Comments received on public discussion draft: The Taxation of Offshore Indirect Transfers - A Toolkit Draft Version 2 October 2018 i Table of Contents BDI -The German Business Representation .................................................................................................. 1 BEPS Monitoring Group ................................................................................................................................ 5 BIAC Tax Committee ................................................................................................................................... 11 Ernest and Martin ....................................................................................................................................... 20 Government of India ................................................................................................................................... 22 Hardeep Singh Chawla ................................................................................................................................ 27 International Chamber of Commerce (ICC) ................................................................................................ 31 International Tax and Investment Center (ITIC) ......................................................................................... 35 PwC ............................................................................................................................................................. 39 Repsol.......................................................................................................................................................... 48 Sergio Guida ................................................................................................................................................ 53 Transfer Pricing Economists for Development (TPED)................................................................................ 54 ANNEX A: Redline comments from ITIC ..................................................................................................... 56 ii COMMENTS ON IMF/OECD/UN/WBG second discus- sion draft: The Taxation of Offshore Indirect Transfers: A Toolkit September 24th 2018 To Platform for Collaboration on Tax General Comments BDI welcomes the revision of the previous draft from August 1st 2017 and the at- tempt to provide further clarification and legal security. We highly appreciate the fact that the draft views the possibility of taxation offshore indirect transfers (OIT) in the country of the underlying asset as an option and that “countries may affirmatively choose – as some have – not to tax such gains on indirect transfers even where they could do so.” (p. 22) To avoid any misunderstanding we suggest that the op- tionality should also be included in the executive summary and not just in the text. Specific comments • Definition and characteristics of location specific rents: Under C. The Allocation of Taxing Rights on OITs: Equity and Efficiency Consid- erations the draft develops the concept of location specific rents (LSR) as an ex- tension of immovable property. (p. 8 and p.20) Beside typical immovable prop- erty like land, buildings, and structures as well as rights related to such property this definition shall also cover licenses and other rights issued by the public. Ad- ditionally the aspect of an economic rent as being “in excess of the minimum ‘normal’ return that the investor requires” (p. 20) is part of the concept. 1 The Taxation of Offshore Indirect Transfers – A Toolkit The definition of an economic rent changed to some extent over time. An ap- proach frequently found in economic theory is that an “economic rent is any pay- ment to an owner or factor of production in excess of the costs needed to bring that factor into production”. 1 In other words this equals the inframarginal part of the return on an investment. Other more traditional definitions like unearned in- come derived from (inherited) land seem to be inappropriate since the subject of the working paper is an investment of a multinational enterprise (MNE). This also makes other definitions inadequate like the lack of opportunity costs be- cause an MNE always has other investment opportunities. However, the marginal rate of return by no means equals the minimum normal rate of return an investor would require. Even a considerably higher rate of re- turn cannot in general be regarded as excessively high as entrepreneurial rents show. 2 In consequence it’s hardly possible to define a normal rate of return. Thus an economic rent cannot generally be regarded as being in excess of the minimum normal return an investor requires; an economic rent not necessarily means an excessively high return as the draft could be interpreted. Additionally the two examples in Box 1 describing the sources of capital gains probably are inframarginal. However, the second example creates double taxa- tion and does not justify taxation even though according to the authors there is an economic rent. Therefore we ask for further explanations of the meaning of economic rent in the draft. Furthermore, normally, public licenses create a natural monopoly. Economic theory did extensive research on natural monopolies and literature provides a vast bunch of recommendations how to prevent inappropriately high rates of re- turn. If a country does not apply these recommendations this could be corrected by taxation. However, we would like to ask why the publishing institutions ask for an ex post instead of an ex ante correction of the market outcome. Additionally the draft states that “LSRs … can be taxed (at up to 100 percent, in principle) without causing any relocation or cessation of activity, or any other distortion”. (p. 20) Here we strongly disagree. Any taxation of OITs affects both the average and the marginal effective tax rate of an investment. While the for- mer affects the location decision the latter affects the size of investment. If an 1 Wikipedia, https://en.wikipedia.org/wiki/Economic_rent, September 2018, a similar definition can be found e.g. in: Wessel, Robert H., A note on economic rent, in: The American economic review, issue 57.1967, 5, S. 1221-1226, Cambridge, Mass. 2 Henrekson, Magnus/Stenkula, Mikael, The Entrepreneurial Rent: The Value of and Compensation for Entrepreneurship, in: Journal of entrepreneurship and public policy, issue 6.2017, 1, S. 11-25, Bingley : Emerald 2 The Taxation of Offshore Indirect Transfers – A Toolkit OIT or any other sale of immovable property is taxed at a higher rate than a sale of an investment that could be shifted abroad this will for sure affect the choice and the size of investment. Furthermore it should be kept in mind that economic theory and empirical evidence clearly showed that barriers to exit – e.g. caused by an extremely high taxation of capital gains in case of disinvestment – are bar- riers to entry. In consequence investment in infrastructure like telecommunica- tion licenses could be lower than publicly desired. The draft also discusses possible extension of the LSR concept e.g. the “access to domestic markets”. (p. 21) The draft does not distinguish whether this holds only for regulated markets or market access in general. We ask for care with re- spect to such extensions. Market access in general is closely linked to the per- mission to serve domestic demand and domestic demand is by definition immo- bile. In consequence the necessary condition in the OECD and the UN MTC that at least 50 percent of the value of the transferred stock or interest has to be derived from immovable factors in the location country could be regarded as satisfied by tax authorities. To provide legal certainty the enumeration of immov- able property in Box 10 (p. 52) should be regarded as comprehensive. Since the draft concedes that “the concept of LSR has not been sufficiently fully devel- oped to be readily captured in legislative language” (p. 21) the draft e.g.could contribute to the definition by additionally making (first) exclusions which sources do not constitute a LSR. • Legal Certainty We welcome the clarification that a taxing right for the country where the under- lying asset is located cannot be supported without an appropriate definition in domestic law of both the taxable asset and the domestic law basis to assert that taxing right. This enables contracting partners to take taxation into account in their negotiations. For that reason we strongly reject the application of a general anti abuse rule (GAAR) for the taxation of OITs. A GAAR normally is designed very general to cover all possible cases of abuse and there are no indicators whether taxes will be due or not. In our opinion the argument that a country might not be aware of it’s taxing rights and the absence of taxation of OITs is not a deliberate policy choice is no longer valid. There are several prominent examples – some of them docu- mented in the draft – each with huge fiscal impact, now two discussion drafts published by the Platform for Collaboration on Tax dealing with the taxation of OITs and a strong commitment of NGOs drawing attention on the existence of 3 The Taxation of Offshore Indirect Transfers – A Toolkit OITs and related taxation issues. For the same reason we welcome the basic advice that taxes should only be levied on a prospective and not retroactive basis. (p. 40) Contrary to the draft we cannot see any reason for a retroactive application and ask for a withdrawal of the passage on page 40, section 3. • Tax enforcement The draft uses simplified stylized examples to demonstrate the application of the toolkit. “The simplified legislative provisions also do not deal comprehensively with more complex issues such as minority shareholders, joint venture arrange- ments, valuation difficulties, treatment of losses, listed securities, and other dou- ble taxation issues that might arise under a given set of circumstances.” (p. 40) Actually minority shareholders or joint venture arrangements are very common and need to be dealt with. The two models described in the draft how to tax OITs both have their specific pros and cons. However, these pros and cons are primarily discussed from the point of view of the tax authorities with respect to tax enforcement and collec- tion. In particular for the complex cases above the impact of both models on the operating entity that generates the LSR e.g. a limitations of liquidity should not be neglected. We would appreciate further input on this issue. 4 The BEPS Monitoring Group Comments on the Public Discussion Draft from the PLATFORM FOR COLLABORATION ON TAX on THE TAXATION OF OFFSHORE INDIRECT TRANSFERS – A TOOLKIT REVISED VERSION These comments have been prepared by the BEPS Monitoring Group (BMG) and the Financial Transparency Coalition (the FTC). The BMG is a network of experts on various aspects of international tax, set up by a number of civil society organizations which research and campaign for tax justice including the Global Alliance for Tax Justice, Red de Justicia Fiscal de America Latina y el Caribe, Tax Justice Network, Christian Aid, Action Aid, Oxfam, and Tax Research UK. These comments have not been approved in advance by these organizations, which do not necessarily accept every detail or specific point made here, but they support the work of the BMG and endorse its general perspectives. They have been drafted by Sol Picciotto, with contributions and comments from Jeffery Kadet, Cristián Garate, Francis Weyzig and Matti Kohonen. We appreciate the opportunity to provide these comments and are happy for them to be published. September 2018 SUMMARY We supported the first version of this report, while making some suggestions for its strengthening. Due to critical comments from commentators, many of whom make use of or advise on indirect transfers as tax efficient exit strategies, this revised version has been weakened such that it is less clear and would be less useful for the many developing countries, and some developed countries, which would benefit. The case studies this revised version includes from developing countries, as well as one which we contribute in this submission, show that the problem concerns avoidance of host country taxation of the capital gains from the transfer of control over a local business as a going concern. The economic analysis presented in this draft provides clear support for the host country’s right 5 to tax the gains (economic rents) deriving from location-specific assets and activities. The legalistic distinction between immovable and moveable assets is archaic and unhelpful, and we recommend a further rewriting of the paper to foreground and develop the economic analysis and redraft the legal advice to ensure that host countries which wish to do so can effectively tax the capital gains from the transfer of control over local business activities. 1. General Comments The comments from the BMG on the first draft applauded the Toolkit, while making some comments for its strengthening. Other commentators, many of whom make use of indirect transfers, were understandably more critical. Changes made in this revised version have made the report weaker such that it will be less clear and useful for the many developing countries that sorely need clear and unambiguous guidance. The revisions attempted to make clearer that the report aims to provide only technical analysis of the issue, leaving it to countries to take a policy decision. However, in our view the issue is still not presented clearly or objectively. Instead it has focused on transfers of immovable assets. It has rejected, without giving reasons, our suggestion that the report should clearly extend to indirect transfers of all assets. The same comment was made by the Government of India and also rejected without giving reasons. Although it has responded to our suggestion that there should be included a discussion of article 13(5) of the UN Model, in our view the discussion now included is inadequate and misleading. The exclusion of any discussion of how measures against offshore indirect transfers (OITs) could extend to transfers of control over any business would fail to provide adequate advice to governments which wish to adopt such a policy. The revision builds on a stylized example of an OIT rather than on extensive case-study evidence increasingly available on the subject, although it does describe three cases. An additional case which is apposite involves Ncell in Nepal, where the government of Nepal argues that capital gains tax should be paid on the sale of a Telecoms asset located on paper in the Dutch subsidiary of TeliaSonera (headquartered in Norway). Interestingly, despite the seller of the Nepal company shares being the Netherlands subsidiary, TeliaSonera claims that that the Nepal-Norway tax treaty should exempt the gain. There is no Netherlands-Nepal tax treaty. Specifically, TeliaSonera’s argument rests on the Nepal-Norway tax treaty’s lack of provision for OITs, 1 and especially the presence of clause 13(5) stating that ‘Gains from the alienation of any property other than those referred to in the preceding paragraphs shall be taxable only in the Contracting State of which the alienator is a resident’. Since 13(1)-13(4) are not applicable, clause 13(5), if it were to apply, prevents source country taxation. TeliaSonera in a media response states: ‘There are no tax obligations in Nepal on the foreign part of the transaction. Instead any taxes levied on the transaction should be paid in Norway, a country which has a double taxation agreement with Nepal in which Nepal has waived its right to tax in favour of Norway.’ 2 1 http://internationaltaxtreaty.com/download/Norway/DTC/Norway-Nepal-DTC-May-1996.pdf 2 https://www.teliacompany.com/en/news/news-articles/2017/ncell/ 6 Nepal contested this interpretation, as it saw the transfer as being between an entity in the Netherlands (and not a Norwegian entity transferor) and a transferee entity in the Netherlands. This was the offshore transfer. As Nepal does not have a tax treaty with the Netherlands, the capital gains realised by the Dutch transferor are not exempted from taxation. The Nepalese tax authority won the case domestically, and a parliamentary inquiry also considered that Ncell should be taxed for its capital gains made on the sale of the telecoms asset. This example reinforces the need for the Toolkit to provide clear guidance for both ‘immoveable’ and ‘moveable’ assets. Without such guidance, developing countries will continue to lose tax revenue. We believe that there should be a reconsideration of our previous recommendation that the Toolkit should clearly present the following options as good technical solutions to the problems experienced by many countries: • that all countries signing the MC-BEPS should accept its article 9(4) for all their covered treaties; • that all countries should renegotiate their non-covered treaties to include article 13(4) of the UN model; and • that all countries should renegotiate all their existing treaties to include article 13(5) of the UN model. 2. Location-Specific Assets From an economic perspective, the argument for taxing capital gains from asset transfers in the source country rests on the concept of location-specific assets. In our view the paper is confusing, because it is predominantly argued in legalistic terms, using the archaic distinction between immovable and movable assets. This distinction is outdated and often imprecise, due for example to differences between common law and civil law countries. Furthermore, measures confined to immovable assets are easily avoided, since with a 2-level structure any immovable asset located in the source country can be transformed into a moveable asset in the resident country. The title of the paper is the taxation of offshore indirect transfers, and it should therefore discuss such indirect transfers of all types of assets. However, as currently written, its focus is on transfers of immovable assets. This seems based on the view that there is ‘wide acceptance that capital gains taxation of OITs of “immovable” assets can be imposed by the location country’ (p. 8, also p. 19). However, the paper later concedes that ‘[t]he rationale, in terms of economic principle, for limiting this treatment to immovable assets is unclear’ (p. 25). In this context it mentions article 13(5) of the UN Model and provides a discussion of cases in India, Peru and Uganda. However, the point about the rationale extending beyond immovable assets is not pursued. Instead, in the context of discussion of taxation of ‘location specific rents’, the issue is stated as concerning the definition of immovable property, which ‘could’ be extended to cover a broader category of immovable property, such as licenses to exploit public goods (p. 52). However, this possibility is not pursued, and the remainder of the discussion assumes that the only concern is with immovable assets. The legal solutions offered indeed depend on an appropriate definition of immovable property (discussed on pp. 51-3). 7 In our view, the approach in the paper presents a misleading analysis of the broader approach to the issue. Under the broader approach the issue is that the sale of the shares owned by the offshore entity (B in the chart on p.13) constitutes a transfer of a controlling interest in the business of Corp. A. This will include a range of assets, but from this perspective it is irrelevant whether they are ‘immovable’: the sale of the business as a going concern is itself the underlying asset. The three examples given as illustrations in the paper all concern this type of transfer of a controlling interest in a business. The examples also demonstrate two key points: • such transfers are frequently made by acquisition of an intermediary entity resident offshore in a treaty jurisdiction; and • the assets of the business are location-specific; this includes not only rights to natural resources but, as the Vodafone, NCell and Zain cases illustrate, government licenses which have been used to build a local customer base and market share. In our view, the legalistic distinction between movable and immovable assets is unsuitable to analyse this issue. As the revised draft accepts, it can be said that a license granted by the government is in a sense immovable. Yet so are other elements of the business, not only its physical plant, but its local workforce and customer base. The example of telecommunications companies is clearly relevant. They have been prominent because mobile telephony has been a booming business. Yet the assets of such a business cannot be regarded as ‘immovable property’ in the legal sense. A better approach is to use the economic concept of location-specific assets. However, this can clearly extend beyond telecommunications. Many types of business can build a customer base and market share using government licenses. This also results from marketing such as local advertising and branding. But from the economic viewpoint it is not only the licenses, the brand-name or goodwill that constitute the assets which a purchaser acquires. From this perspective, it is the business as a going concern that itself constitutes the relevant assets. There are good economic arguments for taxation of the gains from the transfer of control over a business, and for doing so in the country where that business is located. Generally, the acquirer of a controlling block of shares will pay a premium over the prevailing market price for portfolio investments in the same shares, especially when the acquirer is a firm in the same line of business that is making a strategic acquisition that will expand its worldwide business. This recognizes the expectations of higher profits from integration. As the revised draft points out, although transfers due to corporate reorganisations are generally exempt from capital gains tax, such exemptions are usually subject to continuity of ultimate ownership (p. 12). Gains from the transfer of a business to a new owner can legitimately be considered subject to capital gains tax. The draft also points out the economic argument for taxing such gains, since they result from the increase in value of the assets up to the point of the transfer to the acquirer taking account of future taxes likely to be paid, which from an economic perspective captures otherwise untaxed earnings (p. 16). Since this increased value derives from carrying out the business in the country where it takes place, it seems fair that it should be taxed there. In our view the paper should be revised again to foreground the economic analysis, and to discuss the legal issues in the light of that analysis. A legalistic response to the point we have made here could be to limit the paper to indirect transfers of immovable assets. This seems to 8 have been implicit in the rejection, without explanation, of comments suggesting broadening of the discussion to cover all assets. Confining the paper to immovable assets would be an inadequate response to the concerns of source countries that there is an undermining of their tax base due to indirect transfers of control over a business. While it is correct that the paper should leave it to governments to make the policy decisions on what measures are appropriate for them, to exclude discussion of this point would deprive them of the ability to do so. 3. Relevant Legal Measures A broader discussion of preventing avoidance of source taxation through indirect transfers of control of a business would require extension of the legal analysis beyond article 13(4) of the model treaties, and its incorporation into existing treaties through article 9 of the multilateral instrument. The draft has now included, in response to our comments, a mention of article 13(5) of the UN model (p. 33). However, in our view that discussion is unnecessarily dismissive. As the paper now accepts, that article can have helpful anti-avoidance effects in relation to some kinds of dividend stripping and change of residence strategies. Moreover, where the jurisdiction of tax residency of the transferor in an OIT transaction (and thus which DTT might in fact be applicable) cannot be easily determined (such as in cases like Ncell’s transfer of an asset owned by a Dutch subsidiary), a clause that exempts from taxation everything that isn’t mentioned in preceding articles seems overly prescriptive and likely to be to the detriment of source country taxation. This should be mentioned in the paper. A recasting of the paper on the lines we have suggested would affect the analysis of the legal design of implementation measures, which the paper presents as a choice between two models. The revised draft has now changed the emphasis in this choice: the initial draft expressed a clear preference for model 1. The new version has simply deleted not only this statement of preference but also the reasons for it, which in our view were valid reasons well worth bringing to the attention of Toolkit users. Indeed, a recasting of the analysis along the economic lines we suggest would strengthen the rationale for a design based on Model 1. Neither model 1 nor model 2 reflects the complexity of actual cases of OITs, and thus the Toolkit now fails to appreciate the full extent of possible revision to proposed model DTTs as a result of more recent case-studies (such as Ncell) and tax avoidance strategies present in them. In our view Model 1 has clear practical advantages which make it more suitable for developing countries, if properly designed. The country’s policy aims can be achieved purely under domestic law, provided that the law is drafted to be compatible with tax treaties. The draft should explain clearly how this can be done. Furthermore, it should discuss how such law can be drafted to ensure that it can apply to the gains from the transfer of all the location-specific assets comprising the going concern value of the business. As the draft points out (p. 42) ‘the value of the local assets which are deemed to be sold could administratively be determined using assumptions and adjustments based on the price at which the actual shares are sold, on the basis that their value is derived from the value of the local assets’. It is usual under accounting rules applicable to any major asset or business acquisition for valuation work to be performed of the assets acquired (including goodwill) so that financial reporting can be properly completed. It would be very helpful for many countries if the draft could explain this in more detail. 9 4. Downplaying of Motivation for Tax Minimisation New footnote 12 on page 15 reads: Modern complex ownership structures are not necessarily, or even primarily, designed for tax reduction purposes—rather, commercial considerations often underlie these. Nonetheless, one issue does not preclude the other; where business considerations demand forms of complex and indirect ownership, such structures are presumably designed to be as tax-efficient as possible. This was in response to critical comments made by groups and advisors representing the MNEs, private equity funds, and others that are most active in upfront planning of tax-efficient exit strategies. Without question, businesses and investors conduct transactions primarily to achieve business and investment objectives. Any OIT between unrelated persons will reflect those commercial and investment considerations. However, in the vast majority of cases these transactions are specifically structured as OITs in order to achieve zero or minimal taxation in all countries, including of course L, P, and LTJ. The seller will often have initially structured years earlier a specific ownership arrangement that allows an OIT as part of a long-term tax efficient exit strategy. This long-term planning and structuring is not limited to just few high-profile cases named within the report. The Ncell acquisition and later sale is a clear example of this. When this legitimate acquisition was made, the relevant parties created a chain of holding companies that included companies in Norway, the Netherlands, and St. Kitts and Nevis. The acquisition is a real commercial transaction, but the tax avoidance motive in the specific structure chosen is palpable. The existence of legitimate commercial and investment objectives behind any OIT transaction do not somehow cleanse the tax motivation of the OIT. With the above in mind, we strongly suggest that the Toolkit make clear that OIT transactions often will have tax minimisation as one of their goals and that the policy makers and tax authorities within country L must carefully consider this background issue as they consider how to approach OITs and what location-specific assets to include within their coverage. 10 William Morris Chair, BIAC Tax Committee 13/15, Chaussée de la Muette, 75016 Paris France The Platform for Collaboration on Tax Submitted by email: GlobalTaxPlatform@worldbank.org September 24, 2018 Ref: DRAFT VERSION 2: THE TAXATION OF OFFSHORE INDIRECT TRANSFERS – A TOOLKIT Dear Members of the Platform for Collaboration on Tax, Business at OECD (BIAC) is pleased to have an opportunity to comment on the DRAFT Version 2: The Taxation of Offshore Indirect Transfers – A Toolkit issued on 16 July 2018 (the “Revised Toolkit”). We welcome the revision of the previous discussion draft and the effort made by the Platform for Collaboration on Tax (the “Platform”) to provide further clarification and guidance. We were pleased to find a clear statement as to the purpose and status of the Revised Toolkit, namely that it does “not purport to provide binding rules or authoritative provisions of any kind nor does it aim to establish an international tax policy standard of any kind.” 1 We welcome many of the changes made, including the removal of the explicit endorsement of the Model 1 option; the recognition that countries may affirmatively choose not to tax gains on offshore indirect transfers (OITs); the softened language around expanding the application of taxation of OITs to a wider class of assets; and, the reduction of the focus on political pressures. However, we still have several substantive concerns with the Revised Toolkit as drafted. Most notably, by making the Revised Toolkit a more “general” document, it unfortunately does not address the complexities of common scenarios (e.g., reorganisations, minority owners, etc.) faced in practice. This could result in rules being implemented without full consideration of the complex issues involved. This in turn could result in inconsistent and incoherent rules, and in double taxation, increased uncertainty, more disputes, less investment and lower economic growth. Again, we thank you for the opportunity to comment on the Revised Toolkit, reiterate our welcome for many elements of the Revised Toolkit and look forward to the opportunity to engage further before the document is finalised, especially on adding some more detail. More extensive comments on the Revised Toolkit are attached. Sincerely, Will Morris Chair BIAC Tax Committee 1 See Revised Toolkit, page 11. 11 General Comments The Revised Toolkit takes the explicit position that such guidance is “general in nature and in the form of simplified rules-based legal provisions” and does “not deal comprehensively with more complex issues.” 2 We understand that it is impractical and likely impossible to address all the difficulties that may arise under every fact pattern that may apply with regards to OITs. However, the Revised Toolkit should, at a minimum, emphasise the importance of dealing with these technical issues and provide some guidance on the ways in which they could be addressed and the implications of the policy decisions around these options. In our view, providing a complete picture (including all relevant considerations) is critical to providing developing countries with a practical understanding of the full implications of their policy choices. We believe that general guidance on the two proposed models is less relevant and helpful than considering necessary design features that would help deal with economic and juridical double taxation. The lack of such aspects – step-up in basis, options for a deferral or spread of capital gains taxation, ensuring credit systems in countries for taxes already paid, etc. – greatly undermines the value of the Revised Toolkit, especially to resource limited Tax Administrations that will be relying on such guidance. We therefore consider that the Revised Toolkit will be a missed opportunity if it is not revised to take such areas into account. It is critical that the Platform add additional commentary regarding the technical difficulties of dealing with (inter alia) minority interests, corporate reorganisations, treatment of losses, listed shares, group takeovers and valuation. Additional commentary on these items is provided in our specific comments section below. Additional Specific Comments Oversimplification As noted above, while it may be impossible to address all the difficulties that may arise with certain, specific issues, the guidance should emphasise the importance of dealing with those issues, including by using the word “should” instead of “could” throughout the draft when discussing policy choices where there should be a clear preference for addressing an issue (acknowledging, however, that different policy choices might be made by individual countries). The examples and analysis only deal with very simplistic scenarios that are unlikely to reflect the realities of reorganisations (share/asset sales) within large international groups or transformational acquisitions of an international global business. When discussing various options for a local disposition, the Revised Toolkit does not address business or legal consequences at any length. While it is true that tax considerations are likely considered in such analyses, tax considerations are only part of the picture and are likely to be less important than business consideration. For example, businesses are not normally disposing of an asset or business to simply recognise gain. Normally, such dispositions are driven by the business as 2 See Revised Toolkit, page 40. 12 it believes the operations no longer fit within the group’s strategic priorities, or it would be more economically efficient (and deliver the best value to their shareholders and other stakeholders) if another actor were to continue the activities. Also, sellers and buyers have different interests when it comes to a sale or purchase respectively. The buyer may prefer a carve-out of assets, so that it acquires only and exactly what it wants (or limits its potential legal exposure) – whereas the seller may want to dispose of an entire business for legal liability, contracting or simplicity purposes. Moreover, where there is a transformation acquisition, the buyer is acquiring a global business, far from a single asset/assets acquisition (e.g., hundreds of entities in different jurisdictions) – normally also involving thousands of shareholders selling millions of shares through a public offer in the stock- exchange, approval from different national and international agencies, and various business integration plans. In our view, these concerns are much more likely to drive the structure of the ultimate deal versus the specific tax considerations. However, such considerations are entirely absent from the guidance. Business at OECD believes that adding in the business considerations that influence the structure of the arrangement and dealing more comprehensively with tax policy design would support limiting the application of OIT rules to the cases in which taxes are likely to play an important role in structuring the transaction or where tax considerations are likely to be more important or create distortive results. Currently the Revised Toolkit mentions reorganisations and states that the report is not concerned with transfers of this kind. 3 However, we believe the guidance should provide additional clarity by including another subsection to covering reorganisations in Box 5: Change in Control. 4 Similarly, treatment of minority shareholders is mentioned, but not addressed in any level of detail. In our view, publicly-traded stock, reorganizations or group takeovers should be categorically excluded from taxation and there ought to be thresholds to application to avoid certain unreasonable burdens. These types of provisions would highlight the need to consider such exemptions and perhaps serve as acknowledgement that such exceptions are sensible and warranted. We would welcome additional thought and guidance on the question of value and calculation of gain/loss on the OIT. This is an incredibly complicated process of, at a high-level, determining (1) what assets are subject to the OIT regime, (2) calculating the tax basis of such assets, and (3) determining the applicable value of such assets. Often these immovable assets are part of a larger business operation that is being transferred. On this point, we believe further thought should be given to the situations where the inherent value of assets will be realised through locally generated (taxable) profits. For the sake of clarity and a comprehensive guidance for developing countries, the guidance cannot leave unexplained these issues as they are core areas of interest for stakeholders, including Tax Authorities. Another area of particular concern is section “C: The Allocation of Taxing Rights on OITs: 3 See Revised Toolkit, page 12. 4 For example, an additional subsection could be added to Box 5 providing that subsection 3 does not apply to the extent the change in direct or indirect ownership would satisfy country L provisions concerning corporate reorganisations if the transaction were a domestic transaction. 13 Equity and Efficiency Considerations.” 5 In this section, the Revised Toolkit appears to only refer to taxing “gains” when explaining concepts as “fairness,” “efficiency,” etc. We find that this section should include the tax treatment of the issues previously mentioned (reorganisations, etc.) as many of the arguments used to favour and recommend taxation of OITs could also apply to these items. For example, the guidance should note that not all transfers give rise to gains, and transfers that give rise to losses should be treated in the same manner as gains. Tax certainty in this respect plays a crucial role and further guidance and acknowledgement would be welcomed on all such items. Growth and Double Taxation The Revised Toolkit asserts that the sample models have been designed to deal with and prevent double taxation by the location country. 