ACCA - The global body for professional accountants

Of the copious tax provisions in Hong Kong's Inland Revenue Ordinance (IRO), few can match section 61A, the so-called anti-avoidance provision, both in terms of the breadth of its potential reach and its amorphous construct.

Studying this technical article and answering the related questions can count towards your verifiable CPD if you are following the unit route to CPD and the content is relevant to your learning and development needs. One hour of learning equates to one unit of CPD. We'd suggest that you use this as a guide when allocating yourself CPD units.

This article was first published in the June 2009 edition of Accounting and Business magazine.

Since the enactment of section 61A in 1986, there had only been two pre-2007 anti-avoidance related Hong Kong court cases that provided scant guidance on the scope and application of section 61A. Since the end of 2007, however, five court cases have been handed down which address some of the judicial concerns associated with the application of the provision. 

Unlike section 61 (which requires an artificial or fictitious transaction), section 61A can be invoked if two tests are met with respect to any transaction: (i) a transaction has been entered into or effected which has, or would have had but for section 61A, the effect of conferring a tax benefit on a person (relevant person); and (ii) one or more of the persons who entered into or carried out the transaction did so for the sole or dominant purpose of enabling the relevant person, either alone or in conjunction with other persons, to obtain a tax benefit.

The sole or dominant purpose test was generally the focus of the earlier court cases with seven factors generally taken into consideration in determining whether this test was met. This article, however, primarily focuses on how recent case law adds to an understanding of the tax benefit test in light of its growing importance post-2007.

Specifically, the courts have increasingly employed an alternative hypothesis test, whereby challenged transactions are compared against an acceptable reference point to determine the extent of any tax benefit derived. The application of the alternative hypothesis test may lead to many unanswered questions, which may ultimately give rise to tax disputes. 

Tax benefit

It is generally believed that even though a transaction may be found to have met the sole or dominant purpose test, section 61A would not be applicable if there is no tax benefit conferred on any person. Although the term 'tax benefit' is defined as the 'avoidance or postponement of the liability to pay tax or the reduction in the amount thereof', it is not clear how a reference point may be established.

For example, if a taxpayer's liability after a certain transaction is X, how could one determine if X constitutes the avoidance, postponement or reduction of another amount (likely a higher amount) and, hence, whether a tax benefit has been derived? The establishment of this reference point, as discussed below, is the key to the tax benefit test.

Since section 61A was modeled after Australia's anti-avoidance legislation, the landmark Australian case FC of T v Spotless Services Ltd and Anor may be able to shed light on the tax benefit test.

The pivotal issue in that case was whether Spotless attempted to avoid tax by investing its surplus funds in a bank in the Cook Islands rather than in an Australian bank. On an after-tax basis, the return on the Cook Islands deposit exceeded what would have been derived on a comparable deposit placed with an Australian bank, although on a pretax basis, the return on the Cook Islands deposit was less favourable than an Australian deposit.

The court concluded that the dominant factor underlying the taxpayer's investment decision was the pursuit of a tax benefit.

By adopting an alternative hypothesis approach, the taxpayer was considered to have derived a tax benefit when its actual return was benchmarked against the return that it otherwise would have received on a comparable deposit placed with an Australian bank. It was not until more than 10 years after the Spotless case was decided in Australia did the test become prominent for purposes of addressing anti-avoidance issues in Hong Kong.

On 4 December 2007, the Hong Kong Court of Final Appeal (CFA) handed down its decisions in the following cases (all relating to section 61A) in which Spotless was cited as an authority:

  • Commissioner of Inland Revenue (CIR) v Tai Hing Cotton Mill (Development) Limited (FACV No. 2 of 2007)
  • CIR v HIT Finance Limited, and
  • CIR v Hongkong International Terminals Limited (FACV Nos 9 and 17 of 2007).
  • CIR v Tai Hing Cotton Mill (Development) Limited

In the case of Tai Hing Cotton Mill (Development) Limited, Tai Hing and its parent company (Parent) entered into a joint venture with an unrelated party in 1987 to embark on a real estate development project. As part of the overall development plan, Parent sold some of its land holdings to Tai Hing for consideration that was comprised of a fixed amount and a variable amount, which would be determined by reference to the profits derived from the joint venture.

Upon the completion of the development project, the total consideration of the land was determined to be HK$1.1bn, which was in excess of the market value at the time of the transaction by HK$300m. The gain on the disposal of the land to Parent was capital in nature and non-taxable, whereas the payment made by Tai Hing was a deductible stock cost.

The CIR sought to disallow the consideration in excess of the market price (ie. HK$300m) as a deductible stock cost to Tai Hing under section 61A and the case was appealed to the CFA.

