Corporate Tax and Development: opening Pandora’s Box


Reblogging my post from the International Centre for Tax and Development:

Taxation is important for development, not only because taxes provide the revenues to fund public services and infrastructure, but because they are a critical accountability link between governments and citizens.

Regular headlines and report findings tell us that a major problem keeping poor countries poor is that large corporations use clever techniques to avoid paying their fair share of tax:


  • “The money that developing countries lose each year because of the tax antics of big business is very nearly one-and-a-half times what they receive in terms of aid. So, on the one hand, you’ve got relatively rich governments handing out aid, but at the same time you’ve got the multinationals busy taking out as much as they can from those countries.” (Christian Aid)
  •  “If multinational companies were taxed fairly, developing countries could raise an extra $242 billion to tackle inequality.”(Oxfam)
  •  “Developing countries lost US$5.86 trillion to illicit financial flows from 2001 to 2010 and […] corporate tax abuses such as transfer mispricing accounted for 80 per cent of those outflows.” (International Bar Association)
  • “[Developing countries] lose between €660 and €870 billion each year through illicit financial flows, mainly in the form of tax evasion by multinational corporations”(Eurodad)
  • Tax dodging corporations are depriving the world’s poorest countries of billions of dollars they could use to feed their people” (IF Campaign)

It’s a shocking message, but one that is also appealing in that it identifies familiar, satisfying and conveniently accessible culprits for underdevelopment.

It is an argument with undeniable moral power, but most compellingly of all it offers the promise that a few relatively simple international interventions could unlock massive new flows of revenue to governments in the poorest countries. Respected development organisations tell us that stopping this international tax avoidance could release resources on a scale that would solve the most urgent challenges for development:


Who could argue with those numbers?

As Christian Aid point out if you do, you risk looking like you want to detract attention from the consensus on tax dodging and developing countries. And people will probably assume that you are an apologist on the payroll of Big Tax Avoidance.

But still, not looking too closely at the evidence puts those who are concerned with tax and development at risk of conveying misunderstandings, distorting our own priorities, missing chances for constructive engagement, and ultimately risking advocating for poor policy. So I think it is worth taking the data seriously and not just using it for decoration.

So what do we find, if we trace back from the headlines into the footnotes and the data tables of the big ‘tax dodging and development’ reports?

Simple mistakes

austin powers

Its hard to comprehend big numbers

Firstly, a disturbing number of reports that governments of poor countries could benefit from a huge ‘pot of development gold’ from curbing tax avoidance are based on basic mistakes or misunderstandings. One of the most common is where estimates of trillions of dollars of gross illicit flows are misinterpreted as representing a tax revenue loss.

The estimates of illicit flows come from the organisation Global Financial Integrity (GFI), which explains that illicit flows are ‘illegal movements of money or capital from one country to another’. Some of this is motivated by tax avoidance or evasion, but the potential public revenue losses are an order of magnitude less than the overall sums involved.

A large part of GFI’s estimates relate to trade misinvoicing, which they explain is not the same as abusive transfer pricing (a key focus of current tax debates including the OECD BEPs process). However many organisations get confused.

Eurodad (a network of 47 NGOs) for example say ‘In Europe, the loss of income caused by tax evasion and avoidance is estimated to be around €1 trillion per year. When it comes to the world’s developing countries, conservative estimates report that these countries lose between €660 and €870 billion each through illicit financial flows, mainly in the form of tax evasion by multinational corporations’. The €660 and €870 figure here comes from GFI and is a misinterpretation.


This number has been widely misinterpreted

Writing about transfer pricing, Kevin Watkins, Executive Director of the Overseas Development Institute says ‘On one estimate – and probably a conservative one – sub-Saharan Africa was losing US$34bn annually between 2008 and 2010. That’s around 3% of GDP, or almost one fifth of government revenues. … Cutting these losses could fund the investments needed to put millions of children in school, expand opportunities for health care and deliver clean water, it could also reduce dependence on aid.’ This again is a misunderstanding of the GFI figures.

