme^[24].

This might be thought an inappropriate way to calculate tax liabilities, because it assumes that the value of employer-financed accrual only varies with salary and not with other factors – for example, it can be more expensive to provide benefits for members closer to retirement, as there is less time to earn interest on the money paid in. Similarly, contributions will vary depending on the assumed investment returns (or projected economic growth in the case of unfunded public sector schemes). Changes to these assumptions (such as when employer contributions to public sector schemes rose in 2024 to reflect lower OBR growth projections) will feed through fed through to tax calculations. This objection also applies to some alternative ways of valuing accrual, discussed below.

It is implausible that employees would be taxed on their employer’s deficit contributions. Even if deficit contributions could be practically apportioned to individual members, that would mean pensioners being taxed many years after they accrued their benefits when the employer needs to top up funding. In this section therefore we focus on contributions in relation to further benefit accrual.

Option 2: Value the benefits: but how?

Alternatively, a value could be placed on the benefits accrued in a DB scheme. This involves an unavoidable trade-off between fairness and simplicity.

At the simpler / less fair end of the spectrum, the increase in an employee’s accrued annual pension over the year (after adjusting the opening amount for inflation) could be multiplied by the same factor regardless of the employee’s circumstances. This is what currently happens for the Annual Allowance (AA), where a factor of 16 is used.

The AA approach is deliberately crude. It takes no account of how an annual pension starting at 60 is worth more than the same annual pension starting at 65. Nor does it reflect the expectation that it will cost more to promise a given annual pension from 65 to a person now 60 than to a person now 30 (because there is less time for the contribution to earn investment returns).

Although some people incur AA charges (especially in the public sector where cash alternatives to pension are less common), the AA primarily works by deterring a relatively small number of high earners from contributing more in the first place. A simple but inaccurate approach is harder to justify when calculating actual tax bills, potentially for millions of people.

Indeed, when the last Labour Government proposed restricting tax relief to the basic rate for the highest earners not for everyone over the 40% tax threshold) via the High-Income Excess Relief Charge (HIERC), it concluded that a more sophisticated approach was needed. Its legislation paved the way for two-way factors, taking account of the member’s age and normal pension age^[25], but the Coalition Government that replaced it in 2010 considered the HIERC too complicated and reduced the AA instead.

A two-way factor approach could be revived under flat rate relief. But even this would fail to capture elements relevant to a pension promise’s value. This is the case when thinking about how schemes differ (design features such as whether pension increases are capped, the size of a survivor’s pension, or the terms on which members can exercise options such as converting part of their pension to a lump sum; how secure benefit promises appear, based on the scheme’s funding level and employer covenant) and about differences between members (how long they might be expected to live, whether they are likely to have a partner who would qualify for a survivor’s pension when the member dies). Trying to capture much of that would be impossible; the point is that not everything relevant to a pension promise’s value could be included in a tax calculation and policymakers would have to decide what goes in and what stays out.

Collecting the tax

As with DC, it is possible to imagine ways of letting take-home pay and/or pension inputs take the strain.

One difficulty with collecting tax from members is that the increase in the pension value may not be clear until after the end of the tax year. But not all 40% taxpayers (still less those who would be 40% taxpayers if employer-financed pension accrual were a taxable benefit-in-kind) can easily settle large, unexpected bills.

It would also be possible to issue tax bills to members but force schemes to offer a “scheme pays” option; the scheme would pay the tax and reduce the individual member’s pension entitlement to recoup the cost at retirement. Take-up of such an option might be high: it can be more tax-efficient to pay from yet-to-be-taxed resources in the pension scheme than from already-taxed take-home pay.

For unfunded public sector schemes, a “scheme pays” approach reduces the short-term net fiscal gains from increased tax. This is because “scheme pays” diverts money from that the scheme that would otherwise be used to pay benefits. Therefore the Treasury will need to give the schemes the money back in order to meet the pension payments (unchanged in the short term). In the longer term, the pensions paid out for members who have used scheme pays would be smaller, but this fiscal gain would largely come beyond the time horizon used for the Chancellor’s fiscal rules.

If, instead of taxing each year’s savings in retrospect, contributions flowing into a scheme after taxes and top-ups were adjusted to be bigger/smaller than under the current regime, benefit design could be altered in response. If this were done through changes to the accrual rate for everyone, it would cancel out the redistributive effects. Theoretically, it would be possible to translate the extra value of Government top-ups for basic rate taxpayers into higher benefits or a DC balance on top of DB benefits – but this would be extremely unwieldy.

Could flat-rate be DC only?

