Comments on the ATO s paper Intra-group finance guarantees and loans Application of Australia s transfer pricing and thin capitalisation rules

Transcription

1 Level 2 95 Pitt Street Sydney, NSW 2000 Telephone Facsimile tia@taxinstitute.com.au Website ABN th July 2008 Mr Marc Simpson Australian Taxation Office 140 Creek Street Brisbane QLD marc.simpson@ato.gov.au Dear Mr Simpson Comments on the ATO s paper Intra-group finance guarantees and loans Application of Australia s transfer pricing and thin capitalisation rules The Taxation Institute of Australia (Taxation Institute) welcomes the opportunity to comment on the Australian Taxation Office s (ATO) draft paper Intra-group finance guarantees and loans Application of Australia s transfer pricing and thin capitalisation rules (the Discussion paper), released on 3 June Although the paper does seek to provide guidance that addresses the transfer pricing implications of intra-group loans, credit guarantees and other financial transactions, and how such transactions might interact with thin capitalisation provisions, there appears to be a significant gulf between the position in the Discussion paper and both existing Australian tax law and financial market practice. This disparity is significant, given both the deviation of the ATO s proposed treatment of financial transactions from arm s-length principles and its potential for retrospective application. The Taxation Institute is particularly concerned with the lack of objective policy rationale that underpins many of the positions proposed in the Discussion paper. If this is driven by concerns about the integrity of some aspect of the tax rules, then this an issue more appropriately dealt with by legislation. Therefore it should be brought to the attention of Treasury rather being adopted through stretched legal interpretations. The following sets out our key concerns regarding the ATO s interpretation of Australian taxation law and practices in the financial markets, as well as some suggested alternative perspectives regarding the evaluation of financial transactions from an Australian transfer pricing perspective. Interaction between the thin capitalisation and transfer pricing provisions It is noted that the Discussion paper still incorporates the concept of the need for a taxpayer to estimate its arm s-length debt amount for transfer pricing purposes, even if its debt amount is within a safe harbour debt amount determined under Division 820. In our view, the Discussion paper and the ATO s approach is contrary to the legislative intent of Division 820, remaining impractical and at odds with financial market practice.

2 The Discussion paper asserts that, under Division 13, it may be necessary for a taxpayer to determine its notional arm s-length debt amount and subsequently raises three potential alternatives with regard to the portion of the taxpayer s capital structure that is beyond this notional arm s-length debt amount: that the additional debt amount should be priced at an interest rate consistent with the amount of debt that the taxpayer would have had if it had an arm s-length amount of debt; that the additional debt amount may be quasi equity, i.e., legal form debt at a 0% interest rate; that the additional debt cannot be priced based on available comparables and therefore the Commissioner has the right to determine a price on the additional debt under Section 136AD(4). Each of these alternatives clearly contradicts the legislative intent of Division 820, which provides Australian taxpayers with the option to establish their capital structure based upon a predetermined safe harbour debt amount. The ATO s potential application of Division 13 to capital structure creates, in substance, an additional arm s-length capital structure regime and, in many instances, obviates the impact of Division 820. The three potential scenarios highlighted in the Discussion paper could result in a significant increase in tax risk for many taxpayers, particularly in light of the previous lack of guidance on this issue in Australia or globally, as well as the lack of a clearly articulated policy rationale by the ATO. One key source of this increased tax uncertainty is the inherent subjectivity associated with estimating an arm s-length debt amount. Based on our members experiences in assisting Australian taxpayers with evaluating their capital structure using the thin capitalisation regime s arm s-length debt test through the ATO s recommended six-step approach, evaluating the arm slength nature of a taxpayer s capital structure is a difficult, time consuming and expensive exercise. Furthermore, such an analysis is inherently subjective; unlike the estimation of a safe harbour debt amount, a single bright line does not exist beyond which a taxpayer could not, on a stand-alone basis and at arm s-length, no longer fund itself with debt 1. Even if it were possible to definitively estimate an arm s-length capital structure for a taxpayer with a reasonable amount of effort, it is clear that a significant disparity exists between what constitutes an arm s-length capital structure in the ATO s view, relative to actual market practice. For example, the Discussion paper implies that taxpayers with junk bond or sub prime debt (completely unrelated concepts, incidentally) could not fund themselves on a stand-alone basis with debt. However, only a very small proportion of Australian taxpayers have an investment grade credit rating from a major ratings agency. Many Australian taxpayers are not rated investment grade, yet operate viable businesses that can attract both loans from commercial lending institutions and third-party equity investors). Indeed, most taxpayers are not rated at all, but would be rated non-investment grade if they were to obtain a rating. 