How much will it cost us to cut the top rate of tax?

Earlier this week, George Osborne came close to suggesting that by cutting the top rate of tax from 50% to 45% he had increased tax revenues by £8bn. The relationship between that suggestion and the truth is a distant one as I explained here.

Its evident falsity does not, however, make it any less convenient to those whose constituencies, or pocket-books, would benefit from further cuts. And so, predictably, it has precipitated a number of calls for those further cuts – perhaps in the coming Budget.

It is, unfortunately, recondite in these times to want to make policy by  reference to the evidence rather than the heroic assertion of that which is convenient. Be that as it may, here’s some evidence: about who it is who pays the top rate of income tax; how much they benefit from cutting it; and how much it costs the public finances to make that cut.

In this tax year, about 332,000 people will pay the 45% “Additional Rate” of income tax. That’s about half a percent of the population. About 83.5% of them are male. More than 332,000 people earn more than £150,000 per annum – the earnings level at which you begin to pay the top rate – but the tax system offers reliefs to reduce your taxable income, and these reliefs are overwhelmingly accessed by higher earners, as I showed here.

The mean average earnings of someone in the Additional Rate category of earner is a bit over £400,000 pa. That means that more than half of the income of that mean earner is taxed at the Additional Rate. If you are one of the 16,000 people earning over £1,000,000 pa, your mean average earnings of £2.43m pa means that virtually all of your income is taxed at the Additional Rate. If you earn £2.43m, a cut of 5% in the top rate of tax will give you an extra £120,000 in your pocket every year.

In total, Additional Rate payers will pay a bit over £30bn in income tax this year. So if you cut the Additional Rate from 45% to 40% you will give them, collectively, a tax break of more than £3.3bn.

This won’t, however, cut what Government receives by £3.3bn.

Most people aren’t keen to pay tax and when we raise income tax rates we increase the incentive to find ways to avoid that increased burden – either by engaging in tax avoidance, or by retiring early, or working less or leaving the country. When people give up work or work less or emigrate they don’t just avoid the increase in the rate of tax, we also lose the rest of the tax that they pay.

So there are reasons to be cautious about raising rates of tax. But, of course, people who are highly motivated by tax rates tend to live in low tax jurisdictions – they are not in the UK in the first place. As the FT put it earlier this week:


These considerations also act in reverse. When you cut tax rates you can decrease people’s desire to avoid tax, you can increase their inclination to work or work on, and to move to the UK.

We typically call this relationship between the tax take and changes in rates the Laffer Curve. But, as I’ve written here, calling it ‘the’ Laffer Curve is a misnomer because there isn’t only one Laffer Curve – there are different curves for different taxes, at different times and in different economies. And the shape of the curve is affected by a huge number of variables including, in particular, how easy it is to avoid tax.

Back in March 2012, George Osborne announced that he would reverse Labour’s decision to increase the top rate of tax to 50% with effect from 6 April 2013. At the same time as making that announcement HMRC released a paper called ‘The Exchequer Effect of the 50 per cent additional rate of income tax.’ That paper is the only public study of which I am aware into the effects of increasing or decreasing the top rate of income tax in the UK. It concluded that cutting the rate from 50% to 45% would cost around £100m per annum.

However, there is no room for doubt but that cutting the top rate from 45% to 40% would be considerable more expensive.

First, our inclination to alter our behaviour as tax rates change is dynamic. The lower the top rate, the weaker the incentive to change behaviour to avoid it. You can tell this is so because otherwise, by cutting rates to zero, we would raise more than the £163bn we presently receive in income tax. Plainly this is not so. So a cut to a lower top rate (45% to 40%) will cost more money than a cut to a higher one (50% to 45%) – even where exactly the same people are affected.

Second, the Coalition had marked success in tackling personal tax avoidance. One of the arguments as to why cutting rates will not lead to a reduction in tax receipts is that the incentive to avoid tax will weaken. But if it is already extremely difficult for high earners to avoid tax, it cannot sensibly be argued that diminished incentive will have a marked effect.

Third, pay growth in the top percentile of earners – roughly equivalent to those who pay income tax – is strong as this HMRC chart shows:

The number of those earning above £150,000 is projected to grow by 6.5% from 2014-15 to 2015-16 alone. And the aggregate taxable earnings of that group will grow by over 6% (both calculations from data here). As more income becomes subject to that top rate of tax the cost of cutting it increases.