In finding the existence of a sole or dominant purpose under section 61A, the CFA focused primarily on the design of the formula by which the consideration for the land was determined. Instead of a fixed sum, the formula called for a fixed sum plus a variable amount that the CFA claimed served to convert as much of the joint venture's profits into a deductible stock cost as possible. In addition, the CFA was of the view that, in economic terms, the financial implications for the group (ie. Parent and Tai Hing) would have been the same regardless of the amount Tai Hing paid to Parent.

Turning to the tax benefit test, the CFA and Tai Hing had different views as to the reference point to be used in calculating the tax benefit derived. Tai Hing argued that the reference point should have been the group's tax position had Tai Hing not acquired the land from Parent.

Following this line of logic, Tai Hing was attempting to establish a zero liability reference point because, absent the transaction, no tax liability would have accrued.

However, the CFA rejected this as an appropriate reference point. Instead, the court, without challenging or negating the form of the transaction, adopted the alternative hypothesis analysis under Spotless to establish the reference point as if the land had been sold at fair market value. As a result, Tai Hing was found to have derived a tax benefit equal to the tax liability on the excess of HK$1.1bn over HK$800m.

CIR v HIT Finance Limited and CIR v Hongkong International Terminals Limited

In 1994, Hutchison Whampoa Ltd (HW) conducted a group restructuring involving its container terminals and port businesses through the following series of transactions:

  • HIT Holding Limited (HITH) sold its terminals and port assets to Hongkong International Terminals Limited (HITL)
  • HIT Finance Limited (HITF) issued US$1.7bn floating rate notes (FRNs) on the Luxembourg stock exchange to raise funds that would be loaned to HITL for HITL's acquisition of HITH's assets
  • Of the US$1.7bn FRNs, one-third (US$0.5bn) was acquired by third parties and two-thirds (US$1.2bn) by Paribas Asia Limited (PAL), a third party underwriter

After HITH received from HITL the proceeds from the sale of its assets, it paid dividends to its parent, HIT Investments Ltd (HITI), which then loaned the funds to Strategic Investments International Ltd (SII), a group entity, which then acquired the FRNs held by PAL.

The interest income received by SII on its FRNs was sourced offshore and treated as non-taxable, whereas the corresponding interest expense paid by HITL to HITF (and that paid by HITF to SII) was treated as tax deductible. Thus, on a group basis, additional interest deductions and non-taxable interest income were generated simultaneously.

The CIR challenged the portion of HITL's interest deduction that was attributable to the interest ultimately paid to SII under both section 61 and section 61A; the matter was subsequently appealed to the CFA.

Although the CFA did not comment on the applicability of section 61, the court held that section 61A should apply. With respect to the sole or dominant purpose, HITF and HITL argued that a valid business purpose existed for the transaction that generated the disallowed interest (ie. the HIT group's financing of two-thirds of the FRNs).

The asserted business purpose was that without the HIT group subscribing to the two-thirds of the FRNs, the underwriter (PAL) would have had difficulty finding buyers for the FRNs due to the adverse market condition and it would have, therefore, sent 'an extremely adverse message to the market about the HIT group and possibly affected the subsequent HWL bond issue.'

The CFA rejected the taxpayer's argument on the grounds that it did not believe that two-thirds of the FRNs were meant to be 'on offer' to the public and that the 'adverse message' had already been given by the statement in the offer document that 'HW was obliged to give two-thirds of the proceeds back'. Accordingly, the court determined that 'the circular borrowing and repayment through Strategic (SII) were introduced for the sole or dominant purpose of avoiding tax.'

With respect to the tax benefit, the CFA followed Tai Hing in applying the 'alternative hypothesis' analysis first introduced in the case of Spotless. The reference point this time, however, was not the market value of the assets HITH sold to HITL (versus the actual purchase price paid), nor the market rate of interest (versus the actual interest charged on the FRNs).

Instead, the reference point was found to be the non-issuance of the two-thirds of the FRNs. By comparing the 'actual' interest paid and the interest that would otherwise have been paid had the two-thirds of FRNs not been issued, the CFA disallowed the portion of the interest expense claimed by HITL that was attributable to the interest ultimately paid to SII.

Analysis

It is clear from these cases that for section 61A to apply, both the sole or dominant purpose test and tax benefit test will have to be satisfied. However, these cases seem to have focused the analysis more on the latter rather than the former. This approach, arguably, makes sense since in section 61A, the sole or dominant purpose was mentioned only in the context of obtaining a tax benefit; in other words, the sole or dominant purpose test cannot, as a matter of definition, be satisfied if there is no tax benefit.