2 The Discussion paper also appears to overlook a key aspect of junk and sub prime debt. While the paper implies that such circumstances imply that a borrower is undercapitalised, in substance, these are actual circumstances in the financial markets in which commercial lending institutions and investors extend credit to borrowers. In many developed economies, a significant market exists for relatively risky debt. These transactions are not, either in form or in substance, contributions of equity, from a legal or a transfer pricing perspective. The Discussion paper also appears to fail to acknowledge the existence of the market for subordinated credit, where lenders agree to assume increased credit risk (in terms of increased probability of default and/or loss given default), relative to senior lenders, in exchange for an enhanced return, even though the ATO 1 The Discussion paper simplistically outlines a choice between debt funding and equity funding (seemingly based on the legal form of the funding). The real world is more complex with choices to issue hybrid instruments (e.g. redeemable shares, convertible notes, perpetual debt, converting preference shares, mezzanine debt etc.) 2 It should be noted that the terms investment grade and non-investment grade date back to a time when pension funds in the United States and elsewhere faced restrictions on the credit quality of the bonds that they could purchase. Across many industries, noninvestment grade borrowers are commonplace (or even are predominant). 2

3 indicates in Taxation Ruling 92/11 that both credit quality and credit priority (i.e., quality of collateral) impact the pricing of debt transactions at arm s length. Passive association The paper appears to firmly embrace the need for Australian taxpayers to account for the impact of passive association when pricing their intra-group financial transactions, in particular, intra-group loans and credit guarantees. At paragraph 31, the Discussion paper notes that: the fact that the market may analyse the financial arrangements on a group basis may not be determinative since such an approach may defeat the purpose of the statutory rules. However, incidental benefits from association with a larger group where the market accords those benefits without an associated company actually providing an economic service that confers a benefit on the recipient would not usually be a basis for imposing a charge. The Discussion paper claims that independent investors may extend credit to the subsidiary of a multinational on more favourable terms than it would extend credit to an otherwise comparable independent enterprise, in light of the possibility that a parent may rescue a financially-troubled subsidiary, even if it has no legal obligation to do so. At paragraph 75, the Discussion paper presents the argument that because such a perceived increase in credit quality:...flows from the market and not something the parent company has done or provided, the benefit to the subsidiary should be regarded as incidental and attributable solely to its being part of a larger concern and would not warrant a service charge by the parent on the subsidiary. This is stated even though paragraph 74 notes that such an approach arguably is: inconsistent with the arm s length principle embodied in the Associated Enterprises Article of Australia s double tax agreements and Division 13, which are based on the outcome that would be achieved by independent parties dealing wholly independently with each other. The Taxation Institute agrees with the latter perspective. An examination of the context of credit agency guidelines are written from the perspective of a credit analyst lending to a subsidiary in the context of the lender s relationship to the borrower and the global group as well as the lender s commercial context. Those guidelines are not required to take into account the functionally separate enterprise concept that underpins transfer pricing and the arm s-length principle 3. This is a fundamental flaw in the Discussion paper. In addition, we note that the OECD commentary noted in paragraph 75 of the Discussion paper, along with earlier OECD commentary 4, is confined to circumstances where the credit improvement is due solely to affiliation and not as a consequence of any guarantee provided. However, in circumstances where an explicit guarantee is provided, a chargeable intra-group service arises. The Discussion paper does little to reconcile the inherent conflict between the arm s-length principle and the application of the passive association argument to intercompany transactions. If the ATO decides to modify its interpretation of Taxation Ruling 92/11 or issue revised guidance, a move towards accounting for passive association will have a significant impact on the manner which many Australian taxpayers price their intra-group financial transactions. Hence, the ATO should issue definitive guidance on this issue as soon as possible and then only seek to apply it prospectively. In addition, given its importance, it is an area that also needs to be addressed at a multilateral level (e.g., at the OECD) in order to mitigate the risk of double taxation. 