Alex Helliwell, Andy Lymer and Lynne Oats look beyond the GAAR to loan relationship avoidance
What is the issue?
How does the loan relationship TAAR interact with the GAAR and to what extent does it go beyond the GAAR?
What does it mean to me?
It is important to consider the loan relationship TAAR alongside the GAAR. Although they are structured in similar ways, there are a number of differences between them that should make the loan relationship TAAR easier for HMRC to apply.
What can I take away?
Given HMRC’s focus on preventing tax avoidance it is important to ensure that the outcome of any tax planning using loan relationships is consistent with the purpose of the relevant legislation and that the arrangements are commercially driven.
The 2011 Aaronson Report expressed the hope that the introduction of a GAAR (General Anti-Abuse Rule) would simplify tax avoidance legislation (Graham Aaronson, ‘GAAR Study: A Study to Consider Whether a General Anti-Avoidance Rule Should Be Introduced into the UK Tax System’, 2011). However, for many companies in the UK at least, there appears to be only limited evidence to date that this is the case. In this article we review how the GAAR and the Targeted Anti-Avoidance Rule (TAAR) specific to loan relationships continue to interact at present.
At the same time as what would become Finance Act 2013 (which contained the GAAR) was making its way through Parliament, HMRC was consulting on the introduction of a TAAR to prevent companies seeking to benefit from loan relationship related tax advantages. Subsequently, the loan relationship targeted anti-avoidance rule (LR TAAR) was inserted in the loan relationship regime in TA 2009 ss 455B – 455D.
HMRC’s focus on preventing the use of the loan relationship regime as part of what they perceive as tax avoidance arrangements is unsurprising given the number of cases where the loan relationship rules have come before the courts – Stagecoach (Stagecoach Group PLC & Stagecoach Holdings Limited v HMRC  UKFTT 120 (TC),  SFTD 982), Greene King (Greene King PLC & Greene King Acquisitions Limited v HMRC  EWCA Civ 782,  STC 615) and GDF Suez Teesside (GDF Suez Teesside (Formerly Teesside Power Limited) v HMRC  EWCA Civ 2075) to name a but few. Although HMRC have been largely successful in establishing their interpretation of the legislation (the taxpayers in all three of these cases failed to achieve the desired tax reduction), these victories have mainly followed long and protracted legal battles which required the courts to opine on various tax and accounting issues.
However, given the GAAR was at the time in the process of being taken forward, the question of why the loan relationship TAAR was considered necessary arises. Luckily, HMRC answered this very question in their 2013 consultation document where they wrote: ‘The test of “abuse”, for the purposes of the GAAR, is a high threshold, and while the GAAR may apply to certain arrangements … it does not seek to encompass the full range of tax avoidance activity… The need for specific legislation to protect against manipulation of the regime and inappropriate relief in respect of corporate debt and derivatives therefore remains’ (HMRC, ‘Modernising the Taxation of Corporate Debt and Derivative Contracts’ (2013) para 14.3).
If you look at the two pieces of legislation side-by-side, you can see that their structures are very similar. However, there are a number of key differences between their content which have a significant impact on how the legislation may apply. Broadly speaking, the application of the GAAR and the LR TAAR can be broken down into five steps:
1. Is there an arrangement?
Both pieces of legislation contain the same broad definition of an ‘arrangement’ (see FA 2013 s 214, and CTA 2009 s 455C(2)). This is the standard definition of an arrangement that can be found in numerous other areas of the legislation.
2. Is there a tax advantage?
Whereas the term ‘tax advantage’ is broadly defined within the GAAR legislation at FA 2013 s 208, the definition of a ‘loan-related tax advantage’ in LR TAAR legislation (at CTA 2009 s445C(5)) is narrower, focusing on the debits and credits brought into account under the loan relationship regime. This difference is to be expected, given that by definition, the LR TAAR is a targeted anti-avoidance rule.
3. Is the arrangement a tax avoidance arrangement?
Both pieces of legislation contain a motive test whereby an arrangement is only in scope if one of the purposes of an arrangement is to obtain a tax advantage (see FA 2013 s 207(1), and CTA 2009 s 455C(4)). The wording used within the GAAR legislation has been slightly expanded to make it explicit that in applying the motive test, the user is required to have regard to all the circumstances of an arrangement. You can question the practical impact of this addition, as you would expect the courts to consider all circumstances as a matter of course.
4. Is the arrangement ‘abusive’ (for the purposes of the GAAR) or a ‘relevant avoidance arrangement’ (for the purpose of the LR TAAR)?
