EDIT: The piece below doesn’t properly take into account a couple of paragraphs in the article I’m commenting on, so misses the point slightly. I’ll do a follow-up shortly to address that.
I read David Quentin’s article on tax avoidance (Risk Mining the Public Exchequer, http://www.davidquentin.co.uk/Risk-Mining_The_Public_Exchequer.pdf) with interest, but although some of the thoughts are interesting I disagree with the conclusions he draws, the definition of avoidance he comes up with, and his suggested approach to solving the tax avoidance problem.
My disagreement is best done set out by example. Let us take a common uncertain filing position: capital allowances. For a convenient example, I take a query which was recently published in Taxation magazine: a farming company wants to put a borehole and associated water extraction equipment on land it farms, which is owned by the shareholders; can it claim capital allowances?
The advisors clearly consider this to be uncertain, or they wouldn’t have written to Taxation. There are a few areas of uncertainty – is the borehole plant? Is the associated equipment debarred by List B of CAA 2001? Does the fact that the land (and by extension anything added to it) belongs to the landlords mean that the company can’t claim?
So there are number of possible filing positions. They could:
– Claim nothing
– Claim part of it
– Claim it all
Now my first reaction to reading the query was that a full claim would almost certainly be appropriate and rather than being technically unsound would only fail if HMRC got stubborn and the client didn’t want to take them to Tribunal; but I wouldn’t say it was clear-cut. The published replies seem to agree with that stance.
However, claiming nothing is clearly the position of maximum certainty. Claiming part or all gives us a risk that HMRC will challenge the claim, and we are not sure whether we will succeed against such a challenge.
Using Quentin’s methodology, then, we cannot decrease the tax bill without decreasing certainty, so the inflexion point is right at the Y-axis. Alternatively, if we include other capital allowances claims we might say that we have gone as far as we can to the right on the graph already – the position of the Y-axis is somewhat variable, as it depends on how much you know about tax before taking advice rather than anything objective.
Anyway, the inflexion point comes when and if we decide to claim capital allowances on the borehole and associated equipment, rather than not claiming. At that point, we are starting to avoid tax.
If we follow Quentin’s flowchart, the quality of that avoidance depends on whether allowances are actually due, and whether HMRC challenges the claim. If the allowances are actually due, then there is “effective avoidance”; if they are not, then either there is “ineffective avoidance” or the taxpayer pockets cash to which it is not entitled (due to lax enforcement by HMRC).
If we adopt Quentin’s views, then, the prudent approach would be to claim no allowances, in case they are not due.
To me, labelling a capital allowances claim which is not absolutely clear-cut as any sort of avoidance is absurd. HMRC challenge all sorts of capital allowance claims: I once had one argue that a water tank was in item 4 of List B (“A dam, reservoir or barrage…”) on the grounds that the dictionary definition of a reservoir was “a place where a liquid is stored” (or something like that), and so any water tank must be in item 4. He withdrew the challenge when I pointed out that even a fountain pen has a reservoir of ink, and his logic would put it in the same capital allowances position as the Hoover Dam or Kielder – but by Quentin’s argument, claiming capital allowances on that water tank was avoidance.
Indeed, given that a petrol tank in a car is a “reservoir”, one could construct an argument that all motor vehicles are “other equipment associated with the… reservoir”, and so any claim for allowances on any vehicle (or fountain pen) is avoidance. There are a host of other capital allowances claims that could be similarly regarded.
So this definition of avoidance gives false positives and is therefore, in my view, flawed.
Can the flaw be fixed?
I think the problem stems from Quentin’s definition of tax risk as being something which is deliberately created by a taxpayer when he puts himself in an uncertain tax position. This, I think, is only a sub-class of tax risk. Taxpayers find themselves in uncertain tax positions all the time, through no action (or inaction) of their own. I think it would be hard to argue that the farming company which has drilled a borehole has deliberately created a tax risk for itself: it think it is fairer to say that the company has created a commercial position, which has come with an associated tax risk that the company wishes were not there. The risk is simply that the tax position will not give credit for the economic cost sustained by the taxpayer: a risk that there will be an unexpected cost, rather than a risk that benefits you might hope to accrue will not.
This leads me to suspect that Quentin’s analysis works only in a narrow subset of cases: those where the taxpayer’s motives are primarily about tax, rather than commercial. If you do something simply to reap the tax benefits then you’re on his graph. Take an individual taking advice on where to invest cash he has in a bank account. Sitting where it is, there is a tax charge on the interest. Putting some in an ISA reduces that tax charge without introducing any form of uncertainty, and is clearly on the horizontal part of the graph. Investing some in a marketed tax planning scheme is going down that slippery slope. But these are things with no economic difference (a cash ISA is not substantially different from a normal savings account). Investing in an EIS company, with all its uncertain tax benefits (your relief is in part dependent on the future actions of the company, which always gives me the willies when clients invest in start-ups), has economic impacts and would not normally be regarded as avoidance.
So Quentin’s argument seems to boil down to economic substance: if you do something with a tax effect but no economic impact, then unless you are sure the Government is happy for you to do it, you are engaged in tax avoidance.
That doesn’t seem to be radically different from previous definitions of avoidance, to be honest.
One way to perhaps salvage the definition would be to replace “maximum certainty” with “reasonable certainty”: if you are fairly sure that you have the right position, then you’re not avoiding tax; if you think you’re probably wrong but are prepared to take a punt, then you are avoiding. This of course gives a problem in defining “reasonable”, and in gauging whether you’re be above or below that point, but I would feel much more comfortable with it. It is of course very much the US approach under FIN48, if you take “more likely than not” to be the definition of “reasonable certainty”, but I imagine that a prudent person might argue that “likely” or “probable” might be more appropriate – or a more aggressive person that “could” is enough.