I see Ben Saunders has taken up the ABF story with two really interesting posts, the first relating it to Bill Dodwell’s call for a campaigners’ code of conduct, and second to what on earth Zambia was doing signing its bonkers treaty with Ireland. As I’m writing this in a break from struggling to put together a PhD chapter on developing countries’ tax treaty policy, let’s have a look at the code of conduct point.
Ben is right that some campaigners have not exactly responded to this suggestion in a very mature way (although with the recent ‘tax prat’ incident I don’t think either side is exactly upholding the standards of decent debate). But I also agree with Chris Jordan (in the comments here) that the way companies respond to allegations is rarely whiter than white as well: this was the gist of my post last week, in which I felt it unfair that ABF’s misrepresentations of the ActionAid allegations and – in some cases – of its own position had been, for the media, the last word on the story.
I remembered that I wrote a post a few months ago that made a couple of suggestions for where campaigners could behave a bit differently. I thought it would be interesting to reflect on ActionAid’s ABF report in the light of this.
Here’s what I wrote at the time:
There are two things that I think campaigners, and occasionally the media, tend to do which perhaps undermine their case a little.
The first is to jump on every case that suits their argument – or in some cases, that is consistent with how they know companies tend to behave – without checking the facts….it’s important to be clear about the limits of the evidence in each case.
The second mistake that I think campaigners make is not to clarify terms well enough. In particular, ‘tax avoidance’ has a very specific meaning to tax professionals, which is different to how it’s understood by the lay person. For most people, the important distinction is between illegal tax evasion and legal tax avoidance; to them, any tax practice that ‘smells bad’ fits into one of those two categories. But technically, to qualify as ‘tax avoidance’ a practice generally has to fall into a space bounded on one side by the letter of the law, and on the other side by the intention of parliament.
Limits of evidence
The ActionAid report itself is very much “this is what we know, this is what the company says, this is the conclusion we’ve reached,” giving the reader the space to question that conclusion (as some have done). I’m convinced of their case, and I am not always. What readers may not realise is that there are many putative NGO reports like the ABF one that haven’t made it to print, because the evidence isn’t up to snuff. I can think of one where I was sure there was a case, but we couldn’t prove it, one where somebody got the wrong end of the stick and had to be persuaded not to publish, and one that was published despite analytical flaws.
Now, I think Ben is right that one of the big questions hanging over the report is the company’s assertion that the Irish and Mauritian companies pay tax at 28% because of South Africa’s controlled foreign companies regime. This affects allegations of profit shifting, but not of withholding tax avoidance. What to do when a company throws this at you? It will have been difficult for ActionAid, because from the moment they got in touch with ABF, everything the company said in correspondence will have been written with the sole objective of killing or undermining the story, not of ensuring accuracy. I spotted a few sentences in the letters from ABF [pdf] that I could see were clearly written to mislead, just like the way the “we’ve been wronged” quote in the press about the Dutch part of the story only works if you haven’t read ActionAid’s allegations and so can’t spot that it’s based on a blatent misrepresentation of the allegations. So I don’t have much trust in anything ABF says. And the problem is that after a successful bit of presswork it can put the burden of proof onto ActionAid and then go back into its bunker, where it has all the information, and leave us all to speculate. I would have done what ActionAid did in its report, which is to show the evidence it has, and leave it to the reader. And focus on the cost to Zambia, which holds regardless of any saving or not to the company.
Avoidance or not?
There is, I think, a major point and a minor point here. The major point is that ActionAid shouldn’t have gone out with zero tax paid” as its topline. It didn’t need to. It could have led on the $27m figure instead. There’s a particular problem with this ActionAid webpage that says:
Associated British Foods, owners of many household brands like Silver Spoon and Kingsmill has been paying virtually no corporation tax in Zambia – one of the poorest countries in the world. This has cost Zambia an estimated US$27 million since 2007.
The reason ABF paid zero taxes from 2007-12 was mostly, if not entirely, because of capital allowances, not the practices criticised by ActionAid. This problem is about the topline messaging, not the validity of the analysis underlying the report, which does acknowledge the capital allowances upfront. But as Ben points out, the ethical reasons to be accurate may be less significant here than the tactical ones: the top line in all the press work allowed ABF to come out on the front foot with a denial that knocks down the topline message but is irrelevant to the substantive allegations. If I were at ActionAid I’d be kicking myself over this.