Watkins was the lead author on The Africa Progress Panel’s 2013 report which says that ‘Just one single tax evasion technique – trade mispricing – costs Africa an estimated US$38.4 billion per year, more than the continent receives in either international aid or foreign direct investment.’ This figure, repeated by Kofi Annan has gone on to be widely quoted (for example in the media and by NGOs including Oxfam and Action Aid.[1]

Most recently an eminent group of academics called for the UN to ‘stop the loss of trillions of dollars to tax abuse’, drawing on the GFI gross illicit flow figures.

In other cases mistakes are more complex, such as when Christian Aid extrapolated $1billion of lost GDP to Zambia from the declared price on a single shoe box full of copper ore exported from Switzerland. These mistakes are rarely corrected and live on as ‘zombie facts’.

Corporate doesn’t mean what you think it means

Just as celebrities visiting refugee camps attract media attention in a way that domestic humanitarian workers do not, allegations of tax dodging by major global corporations make more headlines than efforts to improve routine local tax administration.

Therefore a common and tempting mistake is to attribute any broad estimate related to ‘corporate tax’ (i.e. taxes due from incorporated business of all sizes) to multinational corporations. So for example the estimated tax losses associated with the GFI illicit flow figures, even when they are correctly quoted, are often incorrectly attributed to multinational corporations (and transfer pricing).

Another figure to which the ‘by multinationals’ label has been attached is ActionAid’s recent estimate that developing countries could raise $104 billion by closing the tax gap on evasion and avoidance and collecting 20% more corporate tax than they do.  This general figure has been interpreted by Oxfam to be ‘the amount of unpaid tax due by multinationals’. Christian Aid’s original $160 billion is also a broad tax dodging estimate  (it is based on a simple percentage applied to all trade) to which the label of ‘by multinational corporations’ is almost reflexively added.

The distinction matters because different policies are needed to address a tax gap relating to a few major multinationals or many smaller businesses, and to transfer pricing or trade misinvoicing.



In some cases information which could help us to understand the relative significance of domestic versus multinational business tax revenues is downplayed. Christian Aid’s 2013 report ‘Multinational Corporations and the Profit Shifting Lure of Tax Havens’ perhaps surprising finds that out of the 9,000 companies operating in India whose published financial data they reviewed, the lowest average tax rates were paid by domestic companies (you have to read the report very carefully to work that out).

Killer facts?

Even more concerning than simple mistakes, and putting the spotlight on the most newsworthy aspects, is the consistent and systematic over-statement of the potential development impacts associated with policies to address tax avoidance.

The various methodologies for estimating corporate tax gaps in developing countries apply a formula to an existing financial flow (GDP, trade, tax revenues) to generate an indicator of the scale of potential gains from changes in tax policy. Some use a basic rule of thumb (governments could raise tax revenues by 20%, governments give away 0.6% of GDP as tax breaks, 7% of trade is mispriced…), others use more complex calculation. But in each case potential new tax revenues are assessed through consideration of currently measured economic activity of one kind or another.


Christian Aid’s Death and Taxes report founded the genre

This means that the big numbers are generated by the big economies –countries such as China, Mexico, Russia, India, Malaysia, South Africa, Indonesia, Poland and Saudi Arabia. However the impression is often given (for example through photographs and case studies) that this money largely relates to the poorest countries.

While many middle-income countries have significant poverty, and every country needs sustainable domestic public revenues, it would be extremely unlikely that large sums of tax revenue raised in emerging economies in Asia, Europe and the Middle East would be spent through the public budgets of least developed countries in Africa and elsewhere.

