A further note – on Real Facebook’s Accounts

The release by UK Facebook of its accounts made some media noise. Heather Stewart of the Guardian covered them here, Richard Murphy here, Vanessa Houlder here, Hugo Rifkind here and your present correspondent observed that UK Facebook was unlikely ever to make a, or a material, taxable profit here.

Most – with a couple of honorable exceptions – of the twitter taxperts rallied round UK Facebook. A number of bad points were taken in its defence – and no doubt one or two good ones as well. I would not want to pretend – for it is a long mile from true – that I have any monopoly on insight.

It does surprise me that this is the invariable and knee-jerk reaction to stories asking whether Big Business is misbehaving. Although not quite as much as it surprises me that many taxperts don’t understand that the consequence of their myopia is poor understand of tax issues: journalists don’t feel able to call on experts that they can’t trust to tell the unvarnished truth. This is a real bugbear of mine: I’ve written about it (in two addresses to my colleagues) here and here. As I put it in the first of those two posts:

Here’s a short prescription. Rather than bemoaning the limited understanding of public and media, we should work to improve it. I speak to a lot of journalists – several a day – and I’ve only ever spoken to one who wasn’t interested in the truth.

But we need to be transparent about the premise from which we proceed. When we act in a professional capacity it’s right that we talk our own book. Everyone understands that we sometimes speak as lobbyists. But it’s important to signal when we do. Otherwise we become part of the problem. If we merely stand on the sidelines and criticise, we don’t merely ignore Gandhi’s injunction to ‘Be the change you wish to see in the world.’ We thwart it.


Anyway, on to UK Facebook.

One of two points taken in its defence was that you can tell that there is no funny business in its accounts because if you look at the rate of corporation tax paid by Real Facebook on its global profits that rate is really rather high. If true this is a decent point: why would you shift profits out of the UK and into another jurisdiction where you’d pay as high or an even higher rate of corporation tax?

But is it true? Well, at first glance, it looks true. Annex 2A of this document, prepared by the EU Commission, states that Real Facebook paid 45.5% tax on income in 2013. And if you go to Facebook’s own accounts, at Note 13 you find this (and I should note that the accounts refers to income tax which is what our American cousins call corporation tax. For ease of use for my UK readers I am going to call it US corporation tax) :


And this:


They’re probably too small to read on the bus but what they show is that in 2014, Real Facebook made provision for $1.97bn of US corporation tax, an effective tax rate of 40.1% on worldwide income. In 2013, the equivalent rate was 45.5% (hence the Commission figure) and in 2012 it was a remarkable 89.3%!

This surprised me a little because, as the Note itself records, the statutory rate of US corporation tax is 35%. And few of us tax professionals are paid to increase above the statutory rate the amount of tax our clients pay. So I did a little digging.

Not that much, mind. Further down in the very same note you find this:

Excess tax benefits associated with stock option exercises and other equity awards are credited to stockholders’ equity. The income tax benefits resulting from stock awards that were credited to stockholders’ equity were $1.85 billion, $602 million and $1.03 billion for the years ended December 31, 2014, 2013, and 2012, respectively.

So alongside the US corporation tax paid, Real Facebook also received what are shyly described as “income tax benefits” arising from share options (I’ll talk about share options for shorthand although there will be other share based remuneration too) it had granted, of £1.85bn.

Hang on a second: so is the tax provided for by Real Facebook in 2014 $1.97bn (at an effective tax rate of 40.1%) or $1.97bn minus 1.85bn (at an effective tax rate of 6%)?

To find the answer to that question, you need to dig a little deeper.

At page 42 you find this:

2014 Compared to 2013. Our provision for income taxes in 2014 increased $716 million, or 57%, compared to 2013, primarily due to an increase in income before provision for income taxes. Our effective tax rate differs from the statutory rate due to non-deductible share-based compensation, operations in jurisdictions with tax rates lower than the U.S., and tax research credits. Our effective tax rate decreased primarily due to a change in our geographic mix of pre-tax income.

The bit that’s emboldened in that quote was emboldened by me. Now I am not an expert in US Accounting Practice or the US Tax Code. But my understanding is that under US Accounting Practice (and the same is true here) an employer get an accounting deduction when it grants most common types of share option based remuneration to employees to reflect the fact that its incurred expenditure. But under the US Tax Code it doesn’t get a tax deduction when it grants those options; instead it gets that tax deduction later on when the options are exercised. And the amount of the deduction depends on how much the option is then worth.

This timing mismatch has a number of consequences. But in particular, it means that a business with substantial staff costs which costs it chooses to meet in the form of share options will always show a high rate of US corporation tax.