3 See the OECD Transfer Pricing Guidelines at paragraph A subsidiary should not be required to pay for benefits derived simply from being part of a concern (i.e. benefits which cannot be related to the specific activities of the parent) for instance the higher credit-standing apart from guarantee agreements resulting from being part of a concern with a good reputation. (Emphasis added). Transfer pricing and Multinational Enterprises Three Taxation Issues (OECD 1984), paragraph 40, page 82. 3

4 Leaving aside the issue of whether it is appropriate to account for passive association when establishing arm s-length terms for intra-group financial transactions, the ATO s potential adoption of passive association as a comparability factor when establishing or evaluating the arm s-length nature of a related-party financial transaction raises a variety of additional key issues. Adjustments for passive association We understand that the ATO has referred to guidelines prepared by Standard & Poor s (S&P) for evaluating the impact of group affiliation on a borrower s credit quality. In adjusting for the impact of group affiliation, it (like the ATO) uses a parental credit rating as a starting point. However, another major ratings agency, Moody s Investors Service (Moody s) generally uses a pure stand-alone credit quality as a starting point when evaluating group affiliation and may provide limited credit quality uplift for any group affiliation, citing the fact that many major multinationals have chosen to not rescue the unguaranteed debt of their overseas subsidiaries when they entered a period of financial difficulty. Under the working assumption that it is appropriate to adjust for passive association (and our view is that it is not appropriate to make such an adjustment, at least under existing Australian transfer pricing legislation and guidance), taxpayers face conflicting approaches as to how they should actually perform such an adjustment. This exercise is further complicated by the paucity of external, market-based comparables available to adjust for the impact of passive association, as well as the fact that most taxpayers will not have access to reliable internal comparables. This paucity of internal and external information creates a real and significant risk that many taxpayers will face a Sisyphean task of trying to establish to the satisfaction of the ATO that the impact of passive association in a particular case is non-existent or small rather than substantial. However, due to the same paucity of information, the ATO is also unlikely to be able to establish reliably in the same case that the impact of passive association is substantial rather than nonexistent or small. This is an unsatisfactory situation and the Discussion paper provides no practical guidance to taxpayers on how the ATO proposes to address this issue in a way that is fair and reasonable. As with using a notional arm s-length capital structure to price debt, the ATO s proposed adoption of a passive association approach to pricing intra-group debt and credit guarantees is both questionable in terms of its consistency with existing law and guidance and its practicality of implementation. If, notwithstanding these concerns, the ATO persists with its views in relation to adjustments for passive association, then in the Tax Institute s view, it should work with industry and professional associations to develop reasonable and pragmatic solutions to address the concerns raised rather than leave such matters to be addressed in the context of audits. Interaction between passive association and thin capitalisation If one were to embrace the concept of passive association, a logical extension of a passive association argument is that the amount of equity maintained by an Australian subsidiary of a major multinational firm would be less relevant than it would be for an otherwise comparable standalone entity. Indeed, using the passive association argument, if investors perceived that a parent would likely rescue an Australian subsidiary, even without any legal obligation to do so, it is difficult to see why the Australian taxpayer would require any equity at all, as investors would be extending credit to the local subsidiary based on its parent s credit quality and not the credit quality of the local borrower (again, using the flawed passive association logic). Of course, Division 820 requires that taxpayers hypothecate an Australian business that does not have any guarantee, security or other form of credit support from overseas affiliates. Hence, when calculating an arm s-length debt amount (or presumably a hypothetical arm s-length amount of debt, even when within a safe harbour debt amount), it is necessary to make adjustments that may increase the taxpayer s hypothetical cost of funds on a stand-alone basis. However, according to the logic of the Discussion paper, when actually pricing intra-group debt, a taxpayer must account for the impact of passive association (which, ceteris paribus, may lead to a 4