Yet this is what is exactly what is often assumed in the calculations used to illustrate the potential development impact of tax transparency measures. Tax gap estimates are commonly converted into development impact headlines such as the number of African midwives salaries that could be paid or ‘lives saved’ in low-income countries. In this way numbers of dead children are used as a mathematical yardstick for large financial sums in a similar way that ‘an area the size of Wales’ is used to talk about rates deforestation

For example, a recent report from ONE aimed at this year’s G8 [correction: G20] states that addressing trade mispricing would easily generate the $28 billion needed to prevent 3.6 million child and infant deaths each year in low-income countries. However, their figures (which they kindly sent me) show that only $1.3 billion of the estimated extra tax revenue would be generated in low-income countries. Therefore 95% of the potential ‘lives saved’ would come about only in the unlikely event that upper middle and high-income developing countries transferred their tax gains to poorer ones (this step is not mentioned in the report).

I think this is shocking. When respected development organisations say that a policy they are advocating has the potential to save lives in the poorest countries, we assume that they mean it could save real lives, not hypothetical lives in a world where Russian tax receipts and Malawian health budgets are perfectly fungible.

It is not just ONE that has used this approach – other ‘killer facts’ on the development impact of tax avoidance also use the same unstated assumption that dollars raised in taxes in higher income developing countries can be spent in poorer countries (some reports such as Christian Aid’s ‘Death and Taxes’ use more complex regression analysis, but they don’t control for obvious confounding factors like GDP).

Losing more than they gain in aid?

The most well-known and well-used killer fact is that developing countries lose more from tax avoidance than they gain in aid (three times more is a common number – it comes from the OECD, but no one knows how they calculated it), other figures are one or two times more.

However these aggregate figures rely on the same assumptions of fungibility as the examples above. What the estimates actually seem to indicate is that in general large, rapidly growing economies are losing more to international tax avoidance than smaller poor countries gain in aid. But this tells us very little about the significance of these figures to either group of countries.

Gross illicit flows are also often compared with aid (overall they are estimated as ten times more than aid).  While curbing illicit flows may have other development benefits, these sums should not be thought of as equivalent to aid or to tax revenues in terms of paying for public services and infrastructure.

So is there really a pot of development gold?

Experts agree that there is significant potential for developing countries to raise their tax receipts, and that some of this relates to multinational companies and international rules and some to domestic revenue collection. However there is little confidence in the numbers involved [ii].


Its an unfathomably large amount, got it?

Nevertheless if current estimates are best we have to go on, they should at least be communicated clearly. One thing that becomes clear once you take away all the showmanship of the killer facts is that the estimates commonly used are simply not that much money. Global numbers in billions are hard to comprehend, but we can make honest and clear efforts to make sense of them on a country-by-country basis. According to the data that ONE sent me (which uses PWC data on national tax rates to estimate the tax revenue losses associated with GFI illicit flows estimates) it looks like most countries where aid contributes a significant proportion of government budgets have estimated trade related tax losses in the region of 15% or less of aid receipts. Not nothing, but not the grand problem-solving amounts we are led to believe.

If you look at what this amounts to on a per capita basis (based on the ONE data and my calculations), Bangladesh could raise $2.77 extra tax for each of its citizens, Ethiopia $6.81, India $9.31and China $4.14. That is dollars; single dollars. Per person. Per year.

Truth and reconciliation

None of this is to say that multinational corporations are off the hook, or that taxation doesn’t matter for development. It matters, particularly as countries grow economically.

There are big point sources of rents – most notably mining, oil and gas concessions – and there is urgent work to be done on revenue transparency and how to negotiate public-private deals which enable sustainable development. There is a strong case for governments to strengthen their own tax systems, and to be more transparent and about revenues, budgets and expenditure. There are clear issues on updating international tax rules for the 21st century.  And effective, workable systems are needed at the international level to prevent tax evasion, money laundering and hiding of the proceeds of corruption.  Most of all there is a need for sustained and sustainable economic growth in developing and emerging economies. Many of the organisations responsible for these strange misunderstood numbers are also doing valuable work in these areas.