6 However, in our view, the analysis does not adequately deal with economic double taxation “in the sense of the same gains being taxed multiple times in the hands of different taxpayers through realisations of gains on intermediate shareholders through multiple tiers of indirect ownership.” 7 In capital intensive businesses (e.g., extractives, infrastructure, manufacturing, etc.), introducing or expanding capital gains taxation is expected to limit opportunities for investors to rationalise portfolios. If investors cannot dispose of assets because of taxation of the gains, then they may be locked into suboptimal investments, which slows growth. This is further aggravated (with economic distortions expected) when capital gains taxation is introduced or expanded without sufficient consideration and guidance regarding the elimination of double taxation. Further, as a general matter, the analysis dismisses double taxation at the residence country by noting that some jurisdictions are moving towards territorial-based taxation systems such that its risk is reduced. 8 However, if there is widespread adoption of Model 1 or 2 and Article 13(4) and 13(5) of the UN Model, then it is quite likely that double or multiple taxation of the same gains may occur. As a consequence, in our view, the utility of Model 1 is overstated as adoption of Model 1 would create a significant risk of double or multiple taxation. Alternatively, the utility of Model 2 is understated, as the reduced (albeit not eliminated) risk for double taxation is not highlighted as a benefit. Further, the issue of double taxation is potentially aggravated by suggestions of overriding existing treaties (breaking agreements previously made in tax treaties). Different countries’ domestic laws view the interaction between treaties and domestic law differently, but treaties have historically been entered into for the avoidance of double taxation to increase cross border trade, investment and jobs (and more recently – as part of a global effort comprising changes to treaties and domestic laws – also to protect countries against double non-taxation). We strongly recommend that the guidance not endorse a treaty override as an appropriate avenue for implementing an OIT. Several references are currently included in the disadvantages sections 9 of the guidance. The guidance should acknowledge that the suggested Models may not be consistent with existing treaties and, if 5 See Revised Toolkit, pages 18-25. 6 See Revised Toolkit, page 40. 7 See Revised Toolkit, page 41. 8 See Revised Toolkit, page 44. 9 See Revised Toolkit, pages 44 and 51. 14 so, treaties should be renegotiated before a contrary domestic law could be implemented with respect to a contrary treaty. Along with this additional language we also suggest complete removal of references to overriding treaties. We believe these potential issues could be lessened by encouraging countries to work cooperatively in designing any new laws with both the Platform and stakeholders (including business) impacted by any law changes. Transparency as part of the legislative process is important to the perceived legitimacy of the ultimate law, striking a balance, and drafting clear and simple rules. As shown by the OECD’s Co-operative Tax Compliance initiative bringing together Tax Administrations and taxpayers, in transparent fashion, often provides quicker and better results for all stakeholders. In any case, more guidance should be included on options to deal with double taxation, both for asset holder countries and for investor countries. Options could include sourcing rules, explicit exemptions, clarity on the availability of foreign tax credits , deferral of taxation (which would then be spread over the asset life of the transferred assets), step up in tax basis or others. Tax Certainty While some helpful changes were made, our members are still concerned with some of the definition and analysis around location specific rents (“LSRs”). We welcome that the language regarding the class of assets subject to OIT rules has been somewhat narrowed. 10 However, there is still significant ambiguity as the Revised Toolkit retains references to LSRs. The concept of an economic rent is not one that economists have universally defined and the examples/general analysis outlined do not provide any clarity on this point as to what could and should constitute an economic rent. With the same aim, our view is that the paragraph at the bottom of page 53 of the Revised Toolkit, concerning expanding the definition of immovable property, should be deleted. The Revised Toolkit, as admitted, is general guidance to assist developing countries to implement policies that are internationally accepted. This paragraph is recommending further thought on expanding the definition of immovable property and raises the real issue of unilateral expansion of the definition. As also expressed above, the guidance should avoid suggesting treaty overrides – creating a definition that conflicts with the negotiated position could override the negotiated position. This is, therefore, inappropriate and should be deleted. It should be noted, that in many countries the definition of immovable property is established in non-tax regulations (e.g., Civil Code) and the transfer of immovable property may trigger other types of taxes besides income taxation (e.g., transfer tax, stamp duties, VAT, local taxation, etc.) as well as legal issues (e.g., registration in the Property Registry), so a wide definition of immovable property for tax purposes could lead to a lack of harmonized legal framework and more burdensome administrative and tax procedures. 10The language “the report finds a strong case in principle, for a wide class of assets, for the taxation of such transfers by the country in which is the asset is located” was removed from the executive summary of the Revised Toolkit. We also welcomed the general changes to the body of the document (e.g., removal of “definition in a sufficiently expansive manner” from page 24, etc.) on this point. 15 Further, the Revised Toolkit goes on to make a general statement that “LSRs … can be taxed (at up to 100 percent, in principle) without causing any relocation or cessation of activity, or any other distortion.” 11 This statement, in our view, is clearly and demonstrably inaccurate, as any taxation of OITs impacts both the average rate and the marginal effective tax rate of an investment. While the former affects the location decision, the latter affects the size of investment. If an OIT or any other sale of immovable property is taxed at a higher rate than a sale of a foreign investment this will impact the choice of investment. In addition, economic theory and empirical evidence demonstrate that barriers to exit – e.g., caused by an extremely high taxation of capital gains in case of disinvestment – are barriers to entry. Consequently, investment in certain industries or products could be lower (e.g., telecommunication infrastructure) if taxed at an overly high rate. Accordingly, we encourage a more careful wording when describing and characterising LSRs, also considering transactions such as the previously referred transformational acquisitions taking place in stock exchanges where the acquisition of listed shares supersedes the location of the immovable property or assets. Further, since the guidance concedes that “the concept of LSR has not been sufficiently fully developed to be readily captured in legislative language,” 12 the draft should contribute to the definition by making explicit exclusions for sources that do not constitute a LSR. A clarification that a taxing right for the country where the underlying asset is located cannot be supported without an appropriate definition in domestic law of both the taxable asset and the domestic law basis to assert that taxing right should be explicitly expressed in the guidance. This enables contracting partners to take taxation into account in their negotiations. We strongly reject the application of a general anti abuse rule (“GAAR”) for the taxation of OITs. A GAAR traditionally is designed, by its very nature, to be extremely broad, with few indicators whether taxes will be due or not. The wide application of GAAR inevitably would lead to tax treaty override and, consequently, be in breach with customary international law. We were encouraged by the reduction of focus on politic pressures in this area – including removal of the separate section on this topic from the prior draft. As with any tax initiative, Business at OECD wants government to act based on all the facts according to applicable tax policy considerations. The introduction of pure politics in this area often makes for bad policy and inappropriate rules. Lastly, we welcome the basic advice added regarding the application of OITs. 13 First, the guidance states that “unless there are strong reasons to do otherwise” either model should only be implemented on a prospective basis. We agree that retroactivity is negatively viewed from both a legal and tax policy perspective, such that we would suggest going a step further by deleting the caveat as currently drafted. 11 See Revised Toolkit, page 20. 12 See Revised Toolkit, page 21. 13 See Revised Toolkit, page 40. 16 Right to Tax / Not Tax We welcome that the Revised Toolkit contains less biased language in favour of taxing OIT by recognising that some countries may knowingly choose not to tax gains derived from OIT. On this point, we believe the guidance could further develop and explain the rationale (pros and cons) behind such decision as it is currently viewed through the lens of taxation. In our view, there are still several misleading and incomplete assertions when referring to the tax treatment of OITs. 14 The Revised Toolkit should put more emphasis on the fact that not taxing an OITs does not mean that gains will remain untaxed in country L (per illustrated examples), as the guidance implicitly suggests currently. In this respect, earnings derived from the asset will still be taxed in country L as earned (including future income reflected in the current gain and any future, incremental appreciation. Taxation Regimes / Compliance Page 48 of the Revised Toolkit addresses withholding (“WHT”) regimes. The Revised Toolkit should further explain that the nature of a WHT is a means of tax collection consistent with an allocation of taxing rights to source and residence countries, rather than an independent tax. In this respect, taxes based on gross revenue can be distortive compared to net income taxes (especially if there is no corresponding law or treaty agreement to give credit), giving rise to complex issues regarding price negotiation between purchaser and seller, tax relief or asset valuation (e.g., if an existing loan is subrogated). This is particularly distortive where the WHT is not simply a collection mechanism, but instead represents the final tax due (as WHT is levied on the entire proceeds even though the asset transfer may result in either minimal gain or a loss). Countries should, therefore, seek for less distortive WHT mechanisms that have some correspondence between the amount of tax and net income. 15 Our members take their compliance and tax obligations seriously and fully comply with respective tax laws. However, when discussing enforcement and collection, the Revised Toolkit states that “compliance with this obligation could be expected to be low.” We strongly recommend that the Revised Toolkit remove implications that a majority of businesses would ignore their tax obligations, as such is simply not shown in practice. 16 Similarly, in the discussion of the pros and cons of Model 1 and 2, we would encourage a similar softening of the language around this point – which, as drafted, appears to provide significant deference to Model 1. In any case, we would like to see a recommendation included that countries add a model calculation to their legislation on how to calculate and determine any additional taxation. Especially in the case 14 For example, we believe that certain language could be changed to eliminate this bias. Current examples of inaccurate misleading statements are the uses of: “tax advantage” (page 15), “tax benefit” (page 16), “untaxed” (Box 1 in page 16), “chosen not to tax” (page 22), “dissatisfaction” (page 25) “avoid tax” (page 34). In these situations, the actual situation is simply a deferral of tax versus an elimination of tax. 15 As an ancillary point, we would also suggest that group takeovers be included within the mentioned exclusions from WHT. 16 Reports have shown high compliance of large MNEs when it comes to non-resident reporting. 17 of certain industries that are often already subject to additional taxation (e.g., extractives), the determination of the gain and filing requirements are unclear and would benefit from a model calculation. Specific Comments per Questions Outlined 1. Has the draft better clarified the economic rationale for taxing such transfers by offshore indirect owners? Some helpful revisions were made in this area as noted above. However, what is largely missing is an initial overview and pros and cons on capital taxation, which is a front loading of taxation on future revenues, as well as pros and cons on OIT taxation specifically. As it could limit the opportunities for portfolio rationalisation with investors as well as opportunities for countries to have the more appropriate and active investors for the source country, these considerations should be assessed. For example, when capital gains are due, the price charged for the asset may be a grossed-up one to ensure return on capital, which will limit the transferability of local assets and the ability to attract new investors. Moreover, when discussing the pros and cons of Models 1 and 2, on pages 43-44 and 51, respectively, such are primarily discussed from the point of view of the tax authorities with respect to tax enforcement and collection (rather than of a broader range of considerations including good tax policy or economics). Further, there are many non-tax responses to addressing high returns (e.g., an exclusive license or right), and using the tax system as an alternative to address them may not be optimal. If other non- tax methods are being used, then we would question whether tax-based solutions should be layered on top of existing rules or are appropriate at all. 2. The new draft does not express a preference for either of the described legislative approaches to taxing these transfers—is this made clear? We agree that the Revised Toolkit does not explicitly express a preference for either of the described legislative approaches (i.e., Model 1 and 2), and we welcome the changes made, specifically to the Executive Summary and the detailed discussion of Models 1 and 2. However, despite taking a more open-minded approach in the Executive Summary and removing explicit recommendations for Model 1, the Revised Toolkit still implies a preference for the Model 1 approach. In our view, this implicit favouritism is shown in the pros and cons discussions on pages 43-44 and 51, as significant issues with Model 1 (e.g., double taxation) are quickly dismissed, whereas improper considerations (e.g., lack of compliance in reporting) are in the spot light for Model 2, as noted above. 18 3. Does the draft adequately reduce any perceived emphasis on such offshore transfers as constituting tax avoidance, and make clear that the economic rationale for so taxing them is not as an anti-avoidance device? In our view, the Revised Toolkit has a reduced emphasis on tax avoidance. 17 However, as detailed above, there are still flaws within the analysis, including oversimplification, a narrow view of tax policy drivers, and broad, counter-productive language. 18 17 For example, the acknowledgement in a footnote on page 15 of the Revised Toolkit that “complex ownership structures are not necessarily or even primarily designed for tax reduction purposes—rather, commercial considerations” is a helpful step towards recognising common business operating needs. 18 See supra note 14. 19 From: ernest@ernestandmartin.com Sent: Tuesday, September 18, 2018 4:32 PM To: Platform for Collaboration on Tax Subject: The Platform for Collaboration on Tax DRAFT Version 2 The Taxation of Offshore Indirect Transfers- [External] Dear Sir, I have gone through the above document and I am of the view that The second norm suggests that the right to tax returns to foreign investors in the form of dividends from a domestic source being accepted, so too should be a right to tax them on returns in the form of capital gains associated with a domestic source. A counterargument is that the asset price and hence the gain reflects accumulated undistributed and future after-tax earnings, which the location country could have taxed in the past and may tax in the future through the corporate income and other taxes (rent taxes in the extractives, for instance). The gain, that is, reflects earnings that the location country has in a sense simply chosen not to tax. ADD: A country may adopt a non distribution tax to tax any dividends not distributed. Which may reduce the capital gain arising from non distribution. Page 39 Designing the enforcement/collection rules. These rules are critical as they support the enforcement and collection of the resulting tax liability. They can include one or more of the following: (a) Notification/reporting and information exchange mechanisms (e.g. domestic reporting requirements supplemented, where appropriate, by international information exchange arrangements); ADD: country by country reporting specifically designed of OIT transactions, to take place immediately the transaction is completed. Page 39 (d) Other legal protections such as restricting the registration, renewal or validity of relevant underlying assets (e.g. extractive licenses) unless applicable notification requirements have been met and/or until it is demonstrated that either: no tax is payable; the relevant tax has been paid; or satisfactory arrangements have been made for the payment of that tax. ADD: “or automatic cessation of any/ all licences” after the word renewal Page 43 Enforcement/collection rules Under this model, the local asset owning entity remains subject to the ordinary compliance rules applicable to resident taxpayers, with no need for specific enforcement and collection rules—or reliance on assistance in collection 20 treaties—to combat the significant difficulties in collecting the tax where transactions take place between two non-residents. Under this model, the tax authority of the location country can use the full suite of its enforcement tools against the local asset owning entity (e.g. apply penalties for a failure to file and pay tax in respect of the deemed gain, and activate the usual enforcement instruments at its disposal, such as seizing or freezing the local assets and potentially selling them to settle an outstanding tax liability). ADD: “any enforcement measures should be restricted to the proportion of the shares transferred or only the shares effectively transfered, so as deprive only the parties who engaged in the transaction and not affect other parties who were not in the transaction. Kind regards, Ernest Muriu Partner Ernest & Martin Associates Certified Public Accountants Haven Court, Block C, Rm 3 Slip Road off Waiyaki Way (opposite St Marks Church) P.O Box 21180 -00505 Nairobi Kenya Office: +254 722 962 194 C1: +254 720 369 054 | C2: +254 724 824 435 Web: ernestandmartin.com 21 India’s Comments on Version 2 “The Taxation of Offshore Indirect Transfers-A Toolkit” Comments of India on the discussion draft Version 2 titled “The Taxation of Offshore Indirect Transfers - A Toolkit” General Comments 1.1 India again welcomes the work undertaken by the platform for collaboration on tax to address the challenges of addressing the base erosion and profit shifting arising from indirect transfer of assets, particularly movable assets, the capital gains on the transfer of which is taxable in the country where the actual assets are situated, and from where the indirect or the derivative assets derive their value. The revised draft acknowledges the various shortcomings in the first draft and makes changes, but there are certain issues on which India had made in its previous comments on the first draft, which have not been addressed. These comments are again reiterated. Another cause of concern is the inaccurate narrative on the Vodafone case in Box 2 on page 26. India considers its inputs to be of immense importance for bringing further clarity in the matter of indirect transfers and strongly urges that the same may be incorporated in the revised draft. 1.2 India considers this work to be of immense importance for developing countries, and essential for their domestic resource mobilization. Thus, India strongly supports this initiative and urges the platform to clearly and unambiguously recognize indirect transfer of assets for minimizing taxes, as an abuse of both the domestic law as well as tax treaties, in the same way as other tax avoidance practices like treaty shopping have already been recognized as impermissible by the Final Report on Action 6 of the BEPS Project. This aspect has been modified to some extent in response to comments made by certain quarters. Commercial considerations may appear to drive a transaction but the main purpose behind the transaction or structural design is usually tax avoidance which is supported by the actual practices. 1.3 India strongly urges the collaboration to highlight in the report that with the inclusion of modification of the preamble of the treaty specifying that the treaty is not intended to permit abusive behavior, such artificial arrangement aimed at minimizing taxes have already become impermissible under the treaties, once they are amended by the Multi-lateral instrument. 1.4 At places, the draft tool-kit appears to be aimed at addressing only the capital gains arising from indirect transfer of immovable assets. India strongly urges that in the tool-kit being developed no difference should be made between the taxation of capital gains arising from immovable property (which in any case has already been dealt in BEPS project) and movable property. Alternatively, the focus of the draft report could be to deal with the capital gains on indirect transfer of movable assets (that have not been given sufficient attention during the BEPS project). 22 India’s Comments on Version 2 “The Taxation of Offshore Indirect Transfers-A Toolkit” 1.5 India would also like to suggest that since the tool-kit is being developed for the developing countries, it is essential that i.e it ensures that their concerns and views are appropriately taken into account. Specific Comments 2. With reference to the issue of “C. The Allocation of Taxing Rights on OITs: Equity and Efficiency Considerations” India would like to point out that it has already been widely accepted in the BEPS project with the endorsement of all countries of G-20, OECD and other associates, that income should be taxed in the jurisdiction where the economic value is created. Since the capital gains of an asset represent the accumulated value of its returns over time, they should be taxed only where the actual assets from which the gains are expected to be derived in the future are located. As pointed out in the draft report, the rights for taxation of capital gains from direct transfer of those assets provide a clear indication that the indirect transfer should be taxed similarly. India would like to point out that since the economic impact of indirect transfer is same as direct transfer, and it is fully recognized that gains should be taxed where the value is created, there cannot possibly be any argument that economically there is a difference between the direct transfer and the indirect transfer of assets. 3. With reference to the counterargument on page 22 that “The increased value of the entity sold may reflect in part managerial and other expertise contributed by the seller, beyond what has been recovered in managerial fees, royalties and other explicit payments ”, India would again like to point out that it is a flawed and self contradictory argument , since the legal entity owning the assets directly and the related entity that derives the capital gains are legally distinct entities and any transactions between them are required to be undertaken at arm’s length price. Once that condition is fulfilled as per the tax treaties, this argument loses relevance. Thus, India reiterates its agreement and support for the conclusion drawn by the collaboration of not taking this flawed argument into account. 4. With respect to the Vodafone case which has been given as an example in Box 2 at page 26 in the report, India will like to bring to notice that the facts in second and the third paragraph of the Vodafone box is not an accurate position. Firstly, as per Indian law the purchaser is required to deduct tax at source while making payment to non-resident seller and secondly, the amendment through Finance Act, 2012 was brought in as a clarification to explain the intent of Indian legislation in situations of indirect transfer. Since the report is in a draft stage, the second and third paragraph in the Box of the report should be amended as under: “As per the Indian Law, the purchaser is required to deduct tax at source while making payment to the non-resident seller. Accordingly, the Indian Tax Authority (ITA) held the purchaser, 23 India’s Comments on Version 2 “The Taxation of Offshore Indirect Transfers-A Toolkit” Vodafone’s Dutch subsidiary, liable for failure to comply with its obligation to withhold tax from the price paid by it to Hutchison on the ground that the capital gains realized by the seller were taxable in India. This sparked a protracted court case, with the Supreme Court of India ruling in 2012 in favour of the taxpayer. The Supreme Court denied the ITA’s broad reading of the law to extend its taxing jurisdiction to include indirect sales abroad, though it took the view that the transaction was in fact the acquisition of property rights located in India. The government of India subsequently brought in a clarificatory amendment with retroactive effect to overcome the technical difficulty arising out of the Supreme Court ruling so as to allow taxation of offshore indirect sales and to validate the tax demand raised against the Vodafone's Dutch subsidiary. The legality of a retroactive effect of the law was subsequently not challenged by Vodafone in the Indian courts and instead it has submitted the action of the government of India to arbitration under the India-Netherlands Bilateral Investment Treaty". 5. The facts mentioned in last paragraph of page 28 of the toolkit “The cases show that the location country may well respond to defeat in court by quite sweeping policy changes . India, for instance, not only changed its domestic law to bring OITs into tax but sought to apply this retrospectively to 1962 (the date of the current income tax act)” is not the correct position. The 2012 amendments in Income tax law were clarificatory in nature to restate the legislative intent in respect of scope and applicability of provisions relating to international taxation and for providing certainty in law due to certain judicial pronouncements which created doubts about the scope and purpose of sections 9 and 195 of Indian Income tax Act. This fact has been recorded in the Memorandum to the Finance Act 2012 on page 19 under the heading “F. RATIONALIZATION OF INTERNATIONAL TAXATION PROVISIONS”. The memorandum to Finance Act 2012 is available in public domain on https://www.indiabudget.gov.in/budget2012-2013/ub2012- 13/mem/mem1.pdf and can be mentioned in the footnote for reference. In view of the above, the suggested changes in the text is as below: “The cases show that the location country may well respond to defeat in court by quite sweeping policy changes. India, for instance, not only changed brought a clarificatory amendment in its domestic law with retrospective effect from 1962(the date of applicability of the current Income tax Act) for taxation of to bring OITs1 into tax but sought to apply this retrospectively to 1962 (the date of the current income tax act)” 6. With respect to the domestic law measures, India agrees with the observation of the collaboration that the tool-kit is only indicative in nature and needs to be adopted in accordance with the specific needs of the applicable Constitutional and legal framework of a tax jurisdiction. 1 Memorandum to Finance Act 2012 available at https://www.indiabudget.gov.in/budget2012-2013/ub2012- 13/mem/mem1.pdf. 24 India’s Comments on Version 2 “The Taxation of Offshore Indirect Transfers-A Toolkit” 7. With respect to the tax treaty measures required for addressing tax avoidance by indirect transfer of movable assets, India would like to point out that where the treaty provides taxation of capital gains from transfer of movable property under Article 13(5) as in provision based on UN Model, such intent to allocate taxing rights to the source state should be preserved by extending the scope of Article 13 (4) to such capital gains in Article 13 (5). India would like to point out that this can be achieved relatively easily by amending the text of article 13(4) as under: 4. Gains derived by a resident of a Contracting State from the alienation of shares or comparable interests, such as interests in a partnership or trust, may be taxed in the other Contracting State if, at any time during the 365 days preceding the alienation, these shares or comparable interests derived more than 50 per cent of their value directly or indirectly from immovable property, as defined in Article 6, situated in that other State or from shares referred to in paragraph 5 . 8. India would also like to point out the following justification for recommending the amendment in Article 13 of the UN Model and tax treaties based on it: • The purpose of Action 6 Report was to prevent the granting of treaty benefits in inappropriate circumstances, and for this purpose, certain changes have been recommended in the OECD Model Convention and Commentary in the Final Report on Action 6. Since that exercise was undertaken keeping only the OECD Model Convention in view, it will be appropriate for the Platform for collaboration on tax, constituted specifically for addressing the concerns related to developing countries to consider treaty abuse concerns that arise only in the UN Model Convention. • Paragraphs 41 to 43 of the Report on BEPS Action 6 address the issue of transactions that circumvent the application of Article 13 (4) of the OECD Model Convention. It does not refer to the UN Model as the practice followed during BEPS was to focus exclusively on the OECD Model Convention, though it was expected that concerns specific to the UN Model can be subsequently dealt by the Committee of Experts. • Unlike OECD Model Convention, paragraph 5 of Article 13 grants certain taxing rights to the country of source for taxation of capital gains from transfer of shares. While Article 13 (4) addresses abusive transactions with a purpose of avoiding tax in respect of paragraph 1 of Article 13, it does not prevent similar abusive transactions in respect of taxation of capital gains under paragraph 5. Since the nature of transactions that circumvent taxation of capital gains under paragraph 1 and paragraph 5 are exactly the same, there appears to be every reason and justification for preventing abusive transactions that circumvent the application of paragraph 5 of Article 13. • Given the emphasis on preventing granting of treaty benefits in inappropriate circumstances and all possible measures being recommended by the global community 25 India’s Comments on Version 2 “The Taxation of Offshore Indirect Transfers-A Toolkit” for addressing abusive transactions for avoiding tax, we should consider measures that will prevent transactions that circumvent the application of Article 13 (5) of the UN Model Convention. The need for such measures is further highlighted by the urgency shown in further strengthening of the anti-abuse provision in Article 13(4) in the OECD Model Convention. 9. India would also like to request the collaboration to consider the possibility of recommending that the concept of “beneficial owner” which has already been introduced in Model Conventions dealing with interest, dividend and royalty income, should also be introduced in Article 13 of the Model Tax Conventions dealing with capital gains, to prevent tax avoidance in respect of capital gains. 26 Hardeep Singh Chawla, Advocate New Delhi, India ______________________________________________________________________ 24th September, 2018 Attn: Platform for Collaboration on Tax (PCT) Submitted by e-mail to: GlobalTaxPlatform@worldbank.org In re: Comments on the Taxation of Offshore Indirect Transfers – A Toolkit (Draft version 2) Dear Sirs, I thank the PCT for inviting public comments on the discussion around Offshore Indirect transfers (‘OIT’). I have been keenly following the Indirect Transfer debate from its early years and I take this opportunity to further elucidate and put forth my brief comments herein under. To introduce myself, I am admitted to practice law in India having been called to the Bar in 2016. Additionally, I have completed the Advanced Diploma in International Taxation from the Chartered Institute of Taxation, UK in the same year. Having worked for two years in M&A and PE Tax practice of a Big4 in India, I am now assisting a Senior Advocate specializing in Direct tax laws. At the outset I wish to clarify that since taxation of OIT is not embedded in our treaty structures and form part of domestic laws, it poses a risk of developing fragmentally / skewed (potentially one or more countries may try to tax an OIT given facts of each case) and risking a transaction to be taxed doubly and/or negating the credit not being available to the investor. In this regard there is doubt cast on the legal basis for taxing an OIT. In my humble understanding, this initiative should form basis of creating a separate legal obligation in the treaties (possibly in the ongoing MLI discussions) to effectuate and give sound legal basis to taxing an OIT. The following represent my personal comments and should not be construed as professional advice. Comments and suggestions ______________________________________________________________________ +91-844 7505 149 | hardeepchawla1@outlook.com 27 Hardeep Singh Chawla, Advocate New Delhi, India ______________________________________________________________________ On the allocation of taxing rights, as a policy measure, country L (as in the illustration) has effectively rendered the tax treaty (allocating rights of taxing capital gains) nugatory between countries P and LTJ. This disturbs the international taxation system in its entirety. An investor between P and LTJ (as the tax treaty networks currently are) should only have to look at the treaty between the two countries, unless OIT forms part of the tax treaty itself. What I am trying to convey is that while there are arguments to support OIT as a measure to retain source country’s right to tax, however, these remains just arguments / policy directives. Unless there is a specified Article to deal with this in comprehensive manner, these effectively remain outside the scope of the tax treaty and in taxing an OIT via a domestic law amendment contrives (at least morally) inter-nation equity wherein countries P & LTJ would have to give credit to the taxes withheld. While this report suggests that domestic law amendments are sine qua non for affecting taxation of an OIT transfer, in my opinion, it’s a combination of domestic law amendment and the treaty, since otherwise the treaty shall override the domestic law of one treaty partner My comments are from a policy perspective only, however, in my opinion the legal basis (as of today) to tax OIT should not be deemed to exist. Immovable property is distinct asset class vis-à-vis capital gains accrued on the whole of the transaction. Transfer of immovable property hence should not be taken as an analogy to tax capital gains accruing from sale of an entity, viz., a company. UN MTC amendment proves this very fact. Remaining on the subject of Inter-nation Equity, India follows that Dividend Distribution Tax (‘DDT’) is form of an ‘additional’ income tax and not a withholding tax. A view accepted by countries such as the UK, Singapore et.al. meaning thereby that DDT levied by the country is outside the scope of credit mechanism and beneficial rate of taxes on Dividend as provided in Article 10 is not applicable since it’s a tax on the company and not the shareholder. This view in my mind is erroneous since Article 10 should specify ‘any’ tax on dividend. The incidence is shifted only via clever drafting, but in effect, the shareholder bears the tax in his hands when such dividend is distributed. As such this point also hinges on the Inter-Nation Equity. Further, India follows a model wherein gains shall be taxable (in India) when the entity being transferred offshore derives substantially its value from India. Substantially is defined to mean more than fifty percent. There could be countries with different limits, for illustration 26% etc., which brings me to my earlier point that these cannot be only left to domestic tax law amendments and should form part of the larger discussion between tax treaty countries ______________________________________________________________________ +91-844 7505 149 | hardeepchawla1@outlook.com 28 Hardeep Singh Chawla, Advocate New Delhi, India ______________________________________________________________________ to give a definite legal basis to it. Herein, this defeats the basic canon of taxation – certainty. Recently in WABCO v. DCIT 1, the Hon’ble High Court of Madras has held that Indian Company has no role to play in transfer of shares which took place outside the limits of India. This was in relation to treating the Indian Company an agent for the OIT transfer. Effectively, discarding the change in control model. As a policy decision that gels well since doing so may result in double capital gains liability. This should weigh in against this model heavily. To attribute value to the MNE and each individual unit forming part of various countries should also form part of the treaty. This is evidently problematic in its analysis but it should list out factors to be considered viz., location, value of natural resources acquired, managerial expertise etc. It’s akin following the alternative method to arm’s length principle where allocation of value is done in each country wherein the MNE operates. Where certain countries chose not to tax such capital gains they should choose ‘may’ clause in the treaty and in absence of a domestic law to this effect this transaction remains tax free. This seems to be a better legal approach then trying to amend the treaty. Therefore, capital gains from direct transfer vis-à- vis indirect transfer achieves parity from an efficiency perspective. Where the MNE has acquired natural resources and that’s being transferred albeit via OIT then there appears to be a good case for taxing such gains. Theoretically. And this should be one of the factors to be analysed as discussed above. The Hon’ble Supreme Court in Vodafone’s case was dealing with the proposition that I am advocating today, that while there is theoretically a good case made out to tax the OIT, in the present legal regime such transfers are problematic following the letter and spirit of the law. Petrotech acquisition in Peru follows largely the same. A few examples with LOB and PPT should be inserted that explain how these may be invoked in case of an OIT situation. To give precedence of GAAR to tax an OIT transaction should not so easily accepted. Where a Specific Anti Avoidance Rule (‘SAAR’) exists in form of domestic tax provisions coupled with the MLI provisions (to prevent treaty override) GAAR should not be resorted to. This increases complexity and compliance in the hands of the investor and even after that does not provide certainty. OIT rules should form part of the treaty and should be very clear. GAAR, 1 High Court of Madras, W.A No. 884/2018 ______________________________________________________________________ +91-844 7505 149 | hardeepchawla1@outlook.com 29 Hardeep Singh Chawla, Advocate New Delhi, India ______________________________________________________________________ as rightly appreciated by the Toolkit is more discretionary and such will just muddle the already muddled waters of OIT taxation. *** Thank you once again for considering these comments and the opportunity to share my views on this issue. I hope you find the comments useful to the discussion. I look forward to a continued collaboration with the PCT. Sincerely, Hardeep Singh Chawla ______________________________________________________________________ +91-844 7505 149 | hardeepchawla1@outlook.com 30 Introduction: The International Chamber of Commerce (ICC), as the world business organization, speaking with authority on behalf of enterprises from all sectors in every part of the world, welcomes the opportunity to comment on the Platform for Collaboration on Tax Revised Report on the Taxation of Offshore Indirect Transfers (OITs) of Assets. ICC commends the work of the International Monetary Fund (IMF), Organisation for Economic Co-operation and Development (OECD), United Nations (UN) and World Bank – under the auspices of the Platform for Collaboration on Tax (―the Platform‖) to collectively produce ―toolkits‖ for developing countries for appropriate implementation of responses to international tax issues under the G20/OECD Base Erosion and Profit Shifting (BEPS) project. ICC seeks to provide a comprehensive business perspective on these issues to assist the Platform in establishing an effective and global solution. ICC submitted comments to the Platform’s initial consultation in September 2017 seeking feedback on the draft toolkit designed to help developing countries tackle the complexities of taxing offshore indirect transfers of assets. ICC welcomes the new draft of the Platform Report on the Taxation of Offshore Indirect Transfers (OITs) of Assets, which attempts to clarify a number of issues. ICC notes that the revised draft contains improvements over the prior draft and highlights these improvements below:  The draft more clearly establishes the status of the document: ―this report and toolkit does not purport to provide binding rules or authoritative provisions of any kind nor does it aim to establish an international policy standard of any kind‖. The aim of the document is to provide more perspective and practical guidance to developing countries to allow informed decisions on whether or not to tax capital gains and OITs. (p.11)  It reduces any perceived emphasis of OITs as constituting tax avoidance as far as it admits that such a transfer may be undertaken for commercial reasons.  