Ten better tax cuts than Reform’s £14bn overtime gimmick

Reform UK’s “hard work bonus” sounds simple: no income tax on overtime above a 40-hour week for workers earning under £75,000. But the likely result is modest extra output, massive relabelling of existing hours as “overtime”, and a bill far above Reform’s £5bn estimate. Our central estimate of the cost is £14bn. For that money we could fund several tax cuts that would do much more for growth, fairness and simplicity.

The policy has all the classic problems of a badly designed tax relief. It draws an artificial line between ordinary hours and overtime. It rewards employees but penalises the self-employed. It treats someone contracted to work 50 hours worse than someone contracted for 40 hours plus ten hours’ overtime. It creates a £75,000 cliff edge where a £1 pay rise could cost almost £6,000. And it invites employers and employees to redesign contracts so that ordinary pay is magically transformed into tax-free overtime.

We described Labour’s increase in employer NICs as one of the worst possible tax increases. Reform UK’s proposal is one of the worst possible tax cuts. We scored it against other plausible £5bn tax cuts on the “bang for the buck” – the long-run GDP boost per £ of revenue forgone from the tax cut. The Reform UK proposal is at the bottom of the list:

The reason is simple: the amount of overtime worked in the UK is fairly small, so the “upside” of attracting more overtime is limited. On the other hand, the amount of time that could be relabelled as overtime is very large: so the “downside” of incentivising people just to relabel their time is very large.

This report quantifies all these effects. Our £14bn central estimate comprises a static cost of £4bn, plus £7bn of people “relabelling” hours as overtime, plus £3bn of additional use of personal service companies, minus additional National Insurance receipts of £0.2bn.

Reform UK have fallen into the trap – more often associated with parties of the Left – of confusing the outcome they want with the outcome that will follow from the actual incentives created by the policy.¹

We’d all be much better off if, rather than introducing new complex tax incentives, we focussed on fixing the taxes we have, removing anomalies that disincentivise work, and (if there is space for tax cuts) thinking carefully which tax cuts have the maximum “bang for the buck”.

The proposal

The basic proposal is to remove income tax² on overtime above a 40-hour week for workers earning less than £75,000. Reform calls it a “hard work bonus”, estimates it would cost about £5bn a year, and says it would be funded from welfare cuts. Reform says about 3.2m workers receive overtime pay and that the £75,000 threshold means about 90% of workers could benefit.

Here’s Nigel Farage’s presentation of the policy:

There’s some more detail in this Robert Jenrick interview (transcript here):

Mr Jenrick gives two examples:

• a Heinz factory worker doing seven extra hours a week would take home about £1,000 extra a year;
• a newly qualified nurse doing five extra hours a week would take home almost £1,300 extra.

Mr Jenrick accepted that the £5bn costing included “some reclassification of unpaid overtime to paid overtime”, and said anti-avoidance rules would be developed with HMRC and the Treasury.

The proposal may be inspired by Donald Trump’s “no tax on overtime” pledge. By the time it became law, the US version ended up much narrower than the original promise: it covers only the overtime premium (e.g. the “half” in time-and-a-half), it is capped at $12,500, it phases out above $150,000 of income, and it is temporary.³ The US version reduces income tax receipts by about 1.1%. Reform UK’s proposal reduces income tax receipts by about 1.6%.

The Trump proposal has been criticised by US right-of-centre analysts usually sympathetic to tax cuts.⁴ That appears to be the early response here – Julian Jessop described the proposal as well-intentioned but unfair, complex, vulnerable to avoidance, and bad for productivity.

The fairness problem

The proposal has an “horizontal equity” problem – because people in a very similar position would be treated very differently.

A plumber contracted to work 50 hours and earning £50k will have a very different tax bill from a plumber contracted to work 40 hours and earning the same amount. Pay will vary wildly depending on the regular hours/overtime mix:

So there’s an obvious incentive created: contract for fewer hours, work more overtime.

Another incentive: don’t become self-employed. A self-employed plumber doesn’t have “overtime” so won’t benefit from the policy.⁵

The £75,000 cliff edge

Reform say the relief applies to workers earning “less than £75,000”.

