ABF’s misleading statements

40 comments

1. I have been trying to make sense of the claim and counterclaim between Associated British Foods and ActionAid too.

   It looks to me like there have been some selective and misleading statements on both sides, which make it hard to work out what is pantomime, and what is really at issue.

   You mentioned in your other post that the research moved far from its initial assumptions. But as someone trying to make sense of the link between the report and the headlines, it feels like there is a “bait and switch” going on.

   The “killer fact” in the campaign communications is that ABF paid no corporate income tax in Zambia in recent years. This is stressed in the report (eg the graph on page 14) and in the tweets and headlines. E.g.: “$123mill made in profit since 2007 “Virtually no corporate #tax” paid in Zambia What’s ABF up to? Our video reveals: http://t.co/K1D5qVNM (https://twitter.com/ActionAid/status/300582532104220672) .

   You say that ABF’s explanation that this is due to capital allowances is completely irrelevant. I’m not so sure. It does directly answer the main charge levelled at ABF which is that they paid no tax on $123m of profit.

   I did enquire to Chris Jordan about this who said that the lost tax revenues that Action Aid is talking about are separate and additional to the a capital allowance. I don’t think this is at all clear in the report. You have to read it with a really fine toothcomb not to come away with the impression that it is seeking to explain why ABF paid no income tax.

   In your post you propose a link between this headline and the complex measures exposed by Action Aid. ” The capital allowance is a fixed sum, which the company will run down progressively as it makes profits. The more that profits in Zambia are reduced by tax haven payments, the longer it will last. ActionAid’s report is quite clear about this.”

   I don’t think the report is at all clear about that link – the capital allowance barely gets a mention. I’m also not sure it’s true – since the tax foregone is withholding tax, not corporate tax.

   I have tried to make sense of “The Bill” on page 33 and how it relates to the campaign “killer fact” and ABF’s explanation.

   First is the US$ 3m from the farming rate change from at 35% to 15% tax rate- I’m no accountant but wouldn’t this be cancelled out by the capital allowance, since that is written off against profits not tax liability? It is not an additional tax saving if they have no taxable profit in a year.

   Next is the tax refund of $6.3m for 2001-2005. There is an argument about the morality of ABF suing Zambia for this, but I think that ABF’s argument in their letter that the refund should be netted against tax paid in those years, not 2007-2012 is fair enough. It doesn’t really belong here.

   Next is the $7.4 m for withholding tax forgone , related to the Irish management company. This does look like an additional tax avoided by ABF in 2007-2012.

   ABF says “The payments are charged at cost, and there is no artificial reduction in profit in Zambia sugar as a result.” You say this is besides the point, as it ignores withholding tax.

   But your argument that the tax haven payments keep profits artificially low implies that this is exactly the accusation that is being made . The report says so on page 14 ” First, large payments are being made to sister companies in tax havens (Step 1), which can be deducted as expenses in Zambia, shrinking the company’s taxable profits in the country (against which capital allowances can then be deducted) while booking profits outside of Zambia. ”

   The 3m and 7.4m withholding tax foregone because of the dog leg interest and dividend via Ireland/Netherland also looks like tax avoided (but again not sure how it is relevant to the income tax headline?) Also if you take the ABF statement at face value, that borrowing directly in Zambia would have made their interest payments go up by the same amount as the withholding tax, then you shouldn’t you reduce the same amount from their profits – I.e. these two figures are double counting.

   So overall I think that a $ 14.8m estimate for foregone withholding tax in 2007-2012 looks supportable, but the other $12m looks shaky. I certainly don’t think these are the final ‘facts’, I could be wrong. But what I don’t see, having read the report and the correspondence carefully, is a clear line from the analysis of foregone witholding tax to the exposé headlines which focus on the zero income tax bill.

   1. Thankyou for this, Maya.

      You’re right about clarity re withholding taxes and corporation tax, I will tidy that up in the post. I’m not saying that there is no profit tax avoidance through the tax haven transactions (I contrast the 26% profit margin in Ireland with the ABF statement that payments are “at cost”). Rather, I’m saying that ABF doesn’t address the withholding taxes at all (in the correspondence yes, but in a very misleading way) which means that it does not seem to dispute a significant chunk of the alleged tax avoidance. Journalists could have pointed this out.

      I also agree with you that an important reason for the corporation tax payments being zero during 2007-12 is the capital allowances, and that ActionAid should have been clearer about this. I’m not saying that there’s a link between tax haven transactions and capital allowances, quite the opposite, I’m saying that ActionAid’s report is right to say that the capital allowances don’t affect its argument about the incentives and tax haven payments. But by bringing in the zero tax payments you’re correct that they conflate the two issues.