Example. Assume for the sake of example that in 2014 I have income of 100, staff costs of 90 all of which I meet by granting share options of that value, no other costs and the rate of corporation tax is 35%. I will have accounting profits of 10, taxable profits of 100, I will pay tax of 35 and I will have an effective rate of corporation tax of 350%.

But, of course, in the real world, you can’t ignore that, in the future, when those share options are exercised, you will get a tax benefit.

Assume that the options in Example were all exercised on 1 January 2015 at a price which reflected my expectation of their value at the date in 2014 when I granted them. I’d then receive a tax benefit of 35% of 90 or 31.5. That tax benefit would be attributable to my activities in 2014. And if you matched it to 2014 I would have paid tax on my 2014 activities of 35-31.5+ 3.5 and my effective tax rate would be 3.5 on 10 or 35%.

Of course, in practice it’s a little more complicated than this example implies. In 2014 we would show on our balance sheet an expectation of what tax benefit we expect to get in the future (an item called a deferred tax asset) which we would adjust upwards or downwards depending on how closely actuality delivered on our expectations (perfectly in my example, less so in practice). And the tax benefit we got might not be attributable only to our activities in the year in which we got it (it is in my example) but in earlier years too. But these complexities shouldn’t obscure the fact that the tax benefit derives from our activities and reduces the amount we have to pay to the tax authority.

In 2014, Real Facebook made an upward adjustment to its deferred tax assets attributable to stock options of $1.85bn. That figure is not wholly attributable to Facebook’s activities in 2014. But it does reflect share options granted by Real Facebook to its employees – which is a cost of Real Facebook doing business. And it does reflect a 2014 upwards adjustment to Real Facebook’s expectation of tax benefits arising from the grant of those share options. And you can’t ignore it and pretend it has nothing to with Real Facebook’s real US corporation tax liability. And if you match it to Real Facebook’s actual provision for US corporation tax in 2014 it drives Facebook’s effective tax rate down from (an improbably high) 40.1% to (a less surprising, for us cynics at least) 6%.

And the argument that you can tell that there’s no funny business in UK Facebook’s accounts because of the high rate of corporation tax paid by Real Facebook simply evaporates.


I said there were two arguments advanced by the taxperts in UK Facebook’s defence. The other is the total tax contribution argument. It is that you shouldn’t focus on the corporation tax paid (or more accurately not paid) by UK Facebook because of all of the other taxes that UK Facebook causes to be paid. In my original piece on UK Facebook’s accounts I said of it this:

In any event, clearly, it’s no defence for X, facing an allegation that it doesn’t pay the appropriate amount of corporation tax, to say: ‘well, I’ve done some other things that the law requires of me.’

The commentariat didn’t think there was any of that going on. Here, by way of example, was one response from a commentator:

I can’t imagine anyone would seek to excuse legal non-compliance in one area by saying ‘but they comply with some other law’…

It’s worth pausing to note that I was making a general observation about how tax avoidance is justified. And that justification certainly is advanced. Speaking yesterday of UK Facebook here Chas Roy-Chowdhury, head of taxation at the Association of Chartered Certified Accountants, said this:


This is, it seems to me, a diversion. The allegation that someone has avoided corporation tax does not seem to me to be answered by pointing to the fact that they comply with other of their legal obligations.

Some facts about Fixed Odds Betting Terminals

Here are some facts about Fixed Odds Betting Terminals.

For bookmakers they “are one of the most profitable forms of high street gambling” (see paragraph 1.190 here). In September 2014 there were 35,059 such machines (see Table 2) generating an average revenue per machine (amounts staked minus prizes) of £46,315 (see Tables 2 and 3).

They are known to be used by drug dealers to launder money, an issue that Government recognises.

A 2005 report by Europe Economics for the Association of British Bookmakers showed that a bookmakers’ gross margin on FOBTs was 2-3% (see paragraph 1.2.4 here), a figure consistent with March 2012 data published by the Guardian (showing a return of 97%). Taken together with the average revenue per machine, this implies an average of over £1.5m per annum is wagered on each machine. Of course, criminals have to pay people to feed this money through machines but they still look (to this uneducated eye) a cheap way to launder money.

The statutory regulator, the Gambling Commission, has recognised that, of those who use machines in bookmakers, 50% show characteristics which might indicate that they might be a group at risk of harm (paragraph 9.11 here). And machine gambling in bookmakers and problem gambling in bookmakers are correlated with low income (see paragraph 9.9 and 9.15 here).

We also know that Licensed Betting Offices (or LBOs) are clustered in areas of high deprivation:


And in areas of high unemployment:


(both of these charts from this report by the Responsible Gambling Trust which works closely with the Gambling Commission). And that substantially all revenue generated by FOBTs is generated in bookmakers (£1.613bn of £1.623bn see Tables 3 and 8 here).

