If you’re going to avoid tax, and at some stage in your life you probably will, you’re likely to do it with your earned income. The reasons for this are pretty clear. Most people have earned income, the difference between tax rates on the employed and self-employed is huge, and the line separating those statuses is both blurred and shorn of logic and economic principle.
Meanwhile, issues around employment and self-employment have leapt to the head of the policy queue. The Office for Tax Simplification has started a review into employment status; the Royal Society of the Arts continues its wide ranging self-employment project; the Office for National Statistics has published a short but influential report noting both that self-employment was at its highest level in forty years and that average earnings from self-employment had fallen 22% since 2008/09; the think tank Centre Forum have described the difference in rates as “without compelling reason” and the leading tax academic Professor Judith Freedman has declared it a “crude tax subsidy” which we should “stop“. On the political front the Labour Party appears publicly to have committed to, in effect, abolishing sole trader self-employment in the construction industry.
What I mean to do in a series of blogs over the coming months is ask some questions about the differing tax treatments of employment and self-employment. What are the differences? How are they justified? How are they arbitraged? Why do employees and employers behave as they do? How does fiscal policy respond to these challenges? And does it respond well?
Where I get stuff wrong, and I will, please tell me. I’ll put the record straight and I’ll try to do it with good grace. If you’d like to contribute to the series, please let me know; the best discussion is a heterodox one.
However, the starting point is probably to observe that, and how, employment and self-employment are taxed differently.
For the sake of simplicity, I consider the situation of an individual, I round to the nearest ’00, I aggregate all taxes on earned income (so not just income tax but also both employers’ and employees’ national insurance contributions. These are, if you like, roses by other names), and I assume no deductible expenditure.
Looked at in this way (and focusing on the position of the typical worker with no other income) the total tax burden on different slices of income (in 2014/15) for the employed and the self-employed (in parentheses) is as follows:
0-8,000 – 0% (0%)
8,000-10,000 – 22.7% (9%)
10,000-41,900 – 40.2% (29%)
41,900-100,000 – 49% (42%)
100,000-120,000 – 66.6% (62%)*
120,000-150,000 – 49% (42%)
150,000+ – 53.4% (47%)
*This band is attributable to the removal of the nil rate band by £1 for every £2 one earns over £100,000.
A few points will immediately be gleaned from the above:
(1) if you were designing a tax system you wouldn’t start here;
(2) no one would describe our rates as smoothly progressive;
(3) no one could describe our rates as particularly progressive;
(4) there’s a huge difference between rates of tax on employed and self-employed income.
(5) whither, now, the Laffer curve?
Follow me on Twitter @JolyonMaugham
Waiting for Godot 
If national Insurance contributions are included then yes, there will be anomalies between self-employed and PAYE workers.
However, NI contributions confer a (albeit increasingly limited) right for PAYE workers to claim unemployment benefits, unlike for self-employed workers. NI contributions do buy a state pension for both groups, however.
National Insurance also creates anomalies between those who receive unearned income and those whose earn (whether through self-employment or PAYE). For example, dividend income is subject to income tax but not to National Insurance, as is income derived from property rents (the latter being regarded as investment income). There may be powerful reasons for this difference to continue but it does create incentives for “tax avoidance”. A cynic might say that the tax system favours capital income over labour income.
Thanks. It’s absolutely a subject that I mean to come back to, whether the differences in tax burden are justified (I don’t think anyone thinks its rational to pretend Class 1 and Class 4 NICs are otherwise than taxes on earnings).
At the moment, the criticisms – from Prof Freedman, the CentreForum and others – are that the benefits difference doesn’t justify the tax differences. For my own part, that’s much too narrow a way of looking at it. We also need to consider the broader economic effects of removing the differential. So far as I am aware, that critically important exercise is yet to be done.
I think including secondary NI the way you do is slightly misleading. There’s no one perfect way to do it, but I think using £113.80 as the denominator gives a better result.
That is, for every £113.80 that is available for employee wages, £25.80 (22.6%) goes in tax/NI in the first band, £45.80 (40%) in the second, and so on.
I’m not an economist and so can’t draw on any theoretical reasons why this should be so, but my experience of SMEs is that they look at the cash in the business available for extraction and then at how much they’ll have in hand.