Writing in the Telegraph last week Fraser Nelson said this:

The top rate of tax is, today, 45p but it was 40p throughout the Labour years. And for good reason: it was the optimal rate, taking the most from the highest-paid. If Osborne were to restore the 40p rate, he’d squeeze far more tax from rich and again demonstrate Conservatism in action.

Now, this is heroic asserting by Fraser. But, if you have regard to the actual evidence, there isn’t room for serious doubt about the effects.

Cutting the top rate of tax will deliver to the top 0.5% of earners a tax cut worth £3.3bn. And it will cost the Exchequer very considerable sums of money. My opinion is that these sums are more likely to be counted in the billions rather than hundreds of millions per annum.

Facebook and UK Corporation Tax (2)

Writing yesterday, I addressed the propensity of certain journalists to ready excitement. I examined what the Facebook deal didn’t mean. Today I want to write a little more on what it does mean. What is the likely thought process that led Facebook to this point? And how much cash is it likely to involve?

First, the thought process.

As David Quentin discusses – in an influential piece of work here – for the sophisticated player tax planning is essentially an exercise in managing tax risk. It involves a choice to seek to reduce your tax liability, which choice carries the the cost of embracing more tax risk. If my assessment of the risk as low is right, I save 100. If my assessment of the risk is wrong, I give back that 100 and I suffer the public embarrassment of fighting and losing a tax case.

I will revisit this equation in future weeks: it doesn’t operate in the public interest. But what I want to cover now is how it affected Facebook’s decision as to where to book UK sales.

Facebook understood – as did Google before it – the diverted profits tax to have backfired. It did not have the consequence of bringing Facebook within the (punitive) 25% diverted profits tax net. What it did do was increase the risk profile attached to Facebook’s tax planning. On the left side of the balance sat the prospective tax savings; on the right sat a now enhanced tax risk. But another factor sat, too, on the right hand side of the balance: the cost attached, especially for a consumer facing business like Facebook’s, to the negative publicity attendant on paying too little tax.

You weigh that tax saving against the increased tax risk and impact of that publicity on your global brand and, ultimately, you make a call as to which side of the balance sits heavier. That would have been the assessment that led Facebook’s to rework part of its UK tax calculation.

But what sat on the left side. What was the actual sum on the left hand side of the balance in my calculation. Can we know?

We would need to know (1) what sales were made by Facebook to UK customers. Of these sales, what subset (2) were effected by UK relationship managers? (You will recall from yesterday’s blog post that only these sales, and not advertising bought by smaller customers online, are affected by the announcement).

Then what profits were generated by selling – as opposed to other parts of Facebook’s value chain (3)? Under existing international tax rules the UK can’t tax what you might think of as ‘manufacturing’ profits here, only ‘retailing’ profits.

Finally, one would apply to those profits (4) our prevailing rate of corporation tax.

As to (1), the latest year of accounts available (for Facebook Ireland Limited’s accounts, where European sales have been booked up to now) gives European sales of €4.8bn in the year to 31.12.14. It doesn’t say what UK sales are, but the relationship of the UK’s GDP to that of the EU (18%) might give you a reasonable proxy for the proportion of sales made in the UK. 18% of €4.8bn would represent €860m. Divide by 1.3 to convert to £ would give you UK sales of £660m.

As to (2), we know that Google generates 60-70% of its sales from 1% of its customers. Let’s assume that the same is true of Facebook: that 1% of customers are serviced by the UK relationship managers, and they generate a (mid-range) 65% of sales in the UK. That would give you 65% of £660m of UK sales, or £430m of sales booked in the UK.

As to (3), worldwide Facebook appears to have a 40% margin (of pre-tax profits/turnover). Attribute, for the sake of argument, 10% of that margin to selling and the remainder to generating and maintaining the technology and intellectual property rights. That would imply a profit on those UK sales of 4%, or £17m.

If you apply to that £17m our 20% rate of corporation tax, you’d arrive at a UK corporation tax bill of £3.4m per annum on those sales. That’s a little less than, but not of a completely different order to, the £4-6m (independently) estimated by Tim Davies, Head of Tax at Mazars, yesterday.

As I explained yesterday, the deal is forward looking. It will apply from April 2016 going. So we won’t see any signs of it in the accounts for the period to 31 December 2015 (which should appear in October 2016). The first sign of it will appear in the accounts for the period to 31 December 2016 (which should appear in October 2017). These will contain 9 months rather than a full year of this new treatment giving rise to an additional corporation tax bill disclosed to us in October 2017 of 75% of £3.4m or £2.55m.

