Time we scrutinised China’s tax treaty practice, too

Democracy in action: David Gauke at Monday's delegated legislation committee session

Democracy in action: David Gauke at Monday’s delegated legislation committee session

On Monday the UK parliament took a total of 17 minutes to scrutinise new tax treaties with Zambia, Iceland, Germany, Japan and Belgium. I’ve complained before about how paltry these debates tend to be, and was all set for another blog along those lines. There was, indeed, much to grumble about. No questions from the opposition about the UK’s renegotiated treaty with Zambia at all, a week after the IMF warned that developing countries should exercise “considerable caution” when entering into tax treaties.

Instead, Labour’s Shabana Mahmood asked how the UK’s treaty making priorities were set, and why there is no treaty with Brazil. The response from the Minister David Gauke was considerably less informative than what I’m sure Mahmood could have found out by asking, say, her colleague Stephen Timms, Gauke’s predecessor.

But something interesting did come up when Gauke was introducing the Zambia treaty. He noted that the withholding tax rates have been reduced in line with Zambia’s treaty with China. And indeed they have. It seems to be China, often regarded as the champion of source state taxation at the UN tax committee, which is responsible for the lower withholding tax rates. I’m going to explain here why I think both the UK and China have questions to answer about these treaties.

The UK-Zambia renegotiation: a missed opportunity

The UK-Zambia renegotiation looks like a ‘balanced package’, meaning that Zambia will have gained and lost in roughly equal measure. Looking at the treaty, I don’t think it can be seen as a win for Zambia.

What it lost was withholding tax rates. Zambian tax on dividends to British portfolio investors will be reduced under the new treaty from 15% to 5%, and tax on royalty payments for the use of British intellectual property will drop from 10% to 5%. This matches what’s in the 2010 Zambia-China treaty [pdf], so it looks like Britain was keen to keep its investors competitive relative to their Chinese competitors.

By way of context, Zambia’s non-treaty rates are much higher, 15% and 20% respectively. We can argue about the economic case for this level of withholding tax, but treaties are not just about rates, they’re about the right to raise rates. It will be five years before Zambia can re-examine these low withholding rates in its treaty with the UK.

What Zambia got in return for the reduced withholding tax rates was the UN concept of services permanent establishment, which will allow it to tax services provided within Zambia by British businesses or individuals. To do so, Zambia won’t need them to have a physical fixed base in Zambia, as it would have done before, but it will need them to be physically in Zambia, furnishing services, for at least 183 days in a given year.

(There are also some modernising changes, which may in practice benefit Zambia more than the UK. This includes simple anti-abuse wording such as the “beneficial owner” clause in the withholding tax articles and a “property rich companies” clause into the capital gains article. It also includes information exchange and assistance in recovery articles. These should be good for Zambia, if it takes advantage of them. The information exchange clause could, for example, allow Zambia to get hold of country-by-country reporting on British companies if that proposal is implemented by the OECD.)

But the overall picture, taking into account the lower withholding taxes, is of a treaty that is still much more disadvantageous to Zambia than one based on the UN model would have been. I’m not even sure it’s a better position than the OECD model. Since Zambia was not nearly as aggressive at negotiating after independence as, say, Kenya, it started this renegotiation from a lower base: already low withholding taxes, no taxing rights over British airlines, limited capital gains tax rights, and no right to tax management fees, to name a few examples.

In a context in which some countries are re-examining their tax treaties with developing countries, and organisations such as the IMF are calling into question the benefit of tax treaties on current terms, there would have been a strong case for the UK to seek not a balanced negotiation, but a reapportionment of taxing rights towards Zambia, in line with the UN model. It’s a real shame that the treaty slipped through parliament on Monday without anyone at least asking about this.

China is driving the falling withholding tax rates

This argument for a more pro-source taxation treaty between the UK and Zambia would be easier to make if the 2010 China-Zambia treaty had been more generous. But in fact the terms of the two treaties are near identical. On the face of it, it seems quite likely that Zambia has been bounced into this renegotiation to help keep British mining, agriculture and manufacturing companies more competitive in the face of competition from China. These companies will benefit from the lower withholding tax rates but are unlikely to be affected by the services permanent establishment quid pro quo.