Apart from that, I agree with what you say.
One other thing that may be relevant, as well as looking at employment/self-employment, would be the incorporation (CT vs IT) issue. This came very much to the fore last year with the partnership taxation condoc, where almost identical mixed structures can now end up with very different tax bills. It’s another way of taxing self-employment (or employment, where IR35 potentially bites), and it’s another area where you can’t easily see a policy rationale for the differing positions.
Thanks. You are quite right on denominator and I will amend after lunch.
True figures are, I think, a lot more complicated, although the pattern would not be dissimilar.
If you are going to include E’er NI then you have to work from a grossed up figure for income. That make sense, because I understand HMRC has evidence that a change to E’er NI will directly feed through to pay – it is the gross cost of employing someone that is relevant. However, both E’ee NI and tax are based on pay, i.e. cost net of E’er NI. That changes both the percentages and the bands if you work on gross figures, or just the percentages if you want to look at total “taxes” as a percentage of income rather than employment cost.
Then there is the question of what else you include. Paul Johnson of the IFS includes High Income Child Benefit charge, since it is a tax unless you give up the benefit, but creates another “funnel” (the graph of the tax rates looks remarkably like an etch-a-sketch of Thomas the tank engine…) between 50k and 60k.
Although it doesn’t impact the difference between employment and self-employment, the next issue is that there are a number of other income-related withdrawals that could be taken into account, most notably tax credits. Those produce total deduction rates that would make bankers squeal in disbelief.
Finally there is the issue already raised about company dividends, but not as a return on capital, rather a different way to route self-employment income. Offering both a complete absence of national insurance while still allowing qualifying years to be accrued if the salary is pitched correctly, it is probably the biggest of the problems.
The answer, as the IFS conference a few weeks ago kept coming back to, is to either merge or align income tax and NI, so they are paid on all income. Whether politicians will ever dare to do so is another matter altogether.
Snap… I was obviously writing while you posted!
Thanks. Not going to get into income tax/negative income tax/benefits thingy because trying to keep it simple. But your point about grossing up is well made and I will correct after lunch. Thank you.
Thanks to each of @MATruman, @Matt_Tax and @ajjacksonCTA for suggesting corrections; now made.
Are your percentages for employees in the higher-rate band correct? They look as if they’re over-stated by 2%.
Well, don’t start from here, but that is one of those jokes! We are, as they say, where we are. Even if we could design a perfect tax system, making changes will create winners and/or losers and so have political costs.
For some peculiar reason, most people don’t seem to mind too much paying “their stamp” (i.e. employee NICs), perhaps because they think it pays for the NHS or some other similarly mistaken notion, but they detest paying income tax. I suspect most employees don’t see employer NICs as a tax that they bear anyway, even if economically it reduces their earnings. So, politically, it is easier to increase “nice” income tax (NICs) and hold or reduce “nasty” income tax.
At least the current system is broadly progressive, particularly if you look at the total income tax paid by a taxpayer as their income goes up, not just the marginal rates: the total amount paid and percentage paid both go up gradually and monotonically. Including benefits can create some awful cliff-edges which really must be addressed. Withdrawing the income tax personal allowance (nil rate band?!) creates a significant distortion, of course.
Don’t you need to include (or at least recognise) the tax position of the employer in an employment scenario? When a person has a realistic choice between employment or self-employment, they will often have some equity in the business, so we also need to think about tax on profits retained for reinvestment, and dividends and capital gains. The only way to avoid tax differences between different business structures would be to tax then all the same way, but again there could be winners and losers.
Thanks Andrew. Am yet to decide whether I want to extend the scope of the debate on from sole traders to (in effect) incorporated sole traders. Will probably wait and see how this goes…
I’m intrigued by the notion that NI “buys a state pension” – actually, I’m not quite sure that’s true any more. Our (in the sense that most of the contributors here are earning) NI is actually paying for someone else’s state pension, as it’s an unfunded scheme. In fact, our NI could be paying for Trident, child benefit, new schools or [insert symbol of state sponsored profligacy here depending on your political affiliation].