The obvious reason is that, realistically, the policy has to have some kind of limit, or very highly paid workers would start claiming overtime. But a cap is a bad way to do it, as it creates a “cliff edge”: just below the threshold, overtime is income-tax-free; at or above it, the exemption disappears entirely.⁶

Take an example: a worker on £60,000 of ordinary salary with £14,999 of overtime. At £74,999 total pay, the overtime is income-tax-free. Their combined income tax and employee National Insurance is about £14,943. Now give them a £1 pay rise. Total pay reaches £75,000 and the overtime exemption disappears. Their tax and NIC jump to about £20,943. The £1 pay rise has made them about £5,999 worse off. The effective marginal deduction rate on that £1 is about 600,000%.⁷

This is much worse than the existing 60% marginal trap created by the £100,000 personal-allowance taper, because a taper is (by definition) smoother than a cliff edge.

We can visualise it here, in a chart showing how net (take-home) pay varies with gross pay for someone with £15,000 of overtime. Right now, that chart is smooth. Reform UK’s policy creates a £75,000 cliff:

This worker loses £6,000 at a penny pay rise that takes them over the £75,000 limit, and doesn’t get back to where they were until they’re earning £85,000. We’ve seen before that this kind of effect has real consequences – people hold back their hours/income to avoid falling over the cliff.

That’s bad for growth.

The obvious fix is to phase out/taper the exemption, which is what the US rules do. That makes the exemption more expensive and adds complication – and you still have a high marginal rate (just not 600,000%).

A better fix is: don’t introduce rules like this.

Where does Reform UK’s £5bn estimate come from?

They don’t say – but we think it’s a simple static estimate, perhaps with a small adjustment for dynamic effects.

A simple way to replicate it is like this:

• ONS data⁸ gives 28.89m employee jobs, with mean weekly overtime pay of £14.20 across all of them (including jobs with no overtime)
• That is about £21.3bn of paid overtime per year.
• Apply a blended income tax rate of 23% and you get ~£5bn.⁹

We also created a more complex (but still approximate) model which uses ONS data¹⁰ to apply 2026/27 income tax bands and the £75,000 threshold properly, and that gives a slightly lower number of about £3.7bn to £4.0bn.¹¹

So £4-5bn is in the right ballpark for what HMRC would lose if behaviour stayed constant. But that is the wrong question. The correct question is: what would the policy cost once people respond to it? And what is the economic benefit of the policy?

Our estimate of the actual impact

People will respond to changes in tax by changing their behaviour, and sometimes by working more or working less. But there’s a much bigger effect. The response of taxable income to tax changes — what economists call the “elasticity of taxable income”, or ETI — is dominated not by real behavioural changes, but by people changing the form, timing and label of their income to reduce their tax bill.¹² So when a new tax boundary is drawn — between ordinary pay and overtime — we should expect these effects to dominate.

We worked with two economists to approximately model these effects. We consider three separately. The first is positive: people genuinely choosing to work more overtime because the post-tax reward is higher. The second is pure cost: hours re-reported or restructured as overtime. The third is also cost: the creation of artificial overtime using personal service companies.

This chart shows the impact of these effects in three scenarios:

Undoubtedly there would be attempts to limit these effects through anti-avoidance rules. But that runs into a basic problem: there is no substantive difference between 50 contracted hours and 40 contracted hours plus 10 hours of overtime. If the rules test whether someone’s working hours changed after the exemption was introduced, then they either prevent the very behaviour the policy is supposed to encourage, or they fail to stop new jobs being designed with fewer contracted hours and more overtime hours. There is no perfect solution: there will always be “leaks”, and attempts to plug them with anti-avoidance would just add complexity for normal businesses. And all these rules would have to be created from scratch – there’s no tax definition of “overtime” today.

1. The positive effect

Removing income tax increases the reward from overtime – and that’s clearly the intention of the policy. For a basic-rate taxpayer the net reward rises from about 72p to about 92p per £1 of overtime. For a higher-rate taxpayer (below the £75k limit), from about 58p to about 98p. These are sizeable effects, and some people will respond by working more hours.

How many more? Modern estimates of the intensive-margin labour-supply elasticity — how much extra work people do when net wages rise — cluster between 0.1 and 0.3.¹³ In plain English, if the take-home reward for an extra hour rises by 10%, the evidence suggests people work only about 1% to 3% more hours.

Applying those elasticities to the eligible UK overtime base gives modest numbers: extra overtime pay of about £0.8bn to £1.3bn a year, extra NIC receipts of about £0.16bn to £0.26bn, and extra output/GDP (the genuine productivity gain) of about £0.9bn to £1.5bn:

It’s surprising to see such small effects, but there’s a simple reason: there’s only about £16.8bn of overtime below £75,000 in the ONS data.¹⁴ There’s just not much space for the policy to have any impact.

