Capital gains tax avoidance: can Uganda succeed where India didn’t?

Zain

Uganda is pursuing Zain for $85m capital gains tax on the indirect sale of its Ugandan subsidiary

I’m writing this post from under a mosquito net on a close Kampala evening. Since arriving on Wednesday I’ve had a whistlestop tour of the issues facing Uganda as it embarks on a review of its tax treaties. So far I’ve met with four tax inspectors, two finance ministry officials, four (count ’em) tax advisers, one academic and three NGO people. I also spoke at an event to launch a a report on Uganda”s tax treaties written by Ugandan NGO SEATINI and ActionAid Uganda.

This post is about “indirect transfers” of assets, where a sale is structured to take place via offshore holding companies, thus escaping capital gains tax. It turns out there is an $85m tax dispute on this between Uganda and the mobile phone company Zain. This is just about the biggest issue in Ugandan tax right now: the tax inspectors are even tweeting about it.

“Indirect transfers” were highlighted in the recent (and generally solid, I thought) OECD report to the G-20 development working group [pdf].* It says:

Developing countries report that the profit made by the owner of an asset when selling it (for example, the sale of a mineral licence) is often not taxed in the country in which the asset is situated. Artificial structures are being used in some cases to make an ‘indirect transfer’; for example through the sale of the shares in the company that owns the asset rather than the sale of the asset itself.

Unfortunately, it is pretty lame on the solutions. As far as I can tell from the G-20 response [pdf], what is going to happen on it is this:

(deep breath…)

As part of its multi-year action plan, the G-20 development working group will consider calling on the OECD, in consultation with the IMF, to report on whether further analysis is needed.

(…and exhale)

I don’t hear the sound of tax positions unwinding.in response to that one.

To remind you, the big daddy of indirect transfer cases is the Vodafone-India dispute. In that case,  according to this handy summary:

In 2007, Vodafone’s Dutch subsidiary acquired the stock of a Cayman Islands company from a subsidiary of Hutchinson Telecommunications International Ltd. (the subsidiary was also located in the Cayman Islands). The purchase price was $11.1 billion. The Cayman company acquired by Vodafone owned an indirect interest in Hutchinson Essar Ltd. (an Indian company) through several tiers of Mauritius and Indian companies.

Like India, Uganda is trying to tax the sale of a mobile phone company when the transaction took place via offshore holding companies:

Zain International BV owned Zain Africa BV, which had equity in 26 companies all registered in the Netherlands, but effectively owning the telephone operator business in as many African countries. One of them, Celtel Uganda Holding BV, owned 99.99 per cent of the Kampala-registered Celtel Uganda Ltd. On March 30, 2010 Zain International BV sold its shares in Zain Africa BV to Bharti Airtel International BV. As all three companies are registered in the Netherlands, and as the transaction was a sale of shares rather than assets, the company said it did not attract capital gains tax.

The cases are of course not identical. For one thing, Uganda is going after the firm that actually made the capital gain. But the Indian jurisprudence is being used in the Ugandan case.

Just last week, an appeal court ruled that the Uganda Revenue Authority does have the jurisdiction to assess and tax Zain on the gain. Zain will now argue that the transaction was exempt. One of its core arguments is sure to be the Netherlands-Uganda tax treaty.

In common with 86% [pdf] of tax treaties signed by developing countries since 1997, this treaty does not contain the UN model treaty provision that would have allowed Uganda to tax gains on the sale of shares in Ugandan companies made by Dutch residents. It may be that Celtel Uganda counts as a ‘property rich’ company because of all its infrastructure assets, in which case Uganda would have been able to fall back on the OECD and UN model provision permitting it to tax those…except (oops!) even that isn’t included in its treaty with the Netherlands. Yes, this treaty is worse for Uganda than the OECD model, never mind the UN.

So instead we come to Section 88(5) of Uganda’s Income Tax Act [pdf] . This is an anti-treaty shopping provision, which denies the benefits of the treaty to a company whose ‘underlying ownership’ is mostly in a third country:

Where an international agreement provides that income derived from sources in Uganda is exempt from Ugandan tax or is subject to a reduction in the rate of Ugandan tax, the benefit of that exemption or reduction is not available to any person who, for the purposes of the agreement, is a resident of the other contracting state where 50 percent or more of the underlying ownership of that person is held by an individual or individuals who are not
residents of that other Contracting State for the purposes of the agreement.

Sounds like Uganda has it in the bag, right? Unfortunately, this matter will turn on whether Uganda’s domestic law can override its treaty commitments. It is quite likely (certain, if you ask Zain’s tax adviser) that a court will decide it cannot. What everyone I have spoken with agrees on (apart, perhaps, from Zain’s tax adviser) is that it would be preferable to have some certainty about this unresolved question.

