Read the 'IOD Response to Government, ConDoc' report
1. The Government is right to propose a narrowly-targeted GAAR. That will allow most taxpayers to have certainty. It will also make it likely that the GAAR will be effective in litigation. A broadly-drafted GAAR would be ambiguous and the courts would probably interpret it in taxpayers’ favour.
2. In response to question 1, the list of taxes seems reasonable.
3. One issue with national insurance is that a single base of earnings may be subject to both primary and secondary contributions, so if an employer were to act to reduce secondary contributions, and the scheme to do so were to be defeated by the GAAR, the employee might suddenly be faced with an unexpected bill for primary contributions, even though the employee had known nothing of the scheme.
4. In relation to inheritance tax, the gap between planning being put in place and a tax charge arising was what led to scope for claims that the pre-owned assets legislation was retrospective. One can understand the Government’s not wanting to get into that sort of argument again. One part of the solution might be an extended transitional regime, under which arrangements put in place before the legislation of the GAAR and not amended since were not subject to the GAAR, although this would assume that it was possible to distinguish reliably between amendments to existing arrangements, and unrelated new arrangements that happened to affect the same estate. Another part of the solution might be to follow the path taken with pre-owned assets: allow people to unwind tax planning arrangements that would be caught, rather than litigate the applicability of a GAAR (or rather than suffer the consequences of following the precedent of a court decision that applied a GAAR to tax planning undertaken by a different taxpayer).
5. On question 3, we are concerned about “reasonable to conclude”. It might be reasonable to conclude that the obtaining of a tax advantage was a main purpose, and also reasonable to conclude that it was not a main purpose, if “reasonable to conclude” means “a reasonable person could conclude”. Should one have “more reasonable to conclude that the obtaining of a tax advantage was the main purpose, or one of the main purposes, of the arrangements, than that it was not”. Going even further, one might have “reasonable to conclude that the obtaining of a tax advantage was the main purpose, or one of the main purposes, of the arrangements, and unreasonable to conclude that it was not”.
6. On question 4, we are not sure whether the concept of a reasonable course of action does any useful work. One sign that it does not is that it is not defined. Things to look at are listed in clauses 2(2) and 2(3), but that does not amount to a definition. Clause 2(4), which is where we find the substantive triggers for application of the GAAR, lists indicators of abuse, not of unreasonableness. Clause 2(5) makes it clear that there may be other triggers, but gives no hint as to what they might be. One can only assume that they are either things of the same general nature as those listed in clause 2(4), or things that make arrangements look fishy in the context of the items listed in clauses 2(2) and 2(3). (It may be that this is all or part of the work that the concept of a reasonable course of action does – providing a route by which those lists of items can give content to clause 2(5).)
7. We therefore suggest that the desired result might better be achieved as follows. (This is only a first attempt at a re-draft, and is no doubt imperfect in many ways.)
2(2) Tax arrangements are abusive if they are arrangements that cannot reasonably be regarded as not being exploitative.
2(3) Arrangements are exploitative if they exploit provisions of the tax law to obtain a tax advantage, in a way that it is plain Parliament would have thwarted if it had considered the point when legislating.
2(4) Each of the following is an indication that tax arrangements might be exploitative.
[Insert what are currently 2(4)(a) to (d)]
2(5) Subsection (4) is not to be read as limiting in any way the cases in which tax arrangements may be regarded as exploitative.
2(6) In deciding whether arrangements can or cannot reasonably be regarded as not being exploitative, the following must be considered:
[Insert what are currently 2(2) (a), (b) (c) and, by way of expansion on (a), 2(3) (a), (b) and (c).]
8. Clause 4(3) and (4) cover consequential adjustments. We do have a concern about clause 4(4)(b), which allows consequential adjustments to affect taxpayers who are not party to arrangements.
9. Our concern is that someone who was unaware of tax planning arrangements, but who may have benefited accidentally by virtue of having entered into transactions with people who had entered into the arrangements, might suffer an unexpected tax bill if the GAAR subsequently nullified those arrangements. In order to avoid that risk, people who were, for example, buying or selling businesses, or providing or receiving finance, might seek indemnities from other parties to cover the risk, or assurances that no tax planning was involved. That could make commercial deals harder to strike. We suggest that parties who were unaware that high-risk tax planning was involved should be safe from consequential adjustments that would lead to their having to pay additional tax.
10. Consequential adjustments may sometimes involve refunding tax to a party that was not involved in arrangements. It will be important for HMRC not to be allowed to refuse such adjustments on the basis that they would constitute unjust enrichment in the VAT sense.
