Reform of the taxation of personal service companies was top of everyone’s list of tips for the Autumn Statement. Reform was widely briefed. But “not yet” was the message. Although it remains on the cards it was not delivered by the Autumn Statement.
Most of what follows I wrote on the morning before the Autumn Statement. I had intended it to be a piece congratulating David Gauke, Financial Secretary to the Treasury, on delivering a much needed policy reform. Obviously I can’t publish that piece. But instead of adding ‘wasted my morning’ to (the bottom of) my (long) list of complaints about the Conservative Government I offer it, mildly revised, as an argument for the reform we need.
To understand why personal service companies are used you need to start with two important facts about the tax system and one about the nature of the employment relationship.
Tax Fact One
First, liability to operate PAYE – to deduct income tax and NICs from payments made to employees – rests with the employer. If the employer (call it ‘XCo’) doesn’t operate PAYE properly, XCo (almost always) carries the can. Even if the consequence of that failure is that the employee (MrY) is better off.
This has the important consequence that every time XCo engages MrY on a ‘freelance’ or self-employed basis XCo takes on risk that HMRC will, later, say that MrY was an employee, leaving XCo with a substantial bill for failing to operate PAYE.
Tax Fact Two
Second, there are lots of advantages for both XCo and MrY to MrY being self-employed.
Above a certain threshold (currently, annualised, £8,112), all income paid by XCo to an employed MrY attracts a 13.8% surcharge. Payments made to a self-employed MrY don’t attract that surcharge.
MrY is also – in cash terms at least – better off. On earnings of between £8,060 and £42,380 an employed MrY will pay NICs of 12% but a self-employed MrY will pay 9%. A self-employed MrY also enjoys a cashflow advantage – and a more generous regime for deducting his expenses.
And the advantages for XCo are not merely cash advantages. It gets to engage MrY with fewer (expensive) employment rights.
XCos are often prepared to share with MrYs some of their advantages in the form of higher pay to encourage them to agree to ‘self-employment’. Sometimes, properly understood, these arrangements are abusive and involve MrY undervaluing his employment law rights – and sometimes they don’t. Indeed, there are many cases where (for this and other reasons) MrY will not work for XCo unless he is engaged on a self-employed basis. There is no hard and fast rule.
The one fact about employment is this.
If MrY has a direct relationship with XCo, he doesn’t get to choose whether he’s an employee or not. Whether he is or not depends on the proper legal characterisation of what he does and what his contract with XCo says. But if MrY is engaged through a personal service company (PSC) to supply his services to XCo he will almost always be self-employed. So XCo and MrY can transform what would be a relationship of employment into a relationship of self-employment by interposing a PSC between them.
The reasons for these tax differences – if there are reasons, and there might be – are poorly understood. I have explored them in some detail here. Many argue forcefully that they should be eradicated. For what it’s worth, for my own part I am not, or not yet, in that camp.
When in 1999 the then Government announced the introduction of IR35 its stated objective was to tackle the use by both engagers and workers of personal service companies to arbitrage tax differences: my Tax Fact Two above.
It sought to achieve its objective by, in effect, ignoring the interposition of a PSC between XCo and MrY. It asked whether, if XCo employed MrY directly he would be an employee or self-employed?
If the answer was “employed” XCo would have an obligation to operate PAYE (see Tax Fact One). And XCo would bear the risk of getting it wrong. At least, that was what was originally proposed.
But XCos didn’t like that risk and they lobbied Government furiously. And the then Government caved and put the liability on the PSC instead.
And that turned out to be a fatal error.
Because, instead of looking at a (relatively small) number of XCos, an overstretched and under-resourced HMRC had to undertake extensive and complex investigations into a (relatively large: tens or hundreds of thousands) number of PSCs.
And if they found a relationship that, ignoring the interposition of the PSC, looked like employment they had to challenge it in the courts.
And each of those cases would have no formal read-across to other PSCs: so HMRC had to litigate them case by case.
And very often, even when HMRC won, it couldn’t collect the tax. The PSC would simply wind itself up and MrY would start a new one. So common was this practice that it acquired a name: ‘Phoenixism’.
Meanwhile, XCo carried merrily on. It continued to have an incentive to engage MrY as self-employed. And so long as a PSC was involved – which the XCo insisted on – XCo enjoyed the benefit of the arrangement and took none of the risk.
And, so far as MrY was concerned, he too could carry on enjoying his share of the rewards. There was only a modest risk of HMRC enquiring – and if it did only a modest risk of any consequences.
The solution to this is remarkably simple. We need to revert to plan A.
The liability needs to rest on XCo. XCo will then show an interest in whether MrY really is an employee. And if he is, XCo will operate PAYE – and MrY will gain employment rights. As the system stands it fails to incentivise anyone to be interested in whether the right tax is paid. It really is as simple as that.
We are talking very substantial sums of money.
The best recent estimates suggest a population of 200,000 personal service companies (see paras 18 and 19 here) used particularly by workers engaged in the oil and gas and IT sectors.
If you assumed (a) average weekly earnings of £800 for those 200,000 and (b) all workers were self-employed, the difference between between Class 1 NICs (paid by the employed) and Class 2 and 4 NICs (paid by the self-employed) would be in the order of £1.2bn.
HMRC have provided an estimate of the difference between all Class 1 NICs (paid by the employed) and Class 2 and 4 NICs (paid by the self-employed) of £2.56bn.
And these figures are before the cashflow advantage and the benefit of the more generous deductibility regime.
Against that the population of MrYs who would be taxed under PAYE if engaged directly by XCos would be much smaller than 200,000. It may well be that a figure of around £500m would be of the right order.
However, it should also be noted that the creation of a new £5,000 tax free band for dividends could open up further and much more substantial opportunities for avoidance which exceed this £500m in scale.