5 downward adjustment in the interest rate charged to an Australian business). This approach appears to have a revenue bias, is contradictory and is of questionable policy merit. Credit guarantees The Discussion paper addresses at length factors that, in the ATO s view, taxpayers should consider when evaluating whether a credit guarantee should be charged and factors that may impact the pricing of a credit guarantee. However, the Taxation Institute recommends that the Discussion needs to consider more adequately, preferably illustrated by example, the similarities and differences between loans and credit guarantees when addressing the transfer pricing and thin capitalisation issues raised by loans and guarantees. Interaction between transfer pricing and s8-1 The Discussion paper asserts that where a loan is provided to a company that is unable to borrow on a stand-alone basis, a question may arise as to the purposes of the loan. The Discussion paper draws an inference that interest expense may not be deductible where the debt funding performs the role of equity and that such a conclusion could be drawn after reviewing the profit earning structure of the business, including shareholder interests. The Discussion paper cites the Sun Newspapers case as being the relevant precedent for this analysis. The Discussion paper then proceeds to discuss the St George Bank case, which is credited with creating an argument that where a subsidiary could not borrow on a stand-alone basis, the debt funding may perform the role of equity. The conclusion reached is that it may then be open to argue that the interest expense should be treated as being of a capital nature. However, in our view, it is arguable that the St George Bank case does not establish that principle. Rather, the St George decision merely applies the Sun Newspapers test that the issue of deductibility is found in the character of the advantage sought from the borrower s perspective, not from the perspective of the parents relationship or concern. In the context of loans, this analysis would be conducted from the borrowers perspective only, and it is just as likely that a court would find that the interest is deductible, assuming it were derived in the ordinary course of business. The discussion concerning the St George Bank case is not relevant to debt interests issued since 30 June This is because s mandates that returns on a debt interest are not prevented from being a general deduction under s 8-1 merely because the return secures a permanent or enduring benefit for the issuer. Reliance on s 136AD(4) of Division 13 The Discussion paper raises the prospect of the ATO using the Commissioner s discretion under s 136AD(4) to determine an arm s length interest rate or an arm s length guarantee fee. The key factor for the application of s 136AD(4) is the lack of information available to the Commissioner to ascertain the arm s length consideration identified in s 136AD(3). The Discussion paper contemplates an absence of comparables to test the terms and pricing of an arrangement (and more broadly, its commerciality). However, the absence of comparables is not by itself indicative of an arrangement that is non-commercial. 5 What the ATO would require to establish a defensible 136AD(4) decision is direct information on how arm s-length parties actually contract with each other in practice. It is not enough to establish a theoretical model (as the Discussion paper has done) without independent evidence of how arm s-length parties contract with each other at each and every step in the analysis. 5 See the OECD Guidelines at paragraph

6 In applying s 136AD(4), the ATO should acknowledge that, depending on how the High Court decides the appeal in WR Carpenter Holdings v C of T [2007] FCAFC 103, it may be required to make full disclosure of all the information it has used in making a Division 13 determination. This has not been its historic practice. The Discussion paper, at paragraph 180, also raises the prospect of using s 136AD(4) to restructure transactions to conform to how independent parties would transact. Again, this is inconsistent with the OECD Guidelines (paragraph 1.36), and relies on clear insight into arm s length behaviour. Given the lack of information contemplated under s 136AD(4) decisions, the exercise of the discretion would be wholly arbitrary. There is therefore the real prospect of double tax arising, particularly where other countries have limited experience in these kind of transfer pricing issues. The ATO may assume that other countries will be convinced by clear theoretical arguments. They may not be. In such subjective areas, the Mutual Agreement Procedure process inevitably leads to taxation by negotiation