Government benefits too, and through it so do we. Gambling duty applies at a rate of 25% to that revenue (totalling in September 2014, £1.623bn – see Table 3 here) to generate a yield of some £400m.

And that leaves net revenue for bookmakers (after machine games duty) of a little over £1.2bn.

So let’s sum up.

Bookmakers gain £1.2bn. Government gains £400m. Those revenues come (in unknown part) from money launderers. They also come from gamblers, 50% of whom show characteristics which might indicate they are at risk of harm. And the poor are disproportionately represented amongst those gamblers, a factor that bookmakers take into account in siting their shops.

I should also recognise that bookmakers say that these gambling winnings support employment and so deliver employment taxes.

All Government is about making choices. We make these choices when we decide how hard to clamp down on the money laundering which primes the UK property market bearing in mind that the structures favoured by money launderers deliver unexpectedly high revenues to Government. We make these choices when the College of Policing decides that an answer to the question how “to deliver policing in an age of austerity” is to accept revenues from the Saudis. We make these choices when we decides how aggressively we want to pursue the low rates of corporation tax that have been described by one academic (reasonably, in my view) as an attempt “to become a tax haven” – but which are expected (reasonably, in my view) to deliver increases in the size of the UK tax base.

I don’t want to stand on a high horse and pretend these choices are easy. But an equation that delivers modest (in exchequer terms) amounts of revenue to the Government, substantial amounts to bookmakers, where that revenue is drawn disproportionately from the pockets of the poor and the unemployed, and which facilitates money laundering, doesn’t look like a good choice to me.

Some brief thoughts on Facebook’s accounts

Facebook UK Limited – which I’ll call UK Facebook – has published its accounts for the year ending 31 December 2014. Heather Stewart of the Guardian has written about them here. I wanted to add a few points of my own.

First of all, don’t be fooled into believing that these accounts tell you anything about what advertising revenue or profit Facebook makes from UK based advertisers, or advertising targeted at UK consumers. UK Facebook’s business is this:


So its business is providing services to what I’ll call Real Facebook.

Second, UK Facebook could have chosen to tell us a little more about the transactions it enters into with Real Facebook. But, as is its right, it didn’t:


I don’t suggest there’s anything legally wrong in it having declined to disclose this material. But it seems – to me at least – a strange decision for a company whose tax affairs are very much in the public eye. It could have taken the view that its reputation would be better served by transparency – but it didn’t. And people are bound to ask the question, ‘why?’

Third, its accounts show a more than doubling of turnover from 2013 to 2014:


Most businesses’ profitability improves when they double their turnover. But not UK Facebook’s: its pre-tax loss actually increased from 23% of turnover in 2013 to 27% of turnover in 2014. That’s not a feature most people would expect to see in a rapidly growing normal business.

Fourth, another notable feature of UK Facebook’s accounts is its huge staff costs:


These are obviously substantial: as a proportion of turnover they were 82% in 2014. But, perhaps even more remarkably, they are static as a proportion of turnover: in 2013 they were also, yep, 82%. Again, that’s not a feature most people would expect to see in a rapidly growing normal business. You’d expect staff costs as a proportion of revenues to decline as a business increases in size.

Now, the most likely explanation for this is that UK Facebook charges its services – largely consisting of its employees – out to Real Facebook on a cost plus basis. I’m not suggesting that there’s anything unlawful about this. But it does rather imply that UK Facebook may well never make a profit. Because what are described in its accounts as its revenues are really just its staff costs multiplied by a number (here 1.22). And that extra 0.22 may well never be enough to cover office costs, fixed assets and so on.

Certainly that 82% (or 1.22 multiplier) ratio suggests that the UK taxpayer won’t ever enjoy meaningful profits from whatever success Real Facebook enjoys in the UK.

Indeed, that seems to be UK Facebook’s own view. Its accounts set out how it treats its deferred tax assets (basically, the right to set past losses against future profits) and states:


And does UK Facebook recognise deferred tax assets?


It does not.

Finally, Heather’s article is attracting some criticism from the twitter intelligentisa for failing to recognise that, by paying remuneration in the UK, UK Facebook is actually increasing its UK tax liability. That criticism seems to me to be misplaced for a number of reasons.

  • Even if you assume that all of the remuneration it pays in the UK is paid to UK resident employees, the net effect of doing so is that its employees acquire an income tax (and modest NICs) liability. And UK Facebook acquires a liability to pay Employer’s National Insurance Contributions of 13.8%. That rate is lower than the rate of corporation tax (likely to be 21% depending on timing) on profits diminished by the payments to employees.
  • You want to be careful not automatically to assume that, because the staff are employed by UK Facebook, it follows that they pay tax on their employment income in the UK. For example, if UK Facebook engages staff who are resident in the US, it is the US, not the UK, that has main taxing rights in respect of the income of those staff. So UK Facebook could be depleting its profits chargeable to UK corporation tax by paying salaries to staff resident elsewhere on which salaries no UK income tax liability accrues and which still deplete the corporation tax liability of UK Facebook.
  • Moreover, the average staff member enjoys a remuneration package with an average cost to Facebook of £238,000.  No one pays their staff that unless there’s pretty vigorous competition for the services of those staff. It’s at the very least very possible that, if Facebook wasn’t paying them here, someone else would be.