It highlights the possibility that not all transfers will necessarily involve a financial gain, confirming the possibility of producing losses.  It states that not all transfers of ownership result in taxable gains or losses, for example, transfers through mergers or acquisitions. (p.12)  It recognises that countries might decide not to tax direct or indirect transfers of assets at national level.  The revised draft recognised a number of relevant design issues for capital gains taxation that consider the need for reciprocity and the concern of innate double RMN/rdm 21 September 2018 31 taxation. For example, if capital gains are exempt, capital losses will often be considered non-deductible. When capital gains are taxed, capital losses will in principle be deductible and the tax base will be stepped up. Deferral for capital gains taxation can be considered, often in cases where the production capability is not moved offshore.  Whether the taxation of OITs is in place or not, it recommends that any new rules only apply prospectively and with appropriate transition provisions (e.g. a step up in basis to the effective date of any new rules). (p.40)  It highlights that a ―more uniform, coordinated and coherent approach to the taxation of OITs, where countries choose to tax them, can make a substantial contribution to coherence in international tax arrangements and enhance tax certainty‖. (p.55) Notwithstanding the improvements noted above, ICC believes that the revised Platform report can be further enriched in order to more effectively benefit developing countries:  The revised draft often implies (direct) capital gains taxation as a given. More consideration could be helpful for developing countries regarding why certain countries choose not to tax capital gains or defer taxation of capital gains in the first place.  The report notes that it does not deal with corporate reorganisations, minority shareholders, joint ventures, treatment of losses or treatment of listed securities. ICC considers these issues to be relevant enough in practice for taxpayers as well as tax authorities that they should be considered in the analysis and elaborated upon.  Although the revised draft includes more indications for the need to consider reciprocity in the design (e.g. consider deductibility for tax losses in case the capital gains are taxable), some of these features are underdeveloped in analysis. They should at least form an integral part of the analysis of both models.  The report does not deal adequately with economic double taxation. It states that it is possible that the ―same gains are being taxed multiple times in the hands of different taxpayers through realisations of gains on intermediate shareholdings through multiple tiers of indirect ownership‖. (p.41)  Despite the statement that the document does not indicate general preference between the two models of taxation - ―the appropriate choice will depend on the countries’ circumstances and preferences‖ (p. 8) - it still appears to favour the Model 1 approach given the statement that ―The merits of Model 1 - relative ease of enforcement and simplicity of the necessary basis adjustment – can be especially appealing for lower capacity countries. (p.45) RMN/rdm 21 September 2018 32  Besides the implied preference, the analysis of Model 2 is underdeveloped. Although hinted at, the practical analysis of features such as dealing with internal reorganisations, general mergers, losses and double taxation is very limited. ICC would like to offer assistance in further analysis of these important design features.  The document is unclear regarding the distinction between an asset and a gain or a loss. The valuation of the asset is not taken into account in the analysis, although the determination of the capital gain or loss is generally the first and often biggest hurdle that foreign investors encounter in developing countries.  ICC recommends a clearer balance between taxation and deduction (if the gain is taxable in the country of operations, the loss must be equally recovered or deductible).  The document should consider well-defined limits (or safe harbours) to the taxation of OITs.  It highlights that ―countries shall not use the strategy of assessing and obliging taxpayers to satisfy a quota + interest + penalties and be subject to lengthy court proceedings. This disincentivises the investment and removes legal certainty‖. (p.28) Conclusion: ICC welcomes the Platform’s decision to revisit the provisions within the Report on the Taxation of Offshore Indirect Transfers of Assets. For practical relevance, to increase tax certainty for tax authorities as well as taxpayers, ICC hopes that further improvements can be made to the Report along the lines of the considerations noted above. While considerable improvements have been made to the document over the prior draft, the Platform may consider opportunities to further develop the report in the areas noted above. Such enhancements would benefit developing economies in their international taxation policy implementation efforts. ICC remains available to provide further input and expertise to support the future steps of the process. ICC welcomes and encourages the Platform’s continued engagement with the business community in order to address pragmatic and effective approaches. RMN/rdm 21 September 2018 33 The International Chamber of Commerce (ICC) Commission on Taxation The International Chamber of Commerce (ICC) is the world’s largest business organisation with a network of over 6 million members in more than 100 countries. We work to promote international trade, responsible business conduct and a global approach to regulation through a unique mix of advocacy and standard setting activities—together with market- leading dispute resolution services. Our members include many of the world’s largest companies, SMEs, business associations and local chambers of commerce. ICC members work through national committees in their countries to address business concerns and convey ICC views to their respective governments. ICC conveys international business views and priorities through active engagement with the United Nations, the World Trade Organization, the Organisation for Economic Co-Operation and Development (OECD), the G20 and other intergovernmental forums. The ICC Commission on Taxation promotes transparent and non-discriminatory treatment of foreign investment and earnings that eliminates tax obstacles to cross-border trade and investment. The Commission is composed of more than 150 tax experts from companies and business associations in approximately 40 countries from different regions of the world and all economic sectors. It analyses developments in international fiscal policy and legislation and puts forward business views on government and intergovernmental projects affecting taxation. Observers include representatives of the International Fiscal Association (IFA), International Bar Association (IBA), Business and Industry Advisory Committee to the OECD (BIAC), Business Europe and the United Nations Committee of Experts on International Cooperation in Tax Matters. www.iccwbo.org RMN/rdm 21 September 2018 34 24 September 2018 To: The Platform for Collaboration on Tax From: The International Tax and Investment Center Subject: Comments on DRAFT Version 2 -- The Taxation of Offshore Indirect Transfers - A Toolkit We applaud the Platform on the improvement in the second draft of “The Taxation of Offshore Indirect Transfers – A Toolkit” (Second Draft). We appreciate that some of the comments made by ITIC as well as other commentators were considered and reflected through revisions in the Second Draft. While we recognize the improvement, we believe there are still additional enhancements that should be made to the document to further the goal of providing developing countries with the most complete analysis of and options for the tax treatment of Offshore Indirect Transfers (OITs). Our suggested enhancements are summarized below and detailed in the sections following the summary. We also address the three questions raised in the 16 July Press Release. Summary 1. In general, the tone of the paper is still quite negative towards the value that investors bring. Page 8 implies that “returns that exceed the minimum required by investors” should be targeted, and page 28 compares the amount of tax at stake in Uganda to 50% of the health budget, ignoring that imposition of the tax should only be a timing difference. More positively, the paper does now recognize that the issue is very complex and that countries should have the right to choose not to tax. 2. Even so, we believe the Second Draft continues to present a bias towards developing countries imposing tax on OITs, stating in the Conclusion “that it is appropriate that location countries have the right to tax OITs”, rather than presenting a more balanced analysis of the advantages and disadvantages and allowing governments, after considering all the alternatives, to reach their own conclusion. There is only a subtle distinction between concluding that it is appropriate that countries have the right to tax versus it being appropriate that they do tax. A country can exercise its choice not to impose tax by either (a) leaving the right to tax out of the treaty or (b) having it in the treaty but leaving it out of domestic legislation. The paper suggests on page 22 that (a) happens mostly by accident, with “authorities simply not aware of or focused on this issue”. However, in our experience most treaty negotiators are quite well informed and we believe that any choice not to tax is deliberate rather than inadvertent. Page 25 of the paper goes on to say that there will be public dissatisfaction if a country chooses not to tax. We believe however that any public reaction www.ITICnet.org 1634 I (Eye) Street, NW, Suite 500, Washington, DC 20006 Tel: +1 202 530 9799 / Fax: +1 202 537 7987 35 Africa • Asia-Pacific • Eurasia • Latin America • Middle East/North Africa depends on the circumstances, how politicians choose to portray the issue and how balanced the issue is portrayed by commentators. 3. While the Second Draft does provide additional clarification on the appropriateness of basis step up in certain circumstances, the full consequences are not disclosed, but should be. In addition, the appropriateness of basis step up in other situations should be addressed to minimize the risk of double taxation. More emphasis should be placed on the fact that imposition of the tax on the OIT should be a timing issue (assuming appropriate basis adjustment) not a question of whether to impose additional tax. 4. The treatment of losses has received a bit more attention in this Second Draft but deserves more. When there are suggestions of going beyond the norm for reaching a gain for taxation and subsequent collection of tax, equivalent treatment should be provided to ensure that a tax benefit can be realized for losses. www.ITICnet.org 1634 I (Eye) Street, NW, Suite 500, Washington, DC 20006 Tel: +1 202 530 9799 / Fax: +1 202 537 7987 36 Africa • Asia-Pacific • Eurasia • Latin America • Middle East/North Africa Questions raised in Press Release 1. Has the draft better clarified the economic rationale for taxing such transfers by offshore indirect owners? The rationale for taxing has been adequately clarified, but what has still not been adequately addressed is the rationale for why a country might choose not to tax OITs. 2. The new draft does not express a preference for either of the described legislative approaches to taxing these transfers – is this made clear? Yes, this is made clear. 3. Does the draft adequately reduce any perceived emphasis on such offshore transfers as constituting tax avoidance, and make clear that the economic rationale for so taxing them is not as an anti- avoidance device? No, combatting tax avoidance is a common theme throughout the draft. See page 10 paragraph 3, page 15 paragraphs 2-4, page 18 paragraph 2, page 21 footnote 25, page 26 illustrative cases, page 42 paragraph 3, page 47 paragraph 1, and all the discussion of tax avoidance without describing the possibility of relevant business purposes for an indirect versus a direct sale, other than to footnote that there are often commercial business reasons for complex structures, but those structures are presumably as tax efficient as possible. Further Comments 1. The draft does not describe why countries might choose not to tax an OIT, despite the acknowledgement that some do not. An analysis describing such reasons is critical to countries trying to evaluate the merits of taxing versus not taxing. Clearly many countries have concluded that it is best not to tax and countries still trying to reach a conclusion on the issue would be well-served to understand the logic behind that decision. In this regard, we suggest that the draft cite the Tax Notes International article on OITs authored by Karl Schmalz, dated October 3, 2016, which provides an informative description of some of the reasons countries might choose to not tax OITs. On page 18 the draft makes the statement in bold type “Since company A remains resident in country L, the transfer has no direct impact on country L’s future receipts of corporate income tax”. This statement is corrected in footnote 20. We suggest that this paragraph be revised to tell the complete story that receipt of the capital gains tax is a timing difference (assuming capital gains tax rates and regular income tax rates are the same) that will result in corresponding reduction to income tax receipts in the future, this being true at the very least in the case of Model 1. This acceleration of tax serves to distort the government’s revenue stream and is one of the reasons governments might choose not to tax the Income (gain) upfront. Box 1 on page 16 purports to identify “Sources of Capital Gains”. One of the most likely cases of creation of a gain results from the discovery and development of natural resources. To the extent this value is realized through a sale, the tax that would have otherwise been paid over the life of the asset, as the value is realized, will instead be accelerated to a current payment of capital gains tax. Assuming basis is fully stepped up for the purchase, no double tax will result, but the capital gains www.ITICnet.org 1634 I (Eye) Street, NW, Suite 500, Washington, DC 20006 Tel: +1 202 530 9799 / Fax: +1 202 537 7987 37 Africa • Asia-Pacific • Eurasia • Latin America • Middle East/North Africa tax is not reaching earnings that are otherwise untaxed - it is being applied to earnings that would have otherwise been taxed when realized. 2. Where country L has extended its taxing authority to reach a gain on the sale of stock not otherwise taxable in country L, because of the presence of assets held by that entity within country L, it would appear appropriate that the basis of the assets in country L would be stepped up to reflect the price paid for the stock, in order to avoid double tax, as the value acquired by the purchaser is realized, or if the assets were subsequently sold directly. 3. In several places the draft references the equity of recognizing losses in those situations where a gain would be taxed. However, the draft also provides that these losses will be subject to appropriate loss utilization rules. If the tax is imposed on a non-resident entity with no income connected to the country, there is a lot of uncertainty about what “appropriate” loss utilization rules might be. We recommend that, especially in the case where the local country company is involved in the collection of tax on a gain, a loss be allowed to reduce the general operating income of the local company. Without a special provision, there could never be an expectation that the loss would have any value. Therefore we recommend the draft suggest a special loss rule to provide for use of the loss by the local country company, against operating income of that company (with step-down in basis of underlying assets) to provide greater symmetry with gains. Please see the Redlined Copy of the Draft for suggested wording changes and line by line comments. Best regards, Daniel A. Witt President www.ITICnet.org 1634 I (Eye) Street, NW, Suite 500, Washington, DC 20006 Tel: +1 202 530 9799 / Fax: +1 202 537 7987 38 Africa • Asia-Pacific • Eurasia • Latin America • Middle East/North Africa Platform for Collaboration on Tax By email to: GlobalTaxPlatform@worldbank.org 24 September 2018 PwC’s Comments on the Revised Draft Toolkit on ‘The Taxation of Offshore Indirect Transfers of Assets’ PricewaterhouseCoopers International Limited, on behalf of the Network Member Firms of PwC (PwC), thanks the Global Tax Platform for the opportunity to provide comments on the revised draft toolkit on the taxation of offshore indirect transfers of assets (OITs). We address our remarks below in relation to the questions raised by the Platform. We do not repeat here comments made in our earlier submission on the original draft toolkit, although there are points that we still think apply to the revised draft toolkit despite the changes made. Although we include some specific examples, we call for the Platform to provide more calculations in the final toolkit in sufficient detail to deal with situations that may occur (i.e. not to simplify things to aid an initial understanding). These calculations should illustrate the commonalities and differences between Models 1 and 2 or in relation to positions already adopted by particular countries. The decision on whether to tax OITs or what method to use in doing so is a policy choice for countries to make. Countries should consider the implications of disagreements between them in the allocation of taxing rights in this area alongside their specific preferences. In particular, the international tax system and global growth would benefit from reducing as far as is possible the potential for double or multiple taxation of the same gains. 1. Has the revised draft better clarified the economic rationale for taxing indirect transfers by offshore indirect owners? 1.1. There are a number of improvements in the wording of the revised draft toolkit from the first draft. However, if any final toolkit is to provide meaningful analysis to developing countries (and others), it needs to elaborate more fully on areas of detail like losses, reorganisations, minority shareholdings and valuations. _____________________________________________________________________________ PricewaterhouseCoopers International Limited 1 Embankment Place London WC2N 6RH T: +44 (0)20 7583 5000 / F: +44 (0)20 7822 4652 PricewaterhouseCoopers International Limited is registered in England number 3590073. Registered Office: 1 Embankment Place, London WC2N 6RH. 39 1.2. The latter would often be required in situations in which the immovable property in question forms part of a larger transaction. Further in that context, in Box 7 on page 46 of the toolkit outlining the taxable asset rule (full and pro rata taxation), it says “in any other case, the amount computed according to the following formula: A x B/C where- A is the amount of the gain; B is the value of the shares or other interests derived, directly or indirectly, from immovable property in Country L; and C is the total value of the interest.” There is no further elaboration on the component of B and C. In the case of Country L having a prosperous property market, and if the equity interests being transferred have a debt component in addition to the immovable property and other assets, the total value of the interest (i.e. C) could be smaller than the interests derived from the immovable property. Greater certainty would result from the toolkit further elaborating on the components of B and C. 1.3. While the revised draft toolkit refers in places to the possibility of multiple taxation, a broader statement about the aim to ensure rules avoid even double taxation may provide more guidance for governments. This may explore the possibilities of a step up in base cost, a credit for previous taxes or some other relief. The problem may be particularly severe in cases where there are tiers of indirect holdings between the ultimate owners and the entity owning the immovable property. Care should be taken not to tax the same transaction at multiple levels if the transfer is affected by an entity several levels below the main holding company: different countries at each level may see it as a taxable indirect transfer. Further, once an ownership change has occurred under Model 1 as presented, any other alienation of the shares in the next three years by the same entity may well also constitute a trigger event resulting in a considerable burden for the taxpayer and tax administration in agreeing market values and tax liabilities perhaps with little tax arising. Consider the following example: • 1/1/2020 regime begins Co A owns 100% of Co B • 1/1/2023 Co A sells 55% of Co B to Co C - trigger event • 1/6/2023 Co A sells 1% of Co B to Co D - trigger event • 1/9/2023 Co A sells 1% of Co B to Co E - trigger event • 1/1/2025 Co A sells 1% of Co B to Co D - trigger event Despite the suggestion in the paper, while the greater availability of participation exemptions may reduce the problem, it may still exist in the short to medium term at least under Model 1. 1.4. The economic case for the scoping of any OIT provision still seems to be weakly set out. The strong indications that it should apply to any assets generating ‘location specific rents’ do not appear to be sufficiently widely supported, including that the term does not seem to be well defined in economic circles. Moves by countries to extend the application of provisions to apply to assets with some locally defined nexus may well lead to differences between source and residence states that would lead to double 40 taxation. To the extent that countries agree OIT rules by employing Article 13(4) of the OECD Model Tax Convention or the UN Model Tax Convention to ‘immovable assets’, as defined in Article 6(2) of the respective Convention, there is potential for double taxation insofar as source countries seek to define ‘immovable assets’ to include a much wider class of assets, and that would then be regarded as a ‘treaty override’ by the residence country. There appears to be a good policy principle for excluding publically traded stock from any provisions of this nature. 2. Is it clear that the new draft does not express a preference for either of the described legislative approaches to taxing these transfers? 2.1. When comparing Model 1 and Model 2, it is worth noting that the amount of taxable gain is likely to be different under the two models. Further study may be needed to reconcile and demonstrate the differences in the two Models. We provide an example in Appendix 1 that illustrates this point. In particular, the revised draft toolkit suggests that gains may go untaxed in the source country without an OIT regime, but expected future earnings from the immovable assets are encapsulated in the values going forward – a fact that may be brought out by more complex examples. 2.2. It would seem appropriate to include a general principle in the toolkit, indicating that no matter which approach a country uses, the taxable gain on OITs should not be greater than the accounting gains. We provide two examples in Appendix 2 to illustrate situations with OITs involving assets where the taxable gain may be greater than the accounting profit. It is noteworthy that the approach for the computation of taxable gain adopted by the Chinese tax authorities is different from the two Models in the toolkit. 2.3. As to China’s practice in taxing OITs on page 62, some descriptions do not appear to be accurate. For example: 2.3.1. “The general rule is that the gain on direct transfers of assets located in China is taxed at a 25 percent rate and offshore indirect transfers are equally taxed when involving the sale of immovable property located in China…. However, if the holding company is situated in a jurisdiction where the effective tax burden is lower than 12.5 percent or where offshore income is not taxed, the Chinese tax authority may disregard the overseas holding company and re-characterize the indirect transfer as a direct one if it determines that there is no reasonable commercial purpose to the offshore transaction other than avoiding the Chinese tax.” 2.3.2. According to China’s Corporate Income Tax Law, it appears that gains derived by non-residents from transfer of onshore taxable assets are generally taxed at 10%, and gains derived by non-residents from transfer of overseas holding companies that directly or indirectly hold interests in China TREs are not taxed in China unless the transaction does not pass the reasonable business purpose test and China invokes its general anti-tax avoidance rule (GAAR). The effective tax burden is just one of the criteria of the reasonable business purpose test. The 41 comments submitted by the State Administration of Taxation of China for the first version of the toolkit include the details of the reasonable business purpose test. 2.4. In some instance, there could be tax demands from entities that have not realised the gains in point, where some sort of deferral or instalment option may be needed to prevent the need for onerous borrowing or sales to raise the necessary funds. 3. Does the draft adequately reduce any perceived emphasis on such offshore transfers as constituting tax avoidance, and does it make clear that the economic rationale for taxing them is not as an anti-avoidance device? 3.1. While a number of changes have been made to place less emphasis on tax avoidance, there is insufficient discussion about the commercial purposes of many transfers and the purposes of countries in designing their tax systems. 3.2. It is recognized within the paper that countries may make a conscious decision that they don’t wish to tax OITs. However, countries have already made decisions and commitments in terms of tax treaties that are broadly about international trading, which they should follow. If one were to suggest they may have been pressurised into such agreements, this may be perceived as derogatory to those countries and their officials or, at least, a weaker point. To suggest that countries set out to override their tax treaties in relation to OITs rather than renegotiate them is not conductive to the development of stronger trading links and global growth. 3.3. The need for countries to make specific provisions in this area if they are concerned about tax avoidance would appear to be considerably diminished with the advent of the principal purpose test (PPT) set out in both Conventions mentioned above. This is also starting to be widely applied through new treaties and to many existing treaties through the multi-lateral instrument (MLI) to effect BEPS treaty recommendations. 3.4. The revised draft toolkit suggests that compliance in relation to various tax provisions may be low. This does not seem to be supported by any evidence and large MNEs and their advisers would be unlikely to sanction such an approach. 3.5.Reference to any provisions adopted by countries in relation to OITs only applying prospectively seems to be a very positive step toward establishing a robust international tax system rather than appearing to target previous perceived behaviour. 42 4. Other comments 4.1. We look forward to the opportunity to engage further in the process of trying to reach greater consensus on the analysis that might be included in any final toolkit. 4.2. If you would like to discuss any element of this response in more detail please do not hesitate to contact me (or Phil Greenfield, philip.greenfield@pwc.com). Yours faithfully, Stef van Weeghel, Global Tax Policy Leader stef.van.weeghel@pwc.com T: +31 (0) 887 926 763 43 Appendix 1 Example of different taxable gain under Model 1 and Model 2 Year 1 Company A1 uses $1000 cash to set up SPV B. SPV B uses all of the cash to set up Co. C in Country C. Co. C invests in immovable properties in Country C at $1000. Year 2 Company A2 contributes $1500 to SPV B to subscribe for 51% of the shares in SPV B. At that time the value of the immovable properties in Country C is $2400 (and the value of SPV B is $2600 – a controlling shareholding would be worth more than a pro rata 51%). The cash of $1500 is spent at the SPV B’s level over the years and never injected into Co. C. Year 4 Company A2 sells its shares in SPV B to a third party with a consideration of $2500. The transaction is taxable in Country C. At the time of this transfer, the fair value of the immovable properties held by Co. C is $4000. Co. A1 Co. A2 49% 51% Country A Country B SPV B 100% Country C Co. C Immovable properties Calculation of taxable gains under Model 1: Country C may deem Co. C to have directly disposed of the immovable properties a) both at the time of the capital contribution by Company A2 and at the time of the disposal by Company A2, or b) only at the time of the disposal by Company A2 a) Gains with [likely] step-up Year 2 taxable gain under Model 1 = 2400 - 1000 = 1400 Year 4 taxable gain under Model 1 = 4000 - 2400 = 1600 b) Single gain Year 4 taxable gain under Model 1 = 4000 - 1000 = 3000 Calculation of taxable gain under Model 2: The taxable gain may be calculated at the level of the shareholding(s) in SPV B and a) both at the time of the capital contribution by Company A2 and at the time of the disposal by Company A2, or b) only at the time of the disposal by Company A2. a) Two gains 44 Year 2 Taxable gain on Company A1 may be = (100% x 2600)-(49% x (2600 + 1500)) – 510 = 81 Year 4: Taxable gain on Company A2 = 2500-1500 = 1000 (or a proportion thereof based on the extent to which the value attributable to immovable assets represents the value of the shareholding disposed of) b) Single gain Year 2 There may be no gains from alienation of shares (or deemed realization of value) Year 4 Taxable gain on Company A2 = 2500-1500 = 1000 (or a proportion thereof based on the extent to which the value attributable to immovable assets represents the value of the shareholding disposed of) By comparing the taxable gains under the two Models, the taxable gains of $3000 under Model 1 are greater than the taxable gains of $1081 or $1000 under Model 2 (and at each individual stage the same principle applies). The taxable gains of Model 1 are also greater than the accounting and economic gain recognized in this scenario – largely by Co. A2 in this case of $1000 (being $2500-$1500). This is because Co. A1 has a latent unrecognized gain in the value of its shareholding in SPV B (partly representing the value of the immovable properties in Co. C but also other assets/ liabilities in Co. C and the assets/ liabilities in SPV B. 45 Appendix 2 Examples of taxable gain greater than accounting profit Case 1: The underlying assets are located in one country Company A uses $1000 cash to set up SPV B. SPV B uses all of the cash to set up Co. C1 (paid-in capital $200), Co. C2 (paid-in capital $300) and Co. C3 (paid-in capital $500) in Country C. After years of operation, Company A sells SPV B to a third party for a consideration of $2000. Say, the Country C tax authorities look through SPV B and deem Company A to have directly disposed of the equity in Co. C1, Co. C2 and Co. C3 and thus the transaction is taxable in Country C. Country A Co. A Country B SPV B Country C Co. C1 Co. C2 Co. C3 Immovable Immovable Immovable properties properties properties At the point of the transaction, the fair value of the equity interest in Co. C1, Co. C2 and Co. C3 is $700, $900 and $400 respectively which are the deemed proceeds for their deemed direct disposals. The accounting and economic profit of Co. A on transferring SPV B = 2000-1000 = 1000 Under Country C' corporate income tax, the taxable gain to Co. A is the difference between deemed sales proceeds for Co. C1, Co. C2 and Co. C3 at the time of transfer and the original paid-in capital contributed by SPV B to Co. C1, Co. C2 and Co. C3 respectively. The taxable gain on deemed disposal of Co. C1 = 700-200 = 500 The taxable gain on deemed disposal of Co. C2 = 900-300 = 600 The taxable gain on deemed disposal of Co. C3 = 400-500 = (100) If the loss on Co. C3 cannot be offset against the other gains, the total taxable income of Company A would be $1100 (derived from the deemed disposal of Co. C1 and C2) which is greater than the accounting and economic profit of $1000. 