As the thorniest problems with flat-rate relief concern DB schemes, one option would be to make this change for DC only. It has been argued that the Government does not need to change tax relief for its own employees when it can adjust salaries and employee contribution rates.

But this is not an easy get-out :

• MPs are in DB schemes, so the message to constituents paying higher rate tax would be “we’re taking away your 40% tax relief but keeping ours”^[26].
• It would be perverse for the tax system to incentivise DB provision for higher earners (who are arguably better able to shoulder risk) and DC for lower earners.
• People can transfer from DB schemes (except unfunded public sector schemes) to DC schemes.

Application to CDC schemes

In a single-employer CDC scheme, the aggregate increase/reduction in contributions might just affect future pension increases for all members; the change need not have the same redistributive effect as in individual DC. For multi-employer CDC schemes, the distribution of basic and higher rate taxpayers would vary by employer and employers would not want to cross-subsidise each other’s workforces.

Footnotes

1. Taxing times: what could go up in the Budget, by Eleanor Langford, the I, 29 August 2024. Return to article
2. Reeves abandons pensions tax raid to spare teachers and nurses by Oliver Wright and Steve Swinford, The Times, 7 October 2024. Return to article
3. Securing social justice: pensions and savings for all by Rachel Reeves, in “The Purple Book: a progressive future for Labour,” edited by Robert Philpot, 2011. Return to article
4. Letter to The Times quoted in “Reeves: Chancellor must choose flat rate tax relief,” by Mark Sands, Money Marketing, 3 March 2026. Return to article
5. “The Everyday Economy,” by Rachel Reeves, March 2018. Return to article
6. Interview with Nick Robinson, Today Programme, BBC Radio 4, 19 June 2024. Return to article
7. See the Office for Budget Responsibility’s “Fiscal Risks and Sustainability Report,” e.g. chart 4.10. Return to article
8. “Expensive and unequal: the case for reforming pension tax relief,” Fabian Society, August 2024. Return to article
9. “Strengthening the incentive to save: ABI response,” September 2015. Return to article
10. Rule 4.2 of “FG22/5,” FCA, July 2022. Return to article
11. In practice, the £100 may have increased owing to investment returns. But this does not affect the comparison of pensions and ISAs or the valuation of the end-to-end top-up. Return to article
12. “Auto-enrolment: your questions answered,” Department of Social Protection, September 2024. Return to article
13. See “Irish Government publishes the Automatic Enrolment Retirement Savings System Bill,” by Brian Mulcair, WTW, April 2024. Return to article
14. “Biden’s Proposal Would Shift the Distribution of Retirement Tax Benefits,” by Garrett Watson, Tax Foundation blog, August 2020. Return to article
15. Literally under net pay arrangements; relief at source delivers the same result by a more complicated mechanism since tax is taken then given back. Return to article
16. The exceptions are those on the personal allowance withdrawal taper, who can contribute £100 at a cost of £40 from net income, and individuals who contribute to schemes using relief at source and do not claim the extra relief. Return to article
17. The numbers would be the same if, instead, contributions reached the scheme gross and the scheme either claimed a net top-up or paid a net tax, depending on the member’s status. But this would require schemes either to know the member’s tax status (to which pensionable salary information will be an imperfect guide) or to have the calculations performed for them by the employer/HMRC. If contributions were invested before the scheme had access to the calculations, assets may have to be sold to pay tax charges (at least in cases where the member would be making no subsequent contributions, e.g., after leaving employment). Return to article
18. “Post on X,” by the Institute for Fiscal Studies, 15 September 2024. Return to article
19. “The Purple Book 2023,” PPF, p.6. Return to article
20. The total employer contributions are sometimes higher than these figures because the employer also pays contributions to cover administration charges or – much more materially – notional deficit contributions. The figures listed here are the total employer contribution given in the previous valuation report but excluding any adjustments for deficits or surpluses. Return to article
21. “NHS Pension Scheme - Valuation Results - Actuarial valuation as at 31 March 2020.” Return to article
22. “Teachers’ Pension Scheme (England and Wales) - Valuation Results - Actuarial valuation as at 31 March 2020.” Return to article
23. “The Armed Forces Pension Scheme - Valuation Results - Actuarial valuation as at 31 March 2020.” Return to article
24. “Civil Service Pension Scheme - Valuation Results - Actuarial valuation as at 31 March 2020.” Return to article
25. “Finance Act 2010,” Schedule 2, paragraph, 213L. Return to article
26. If the Government did not like the optics of this, it could look to move MPs into a DC scheme, but decisions around their remuneration currently sit with the Independent Parliamentary Standards Authority. Return to article