(The one way ActionAid can defend this decision is that at there will come – may already have come – a point at which the capital allowances are used up and the company starts paying taxes again, and that point will come later because of the “profit shifting” it alleges).
The other point is about which of the alleged practices constitute avoidance – and indeed what exactly is being avoided. Here I think ActionAid is on a surer footing.
First, there’s $9.3m from the tax incentive. That’s not tax avoidance, but nor does ActionAid say it is. Even the much-criticised infographic is quite clear about that. Is it legitimate to criticise a company for using the courts to get a tax incentive that it appears the government did not intend it to be eligible for? Absolutely. Certainly not ‘unscandalous’ as Maya calls it. ABF argues in the correspondence that the refund it got for the retrospective application of the tax incentive shouldn’t be counted as a cost to the government over 2007-12. I could be persuaded either way on this one, but it’s a judgement call, not a matter of fact.
Next, $10.4m in withholding tax avoidance. I think it needs emphasising more that not only is this treaty shopping, but it’s also undisputed by everyone: ABF only manages to deny it in the Telegraph by pretending ActionAid is alleging the avoidance of Zambian corporation tax instead. Ben wonders whether this is ABF’s fault or the Zambian government’s, but I think treaty shopping is quite clear cut letter and not intention of the law, isn’t it?
Finally the $7.4m through the management fee payments. Now I think the report is deliberately ambiguous on this: approximately $7.4m of Zambian tax is avoided. If the services are really provided ‘at cost’, that’s withholding tax; if they’re not provided at all, it’s corporation tax (the saving just happens to be the same either way); most likely, since there’s a 26% profit margin in Ireland, the truth is somewhere inbetween, but the report is upfront about not knowing which it is. This is the bee in Maya’s bonnet, but I just don’t think it’s that big a deal that ActionAid decided to simplify matters by saying it was corporation tax. I might have taken 26% as corporation tax and 74% as withholding tax, but the cost to Zambia would be the same. The only differences it makes are:
- to the “ABF paid no corporation tax in Zambia because of profit-shifting” topline, which shouldn’t be there anyway; and
- to the strength of ABF’s controlled foreign company rules argument, which only works on the corporation tax side.
The reason I’m not so bothered about whether this would be some kind of fraud is my experience from research in Ghana. How this stuff seemed to work there was as follows. The investment promotion authority authorised a contract for, say, management fees up to a certain percentage of turnover. The tax authority didn’t ask whether there was any substance underlying the payments, it just checked that the fees were leaving Ghana as per the contract. It was just understood that you could claim up to a certain amount without having to justify it. I don’t know if that is what is happening in this case.
What about this code of conduct?
In a week when the OECD published a report that criticised NGOs for blaming all tax avoidance on transfer pricing, ActionAid published a report in which most – possibly all – of the allegations are about treaty shopping and tax incentive, not “profit shifting”. That’s a welcome departure. But then they shot themselves in the foot by leading with a topline that wasn’t really correct. ABF came out with a very well put-together denial that called ActionAid out on this one point but mostly downright misleading. Journalists printed that unquestioningly.
If I were drafting a ‘code of conduct’, I might include:
- being clear about the limits of your evidence;
- ensuring that your campaign communications and media releases, not just your report, reflect this;
- being clear when your allegation is of tax avoidance and when it’s of something else, such as pursuing a tax incentive in this case;
- giving the company a right of reply, and publishing the correspondence;
- have an independent tax professional look over the report and the company’s response, and comment on it before publication;
- publish that independent commentary, including the commentary on the company response.
In this case, ActionAid did 1,3,4 and 5 (yes, they had a former tax inspector look over the report). Their error was in not getting 2 right.
As for 6, I think the problem is that the debate has become so polarised, finding tax professionals who commanded the confidence of both sides would not be easy. As I have written more than once, while mistakes do happen on the campaigning side, I find that the tax profession often have a tendency to confound differences of opinion about how the world should be with factual errors about how it is.
So maybe now I am suggesting that a code of conduct for NGOs would need to be supplemented by a) a code of conduct for companies responding to avoidance allegations (why not have these independently verified too?), and b) a code of conduct for tax professionals assisting NGOs and companies as part of their code of conduct. Or would that be too meta?