A more subtle problem: more hours is not the same as more output. A long literature finds that productivity per hour falls as weekly hours rise, particularly above 40-50 hours.¹⁵ A policy that incentivises the 41st, 50th and 60th hour of a worker who is already full-time is not obviously the best way to raise UK productivity. We do not attempt to account for this effect, so the output/GDP figures above may be overstated.

2. Hours re-reported as overtime

While the potential economic upside is limited, the downside risk is much larger: there are many more people whose existing hours could be reclassified as overtime, or whose future contracts could be written with fewer basic hours and more overtime.¹⁶

This is not a hypothetical. Robert Jenrick has said “I don’t expect that’s the way it’s going to work and there’ll obviously be anti-avoidance”.¹⁷ Anti-avoidance is invariably what governments say when a relief opens an obvious door; it almost never fully closes it, but always adds complication.¹⁸

There is a very large potential for such effects here. We estimated that ¹⁹, about £868bn of employee pay sits below the £75,000 threshold. If even 1% of that pay were relabelled as tax-free overtime, the income-tax cost would be about £2.2bn. If 5% were, it would be about £10.8bn. If 10% were, about £21.2bn.

We can do better than guesswork. France ran almost exactly the same policy as Reform UK suggest back in 2007 – the “loi TEPA” – and created a wide exemption from both income tax and (importantly) employee social-security contributions on overtime hours.

The French scheme had a significantly greater incentive effect than Reform’s proposal — it removed social-security contributions as well as income tax; moreover French income tax rates were significantly higher than UK rates. However people had fewer opportunities to game the system, because France had a statutory 35-hour working week.

An analysis by Cahuc and Carcillo (2014)²⁰ found no significant effect on the actual hours worked, but a sharp rise in declared overtime, concentrated among skilled workers whose hours were hard to verify.

Overtime declared by skilled French employees rose by 1.3 hours per week more than their actual hours rose.²¹

If the French declared-overtime effect was simply transplanted onto the UK, the additional revenue cost would be in the order of £1.9bn (the 0.4-hour all-worker French estimate), £3.4bn (the 0.7-hour skilled-worker estimate), or £6.3bn (the 1.3-hour stress case for hard-to-verify roles), applied across eligible UK full-time jobs.

But we should expect a worse result in the UK than that simple transplant suggests, for two reasons that pull in opposite directions, and which we can (very approximately) quantify:

• The per-£ incentive is smaller in the UK. France exempted both income tax and employee social contributions, worth roughly 33p of tax saved per €1 of overtime for a typical worker. Reform exempts only income tax: about 20p per £1 for a basic-rate worker, 40p for a higher-rate worker, and around 21p on average across the eligible under-£75k full-time base, using ONS data²².²³ So the UK per-£ incentive to mis-label pay is around two-thirds of the French one.
• The addressable workforce is much larger in the UK. The French 35-hour week was a statutory labour-law boundary enforced by URSSAF and the inspection régime: any hour declared as “overtime” had to be a real hour worked beyond a contractual 35h limit. That confined the French response to a narrow group — roughly 15% of private-sector employees, the cadres and professions intermédiaires with hard-to-verify hours. The Reform proposal has no equivalent statutory anchor at 40 hours. Any UK full-time contract under £75,000 can be drafted or redrawn so that pay above 40 hours is “overtime”, which plausibly puts 30%–70% of eligible workers within reach of the relabelling channel.

Combining the two — smaller per-£ saving but a 2x–5x larger addressable population — gives a UK “distortion potential” roughly 1.3x to 3x France’s.²⁴

If we scale the French IPP declared-overtime numbers by that ratio rather than transplanting them, the implied UK declared-hours channel is around £2.5bn (low), £7bn (central), or £19bn (stress) per year — i.e. roughly double the unadjusted French numbers across the range. Both versions are shown below.

This is all highly approximate, and should be regarded as indicative rather than a forecast. However even the central figure here is more than the entire £5bn Reform attributes to the policy — and this is just one effect.

3. Directors and personal service companies

The biggest challenge to this policy comes from a category of taxpayer who could exploit this relief almost trivially: owner-managers of single person “personal service” companies.²⁵ PSCs have always responded very quickly to potentially favourable tax changes, and we expect they would do here.²⁶

There are at least three distinct mechanisms through which people could restructure to use the exemption.