The IMF report talked about “strategic spillovers” from tax policy, in which one country’s policy pushes other countries towards a response. I’m now starting to wonder if China’s negotiating stance might be having just such a strategic spillover, contributing to the decline in withholding tax rates in treaties also picked up by the IMF. Below you’ll see that China’s treaties with sub-Saharan countries have the lowest withholding taxes in a sample of countries investing into Africa.

Withholding taxes in treaties with sub-Saharan countries, 1973-2012

Withholding taxes in treaties with sub-Saharan countries, 1973-2012

China’s treaties are newer than other countries’, so what if this trend is just an artefact of the general decline? Not so if we look at just the last 20 years, where the story is much the same, with only Mauritius (which has several zero withholding tax treaties) having a more advantageous treaty network.

Withholding taxes in treaties with sub-Saharan countries, 1993-2012

Withholding taxes in treaties with sub-Saharan countries, 1993-2012

Let’s now test whether those three Chinese treaties in Africa are typical of China’s treaties more generally. This time we’re looking at all low-income countries. Not only is China the largest signatory of treaties with this group among my sample, it also emerges as one of the most demanding negotiators.

Withholding tax rates in treaties with low-income countries, 1993-2012

Withholding tax rates in treaties with low-income countries, 1993-2012

Withholding taxes are of course only one part of the source-residence balance in a tax treaty. I took a quick look at the China-Africa treaties, and – aside from the services permanent establishment.- there is no sign that they include pro-source provisions such as withholding taxes on management fees, or a “limited force of attraction”. It’s been well-documented that China favours expansive source taxation in its treaties with outward investors, while denying them to capital-importing developing countries.

The UK-Zambia treaty seems to be an example of a strategic interaction between two countries, one (the UK) with a longstanding investment base in Zambia, and the other (China) posing a threat to that investment. It’s all very well to criticise countries like the UK for not being more generous in negotiations with developing countries, but in doing so, critics should be careful not turn a blind eye to countries outside the OECD, who may even be the ones leading the race to the bottom.

Thoughts on Deloitte and the China-Mauritius-Mozambique route

Sunday’s Observer carried a story, prompted by ActionAid, based on a presentation given by Deloitte to a group of Chinese investors. The presentation explained how to avoid withholding tax and capital gains tax in Mozambique by routing the investment through Mauritius. It’s great to see this kind of common or garden tax arbitrage highlighted and controversialised.

What caught my eye was Deloitte’s response. The company said:

It is wrong to describe applying double tax treaties, such as the treaty between Mauritius and Mozambique, as tax avoidance. Such treaties are freely negotiated between the Governments of the countries involved.

Double tax treaties exist to enable the countries concerned to strike a balance between the need to encourage investment, including cross-border investment, to raise tax revenue, and to work together with other countries who have the same legitimate concerns to raise revenue and promote business.

The absence of such treaties could result in a reduction of investment, and less profit subject to normal business taxes in the countries concerned.

Any discussion of tax treaties by tax professionals would typically be around the technical and administrative aspects of the treaties and not an expression of favour of any particular country at the expense of any other country.

Leaving aside the questionable empirical basis of the tax treaties-investment link, what interests me is the way the statement completely glosses over the difference between “applying double tax treaties”, and treaty shopping, i.e. structuring investments through an intermediate jurisdiction like Mauritius in order to obtain more preferential treaty benefits. It raises a couple of questions for me.

First, is that obfuscation just a good PR strategy, or is it the case that tax advisers don’t see this distinction as valid? Is advice on tax treaties always aimed at getting the best treaty rates available, with no conceptual (or ethical) difference between a direct investment and one via a tax treaty conduit?

Second, how can we say whether this is tax avoidance or not? At first sight, I’d argue that the intention of the Mozambique-Mauritius treaty is to provide benefits to Mauritian investors in Mozambique, and vice versa, while the absence of a Mozambique-China treaty reflects those governments’ intention that a Chinese firm investing in Mozambique should incur taxes at the non-treaty rates. In that interpretation, tax treaty shopping contravenes the intention of the treaties and is tax avoidance.