and does not provide certainty or resolution in real time. We suggest that the ATO needs to build international consensus with key treaty partners and the OECD before implementing the approach. Requirement for outcomes to make commercial sense The Discussion paper says that the ATO may need to apply the transactional net margin method ( TNMM ) using data as to operating profits (net of financing expenses) for comparable uncontrolled enterprises to determine an arm s length outcome for the borrower. It goes on to say that the amount derived from the use of the TNMM would then need to be recalculated as a varied interest rate on the relevant debt funding as appropriate. No other guidance is provided. We also note that the paragraphs from TR 97/20 referred to by the ATO pre-date the introduction of the thin capitalisation rules in Division 820 and further that no reference has been made to the views expressed by Justice Downes in Roche Products Pty Limited v C of T [2008] AATA 261 in relation to the application of TNMM in the context of Division 13. Finally, the Discussion paper fails to acknowledge that in many situations the presence of related party loans or a finance guarantee from a related party will form just a part of a number of crossborder related party dealings (eg trading stock, services, licence fees). Application of TNMM in the way suggested by the ATO would therefore be problematic. Additional guidance should therefore be provided by the ATO as to how such an approach would apply in practice. Potential for retrospective application At paragraph 31, the Discussion paper notes that: Nothing in this discussion paper is intended to be contrary to [TR 92/11] or indicate that the Tax Office may be considering changing its views on determining an arm s length interest rate from those stated in TR 92/11. This statement appears to us to be disingenuous. TD 2007/D20 and the Discussion paper represent the first occasion on which the ATO has expressed such views. Taxpayers and ATO auditors alike have been determining the arm s length interest rate on related party loans based on a company s actual capital structure and in light of ATO guidance contained in TR 92/11 for more than 15 years ago. The proposed retrospective application of the approach outlined in the Discussion paper means many companies could face transfer pricing adjustments irrespective of the fact that: the capital structure of their Australia subsidiaries may have been determined on the basis of satisfying the safe harbours within the thin capitalisation rules; and related party debt may have been priced in good faith having regard to standard transfer pricing approaches and ATO guidance existing at that time (eg TR 92/11 and TR 2002/2). 6

7 Retrospective application of the approach outlined in the Discussion paper would be harsh and unreasonable, especially given there is no statute of limitation for Division 13 adjustments. Taxpayers are legally entitled to rely on the ATO view stated in TR 92/11 in respect of related party loans already entered into until the view stated in that ruling is withdrawn or changed 6. Accordingly, in the Taxation Institute s view, the approach outlined in the Discussion paper should only be applied on a prospective basis. Recommendations Given the threshold issues raised around the Discussion paper, and the tax policy objectives of the transfer pricing and thin capitalisation rules, the ATO needs to develop a different approach to price intra-group loans and guarantees from that espoused in its draft paper. Such an approach must evaluate pricing of the actual transactions that are the focus of Division 13 and Thin Capitalisation rules, as well provide a basis for consensus with trading partners and other revenue authorities. Consensus clearly also needs to be established with the OECD. Our preferred approach is as follows: 1. Division 13 is applied to the actual capital structure of the taxpayer. Then determine the arm s length pricing of interest on the loan and any explicit guarantees. 2. Division 820 is applied to limit debt deductions (interest and guarantee fees) in the event actual debt levels exceeds the maximum allowable debt. 3. Explicit guarantees should be evaluated on the basis that they allow the guaranteed entity to reduce the cost of its debt. 4. The ATO s approach to have application only from the date a final public ruling has been issued, and must not have retrospective application. The Taxation Institute would be pleased to address the NTLG Transfer Pricing Sub-Committee directly when it meets to consider the submissions in relation to the draft paper. In the meantime, if you would like to discuss any of the issues raised in our submission or require further assistance or information, please contact the Taxation Institute s President, Sue Williamson, on or the Taxation Institute s Senior Tax Counsel, Dr Michael Dirkis, on Yours faithfully Sue Williamson President 6 With respect to inconsistency see s357-75(1) of Schedule 1 to the Taxation Administration Act 1953 and with respect of a change in the Commissioner s general administrative practice see s358-10(2) of Schedule 1 to the Taxation Administration Act