      However, In all the statements and correspondence, ABF just says that the allowances are ‘substantial’ without giving a figure. Are they so substantial that, paying tax at 35% rather than 15%, and without the tax haven transactions that ActionAid questions, they would still have wiped out the whole tax bill? I presume that ABF would have said so if this were the case. Even if so, the practices criticised by ActionAid mean that tax payments will be zero for longer than they otherwise would be (and that may include some of the years from 2007-12).

      The other detailed points you make are interesting and worth looking into, but probably not in the comments field as it will get too confusing!

2. Martin –

   On your last question – I thought capital allowances are where you write off the cost of developing an asset against taxable profits i.e. before the tax is calculated, not against your tax bill. Your don’t write it off faster if the tax rate is higher?? Maybe its different in Zambia. Otherwise I don’t understand how the 15% tax rate extends the tax free period from the capital allowance deduction.

   1. I see your point. I should have referred only to the tax haven payments, not to the tax incentive. There’s also a footnote to note 7 in the 2012 accounts (http://www.illovosugar.co.za/About_Us/grouplistedsubsidiaries/ZambiaSugarPlc.aspx) showing derecognised capital allowances in 2012, on the grounds that they’re unlikely to make enough profit to use them up before they expire after 5 years.

3. OK, got it, so the saving on the 35-15% tax cut was only in 2007, and is not on top of the capital allowance, and will kick in again when that ends. Still not sure about the back tax and the potential double counting between the dog-leg on interest payments and the foregone witholding tax on resulting profits?

   The main question though is still why front-end the report with headlines, graphs, comparisons etc.. that are all about the zero corporate income tax, when the back end analysis is mainly not about corporate income tax at all, but a different set of taxes forgone?

4. Hi Maya, it’s really great to have this level of real debate going on – thanks!

   This is one of the reasons we were keen to publish all of our correspondence with ABF, so that others can scrutinise the ‘working out’ on both sides. I’m not sure I totally agree with all of your analysis (I suppose I wouldn’t, since I’m ActionAid’s tax policy adviser!) and have tried to explain why below.

   CIT vs withholding taxes

   You are absolutely right that there are two separate things going on, according to our analysis:

   (i) some reduction of taxable profits through payments to insubstantial related party companies in tax havens, reducing CIT liabilities in Zambia (‘Step 1’)

   (ii) some avoidance of withholding taxes on interest payments and dividends (‘Step 1’, ‘Step 2’, ‘Step 3’)

   In other words, we are *not* simply talking here about transfer pricing abuse or ‘profit-shifting’, as ABF has alleged (one of the interests of this case study is that it shows how avoidance via treaty shopping and the use of conduit entities may be equally problematic, but has received much less media attention than TP).

   I hope our report has been clear about this distinction, stating (on the very first page of the report): “Some of these transactions reduce Zambia Sugar’s taxable profits, while the structure of others avoids the Zambian taxes ordinarily levied on such foreign payments themselves.”

   And again on page 2: “[we have] identified four strategies that have significantly reduced Zambia Sugar’s taxable profits to begin with, and that have avoided separate Zambian taxes on the company’s financing and dividends.”

   If it still isn’t clear enough that we’re talking about two separate phenomena, that’s our bad.

   But while all three ‘steps’ avoid withholding taxes by routing payments through jurisdictions with clement tax treaties, we *do* stick by the analysis that ‘step 1’ – the payment of management fees and export agency commission to Ireland and Mauritius – is reducing taxable profits in Zambia, and booking some profits in a lower-tax jurisdiction instead (at least in Ireland – we can’t see the Mauritian accounts). These are payments being made to companies which, to the best of our investigative ability, we established had no actual operations in Ireland or Mauritius. We’ve calculated the tax loss in this step based on CIT foregone. The connection between reducing taxable profits and capital allowances written down against those profits *is* therefore relevant here.

   We’ve also, I think, been conservative at this step, estimating the CIT loss in ‘Step 1’ based on the actual CIT rate applied to Zambia Sugar’s taxable profits (15% up to 2012), rather than the rate they would have suffered had the company not wrung an extra incentive out of the Zambian government in court (around 22-26% depending on their split of farming, sugar production and export income in any given year). So we *haven’t* costed in the compound effect of reducing taxable profits PLUS the special 15% tax rate, which would be greater still.