46 Case 2: The underlying assets are located in different countries Company A uses $1000 cash to set up SPV B. SPV B uses all of the cash to set up Co. C1 in Country C (paid-in capital $200), and Co. C2 in Country D (paid-in capital $800). After years of operation, Company A sells SPV B to a third party for a consideration of $2000. Say, the Country C authorities look through SPV B and deem Company A to have directly disposed of the equity in Co. C1 and thus the transaction is taxable in Country C. In addition, the Country D tax authorities also decide to tax this transaction for Company A’s offshore indirect transfer of the assets under Co. C2. Country A Co. A Country B SPV B Co. C1 Co. C2 (Country C) (Country D) Immovable Immovable properties properties At the point of the transaction, the fair value of the interest in C1 and C2 is $900 and $1100 respectively. The accounting and economic profit of Co. A on transferring SPV B = 2000-1000 = 1000 Under Country C’s corporate income tax, the taxable gain to Company A is the difference between the deemed sales proceeds (fair value) for the equity interest of Co. C1 at the time of transfer and the original paid-in capital contributed by SPV B. The taxable gain on deemed disposal of Co. C1 which is subject to Country C tax = 900-200 = 700 Under Country D’s domestic income tax law, the computation of taxable gain is similar to the formula in Model 2 of this toolkit. The taxable gain of transferring Co. C2 which is subject to Country D tax = 1000 x 1100/2000 = 550 In the above case, the total taxable gain recognized in Country C and Country D is $1250 which is greater than the accounting and economic profit of $1000 derived by Company A. 47 48 49 50 51 52 From: Sergio Guida Sent: Monday, September 24, 2018 5:45 AM To: Platform for Collaboration on Tax ; Platform for Collaboration on Tax Subject: Final comments / revised version of the PCT's toolkit "Taxation of Offshore Indirect Transfers of Assets" [External] Dear Sirs, I'm really happy to give you the following answers: Questions to consider 1. Has the draft better clarified the economic rationale for taxing such transfers by offshore indirect owners? Yes, now it is clearer. 2. The new draft does not express a preference for either of the described legislative approaches to taxing these transfers—is this made clear? Yes, I'm convinced of it. 3. Does the draft adequately reduce any perceived emphasis on such offshore transfers as constituting tax avoidance, and make clear that the economic rationale for so taxing them is not as an anti-avoidance device? I perceive definitely this attitude. Thanks a lot for your kind attention, With my very best regards, Dr Sergio Guida 53 Transfer Pricing Economists for Development Paris France www.tped.eu September 25th, 2018 The Platform For Collaboration on Tax Discussion Draft The Taxation of Offshore Indirect Transfers – A Toolkit Subject: Comments on the Discussion Draft v2 on The Taxation of Offshore Indirect Transfers (“OITs”) Dear Madam/Sir, The purpose of this letter is to present TPED’s comments after the release by The Platform For Collaboration on Tax of the second version Draft on the Taxation of Offshore Indirect Transfers (“the Draft” or “the Draft Report”). Firstly, we would like to commend the Secretariat for the opportunity of submitting a second round of comments, as well as for the comprehensive document describing how initial comments had been taken into account. In line with the Association’s focus, TPED’s comments focus on the economic aspects of the Draft, but also legal and tax factors have been taken into account. Our comments focus on the following two aspects, which reiterate some of our comments already submitted in TPED’s initial Response dated 19 October 2017. 1. As already highlighted with our comments to the first draft version, it seems the Draft Report is creating misalignments between conceptual definitions and underlying economic theory (see, for instance, the inconsistency on the various diverging definitions of location specific rents). We would appreciate that for the sake of clarity the economic concepts be well defined and consistently understood. Since most of the terms used do not have a legal definition, at minimum, terms that have already been subject to rounds of negotiations, discussions and agreements should be leveraged, such as intangibles, rents, etc. For instance, intangibles and rents have been consistently defined by the OECD and the UN in a transfer pricing context. We acknowledge the Response by the Secretariat that the “definitions used are not necessarily those used anywhere else in domestic laws or international documents”. Still, we believe that the definitions, that the 54 TPED – www.tped.eu Page 2 October 19th, 2017 Draft suggests – different from the international transfer pricing consensus (for instance the intangibles definition1), or the absence of definition (for instance the rents), create confusions. 2. The valuations underlying the 50% and pro-rata calculations can indeed be complex, as the revised Draft acknowledges it. It would be useful to define: ”relevant assets relating to the immoveable property” (page 42) in : “In practice, it is recognized that these valuation exercises are complex to undertake, particularly where relevant assets relating to the underlying immoveable property derive their value from commodity prices, centrally provided inputs (e.g. management and technical expertise) and other group shareholdings”. It should be further clarified that the immoveable property is an asset different from other assets of the company, and that the valuation of this asset should serve as a basis to the 50% and/or pro-rate assessment. Unless the draft suggests that related assets of the immoveable property consist of all the assets resulting from the immoveable property. For instance, the grant of a government right allows a business to start, and clients / customer relationships to be secured. Client list has a value, which from a financial and economic standpoint may be different from that of the right. Shall these values be combined in the context of an OIT evaluation? The evaluations of the “land/buildings/structures” which currently form most of the scope of immoveable property did not give rise to similar valuation challenges, given the more direct / straight-forward valuations methods involved. With a broader definition of immoveable property to “rights”, as such to “intangibles”, the scope of the immoveable property to be valued needs to be carefully delineated. We thank you again for the opportunity of providing comments and remain at your disposal for further comments. Best regards On behalf of TPED2 Sébastien Gonnet, TPED, NERA Economic Consulting Paris Giammarco Cottani, TPED, Ludovici & Partners, Milan 1“ Intangible Property. For purposes of this report, this term is defined herein as property which has no physical presence, for example, a financial asset such as corporate stock; intellectual property; business goodwill” (page 6). 2 The views expressed are those of the authors, not necessarily those of TPED or its other members. 55 TPED www.tped.eu ANNEX A: Redline comments from ITIC 56 The Platform for Collaboration on Tax DRAFT Version 2 The Taxation of Offshore Indirect Transfers— A Toolkit International Monetary Fund (IMF) Organisation for Economic Co-operation and Development (OECD) United Nations (UN) World Bank Group (WBG) Commented [ITIC1]: We recommend a more complete disclaimer be added to the front page, or in the Executive Summary, rather than not appearing until page 11 or 12 of the document at the end of the This document has been prepared in the framework of the Platform for Introduction. Collaboration on Tax (PCT) under the responsibility of the Secretariats and Staff of . the four organisations. Neither this draft nor the final report should be regarded as providing binding rules or authoritative provisions, nor aim to establish an Formatted: Font: 10 pt, Not Bold, Not Italic international policy standard, of any kind. This draft and the final report are Formatted: Font: 10 pt, Not Bold, Not Italic intended to describe an international taxation issue of particular concern to Formatted: Font: 10 pt, Not Bold, Not Italic developing countries, and to provide practicable guidance to them on options for how to address that issue, should they choose to do so. The draft and the final report do not represent the officially endorsed views of those organisations or of their member countries. . Formatted: Font: 10 pt, Not Bold, Not Italic The toolkit has benefited from comments submitted during a public review period, August–October, 2017. The PCT partners wish to express their gratitude for all submissions received. CONTENTS ACRONYMS 5 GLOSSARY 6 EXECUTIVE SUMMARY 8 INTRODUCTION 10 ANALYSING OFFSHORE INDIRECT TRANSFERS 12 A. The Anatomy of Offshore Indirect Transfers 12 B. Revenue Implications 16 C. The Allocation of Taxing Rights on OITs: Equity and Efficiency Considerations 18 TAX TREATIES AND OFFSHORE INDIRECT TRANSFERS 29 A. OITs in the Model Treaties 29 B. Article 13.4 in Practice 33 C. OITs and the Multilateral Convention 35 IMPLEMENTATION CHALLENGES AND OPTIONS 38 A. Overview of Legal Design and Drafting Principles: Two Models 38 B. Model 1: Taxing the Local Resident Asset-Owning Entity under a Deemed Disposal Model 41 C. Model 2: Taxing the Non-resident Seller 45 D. Defining “Immovable” Property 51 CONCLUSIONS 54 A. United States: Dispositions of U.S. real property held by foreign investors 59 B. Peru 62 C. China 62 BOXES 1: Sources of Capital Gains 16 2. India—The Vodafone Case 26 3. Peru—The Acquisition of Petrotech 27 4. Uganda—The Zain Case 27 5: Change in Control 41 6: Source rule 45 7: Taxable asset rule: Full and pro rata taxation 46 8: Enforcement/collection rule: Withholding tax 49 9: Enforcement/collection rule: Notification and agency taxation of non-portfolio interests 50 10: Sample definition of immovable property 52 Figures 1: Stylized Example of an OIT Structure 13 3 2: Article 13.4 in DTTs 35 3: Proportions of Countries’ DTTs including Article 13.4 35 Tables 1. Allocation of Taxing Rights Between Countries on Transfers of Assets Under 29 APPENDICES A. CONSULTATIONS 56 B. COMPARING DIRECT AND INDIRECT TRANSFERS 57 C. EXAMPLES OF COUNTRY PRACTICES 59 D. ARTICLE 13.4 IN PRACTICE—AN EMPIRICAL ANALYSIS 64 APPENDIX TABLES D1 Descriptive Statistics 65 D2: The Likelihood of Including 13.4 in a DTT, Estimation Results 66 REFERENCES REFERENCES 68 4 ACRONYMS BEPS Base Erosion and Profit Shifting CGT Capital Gains Tax DTT Double Tax Treaty DWG Development Working Group of the G20 GAAR General Anti Avoidance/Abuse Rule IMF International Monetary Fund MTC Model Tax Convention LOB Limitation on Benefits LSR Location specific rents MC Multilateral Convention to Implement Tax Treaty Related Measures to Prevent BEPS (Multilateral Instrument) OECD Organization for Economic Co-operation and Development OIT Offshore Indirect Transfer PCT Platform for Collaboration on Tax PE Permanent Establishment SAAR Specific Anti-Avoidance/Abuse Rule UN United Nations WBG World Bank Group 5 GLOSSARY Asset: Something of financial value. Commissionaire arrangement. An agreement through which a person sells products in a given State in its own name but on behalf of a foreign enterprise that is the owner of the products. Direct Transfer. The disposition of a direct interest in an asset, in whole or in part. Direct Interest. Ownership in regard to any particular asset in which there are no intervening entities between the owner in question and the asset in question. Entity. An organization or arrangement such as a company, corporation, partnership, estate, or trust. Immovable Assets. See discussion in text. Indirect Interest. Ownership interest in an asset in which there is at least one intervening entity in the chain of ownership between the asset in question and the owner in question. Indirect Transfer. The disposition of an indirect ownership interest in an asset, in whole or in part. Intangible Property. For purposes of this report, this term is defined herein as property which has no physical presence, for example, a financial asset such as corporate stock; intellectual property; business goodwill. Interest. Effective ownership, in full or in part, of an asset. Limitation on benefits. A treaty provision that seeks to limit tax treaty benefits to genuine residents of the other contracting state. Location Specific Rents. Economic returns in excess of the minimum “normal” level of return that an investor requires – “rents”—which are uniquely associated with some specific location (and can thus be taxed without in theory having any effect on the extent or location of the underlying activity or asset). Model Tax Convention. A model (or template) that can be used as the basis for an actual Tax Treaty negotiated between two countries. There are two primary Model Tax Conventions, one prepared by the UN, and one prepared by the OECD. The two Model Tax Conventions are largely the same, although they differ in a few significant specifics. Multilateral Convention. The Multilateral Convention to Implement Tax Treaty Related Measures to Prevent BEPS is an instrument developed under Action 15 of the G20-OECD Base Erosion and Profit Shifting Project to facilitate and coordinate changes in treaty arrangements. Offshore Indirect Transfer. An indirect transfer in which the transferor of the indirect interest is resident in a different country from that in which the asset in question is located. Onshore Indirect Transfer. Any indirect transfer other than offshore Permanent Establishment. A concept used to determine when an entity has sufficient connection with a country to allow that country to subject to tax the entity’s net business profits attributable to that Permanent Establishment in that country. 6 Principal purposes test. A rule under which if one of the main purposes of an arrangement is to obtain tax treaty benefits, these benefits would be denied unless that granting these benefits would be in accordance with the object and purpose of the provisions of the specific tax treaty. Residence Country. The country in which the person or entity that derives income or capital gain is a resident for tax purposes. Round Tripping. Describes a chain of transactions in which the beginning and end of the chain are in the same country (and normally with the same taxpayer), but intermediate transactions take place through other entities located outside the country. Source Country. The country within which income or capital gain is deemed to arise. Sometimes referred to here as the ‘location’ country. Tax Basis. The original value of an asset for purposes of taxation. Tax basis is typically the original purchase price (plus direct purchase expenses), minus (for business assets) any deduction for depreciation that has been taken by the business for income tax purposes. Tax Treaty. Also known as a Tax Convention or Agreement. A tax treaty, which is usually concluded between two or more countries, prescribes which country has the right to tax the income of an entity or individual that operates in more than one country, so that the income will either not be subject to tax in both countries or, if it is, relief is granted to eliminate double taxation to the extent possible. Transfer of an interest. A change in the ownership interest of an asset, in whole or in part, whether between independent or related parties. Transferor: Person or entity transferring an ownership interest in an asset. Withholding Tax. As used here, this refers to a tax levied by a source country at a flat rate on the gross amount of dividends, interest, royalties, and other payments made by residents to non-residents. 7 EXECUTIVE SUMMARY The tax treatment of ‘offshore indirect transfers’ (OITs)—in essence, the sale of an entity owning an asset located in one country by a resident of another—has emerged as a significant issue in many developing countries. It has been identified in IMF technical assistance work and scoping by the OECD, but was not covered by the G20-OECD project on Base Erosion and Profit Shifting (BEPS). In relation to the extractive industries, OITs are also the subject of work at the UN. The country in which the underlying asset is located may wish to tax gains realized on such transfers—as is currently generally the case for direct transfers of immovable assets. Such treatment might reasonably be applied to a wider class of assets, to include more of those generating location specific rents—returns that exceed the minimum required by investors and which are not available in other jurisdictions. This might include, for instance, telecom licenses and other rights issued by government. The report also recognizes, however, that gains on OITs may be attributable in part to value added by the owners and managers of such assets, and that some countries may choose not to tax gains on OITs for a number of reasons. The provisions of both the OECD and the UN Model treaties suggest wide acceptance that capital gains taxation of OITs of “immovable” assets can be imposed by the location country. It remains the case, however, that the relevant model Article 13(4) is found only in around 35 percent of all Double Tax Treaties (DTTs), and is less likely to be found when one party is a low income resource rich country. To date, the Multilateral Convention has increased the number of tax treaties that effectively include Article 13(4) of the OECD MTC. This impact is expected to increase as new parties sign the MC and amend their covered tax treaties to include the new language of Article 13(4). Whatever treaties may or may not come into play, however, such a taxing right cannot be supported without appropriate definition in domestic law of the assets intended to be taxed and without a domestic law basis to assert that taxing right. Some have asserted that tThere is a need for a more uniform approach to the taxation of OITs by those countries that choose to tax them. Countries’ unilateral responses have differed widely, in terms of both which assets are covered and the legal approach taken. Greater coherence could enhance tax certainty. Others suggest that the decision to tax OITs is no different from any other decision of a particular country with respect to its sovereign taxing choices, and that tax certainty is more likely enhanced by a country simply making it clear in its own statutes, regulations, and treaties as to its position on OITs. For countries desiring to tax gains on OITs, tThe report outlines two main approaches to the taxation of OITs by the country in which the underlying asset is located—provisions 8 for which require careful drafting. It identifies the two main approaches for so doing and provides, for both, sample simplified legislative language for domestic law in the location country. One of these methods (‘Model 1’) treats an OIT as a deemed disposal of the underlying asset. The other (‘Model 2’) treats the transfer as being made by the actual seller, offshore, but sources the gain on that transfer within the location country and so enables that country to tax it. The report expresses no general 9 preference between these: the appropriate choice will depend on countries’ circumstances and preferences. 10 INTRODUCTION This report and toolkit is one of several that respond to a request by the Development Working Group (DWG) of the G20 to the International Monetary Fund (IMF), Organization for Economic Cooperation and Development (OECD), World Bank Group (WBG) and the United Nations (UN)—the partner members of the Platform for Collaboration on Tax—to produce “toolkits” for developing countries for appropriate implementation of responses to international tax issues under the G20/OECD Base Erosion and Profit Shifting (BEPS) project, as well as additional issues of particular relevance to developing countries that the project does not address. The issue taken up here is the capital gains tax treatment of offshore indirect transfers of assets (OITs): 1 the sales, that is, not of underlying assets themselves but, in some other jurisdiction,. of some entity owning those assets. There has been quite widespread concern among developing countries that OITs might be used to avoid, inappropriately, capital gains taxation in the country where those underlying assets are located. This issue, not covered in the BEPS project, was identified by developing countries as of particular significance for many of them, especially, but not only, in the extractive industries. Its significance has also been stressed by the IMF (see IMF 2014, which draws on several cases arising in IMF technical assistance work), the OECD (see OECD 2014a and 2014b, which identify high priority international tax issues in low income countries), and the UN.2 While this issue has long been recognized, it has become of much greater importance in recent years. The aim of this report and toolkit is to provide analysis of and options for the tax treatment of OITs. To these ends, it addresses several questions: (i) What considerations arise in deciding whether or not such transfers should be taxed in the country in which the underlying asset is located? (ii) To which types of assets do these considerations suggest that any such taxation should apply? (iii) How can such taxation, if adopted, best be designed and implemented as a practical, legal matter? The issues at stake are highly complex, in terms of both the underlying economics and in their legal aspects. In addressing them, this toolkit draws on the existing literature3 and on IMF technical assistance work with developing countries, and reflects responses to public comments received from business, civil society, and country authorities on a first draft of the report.4 It does not set 1 Terms italicized on first use, other than company names, are explained in the glossary. 2 Treaty-related capital gains tax issues were also identified as a concern by respondents to a UN questionnaire on BEPS priorities for developing countries (Peters, 2015). See also United Nations, 2017, Handbook on Selected Issues for Taxation of the Extractive Industries by Developing Countries , which has a chapter on OITs: available online at http://www.un.org/esa/ffd/wp-content/uploads/2018/05/Extractives-Handbook_2017.pdf. 3 Including notably Tolenado, Bush and Mandelbaum (2017), Burns, Le Leuch and Sunley (2016), Cui (2015), Kane (2018) and Krever (2010). 4 See Appendix 1. 11 out a single, definitive approach suitable in all circumstances. The aim rather is to identify practicable options, leaving the ultimate decision to each developing country on the basis of its own view andwith a particular view to the circumstances of developing countries. This report is structured as follows. The next section provides an introduction to OITs, sets out a highly simplified stylized example to illustrate the issues that their tax treatment raises, and provides an analysis of the economic considerations that inform answers to the questions of whether a country wishes to tax OITs and, if so, of which types and how Section III describes some recent cases that highlight these concerns, reflecting the variety of current unilateral country rules, and Section IV then focuses on the treatment of OITs as they are currently addressed under the two primary model tax treaties—of the United Nations and the OECD—and discusses the important possibilities created by the OECD’s new Multilateral Convention (MC). Section V then considers in detail issues of implementation raised by two existing approaches to the taxation of indirect transfers. The final section presents conclusions. Appendices provide further detail on the empirical analysis and on selected country experiences. This report and toolkit does not purport to provide binding rules or authoritative provisions of any kind nor does it aim to establish an international policy standard of any kind. Rather, it is intended to describe an international taxation issue of particular concern to developing countries, and to provide practicable guidance to them on options for how to address that issue, should they choose to do so. As such, the report represents the analysis and conclusions of the tax staffs of the four partner organizations, and does not represent the official views of the organizations’ member countries or Management. 12 ANALYSING OFFSHORE INDIRECT TRANSFERS This section explains what is meant by an ‘offshore indirect transfer’ (OIT)—using a simplified stylized example that will be used throughout the toolkit 5—discusses the revenue implications, and considers key conceptual considerations related to the taxation of OITs. A. The Anatomy of Offshore Indirect Transfers Definitions and a simple example By an indirect ownership interest is meant here an arrangement under which there is at least one intervening entity between the controlling owner and the asset in question. A direct interest, in contrast, is one in which there are no intervening entities. Figure 1 illustrates a stylized three-tiered ownership structure. In the terminology just established, Corporation A has a “direct” interest in “Asset”; Corporation B and its ‘parent’ Corporation P1 both have “indirect” interests in “Asset.” Moving up the tiers, Corporation B has a direct interest in the shares of Corporation A, and Corporation P1 has an indirect interest therein. A “transfer” is a change in the direct or indirect ownership of an asset, in whole or in part, whether between independent or related parties. Transfers of ownership may give rise to a taxable capital gain (or loss), and this is at the heart of the concerns in this report. Of course, not all transfers of ownership result in taxable gains (or losses), aside from the issues discussed herein. Transfers through mergers or acquisitions, in particular, may not be taxable events, even if the asset has appreciated (or depreciated) in value if the transaction satisfies domestic tax rules regarding tax-free restructuring or reorganization. Generally, tax-free reorganization rules require that there be very substantial continuity of ultimate ownership to obtain the benefit of the postponement of realization of gains at the time of the transaction. This report is not concerned with transfers of this kind. Transfers can be ‘direct’ or ‘indirect’, the meaning here being that: A direct transfer involves the disposition6 of a direct ownership interest in an asset, in whole or in part. 5 Of course, corporate structures in the real world are generally far more complex than this example- at any point in the ownership chain there may be multiple owners, and complex cross ownership arrangements are common. The model example, however, serves to bring out as simply as possible the core considerations at issue. 6 TransfersSales also include installment sales and those subject to an “overriding royalty;” in both cases, a series of payments is made to the seller (transferor) after the transfer takes place. See Burns, Le Leuch, and Sunley (2016). 13 An indirect transfer involves the disposition of an indirect ownership interest in an asset, in whole or in part. It is the underlying asset that is being indirectly transferred.7 Figure 1: Stylized Example of an OIT Structure Corp. P1 Corp. P2 Shares of B Shares of A P LTJ Corp. B After Sale Only Countries Shares of A (Before Sale) Corp. A L Asset Note: In this transaction, corporation B, resident in LTJ, sells its shares in corporation A to corporation P2, resident in P. This is a direct transfer of the shares in A, and an indirect transfer of the asset held by corporation A that are located in country L. More complex patterns are of course possible, and indeed common. It could be for instance, that corporation B is disposed of by a corporation C (not shown) interposed between corporations B and P1; this would be an indirect transfer of both the shares in B and the underlying assets held by A. 7 So, for instance, a direct transfer of shares in a company owning some real asset is an indirect transfer of that underlying real asset. 14 Tax treaties typically create a distinction between two classes of asset that is critical for this report: Immovable assets: The precise definition of this term is a matter for national law, which may or may not be modified by, and for the purposes of, any tax treaties to which the country is a party. It typically includes land, buildings, and structures as well as rights related to such property (which may include agricultural, forestry, and mineral rights).8 As discussed later, the definition of immovable assets could also include government created licenses to provide specific products or services (e.g. telecommunications) to specified geographic locations, although this is not common. Movable assets: For purposes of this report, by this is meant any asset not classed as immovable. This may include not only other physical property, but intangibles (such as intellectual property or goodwill), and financial assets (e.g., stocks, bonds). Under existing arrangements, in both treaties and domestic laws, the location of the asset and the residence of the disposing party (the ‘transferor’) both play a role in determining which taxing jurisdiction (or jurisdictions) may claim the right to tax transfers. The provisions of various countries’ tax laws in this regard differ widely. For clarity in discussing the complexities of indirect transfers, we define:9 Offshore transfers as transfers in which the transferor is resident for tax purposes in a different country from that in which the asset in question is located, and the transferor does not have a permanent establishment in the country in which the asset in question is located. Onshore transfers as all other transfers. Structuring transactions Imagine in Figure 1 that the owners of P1 want to realize a capital gain reflecting an increase in the value of the underlying asset; and that the owners of P2 wish to gain control of that asset.10 The tax rules of (at least) four countries come into play in shaping the tax treatment of this transaction (along with any applicable treaties): that in which the underlying asset is located (L), that in which the seller is resident (LTJ), that in which the parent of the seller (P1) is resident (P), 8 The definition of immovable property is set out in Article 6 of the model treaties. 9 These terms may have meanings different from those used here in the domestic laws of different countries. 10 We assume throughout, except where indicated, that buyer and seller are unrelated, and so set aside issues related to transfer pricing and, as noted above, the treatment of corporate reorganizations that arise if they are related. 15 and that in which the buyer (P2) is resident. More complex cases can certainly arise,11 but this simple example captures the key concerns. One way to realize the gain would be for P1 to arrange a direct sale of the asset by corporation A. This will generally create a tax liability for corporation A to in country L, being a straightforward domestic (onshore direct) transfer. And generally in such a simple asset transfer case, the basis of the asset would be stepped up to reflect the purchase price such that future income taxes would be correspondingly reduced (assuming the tax rate on the gain on the asset sale is the same as the tax rate on the income generated from such assets). See the discussion below on Revenue Implications. The tax efficient strategy for P1 may be to instead arrange for the sale to be made indirectly by an entity resident in a country (LTJ) that applies a low tax rate to capital gains. 12 In Figure 1, this is shown as the sale by corporation B, resident in low tax country LTJ, to Corporation P2, resident for tax purposes in country P, of its shares in corporation A. Any tax advantage from eliminating the tax otherwise payable in L may be offset later by taxation under the tax rules of the seller’s parent’s country P and by increased income taxes in L. since, unlike in the simple asset Formatted: Font: Italic transfer case, there would not be a basis step up to the assets indirectly transferred. But anything short of immediate taxation in P, may not substantially neutralize the timing tax advantage of selling the asset indirectly in LTJ rather than directly in L, although such immediate taxation in P Commented [ITIC2]: Note that the tax advantage is could lead to economic double taxation.. simply a deferral of tax in L, not an elimination of the tax. This timing effect is noted below in the revenue implications section, where it is stated “The revenue The transaction also has tax consequences for the purchaser, P2, since the amount paid for the issue for country L is thus one of timing” shares13 of company A becomes the tax basis relative to which any capital gains (or losses)14 on a Formatted: Font: Italic future sale of those shares will be calculated. First, P2 loses the asset basis step up it would have Formatted: Font: Italic received had it purchased the assets directly, and thus pays higher income taxes in L over time. With respect to the possible future sale of the shares, iIf the underlying asset is expected to Formatted: Font: Italic decline in value—as a result of true economic depreciation, perhaps because the underlying asset is a right with some expiration date—the expectation is of a future capital loss on the shares; and the value of that for tax purposes will be maximized by locating the loss in an entity located in a high tax jurisdiction (because it generates a deduction with no offsetting charge). If, on the other hand, the underlying asset is expected to increase in value, the tax minimizing strategy is to locate the company which acquires company A in a low tax jurisdiction. It may be possible for residents of the country in which the underlying asset is located to use this structure for ‘round-tripping.’ 15 Since the same logic applies when the country in which the ultimate owner resides, P, is the same as that in which the asset is located, L, capital gains tax that would be payable on a domestic sale in L can—in the circumstances assumed in Figure 1—be 11 There may be many further companies interposed along the chain of entities, between A and B; and title may actually pass in another (fifth) country. 12 Modern complex ownership structures are not necessarily, or even primarily, designed for tax reduction purposes—rather, commercial considerations often underlie these. Nonetheless, one issue does not preclude the 16 other; where business considerations demand forms of complex and indirect ownership, such structures are presumably designed to be as tax-efficient as possible. Commented [ITIC3]: The implication of this seems to be that tax efficiency is bad or suspect—it may instead be the difference between doing a project or not, and 13 The acquirer might prefer to acquire the asset directly, since immovable property will generally qualify for thus employing a lot of people or not. Tax efficiency depreciation allowances and so, in many cases, yield deductions sooner than basis in shares that can be set off often takes advantage of tax provisions designed against future gains. explicitly to promote investment—and thus, when an 14 Such losses, importantly, may be usable to offset gains on other assets. investor reacts to this, is that somehow improper, as 15 Kane (2018) stresses the potential importance of this in relation to indirect transfers. the flavor of this statement suggests? 17 avoided by instead selling indirectly offshore. 16 Any tax benefit from this would be negated, however, if country L taxes its residents on capital gains realized by controlled non-resident entities—unless that gain is illegally concealed from the tax authorities in L. Commented [ITIC4]: This completely ignores the fact that Country L taxes the income from the asset as well. This example is highly stylized: as discussed in detail below, the tax treatment of indirect transfers If it taxes the gain up front—negating the so-called tax benefit to the seller-- it needs to restore that “benefit” in practice will depend on details of both domestic law in the countries involved and any tax to the purchaser—or it ends up engaging in double treaties between them (which may for instance allow country L to tax the sale by company B). taxation. However, many indirect transfers are in practice structured so as to bring the features assumed in the example of Figure 1 into play. B. Revenue Implications The revenue issues at stake in considering OITs are complex, and can be quite case-specific. As throughout this report, the intention here is not to provide an encyclopedic account of all possibilities, but to bring out core considerations at work. As general background for this and later discussion, Box 1 considers the general nature of capital gains and, in particular, how they arise. Box 1: Sources of Capital Gains Capital gains derive in large part from changes, between the initial purchase and sale, in expected future after-tax payments to the owner of the asset.17 Both aspects of this are important: • While capital gains can sometimes be fully anticipated,18 in the cases with which this report is principally concerned they typically arise from unexpected changes in future net distributions, perhaps as a result of a resource discovery or an increase in commodity prices—which in turn are often changes in location specific rents, a concept discussed further below. • Since the value that any actual or potential holder places on an asset can be expected to take into account any future corporate, withholding or other taxes due—including capital gains tax on any future sales—capital gains tax reaches income not taxed by these other instruments. Viewed in one way it is a form of double taxation. More economically relevant, however, it is a way to capture changes in earnings that are otherwise untaxed. Commented [ITIC5]: Box 1 is not accurate. What is 16 This has been a concern, for example, with the treaty between India and Mauritius, under which gains realized in the basis for saying that higher than expected future the latter on transfers of Indian entities are untaxed. This is widely believed to be one reason why around 25 percent earnings are “otherwise untaxed?” These earnings are of foreign direct investment in India in recent years has been routed through Mauritius (IMF, CDIS 2010-2015)— in fact taxed as ordinary income in the country in which though it is unclear how much of this is round-tripping. In May 2016, a protocol amending the treaty was signed. they are earned—so taxing them under a capital gains The new article 13 will allow taxation of capital gains on the alienation of shares of a company resident of a tax-- UNLESS a step up in basis is permitted on the contracting state to be taxed in that state; the applicable rate will be 10 percent beginning 2019. Shares purchased underlying taxable income generating assets in the prior to April 1, 2017, will continue to be exempted from such tax. country-- is clearly economic double taxation—which is the “more economically relevant” point. 