But it’s more tricky than that. It’s technically fairly easy to include an anti-abuse clause in a treaty, to spell out this intention. If there isn’t one, it might be because of poor negotiation by Mozambique, or Mauritian unwillingness to renegotiate, but it might also be that Mozambique intends investors from other jurisdictions to use Mauritius as a route to invest, since this reduces the perceived need to negotiate treaties with every potential source of investment. After all, Mozambique only has a handful of treaties, and its Mauritius treaty was signed just after Mauritius adopted the offshore regime that creates the problem. Meanwhile, China is a pretty aggressive negotiator, and those African countries that have signed treaties with China seem to have ended up with less taxing rights than Mozambique has from its treaty with Mauritius. So Mozambique may actually be better off letting Chinese investors exploit its treaty with Mauritius, rather than negotiating a treaty with China. Though it would be better off still in terms of taxing rights if it had neither!

And what about China? It only has a few treaties with African countries, but it does have a treaty with Mauritius. China has foreign tax credits, so the less its multinationals pay abroad, the more revenue it gains (of course a lot of its overseas investment is by state-owned enterprises, and there again, tax savings abroad go straight into government coffers). So maybe China doesn’t see any urgent need to change the status quo. That said, at least half of the African countries with which it has signed tax treaties also have treaties with Mauritius, which suggests a preference for its multinationals investing directly.

One of the main issues in all this is that we don’t know what developing countries intend when they negotiate treaties. To make matters worse, in many developing countries, including (I think) Mozambique, the executive has historically had the power to ratify treaties. So there’s no ‘will of parliament’ to look for, and no public record of any debate among decision-makers. Mozambique’s treaty was signed almost 20 years ago. If my experience talking to officials in other countries is anything to go by, it will be hard to find anyone who can remember how and why this treaty came about. The tax landscape has changed in the meantime, as has the economy, not least with the growth of cross-border services. This would be a good time for Mozambique to review its treaty network.

UN transfer pricing manual: what Brazil, India and China do differently

‘Country Practices’. The title of Chapter 10 of the new United Nations Practical Manual on Transfer Pricing [pdf] doesn’t exactly set the pulse racing. But as I noted in my blog on the manual as a whole, this document is politically very significant.

It’s probably the only detailed description of Brazil, China, India and South Africa’s approach to transfer pricing, both how they follow and how they differ from the OECD methods. Significantly, these contributions are expressed not just in terms of the legal and administrative realities, but also the policy objectives underlying them. Chapter 10 functions as a comprehensive critique of the OECD guidelines – almost a manifesto – endorsed and in most cases written by tax officials from some of the world’s most powerful economies.

So what’s in it? Here are some digested highlights – but you should read it all.

Brazil: fixed margins

Since the mid 1990s, Brazil has been the world’s transfer pricing maverick. It makes no claim to follow the OECD guidelines, although the core of its approach is analogous to some of the methods outlined in those guidelines. Brazil views its approach as consistent with the arm’s length approach, since it’s another way to approximate the price that would be paid between companies trading at arm’s length. Many others disagree with this view.

Under the OECD approach, the taxpayer (and tax authority, if it is challenging a taxpayer) needs to identify a ‘comparable’ company or transaction for each and every transfer price under assessment, subject to some adjustments. The simplest method is to find two independent companies that are trading a similar good or service, and use that price. Brazil adopts this method too.

If that can’t be done, some other OECD methods (Cost Plus and Resale Minus) use comparable profit margins, applying them to the price paid when the good or service under consideration is eventually bought from or sold to a third party. (There are other OECD methods that are even more complicated, but they’re not used in any form by Brazil). The Brazilian approach uses the same idea, but prescribes fixed profit margins.

In a simple example, imagine a Canadian mining company operating in Brazil. There are two subsidiaries, a mine in Brazil, and then an intermediary in Bermuda that buys the minerals and sells them on to third parties. To apply the OECD’s Resale Minus approach, you’d find a comparable commodity trader that buys minerals from third parties and sells them on to other third parties. You’d look at the profit margin it makes, then apply that same profit margin to the actual price at which the minerals from Brazil were sold on by the Bermudan company to a third party, to obtain the price at which the Bermudan company bought the minerals from its sister company in Brazil.