   And while Step 1 *is* calculated a CIT loss, by routing management fees (to related or unrelated parties) via Ireland the company *also* avoids the 15% withholding tax ordinarily levied on management fees. Were we to calculate the tax lost from ‘Step 1’ based on the WHT loss rather than CIT loss, the tax lost would actually be greater, since Zambia Sugar’s ‘farming’ CIT rate dropped to 10% in 2012, while WHT remains at 15% and is rising to 20% in 2013.

   Effect of capital allowances

   We hope we’ve been upfront about this, putting the capital allowances argument right at the top of the report (p.2) where we say: “generous capital allowances – the subject of current Zambian government scrutiny – may significantly reduce the company’s tax liability.”

   And again on p. 14: “This [writing down capital allowances] is undoubtedly a plausible explanation for some of the discrepancy between the usual Zambian rate of corporate income tax and the far lower tax rate applied in practice to the company’s (continually growing) pre-tax profits.”

   As Martin says, ActionAid is not arguing that capital allowances aren’t responsible for a significant reduction in the effective rate of CIT levied on Zambia Sugar’s pre-tax (accounting) profits. Rather, we’re saying that it doesn’t affect our analysis of the tax losses we calculate, because:

   (i) capital allowances are irrelevant to the WHT avoidance (‘Steps 1-3’), which is not dependent on profits;
   (ii) capital allowances also don’t reduce the tax advantage of reducing taxable profits in other ways (e.g. ‘Step 1’). If you’ve got $200 of taxable profit in Zambia taxed at 15%, and you reduce your taxable profits to $180 by e.g. paying management fees to an Irish company with no staff there, then you’ve reduced your Zambian tax liability from $30 to $27 (i.e. by $3). If you’ve also got $50 of capital allowances to write down against your taxable profits in a given year, then your CIT liability is $22.50 without your Irish payments, $19.50 with them. In each case your Irish payments have reduced your Zambian tax bill by $3 – no more and no less than if you weren’t writing down any capital allowances.

   Borrowing costs reduced by the ‘dog-leg’

   Unfortunately we’ve no evidence what the borrowing costs would have been without the WHT avoidance. All we can know is that Zambian WHT is being avoided, to the benefit of Zambia Sugar and the lending banks.

   N.B. Using conduit companies to cancel WHT is a standard avoidance technique known as ‘treaty shopping’ – the OECD produced a useful report on this in 2003.

   Effect of the farming rate change from at 35% to 15% tax rate

   You’re absolutely right that in the years when capital allowances are claimed against taxable profits, this will reduce or even cancel out the effect of the rate change (if your taxable profits after allowances is zero, then your CIT liability is zero whether it’s charged at 35% or 15%). And we don’t know how much capital allowances have been claimed in each year under scrutiny. So to avoid double-counting we therefore use only the reduction of tax liabilities from the rate change declared by Zambia Sugar itself in its accounts’ tax reconciliation note. These notes declare an effective rate reduction of 10.8% in 2007/8 ($1.8m), and 4.7% in 2011/12 ($1.2m) when the ‘farming’ rate dropped again from 15% to 10%. In the interim years – the likely years when capital allowances are being claimed – we’ve assumed no tax advantage from the rate change at all. As a result, $3m tax loss over 6 years is pretty modest: when capital allowances are not being claimed, even assuming that Zambia Sugar’s profits get no larger than at present, then the Zambian tax loss from this rate change will be some $3.6m every year.

   Tax refund for 2001/2 to 2004/5

   We’ve netted this refund against the company’s 2007-9 tax payments because this is how Zambia Sugar’s own ‘cash tax’ calculations present them in their own accounts. These are the years in which the refunds take place, so the refunds are set against tax payments in those years. ABF actually agreed with us that using the ‘cash tax’ measure is appropriate: they said “we agree that cash tax paid is a good, clear and understandable measure of one element of a company’s contribution to the local economy.” If there’s a problem with this approach, then it’s a problem with the accounting convention for cash tax, not our methodological invention.

   There’s also a more ‘real-world’ reason for counting this refund as a tax loss in 2007-12. What we are seeking to show in our report is the Zambian tax consequences of tax activities undertaken by Zambia Sugar since ABF took over its parent company in 2006. The refund is a consequence of a tax ruling won in 2007 (a ruling obtained by aggressive litigation which presented Africa’s largest industrial sugar factory as nothing but a farmer). It’s not tax avoidance of the financial engineering kind, and we say this clearly in our report. But it is the aggressive pursuit of a tax incentive in a desperately poor country. And it’s a real loss to the Zambian exchequer in those years – real money, concretely paid out by the Zambian revenue Authority in 2007/8 and 2008/9. I’m afraid I don’t agree, therefore, that this calculation is in some way ‘shaky’.

