Taxation logo taxation mission text

Since 1927 the leading authority on tax law, practice and administration

Lack of evidence

16 February 2010 / Mike Truman
Issue: 4243 / Categories: Comment & Analysis , Companies
The claim that poor countries lose $160 billion in tax from ‘transfer mispricing’ has been repeated endlessly. MIKE TRUMAN finds it hard to justify


  • Two Christian Aid reports claim $160 billion tax lost.
  • Raymond Baker’s 7% claim does not relate to TNCs.
  • Problems of methodology in Simon Pak’s study.
  • Real shortfall is homegrown tax evasion.

Nine months ago, outside the Park Lane Hilton on the night of the Taxation Awards, there was a small tongue-in-cheek ‘Alternative Tax Awards’ run by Christian Aid.

They were publicising their ‘Big Tax Return’ campaign, and inviting people to petition the major accountancy firms to support country-by-country reporting, so that multinational firms would have to show publicly exactly how much profit they made in each country.

This would prevent, Christian Aid said, the mispricing of goods and services within multinationals that led to ‘at least $160 billion’ of tax being lost by developing countries.

I subsequently took up an opportunity to debate the issue with John Christensen from Tax Justice Network, which was involved in the research behind it. I expected to find a lot that I agreed with.

In fact, I ended up thinking that some very strident claims had been built on very little evidence. I’ve steered clear of covering it in Taxation so far because multinational taxation is not normally an area we cover .

But the claim has been widely repeated as fact, and the issue has been in the pages of Tax Adviser recently, so I couldn’t resist having my own say.

There were two main reports in which Christian Aid laid out their case. The first was Death and Taxes published in March 2008, the second was False Profits published a year later in March 2009. Both reports are available on the Christian Aid website as pdf files.

Mister 7%

Death and Taxes calculates the tax lost to ‘false invoicing and abusive transfer pricing’, and then calculates how many deaths could have been prevented if developing countries had been able to collect and spend those taxes.

It sets out its methodology for calculating the lives that could be saved in a closely worded appendix, using what appear to be sensible statistical methods.

Christian Aid also explain what they mean by ‘false invoicing’ and ‘abusive transfer pricing’, which they more often refer to as ‘transfer mispricing’.

False invoicing is where a business either inflates the price invoiced on the goods they import to a developing country, or reduces the price invoiced on exports, in both cases for transactions with third parties.

Transfer mispricing is where two related affiliates of the same trans-national company (TNC) similarly agree to overprice imports or underprice exports. In both false invoicing and transfer mispricing, the result is a reduced profit being declared in the developing country, and therefore a reduced tax take.

The problem is in the figures that they use. To get to the figure of tax lost, they need to know how much of the volume of trade is illicit capital movement caused by false invoicing and transfer mispricing. This is stated as 7% of trade volumes. In other words, the true amount earned by the country in such trade is understated by 7%.

So where does this figure come from? It comes from some work by Raymond Baker, most recently set out in his book Capitalism’s Achilles Heel. He conducted 550 interviews in the early 1990s with officials from trading companies in eleven different countries, on conditions of anonymity.

He said, ‘mispricing in order to generate kickbacks into foreign bank accounts was treated as a well-understood and normal part of transactions’, and drew his 7% estimate from this evidence.

Now, there are several criticisms that can be made of this with regard to false invoicing (transactions between third parties). The interviews are nearly 20 years old, they are necessarily secret so their suitability as a sample cannot be verified, and so on.

But let’s accept for the sake of argument that this is a valid estimate of false invoicing. What does this say about the main target of Christian Aid’s campaign, transfer mispricing within TNCs?

In my view, it says absolutely nothing about it. The whole point of the false invoice is to divert some of the value of the transaction into a bribe.

What TNC is going to sit back and accept that two of its subsidiaries should artificially alter prices so that they can generate a kickback for their employees?

That’s not to say it won’t happen, but the TNC can be expected to have procedures in place to try and stop it.

The alleged motive of the TNC is entirely different: tax evasion. Again, while that almost certainly happens, the figure for false invoicing will not tell you anything about its incidence.

Baker himself says that he has done no formal investigation in this area, he simply cites his own experience of seeing exaggerated intra-company pricing, which he says exceeds that in unconnected entities. It is, in other words, a guess, informed by personal experience.

Mister 25%

False Profits seems initially to back up the Raymond Baker figures. Although only covering trade between the EU/US and the rest of the world, the research looks at around 10 million records of trade data between 2005 and 2007.

It categorises them by common trade classifications, and then estimates which are overpriced imports to developing countries and which are underpriced exports from developing countries.

Again, this is taken as showing the amount of profit illicitly moved from those countries, from which lost tax can be calculated.

However, my eye was caught by a particular country’s figures. Part of the total that is being interpreted by Christian Aid as an illicit capital outflown was about 600,000 euros said to have been illicitly moved out of Andorra, presumably to avoid tax.

Except that, of course, Andorra doesn’t have any direct taxes. Why would anyone want to reduce their Andorran profits, particularly for tax reasons?

The problem, once again, is the methodology. What this study does is to put all the transactions for a particular classification, as recorded by the US and the EU trade data, in a line from the lowest to the highest.