Brazil’s Resale Price Method is similar, except that rather than looking for a comparable, you apply a fixed margin of (usually) 20% to the actual price at which the minerals were sold on to third parties. In other words, the transfer price from Brazil to Bermuda would be 80% of the resale price from Bermuda to a third party. Last year Brazil made its method a bit more fine-grained, by setting out different margins for different sectors, based on data about each sector.

The Brazilian section of Chapter 10 argues that this method is easier to apply and provides more certainty than the OECD approach. It acknowledges that the approach may create double taxation because it’s not compatible with other countries that use the OECD methods, and that, “it is unavoidable that some Brazilian enterprises will be taxed at (higher or lower) profit margins not compatible with their profitability.”

China, India and South Africa

“As a developing country, China faces a number of difficult challenges, to many of which ready answers have not been found from the OECD guidelines,” notes the Chinese section of Chapter 10, written by two senior tax officials. South Africa’s submission concurs:

Whilst the OECD Guidelines have been particularly useful in providing a conceptual understanding of what is the nature of the arm’s length principle, there are instances when the Guidelines fail to address the more practical aspects of how to apply the principle

According to all three of these countries, the difficulty is with how to implement the arm’s length principle. And their sections of the manual set out many practical problems with implementation. But I think that there’s something more than that at stake here. These countries (or at least India and China) want to claw back a larger share of the tax base, which is about changing the apportionment between themselves and other countries. They may argue that their position is a more accurate implementation of the arm’s length principle, but to do so is to understate what they’re trying to achieve. Below are three examples.

Location specific advantages

The premise of the Location Specific Advantage (LSA) approach is that investments by multinational companies in these countries are more profitable than those in other countries, as a result of LSAs such as a cheap, comparatively skilled labour force, a large, relatively untapped consumer market, and more lax environmental legislation. This needs to be take into account when identifying and using comparables that do not have these same advantages.

To determine the transfer price by reference to a comparable transaction or company in another country, “China takes the view that there may be instances where the differences in geographical markets are so material that it warrants comparability adjustments to bridge the differences.” In other words, China revises up the profit made by Chinese subsidiaries, and hence the tax charge, when comparing them to other countries, because it thinks the LSAs it offers make investments in China more profitable than those in other countries.

South Africa says it shares this concern:

There are many instances where unique dynamics exist within the South African market enabling South African subsidiaries to realise higher profits than their related party counterparts in other parts of the world, or than are evidenced by comparable data obtained from foreign databases…Building on the practice followed in India and China, the SARS is currently considering its approach to location savings, location specific advantages and market premiums etc. within certain industries and such factors will be addressed when conducting audits.

The Indian section also discusses location specific advantages. Its argument, however, relates to whether local comparables, as well as foreign ones, are valid. Although the chapter doesn’t say this, I think India’s argument is also that the arm’s length principle doesn’t work here, because there is no arm’s length scenario that captures the location specific advantages:

Hypothetically, if an unrelated third party had to compensate another party to the transaction in a low-cost jurisdiction by an amount that was equal to the cost savings and location rents attributable to the location, there would be no incentive for the unrelated third party to relocate business to a low-cost jurisdiction.

What makes the LSA concept interesting is not merely that these countries are laying claim to a larger share of MNCs’ tax base than the OECD guidelines attribute to it. They are saying, quite explicitly in China’s case, that under free market (“arm’s length”) conditions, they don’t receive a fair share of the profits from inward investment, because the LSAs aren’t fairly priced by the market. China is using the tax system to correct for what it sees as unfair conditions in the global economy.

Below is a presentation by another Chinese official at a Tax Justice Network conference last year. Scroll through to slide 20 for the part on LSAs.


Transfer pricing is supposed to start from a ‘functional analysis’ that takes into account functions, assets and risks to determine how profit should be allocated. But intangible assets (and risks, see below) can be more easily moved than functions and tangible assets. China and India are concerned that multinationals tend to characterise subsidiaries in their economies as exploiting foreign-owned intellectual property, on which they must pay royalties, in order to deflate the profits made there.