But the devil is in the detail. The public and civil society should participate in ongoing developments on tax reform and economic policy: it shouldn’t just be left to ‘the experts’ such as the OECD and IMF (although their attention to tax and development issues is welcome), or to corporate lobbies. But we can’t engage meaningfully with complex issues, or scrutinise the powerful, if the organisations that help us to do this on allow inaccuracies and overblown claims to undermine understanding, whether internationally, at a national level, or indeed within their own organisations.

Communicating ambiguously, defending the continued use of out-of-date estimates, allowing them to be misunderstood and dressing them up with development impact calculations which go beyond the heroic to the implausible, all reinforce the impression that there is a huge pot of development gold associated with tax avoidance. But it does nothing to help understand what the real potential gains of policy changes could be.

Furthermore, the overall addiction to these numbers damages organisations as it undermines their own internal integrity and understanding of issues, and ultimately our trust in them.

There are positive signs that a next generation of analysis and informed public debate could emerge. The Financial Transparency Coalition brings together many of the civil society organizations internationally working on issues of tax avoidance and illicit financial flows and offers a forum for building and sharing knowledge and potentially for holding each other accountable and preventing misunderstandings being spread. The set of NGOs (and donors) working on tax and development is growing to include those focused on transparency and accountability, who are committed to the ‘eat your your own dogfood’ principle on openness. Organisations with serious research chops such as the Center for Global Development and the CHR Michelsen Institute and university based institutions like the International Centre for Tax and Development and the Oxford University Centre for Business Taxation bring academic discipline. Tax professionals and their organisations are increasingly willing to contribute to constructive debates and technical collaboration.

My own experience is that while historically any critique of tax dodging figures has been seen as something to be rebutted [* NB: scroll down to comments] and rejected, organisations such as Oxfam and Action Aid are becoming more willing to revisit claims which are clearly mistakes.  And recently when I brought it to the attention of Alex Cobham and Owen Barder at the Center for Global Development that they were making some really weird claims about trade mispricing of copper by Swiss traders they took my concerns seriously and are now working on revising the report.

The mistakes and misinterpretations I have outlined here are not simply typos, but seem to indicate that something has gone wrong in organisational learning and accountability.  Something as simple as a multi-sector group to offer peer review could raise the level of understanding and debate considerably.

Some argue that without the attention gained through the overblown promise of trillions of dollars, and millions of lives saved, none of this positive attention would have come about. That may be true. But I believe misleading and misunderstood figures have now become a barrier to constructive engagement between NGOs, academia, business and the tax profession, and to developing more robust analysis and ways forward.

We have to get beyond pretending there is a huge pot of development gold for the poorest countries hidden in the Pandora’s box of tax data. There is no excuse for continuing to sensationalise figures, or to allow errors to go uncorrected, and every reason to try to develop common language and understanding of data to enable constructive engagement.

According to legend after Pandora opened the box and all the bad stuff flew out what was left inside was hope. The hope of democracy is that governments, businesses, citizens and their organisations can negotiate to develop institutions and systems to enable healthier and more prosperous societies. I think that they are served better if they have the best information to guide them.

[*] Action Aid corrected this report

[ii] Experts as diverse as The Oxford University Centre for Business Taxation  (for DFID) and James Henry (for The Tax Justice Network) have expressed concern that the current estimates are not robust.

[* Updated: October 20 2014]

7 Responses to “Corporate Tax and Development: opening Pandora’s Box”

  1. 1 iain

    An excellent piece of work that starts to to delve into the depths of the issues and uncovers some very worrying and apparently fundamental mistakes.
    Have the various authors of the Reports scrutinised been able to provide any responses, in the way you mention Alex Cobham and Owen Barder? I wonder if there has been some sort of domino or round robin effect, whereby one (mistaken) piece of analysis has been accepted and then used as the basis for the next Report, and so on.
    If the true issue is the extent of domestic tax loss (which is a far wider issue that taxes lost via international businesses) then maybe the best approach is to try to do what some other countries have done and analyse (as best they can) their Tax Gap. Both in terms of an operational gap (ie how effective is the country at collecting what it is meant to under the law) but also the policy gap (ie what parts of the potential tax base are not touched by collection).