17 More precisely, taking the price of an asset to be the present value of expected net distributions to the owner, the capital gain on an asset purchased at time 0 and sold at time T is the amount by which net present value of distributions subsequent to T expected at time T exceeds that expected at time 0, with the latter discounted back to time 0 less (b) the net distributions that were expected at time 0 between then and time T. (This is of course a simplification of complex valuation issues: potential investors may have different expectations, for example, and/or may face different tax treatment on distributions.) 18 The value of an asset that derives from a certain payment at a fixed date in the future, for instance, will on at that account increase as that date approaches. 18 Revenue effects from the transfer itself Consider the two broad possibilities that the owner of an underlying asset on which a capital gain has accrued has for realizing that gain: A direct transfer of the underlying asset itself, which will be subject to tax in country L. Within this option, there is a choice as to whether to sell that asset now or in the future, with the latter having the advantage for the taxpayer of deferring the liability on that gain but the detriment of deferring the cash from a current sale. An indirect transfer, selling an entity that owns the underlying asset. The purchaser, we assume for purposes of this comparative analysis, will eventually sell the underlying asset19 (or it will expire with zero value). In these circumstances, the aggregate nominal value of tax receipts in country L, cumulated over time, is independent of how the underlying asset is transferred. In all cases, the underlying asset is eventually sold, and corresponding revenue collected on the accrued gain. (Of course, there may be further changes in the value of the underlying asset, but these simply imply further charges (or losses) to be combined with that initial accrued gain. If the asset expires with zero value, for instance, there is a future capital loss that offsets the gain accrued at the time of sale). The revenue issue for country L is thus one of timing, rather than the directness or otherwise of the transaction—but the concern can be very sizable one. The longer the sale of the underlying asset is postponed, the lower in present value are country L’s receipts. This timing effect is a consideration of some importance for governments of lower income countries that face constraints on their borrowing capacity. At six percent interest, for instance, a delay of ten years in receiving revenue of $1 billion reduces its present value by around $450 million. Of course, if the owner of the asset is expecting a greater rate of return than the country’s borrowing rate, there is an economic penalty to the investor which is much greater than the benefit to the country—i.e., there is an overall economic loss due to the taxation of the transfer up front versus a reliance on taxation of the future revenues to come. Thus, an investor seeking a 10% return would see the detriment of paying tax on the transfer upfront or delaying the tax for the same ten year period as above as reducing its present value by over $600 million. Given that the revenue issue is one of timing, a number of reasons may exist for a country to Formatted: Font: Not Bold decide not to immediately tax natural resource asset transfers, whether directly or indirectly Formatted: Justified, Indent: Left: 0.55", Right: 1", made. These may include a desire to facilitate free transferability of such assets as a means of Space Before: 6 pt, Line spacing: Multiple 1.1 li, No promoting the most efficient development and operation of the asset and maximizing overall widow/orphan control value for the country. (Norway provides a prime example of this—not distorting in any way the decision to hold on to the property versus selling to a new owner.)1 Alternatively, the overall 1 “Under regulations adopted based on section 10 of the PTA [Norway’s Petroleum Tax Act], capital gains arising from the transfer of assets that are allocated to the petroleum tax regime are not taxable and losses non-deductible (neither when calculating ordinary petroleum tax nor special tax). Moreover, the buyer will take over the seller’s tax balances (including the basis for uplift) and other tax p ositions and 19 fiscal regime for a natural resource project is often quite different from non-resource businesses, and specifically designed to provide a mix of revenues over the very long life of the project. Taxing gains, given the other government take provided for in the fiscal terms, may make the country’s overall fiscal regime non-competitive with other countries, and change the timing of the revenues from that contemplated.2 The question then is whether indirectness can increase the attractions, in tax terms, of deferring sale of the underlying asset. More generally, does taxation distort an initial owner’s choice between, on one hand, a direct sale of the underlying asset today and, on the other, an indirect sale today with direct sale of the underlying asset (by the purchaser) deferred? Appendix B explores this issue. In the stylized setting there, the conclusion is that the possibility of distortion turns on the comparison between the rates at which the gain realized on the indirect sale of an entity will be taxed and (assuming the purchase is financed by borrowing) the rate at which the purchaser can deduct interest income. If the two rates are equal, then the two sale options yield the initial owner exactly the same amount: the tax benefits of deferring sale of the underlying asset are reflected in the price that the purchaser of the shares is willing to pay, and is amplified by the ability to deduct interest paid on the debt incurred to make the purchase; but those tax induced increases in the price at which the entity can be sold increase the initial owner’s liability to capital gains tax on the share transfer. If these two tax rates are equal, the benefits of deferral are exactly 19 If this is not the case, comparison with direct sale is moot. stand in the shoes of the vendor.” See United Nations, 2017, Handbook on Selected Issues for Taxation of the Extractive Industries by Developing Countries, supra, at p. 102 Box IV. 4, “An example of asset sale taxation: Norway.” Formatted: Underline 2 See, e.g., Karl Schmalz, “Capital Gains Issues in the Extractive Industries,” Tax Notes International (3 October 2016) pp. 91-96. 20 neutralized by the capital gains tax on the share transfer. If, however, the rate of tax on the share transaction is low relative to the rate at which interest is deducted—a plausible case—then the indirect route, with sale of the underlying asset deferred, is tax-preferred by the initial owner. While the revenue issue for the location country is thus essentially one of timing, it is reasonable to conclude that indirect transfers conducted in low tax jurisdictions may have the effect of amplifying tax distortions towards delayed sale of the underlying asset. Commented [ITIC6]: Taxing the capital gain on a sale in country is what creates the tax distortion of sell Effects on other tax payments versus retain and operate. Compare the example of Norway where free transferability exists and the sell Since company A remains resident in country L¸ the transfer has no direct impact on country versus hold decision is not distorted in any way by tax. L’s future receipts of corporate income tax20 (or, in the case of the extractive industries, any Can this point be balanced out a bit. royalties or rent tax) from A. (There may be indirect effects from changes in the commercial and Commented [ITIC7]: See item #1 in Further Comments section of cover letter. financial operations of A as a result of changes in its ultimate ownership, but we leave such effects aside in this discussion.) The same is likely to be true, in practice, of L’s receipts from any post-sale withholding taxes on dividend, interest or other payments made by corporation A to its new direct owner. In Figure 1, A’s new direct owner P2 is resident in a country different from that of the initial direct owner B. In that case, different withholding tax rates may apply, with consequent effects on country L’s revenue. It seems to be more common in practice, however, that the transfer takes the form of the sale of B by a company interposed between B and the initial parent P1. Company B thus remains the direct owner of A, and there is then no change in the withholding taxes payable.21 C. The Allocation of Taxing Rights on OITs: Equity and Efficiency Considerations A threshold question is whether or not the country in which an asset is located should have primary taxing rights on its indirect transfer abroad—and, if so, to precisely which assets this should apply. In taking up this question, the analysis here goes beyond the possibility of taxing indirect transfers solely as a back-up method to combat tax avoidance, and sets out key Unless, that is, the sale leads to a step up in basis and the asset is depreciable (as it generally would be under 20 Model 1). 21 It might seem that realizing a lightly taxed capital gain provides a way in which to avoid withholding tax on distributions of previously accumulated retained earnings (on which, being undistributed, no dividend withholding has been collected). But those retained earnings presumably have a value to the purchaser only in so far as they can, at some point, be paid as dividends: at which point the withholding tax will apply. The equity placed in Company A is in a sense trapped, in that the future dividends that ultimately give it value —even if derived from past retentions—will be subject to withholding when paid. On this ‘trapped equity’ view, see for instance, Auerbach (2002). 21 considerations in deciding an appropriate allocation of taxation rights on gains realized indirectly on domestic assets.22 Several (inter-related) issues of economic principle come into play—leaving aside, for the moment, the current practices and legal concepts discussed below. These include: inter-nation equity, in assuring an allocation of revenues meeting some notion of fairness between countries; efficiency, in ensuring that assets are used in the most productive ways; and, not least, political economy—which, given the high profile of many OIT cases, has driven many recent developments in this area. Beyond some basic matters of practicability, issues of implementation—ensuring that tax is collected at reasonable cost to both tax administrations and taxpayers themselves —are deferred until Section 5 below. Inter-Nation Equity Views differ on what ‘fairness’ means in the allocation of taxing rights across countries, but three current norms point to some possibility for consensus in relation to the location country’s right to tax OITs: • Capital gains on onshore direct transfers of tangible assets are taxable by the country in which the asset is located (even though the seller—and, likely, also the purchaser—may be non- resident); Dividends received by a parent company abroad may be subject to tax through withholding by the country in which the paying company is resident;23 • It is quite widely accepted—as reflected in the model treaties discussed below—that the country in which an ‘immovable’ asset is located is entitled, if it so chooses, to tax gains reflecting increases in the value of that asset—though not all countries do so. The first norm points to a view that the country in which an asset is located should be entitled to tax gains associated with it—at least to the extent that those gains are not attributable to value-enhancement provided from abroad (a natural resource deposit has little value, for instance, until it is ‘discovered’). Establishing the extent of any such contribution, however, could of course be problematic; this point is taken up below. 22 Location countries may, having achieved taxing rights over these transfers, elect not to exercise those rights, or not to do so in full in order to promote their business environment—just as many countries elect to grant tax holidays and exemptions in the hope of attracting foreign investment. Whether such incentives are effective, or necessary, is the topic of an earlier Platform toolkit (“Effective and Efficient Use of Tax Incentives for Investment in Lower Income Countries,” 2015). In any event, countries cannot make such a choice if they do not have the underlying right in the first place. 23 Except, for example, by the Parent Subsidiary Directive within the European Union (we leave aside specific intra-EU issues in this report) or by domestic legislation or treaty provision in other countries. 22 The second norm suggests that the right to tax returns to foreign investors in the form of dividends from a domestic source being accepted, so too should be a right to tax them on returns in the form of capital gains associated with a domestic source. A counterargument is that the asset price and hence the gain reflects accumulated undistributed and future after-tax earnings, which the location country could have taxed in the past and may tax in the future through the corporate income and other taxes (rent taxes in the extractives, for instance). The gain, that is, reflects earnings that the location country has in a sense already taxed (i.e., prior after-tax earnings not yet distributed) or will tax in the future (the expected after-tax earnings the asset will generate over its remaining life—but which is subject to taxation as earned)simply chosen not to tax. But this counterargument is not wholly compelling, especially in the case of low- capacity countries. Dividend tax rates may be constrained by tax treaties—though that could be interpreted as simply another way in which country L has chosen not to not to tax future Commented [ITIC8]: This does not seem to follow. earnings. Perhaps more persuasively—a point taken up later—the exploitation of avoidance The new owner will still be taxed on undistributed opportunities may diminish the effective power of the country in which the underlying assets are earnings that are subject to the dividend tax, future located to tax future earnings. In the limit, for a country that cannot effectively tax either the earnings are directly taxed, and their after-tax amounts earnings of the acquired entity or the dividends paid to a foreign parent, taxing the gain on asset will also be subject to the applicable dividend withholding tax—see the previous comments on transfers, direct or indirect, may be its surest prospect of raising revenue on the associated withholding taxes which concluded, “… Company B thus earnings. remains the direct owner of A, and there is then no change in the withholding taxes payable.” The third norm highlights the importance of the concept of ‘immovability,’ and the question The comparisons should always be to what tax the country would have received if there had been no of why it should matter for tax purposes whether an asset is movable or not. The distinction transfer, or if there had been a direct transfer—with is not one that comes naturally to economists, who simply conceive of assets as things that have basis step up. value because they have the potential to generate income—putting intangibles like patents, or a brand name, on a par with, for instance, natural resources. There appear to be three possible Commented [ITIC9]: This is hard to understand—and rationalizations—related, but distinct—for the importance given to the distinction: the following sentence just seems to posit a concern that does not exist—especially with respect to extractive assets. We would hope this would be Pragmatically, immovability facilitates the collection of tax, since the asset can be clarified. seized in the event of non-payment, with no risk of its fleeing abroad. Why would a country not effectively tax the earnings of • Immovability of an asset may imply that its value reflects, to some degree, its the acquired entity—if there had been no transfer, would you somehow make this same point? And why location. That value may, more precisely, reflect location specific rents (LSRs): receipts, that can’t it tax the dividends paid to a foreign parent? is, which are in excess of the minimum “normal” return that the investor requires, with these Again, is this any different from the previous parent? ‘rents’ being uniquely associated with a particular location. LSRs are in principle an ideal Commented [ITIC10]: Doesn’t this need to be object for taxation, because they can be taxed (at up to 100 percent, in principle) without qualified? “Up to 100% in principle” of profits in excess causing any relocation or cessation of activity, or any other distortion—and so provide a of a risk adjusted rate of return to the investor on its invested capital? fully efficient tax base (dominated, on efficiency grounds, only by taxes that serve to correct And doesn’t it further need to be qualified given that a some externality). While this in itself is an efficiency argument (not speaking directly to big assumption is that these types of investments are the question of which government should receive the revenue), in practice there is also not available elsewhere—i.e., in other countries—which widespread if usually implicit recognition that it is appropriate for revenue from taxing given the current worldwide resource base is not really correct. what are manifestly LSRs to accrue to the government of the place of location. The most Without appropriate qualifications to these statements, obvious examples of such assets are often thought of—and in the resource case generally this language can be quite misleading. We would are—owned collectively by the nation. suggest eliminating it, or fully explaining the economic theory—which seems beyond the scope of this paper. 23 The best way to tax such rents is by a tax explicitly designed for that purpose, and indeed there is extensive experience with a variety of such ‘rent taxes’, including though not only 24 in the extractive industries.24 These taxes are not, however, invulnerable to profit shifting of various kinds, particularly in lower income countries.25 The ability to tax capital gains arising from changes in the value of such rents can therefore be a useful backstop when the implementation of such taxes is imperfect —though clearly inferior to an ability to effectively tax them as they accrue. • A third rationale for the right to tax gains on local immovable property is grounded in the benefit theory of taxation—i.e., that taxes are in the nature of payments for public services provided by government, which help maintain the value of local economic factors, including local immovable property. This argument is also sometimes used as a rationale for the corporate tax itself, and not especially compelling in that gains (or profits) may be a poor proxy for the benefits received. In economic terms, the concept of ‘immovability’ might be most meaningfully thought of as proxying for the possibility of location specific rents—with implications for how the term should be defined. This view suggests an expansive definition of ‘immovability’ capable of capturing at least the most likely sources of significant LSRs. This, however, is much easier said than done: the concept of LSR has not been sufficiently fully developed to be readily captured in legislative language. But while LSRs can be difficult to identify in general, in some cases they are reasonably obvious. They are often associated, in particular, with government-created rights— notably in the extractive industries and telecoms. Many of the cases that give rise to concerns in relation to indirect transfers revolve around rights that are explicitly tied to particular locations— with their value being made visible by the transfer itself.26 LSRs could also arise, for instance, from access to domestic markets, but this can be difficult to gauge and distinguish from rents associated with brand names or intellectual property. And of course the fact of a company being resident for tax purposes in a particular country clearly does not imply that its value substantially derives from LSRs arising there. What these considerations suggest is that any definition of immovability that proceeds by positive listing should anticipate, so far as is possible, likely sources of significant LSRs—and there are signs that, though not expressed in those terms, this is increasingly the case. Definitions 24 See for instance several of the contributions in Daniel and others (2010). Commented [ITIC11]: We would simply renew our 25 See for example, Beer and Loeprick (2017), who find evidence of extensive profit shifting in the sector, with causes for concern in citing this study, which we believe signs that developing countries are especially vulnerable. For evidence of their greater vulnerability to profit is highly flawed as explained in our comments to the shifting more generally, see for instance Crivelli, de Mooij and Keen (2016). first draft. Further, since over 80% of the data are from 26 Indeed, this is evident, to some degree, in the national responses to indirect transfer cases, which have focused the UK, Norway, the Netherlands, and Russia, citing it further to support that there are “signs that developing not on reducing the domestic taxation of direct transfers—as one would expect to be the case if there were no countries are especially vulnerable” is even more location-specific value to the underlying asset—but to seek to extend taxing rights. Without the existence of LSRs, suspect. Finally, to add the second citation, which that is, one would expect low taxation of indirect transfers to spur more intense tax competition in the treatment involves a study that explicitly excludes the extractive of gains on transfers rather than, as seems to be the case, the opposite. industries, without clearly stating that, seems inappropriate. 25 have come to more commonly include, for instance, not just the right to extract natural resources but the full range of licenses that may be associated with their discovery and development. There are two counterarguments to this emphasis on location country taxation: • Any gain reflects underlying income that the location country has chosen not to tax except as earned. It may be, however, that the capital gains charge is that country’s preferred method of taxing that income—or even, in the case of some developing countries, that when the law was drafted and treaties were signed, the authorities were simply not aware of or focused upon this issue. Otherwise, that future income is at risk of Commented [ITIC12]: This seems to sell short treaty non-taxation, whether (as discussed above) for timing reasons, or because of negotiators from developing countries--As noted in our overall comments, our experience is most treaty imperfections in other tax instruments, especially in developing countries. This makes negotiators are quite well-informed and we believe that taxation of gains a worthwhile, albeit very imperfect, additional tool. any choice not to tax is deliberate rather than inadvertent. It should be recognized, however, that countries may affirmatively choose—as some have—not to tax such gains on indirect transfers even where they could do so. 27 This may be seen, for instance, as a way to attract foreign investment or, in the case of natural resource taxation, part of an overall fiscal regime that imposes other taxes and fees such that adding taxation of indirect transfers would make the overall regime non- competitive. • The increased value of the entity sold may reflect in part managerial and other expertise contributed by the seller, beyond what has been recovered in managerial fees, royalties and other explicit payments. This suggests that the gain might therefore be properly taxed where the seller resides (so ensuring, in efficiency terms, that the seller’s decision as to the country in which it chooses to undertake such value-adding activities is not affected by the tax system). It may indeed be that there are company-specific as well as location-specific rents at work, and one might argue that the latter are naturally taxed where the company is resident. 28 The many countries operating dividend exemption schemes, however, have effectively indicated no desire to do so. More generally, how compelling this argument is may well depend on the circumstances of the case, being less plausible when the selling entity has few substantive functions. Moreover, the possibilities for structuring indirect transfers means there can be no presumption that the jurisdiction in which the gain is realized is that in which the underlying expertise or financing was ultimately provided. One might then think of some form of substance test, though this as always runs the risk of creating its own distortions, with resources allocated simply to meet the requirements of such a test and not for reasons of productivity. 27 The United States, for example, limits the reach of the Foreign Investment in Real Property Act (FIRPTA) to transfers occurring either directly, or at the first tier of ownership of the asset. Norway has affirmatively declined to tax such transfers in the resources sphere. 28 There are issues here, which we leave aside, as to the relevance of companies’ residence as a basis for taxation, 26 given the increasing disconnect between that and the residence of final shareholders. 27 The weight of argument creates a strong equity case for a presumptive primacy of source country taxing rights in relation to gains on immovable assets, defined to apply to sources of location specific rents. Efficiency A general principle of good tax design is that the tax system should, so far as is practicable, not distort investors’ decisions: unless there is good reason to do so, taxation should not lead businesses to change their commercial decisions.29 The reason for this is that any such changes mean that resources are being used in ways that are socially inefficient, but are privately profitable only because of taxation. 30 While efficiency considerations point firmly to the taxation of rents of various kinds, beyond that the literature on efficiency criteria to guide international tax arrangements provides few practicable insights. The prescription that rents are an efficient object of taxation is a very general one. As for other forms of taxation (that is, ones that may distort decisions), there is a large literature on their efficient design in international settings—focusing here on collective rather than national interests. This literature, however, has produced few (if any) agreed practicable policy prescriptions. For example, if the concern is to avoid distorting how parent companies choose to allocate their productive capacities across different countries then residence-based taxation is appropriate31 (since whatever was the most profitable choice before tax will also be the most profitable after tax). But if, on the other hand, the concern is to ensure equal within-country treatment of all potentially active companies, wherever they are resident, then source-based taxation is needed.32 Theory offers little guidance as to which view is the more appropriate from a collective perspective, 33 Two considerations, however, do point to significant efficiency considerations in this context.34 29Leaving aside the cross-border issues of interest here, several non-neutralities arise more generally in relation to capital gains taxation (in relation, for instance, to the distortions arising from taxing gains on realization rather than accrual). These are not addressed in the discussion here. 30 Strictly, it is worth noting, efficiency considerations relate only to the tax rules applied, and are in themselves essentially silent on which country should receive the associated revenue. Revenue sharing on indirect transfers seems a sufficiently remote possibility, however, for it to be ignored here. 31 Ignoring here the possibility of changing the place of corporate residence. 32 This latter is akin to the notion of ‘capital ownership neutrality’ advocated by Desai and Hines (2013). 33 See for example Appendix VII of IMF (2014). S 34 There are other dimensions of neutrality that should in principle also be considered. These include, for instance, the financing of the entity operating the underlying asset (Company A in Figure 1). To the extent that the dividends it pays are taxed more heavily than are capital gains on its sale, this gives a tax incentive to finance the operations of that entity by retaining earnings rather than by injecting new equity—which might, for instance, imply slower growth of its operations (Sinn, 1991) This would be alleviated by taxing dividends and capital gains at the same 28 The most fundamental efficiency argument for the country in which assets are located to tax both indirect and direct transfers is as a way to tax LSRs—albeit imperfectly. The Commented [ITIC13]: We respectfully suggest this is a preferability in principle, but limitations in practice, of explicit rent taxes were stressed above. bit of a stretch to continue to make this argument. What about a case where there is no transfer, direct or Auctions are another possible tool for rent extraction, and have been widely used, for instance, for indirect. That is the comparison that should be made. petroleum rights; but these can be subject to problems of asymmetric information and thin markets (being rarely used for instance, in relation to hard minerals).35 On efficiency grounds, as well as those of inter-nation equity, taxing gains can be a useful supplementary device where—as in many developing countries—other methods of taxing LSRs are imperfect. One natural requirement for neutrality is that direct and indirect asset transfers be treated identically for tax purposes. That is, transferring an asset or transferring shares deriving their Commented [ITIC14]: We strongly agree that the value from that asset, to the extent that they represent the same transfer of ownership, should— analytical approach of testing against what would happen in a direct case is proper. If the result of all else equal—attract the same tax treatment. Otherwise there will be an incentive to distort arguably extending the scope of taxation outside of the transactions as a result of the differences. country is to achieve a different result from the same case in a direct asset transfer, then arguably something Given the current norm—that the country L in which immovable assets are located has the is wrong with that approach to taxing OITs—with one right to tax direct transfers—such neutrality is most likely to be achieved by taxation of exception. A country may decide to tax a direct transfer while not indirect transfers in L. In principle, neutrality along this dimension could instead be achieved by taxing an indirect transfer—the notion being that the the location country forgoing any claim to tax either direct or indirect transfers, leaving this instead increased complexities in taxing OITs, given only a to the country in which the seller is resident. This, however, simply seems unlikely to happen—and timing benefit, create a call on administrative resources that is not justified given the present value of any it may be undesirable that it should, if this is a less distorting source of revenue for L than the benefit. available alternatives. That leaves the simplest route to neutrality: taxation of indirect transfers by Commented [ITIC15]: But not necessarily “practicality”, the country in which the asset is located. as noted in the comment above. And again, the income will still be taxed—just as earned versus up front. Assessment The arguments are not all in one direction, but on balance the analysis above suggests it to be appropriate that the location country have the right to tax capital gains associated with transfers of immovable assets to the country in which the assets are located, regardless of whether the transferor is resident there or has a taxable presence there (although the decision of whether or not to exercise that right is a different analysis).36 In equity terms, this mirrors the generally recognized right in relation to direct transfers; in efficiency terms, it provides one route to the taxation of location specific rents—highly imperfect, but potentially valuable when preferred instruments are unavailable or weak—and fosters neutrality between direct and indirect transfers. General agreement on the scope of such a right—and well-developed models of rate. That does not necessarily mean that both types of income should be taxed by the same country, but is most naturally achieved by the location country taxing gains just as it does dividends. How significant a concern this is, however—compared for instance to what is often a very marked tax preference for debt finance—is unclear. 35 On both rent taxes and auctioning in the extractive industries, see Daniel, Keen and McPherson (2010). 36Others have reached a similar conclusion. Cui (2015, p.154), for instance, takes the position that “too much of the international tax discussion recent decades has been centered on whether non-residents should be taxed on 29 capital gains, rather than how they are to be taxed.” 30 implementation—would help to avoid uncoordinated measures that jeopardize the smooth and consensual functioning of the international tax system and give rise to tax uncertainty. The rationale, in terms of economic principle, for limiting this treatment to immovable assets is unclear. Much current practice is already sharply at odds with this; while primary taxing rights are frequently given to the source country in relation to immovable property but to the residence country in the case of equity participation in other businesses, there are some notable exceptions, such as the cases of Peru and India discussed later. Indeed, Article 13(5) of the UN Model Tax Convention (MTC), discussed in section IV below, extends location country taxation up one tier of ownership, to gains on any company shares.37 What emerges clearly is the importance to the location country of clearly defining ‘immovable assets’. Considerations of inter-nation equity, efficiency and practicability converge to suggest that the scope could include all assets with the potential to generate significant location-specific rents and over which the government can exercise sufficient control to ensure collection. However, a country may choose a narrower definition. Moreover, while the location country may choose not to exercise its right to tax OITs, experience—exemplified by the cases discussed in the next section—shows that not doing so can provoke intense domestic dissatisfaction unless care is taken to communicate the rationale for not doing so (including that there are other revenues, and that total revenues may actually be enhanced over the life of the project by not taxing an OIT). These assets are Commented [ITIC16]: This and the next point are commonly highly visible, with strong salience for the general public—perhaps reflecting a highly more political versus economic issues, and the reason for some suggested language additions to balance this publicized resource discovery, for example—and are often finite resources owned by the nation out. (in the case of extractive resources) and/or created by the government (in the form of licenses or other rights). And, as will be seen shortly, the sums at stake can be large. This underscores the need for politicians and commentators to be well informed and able to discuss the pros of not taxing an OIT, should that be the country’s choice. If public This dissatisfaction, nevertheless, is significant, it can lead to unilateral legislative actions—which may (and do) differ across countries—that exacerbate tax uncertainty, with harmful effects for investors, taxpayers and governments. 31 37 More precisely, this provision allows state L to tax the sales by non-residents of shares in companies resident in L There is no comparable provision in the OECD MTC. 32 THREE ILLUSTRATIVE CASES Three highly publicized OITs are described in Boxes 2 to 4: 38 Vodafone’s purchase of a substantial interest in a mobile phone operator in India, the indirect sale of the Peruvian oil company Petrotech Peruana, and the indirect sale by Zain of various assets in Africa including a mobile phone operator in Uganda.39 All of these transactions have (at least so far, as appeals continue) raised the issue of whether multinational groups can ultimately escape taxation of gain on a OIT in the country in which the underlying assets were located, and ensure no or light taxation of the gain elsewhere, by arranging that the transfer be effected as a sale by an entity not resident where the subsidiary holding the underlying asset is located. Box 2. India—The Vodafone Case In 2006, Vodafone purchased Hutchison’s participation in a joint venture to operate a mobile phone company in India (the owner of an operating license), for nearly US$11 billion. This transfer was accomplished by Hutchison, a Hong Kong-based multinational, selling a wholly owned Cayman Islands subsidiary holding its interest in the Indian operation to a wholly owned subsidiary of Vodafone incorporated, and for tax purposes resident, in the Netherlands. The transaction thus took place entirely outside India, between two non-resident companies.1 The Indian Tax Authority (ITA) sought to collect US$2.6 billion tax on the capital gain realized by Hutchison on the sale of the Cayman holding company. Given that Hutchison no longer had assets in India after the transaction, the ITA sought to collect the tax from the purchaser, Vodafone’s Dutch subsidiary, arguing that it had the obligation to withhold the tax from the price payable to the seller. This sparked a protracted court case, with the Supreme Court of India ruling in 2012 in favor of the taxpayer. The Supreme Court denied the ITA’s broad reading of the law to extend its taxing jurisdiction to include indirect sales abroad, though it took the view that the transaction was in fact the acquisition of property rights located in India. The government of India subsequently changed the law to allow taxation of offshore indirect sales and tried to apply the new provision retroactively, in a second attempt to collect the tax from Vodafone’s Dutch subsidiary. The legality of a retroactive effect of the law was subsequently submitted to arbitration by the taxpayer under the India-Netherlands Bilateral Investment Treaty1. (The treaty was unilaterally terminated by India in December 2016, but this does not affect ongoing disputes). Several years after their appointment, arbitrators selected by the parties finally agreed on choosing a chairman of the tribunal. 