For example, China takes the view that the value of a marketing intangible (such as a brand name) is inextricably linked to the market in which it is sold. In a memorable example, Head & Shoulders is a popular shampoo in China, and a foreign brand. But when it first arrived on the Chinese market, most consumers didn’t know what the words “Head and Shoulders” meant. “Over time,” says the manual, “the local Chinese affiliates acquire the skill and experience from operations in China, and may even contribute to the improvement of the MNE’s original intangibles. The issue in this scenario is whether the local Chinese affiliates should be entitled to additional profit, and if so, what is the appropriate method to calculate the additional profit?”

The Indian approach is similar:

Indian subsidiaries/related parties (which are claimed as no risk and limited risk bearing distributors by the parent MNE in order to justify low cost plus return) have incurred and borne huge expenditure on development of marketing intangibles. These entities generally incur very large losses or disclose very nominal profit as evident from their return of income

Instead of a flow of royalty payments out of India, it argues, the parent company should be reimbursing its Indian subsidiary for the local marketing intangibles created, without which it couldn’t sell into India.

Risks, and the formulary apportionment bombshell

India is concerned that the OECD approach overvalues the role of risks in allocating profits. Risk plays an important role in the functional analysis that underpins transfer pricing, because the higher the risk taken by one part of a company, the higher the transfer price it can charge. The Indian section says risks are “a by-product of [a company’s] performance of functions and ownership and the exploitation or use of assets employed over a period of time.” It “does not agree with the notion that risk can be controlled remotely by the parent company” and says that if “important strategic decisions” are taken in India then “the allocation of risk to the parent MNE is not only questionable but is devoid of logical conclusion.” Ouch.

The Chinese section concurs, arguing that “a risk‐based approach may place insufficient regard for the fact that there are sizeable assets located in China (i.e. the work force and factory plants).” It goes as far as to conclude that, in the case of the electrical manufacturing sector, “In this case, the assets and the people should largely dictate where the group’s profits should stay, and a global formulary approach should be a realistic and appropriate option.”

This is quite a bombshell. The chapter effectively argues that, in order to reduce the distorting effect of risks and intangible assets in transfer pricing, there is a case for the limited use of the formulary system advocated by Tax Justice Network, distributing the profits based on only on the assets and people in each jurisdiction. Wow. That will no doubt get tax justice activists excited, but to me what’s more interesting is the open discussion about which factors to take into account in allocating profits, and which to disregard. That’s not just about refinements to the technical implementation of the arm’s length principle: it’s politics.

What did we learn from yesterday’s BEPS ‘tweetchat’?

140 character exchanges are unlikely to be the most revealing and nuanced on a topic as complex as the OECD’s ‘base erosion and profit shifting‘ project. But it was interesting that the OECD decided to reach out in this way, and also that questions did not just come from the usual suspects – the Institute of Chartered Accountants in England and Wales, for example.

Here’s a storify of the questions and answers. While, as I say, you can’t read too much into a single tweet, a couple were worth mentioning.

Continue reading

On the BRICS’ choice of diplomatic language at the OECD

I’ve just been reading through the 2010 update to the OECD model tax treaty [pdf] and in particular the different positions set out by non-OECD countries. This is where countries can put on record their objections to the model treaty, so that potential treaty partners know what they’re getting themselves into. I think the difference of language used by Brazil, China and India is quite illustrative of their approach to diplomacy in international tax.

Continue reading

Five people who were missing from the Global Tax 50

We’re invited to comment on International Tax Review’s Global Tax 50. It says a lot about these annual trade press lists that four years before I appeared in the International Tax Review list I had featured in its equivalent in Drapers, which covers the fashion industry! Still, it’s a fun game, so let’s play it. Here’s who I think they missed out.

Continue reading

More on Starbucks tax affairs in the Financial Times

A great letter in this morning’s FT from a tax lawyer demonstrates not one but two points about the transfer pricing of royalty fees. The first is the way in which companies game the system:

The role of the tax expert was to identify the highest level of royalty that could be defended in attritional correspondence with the Inland Revenue. The lawyer’s role was to reach for the intellectual property precedents and draft licensing agreements, which bore the imprimatur of arms-length contracts although they were in reality no such thing.

Sometimes discussion of transfer pricing by campaigners obscures the difference between the blatent use of prices that are much lower or higher than should be permitted (this is true ‘transfer mispricing’, which is more akin to fraud) and the more common manipulation of prices to get the best tax position possible within the ‘arm’s length’ range.

Continue reading