• New companies. Self-employed people could incorporate so that what was previously self-employment profit becomes director remuneration, with part of it labelled as tax-free overtime. Many self-employed would regard that as fair; after all, they are just equalising their position with their employed peers. Our very approximate estimate is that the impact could be anywhere between £0.1bn and £4bn.²⁷
• Dividend substitution. Existing PSCs could replace taxed dividend extraction with income-tax-free “overtime salary”. Our very approximate estimate of the impact is between £0.2bn and £1.9bn.²⁸
• Wage recharacterisation. Existing director salary could be relabelled: 50 hours of salary becomes 40 hours of ordinary pay plus 10 hours of tax-free overtime. Our very approximate estimate of the impact is between £0.33bn and £2.7bn.²⁹

It’s important to note that the dividend and wage effects would not necessarily be artificial. Self-employed people often work much more than 40 hours. The “overtime” would in many or most cases be real. People using this strategy could – entirely fairly – say they were just trying to achieve the same result as an employed individual.

Adding those illustrative channels gives a PSC/legal-form exposure of roughly £0.7bn to £8.5bn a year if the loophole were not closed. We should repeat that these are approximate estimates and not forecasts.

The obvious response would be to simply exclude directors from ever benefiting from the policy. This is not as straightforward as it sounds – the boundary between a PSC and a “real” business is not always sharp and well-defined. People would game any simple exclusion. For example, a rule that simply says there is no overtime exemption for a director/employee could be fooled by a plumber appointing her husband as the sole director.

Anti-avoidance also creates complication for “normal” small companies that wish to take advantage of the exemption. For example: you deal with the “husband” scenario by excluding companies with connected employees and directors. People then try using other relatives, nominees, partnerships, umbrella-style structures, or several PSCs bundled into one trading company. So the rules have to look through legal form and ask who really controls the business, who benefits from the work, whether several companies are in substance one business, and whether the pay is genuinely overtime or just relabelled salary. That kind of anti-avoidance arms race is familiar in tax: each simple rule creates a workaround; each workaround requires a more complex rule. The result is that ordinary small companies have to take advice on whether they are caught by rules aimed at avoidance.

Tax cuts: getting the most “bang for the buck”

We’ve analysed a range of possible £5bn tax cuts. All have more economic impact than Reform UK’s exemption for overtime. The chart below shows the range of possible effects:

This understates the problem with Reform UK’s proposal because, for the reasons explained above, their “£5bn tax cut” in reality will cost significantly more. So the chart in the introduction to this article presents the same data differently to show the “bang for the buck” – the maximum GDP benefit per £ of revenue foregone from the tax cut (taking the mid-point cost of the tax cut).

(And apologies; the headline says “ten” but there are actually eleven.)

Looking at each of the potential tax cut in turn, with approximate estimates of the cost and GDP impact:

