Further response to David Quentin on tax avoidance

Apologies for the late posting of this: I’m catering at a scout camp over the bank holiday weekend, and preparing for that has taken more of my spare time recently than I anticipated.  I am also likely to be rather incommunicado (in said field) for the next week or so: further apologies for the resulting delay in replying to any comments.

This post has been put together in fits and starts, so yet more apologies if it doesn’t quite hang together.

Sorry about apologising so much, too 🙂

My fundamental point is that Quentin’s model creates false positives, and to eliminate them he makes concessions which then create false negatives.

False negatives

Following on from my previous post, David Quentin (and Iain Campbell) pointed out to me that I was not taking into account what Quentin said in his paragraph 33:

“It should be emphasised that we are talking here about tax risk that has been deliberately created qua tax risk, and not tax risk which has arisen as a result of deliberate but non-tax-motivated behaviour. If you do something in pursuit of your commercial objectives and it gives rise to tax risk, you are not in the “avoidance” zone. You are over on the left of the curve in the zone where you want tax advice to eliminate tax risk, not create it.”

He followed this up on his blog by saying, as a response to my previous post:

“If the tax adviser told it to do something different (e.g. some sort of structured work-around) because that would increase the certainty of being able to claim capital allowances, this would be an upwards movement in “tax advice risk space” (as described in paragraph 12 in my paper) and would not constitute tax avoidance”.

I cannot agree with this: I feel that if a tax advantage is not clearly due, and a taxpayer takes steps with the sole purpose of making as sure as possible that the tax advantage is actually obtained, then this would frequently be regarded as avoidance.

In fact, the phrase “structured work-around” could be used as a fairly close definition of “avoidance” in most non-tax contexts (I avoid problems with accounts and tax software by working around them in Excel, for example).

Taking Google’s situation of selling from Ireland to the UK (rather than selling in the UK) as an example: if the aim is to ensure that for tax purposes sales are recorded in Ireland rather than the UK, and Google issues instructions to its staff as to how they should act to ensure that they keep the sales in the desired place, then issuing those instructions (which may have no, or even adverse, economic impacts) looks to me rather like an avoidance step – but it would not be under Quentin’s analysis. So here we have a false negative.

One could simply say that being careful to minimise the tax risk you create is still creating risk, so overall there is a downward movement and hence avoidance. But I think that the way the model walks through the tax risk space means looking at individual steps, not at the big picture; and I think that outside observers would tend to do so.

False positives

I also think that Quentin is seeing too much of a clear divide between doing “something in pursuit of your commercial objectives [which] gives rise to tax risk” on the one hand and a company doing something “because its tax adviser told it to” on the other. It’s true that in many cases clients do go ahead merrily with their original plans and we then try to make sure the tax is as certain as possible (documentation! Please, somebody think of the documentation!), but ideally one does rather hope that commercial actions do change as a result of what tax advisors say. Tax being a deal-breaker is not uncommon, and tax affecting the form of a transaction is (fairly) routine. I’m frequently asked whether a client should buy or lease a car personally or through their company, for example – people are perfectly prepared to change the economic form based only on tax factors (usually without specifying the car – which I regard as a somewhat crucial variable in the equation – but hey ho).

OK, that is probably a poor example given that as the eventual tax treatment is fairly certain (assuming mileage records are rigorously kept) I doubt it would count as avoidance under Quentin’s model (or many others), but the general point stands: it is often hard to disentangle behaviour which is “tax-motivated” from that which is not.

The borehole example may be better: one point of capital allowances (see the annual investment allowance in particular) is to induce businesses to make investments they otherwise wouldn’t make. So if the business is umming and ahhing about whether to bore the hole, deciding to do so because the advisor has told it that allowances will make it worthwhile sounds like the sort of step Quentin does include in his model:

– It is done due to the taking of tax advice; and
– It creates a tax risk

Quentin would then take it out of “avoidance” territory because it’s done in the pursuit of commercial objectives, although the fact that it would not have been done in the absence of the tax advice might appear to bring it in.