For purposes of the tax benefit test, it is also clear from these cases how the alternative hypothesis analysis is applied is central to the question of whether a tax benefit exists. Undoubtedly, the cornerstone of the alternative hypothesis analysis is the finding of an appropriate reference point against which the tax benefit, if any, can be measured. In the case of Spotless, the reference point adopted was a comparable bank deposit in Australia. Would it have been possible to establish a tax-free Australian bond or financial instrument as a reference point in the Spotless case? In addition, could the reference point in Spotless have been a deposit or financial instrument that was situated or issued in a jurisdiction outside Australia?

These questions point to a broader question: are there any bright line parameters based on which a reference point should be determined? Would such parameters include geography (since the taxpayer in the case of Spotless was an Australian resident, should the reference point also be a deposit placed with an Australian bank?); the nature of subject matter (since the subject matter in the Spotless case was a bank deposit in the Cook Islands, should the reference point also be a bank deposit?), etc.

Assuming these parameters were on the court's radar when Spotless was decided and would have been considered appropriate at that time, would the same parameters be considered appropriate today, more than a decade after the case of Spotless? Or should some other criteria (with or without regard to parameters) apply for purposes of determining the reference point?

In analysing the alternative hypothesis, Lord Hoffman in Tai Hing made reference to Australia's anti-avoidance legislation that 'the Commissioner may assess the taxpayer on the income which "might reasonably have been expected to be included" in his income if the scheme had not been carried out', and then reasoned that the CIR in Hong Kong should adopt a similar methodology (ie. the alternative hypothesis analysis). In this connection, Hoffman made an important point with respect to the application of the alternative hypothesis analysis:

'[T]he Commissioner… would not be entitled… to make an assessment on the hypothesis that the taxpayer had entered into an alternative transaction which attracted the highest rate of tax. That would not be a reasonable exercise of the power. But she may adopt the hypothesis which the evidence suggests was most likely to have been the transaction if the taxpayer had not been able to secure the tax benefit.'

In this light, it seems clear that the appropriateness of the reference point rests on its reasonableness and 'whether it is mostly likely to have been the alternative transaction' as opposed to whether it would generate the highest rate of tax.

Returning to Spotless, it may have been reasonable to assume that more than a decade ago, a plain vanilla bank deposit in an Australian bank was the most reasonable and likely reference point at the time given the limitation or lack of investment choices then and that the world was not as 'flat' as it is today. In today's technological age, large sums of money may be moved around the globe at one's fingertips to chase the highest rate of return possible from a wide array of financial instruments.

Therefore, if Spotless were to be decided today, could the reference point be any different given the potentially much larger range of possibilities? In other words, could there be a reference point that is not an Australian bank deposit but could prove to be more reasonable and likely and yet would not attract the highest rate of tax?

In the case of HITF and HITL, if the CFA would respect the transaction, whereby HITH sold its terminals and port assets to HITL, it would be reasonable to assume that the court would also recognise HITL's need to obtain financing for the acquisition. It is precisely this financing part of the transaction where the CFA believed a reference point should be established for the purpose of measuring any tax benefit that might have been derived by the taxpayer. The court ruled that the reference point should be the non-issuance of the two-thirds of the FRNs, thus implying that the HIT group should have deployed its own funds (ie. without issuing the two-thirds of the FRNs) to finance the acquisition.

For example, the financing part of the transaction could have been replaced by the following transaction: HITH could have distributed the sales proceeds from the disposal of its terminals and port assets as dividends to HITI, which could then have used the same to either capitalise HITL or loan, on an interest-free basis, to HITL to enable it to meet its financial obligation with respect to the acquisition. The question is, again, whether this reference point is considered reasonable and most likely under the circumstances.
In the area of mergers and acquisitions, it is not uncommon (in fact, more the rule than the exception) that a corporate acquirer would borrow the funds required to acquire a target from third party lenders. In many cases, the typical lender would be a consortium of multinational banks, which would charge interest at rates that are commensurate with the risk profile of the acquirer/transaction. It is intuitive that the corporate acquirer, even if it has sufficient funds of its own to acquire the target without resorting to borrowing, could still choose to borrow and for legitimate business reasons (for example, the acquirer may have other productive uses for its own funds). This 'back of the envelope' analysis is intended to establish the notion that the fact that a corporate acquirer borrows from a consortium of multinational banks (for this purpose, assume these are overseas banks) can be viewed as both reasonable and likely (and possibly most likely given the usual practice in this area) regardless of whether the acquirer already has the funds required to make the acquisition.