Apparently, the notion that their NI was being spent, not saved, came as rather a rude awakening to subjects in recent PwC research into taxes; transparency, and public engagement have a long way to go. Spreading the news about blogs like this would be a good way to help!
Thanks Jason. I just don’t know tbh. But I might see if I can track down someone – @strongerinnos comes to mind – and see if he might answer the point…
Yes, the state pension is unfunded – it is paid out of current income.
However, an unconditional right to receive a payment from current income in one’s old age depends on having paid NI for 35 complete fiscal years. This entitlement is reduced pro-rata for fewer years of contribution.
The method of financing state pensions does not affect entitlement.. NI contributions establish an entitlement and are not a means of funding state pensions (although they may help).
So it is not wildly inaccurate to describe NI contributions as buying a state pension.
An observation on “how we got here” might also shed light. For many years the tax system levied considerably higher rates on unearned income – the last Investment income surcharge was 15% before it was abolished. (I don’t have the date handy). Gordon Brown made very significant changes to dividend taxation with the abolition of ACT, creating a separate tax band for divis, and reducing the effective rate for basic rate taxpayers (and indeed others) as the rate of corporation tax on small profits fell under his watch.
At the same time, NIC was seen as the easy option for any Government keen to obscure tax rises from voters, and hey presto, a complete volte face over a period of around 25 years, so that earned income is now taxed at a considerably higher rate than investment income. Depending on your political viewpoint, you might regard this as idealogically questionable, but fiscally the vast majority of income is earned, so it balances the books quicker.
While I am here, I will add one further observation. I hear frequently the view that becoming self employed is a route to avoid tax, and after almost 30 years in practice I can say that I have NEVER met anyone for whom that is the case. The current surge in self employment for low earnings is most likely the over 50’s who are unable to find employment, but are at a stage in their lives when they can afford to do something different for modest return – lifestyle businesses if you will. They are none the less very committed to success, and enjoying a new lease of life. Many of my West Country clients have moved away from lucrative jobs in London to a slower pace of life, bringing up a family and running their own small business.
Thanks Rebecca. That point about the flip between rates on earned and unearned income is a rather interesting one. It’s a point been made by a number of commentators – I will try to tackle it in a later post (unless there are other takers in the meantime)…
On self-employment and avoiding tax, this is certainly something I will pick up. My own experience is that the picture is complex. Sometimes people contract in such a manner as to become self-employed for tax purposes; sometimes ‘engagers’ are compelled to allow them to do this if they wish to retain labour; but of course the converse is also often true. Shifting in particular those workers who have no economic bargaining power to self-employed contracts strips those individuals of employment rights and saves the engager 13.8% on its wage bill…
Someone has kindly offered to deal with how courts are scrutinising the mechanisms for de-naturing employment contracts in a coming blog post.
Historically wasn’t there some justification in that you couldn’t go straight from self-emplyment to the dole? So you could claim unemployment benefit as soon as you were made redundant, but if your work dried up as a sole trader, you couldn’t.
Yes, but since the 1960s you have been able to claim some form of means-tested benefit in that situation. The only advantage of unemployment benefit was that it wasn’t means tested, so you also got it if your partner was working etc. I’m not entirely up to speed on the current rules, but I think the distinction is similar.
The amount by which the self-employed underpay each year, compared to the employed, and even if the differential benefits are taken into account, was given as £2.8 bn by Judith Freedman. Flat pension in future means that the differences in entitlement are really minimal now.
Thanks Mike. V helpful as always.
I mean to address, probably in my next post, the question whether it is only by reference to benefits that one might seek to justify what you describe as an “underpayment” by the self-employed.
Hmmm…can’t reconcile the 71% on incomes in the £100k-£120k band for employed.
I get – (say) £1,000 available ,
Employers’ NI = £121 (1,000/113.8*13.8)
Salary £879
Tax £527
Employee’s NI £18
Total Tax & NI = £666 Therefore 66.6% (an appropriately diabolical percentage)
Have I missed something?
Only the opportunity to point out – explicitly – that no arts graduate should be allowed, unsupervised, near numbers. Thanks for your correction – it’s much appreciated.
Pingback: Waiting for Godot | The Challenges of Taxing Employment (II) False Self-Employment