So Quentin has to say that if the action taken by the taxpayer has a significant non-tax impact then it’s not within his model, even if it does have a tax motivation. That is, you are within the model if you have a tax motive but can be taken out by a non-tax one. That would take the car and borehole examples completely out of the avoidance question. But this seems to be simply returning to the “no economic substance” test, and renders Quentin’s risk test irrelevant – or at least, as suggested above, overly narrow and leading to false negatives. It also needs a definition of “significant non-tax impact”, if leasing a car personally rather than through the company is significant but locating your European headquarters in a particular jurisdiction is potentially negligible.

Or, looking at the same example another way: if you would put the borehole in anyway, then you have a choice: you can claim allowances, or you can not. If you claim them because your advisor says you have a shot at getting them, although the position is not clear, that would seem to be something done which is very clearly tax-motivated and creates a risk (of interest and potentially penalties, as well as the tax underpaid), and which cannot have a commercial purpose in itself because the borehole is already sunk.  The claim seems to be avoidance in Quentin’s model, even if the sinking of the borehole is not.

Similarly with undertaking R&D because the tax credit makes it worthwhile: this seems to count as avoidance, if it’s not clear that the credit will be available, and is another potential false positive.

Now I am actually OK with calling this avoidance: in my mental model of avoidance (see below), any step you take to reduce your tax bill is avoidance, whether you use an ISA or Cup Trust; the difference for me is whether the avoidance is acceptable or not, which is an entirely different question. But Quentin seems to regard all avoidance as unacceptable, or at least undesirable (apologies if I am misreading here).


So overall I think that Quentin’s model is interesting, and should certainly be taken into account when looking at avoidance; but it seems not to count some cases of avoidance as such, and it would seem to class some normal tax planning as avoidance, and so it doesn’t seem to be complete. It is too narrow in one dimension (looking only at cases with little economic impact), and too broad in another (covering all cases where there is some risk).

It seems to work better as a potential hallmark of avoidance, but one which is neither necessary nor sufficient.

Perhaps to improve the model the tax risk space needs to be three-dimensional, to include economic impact? Travelling down and right on Quentin’s graph is objectionable if you stay in those two dimensions; but if you also posit a z-axis of increasing economic impact, is travelling south-east acceptable so long as you go uphill while doing it? If so, does the steepness of the slope matter?


I think the better model of avoidance is to look at a two-step process:

1) Define avoidance as the situation where the tax bill has been reduced as a consequence of an action taken by the taxpayer (this is very broad);
2) Classify avoidance as acceptable (ISAs, R&D) or otherwise (BEPS, IR35)

In this scheme, the question of whether an action taken serves only to reduce tax and increase tax risk, as set out by Quentin, would be taken into account when looking at part 2: was the avoidance undertaken for acceptable reasons?

The sort of transaction Quentin is targeting reduces the tax bill with no concomitant economic impact but just an increase in tax risk, so one cannot say the tax consequence is justified; this is unacceptable avoidance (or perhaps “abuse”, depending on how egregious it is).

The borehole example, on the other hand: investing in plant gives one allowances, which reduces the tax bill, so tax has been avoided; this is done in a way which does not simply increase tax risk but has economic effects, the tax consequences of which are clearly in accordance with Parliament’s intentions; so this is perfectly acceptable avoidance (or “planning”, perhaps).