1 The relevance of this tribunal on tax matters is still a matter of dispute. is true that these are three of the, if not the, largest and best known instances of this issue. However, for that 38 It reason their details are also the most public, and the best explored—and they have had perhaps the greatest impact on future actions of the host countries and the broadest ramifications. But experience shows that many other countries face this issue, albeit with less spectacular publicity. 39 Other examples are in IMF (2014) and Burns, Le Leuch and Sunley (2016). 33 Box 3. Peru—The Acquisition of Petrotech In 2009, Ecopetrol Colombia and Korea National Oil Corp purchased a Houston-based company (Offshore International Group Inc.) whose main asset was Petrotech Peruana (the license-holder), a company incorporated and resident in Peru and the third largest oil producer there, for approximately US$900 million, from Petrotech International, a Delaware incorporated company. Since Peru’s income tax law at the time did not have a specific provision taxing offshore indirect sales, the transaction remained untaxed there. The potential foregone tax revenue for Peru was estimated at US$482 million. Petrotech i International, a resident of the U.S.A., would be taxed in the U.S. on the corresponding capital gain. The case triggered a Congressional investigation in Peru that eventually led to a change in the law. Currently, all offshore indirect sales of resident companies are taxed in Peru, regardless of the proportion that immovable property belonging to the Peruvian subsidiary may represent in the total value of the parent company (Article 10, Income Tax Act, Peru; see Box A.1, Appendix B,). Some limitations apply: the portion of the parent company subject to sale must derive its value at least 50 percent from Peruvian assets, and at least 20 percent of the Peruvian assets must be transferred in order for the transaction to be taxable in Peru. Box 4. Uganda—The Zain Case In 2010, a Dutch subsidiary of the Indian multinational Barthi Airtel International BV purchased from Zain International BV, a Dutch company, the shares of Zain Africa BV (also a Dutch company) for US$10.7 billion, which owned in turn the Kampala-registered mobile phone operator Celtel Uganda Ltd. (among other investments in Africa). 40 The Uganda Revenue Administration (URA) held Zain International BV liable for the corresponding capital gains tax, amounting to US$85 million. Uganda’s Appeals Court ruled—in sharp contrast to the decision of the Supreme Court of India in Vodafone—that the URA does have the jurisdiction to assess and tax the offshore seller of an indirect interest in local assets (overturning an earlier ruling by the High Court of Kampala.) 41 However, the taxpayer interprets the tax treaty between Uganda and the Netherlands as protecting the Netherlands’ exclusive right to tax such transaction. This is an issue of some potential significance since some anti-avoidance rules in domestic law could be viewed as supplementary to the treaty, not an override; it is currently unresolved. 40 Zain International BV belongs ultimately to the Zain Group, whose main shareholder is the Kuwait Investment Fund. 41 See Hearson (2014) and The East African “Court gives URA nod to seek taxes on sale of Zain assets in Uganda”, September 13, 2014, at http://www.theeastafrican.co.ke/news/URA-taxes-on-sale-of-Zain-assets-in-Uganda/- /2558/2451578/-/item/0/-/6hm2he/-/index.htmlhttp. 34 The amount at stake is in all three cases very large: that in Zain is in the order of 5 percent of total government revenue (and, for example, nearly 50 percent of public spending on health); that in Vodafone is around 2 percent of central government revenue (and almost 8 percent of all annual income tax revenues).42 Another common feature is that the indirectly-transferred asset in question was a business whose value derived from a concession granted by the government of the country in which the underlying asset is located. Value is thus manifestly tied to particular jurisdictions, and largely consists of what are recognizably location-specific rents deriving from some government-issued license. In all three cases,43 the country in which the underlying asset was located lost in court—or at least has not yet clearly won. The reasons differed, however: insufficiency of the domestic income tax law to reach such transfers in India and Peru, potential override of a treaty (one that does not contain provisions along the lines of Article 13.4, discussed in Section IV below) in Uganda. In all cases, governments and many civil society organizations argued that developing countries had been denied (or had inadvertently denied themselves) a fundamental (and substantial) source of revenue. This was especially problematic politically when it could be shown Commented [ITIC17]: If governments and especially that the subsidiary being indirectly sold had previously paid little, if any corporate income tax, as civil society organizations had more correctly explained the issue as one of only timing, the "public outcry" was pointed out in a congressional investigation on Petrotech ordered in Peru. Public outcry in would no doubt have been much less. several of these cases was considerable. In Peru, for example, the transaction became linked with corruption scandals, leading to the dismissal of Prime Minister and Cabinet. The cases show that the location country may well respond to defeat in court by quite sweeping policy changes. India, for instance, not only changed its domestic law to bring OITs into tax44 but sought to apply this retrospectively to 1962 (the date of the current income tax act). Peru Commented [ITIC18]: If the draft is taking the position and Chile amended their domestic laws to bring into tax offshore transfers related to all assets that governments should have the "right" to tax OIT's based at least in part on equity, we believe an equally located in their countries—not just those deriving value from immovable property located there. strong position should be taken against the right of a Such unilateral responses are understandable and may reflect different legal systems in different country to retrospectively apply a change to the law to countries. Not least because of their diversity, however, they risk introducing even more tax OIT's. Taxpayers must be able to rely on existing law to conduct their activities. Retrospective application of incoherence and uncertainty in international taxation than already exists. the change in law in the cited case would prevent the company from having appropriately negotiating the sale and if known at the time of sale might well have resulted in the sale not taking place. 42 Other examples are given in Appendix VI of IMF (2014). 43 In other cases—Heritage in Uganda, Las Bambas in Peru, for instance—tax has been recovered by the location country. 44 Cited by Cui (2015, p.146), the amendment reads: “any share or interest in a company or entity registered or incorporated outside India shall be deemed to be…situated in India, if the share or interest derives, directly or indirectly, its value substantially from the assets located in India.” 35 TAX TREATIES AND OFFSHORE INDIRECT TRANSFERS This section reviews the treatment of OITs envisaged in the model tax treaties,45 reports on an empirical analysis of current treaty practices, and describes the 2017 Multilateral Convention to Implement Tax Treaty Related Measures to Prevent Base Erosion and Profit Shifting. This review makes clear that for indirect transfers, in relation to assets defined as “immovable” under the models, primary taxing rights are accorded by these agreements to the location country. A. OITs in the Model Treaties Model treaty practices, for both moveable and immovable assets, are summarized in Table 1. (The practices of specific countries, of course, may be quite different) Table 1. Allocation of Taxing Rights Between Countries on Transfers of Assets Under Model Tax Conventions Type of Type of Transfer Property Offshore Direct Offshore Indirect If more than 50% of value of transfer is Immovable (directly or indirectly) derived from Assets in Country L immovable assets: Country L1 Country L Otherwise: Country R Seller has PE in Country L to which the assets are allocated: Country L Movable Assets in “Substantial” Ownership Interest1: Country L UN MTC (Article 13.5): Country L OECD MTC: Country R Other Cases: Country R1 Legend: L = Country where underlying asset is located R = Country where seller resides Where two countries have a claim to tax the capital gain arising from the transfer, they may by treaty establish which has the primary right to tax, with the appropriate relief mechanism in the other in order to avoid double taxation. In the absence of a treaty to that effect, double 45 Toledano and others (2017) rightly note too the need to ensure that countries’ intended treatment of OITs is not inadvertently undermined by provisions in bilateral investment treaties or concession agreements. 36 taxation may occur, though taxpayers would presumably avoid structuring transactions in ways subject to such treatment. Both MTCs provide that direct transfers of immovable property may be taxed by the country in which that property is located (Article 13(1); identical language). Gains on indirect transfers are dealt with in Article 13(4) of each MTC.46 In the OECD version, prior to its 2017 update,47 this reads: Gains derived by a resident of a Contracting State from the alienation of shares deriving more than 50 percent of their value directly or indirectly from immovable property situated in the other Contracting State may be taxed in that other State. The 2017 update of the OECD MTC includes the following amended version of Article 13(4): Gains derived by a resident of a Contracting State from the alienation of shares or comparable interests, such as interests in a partnership or trust, may be taxed in the other Contracting State if, at any time during the 365 days preceding the alienation, these shares or comparable interests derived more than 50 percent of their value directly or indirectly from immovable property, as defined in Article 6, situated in that other State. In the 2011 version of the UN treaty, the core provision of Article 13(4) is that: Gains from the alienation of shares of the capital stock of a company, or of an interest in a partnership, trust or estate, the property of which consists directly or indirectly principally of immovable property situated in a Contracting State may be taxed in that State. A sub-clause defines ‘principally’ by a 50 percent threshold test similar to that in the OECD version.48 The 2011 UN version continues: In particular: (a) Nothing contained in this paragraph shall apply to a company, partnership, trust or estate, other than a company, partnership, trust or estate engaged in the business of management of immovable property, the property of which consists directly or indirectly principally of 46 The UN MTC has included this provision since its inception in 1980, and the OECD MTC has included one since 2003. 47 Superseded model treaty provisions remain relevant in that they may be reflected in still-current treaties. 48 The U.N. text parallels wording in U.S. domestic law on indirect sales of immovable property and U.S. commentaries to the OECD MTC. 37 immovable property used by such company, partnership, trust or estate in its business activities. (b) For the purposes of this paragraph, “principally” in relation to ownership of immovable property means the value of such immovable property exceeding 50 per cent of the aggregate value of all assets owned by the company, partnership, trust or estate. The 2017 version of the UN Model has the same language as the 2017 version of the OECD Model, reflecting a blending of the previous provisions from both models, including adaptations designed to prevent abuse. In both MTCs, the definition of immovable property is first found in Article 6: The term “immovable property” shall have the meaning which it has under the law of the Contracting State in which the property in question is situated. The term shall in any case include property accessory to immovable property, livestock and equipment used in agriculture and forestry, rights to which the provisions of general law respecting landed property apply, usufruct of immovable property and rights to variable or fixed payments as consideration for the working of, or the right to work, mineral deposits, sources and other natural resources; ...” Clearly, this leaves much scope for more precise definition in domestic law, which varies quite widely.49 The basic rule for indirect transfers of immovable property has thus been quite similar in the two MTCs, and is now the same.50 This similarity reflects a commonality of broad intent in the two provisions. Both allocate the primary right to tax to the location country when a transfer by a 49 Krever (2010) notes that “[a]s a general rule, civil law jurisdictions seem content to limit the meaning of immovable property, at its narrowest going little beyond tangible real estate, while natural-resources-rich common law countries have the broadest definition” (p.223). As an example of an expansive approach, he also cites (p.237) the definition of “taxable Australian property” as including ”any authority, li cense, permit or right under an Australian law to mine, quarry or prospect,…a lease of land that allows the lessee to mine, quarry or respect,….an interest in such an authority, license, permit, right or lease…and any rights that are in respect of buildings or other improvements…on the land concerned or are used in conjunction with operations on it.” See also Box 9 below and the discussion there. 50 Until recently, the UN version was broader in applying to forms of title other than shares (the text to this effect, emphasized above, having been introduced in 2001). Under Action 6 of the BEPS work, however, agreement was reached to amend the OECD MTC to eliminate this difference. One point relating to both is that the 50 percent test relates to the proportion accounted for by the immovable property in the total value of the title being sold, not the share of the gain: so taxing rights may be allocated to a country other than that in which the majority of the gain arises. Some countries see this “blunt” aspect of the rule as an advantage, in discouraging abuse, and as minimizing potentially complex disputes on valuation issues as to property distributed internationally. In any case, domestic law may limit the liability to profits proportionate to the amount of immovable property in the taxing countries, and therefore may not seek to fully exert the treaty taxing rights. Location countries may want to consider the balance of complexity versus “bluntness” in drafting the relevant rules on valuation and apportionment. 38 non-resident occurs in the other state,51 but restrict this possibility to the specific case when the transfer is directly or indirectly of ‘immovable property.’52 The exclusion from taxation of indirect transfers involving certain types of entities whose property consists principally of immovable property used by them in their business activities as provided in Article 13(4)(a) of the UN MTC prior to 2017 was potentially limiting, but has now been changed. 53 Article 13(5) of the UN MTC (which has no parallel in the OECD MTC) extends the reach of offshore taxation beyond immovable property, however defined. Before 2017, Article 13(5) read as follows: “Gains, other than those to which paragraph 4 applies, derived by a resident of a Contracting State from the alienation of shares of a company which is a resident of the other Contracting State, may be taxed in that other state if the alienator, at any time during the 12-month period preceding such alienation, held directly or indirectly at least percent (the percentage is to be established thorough bilateral negotiations) of the capital of that company.” The 2017 UN Model includes the following changes to Article 13(5) (changes highlighted): Gains, other than those to which paragraph 4 applies, derived by a resident of a Contracting State from the alienation of shares of a company, or comparable interests, such as interests in a partnership or trust, which is a resident of the other Contracting 51 Since the articles place no explicit limits on that power to tax, it must be interpreted that the country may do so even if the indirect sale takes place in the other State 52 The OECD MTC considers other important exclusions, which relate more to implementation complexities than to conceptual issues. For example, it could exclude from taxation alienators holding below a certain minimum level of participation in the entity; or the sale of shares of companies listed in an approved stock market, or gains from transfers of shares in a corporate reorganization. Commentary 28.7 to OECD MTC; OECD (2010). 53 The Commentary to the 2011 UN Model did not address the interpretation of this exclusion and so the scope of its application was unclear. At worst, exempting from tax in the location country an indirect transfer of immovable property (complying with the more than 50 percent value rule) when it involves property that is being principally used in the business activities of the entity sold—including for example a hotel or mine—as Article 13(4)(a) of the UN model treaty may be argued to do—could have gone too far in limiting taxing rights, especially for developing countries, as it could have involved sectors in which sizeable economic rents are concentrated. An alternative interpretation commonly put forward was that the “business activities” exclusion only applied where the non - resident seller of the shares used the relevant immovable property in its own business activities as compared to that property being used by the asset owning entity whose shares are being transferred. Due to this uncertainty, the UN Manual for the Negotiation of Bilateral Tax Treaties between Developed and Developing Countries itself notes that “...in practice, this provision is not commonly found in treaties negotiated by developing countries... since gains from the alienation of interests in entities that own and run mines, farms, hotels, restaurants, and so forth, are not covered by this paragraph.” United Nations (2016). 39 State, may be taxed in that other State if the alienator, at any time during the 365 days preceding such alienation, held directly or indirectly at least per cent (the percentage is to be established through bilateral negotiations) of the capital of that company. This allocates to country L taxing rights over the gain derived by a non-resident of country L from the disposal of shares (or comparable interests) of a company, partnership or trust that is itself resident of country L. However, this treaty provision extends only to shares or comparable interests in entities that are resident of L. That is, it only applies to offshore direct ownership of such entities. For that reason, while Article 13(5) may help address certain tax avoidance arrangements (e.g. certain dividend-stripping or change of residence strategies),. it is not suitable as a provision to ensure the source taxation of gains on indirect transfers. According to the UN Manual for the Negotiation of Bilateral Tax Treaties between Developed and Developing Countries, treaty practice varies with regard to the use of Article 13(5): some countries explicitly exclude gains on listed shares, others restrict the scope to gains realized by individuals who were previously residents of the source State and many countries do not include this provision at all in their treaties. . B. Article 13.4 in Practice54 About 35 percent of all DTTs include Article 13(4),55 with an explicit reference to gains that derive their values indirectly from immovable properties (Figure 2). The percentage of DTTs that contain a provision relating to capital gains on shares deriving value from immovable property, counting also those without the word ‘indirectly,’ is about 60 percent (Wijnen and de Goede, 2014).56 The inclusion of Article 13.4 is slightly less common in DTTs that involve low or lower-middle income countries, at around 31 percent (Figure 1). This is a noticeably lower proportion than identified in previous work by Hearson (2016), who finds that, in a sample of 537 DTTs involving developing countries, about 51 percent contain a provision regarding capital gains from shares in 54 The details underlying the analysis in this section are in Appendix C. Data are as of 2015. 55 By “including Article 13.4” is here meant, more precisely, the inclusion of an article akin to 13.4 of the model treaties with an explicit reference to ‘indirectly.’ 56 Wijnen and de Goede (2014) look at about 1,800 DTTs tax treaties and amending protocols concluded during 1997-2013. 40 immovable property. The difference reflects the use here of a larger sample (close to the universe) and the imposed search criterion for ‘indirectly’. The likelihood that Article 13.4 is included in a treaty57 is significantly: Lower if one of the treaty partners is a resource-rich low-income country, by about 6 percentage points. This is a striking finding, and in light of the discussion above, a troubling Commented [ITIC19]: How is this conclusion one. Examples of DTTs that involve resource-rich low-income countries and omit Article 13.4 reached—since “striking” connotes both “highly significant” and “bad?” This could have been a include Uganda-Mauritius (concluded in 2003), Malawi-Norway (2009), Trinidad and Tobago- conscious choice by a country given the benefits it sees Brazil (2008). to promoting free transferability and the overall mix of its fiscal regime. Lower if one of the treaty partners is a low tax jurisdiction,58 by about 13 percentage points. This too is troubling, in the sense that these are likely to be cases in which the opportunity to avoid tax in the location country by transferring indirectly is most attractive. Commented [ITIC20]: It is not avoided—only deferred. Higher, the greater is the difference between the rates at which capital gains are taxed in the treaty partners.59 This, on the other hand, suggests an awareness of the high tax treaty partner to the opportunities for avoidance through OITs. The effect, though, is quite modest: Commented [ITIC21]: deferrral a 10-percentage point difference between capital gains tax rates increases the probability of including Article 13.4 by only about 4-percentage points, on average. Increasing over time. Almost no countries with multiple treaties have Article 13.4 in all of them—implying a vulnerability through OITs structured to exploit treaty provisions even where the location country L was evidently aware, of the possibility of imposing tax on such gains. As shown in Commented [ITIC22]: This comment seems rather Figure 3, several countries do not have Article 13.4 in any of their treaties (it appears in none of unbalanced (e.g., using terminology like “vulnerability” and “exploit”). Isn’t it just as likely that this provides Gambia’s seven, for example) or have it only in relatively few (less than 20 percent, for instance, for solid evidence of conscious decisions by countries not Kuwait, Nigeria, and Papua New Guinea). to tax OITs? About 22.5 percent of the universe of treaties include Article 13.5 of the UN model treaty. There is no clear observed pattern linking the inclusion of Articles 13.4 and 13.5: some treaties include only one, some contain both. 57 This is of course a backward-looking exercise, so does not necessarily speak to the likelihood of future inclusion. 58 In the sense of being included in the list of Hines and Rice (1994). 59 There is considerable heterogeneity in capital gains tax rates across countries. There are 35 countries that charge no taxes on capital gains of corporations (e.g., Hong Kong, Malaysia, Singapore, and some resource-rich countries such as Bahrain and the UAE). At the other end of the distribution, there are 10 countries that impose a rate of 35 percent (e.g., Argentina and the United States) and the highest rate is 36 percent in Suriname. 41 Figure 2: Article 13.4 in DTTs All DTTs and 'Article 13.4' DTTs of Low Income Countries and 'Article 13.4' 3000 1000 Cumulative Number of Treaties Cumulative Number of Treaties 800 2000 400 600 1000 200 0 0 1947 1957 1967 1977 1987 1997 2003 2009 2014 1947 1957 1967 1977 1987 1997 2003 2009 2014 UN, OECD, or a similar version of articel 13.4 Total UN, OECD, or a similar version of article 13.4 Total Note: low-income countries are lower middle-income resource-rich country as defined by the income classification of the World Bank. Annex A describes the DTTs. Figure 3: Proportions of Countries’ DTTs including Article 13.4 40 30 Number of Countries 20 10 0 0 10 20 30 40 Shares of Treaties with Article 13.4, Percent C. OITs and the Multilateral Convention The Multilateral Convention to Implement Tax Treaty Related Measures to Prevent Base Erosion and Profit Shifting (also known as the Multilateral Instrument or “MLI”) is the outcome of BEPS Action 15, which called for the development of a multilateral instrument to implement efficiently BEPS tax treaty related measures. Accordingly, the Multilateral Convention modifies existing bilateral tax treaties between Convention parties to meet the BEPS treaty-related minimum standards, i.e., the prevention of treaty abuse under BEPS Action 6 and the 42 improvement of the dispute resolution mechanisms under BEPS Action 14. At the same time, the MC facilitates the implementation of other tax treaty measures developed in the BEPS Project, e.g. measures against artificial avoidance of permanent establishment status through commissionnaire arrangements. For countries party to the Convention lacking a provision in their existing tax treaties equivalent to Article 13(4) of the 2017 OECD MTC, Article 9(4) of the MC effectively incorporates such a provision into their tax treaties, which are modified by the MC under international law, unless the country opts out of Article 9(4). By opting for this provision, the jurisdiction in which immovable property is situated would be allowed to tax capital gains realised by a resident of the treaty partner jurisdiction from the alienation of shares of companies that derive more than 50 per cent of their value from such immovable property. For countries that already have in their tax treaties a provision related to the taxation of capital gains realised from the alienation of shares, the MC offers two options for enhancing it. First, Article 9(1) of the MC allows parties to modify their covered tax treaties by introducing a testing period into Article 13(4). Accordingly, Article 13(4) will refer to a period of 365 days preceding the alienation of shares for determining whether the shares derive their value principally from immovable property. Additionally, Article 9(1) of the MC offers the parties the possibility to enlarge the scope of Article 13(4) of the OECD MTC by expanding the type of interests covered. As a result, interests comparable to shares, such as interests in a partnership or trust, would be also included in the wording of Article 13(4)60. Unlike a protocol that amends a single tax treaty, the MC modifies all existing tax agreements identified by the various treaty partner countries signing the MC. In particular, the provisions of Article 9 of the MC shall apply, when relevant, in place of or in the absence of provisions of the relevant tax treaties on gains from the alienation of shares or other comparable interests, unless the signatory has opted not to apply Article 9. Notably, merely signing the MC does not mean that the signatory’s tax treaties will effectively include the provisions of Article 9 of the MC. The MC allows parties signing the convention to reserve their right not to apply any of the provisions included in Article 9(1) or not to include the language of Article 9(4). 61 In addition, as explained in Section V of this toolkit, the location country must have enabling provisions in its domestic law for effectively imposing tax with respect to a capital gain derived from an OIT. Beyond the specific provisions included in Article 9, Part III of the MC introduces additional measures to prevent treaty abuse that may also be relevant to preserve the taxing rights of 60 In the 2014 version of the OECD MTC, the option to cover gains from the alienation of interests in other entities such as partnerships or trusts is provided in paragraph 28.5 of the Commentary on Article 13. 61 In the first round of signing of the MC, in June 2017, half of the participating 69 countries reserved on the relevant provision—meaning that they would not include it in renegotiating their treaties. 43 the location country in an OIT situation. In particular, Article 7 of the MC contains the so-called principal purpose test and the limitation on benefits provisions (‘LOB’). Those measures allow parties to a covered tax treaty to deny treaty benefits if obtaining those benefits was one of the principal purposes of the relevant transaction or when a resident of a contracting state does not meet the conditions established under the LOB. This toolkit provides further analysis of the application of general or specific anti-avoidance rules to OITs in Section 5(A). Finally, the MC may also be a future means of addressing further developments in the taxation of OITs. In this respect, the design of the MC offers the parties the possibility to amend the instrument in the future through the mechanism established in its Article 33. This mechanism would thus allow the inclusion of new tax treaty measures to safeguard the taxing rights of the source country in relation to OITs. To date, the MLI will effectively modify a substantial number of tax treaties to add the language of Article 13(4). On the basis of Provisional MLI positions, it is expected that Article 9(4) will effectively be added to more than 115 covered treaties. This number is expected to increase in the future as new signatories opt for the provisions of Article 9(4) of the MLI. At the same time, the MLI will modify around additional 70 covered treaties by bringing them into line with the new wording of Article 13(4). 44 IMPLEMENTATION CHALLENGES AND OPTIONS The fundamental legal/structural issue with OITs is that contractually the underlying asset does not change hands, so there is formally no capital gain directly realized in respect of it in the country where it is located. What changes hands is stock or a comparable interest in an entity that holds the asset either directly or indirectly, but the stock or interest is—in the cases in question— held and transferred in another country, either in the country of residency of the seller or in a third country. Various situations can then occur. This section outlines the implementation options and challenges associated with the taxation of OITs, should a country choose to do so. It focuses on the taxation of gains relating to immovable property (including for instance mining rights) situated in a source country (location country). This section also offers some guidance in relation to the taxation of gains relating to a substantial shareholding in a company resident in the location country. In both model treaties, as seen above, a taxing right arises when over 50 percent of the value of the transferred stock or interest derives from immovable property in the location country. In order to determine whether the value of the interest is principally (more than 50 percent) derived from that immovable property, a comparison is ordinarily required to be made of the value that the immovable property (relevant asset) bears to the value of all the property owned by the entity (all assets) without taking into account debts or other liabilities. Even where this test is met, since an OIT occurs outside the location country (between a non- resident seller and a buyer who may also be a non-resident), the location country may face significant difficulties in collecting the tax.62 This section of the report aims to provide practical guidance in relation to the design of possible legal instruments for imposing a tax liability on the gain, which will subsequently enable the tax authority to enforce and collect the tax. The legal instruments outlined in the section consist mostly of sample domestic legislative provisions. This is because, even if any relevant treaties preserve the location country’s taxing rights along the lines of model Article 13(4), it is essential that the location country has enabling provisions of this kind in its underlying domestic law. A. Overview of Legal Design and Drafting Principles: Two Models It is critically important that, if a country desires to tax OITs, the domestic tax law framework contain an indirect transfer taxing rule as well as appropriate enforcement rules to collect the resulting liability. A treaty cannot create such taxing rights or enforcement mechanisms if they do not exist in domestic law. 62 Toledano and others (2017) cite examples in which the authorities of the location country simply did not know that an indirect transfer had occurred (in one case, even though the government held a 20 percent stake in the 45 transferred entity). 46 There are a number of key design aspects to consider when seeking to tax an OIT. Each aspect has its own legal design considerations and sensitivities in the context of a location country with a tax system that conforms to the existing international norms of residence and source. The key aspects can be summarized as follows: 1. Designing the tax liability rule: There are two common models in this regard: (a) Model 1 (taxation of a deemed direct sale by a resident): This model seeks to tax the local entity that directly owns the asset in question, by treating that entity as disposing of, and reacquiring, its assets for their market value where a change of control occurs (e.g. because of an offshore sale of shares or comparable interests). Model 2 (taxation of the non-resident seller): This model seeks to tax the non-resident seller of the relevant shares or comparable interests via a non-resident assessing rule. Model 2 must be supported directly or implicitly by a source of income rule. This source of income rule provides that a gain is sourced in the location country when the value of the interest disposed of is derived, directly or indirectly, principally from immovable property located in that country. A source rule relating to gains from the disposal of other assets may also be considered, as discussed above in Section II —including substantial shareholdings in resident companies. A source of income rule may be further supported by a taxable asset rule dealing with such matters as whether taxation only applies to disposals of substantial interests (such as a 10 percent shareholding rule) to exclude from the scope of tax changes in ownership of portfolio investments. The rule can also prescribe whether the entire gain will be subject to tax when the value of the indirect interest is less than wholly derived from local immovable property or, alternatively, whether the gain will be subject to tax on a pro rata basis. Each legal design option under this rule is discussed and explained further below. 