But that’s not all.

We also know – see my post of yesterday – that Facebook UK has approximately £10m of ‘deferred tax assets’ (effectively credits against future tax liabilities, and here comprised of carried forward losses and capital allowances). These would need to be depleted before any cash moves from Facebook UK to HMRC. £2.55m in the year ending 31 December 2016 would leave just about enough to clear the tax bills for the years ending 31 December 2017 and 31 December 2018. So the first cash changing hands in consequence of this deal would be in the year ending 31 December 2019 and we’d find out about it in October 2020.

Of course, this calculation makes a number of assumptions which may or may not be right. If Facebook’s UK revenues grow faster the deferred tax assets would be depleted earlier. Earlier depletion might also result from HMRC’s existing enquiries into Facebook UK’s affairs bearing tax fruit. Or the assumptions made in my calculations might just be off.

But, even bearing these in mind, the BBC’s Businesss Editor Kamal Ahmed’s blog post of yesterday which led with:

Facebook is set to pay millions of pounds more in tax in the UK after a major overhaul of its tax structure.

might look, in the cold light of day, a little excitable. “Set,” perhaps, but not for a good few years to come.


Facebook and UK Corporation Tax

This morning Facebook UK announced – via the BBC’s Kamal Ahmed – that it would start booking sales in the UK. Here’s what they told the BBC:


What this means in practice is that Facebook UK’s taxable UK profits will cease to be, as previously they were, a simple mark-up of staff costs (a fact I revealed here).

Instead they will be, in part at least, a function of the value that is generated in the UK by those staff. In principle this should mean higher taxable UK profits.

So, welcome news. But (and you knew I was going to say this) let’s not get carried away.

First of all, Facebook UK Limited appears from its accounts to have three activities in the UK:


being sales support; marketing services; and engineering support.

The “marketing services” and “engineering support” – which you might think of as maintaining the Facebook machine and developing new bits of it – activities will continue to generate UK taxable revenue only on the basis of what the staff cost plus a small mark-up. We should expect these bits to lead to materially no UK tax liability now or in the foreseeable future.

What about what was previously described as sales support? Facebook UK has two types of customer. Those who have UK client managers and those who don’t. It is only in respect of those who have UK client managers that a proportion of the value of those sales will fall within the UK tax net. Sales to those who don’t have UK client managers will continue to be taxed in the Ireland.

So what will now be taxed in the UK is the value added by a sub-set of a sub-set of Facebook UK’s staff.

Second, it’s important to note that, like Google, none of the profits that the Facebook Monolith – i.e. if you assume a world in which what is economically a single entity is taxed like a single entity – makes on UK sales will be subject to UK corporation tax. Unlike Google we can’t know precisely how big those UK sales are – but they are very likely to be billions. What is taxed instead is a combination of fragments of deemed profits on UK costs and deemed profits on deemed value added by a limited number of UK staff. You don’t need me to tell you this ain’t what a sensible corporation tax system looks like.

Third, Facebook has considerable reliefs available to carry forward. You can see that from here:


Previously Facebook UK didn’t bother to recognise these sums (effectively comprising rights to reduce future tax bills) as assets because they didn’t think they’d ever have any UK corporation tax liability to pay. This might now change – but regardless we should expect it to be a good while before today’s deal results in any real cash moving from Facebook UK to HMRC.

So although this morning’s announcement has been well handled in PR terms, it seems to me that it is unlikely to be especially meaningful in terms either of actual cash passing from Facebook or in terms of its tax liability. Facebook UK’s last published accounts disclosed a payment of £4,327. I wouldn’t bet on today’s announcement resulting in any meaningful change.


Another important and related aspect to note is that what Facebook UK have announced today is a forward looking change – effective only from April 2016.

Facebook UK Limited have refused to confirm or deny that they are under revenue enquiry for years up to now – but it’s clear from their accounts that they are.


Of course we can’t know that these claims relate to how they fragment their sales and engineering businesses to minimise their tax liability. But the Google experience suggests that it’s a pretty decent guess they do.

And nothing in today’s announcement deals with these enquiries. Nor the period after them, from April 2015 to April 2016, when the Coalition’s diverted profits tax applied.

But likely it won’t be until October 2016 at the earliest – when we see Facebook UK’s accounts to 31 December 2015 – that we gain any clarity about past years.


What, finally, about the Diverted Profits Tax?