• Abolish stamp duty on shares (SDRT and stamp duty) – cost ~£3.8bn. This is a textbook bad tax: stamp duty depresses share prices, raises the cost of capital for listed UK firms, and gives an unjustified advantage to debt over equity. People often wonder why UK markets underperform and struggle to attract IPOs – here’s one reason. Oxera estimates a one-off 4% rise in UK equity valuations (~£99.8bn) and a permanent GDP boost of 0.2%–0.7% (£5bn–£17bn a year). Here we cautiously assume a £3.8bn cost, but Oxera suggested the net cost could be very small or even zero. ³⁰
• Abolish (or smooth out) the £100,000 personal-allowance taper – cost a few £bn. The taper creates a 62% marginal rate between £100,000 and £125,140, rising to 71% with Plan 2 student loans and ~78% for a Scottish graduate with three children, before falling back to 45%. It’s one of the worst features of our tax system, and incentivises doctors, IT contractors and other skilled workers to cut their hours.³¹ Plausibly boosts long-run GDP by up to £5bn/year.³²
• Cut residential stamp duty land tax (SDLT) – every £5bn saved buys you roughly 40% of SDLT. The Mirrlees Review described SDLT as “a strong contender for the UK’s worst-designed tax” – since then it has gone up three times.³³HMRC’s own analysis suggests a 1 percentage point change in the effective rate moves residential transactions by 5–7%, with knock-on effects on labour mobility, downsizing and housing supply. See our suggestion for how to phase abolition in without raising prices.³⁴
• Cut employer NICs by ~0.5 percentage points (15% → 14.5%) – cost ~£5.6bn.³⁵ We described Labour’s increase in employer NICs as one of the worst possible tax increases. Long-run incidence falls almost entirely on workers, in the form of depressed wages for higher-paid workers, and fewer lower-paid jobs.³⁶ The OBR forecast that Labour’s £25bn employer-NICs rise would knock 0.1% off potential output; reversing part of it would do the opposite, with about 50,000 fewer “lost” jobs per percentage point.
• Cut the main rate of employee NICs by 1p (8% → 7%) – cost ~£5.35bn.³⁷ A clean, universal cut to the marginal tax rate on labour for every employee earning above the primary threshold. Unlike the overtime gimmick, it applies to every extra hour worked, including the hours of part-timers and the self-employed equivalent through Class 4 alignment. It cannot be gamed by relabelling – in fact the opposite: reducing employee National Insurance reduces the incentive to relabel employment as self-employment. Long-run GDP benefit of up to £2.7bn/year.³⁸
• Extend full expensing to leased plant and machinery and to structures – cost a few £bn.³⁹ Investment incentives address a real economic distortion (the tax bias against equity-financed and long-lived capital), rather than a labelling convenience. Potential GDP benefit of up to £3bn.⁴⁰
• Cut the basic rate of income tax by ~0.75p (20% → 19.25%) – cost ~£5.2bn. Affects ~30 million taxpayers, well-targeted at the working population the overtime policy claims to help, and doesn’t create avoidance opportunities. GDP boost of up to £1.7bn.⁴¹
• Raise the personal allowance by ~£600 (£12,570 → £13,170) – cost ~£4.9bn. Takes ~300,000 of the lowest earners out of income tax altogether, partially undoing the fiscal drag built up since the 2021 freeze. Progressive, simple to administer, impossible to avoid by relabelling, and a GDP boost of up to £1.7bn. ⁴²
• Abolish the apprenticeship levy and replace it with a real skills budget – cost ~£4bn. In economic terms it functions as an extra 0.5% payroll tax on large employers with weak links to actual training.⁴³ Replacing it with a properly designed skills levy or direct skills budget could plausibly boost productivity rather than depress it, with a long-run GDP boost of up to £1.7bn.⁴⁴
• Cut the Universal Credit taper from 55% to ~48% and raise work allowances – cost roughly £3–5bn. Lower-paid workers in UC currently face effective marginal rates of 70%+. A taper cut is the single best-targeted way to make work pay at the bottom of the income distribution – exactly the people Reform’s overtime policy mostly excludes. Potentialy GDP impact of up to £2bn.⁴⁵
• Abolish the High Income Child Benefit Charge (HICBC) – cost ~£1.5bn. HICBC creates an effective marginal rate of ~60–70%+ for parents earning £60,000–£80,000 (worse with more children), based on the higher earner’s income only. It is widely regarded as unfair and borderline-unworkable in its current design and is a serious work disincentive for parents. Combine with smoothing the £100,000 free-childcare and personal-allowance cliffs and we’d have a coherent ~£3–5bn “make work pay for parents” package. Plus a potential GDP boost of up to £0.9bn.⁴⁶

People will have different views on the best use for £5bn. Our point is that there are at least ten serious, evidenced-based uses of £5bn that would do more for growth and fairness than a tax-free overtime label. Reform UK has, on the available evidence, picked one of the very worst.

---

Many thanks to K and B for economic modelling and P for the initial draft.

The code for the modelling and charts is in our GitHub repo.

Further reading

Why did HMRC not charge Angela Rayner a penalty?

UPDATED 5pm: We have now spoken to Graham Aaronson KC, who advised Angela Rayner. He says that HMRC’s decision is explained by facts not…

Tax will go up but public services will get worse. This is why.

UK tax as a percentage of GDP is on course to be the highest since 1945. Over the next four years it will increase…

Inside a fake tax refund factory

Updated 3.45pm on 7 April 2026 with Welcome Accountancy’s legal threat. The internet is full of people promising effortless tax refunds. For many, the…

We defeated an £8m libel claim – and it shows libel law and the Tax Bar need reform.

Last month, the High Court threw out an £8m libel claim brought against me by tax barrister Setu Kamal. The Court held that key…

Finance Monthly fabricated an interview with me

Earlier this week, Finance Monthly published an article presented as an interview with me. No such interview took place. I was not contacted. I…

Mandelson and the Epstein emails: could he be prosecuted?

As has been widely reported, Peter Mandelson forwarded confidential Government documents to Jeffrey Epstein, and advised the CEO of JPMorgan to “mildly threaten” the…