Given this is the critical test in both pieces of legislation (see FA 2013 s 207(2) and CTA 2009 s 455B(3)), it is important to consider the differences between the tests applied by the GAAR and the LR TAAR legislation in further detail. The most important difference is the shifting burden of proof from the taxpayer to HMRC in GAAR cases. That is to say, in GAAR cases, HMRC are required to prove that an arrangement ‘cannot reasonable be regarded as a reasonable course of action’ whereas, in LR TAAR cases, it is for the taxpayer to prove that the outcome of an arrangement is positively consistent with the purposes of the relevant legislation. In theory, this will make it easier for the HMRC to apply the LR TAAR to an arrangement, compared to the GAAR. However, whether this is indeed the case in practice depends on the approach taken by the courts in GAAR cases. The judgments in recent loan relationship cases show that, in general, judges view perceived loan relationship related tax avoidance schemes unsympathetically and therefore are likely take a broad view of the double reasonableness test in such cases. As such, in practice, the LR TAAR may not be significantly easier for HMRC to apply compared to the GAAR.
It is also interesting to note that, although the authors of the Aaronson Report considered that the inclusion of the double reasonableness test within the GAAR legislation was an important safeguard for taxpayers, other commentators, such as Watson, considered the lack of the double reasonableness test in the LR TAAR legislation a positive development as it avoided the ‘hopelessly vague “reasonable course of action”.’ (William Watson, ‘The Regime TAAR for Loan Relationships’ Tax Journal, 2 October 2015).
A final point to consider is whether the inclusion of the word ‘abuse’ in the GAAR has a significant impact on how the GAAR applies, compared to the LR TAAR where the term is omitted. Arguably, the use of the word ‘abuse’ merely adds to the terminological quagmire, for at the end of the day, the question the courts must consider is whether a piece of tax planning is acceptable in the context of the relevant legislation or not. That is not to say that the inclusion of the word ‘abuse’ is redundant for, as the Chartered Institute of Taxation pointed out: ‘framing the GAAR as an anti-abuse rule is correct as that should signal its target (itself an important feature)’ (Economic Affairs Committee, ‘The Draft Finance Bill’ HL 2012-13,139 Ev 78). Ultimately, the role of the term ‘abuse’ will only become clear once GAAR cases are heard by the courts (which is yet to occur, although decisions of the GAAR Advisory Panel are beginning to emerge and may provide some clarity).
5. Is an indicator present of abusiveness (for the GAAR) or non-exclusion (for the TAAR)?
Both the GAAR (at 207(4)) and the LR TAAR (at 455D(1)) contain a list of potential outcomes of an arrangement which, if present, would indicate that the GAAR or LR TAAR would apply (unless that outcome was the anticipated result of the relevant provision when enacted). As would be expected, the list in the GAAR legislation is relatively generic, whereas the list in the LR TAAR legislation is longer and tailored to outcomes that HMRC are attempting to prevent.
One unique feature of the GAAR is that it provides protection for taxpayers who undertake forms of tax planning which were standard practice when they were entered into and were accepted by HMRC.
There are other differences between the GAAR and the LR TAAR, such as the point that when determining whether the GAAR applies the courts are not constrained by Pepper v Hart  STC 898 (UKHL) and the additional procedural hurdles HMRC need to comply with to apply the GAAR, however these are beyond the scope of this article.
Only time will tell whether there will be a significant divergence between application of the GAAR and the LR TAAR to loan relationship avoidance schemes, and what impact this has on simplifying assessment of tax avoidance, but preventing taxpayers from entering such schemes and defeating those schemes is clearly a priority for HMRC. As such, taxpayers entering arrangements which could be seen as being such schemes should carefully assess the risks involved before doing so. In addition, if taxpayers are intending to rely on the motive test to prevent the GAAR and LR TAAR from applying, they must ensure arrangements that give rise to a beneficial tax outcome are commercially driven, with tax remaining a secondary consideration. Drawing on a memorable quote from the First Tier Tax Tribunal’s judgement in AH Field (a recent tax avoidance case involving loan relationships): ‘[o]bviously if the tax advantage is mere “icing on the cake” it will not constitute a main object,’ (Inland Revenue Commissioners v Trustees of the Sema Group Pension Scheme  EWHC 94 (Ch),  STC 276 ) however, tax avoidance will be the main purpose or one of the main purposes of the scheme if it is ‘more than mere icing’, for instance the scheme in AH Field producing ‘a preponderance of icing and very little cake.’ (AH Field (Holdings) Limited v HMRC  UKFTT 104 (TC) ).
This article is based on a MSc/CIOT Fellowship submission as part of the University of Birmingham and CIOT linked programme.