It then says that anything in the bottom 25% is underpriced and anything in the top 25% is overpriced. It therefore follows that if there is any variation in price at all, this method will result in a quarter of the transactions being treated as priced too low and a quarter as priced too high.

I put these points by email to Professor Simon Pak, who carried out the research. On Andorra, he pointed out that there were other reasons why transactions might be mispriced apart from tax.

This is true, although the main one I know of creates an artificially high price for Andorran tobacco exports (yes, Andorrans farm tobacco in their mountain hide-out. The reasons are complicated and probably best not enquired into too deeply).

However, Christian Aid’s use of the research is all about tax paid by multinationals.

On the use of an inter-quartile range, Professor Pak pointed out that this was the measure used by the US tax authorities for transfer pricing. It is, but only once a truly comparable set of figures has been obtained. Even though these are the most detailed data available, they are still comparatively broad trade classifications.

They also cover the whole of a year. As the report itself says, right at the end, if prices are volatile this can legitimately cause wide variations from the interquartile limits.

It cites crude oil as an example, where the spot price varied between $36 and $145 a barrel in 2008. And yet far more prominently, in the main body of the report, Christian Aid says:

‘In Nigeria, £501m was lost from its burgeoning mineral fuel and oil industry [in 2007] … What made that figure all the more surprising was that the previous year, only £145m was lost in that trade category.’

In fact, a rough piece of spreadsheet work that I did using some spot prices for Brent Crude shows that you would indeed expect the 2007 figure to be between three and four times higher than 2006 simply because more volatile prices create the impression of much greater underpricing.

The debate so far

In the space available I can only comment briefly on other contributions to the debate, but I would urge you to go and read them.

Bill Dodwell, in November’s Tax Adviser made many of the points that I have above, but also pointed out that False Profits calculated that the US and Japan were the two biggest ‘losers’ from transfer pricing, which as he says is simply not plausible, given their aggressive attitude towards it.

Dr David McNair, from Christian Aid, responded to Bill’s article in the January issue of Tax Adviser. He repeats the original assertions about Raymond Baker’s and Simon Pak’s work supporting each other, without addressing any of the problems above.

He does address another problem raised by Bill, that if you are going to look at the capital outflows from one end of the price distribution you also have to look at the capital inflows from the other.

Dr McNair says that netting the two off produces misleading results, because it makes Africa look like a net recipient of illicit inflows, which is ‘inconsistent with its continued dependence on bilateral aid’.

Indeed it is, but doesn’t that also suggest that the methodology itself is simply wrong if it produces such an absurd conclusion?

Dr McNair also says that there was considerable discussion at a recent World Bank conference on illicit flows about the $160 billion figure, but that no alternative was offered and no compelling challenge made to it. Whatever was said at the conference, there certainly have been alternative figures put forward.

A study by the Oxford School of Business Studies in June 2009, carried out by Clemens Fuest and Nadine Riedel and available on the DFiD website, gives $160 billion as the highest estimate of the studies they considered, with $35 billion being the lowest (from an Oxfam study in 2000).

The case of Ghana

What the Oxford study suggests is that work should be done at a more detailed level, looking at the way tax avoidance and evasion is actually occurring in individual developing countries.

Interestingly, a 2009 report produced jointly by Christian Aid and the Tax Justice Network does precisely that for Ghana.

It is unable to produce any new data about tax evasion by TNCs, instead it reproduces the arguments from the previous two global reports and then applies those principles to Ghana.

However, what it does do is to produce some very revealing data about where the tax revenues actually come from in Ghana.

Individuals and companies make roughly equal contributions to the direct tax take in Ghana, each at just under 14% of total taxes collected. In developed countries, corporate tax generally raises much less than personal taxes. That suggests that it is personal tax which is undercollected, not corporate tax.

Within personal taxes, although the rate of collection by withholding taxes has increased significantly, the amount collected from the self-employed has not, and in 2007 was only 11% of total income taxation.

The shadow economy in Ghana is estimated, the report says, at 38% of total economic activity, and over 80% of employees are in the informal sector.

Presumably such businesses are not only avoiding direct tax on their own profits, they are also avoiding payroll withholding taxes and indirect taxes.

So who is actually paying tax in Ghana: their own and the taxes they have withheld?

In general, it is the large companies, particularly the TNCs. While the report raises a question mark about companies involved in the extractive industries, it acknowledges that ‘relative to some parts of the economy, formal sector business actually contributes a reasonably large share of national revenue’.


So is country-by-country reporting, campaigned for by Christian Aid and the Tax Justice Network, really the answer? I don’t see that it is.

I don’t necessarily oppose a limited version of it, perhaps concentrating on disclosing profits made in ‘secrecy’ jurisdictions, but I see no great benefit for Ghana in MegaCorp Inc having to itemise, allocate and audit its profits in a hundred different countries round the world.

Far more important, it seems to me, is the other part of the package now gaining traction with the Financial Secretary to the Treasury, Stephen Timms (though whether that will make any difference after 6 May is anyone’s guess).

That is technical help for developing countries to put in more effective tax systems, and the sharing of information obtained from tax information exchange agreements.

TNCs make a nice bogeyman to blame for low tax revenues in the developing world, but the evidence suggests that the real problem is locally grown.

Issue: 4243 / Categories: Comment & Analysis , Companies
back to top icon