The URA has recently begun denying treaty benefits under section 88(5), and until now taxpayers have accepted its reasoning. But, speaking in genera terms at the SEATINI/ActionAid public meeting, a tax official said that the URA doesn’t know if its position will stand up to a court challenge. Tax advisers in the private sector say that, as well as the question of treaty override, the meaning of “underlying ownership” needs to be clarified. Because the Zain case has so far been fought on technicalities, “we were robbed of the opportunity to see how it [Section 88(5)] would work in practice,” one told me.

Perhaps the next stage of the Zain case will answer this question. If it does, it should give some welcome guidance to developing countries struggling with these indirect transfers. If they can’t use their domestic law to override their treaties, they will need to insert an anti-abuse clause into their treaties, strengthen their source taxing rights, or consider cancelling them.

This brings us back to BEPS, and the action on tackling treaty abuse. The OECD is proposing a limitation of benefits clause based on that used by the US, which is similar to that in Uganda’s domestic legislation, only a lot more detailed about who is ruled in and out. This would do the trick, but the challenge would be getting it into treaties that have been already signed.

To solve that, the OECD is pushing a multilateral convention to modify treaties all at once, built on a flexible level of commitment. It concedes [pdf] that the multilateral instrument “has not been identified as high priority by developing countries.” For it to work for them, I think it would need two things:

1. Genuine flexibility so that developing countries can opt into only the bits they want, such as the anti-abuse clause.

2. Willingness on the part of high-risk jurisdictions for treaty shopping (in Uganda’s case the Netherlands, Mauritius, and perhaps now the UK) to opt in to the anti-abuse clause as well.

For Uganda, it might not make sense to wait for this, since we are only talking about two or three treaties. It could ask its partners for a protocol containing a limitation of benefits clause right now. Or, of course, it protect itself and raise more revenue by strengthening all its treaties’ capital gains articles, as the UN model provides for in the first place.

*thanks to @psaintamans for the link!

Some heretical thoughts on automatic information exchange and developing countries

Information exchange agreements are one of the main tools open to governments in the fight against tax evasion. They allow tax authorities to investigate tax(non-)payers’ affairs across borders, uncovering unreported income and piecing together tax planning structures. They’re also the main focus of coercive efforts by big economic players to “crack down on tax havens.”

As in many areas of international economic policy, what developing countries need isn’t the same as what developed countries do. So while tax justice campaigners are celebrating the G20’s weekend declaration on information exchange, I’m a little more sceptical about what it will mean for tax and development. This post is also by way of a comment on Owen Barder and Alex Cobham’s proposed G8 initiative on tax transparency.

Quick background

Information exchange comes in three flavours:

  • On request, the current lowest common denominator in agreements, where one country’s tax authority must fill in a form for each taxpayer about whom it wants information, setting out and justifying its demand from another country.
  • Spontaneous, where tax authorities pass on information of interest to another that they uncover in the process of an investigation.
  • Automatic, touted by campaigners as the gold standard, under which information is transmitted in bulk between authorities. This has the advantage that a tax authority will find out about a taxpayers’ undeclared overseas income without needing to suspect its existence beforehand.

Taxpayer confidentiality means that countries can’t usually do any of this without the legal authority of a treaty. Treaties can come in two forms:

  • Bilateral, either as part of a double tax treaty or as a standalone tax information exchange agreement
  • Multilateral, of which there are several examples including an EU Directive and the Multilateral Convention discussed below.

G20: what’s going on

So, to the G20. Saturday’s Finance Ministers’ communiqué [doc] says:

we also strongly encourage all jurisdictions to sign or express interest in signing the Multilateral Convention on Mutual Administrative Assistance in Tax Matters and call on the OECD to report on progress. We welcome progress made towards automatic exchange of information which is expected to be the standard and urge all jurisdictions to move towards exchanging information automatically with their treaty partners, as appropriate.

This latter sentence is a gradual strengthening of language, from this in November 2011 [doc]:

We welcome the commitment made by all of us to sign the Multilateral Convention on Mutual Administrative Assistance in Tax Matters and strongly encourage other jurisdictions to join this Convention. In this context, we will consider exchanging information automatically on a voluntary basis as appropriate and as provided for in the convention

Via this in June 2012 [doc]:

We welcome the OECD report on the practice of automatic information exchange, where we will continue to lead by example in implementing this practice. We call on countries to join this growing practice as appropriate and strongly encourage all jurisdictions to sign the Multilateral Convention on Mutual Administrative Assistance.