Applying this analysis to the case of HITF and HITL, one could argue that as far as trying to establish the reference point solely for the financing component of the transaction, borrowing from third party lenders can very well be considered as an appropriate reference point. Assuming such a reference point is adopted, the tax benefit would likely be measured as the excess of the interest paid by HITL (and HITF) with respect to the two-thirds of the FRNs over the 'hypothetical interest' that would have been paid on third party loans. The use of this reference point, relative to the non-issuance of the two-thirds of the FRNs, may be preferred due to the following:

It may be argued that the court in Tai Hing had more or less left the original transaction intact and had only introduced an alternative transaction with different terms or features, ie. the alternative consideration for the land was held to be the fair market value at the time of the transfer rather than the amount derived from the formula pre-agreed between the parties.

In other words, the alternative transaction was arrived at based on the form of the original transaction but with different terms or features. (To this end, the CFA in HITF and HITL and Tai Hing opined that it would not treat the original transaction as if it had not occurred in applying the alternative hypothesis analysis.) Therefore, to have the non-issuance of the two-thirds of the FRNs' as the reference point vis-à-vis the issuance of the two-thirds of the FRNs seems to have run contrary (at least perceptibly) to this principle.

Of course, one may counter that this principle should apply only to certain parts of the transaction (eg. the sale of assets by HITH to HITL) but not to all parts of the transaction. If this is the case, however, where should the line be drawn as to what parts of the transaction should be respected and what parts should not be respected? On the other hand, one may also argue that notwithstanding the foregoing, under section 61A(2)(a), the CIR is empowered to disregard a transaction if the sole or dominant purpose of the transaction is to obtain a tax benefit, thus concluding that there should be sufficient authority for the IRD or court to disregard all or parts of a transaction for purposes of section 61A.

This argument, if allowed, conceivably could lead one to a circular thought process since there has to be a tax benefit to begin with (as a pre-condition) before a transaction can be disregarded under section 61A(2)(a), but the discussion at hand is whether all or part of a transaction can be disregarded for purposes of determining if there is a tax benefit.

Using the third party loans as the reference point, in this instance, would help uphold the 'form' of the original transaction while preserving the integrity of the alternative hypothesis analysis in the sense that only different terms or features are highlighted and compared without negating the form of the transaction itself, which the CFA had seemingly tried to avoid. Moreover, if borrowing is entirely ruled out (whether it is in the form of the two-thirds of FRNs or third party borrowings), it would seem that the taxpayer is being compelled into accepting an alternative that would attract the highest rate of tax, which is clearly not what was intended by the CFA or the IRD.

Conclusion

As apparent from the cases discussed above, the alternative hypothesis analysis has become the cornerstone of the tax benefit test and thus the application of section 61A. However, establishing an appropriate reference point may prove to be elusive given the potential wide range of possibilities and the inherent difficulty in evaluating which possibility is reasonable and most likely under the facts and circumstances of each case, especially if there is a preference for the form of the transaction to be respected to the extent possible.

The significance of Ramsay (and how it may be applied) could be illustrated by the following hypothetical scenario in HITF and HITL: assuming the reference point of third party loans has been adopted as described above and it is found that the taxpayer did not derive any tax benefit and therefore section 61A would not apply.

But, if Ramsay could be applied to eliminate certain steps (or by way of statutory interpretation) of what would generally be regarded as a round trip transaction, the interest expense in question could still run the risk of being treated as non-deductible. Unfortunately, the CFA did not address this issue in the cases of Tai Hing or HITF and HITL.

In addition, a literal application of the alternative hypothesis analysis may potentially lead to unintended results. Suppose a well-connected Hong Kong taxpayer wants to diversify his real estate investments by investing in rental property in Dubai, which happens not to tax such income (if it is located in a tax-free zone). Would the investment diversification strategy cause the taxpayer to incur a section 61A problem under the alternative hypothesis analysis?

Finally, it is not entirely clear how section 61A and the Ramsay principle should be reconciled and applied concomitantly. Although the IRD is of the view that the Ramsay principle can be applied independent of sections 61 and 61A, the Court of Appeal in the case of Tai Hing made the following remark when the CIR sought to invoke Ramsay to interpret section 16 in an attempt to disallow the portion of the taxpayer's stock cost believed to have been inflated:

'The judge's (High Court) approach to the issues on section 16 which centred on what he considered was a purposive construction were in my view wrong… What the judge purported to do went far beyond giving a purposive construction to section 16 but was tantamount to applying Ramsay doctrine in circumstances where there was no justification to do so.'

What seems certain from the above analysis of section 61A is that there is a high level of uncertainty that may make the taxpayer's life (and those of the taxpayer's advisers) difficult when the taxpayer attempts to navigate often uncharted territories, where genuine business/investment decisions, normal tax planning arrangements and tax avoidance schemes may overlap.

Although more clarity and guidance would be welcomed, it remains unclear as to whether such guidance would even be possible given the nebulous nature of the subject matter.

Patrick Yip and Finsen Chan, Deloitte Touche Tohmatsu, Hong Kong

Last updated: 16 Jul 2014