2. Designing the enforcement/collection rules. These rules are critical as they support the enforcement and collection of the resulting tax liability. They can include one or more of the following: (a) Notification/reporting and information exchange mechanisms (e.g. domestic reporting requirements supplemented, where appropriate, by international information exchange arrangements); (b) Withholding tax mechanisms (e.g. on payment of the purchase price); (c) Mechanisms imposing a tax payment obligation on a relevant local entity (e.g. as agent of the non-resident seller); and/or (d) Other legal protections such as restricting the registration, renewal or validity of relevant underlying assets (e.g. extractive licenses) unless applicable notification requirements have 47 been met and/or until it is demonstrated that either: no tax is payable; the relevant tax has been paid; or satisfactory arrangements have been made for the payment of that tax. A General Anti-Avoidance Rule (GAAR) could be applied as a rule of last resort to tax a gain from an OIT in appropriate circumstances. However, this sort of rule can be quite difficult for countries with weak administrative capacity to apply successfully. Some countries have adopted taxing mechanisms that operate in a similar way to a specific anti-abuse rule by seeking to, in effect, collapse the multiple tier holding structure and treat the ultimate non-resident seller of the interests as the seller of the local assets, who realizes a gain with a local source. This is the type of approach adopted in China. The successful application of such a rule ultimately depends on: (i) the design and drafting of the particular anti-abuse rule, which is often less rules-based and more discretionary in its application; and (ii) the capacity of the tax authority to appropriately apply such a specific anti-abuse rule in an even handed and predictable way.63 This type of rule would only reach the gain in question if intentional tax avoidance regarding the transaction could be shown. Such rules would therefore not provide that the gain in question should be taxed as a matter of principle on the basis of a substantive right to tax in the location country and would be much more limited in scope. Such rules may, however, be practical in terms of effectively limiting taxation to Commented [ITIC23]: It would be a conscious decision a certain cutoff point up an ownership chain, since it is likely that the higher up in the chain the of the country to go this way—it is not an admission of not having the taxing right if it chooses not to tax transfer occurs, and the broader the scope of the assets and subsidiaries affected, the less likely under its statutes. the transfer is motivated to avoid a particular country’s tax and the more likely it is motivated by general business concerns. Some type of cutoff point could also be considered even where a country decides to enact specific taxation rules with respect to taxation of OITs. The sample domestic legislative provisions set out throughout this section are general in nature and in the form of simplified rules-based legal provisions. Importantly, they do not take into account the individual circumstances of any particular tax system, nor have they been adapted for all relevant circumstances (e.g. corporate reorganizations) and other common concessions that typically apply (e.g. under a capital gains tax regime, in circumstances when it is considered appropriate to defer the recognition of taxable gain). The simplified legislative provisions also do not deal comprehensively with more complex issues such as minority Commented [ITIC24]: This is a real illustration, and in shareholders, joint venture arrangements, valuation difficulties, treatment of losses, listed fact a recognition, of very valid reasons why a country may decide to defer taxation. securities, and other double taxation issues that might arise under a given set of circumstances. Further, unless there are strong reasons to do otherwise, either model should only be implemented Commented [ITIC25]: New provisions in the law (as on a prospective (and not retroactive) basis (e.g. to transactions taking place after the change is opposed to corrections to or clarifications of prior law) should never be allowed retroactive application. To do announced, as opposed to applying to tax years before the announced change), and appropriate so deprives taxpayers of the ability to manage their transitional arrangement could also be considered (such as deeming the market value cost base business. of relevant assets to be that at the time of commencement of the new taxing model). The ultimate set of provisions to be adopted in the location country to enable it to tax OITs would need to take into account the specific legal tradition and system, as well as the political and administrative structure and fiscal policies of the country concerned. These sample legislative provisions have been designed and drafted to prevent legal double taxation by the location country—that is, to prevent the gain on an asset transfer being taxed 48 twice by the location country in the hands of the same taxpayer. This reflects common international Commented [ITIC26]: Shouldn’t this be to prevent the practice in this context. Consideration could also be given to designing and drafting legislative same income from being taxed twice—irrespective of the taxpayer? Shouldn’t the rules be designed to prevent double taxation economically—especially when the comparison is to a direct asset sale in a country? 63 See Waerzeggers and Hillier (2016). 49 provisions that limit double taxation in the sense of the same gains being taxed multiple times in the hands of different taxpayers through realizations of gains on intermediate shareholdings through multiple tiers of indirect ownership. To achieve this, the tax cost of relevant assets (e.g. each intermediate shareholding) must be reset (stepped-up) to market value each time a relevant taxable realization occurs or, alternatively, the law can provide for the non-recognition of a gain on each intermediate asset.64 Even though provisions of this kind would be more comprehensive, they would also be more complex to apply and administer, and so are not reflected in the sample domestic legislative provisions set out in this section. For the purposes of the sample provisions provided in this section, the location country is referred to, as above, as Country L. B. Model 1: Taxing the Local Resident Asset-Owning Entity under a Deemed Disposal Model This model seeks to tax the local asset owner on the basis that the asset it holds has undergone a change of control because of an offshore sale of an entity that owns the local asset owner, directly or indirectly. Under this model, the tax liability with respect to the gain realized by the non-resident seller is (unilaterally) triggered for the local resident asset-owning entity under a specific set of domestic legislative provisions, without primary reliance on the international source of income or broader international taxation rules (such as tax treaty allocation rules). This approach has been adopted in a number of source countries, such as Nepal, Ghana and Tanzania. A sample set of legislative provisions underpinning this domestic deemed disposal model is set out in Box 5. Box 5: Change in Control (1) Subsection (3) applies when the direct or indirect ownership of an entity mentioned in subsection (2) changes by more than 50 percent as compared with that ownership at any time during the previous three years. (2) An entity to which subsection (1) applies is an entity in respect of which, at any time during the 365 days preceding the relevant change in underlying ownership, more than 50 percent of the value of the shares or comparable interests issued by that entity is derived, directly or indirectly, from immovable property in Country L. (3) Where this subsection applies, the entity is treated as: (a) realizing all its assets and liabilities immediately before the change; (b) having parted with ownership of each asset and deriving an amount in respect of the realization equal to the market value of the asset at the time of the realization; (c) reacquiring the asset and incurring expenditure of the amount referred to in paragraph (b) for the acquisition; (d) realizing each liability and is deemed to have spent the amount equal to the market value of that liability at the time of the realization; and (e) re-stating the liability for the amount referred to in paragraph (d). 64 The latter is the method of implementation used for the Chinese provision. 41 The application of this model can be summarized as follows: • The model operates to tax the unrealized gain in the hands of the local asset owning entity. That entity will typically be a resident of the location country, giving that country the right to tax on both a residence basis and a source basis. The model seeks to tax the accrued gain on the entire asset when a change of control occurs from the sale of interests whose value is principally (e.g. more than 50 percent) derived from the asset—the local immovable property. Losses should also be recognized where there are no accrued gains, and should be subject to appropriate loss utilization rules (as applicable). The tax liability is triggered by a change of control, irrespective of whether that change occurs because of an offshore or onshore sale of shares or comparable interests. A technical direct change of control because of a corporate reorganization should not trigger the tax liability. • The tax liability is also triggered irrespective of the size of the interest that is sold to bring about the relevant change in control. That is, no threshold of ownership of the asset is set—for example, by saying that the tax liability would only be triggered if the change of control arises from the sale of an interest of 10 percent or more in the asset. This is done to limit tax avoidance opportunities which would arise if a significant interest threshold were adopted under this model. Change of control is determined by reference to direct or indirect ownership, which enables the tracing through of intermediate holding entities between the local asset owning entity and the ultimate issuer of the shares which are the subject of the actual sale. • Where a change of control occurs, the model treats the local asset owning entity as disposing of its assets for their market value. The value of the local assets which are deemed to be sold could administratively be determined using assumptions and adjustments based on the price at which the actual shares are sold, on the basis that their value is derived from the value of the local assets. An apportionment rule could be adopted and applied so that the price paid for the shares (assuming the share sale occurs on arm’s length terms) is appropriately allocated amongst the assets held by the local entity. In practice, it is recognized that these valuation exercises are complex to undertake, particularly where relevant assets relating to the underlying immoveable property derive their value from commodity prices, centrally provided inputs (e.g. management and technical expertise) and other group shareholdings. Additionally, the location country may support its domestic legislation based on Model 1 by imposing a reporting obligation of the price of the shares to the resident entity. However, the nature of the disposal is only a deemed (as compared to an actual) disposal for tax purposes. Therefore, the local asset owning entity will still be the legal owner of the assets after the disposal is deemed to take place. In order to protect against double taxation, the model treats the local asset owning entity as reacquiring the assets for their 42 market value. This means that its tax cost in those assets is stepped up to market value— which is important to ensure that double taxation does not arise in the location country as the assets are used in generating income or in the event that another subsequent change of control occurs. Commented [ITIC27]: Double taxation is a concern on both continuing operations of the assets OR on a Liabilities are also reset under this model for ease of administration. This means that the subsequent change of control. entire balance sheet is reset instead of resetting only the assets of the local asset owning entity, which would otherwise leave liabilities to be recognized at their historic value and complicate compliance. No gain or loss on a liability would be expected to be realized in the ordinary case under this model where the market value of that liability was equal to its face value. Commented [ITIC28]: Isn’t this just saying no gain or loss on a liability where there is no gain or loss? Maybe For completeness, and as noted above, the model constitutes a simplified set of legislative better to say that there will be gain or loss on liabilities, unless this is the case. provisions that do not deal with more complex issues such as corporate reorganizations, minority shareholders, joint venture arrangements, valuation difficulties, listed securities, and the treatment of losses. The ultimate set of provisions to be adopted in the location country will need to be adapted to reflect the individual circumstances of the country concerned, including its domestic and international tax policy settings. Enforcement/collection rules Under this model, the local asset owning entity remains subject to the ordinary compliance rules applicable to resident taxpayers, with no need for specific enforcement and collection rules—or reliance on assistance in collection treaties—to combat the significant difficulties in collecting the tax where transactions take place between two non-residents. Under this model, the tax authority of the location country can use the full suite of its enforcement tools against the local asset owning entity (e.g. apply penalties for a failure to file and pay tax in respect of the deemed gain, and activate the usual enforcement instruments at its disposal, such as seizing or freezing the local assets and potentially selling them to settle an outstanding tax liability). The key advantages of this model are: Greater ability to enforce and collect the tax liability as the taxable gain is deemed to have been realized by the local asset owning entity (as compared to a non-resident). This means that the tax authority can use the full suite of its enforcement tools against the local asset owning entity. Double taxation in the location country should not arise either as income is earned in the future or when another subsequent change of control occurs, as the basis of the local assets which are deemed to be disposed of is stepped up to market value in the hands of the local asset owning entity both for depreciation and for gain and loss purposes. The gain under the deemed disposal model consists of a locally sourced gain realized by a local resident entity. Therefore, the taxing right of the location country should not be affected by a tax treaty. 43 The key disadvantages of this model are: Some may argue that there would still be tax treaty limitations if a tax treaty were in place, on the grounds that the imposition of the tax is in substance source country taxation triggered by an offshore sale of interests (e.g. shares). However, some consider that this argument could be materially countered by adopting treaty anti-abuse provisions such as those reflected in the BEPS minimum standard to counter treaty shopping. This could include a limitation on benefits (LOB) article and/or principal purpose test (PPT), either in a tax treaty or as a domestic law override of a tax treaty, if the non-resident seller is a company situated in a treaty country for the purpose of obtaining treaty benefits (assuming there are no constitutional limitations to doing so through domestic law). The PPT takes the form of a SAAR to prevent treaty shopping, while an LOB rule can take many forms, but an increasingly common formulation is to require 50 percent or more of the ownership of the resident of the other contracting state to be held by an individual or individuals who are themselves residents of that other contracting state before becoming eligible to access tax treaty benefits. However, countries take different views about the ability to effectively administer an LOB domestically. • There could be possible double taxation as the residence country of the offshore seller of the transferred interest could tax the gains realized by that seller from the sale. And if so, there would be no foreign tax relief available in the country of the seller, because the tax liability in the location country arises for the local asset owning entity, and not the offshore seller, who under this model is not taxed at all by the country of location/source. However, these double taxation concerns would be reduced where the residence country of the offshore seller operates under a territorial system of taxation or otherwise excludes such foreign gains from its domestic tax base (e.g. through a participation exemption, or because the residence country is a no or low tax jurisdiction). There is a current trend by residence countries to adopt—or move towards—a more territorial system of taxation. Since the entity that directly owns the asset does not receive the money from the transfer of the shares or comparable interests, difficulties may arise regarding the effective collection of taxes when that entity lacks the liquidity required to pay the tax liability. Practically speaking, however, it is expected that the parties (particularly the purchaser) would take steps to ensure that the local asset-owning entity had sufficient funds to discharge its tax liability, in order to prevent the tax authority from taking enforcement action against locally held assets. This approach undermines the separate legal entity distinction between the local asset holding entity and its relevant tiers of parent entities. Further, absent overarching shareholder agreements, this approach would also result in continuing minority shareholders becoming exposed to the underlying tax liability in circumstances where that tax liability is triggered through the sale by a majority shareholder of their own controlling shareholding. However, those continuing minority shareholders would also benefit from 44 the step up to market value of the local assets (including any depreciable assets) in the hands of the local asset owning entity. This approach, as a practical matter, requires the local asset-holding entity to monitor changes in its own ownership. The merits of Model 1—relative ease of enforcement and simplicity of the necessary basis adjustment—can be especially appealing for lower capacity countries. Under this model, it should be noted, the source rules of the location country L need to be designed and drafted in a manner that would not result in the OIT that triggered a change of control having a source in Country L. If the actual sale of the offshore interests were held to be sourced in Country L, double taxation would arise on the same transaction in the same location as a result of both deemed resident taxation and actual non-resident taxation in Country L. A tax liability rule which is designed and drafted to impose the primary tax liability on the non-resident seller of the relevant interest instead of on the local asset owner (under a non-resident taxation model) is discussed hereafter in the context of Model 2. C. Model 2: Taxing the Non-resident Seller Under this model,65 Country L seeks to impose tax on the non-resident seller on the basis that the transfer gives rise to a gain with a local source in Country L. Where countries have resolved to tax OITs, this model (or a variation thereof) has been the one most commonly adopted. Under this model, the source rules become critical for triggering the tax liability in the location country. This is because a non-resident is ordinarily66 only subject to taxation on income derived from sources in the particular location country. By way of example, a sample source rule along the lines shown in Box 5 below could be considered when seeking to impose a liability on a non- resident in respect of a gain realized on the sale of an indirect interest in immovable property situated in the location country L. Box 6: Source rule The following amounts are derived from sources in Country L: (a) A gain arising from the alienation of: (i) immovable property in Country L; (ii) shares or comparable interests, if, at any time during the 365 days preceding the alienation, more than 50 percent of the value of the shares or other interests is derived, directly or indirectly through one or more interposed entities, from immovable property in Country L; 65 See also Toledano and others (2017) on the design and implementation of Model 2. 66 The assumption here is that the tax system of the location country is structured using international norms like those embodied in the model treaties—but included in the location country’s domestic law/source of income rules. 45 As noted above, the source rule may be combined with a taxable asset rule. The practice as embodied in Article 13(4) of the OECD and UN Model MTCs is to allow the taxation of the entire gain when the value of the indirect interest is principally (e.g. more than 50 percent) derived from local immovable property. Where this is the case, reliance may simply be placed on the source rule (above). Alternatively, a taxable asset rule may confirm and support this treatment by similarly providing that the entire gain is taxable when the value of the indirect interest is principally (e.g. more than 50 percent) derived from local immovable property. The taxable asset rule could alternatively be designed and drafted to apply on a proportionate basis (e.g. taxing only those gains attributable to the local immoveable property, as distinct from the entire gain), or on a modified pro rata basis where a lower threshold is met (e.g., 20 percent of the gain is derived from local immovable property rather than 50 percent or more). For example, a modified pro rata mechanism of this nature has been adopted in Kenya for the extractives sector. Where such a lower threshold is adopted, it would generally be appropriate to impose tax only on a proportionate basis. A sample set of domestic legislative provisions demonstrating the two approaches is outlined in Box 7 below. Box 7: Taxable asset rule: Full and pro rata taxation (1) The chargeable income of a person includes gains from the realization of shares or comparable interests, if, at any time during the 365 days preceding the realization, more than 20 percent of the value of the shares or other interests is derived, directly or indirectly through one or more interposed entities, from immovable property in Country L. (2) For the purposes of subsection (1), the amount of the gain to be include in chargeable income is- (a) if the shares or other interests derive, or derived at any time during the 365 days preceding the realization, more than 50 percent of their value, directly or indirectly, from immovable property in Country L, the full amount of the gain; or (b) in any other case, the amount computed according to the following formula: A × B/C where- A is the amount of the gain; B is the value of the shares or other interests derived, directly or indirectly, from immovable property in Country L; and C is the total value of the interest. The development and application of any domestic legislative provisions will need to take into account any existing tax treaty obligations. However, as discussed above, the taxing right over gains realized on offshore indirect transfers which are principally (e.g. more than 50 percent) derived from local immovable property is generally preserved in Article 13(4) of the OECD and UN MTCs. Both MTCs permit the location country to capture gains from the sale of relevant interposed holdings at different tier levels. It is important that the domestic legislative provisions of the 46 location country are designed and drafted to preserve this taxing right over relevant interests which derive more than 50 percent of their value, directly or indirectly, from immovable property in the location country as permitted by the MTCs. There is also a question as to whether the taxable asset rule should define—and potentially narrow—the scope of the interest which is to be subject to tax. Three further options arise in this regard: (i) imposing a tax liability in relation to the disposal of all interests (including interests even representing less than a de minimis interest in the asset), as long as the value of the interest disposed of derives more than half its value from that asset; or (ii) imposing liability only on the disposal of more significant interests (e.g. interests of 10 percent or more of the asset); and/or (iii) whether a back-up threshold based on the nominal value of the interest should also apply (e.g. apply the rule only to interests with a value of $1 million or more). For example, if a percentage of interest threshold were to be adopted, a 10 percent threshold could be considered as it is the international norm for distinguishing between a non-portfolio and portfolio investment. If adopted, such thresholds could help minimize compliance costs and ease administration. However, implementing a threshold interest requirement needs to be carefully drafted so as to preserve the policy intent of the threshold and combat tax avoidance opportunities through staggered sell- downs (i.e. selling multiple parcels of shares each comprising an interest of less than 10 percent). Finally, countries may provide exemptions to the application of Article 13(4) of the OECD and UN MTCs to certain capital gains for different reasons. As explicated in the Commentary on Article 13(4) of the OECD MTC, these exemptions may refer to gains derived from the alienation of: (i) shares of companies listed on a stock exchange; or (ii) shares in the course of a corporate reorganization; or (iii) shares which derive their value from immovable property where a business is carried on; or (iv) shares held by pension funds; or (v) a small investor’s interest in a Real Estate Investment Trust (REIT). Further, to the extent that a loss arises (instead of a gain), to prevent double taxation, that loss shouldcould also be recognized in Country L. If the loss is and be subject to appropriate loss utilization rules that limit the ability to offset an earlier taxed gain, as can often be the case, taxpayers are inappropriately double-taxed. Enforcement/collection rules Non-residents subject to tax in Country L under this method would normally be required to file tax returns in Country L where a taxable gain is realized in relation to the OIT. However, compliance with this obligation could be expected to be low. Even though the tax authority in Country L has certain enforcement instruments at its disposal (as noted above), these can be difficult to apply in the case of a tax liability of a non-resident (as compared to a tax liability of a resident), particularly when the sale proceeds from disposing of the interest have left, or were never in, the location country, and there are no other assets directly owned by the transferring offshore entity in the location country to meet or secure the tax liability. Therefore, appropriate supplemental enforcement and collection mechanisms need to be designed, drafted and implemented for this situation. Certain legal protections can be developed to support the enforcement and collection 47 efforts of the tax authority, as well. Measures of this kind could consist of restricting the 48 registration, renewal or validity of relevant underlying assets (e.g. extractive licenses) by governmental registration bodies or other registration and issuing entities, unless applicable notification requirements have been met and/or sufficient evidence has been furnished to demonstrate that either no tax is payable, the relevant tax has been paid, or satisfactory arrangements have been made for the payment of that tax. Withholding Several countries use a withholding mechanism to collect tax with respect to a non-resident seller’s gain. A specific withholding tax regime can be designed and drafted to apply to payments to a non-resident seller. Withholding taxes can represent all or a portion of the tax liability (or possibly an estimate) of the recipient of the payment. The tax must be withheld from the payment by the payer and paid to the tax authority in the location country. A regime may be designed to impose withholding of tax by the payer, as either a final or non-final charge on the payee. A final withholding tax represents the final tax liability for the person receiving the payment withheld upon. Final withholding tax regimes are common for gross payments of dividends, interest and royalties made to non-residents. In contrast, a non-final withholding tax is collected as an estimate of the recipient’s final income tax liability. The recipient is ordinarily still required to file a return and pay any outstanding balancing amount after claiming a credit for the amount of tax withheld (or receive a refund, if the withheld amount exceeds the tax due). Typically, a withholding tax regime applicable to OITs would be designed as a non-final regime. A withholding tax regime applies to OITs in a number of jurisdictions, including the U.S., Canada, India, China and Australia. The withholding tax regime could be designed to exclude withholding in certain circumstances in order to minimize compliance costs. This could include, for example, transactions below a predetermined de minimis threshold (an option noted above); transactions related to listed securities on a stock exchange; transactions wherein a clearance certificate is obtained from the tax authority in Country L to confirm that no amount is required to be withheld in the particular circumstances (for example because the asset is being sold for a loss etc.). A sample withholding tax regime is shown in Box 8. There are a number of issues in relation to the adoption of a withholding tax regime in the context of OITs. For example, if—as will typically be the case—the purchaser is also a non-resident then similar non-compliance risks arise. As noted, the withholding tax can only be collected as an estimate of the seller’s final income tax liability (as the actual quantum of the seller’s gain is unlikely to be known by the purchaser) and so withholding necessarily increases the compliance burden for the purchaser (who is subject to the withholding obligation) and the seller (who needs to file a tax return and determine any outstanding balancing amount or refund after claiming a credit for the amount of the tax withheld)—although this burden could be manageable. In this regard, it is often considered that the risk of non-compliance with a withholding tax obligation in the context of OITs (particularly where the purchaser is also a non-resident) is minimized by the likelihood that 49 a prudent third party purchaser will not acquiesce or facilitate the avoidance of the seller’s tax liability (and therefore is more likely to comply with its withholding tax obligation). Further, a failure to withhold would expose the purchaser to penalties and potential seizure of the local asset by the tax authorities, and would also result in the seller making a windfall gain if the purchaser were unable to recover the penalty amount from the seller. In this sense, the withholding tax mechanism creates an interest in the purchaser to assure tax compliance of the seller in respect of the transaction. Box 8: Enforcement/collection rule: Withholding tax (1) A person must withhold tax at the prescribed rate when: (a) the person pays an amount to another person (recipient) in acquiring shares or comparable interests; and (b) more than 50 percent of the value of the shares or comparable interests referred to in paragraph (a) is derived, directly or indirectly through one or more interposed entities, from immovable property in Country L. (2) The person (withholding agent) must pay the amount to the tax administration on or before the day that the withholding agent becomes the owner of the shares or comparable interests and must file a statement in the manner and form prescribed. (3) A withholding agent who fails to withhold tax in accordance with this section must nevertheless pay the tax that should have been withheld in the same manner and at the same time as tax that is withheld. (4) Where a withholding agent fails to withhold tax from a payment as required by this section- (a) the recipient is jointly and severally liable with the withholding agent for the payment of the tax to the tax administration; and (b) the tax is payable by the recipient immediately after the withholding agent becomes the owner of the shares or comparable interests. (5) A withholding agent who withholds tax under this section and pays the tax to the tax administration is treated as having paid the amount withheld to the recipient for the purposes of any claim by the recipient for payment of the amount withheld. (6) A withholding agent who fails to withhold tax under this section but pays the tax that should have been withheld to the tax administration in accordance with subsection (3) is entitled to recover an equal amount from the recipient. (7) The recipient is treated as having paid any tax- (a) withheld from the payment under this section; or (b) paid in accordance with subsections (3) or (4). (8) A recipient is entitled to a tax credit in an amount equal to the tax treated as paid under subsection (7) for the year of assessment in which the payment is derived. 50 Notification and agency taxation In the absence of adopting a withholding tax regime, two other enforcement and collection measures may be considered for the purpose of putting the tax authority in the best position to be aware of the disposal and then being able to subsequently enforce and collect the tax. These involve designing and imposing the following two obligations: (a) a notification/reporting obligation; and (b) a payment obligation for an entity in the location country as agent for the non-resident. The notification/reporting obligation is important not only for raising an assessment, but also for exploring other available avenues for recovery of any unpaid tax on the transfer, such as through an Assistance in the Collection of Taxes article under applicable tax treaties or though the OECD Multilateral Convention on Mutual Administrative Assistance in Tax Matters. Further, imposing a payment obligation on a resident person or entity as agent for the non- resident enables the tax authority to use the full suite of its enforcement tools against that resident. Legislative mechanisms of this kind have recently been adopted in Kenya and Fiji for the extractives sector. A sample set of legislative provisions is set out in Box 9. For illustrative purposes, they are shown as triggered with respect to holdings of non-portfolio interests of 10 percent or more. Box 9: Enforcement/collection rule: Notification and agency taxation of non-portfolio interests (1) Subsection (3) applies when the direct or indirect ownership of an entity mentioned in subsection (2) changes by 10 per cent or more. (2) An entity to which subsection (1) applies consists of an entity in respect of which, at any time during the 365 days preceding the relevant change in the direct or indirect ownership, more than 50 percent of the value of the shares or comparable interests issued by that entity is derived, directly or indirectly, from immovable property in Country L. (3) Where this subsection applies, the entity: (a) must immediately notify the tax administration, in writing, of the change; and (b) is liable, as agent for any non-resident disposing of the interest to which the notice under paragraph (a) relates, for tax payable by the non-resident under this Act in respect of the disposal. (3) Subsection (3)(b) does not apply to the disposal of shares quoted in any official list of a recognized stock exchange in Country L.* (4) Any tax paid by the entity on behalf of a non-resident under subsection (3) is to be applied against the tax liability of the non-resident under this Act. * This exclusion reflects that relevant parties are unlikely to be in a position to establish by legal agreement how the ultimate burden of the tax is to be borne by them where the shares are sold on a stock exchange. 51 Pros and cons of Model 2 The key advantages of Model 2 are: • It more closely preserves the separate legal entity distinction between the local asset owning entity and its relevant holding entity/parent. • Relief of double taxation is preserved in the country of residence of the seller, as foreign tax relief should remain available because the offshore seller is primarily liable for the tax payable sourced in the location country on the gain realized from the sale. The key disadvantages are: • Reduced ability to enforce and collect the tax liability as the taxable gain is realized by the non-resident seller (as compared to a local entity under the deemed disposal model)— although this could be aided by using the withholding/agency collection mechanism. • The agency approach assumes that the direct owner in country L can always make itself aware when there has been a transaction resulting in a 10 percent or greater change in the underlying ownership of the entity. • Double taxation can effectively arise on a subsequent sale of interests in other entities that indirectly hold the assets because shares or interests in those entities are not stepped up to market value. However, this is a general feature that typically arises when there are multiple tiered holding structures, whether domestic or cross-border. • Economic double taxation will occur unless the assets in the country are stepped up to Formatted: Bulleted + Level: 2 + Aligned at: 0.85" + the sales value, as in Method 1. This can distort the taxation results as compared to the Indent at: 1.1" base case of a direct sale (or a non-sale, continued operation scenario). • Even with appropriate domestic legislation, under this model the taxing right of the location Country L could (unless there was a treaty override) still be limited by an applicable tax treaty, if the relevant treaty does not include an article similar to Article 13(4) of the OECD or UN Model MTC. D. Defining “Immovable” Property In all of the foregoing approaches, an appropriate definition of “immovable property” is critical for the effective application of the chosen tax liability rule and associated enforcement and collection rules. A definition of “immovable property” with appropriate clarity will be equally relevant for Model 1 and Model 2, and each of those models is capable of having a even greater reach in circumstances where that definition is extended to cover a broader category of “immovable property” than is traditionally the case, particularly where narrow definitions have been adopted. See footnote 49, infra. Natural resource rich countries might in particular want to adopt a definition closer to that of Australia’s (.as set out in footnote 49, infra). A 52 sample definition of “immovable property” along the lines of the Australian one is set out in Box 10 below: Commented [ITIC29]: We recommend the definition in box 10 include only subparagraphs a, b, and c. Subparagraph d is overly broad, and arguably involves information that would fail to meet the test of an asset generating location specific rents. 