I understand that, like Google, Facebook regards its tax planning as robust enough to withstand a Diverted Profits Tax challenge. But they do accept that it adds tax risk to their business model. And it is a combination of that added tax risk and a response to public perception of them as tax avoiders (despite HMRC’s hugely unhelpful Policy Statement of earlier this week) that is causing them to book at least some real profits in the UK.

This is a real credit to those who have been campaigning for tax justice – through our actions we can achieve  what Government is either unable or unwilling to.


HMRC redefines tax avoidance to exclude Google, Facebook, Amazon

Yesterday HMRC published an “Policy Paper”: ‘Taxing the profits of companies that are not resident in the UK.’ You can read it here.

It contains this extraordinary assertion:


In effect, say HMRC, if it is “the way that Corporation Tax works” then it is not “avoidance”.

There are a lot of problems with this statement.

The most glaring of them is that it has as its consequence that there is no such thing as tax avoidance. If the structure works it’s not tax avoidance. And if the structure doesn’t work, by definition it doesn’t avoid tax, and so it can’t be tax avoidance either.

Another is that it is contradictory to the definition of avoidance that HMRC itself adopts for the purposes of calculating the Tax Gap.


For this purposes of calculating the Tax Gap, HMRC say (and this time rightly) that even if a structure does deliver a tax reduction it can still be avoidance – “where it serves little or no purpose other than to produce a tax advantage.”

But the most extraordinary thing of all is that HMRC is going out of its way to provide political cover for businesses which engage in abusive tax practice.

Where is the public interest in HMRC saying, publicly, that it is not avoidance for businesses to establish with a view to minimising their tax liability these highly artificial structures?

Why on earth is HMRC acting as public relations agency for Google, or Facebook, or Amazon?

Did cutting the top rate really raise £8bn?

Speaking in Parliament today, George Osborne said this:


(thanks to the Guardian’s Andrew Sparrow for the report).

I don’t have the number to which Osborne refers but it is broadly in line with what was forecast in May 2015 which showed a projected increase in income tax paid by additional rate taxpayers of £7.1bn.

Does this increase vindicate, as Osborne suggests, to the tune of £8bn of extra receipts the decision to cut the 50p rate?

Reader, it does not.

To understand the effects of cutting the rate you’d need to understand what receipts would have been if Osborne hadn’t cut the rate with effect from 2013/14 and compare them to the actual receipts for that year.

No one has done that exercise since the cut but HMRC did some projections beforehand.  It calculated that cutting the 50p rate to 45% would cost money, some £360m over five years.


So how do we explain that £7bn increase in receipts?

As Osborne well knows, if you tell people in March 2012 that you’re going to cut their tax bill by a tenth (from 50% to 45%) in a year’s time, people will choose to delay payment until April 2013 when their bills will be lower. And they did.

First, we knew they would do this at the time:


Second, HMRC said in May of last year that it had happened:


In other words, tax receipts were artificially low in 2012-13 (because people delayed receiving income until rates fell) and were artificially high in 2013-14 (when those delayed receipts were received). Combine those two numbers and you may well explain your £7bn jump.

Third, you can buttress the point if you look at the change in the composition of receipts from Additional Rate Payers in 2012-13 to 2013-14.

  • There was little point to them delaying basic rate employment income: the percentage tax bill on that remains static. And in fact the basic rate employment income receipts actually fall from £1.78bn to £1.73bn.
  • There was little point to them delaying higher rate employment income: the percentage tax bill on that also remains static. And the higher rate employment income receipts rise by a modest 7% from £11.2bn to £12bn.
  • But on additional rate employment income, where their tax bill falls by a tenth, there is an increase of over 19%.

That delaying tactic is likely to explain most or all of that £7bn difference. At paragraph A.26 here HMRC forecast that £6.25bn of income would be moved from 2012-13 to 2013-14. A £6.25bn reduction in 2012-13 receipts plus a £6.25bn increase in 2013-14 receipts would give you a difference in expected tax receipts of 45% of £12.5bn or £5.625bn in tax. (That calculation makes the sensible assumption that only income benefiting from the cut – i.e. taxable at 45% rather than 50% – is pushed forward a year).

But it’s not only that Osborne has been a little economical with the truth. It’s not only that, on all the available evidence, his tax cut actually cost money. It’s that the whole episode signals a terrible indictment of Government policy.

Osborne could have taken measures to prevent these delaying tactics – which remember only benefited those earning over £150,000 per annum – but he didn’t.

And this cost the country £2.4bn in 2012-13: see Table A3. (Although it should be noted this figure will unwind in part in later years.)