One thought on “Further response to David Quentin on tax avoidance

  1. 1. Following Andrew I have to start by apologising for the fits and starts that have delayed a response to Risk Mining longer than some 140 characters on Twitter. I’m posting here because I can’t find a way to post directly onto David’s blog. One side effect is that this posting has grown with time as I’ve had a chance to reflect on the issues, and on the later posts from David and Andrew. So, forgive the length, or skip to the final paragraph.
    2. First, two general remarks. Although avoidance clearly exists I have some conceptual difficulties on the very term ‘tax avoidance’. As has been said elsewhere we can ask if something succeeds whether it can be ‘avoidance’ (e.g. Lord Hoffmann, “Tax avoidance in the sense of transactions successfully structured to avoid a tax which Parliament intended to impose should be a contradiction in terms. The only way in which Parliament can express an intention to impose a tax is by a statute which means that such a tax is to be imposed. If that is what Parliament means, the courts should be trusted to give effect to its intention. Any other approach will lead us into dangerous and unpredictable territory.” ‘Tax Avoidance’ (2005) British Tax Review, 2, 197, p.206). Or even the more cynical approach, based on the old epigram, ‘Treason doth ne’er prosper: what’s the reason? Why, if it prosper, none dare call it treason.’
    3. It is also often assumed that if “avoidance” is stopped then the gains to the Exchequer can easily be quantified. I am not entirely comfortable with this way of establishing the costs of avoidance. Too often it seems to be presented as a binary position – it’s either the “full tax” or the “net tax after avoidance”. This is indeed a measure, and perhaps an easy one, but if avoider A had realised their avoidance scheme was too aggressive, or not competent, who knows whether they would have adopted different ‘tax planning’. And if Avoider A loses, publicly, then potential Avoider B may not be deterred from avoidance, but adapt A’s avoidance, or be less “aggressive”, so as to escape successful challenge.
    4. Anyway, all of the above is by way of background. I confess to feeling like a stranger in a strange land when I read of “tax advice risk space”, or maybe lacking a good grounding in quantum physics. It seems that the discussion is couched in terms of binary positions. So perhaps it is to physics we should turn to, rather than law or philosophy to discuss what is or is not tax avoidance. Is the “space-time continuum” a richer model because it operates outside the flat two-dimensional framework that David’s model sets out? I think those additional dimensions can provide some increased explanatory power, for example by introducing the sort of considerations that Andrew refers to (and which David feels unnecessary). I believe you cannot have a simple dichotomy between “pursing commercial objectives” and entering the tax advice risk space (changing/creating risk) through obtaining and acting on professional advice.
    5. In fact, the business itself may often have an understanding (or flawed understanding) of tax matters and its decisions motivated by a mixture of motives. Indeed, on more than one occasion, I have seen where a proprietor has set out to do something based on a misunderstanding of the law or accountancy (or both), with tax assumptions. The taking of tax advice has forced them to accept that more tax would be paid than they had hoped.
    6. But, more importantly, I think this binary position is shown in the implications of para 3. I do not think these have been fed through into subsequent discussions, as exemplified in the flow chart at para 31 – “is the filing position correct as a matter of law?”
    7. I have suggested elsewhere that this is really a Schrödinger Box. This is for the reason David sets out in para 3: “As at the point of filing (i.e. after the relevant taxpayer behaviour has taken place but before the outcome has eventuated) there is no way to determine whether or not the tax liability is understated.” So nobody can, at the point of filing, navigate their way through that decision box. It is simply not possible to say, at that point, that by accepting and acting on tax advice the taxpayer has filed an incorrect return. This is regardless of whether that advice eventually, under challenge, does or does not prove to be effective or ineffective avoidance.
    8. I am also not sure if some of the other conclusions follow, e.g. para 43:
    “But deliberately underdeclaring your tax, as with only creating the possibility that you have underdeclared your tax, throws an enforcement burden onto the tax authority which, depending on whether or not a successful tax authority challenge takes place, may result in your not paying tax which is legally payable. And to that extent evasion and avoidance are the same thing.”
    9. The taxpayer puts forward a view, often based on professional advice, as to the tax consequences of a transaction or structure. If, at the time of filing, it is impossible to know if your position is correct or incorrect, then (putting aside the gross cases more akin to evasion) you are surely under no obligation to take a view that might result in overstating your tax. Nor should the tax authority expect to operate a system in which this was the norm. Under the presumption of taxpayer honesty is not the onus on the authority to be resourced to carry out such challenges at the level and depth it sees fit to ensure good compliance? Indeed, this seems on all fours with David’s position set out earlier in para 35 “A policy of challenging deliberately-created exchequer risk where the filing positions are nonetheless more likely than not to succeed would create a gigantic public-sector industry of largely fruitless tax litigation. Even if tax authorities wanted this and had the resources to bring it about, it could not conceivably be politically acceptable because of the corresponding waste of taxpayer resources in resisting these tax authority challenges.”