I’m not sure of the difference between “call on” and “strongly encourage” in diplomatic language, but these evolving declarations suggest a couple of trends in the G20’s coercive agenda. First, automatic information exchange is moving from something voluntarily towards becoming “the standard.” Second, the OECD is now going to be reporting back to the G20 on who has joined the Multilateral Convention on Mutual Administrative Assistance in Tax Matters, an OECD/Council of Europe collaboration that is now open to all comers. (I wrote a sceptical post about the Multilateral Convention when I was at ActionAid, and while I won’t recap the points I made there, I was never reassured about the convention’s governance.)

It seems, then, that a direction of travel has been mapped out towards an international standard in which jurisdictions exchange information automatically through the Multilateral Convention, monitored by the OECD and under pressure from the G20. This is quite consistent with Alex and Owen’s proposal, under which automatic exchange among the G8, their dependencies and overseas territories, would be “the foundation of a global system for automatic exchange of tax information between all countries who wish to participate in it.”

Coercion and developing countries

We can be sure that the G20 pressure is targeted at tax havens, not developing countries. Note that Alex and Owen, above, emphasise countries “who wish to participate in it.” In practice, however, many developing countries will inevitably be caught up in the momentum. They are already participants in some of the key international instruments. For example, in sub-Saharan Africa, where tax administrative capacity is weakest:

So what’s the risk? I’m worried that, if they adopt the automatic information exchange standard any time soon, overstretched revenue authorities in developing countries will have to undergo a massive reprioritisation of effort towards putting in place systems to enable them to supply information automatically. This when they already struggle to find the capacity for, for example, transfer pricing audits. It’s hard to imagine them finding the resources to both comply with the demands placed on them by automatic exchange, and to make use of the data they receive automatically. (The same concern, at a lower level, might apply to developing countries’ existing commitments to information exchange on request: what’s the compliance burden placed on a developing country by the requests it receives from treaty partners, and by the Global Forum’s peer review process?). Alex and Owen have this covered when they suggest that:

The authorities of developing countries will be able to access the information shared by our tax authorities from the outset, even if their own tax authorities are not yet themselves in a position to share information automatically.

I’ve also heard people argue that automatic information exchange might be the kind of motivator that tax authorities and governments in developing countries need to push them to get serious about tackling tax evasion, given that many of its biggest exponents are influential members of elites. Well, that’s possible. But wouldn’t it make more sense to abandon talk of information exchange with developing countries and focus on information transfer? That way they’d have all the information without such a large compliance burden. And an asymmetrical arrangement is easily justified from a development perspective.

Source-based information exchange?

Aside from the compliance burden, another question is the basis on which information is exchanged. Automatic exchange is designed to benefit the country in which a taxpaying individual or company is resident, since the information flows to there from the other countries in which she earns her income. This is also the basis of Alex and Owen’s proposal. It would certainly be useful for developing countries in the context of tax haven abuse by their individuals and medium-sized domestic companies, but it is of no use to them in investigating the tax affairs of foreign investors. Maybe that’s not the brief, but Alex and Owen’s proposal does say:

We attach particular priority to ensuring that developing countries are able to access and use this information to enable them to collect taxes which are legally due from companies and individuals within their jurisdiction.

This got me thinking. What would an automatic information exchange system designed to support source-based taxation of multinationals look like? There is at least one piece of information that could flow in the opposite direction, from residence to source country.

Consider Tax Authority A investigating Company A, and Tax Authority B investigating Company B. Company A sells widgets to Company B. Under existing automatic exchange standards, no information would be exchanged, unless Company A was owned by Company B, or vice versa. Alex and Owen’s proposal extends this, as follows:

all participating authorities will collect data from financial institutions and companies about income, gains, and property paid to non-resident individuals, corporations, and trusts, which they will provide automatically to the jurisdiction in which that individual or company is resident.

Under this version, Tax Authority B would transfer information to Tax Authority A about the payment made by Company B to Company A for the widgets. It’s an improvement, because at the moment companies often force tax authorities to jump through legal hoops to obtain this information via treaty requests. That’s because even if Companies A and B are part of the same multinational, they are separate companies, and Tax Authority A can’t ask Company A for Company B’s financial information.

The one complication is that neither tax authority in this scenario would necessarily know whether Company A and Company B were related parties. Sometimes tax authorities in developing countries struggle to prove that an overseas party with which the company under audit is transacting is a related party at all, especially if their tax law defines related party in a way that is hard to prove.

This information would in fact be available to Tax Authority C in the multinational’s parent country, which should (or at least could) have a list of all parent Company C’s subsidiaries overseas. Source-based information exchange, limited to the case of multinational companies, would complement residence-based exchange. It would involve the tax authority in the country of the parent transferring information it gains through auditing the parent on to tax authorities in the countries where the multinational works.