    • 2 Maya

      Hi Iain

      I’ve been in touch with most of the authors of the reports mentioned and many of them commented on drafts of the article. I will leave it to them to respond!

      You are right some of these are “round robin” mistakes – and organisations sometimes correct them if you point them out – but they tend to be treated as typos or sub-editing mistakes, rather than recognised as signs of misunderstanding or dysfunction in the debate.

      Others are more worrying contortions of the available evidence.

      I think your idea of a common approach to defining, estimating and communicating operational and policy tax gaps sounds like a good one.

  2. I want to thank you for taking the time to work through all of the headlines and to ask the basic question, “What was your method for arriving at this number?’.

    I find that most of the “discussion” around tax policy is too muddled with preconceived political positions, rather than starting from a commonly agreed nomenclature and given facts. In addition, the discussion also seems to ignore implementation practicality and unintended consequences when suggesting policy changes.

    There is certainly a need for a continuing logical and practical modification of domestic and international tax policy to deal with the ever-changing nature of economic growth and trade. However, with an environment fundamentally tainted by unsubstantiated and basely numbers (and vicious ad hominem attacks to those who challenge them), much of this discussion quickly degenerates into a debating mosh pit.

  3. 4 Mike

    Part of the problem, surely, is that some of the largest sums of foregone tax revenue** from multinational corporate tax behaviour in truly low-income countries are invisible in trade-based or profits-based estimates?

    One example is the avoidance of tax on capital gains from sales of subsidiary companies or their assets (mines, factories etc.), by selling them indirectly from one offshore holding company to another. Single transactions of this kind, if taxed, could in some cases generate sums of revenue that are truly significant for small states and economies. The Hutchison Essar deal in India is the axiomatic case, but there are examples from much smaller LICs and LMICs (Uganda, Mauritania). The recent IMF report on corporate tax spillovers cites a 2010 case from Mauritania where, the IMF claim, a $4bn gain from the sale of a Mauritanian gold mine from one Canadian company to another via the Bahamas went untaxed in Mauritania (Mauritania generally levies a 25% tax on corporate capital gains, so the likely foregone tax revenue is significant by any country’s standards).

    This is not to say that such revenue is easily gathered: many LICs and LMICs don’t have capital gains tax; and even where they do it’s often domestic tax laws without adequate coverage of indirect transfers of interest, or residence-heavy capital gains articles in tax treaties, that prevent such indirect transfers being taxed in the country where the asset is located. But, without making any global claims, such examples do show that there are some significant ‘pots’ of potential multinational-corporate tax revenue even in ‘genuinely’ poor countries, which simply may not show up in the kinds of estimates you’re citing above.

    **(I’ll leave it up to you whether you want to call it tax avoidance, or something else!)

    • 5 Maya

      Thanks for the thoughtful comment Mike – I am sorry I missed it and left it stuck in moderation for so long!

  4. 6 Maya

    ONE have responded to my concerns about the calculation of the ‘some 3.6 million deaths could be prevented each year in the world’s poorest countries statistic here:

    They say “..we recognize, as we’ve heard in comments, that the hypothesis that such deaths would in practice be prevented turns on some large assumptions. For while the largest illicit financial flows (and hence potential recovered revenues) are to be found in the richer developing countries, the greatest need for investment in health systems are in the very poorest.

    There are a lot of moving parts to complex problems such as corruption. We can’t claim the same level of certainty as a Randomised Control Trial for a medical intervention, or even a modelling approach where the data is stronger and more consistent. But in pushing for common sense solutions which have the support of experts, governments and multilateral institutions we have to address the question of what kind of assumptions are reasonable to demonstrate why it is important to address the problem.”

  1. 1 Constructive engagement on the ‘tax big numbers’ | Hiya Maya

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: