I wrote the following article for Taxsutra in India a week or so ago after it was announced India was to have a comprehensive general anti-avoidance rule. It pretty much applies here too though so I thought I'd share it:
The news that the Indian government is planning to introduce a general anti-avoidance provision and measures to tackle tax haven abuse are both extraordinarily welcome. Many budgets, the world over, are forgotten a day or so after they have been delivered. For India this one is likely to go down in the annals of tax history.
The history of general anti-avoidance rules is chequered. Some, and here I think mainly of Australia and South Africa, have worked. Less noticed, but perhaps most effectively of all, has been the curious dependence of some tax havens, such as Jersey, on such provisions to ensure they offer what looks like a simple tax code that they can, none the less, impose upon their resident populations, many of whom are taxable on their world wide incomes but are aware of multitudinous ways to hide that income from view! In other cases, and Canada is the most often referred to, experience of general anti-avoidance provisions has been poor. It is therefore good to note that India has noted some of such issues that have caused GAARs to fail whilst clearly having a focus on what it needs from a GAAR.
Let's be clear what a general anti-avoidance rule should do. It should say, as clearly as possible, that if a step is added into a transaction or series of transactions without apparent primary motive bar the avoidance of tax then that step should be ignored for the purposes of computing the tax liability arising on the transaction in question. The behaviour that is being challenged is the desire to subvert the apparently legal, but none the less abusive, actions of those in the tax profession who make it their business to scour legislation and regulation, nationally and internationally, and both within the tax code and in other connected areas, such as accounting regulation, to exploit legal loopholes on behalf of their clients knowing that the combined construction that they put upon the law and regulation was never the intent of the domestic parliament whose tax revenue is threatened as a result. This can, of course, in very many ways be said to be the legislative embodiment of the Ramsey principle, much discussed in tax litigation in common law countries around the world, and not least in India of late.
Those who oppose general anti-avoidance rules, and who opposed Ramsey, argue that the form of contracts must be honoured in tax law. It is a curious argument to an accountant like me. I am well aware that the form of a multinational corporation, or indeed the form that the affairs of many a wealthy family takes, is far removed from its substance. It is this substance versus form argument that is at the core of the GAAR debate. Those, like me, who argue for GAARs suggest that it is the substance of the issue that matters. If, then, it is possible for a multinational corporation to so structure its affairs in almost any contractual style that it so wishes to minimise its tax bill but the substance is that any such structuring is simply an artifice because the whole structure is under common control and the outcome of the supposed contractual arrangements is pre-ordained to achieve a particular result (and this is, almost invariably true) then the contractual structure is but an artifice too because the arrangements are not between independent parties but are instead designed with but one aim. The fact that some of the parties may appear independent, whether as trustees, orphan entities or even as genuine third parties who have however contracted to lend their name to the pre-agreed purpose in exchange for a fee does not change this fact: the contractual arrangement to act in pre-agreed fashion is what is critical. The substance if such arrangements is ultimately not commercial; it is tax driven. It is proving this substance that is key to any assessment of a case to apply a GAAR to: if it cannot be shown that there was a pre-agreed agreement as to the nature of the transaction that was designed to achieve the tax goal, or unless it can be shown in the case of repetitive arrangements (and here I have derivative arrangements in mind, which are now for many multinational companies the favoured mechanism of choice for transferring profits offshore) that such is the pattern of events that there was no commercial substance to the arrangements because a predictability of outcome appears sufficiently prevalent in the events, then it is unlikely that a GAAR accusation will succeed.
All this takes effort. That is why the suggestion that there be a de minimis anticipated tax loss before the GAAR can be applied seems sensible in India. That also prevents the GAAR being used to attack routine transactions which should be addressed by legislation in their right. Such a situation might, for example, relate to the simple decision to incorporate. If this gives a tax advantage (and candidly, that is often the intent, limited liability frequently being a secondary consideration) it would be perverse to apply a GAAR for that reason alone. The law encourages incorporation and therefore the GAAR should not apply. It is for the same reason that the recommended use of a panel who must be convinced of the relevance of applying the GAAR is also appropriate. Current proposals in the UK also have this requirement, although when (as I was) I was consulted on this issue by Graham Aaranson QC on this matter during the preparation of his recommendations I suggested the three person panel should comprise a tax official, an informed lay member and be chaired by a person with judicial status. He did not include that suggestion, but I will repeat it none the less for India as I think the role of a person who can clearly be seen to be independent is important in this process.
That though does bring me to the other issue that I think vital if any GAAR is to succeed. This is the essential requirement that the GAAR include very clear instruction to judges on the basis of interpretation they must use when applying the legislation. It is precisely because Canadian judges appear to have ignored the fact that applying a GAAR requires purposeful interpretation of the statute under consideration that basically undermined the GAAR in that country. In that case it is essential that the GAAR gives such direction to judges that they will always seek to endeavour to interpret the spirit of the law or laws that have been transgressed when making their judgements and it may also be appropriate, in the absence if purposefully written legislation, to specify what they make not consideration when determining what they think that purpose might be. This could include ministerial statements at the time the legislation was introduced, press releases, consultations and of course parliamentary proceedings. Whatever is decided, without this the risk that a GAAR simply becomes another rule of judicial interpretation and is neutered as a consequence in the way Lord Hoffman neutered Ramsey in the Westmoreland case is all too likely.
There is no doubt in my mind that a comprehensive GAAR that includes sensible protections, reasonable but not excessively onerous burdens of obligation on a tax authority to prove its case and protection against it being used in trifling matters or to tackle issues where new primary legislation is really required is an essential part of any modern tax system where a government wishes to prevent abuse of its territorial domain, its right to collect tax from its residents and where it wishes to stop the persistent abuse of the form of legislation (a form that is, ultimately inevitable) by those who will construct transactions whose substance is to simply undermine the right of the state to tax in a modern democracy. India is taking an important step forward in this respect, and all who believe in that right of India to tax in the name of its democracy and people should also welcome this move.