53 Box 10: Sample definition of immovable property “Immovable property” includes* a structural improvement to land or buildings, an interest in land or buildings or an interest in a structural improvement to land or buildings, and also includes the following – (a) a lease of land or buildings; (b) a lease of a structural improvement to land or buildings; (c) an exploration, prospecting, development, or similar right relating to land or buildings, including a right to explore for mineral, oil or gas deposits, or other natural resources, and a right to mine, develop or exploit those deposits or resources, from land in, or from the territorial waters of, Country L; or (d) information relating to a right referred to in paragraph (c). * This definition has been drafted on an inclusive basis and presupposes that it would cover within the ordinary meaning of “immovable property” all traditional notions of real property (e.g. land, buildings and mines etc.). Confining the definition of immovable property to traditional notions of real property in the form of land and buildings may be thought too restrictive. Such a narrow definition would not be sufficient to enable Country L to trigger its taxing right over gains made in the context of extractive industries, such as gains from licenses to explore for, develop, and exploit natural resources located in County L. Countries should therefore consider defining immovable property in their domestic laws to include at least: • Real property (in the narrower sense); • Mineral, petroleum, and other natural resources; and • Rights (such as those embodied in licenses) to explore for, develop, and exploit natural resources, as well as information relating to those rights. The domestic law definition of immovable property will also be important in the context of the application of a tax treaty. This is because the basic rule under the OECD and UN MTCs is that the term “immovable property” has the meaning under the domestic law (tax or other law) of the contracting state in which the property is located. The definition could be further extended to cover a broader category of “immovable property” that Country L might think it appropriate to tax. As argued above, this could include—for example—gains arising in relation to location specific rents clearly linked to national assets, such as from licenses to exploit public goods (e.g., electric, gas, or other utilities; telecommunications and broadcast spectrum and networks etc.).67 may be that an enhanced ability to tax rents generated by government restrictions would have an adverse 67 It effect in encouraging the imposition of such restrictions. This is a reasonable and significant concern. Such 54 Consideration could be given to extending the definition even further, to cover rights to receive variable or fixed payments in relation to extractive industry rights or government issued rights with an exclusive and territorial quality. This would also ensure that gains relating to any subsequent assignments derived from those underlying rights granted by or on behalf of the government of Country L would also remain within Country L’s tax base. It is clearly the case, however, that the concept of location specific rents is much easier to conceive of in economic terms than it is to convey in legal language. This is an area in which further thought is needed. In addition, it will also be important that a country complies with its good faith obligations with respect to the interpretation of tax treaties if that country decides to expand its definition of immovable property when existing tax treaties are in force. Although a country may change the definition of a term used in its domestic legislation which is also used in treaty provisions but which is not specifically defined for the purposes of the treaty, countries should ensure that any modifications or extensions are compatible with the context—and negotiated position—of their existing tax treaties that may be in force at the time of any modification or extension. Treaty partners should consult with each other in this regard. incentives already exist, however: Auriol and Warlters (2005) find evidence that governments in lower income countries tend to create barriers so as to concentrate profits in easily-taxed large firms. The prior practical issue, in any case, may be securing revenue where such the granting of such rights is used to regulate major natural monopolies. 53 CONCLUSIONS This report and toolkit does not purport to provide binding rules or authoritative Formatted: Font: Not Italic provisions of any kind nor does it aim to establish an international policy standard of any kind. As such, the report represents the analysis and conclusions of the tax staffs of the four partner organizations, and does not represent the official views of the organizations’ member countries or Management. Commented [ITIC30]: This abbreviated disclaimer helps put the following “conclusions” in the context of the document This report and toolkit concludes that it is appropriate that location countries have the right to tax OITs, although the decision of whether or not to elect to tax should be made by each Formatted: Font: Not Italic country based on its own facts and circumstances, , at least for assets that are likely to Commented [ITIC31]: Added to make it clear that the embody, primarily and substantially, location specific economic rents, including those draft does not conclude that countries should tax an OIT. traditionally thought of as “immovable.” (The former might include, for instance, natural resources, both the physical assets and associated rights; and the rights to location specific telecom or other licenses). This is so, moreover, regardless of whether equivalent tax in regard to the transfer would be paid elsewhere. That is, the reasoning behind this presumption implies that Commented [ITIC32]: But the tax will be paid in the such taxation could apply not only as an anti-avoidance device to combat “double non-taxation,” country—the paper acknowledges that only a timing effect is involved. These arguments go much further but rather could constitute a fundamental aspect of the country’s tax laws. than that. And how can double non-taxation arise here again when dealing with a timing difference? The rationale for this approach—which would in effect place the approach adopted in Article 13(4) of both model tax treaties on a firmer economic basis—rests on several pillars. In equity terms, it mirrors a quite generally recognized right in relation to direct transfers of immovable assets. In efficiency terms, it provides a backstop to the taxation of location specific rents which, especially in low income countries, other instruments may be able to achieve only imperfectly, and also fosters neutrality between direct and indirect transfers. And it responds to pressure sin the case of the sale of salient national assets which have led to uncoordinated measures that jeopardize the smooth and consensual functioning of the international tax system and can give rise to tax uncertainty. This is not to say that location countries should always tax related OITs. Therey may behave good reasons for a country not to exercise a right to tax OITs—depending, for example, on their administrative capacities, particularly given some of the complexities noted regarding implementation and enforcement (and coupled with the fact that such OIT taxation will generally only affect the timing of revenue), current versus future y, revenue needs and budget objectives, and the desire to attract foreign investment—to choose not to, as some countries now do. The provisions of both MTCs suggest quite wide acceptance of the principle that capital gains taxation of OITs of “immovable” assets can be allocated primarily to the country in which they are located. As of 2015, however, Article 13(4) appeared in only around 35 percent of all DTTs, and was less likely to be found when one party is a low-income resource rich country. To date, the Multilateral Convention has had a positive impact by increasing the number of tax treaties that effectively include Article 13(4) of the OECD MTC. This impact is expected to increase, as new parties may decide to negotiate or re-negotiate treaties based on the 2017 OECD and UN 54 formulations of that provision and/ or to sign the MC and modify their covered tax treaties to include the new language of Article 13(4). The report also stresses, however, that, whatever treaties may or may not come into play, such a taxing right cannot be supported without appropriate definition in domestic law of the assets intended to be taxed, and without a domestic law basis to assert that taxing right. Sample simplified legislation for such rules was provided in the text, but actual legislation to deal with all of the issues involved in exerting this taxing right will take considerable care, analysis, and careful drafting. . A key issue in that context is the appropriate definition of “immovable” property. The concept is not one that is especially meaningful in economic or even administrative terms. The analysis here suggests that a more useful conceptual approach is to identify and capture within 55 the definition those assets whose value derives in large part from location specific rents. While it would be preferable to tax such rents directly—as indeed countries are routinely advised to do—, imperfections in the design and implementation of such taxes can leave a valuable backstop function for taxation of the gains associated with increases in the value of such rents. This view of Commented [ITIC33]: As noted throughout, this the underlying economics points towards the possibility of using an expansive definition of argument is still hard to understand, particularly with respect to extractives. immovable assets to include a wide range of transfers related to rights bestowed by government that are capable of generating substantial income. The central practical issue raised by OITs is enforcement of taxation by the country in which the asset is located—provisions for which require careful drafting. The report outlines the two main approaches for so doing—which in legal terms are quite different. One of these methods treats such an OIT as a deemed disposal of the underlying asset. The other treats the transfer as taking place by the actual seller, offshore, but sources the gain on that transfer within the location country—thus permitting the country to tax it. This report has provided sample simplified legislative language for domestic law in the location country for both. Countries are responding to the issues they have encountered in respect of OITs in very different ways. The measures they have adopted differ both in terms of which assets are covered (immovable property, narrowly or broadly defined; other assets like telecoms; intangibles such as corporate stock issued in regard to a domestic company but held by a non-tax resident), and in terms of the method used to impose the tax as a legal matter. A more uniform, coordinated and coherent approach to the taxation of OITs, where countries choose to tax them, can make a substantial contribution to coherence in international tax arrangements and enhanced tax certainty. This report is intended to help progress to those ends. 56 Appendix A. Consultations The first draft of this report was posted for public comment from August through late October, 2017, in English, French and Spanish. During that time, extensive written comments from 18 private sector organizations, civil society groups, country governments, and individuals were received and posted by the Platform on line: BIAC The Business and Industry Advisory Committee at OECD BMG Base Erosion and Profit Shifting Monitoring Group (a consortium of 7 civil society organizations) CBI Confederation of British Industry (London) China (State Administration of Taxation, P.R. China) Deloitte LLP (London) ICC International Chamber of Commerce India (Government) ITIC International Tax and Investment Center (US and other) Jubilee USA An alliance of 700 faith groups (US) KPMG KPMG International (UK) Philip Baker (UK) PwC PricewaterhouseCoopers International Limited (London) Repsol (Spain) Sergio Guida CPA (Italy) SVTDG Silicon Valley Tax Directors Group (US) TEI Tax Executives Institute (US) TPED Transfer Pricing Economists for Development (Paris and Vienna) USCIB United States Council for International Business (US) 57 Appendix B. Comparing Direct and Indirect Transfers68 Suppose the underlying asset has market price 0 = 0 (for simplicity) when acquired by the initial owner, price 1 in period 1, and 2 = (1 + )1 in period 2. If the owner sells the asset itself in period 1, this yield income net of the capital gains tax in the location country charged at rate G, of 1 − (1 − 0) = (1 − )1 (1) Alternatively, the owner might sell the shares of the company owning the underlying asset in period 1, yielding proceeds, assuming the shares to have initially had zero value, of 1 − (1 − 0) = (1 − )1 . (2) where Z is the rate of tax levied on the share transaction in the jurisdiction in which the company sold is resident. To determine the share price 1, suppose that in period 2 the purchaser will sell the underlying asset, incurring tax in the location country, sell the shares of the company acquired at stock price 2, and repay (with deduction of interest at tax rate T of the debt incurred to finance. This gives net cashflow to the purchaser in period 2 of69 (1 − )2 − (2 − 1) − {1 + (1 − )}1 (3) where R denotes the pre-tax rate of interest. Setting this to zero, and assuming the company has no other underlying assets, so that 2 = 0, the largest amount that the purchaser is willing to pay is (1 − )2 1 = (4) 1 − + (1 − ) Substituting from (4) into (2), the net receipts from an indirect sale in period 1 are thus (1 − )(1 − )(1 + )1 (5) 1 − + (1 − ) 68 The analysis here is motivated by and to some degree elaborates on that in Kane (2018). 69 Note that there is no step up in basis of the underlying asset consequent upon the share sale. For simplicity, it is assumed that the rate of capital gains on the initial and subsequent shares sales are the same. 58 Comparing this with (1), the indirect sale therefore yields more to the initial owner of the underlying asset then does the direct sale if and only if (1 − )(1 − )(1 + )1 (6) > (1 − ) 1 − + (1 − ) 1 which reduces to the condition (1 − ) > (1 − ) (7) Some observations that follow from this result: The rate of capital gains tax on sale of the underlying asset, , is immaterial to the comparison. This is the counterpart of the point highlighted in the text that the issue for the location country (in respect of the gain on the underlying asset) is essentially one of timing, given that the tax will be paid at some point.70 If the price of the underlying asset increases at the pre-tax rate of interest R, then the initial owner is indifferent between the options of immediate direct and indirect sale if and only if = , meaning that capital gains on shares are taxed and interest deductible at the same rate. • If no tax is payable on capital gains in share transactions ( = 0) and the rate at which price of the underlying asset increases is equal to the general rate of inflation, then the indirect sale is preferred by the investor so long as the real after-tax interest rate (1 − ) − is strictly positive. 70 Kane (2018) notes the analogy with the trapped equity view of dividend taxation referred to in . 59 Appendix C. Examples of Country practices There is considerable diversity in countries’ approaches to taxing OITs. Many OECD countries naturally follow their MTC, but not all. Mexico’s approach, for instance, is closer to the UN MTC: it taxes capital gains realized by foreign residents on the transfer of shares issued by domestic companies, regardless of where the title is passed, if more than 50 percent of the value of these shares derives from immovable property situated in Mexico.71 Other countries deviate from both the OECD and the UN MTC. The U.S., Peru and China, to name a few, exemplify this diversity. For instance, the transfer in the Zain case described in Box 3 (p.23) would not be taxed in the U.S., but would be in Peru, and might or might not be in China. A. United States: Dispositions of U.S. real property held by foreign investors Weaknesses of a pure residence taxation model The U.S. income tax originally followed the premise that, in the absence of a U.S. trade or business, business profits of foreign residents should be taxed in their place of residence, defined in the case of individuals by a physical presence test (at least 183 days during 12 consecutive months). However, the law allowed numerous avenues to avoid the U.S. capital gains tax when there was a U.S. trade or business. For example, the payment for the sale of an asset could be timed to occur after the entity engaged in the U.S. trade or business had been liquidated, so that the capital gain would be realized when the foreign resident had no U.S. business to be connected to. Also, foreign residents could exchange the U.S. property for another of the same kind abroad and this would not qualify as a realization of a capital gain72. Alternatively, the foreign resident could hold the property in a domestic (or foreign) corporation and sell the stock of the corporation instead of the underlying property; in other words, it could avoid the tax through an indirect sale (either onshore or offshore).73 The farmers’ lobby Of particular concern was that foreign investors could have a resident tax status during the operational stage of the business, obtain a net income basis tax regime for that period of time, minimizing taxable profits (through expense deductions), and switch to a non-resident tax status when selling the appreciated asset, avoiding the capital gains tax at that point.74 Assuming that 71 Ley del ISR (Mexico), art 161. 72 Brown (2004), p.297. 73 Petkun (1982), p.13 74 Presumably the value of the asset could reflect undistributed accounting profits. 60 foreign investors paid no (or little) capital gains tax abroad on the sale of the U.S. property, they would have an advantage over U.S. businesses. Farming lobbies in the U.S. made this point forcefully in the late seventies,75 claiming that “… foreign investors in U.S. farmland get such good breaks they often can afford to outbid American farmers who want to expand their holdings”.76 The National Farmers Union was particularly concerned about tax treaties that granted additional avenues to avoid the capital gains tax. In their view, for example, the treaty being renegotiated with the UK at the time would contain “… a provision which would invite large-scale state income tax avoidance by foreign interests dealing in oil, grain, and commodities or investing in U.S. farmland”.77 The issue at hand was that the treaty prevented the U.S. from taxing foreign investors on the gain from the disposition of U.S. capital assets.78 U.S. tax on capital gains obtained by nonresidents The current Foreign Investment Real Property Act (FIRPTA) was enacted in 1980 to remove the perceived competitive advantage favoring foreign investors in the U.S. real property market.79 Under this law non-resident aliens would no longer be able to avoid U.S. tax on gains upon the direct sale of real property in the U.S. The statute defines real property as mines, wells and other natural deposits, ownership of land (or improvements), and options to acquire land80, and it taxes the sale of all directly held U.S. real property interests (RPIs), including those held by foreign residents, not just those for which the taxpayer received net basis taxation.81 It does not include, however, stock regularly traded on an established securities market, regardless of how much of its value may be represented by U.S. real estate holdings.82 FIRPTA taxes gain on disposition of the following defined U.S. RPIs: 1. Direct interests in real property located in the U.S.; 2. Interests in a domestic corporation which holds substantial U.S. real property83; 75 Petkun (1982), p.14 76 Spokane Daily Chronicle, May 8, 1978. See also Brown (2004), p.298. 77 Citing an unpublished working paper by the US Department of Agriculture, the press explained that, for example, a “… German investor often possesses the advantage of escaping from all capital gains taxes and does not relinquish the privilege of being treated identically with U.S. taxpayers in other respects”; meaning that the German investor did not get taxed on its gross operational income if it had consistently been considered as a passive foreign investor. Spokane Daily Chronicle, May 8, 1978 78 Petkun (1982), p. 27 79 FIRPTA principal provision: IRC, S 897. 80 Note that in the U.S. landowners also own what lies underground beneath their property. 81 Brown (2004), p. 305. 82 Petkun (1982), p.21. 83 A ‘real property holding corporation’ is defined as holding majority real property, which is marked to market. 61 3. Interests in domestic or foreign partnerships, trusts or estates with U.S. real property. Also, FIRPTA overrode treaties that exempt foreign residents from a capital gains tax on their U.S. RPIs in any of those three cases.84 FIRPTA does not alter the basic principle governing U.S. taxation of non-residents: all gains (and losses) from dispositions of directly held U.S. RPIs are treated as income effectively connected with a U.S. business and the foreign investor disposing of a U.S. RPI is deemed to be engaged in a U.S. business and thus taxed accordingly. Importantly, however, a foreign corporation can hold U.S. real property and the disposition of its stock by a foreign investor is not subject to U.S. tax; FIRPTA does not reach foreign indirect sales of U.S. property held by a foreign corporation.85 Indirect effect of FIRPTA on foreign investors According to some analysts, the fact that nonresident investors are not taxed on the capital gains from disposing shares in a foreign corporation holding U.S. real property does not mean that FIRPTA left a loophole.86 The argument is as discussed in the text in connection with the revenue consequences of OITs, and analyzed further in Appendix B: that transferring a foreign holding company that owns an appreciated U.S. RPI without paying the corresponding capital gains tax transfers the tax contingency to the purchaser, who will inherit the original cost basis of the underlying asset. Assuming that this underlying asset will eventually be directly disposed of in the local market in a subsequent transfer, the purchaser will consequently discount the price of the foreign shares representing indirect ownership of in their ability to transfer U.S. RPIs without recognizing taxable gain, the buyer of the foreign stock will the U.S. RPI. The analysis in Appendix B suggests that this will leave the initial seller exactly indifferent between direct sale and indirect sale of the U.S. corporation to a non-resident only under particular configurations of tax and other parameters. In the understanding that foreign corporations are restricted pay a lower price to take into account the future tax liability; “this pricing adjustment, if it occurs, will result in imposing the U.S. income tax on the seller indirectly”.87 In the words of another analyst: “… while the sale of stock in a foreign corporation holding U.S. realty is not taxable under FIRPTA, the foreign seller can be expected to 84 Starting 4 years after the enactment of FIRPTA in 1980, before article 13 (4) was introduced in the MTC. 85 Petkun (1982), p. 23; see also Doernberg (2012), p.112. 86 Petkun (1982), p. 23. 87 Petkun (1982), p. 23. 62 bear an indirect tax due to receipt of a reduced sales prices reflecting the corporation’s future tax liability.”88 B. Peru After the contentious case of Petrotech (described in Box 2), Peru passed legislation taxing all OITs, not just those whose value arises from immovable property located in Peru. The sale of an interest of any nonresident company whose value results at least 50 percent from shares of companies residing in Peru would be taxed in Peru. At least 10 percent of the parent foreign resident assets must be transferred for the tax to take effect89, thus, sales of retail investors abroad would not be affected. Box A.1: Peru’s Income Tax Law on offshore indirect sales of assets Art. 10.- “… it is also income from Peruvian source: e) That obtained from the indirect sale of shares or participations representing capital of legal persons residing in Peru. An indirect sale occurs when shares of participations representing the capital of a non-resident legal person that, in turn, is the owner - directly or through one or more intermediaries- of shares or participations representing the capital of a legal person resident in Peru, if … concurrently … 1. In any of the twelve months prior to the sale, the market value of the shares …of the resident entity … represent … at least fifty percent of the market value of all shares … of the non-resident entity. …………….. 2. In any period of twelve months, shares sold by the non-resident… represent at least ten percent of the capital of the non-resident entity. An indirect sale also occurs when a non-resident entity issues new shares …resulting from an increase in subscribed capital, new capital contributions …or a reorganization that diminishes their value below the market benchmark. In all cases, whenever the share sold, or the new shares issued … belong to an entity residing in low tax jurisdiction; it will be treated as an indirect sale.” C. China China’s approach to taxation of capital gains on transfers of interests differs in being structured as an anti-abuse provision. The general rule is that the gain on direct transfers of assets located in 88 Brown (2004), p. 299. The first condition was introduced with the Law No. 29663, February 2001; the second condition with Law 89 No.29757, July 2011. Peru’s domestic legislation taxing OITs can be overridden by double taxation treaties. 63 China is taxed at a 25 percent rate and offshore indirect transfers are equally taxed when involving the sale of immovable property located in China. In other cases, the taxation right on the indirect transfer of equity investment is sourced to the location of the investing enterprise. This means that a nonresident enterprise which owns another nonresident holding company which in turn invests in a Chinese company, would not be taxed in China on the gain resulting from the transfer of shares of the holding company; the capital gain is sourced in the location of the holding company.90 However, if the holding company is situated in a jurisdiction where the effective tax burden is lower than 12.5 percent or where offshore income is not taxed, the Chinese tax authority may disregard the overseas holding company and re-characterize the indirect transfer as a direct one if it determines that there is no reasonable commercial purpose to the offshore transaction other than avoiding the Chinese tax.91 Considerations determining whether a transaction fails the reasonable business purpose test include, for example, if i) the value of the asset directly transferred derives at least 75 percent (directly or indirectly) from Chinese taxable property; ii) the nonresident enterprise does not undertake substantive functions and risks; the tax consequence of the indirect transfer in the foreign country is less than the Chinese tax payable if the sale had been made directly.92 The Chinese approach to indirect transfers of assets is relatively defensive and discretionary: it taxes OITs when it deems that they have been structured to avoid the Chinese tax, without being taxed commensurately by another jurisdiction 90 Rule originally established in Notice No. 698, December 10, 2009 and replaced by Public Notice (2015) 7. 91 Article 47 EIT Law. Some exceptions apply, for example, sale of shares in the stock exchange. 92 Wei (2015). 64 APPENDIX D. ARTICLE 13.4 IN PRACTICE—AN EMPIRICAL ANALYSIS This Appendix describes and explores the presence or absence of Article 13.4 in (essentially the universe of) double tax treaties as of 2015.93 It also looks at country characteristics that affect the likelihood of including this provision in a DTT. Data and Variables The analysis covers 3,046 DTTs—which is almost the entire universe of active DTTs in 2015. Of these, 2,979 were recovered from the International Bureau of Fiscal Documentation (IBFD); the rest were recovered by internet search or from the ActionAid tax treaties dataset.94 About 35 percent of these treaties (973) include a provision that entitles the source country to tax gains from alienation of the capital stock of an entity the property of which consists directly or “indirectly” principally of immovable property. (We are unable, however, to distinguish meaningfully between adoption of UN and OECD versions). Further, about 35 percent of the treaties that involve at least one resource-rich country include article 13.4 as defined in this section (291 of 834 treaties) (Figure A1). Additionally, about 38 percent of the treaties that involve at least one low tax jurisdiction include Article 13.4 (Figure C.1). In modelling the likelihood of Article 13.4 being included in a treaty (conditional on the existence of a treaty) we make use of the following variables (summary statistics being in Table C.1): • Article 13.4, the dependent variable, is a dummy equal to one if a DTT includes article 13.4 and the word “indirectly” (using the UN or the OECD version or similar variants), and zero otherwise. Resource-rich low-income is a dummy equal to one if at least one signing country is a low- income or lower middle-income resource-rich country (as defined by the income classification of the World Bank and if revenues from resources exceed 10 percent). • CGTi – CGTj is the difference (in absolute value) between the concluding countries’ tax rates on capital gains. If a country’s tax code distinguishes between CGT for corporations and for personal tax purposes, we use the corporate CGT. The source is country reports 93 Ifa treaty contains this provision, in some cases of course it appears under a different number in the treaty (e.g., 13.2, 13.5, or 14.4). Further, it should be emphasized that, in this analysis, if a provision on the treatment of gains from immovable property is present in a DTT but does not explicitly state the word “indirectly”, it is not referred to as Article “13.4”. 94 At http://www.ictd.ac/datasets/action-aid-tax-treaties-datasets 65 published by Ernst and Young and Deloitte. A larger difference in CGT makes the use of a DTT for tax planning more attractive. Low tax is a dummy equal to one if the jurisdiction is a low tax jurisdiction in the sense of being included in the list of Hines and Rice (1994). There are 454 DTTs that involve such countries (almost 1/6 of the total).95 • Low tax × Low Income Res. is an interaction term between Resource-rich low-income and Haven. This variable enables us to test whether the role of low tax jurisdictions depends on the income level of the partner country. The idea is that Low tax can have an impact on Article 13.4 only if (i.e., conditional on the observation that) one signing country is a low- income country, but not if that country is an advanced economy. It is ultimately an empirical question whether or not this is the case. Year is the year of concluding the treaty. Essentially, it is a trend variable spanning from 1947 to 2015. Table D1: Descriptive Statistics VARIABLES N mean SD min max Article13.4 3,046 0.319 0.466 0 1 Year 3,046 1997 12.51 1947 2015 CGTi – CGTj 2,993 11.05 8.512 0 35 Low tax 3,046 0.149 0.356 0 1 Resource-rich low-income 3,044 0.105 0.306 0 1 Low tax × Low Income Res 3,044 0.00591 0.0767 0 1 95 That list includes 50 jurisdictions: Andorra, Anguilla, Antigua and Barbuda, Aruba, Bahamas, Bahrain, Barbados, Belize, Bermuda, Cayman Islands, Cook Islands, Costa Rica, Cyprus, Djibouti, Dominica, Gibraltar, Grenada, Guernsey, Ireland, Isle of Man, Jersey, Jordan, Lebanon, Liberia, Lichtenstein, Luxembourg, Macao, Maldives, Malta, Marshall Islands, Mauritius, Micronesia, Montserrat, Nauru, Netherlands Antilles, Niue, Panama, Samoa, San Marino, Singapore, St. Kitts and Nevis, St. Lucia, St. Martin, Vincent and the Grenadines, Switzerland, Tonga, Turks and Caicos Islands, Vanuatu, and Virgin Islands (British). 66 Analysis and Results Table A2 presents estimation results from two models using Article 13.4 as the dependent variable: A Linear Probability Model (LPM) in columns (1) to (3) and a logit model in columns (4) to (6). The results, shown in Table C.2, are discussed in the text. Table D2: The Likelihood of Including 13.4 in a DTT, Estimation Results (1) (2) (3) (4) (5) (6) The Dependent Variable Article13.4 Model LPM Logit Resource-rich low-income -0.0592** -0.0593** -0.0657** -0.065** -0.064** -0.079*** (0.026) (0.026) (0.0276) (0.027) (0.027) (0.027) CITi-CITj 0.0030*** 0.0042*** 0.0021** 0.0037*** (0.000) (0.000) (0.0009) (0.001) Low tax -0.133*** -0.164*** (0.024) (0.024) Low tax × Resource-rich -0.162* -0.164 low-income (0.084) (0.146) Year 0.0136*** 0.0135*** 0.0141*** 0.017*** 0.016*** 0.017*** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) Constant -26.88*** -26.76*** -27.80*** -180.5*** -177.4*** -195.8*** (1.063) (1.072) (1.092) (9.628) (9.609) (10.64) Observations 3,044 2,971 2,971 3,044 2,971 2,971 R2 0.134 0.135 0.146 Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1 As estimated coefficients in non-linear models cannot be interpreted as marginal effects, columns (4) to (6) display marginal effects (except for the constant). 67 Figure D1: Article 13.4 in DTTs with Resource-Rich Countries or low tax jurisdictions 68 REFERENCES Arnold Brian J. and Michael J, McIntyre, 1995, International Tax Primer (Den Haag, The Netherlands: Kluwer Law). Auerbach, Alan, 2002, “Taxation and Corporate Financial Policy,” pp. 1251-1292 in Alan J. Auerbach and Martin Feldstein (eds), Handbook of Public Economics, Vol. 3 (North Holland: Amsterdam). Auriol, Emmanuelle and Michael Warlters, 2005, “Taxation Base in Developing countries,” Journal of Public Economics, Vol. 89, pp.625-646. Beer, Sebastian and Jan Loeprick, 2017, “Profit Shifting in the Oil and Gas Industry.” Energy Economics, 2017, vol. 61, 186-198. Fred B. Brown, Fred B., 2004, “Wither FIRPTA?” Tax Lawyer, Vol. 57, No2, p.297. Burns, Lee, Honoré Le Leuch and Emil Sunley, 2015, “Taxing Gains on Transfer of Interest,” forthcoming in Philip Daniel, Michael Keen, Artur Swistak and Victor Thuronyi (eds), International Taxation and the Extractive Industries (Routledge) Cope, Charles W. and Parul Jain, 2015, “Taxation of Indirect Share Transfers: Recent Development in India and related policy considerations”, Tax Notes International Crivelli, Ernesto, Ruud de Mooij and Michael Keen, 21016, “Base Erosion, Profit Shifting and Developing Countries,”, FinanzArchiv, Vol. 72, pp. 268-301. Cui, Wei, 2015, “Taxation of Non-residents’ Capital Gains,” pp. 107-154 in Hugh Alt and Brian Arnold (eds) United Nations Handbook on Selected Issues in Protecting the Tax Base of developing Countries (United Nations). Daniel, Philip, Michael Keen and Charles McPherson (eds), 2010, The Taxation of Petroleum and Minerals: Principles, Practices and Problems (Routledge). Desai, Mihir A. and James R. Hines Jr. 2003, “Evaluating International Tax Reform,” National Tax Journal, Vol. 56 (September), pp. 409–40. Doernberg, Richard L., 2012, International Taxation in a Nutshell, West publishing. Hearson, Martin, 2014, “Capital Gains Tax Avoidance: Can Uganda Succeed Where India Didn’t?”, September 23, available at https://martinhearson.wordpress.com/2014/09/23/capital- gains-tax-avoidance-can-uganda-succeed-where-india-didnt/ Hearson, Martin (2016), 'The ActionAid Tax Treaties Dataset'. Brighton: Institute of Development Studies. http://ictd.ac/datasets/action-aid-tax-treaties-datasets Hines, James R. and Eric M. Rice, 1994, “Fiscal Paradise: Foreign Tax Havens and American Business,” Quarterly Journal of Economics, Vol. 109, pp. 149-182. International Monetary Fund, 2014, “Spillovers in International Corporate Taxation,” Available at www.imf.org/external/np/pp/eng/2014/050914.pdf 69 International Monetary Fund, Organisation for Economic Cooperation and Development, World Bank Group and United Nations, 2015, Options for Low Income Countries’ Effective and Efficient Use of Tax Incentives for Investment. Available at https://www.imf.org/external/np/g20/pdf/101515.pdf Kane, Mitchell, 2018, “Offshore Transfers: Policies and Divergent Views,” forthcoming in the Bulletin of International Taxation. Krever, Rick, 2010, “Tax Treaties and the Taxation of Non-residents’ Capital Gains,” pp.212-238 in Arthur Cockfield (ed.) Globalization and its Discontents: Tax Policy and International Investments (University of Toronto Press: Toronto). Organization for Economic Cooperation and Development, 2014a, Report to the G20 Development Working Group on the Impact of BEPS in Low Income Countries, Part 1 available at http://www.oecd.org/tax/part-1-of-report-to-g20-dwg-on-the-impact-of-beps-in-low- income-countries.pdf. Organization for Economic Cooperation and Development, 2014b, Report to the G20 Development Working Group on the Impact of BEPS in Low Income Countries, Part 2 available at https://www.oecd.org/g20/topics/taxation/part-2-of-report-to-g20-dwg-on-the-impact- of-beps-in-low-income-countries.pdf. Peters, Carmel, 2015, “Developing Countries’ Reactions to the G20/OECD Action Plan on Base Erosion and Profit Shifting,” Bulletin for Fiscal Documentation, June/July, pp.375-381. Petkun, Lisa B., 1982, “The Foreign Investment in Real Property Act of 1980”, Penn State International Law Review: Vol 1: Issue 1, Article 2. Sinn, Hans-Werner, 1991, “The Vanishing Harberger Triangle,” Journal of Public Economics, Vol. 45, pp. 271-300. Toledano, Perrine, John Bush and Jacky Mandelbaum, 2017, “Designing a Legal Regime to Capture Capital Gains Tax on Indirect Transfers of Mineral and Petroleum Rights: A Practical Guide,” International Senior Lawyers Project and Columbia Center of Sustainable Development; at http://ccsi.columbia.edu/files/2017/10/Designing-a-legal-regime-to-capture-capital- gains-tax-CCSI-ISLP-Oct-2017.pdf. United Nations, 2015, Note on Capital Gains Taxation and Taxation of Indirect Asset Transfers, available online at http://www.un.org/esa/ffd/wp- content/uploads/2015/10/11STM_Attachment2_Cgt.pdf. United Nations, 2016, Manual for the Negotiation of Bilateral Tax Treaties between Developed and Developing Countries, United Nations, p.108 (New York, 2016) Waerzeggers, Christophe and Cory Hillier, 2016, “Introducing a General Anti -avoidance Rule (GAAR)—Ensuring that a GAAR Achieves its Purpose”, Tax Law IMF Technical Note 2016/1, IMF Legal Department. 70 Wei, Jinji, 2015, “China Extends Tax Rights on Nonresidents’ Capital Gains from Indirect Transfers”, International Tax Notes, February 9. Wijnen and de Goede (2014), “The UN Model in Practice 1997-2013,” Tax Treaty Monitor, Brian J. Arnold in Cooperation with IBFD’s Tax Unit Treaty. 71