11. We are not sure that the drafting of clause 6(2) is ideal. Is the intention to apply the entire tax code excluding the GAAR, to look at the result, and then to see whether the GAAR should over-ride that result? Clause 6(2) might not achieve precisely that, because it might lead to the GAAR’s being invoked when applying the rest of the tax code. That is, it might allow the GAAR to interfere with the application of priority rules in relation to the rest of the code. We suggest that clause 6(2) might be re-worded as “But no priority rule shall operate to exclude the application of the GAAR, or to subordinate its application to the application of other rules”.
12. On question 7, a range of measures might help to address the transitional problem. One possibility would be to distinguish between arrangements in which substantial commitments had been made before 1 April 2013 (they might be excluded from the GAAR), and those in which only something minor and easily reversible had been done before 1 April 2013 (they might be subject to the GAAR). Another possibility would be to incorporate that distinction, not in the commencement clause but in the rules that would determine what adjustments would be just and reasonable. (Someone who could easily have reversed arrangements after 1 April 2013, but chose to press ahead, could expect a full adjustment. Someone who was already heavily committed by that date might be let off all or most of the adjustment, except to the extent that he had not taken an opportunity to scale down the arrangements and minimise the tax at risk.) A third possibility would be to offer taxpayers ways to unwind their schemes at no tax cost (and also no tax advantage).
13. Clause 5(3)(a) allows guidance, statements and other material to be considered by the court or tribunal. We agree with this general approach, but Ministers are apt to describe schemes as abusive, or to use similar language, in the context of political debate. Should such mere expressions of opinion be excluded from consideration by the legislation, or can we rely on the courts and tribunals to take no notice of them? They should certainly not be taken into account, because taking them into account would make the application of the GAAR partly a function of the personal and political opinions of Ministers, and the law should not operate like that.
14. Paragraph 5.10 proposes the application of the usual penalties. Given that the GAAR will be a novelty, we believe that if a taxpayer self-assesses that it does not apply, and it is then found that it does apply, the threshold applying a penalty for the return’s being false in this respect should be set very high. Most mistaken failures to self-apply the GAAR should be treated as innocent errors, and should attract no penalty.
15. Chapter 6 discusses the advisory panel. We appreciate that this will not be a quasi-court. Nonetheless, it will be important for it to treat HMRC and the taxpayer in the same way. Both HMRC and the taxpayer should communicate directly with the panel. If submissions from one of them are shown to the other, then submissions from the second one must also be shown to the first. Both parties should be limited to written communications with the panel, or both should be invited to any meeting with the panel. And any communication from the panel should be sent to both parties at the same time.
16. On paragraph 6.20, the panel should be free to publish anonymised opinions if it so wishes. HMRC should be free to advise whether a given opinion had been sufficiently anonymised, but the panel should make the final decision. Given that the panel would not be a fixed group of individuals, such decisions could be made at intervals, by all the people who had been members of the panel since the last time a report was published, or they could be made in relation to each opinion, by the panel members who had given it.
17. On paragraphs 7.6 and 7.8, the advisory panel (perhaps deemed to consist of all people who have participated in it over the preceding year) must have the opportunity to amend the guidance (with amendments to be agreed with HMRC), as well as to review and approve it. Otherwise, the guidance will become what HMRC would like it to be, which is not necessarily what it should be.
18. Our final comment is a very general one. A leading concern of commentators is that a GAAR will generate uncertainty. It is possible to test a GAAR on this point – not quite scientifically, but as close as one can get in such matters – and we consider that the test should be carried out. It would proceed as follows.
(a) Identify a dozen schemes, some of which HMRC thinks would be caught by the proposed GAAR, and some of which HMRC thinks would not be caught. They should all be difficult cases, in which it is not obvious whether they would be caught.
(b) Ask a dozen tax lawyers to participate in the test. These should be people who have had no involvement at all with the process so far, either in conversations with Graham Aaronson or his team, or in discussions with HMRC or the Treasury.
(c) Give each of the lawyers the draft GAAR, the details of the dozen schemes, and no other guidance whatsoever. Ask them to say, individually, which schemes they think would be caught by the GAAR (assuming that they were not caught by any more specific legislation – they should not be asked to check that point), and which would not. They should all work separately, with no conferring whatsoever. They should be required to reach decisions on all schemes, even if they want to say “I don’t know”.
(d) See whether they all give much the same answers, or widely different answers. If the former, uncertainty should be low. If the latter, it will be high. The answers to be considered for similarity and difference should be the conclusions (“schemes 1, 3, 4, 7 and 12 would be caught, but not the others”), and not the reasoning, because only the conclusions could be compared objectively.
(e) Advice should be taken from statisticians in advance of stage (d) on how to measure similarity and difference in answers. The method of measurement should be fixed in every detail before the answers come in.
(f) The schemes, the method and the data on answers should all be published in full, to facilitate public debate.