In any event, clearly, it’s no defence for X, facing an allegation that it doesn’t pay the appropriate amount of corporation tax, to say: ‘well, I’ve done some other things that the law requires of me.’


Can we simplify the tax system?

The UK’s tax code – in 2015-16, because next year it will be longer – runs to 22,298 single spaced, small font, heavily footnoted pages. That’s two-thirds the page-count of the 32 volume Encyclopaedia Britannica, which affected to summarise the sum total of human knowledge.

The immediate consequences of this state of affairs are uniformly negative. Complexity clogs the ability of business to grow; reliefs, poorly attuned to the behaviour they’d like to incentivise, distort the decision-making of consumers and businesses alike; new technologies enable arbitrages that reward fiscal and punish commercial efficiency; and enormous compliance costs barnacle the journey to profit.

We have an Office for Tax Simplification. Setting it up, the Coalition delivered on a 2010 Conservative Manifesto Pledge, one which grew out of a Geoffrey Howe chaired report entitled ‘Making Taxes Simpler’. That report took “as its starting point (and rightly so) the proposition that the UK direct tax system cries out for simplification and reduction in scale.” Its main recommendation was the establishment of an Office for Tax Simplification.

But in each and every one of the five years the Office for Tax Simplification has existed the tax code has grown by an average of 900 pages.

How did we get here? And is there any room for optimism?

In April this year, Andrew Tyrie MP, Chair of the Treasury Select Committee, called a meeting with a small group of hand-picked leaders from the tax field – academics, lawyers and accountants – to discuss proposals to improve the functioning of the Office for Tax Simplification.

The tax profession is pretty uniform in its desire for progress – whatever you might read elsewhere – and suggestions weren’t slow to come.

But they foundered on a single, political, reality.

Achieving tax simplification isn’t a technocratic exercise. It creates winners and losers. And losers make a lot of noise – remember the so-called Granny Tax? A simplification measure to freeze the personal allowances of over 65s and over 75s until the rest of us caught up – and the choice of who they are is an intensely political one. And those political choices will often fail to coincide with what, in purely abstract terms, good tax policy looks like.

Indeed there are occasions when it will stand in direct opposition to it. A good recent example of this is the Google Tax.

Business hated it: it tore up the Coalition’s Corporate Tax Roadmap. The tax profession was no more enthusiastic: the ACCA, ICAEW, CIOT and others lined up to slam it as radical, introduced without proper consultation and pre-empting international measures to tackle avoidance. All in order to raise washers. Nevertheless, and unopposed by Labour, it became law.

It was dictated by politics. The (unfounded) vulnerability of the Coalition to allegations it was soft on tax avoidance rendered the Google Tax politically necessary. And no appeal to process, or simplification, or anything else would persuade the then Government otherwise.

Understand this and you’ll understand that meaningful tax simplification involves the exercise of real political will. As Andrew Tyrie but few of his attendees recognised, politicians won’t delegate their most powerful mode of electoral patronage to an unelected technocracy. And nor, you might think, should they. And the abstract advantages of good tax policy aren’t, on their own, sufficient to cause politicians to act.

So what, then, is left?

There will be occasions when political and technocractic aims are coincident. And when those occasions arise we can expect a rationalisation of our tax system – but we shouldn’t kid ourselves that the driver is a technocratic one. We can hope for, indeed expect – for there is an uncertain but steady drift in this direction – a process that produces better quality legislation. But I don’t see substantial reductions in page numbers in our future.

Perhaps we are better off working with this?

We live in a complex world. The modes through which we transact collectively multiply. The competition between businesses for a competitive edge takes in the attractant force of lower effective tax rates on the pricing of capital. Human nature is that we’d prefer that tax liabilities slipped from our roofs and onto those of our neighbours. And our Government jockeys with its neighbour and competitor nations for a greater share of the tax base. Perhaps the tax code we have is what the world we have demands?

But I will offer this prescription.

The overwhelming majority of taxpayers – individuals and businesses alike – don’t need the complex tax framework I’ve described. What they need is something which is rational, workable, and knowable. Perhaps that’s where the Office for Tax Simplification should focus its attention. Maybe the simplification of our tax system looks like its division into two tax systems: a short one for taxpayers who eschew and a longer one for taxpayers who embrace complexity?