    10. I know David is unwilling to accept Andrews’ idea of “tax-motivated” behaviour being distinguished from behaviour based on “economic substance”. I confess I can see some attractions in Andrew’s views. I think there is, for example, a significant difference between structuring a business, or business transactions, in such a way as to give rise to less tax than a different structure, than in schemes/structures based on tax advice that, as it were, sit on top of the underlying business profits, and seek to avoid or reduce the tax payable. And in terms of challenging taxpayer behaviours, and possibly applying sanctions to it, I see no reason why one filter might be to separate out the two sorts of avoidance. (Indeed, if the taxpayer has not taken tax advice, so could not have acted on it, would David’s model apply to treat as avoidance those instances where the taxpayer copies other schemes or devices they are aware of? I suspect he would argue it was, but it is a passive role for tax advice and no advisor.)
    11. In this connection I think para 6 is also relevant. Only some forms of transactions and at certain stages can be within its scope. Surely the vast majority of transactions carried out by a business are things like the provision or purchases of goods and services, where those involved are simply customers or providers with no discussion about the form of the transaction. The model seems more appropriate to a tax significant transaction, or a series of such transactions. I think in practice this is more akin to the type of avoidance designed to reduce the tax payable on business profits. (Of course, there can be schemes that have the appearance of business transactions but here a commerciality test might work well.)
    12. Even if we accepted David’s model in full I am not as convinced as he that it would deal with that class of case described as using legislation or the international tax system to one’s advantage (domestic legislation and/or double taxation treaties together with the norms of international taxation result in taxpayers being able to take advantage of opportunities. See http://www.sbs.ox.ac.uk/centres/tax/Documents/reports/TA_3_12_12.pdf). David remarks (about Google) that “Effecting transactions cross-border so as to report trading income outside the jurisdiction when you have an operational footprint within the jurisdiction manifestly introduces a tax risk factor and so my model returns a (true) positive in those circumstances.”
    13. But where is the evidence the structure was put in place, not from commercial/business considerations, but from acting in accordance with tax advice that created risks.? The EU Single Market creates complexities of its own in terms of free movement of capital, incorporation, etc. And any exporting/multi-national business creates an operational footprint in countries outside its own (although in a digital economy the nature of that footprint is much less tangible or certain). If the act of providing goods or services cross-border creates a tax risk, does that arise from following tax advice, or from the commercial decision? In other words, does the seeking and following of tax advice in such circumstances fall under para 33?
    14. In the PAC hearings Google maintained that it set up a HQ in the Irish Republic for a range of reasons (http://www.publications.parliament.uk/pa/cm201314/cmselect/cmpubacc/112/112.pdf, see for example Q7). But clearly there was tax advice on how they did this. As I see it the tax risk was not whether these structures were legally allowed, or inherently ‘risky’, but whether, in practice, Google operated in a compliant way. But this is the same for any operating practice where the tax outcome depends on the facts, e.g. whether a person has been taken on as an employee or is self-employed, who owns a particular asset for which allowances have been claimed, incurring deductible expenditure before a cut-off date, etc. I think David may argue that such risks only arise as risks because of the nature of the tax advice. Instead, I would regard this as comparable to the sort of “structured work-arounds he refers to as reducing tax risk. (To date I have not read anyone who doubts that is possible to set up as Google did, but a lot of criticism that it should ever be permissible.)
    15. I also think that if we accepted David’s model then it is actually very close to Andrew’s counter-proposal of “Define avoidance as the situation where the tax bill has been reduced as a consequence of an action taken by the taxpayer”, simply by accepting that taxpayer “action” can include getting and acting on tax advice. But if this very wide definition is accepted then does it not denature the term “avoidance”? Indeed, does it become so broad that it might be as useful to replace it with other words, like ‘mitigation’ or ‘planning’? If avoidance is to mean anything it must surely be less wide in its definition?
    16. Overall, my thanks to David (and Andrew) for getting us to re-examine the way we describe avoidance. Personally, I think the introduction of “tax advice risk space” is a useful additional way of examining transactions or structures when considering if there is or is not ‘avoidance’. But it is not a substitute for other forms of tests, such as ‘commerciality’, and I do not think it catches the use of anomalies in the international tax system.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s