   ”0.5% on pre-tax profits”

   You’re absolutely right that the divergence of the effective tax rate on Zambia Sugar’s pre-tax profits must be in part a consequence of capital allowances. We say this explicitly in our report: ““This [writing down capital allowances] is undoubtedly a plausible explanation for some of the discrepancy between the usual Zambian rate of corporate income tax and the far lower tax rate applied in practice to the company’s (continually growing) pre-tax profits.” Any reduction of taxable profits themselves will impact the pre-tax profit figure itself, not the tax rate it suffers. And obviously the avoidance of withholding taxes won’t show up in the effective CIT rate on pre-tax profits.

   It’s for this reason that we haven’t simply calculated our headline figure for the amount of tax lost by the Zambian exchequer by applying 35% to the company’s pre-tax profits and subtracting 0.5%.

   However, the discrepancy over this period between the headline farming rate (15%) and the effective rate on the company’s pre-tax profits (0.5%) is also substantially the result of the refund from the farming rate change in 2007/8 and 2008/9, which *is* part of our analysis. And of course reducing taxable profits through offshore payments (‘Step 1’) to start with, will also make the impact of capital allowances on the ‘pre-tax profit CIT rate’ larger in any given year.

   In short, the 0.5% figure is likely a consequence of capital allowances incentive + tax incentive refund + pre-tax profit reduction.

   Perhaps we need to be clearer that we don’t say the tax loss from the activities we identify is the same as the difference between applying 35% and 0.5% to the company’s pre-tax profits.

   But the 0.5% figure is also significant for more fundamental reasons. If tax incentives mean that a massively profitable multinational company has an effective tax rate lower than a market trader which sells its products, have those incentives gone too far? (Indeed, while we focus on Zambia Sugar’s company-specific tax incentives, Zambia’s large capital allowances for investors are not uncontroversial incentives either, as our report notes: in the last budget the government sought to reduce some capital allowances particularly in the mining sector – the first wave in a wider Zambian review of tax breaks and incentives granted to big companies across all sectors during 2013).

   This wider question of the overall effects of Zambia’s tax system – the compound effects of general tax breaks, company-specific tax breaks and tax avoidance techniques – may not be so significant for Western campaigners seeking specifically to constrain tax avoidance by multinationals headquartered in their own jurisdictions. But it’s highly significant for Zambian citizens currently debating the Zambian tax system – a debate in which our colleagues at ActionAid Zambia, with whom we wrote this report, are fully engaged (see our report’s recommendations to the Zambian government on pp. 36-7).

   ***

   We’re delighted, again, to be debating these issues in detail. It’s through this kind of debate – here and in Zambia – that we’re going to move towards a fair tax deal both for the Majority World and for multinationals.

5. Thanks for the long reply 

   I must say I’m still somewhat confused. 

   You say the calculation of tax savings for the Ireland management company is done on the basis of  reducing taxable profits in Zambia, and are calculated based on CIT foregone. 
   But the report says (on page 19) says it is WHT”. While this complex and contradictory network of transactions makes it difficult to determine the substance and ultimate destination of Zambia Sugar’s payments to Ireland, we can nonetheless assess the straightforward losses to Zambia from offshoring these expensive management functions: we estimate that payments to Zambia Sugar’s Irish sister company, deducted from its own taxable profits and avoiding 15-20% Zambian withholding tax, have reduced the company’s Zambian tax bill by around US$1.2 million (ZK6 billion) a year: US$7.4 million (ZK32.6 billion) since 2007.” Which is it?

   On effect of capital allowances, yes it is mentioned but why is it not included in the estimates anywhere – as this is the largest single tax break for the period, not putting a number on it seems odd, and can only have the effect of playing down this factor. Why?

   Did you take a view as to how much of the gap between 15% and 0.5% is accounted for by the capital allowance? Do you assume it is the residual, or do you think there are other unexplained reasons for the low CIT rate?

   I see how in your example (ii) capital allowances and tax rate reductions could both reduce taxes. But as i understand it ABF are not saying they have written off 25% profit through capital allowances , but something closer to 100%. If so, this seems like a main area of disagreement on fact between you and ABF, but it is not explicit anywhere in the report. 

   “Unfortunately we’ve no evidence what the borrowing costs would have been without the WHT avoidance.  All we can know is that Zambian WHT is being avoided, to the benefit of Zambia Sugar and the lending banks.” – surely you could ask an economist and make an assumption? It looks like the  current implicit assumption is that zero of that WHT would have been passed on from the banks to AFB? How did you arrive at this?

   Effect of the farming rate change from at 35% to 15% tax rate – fair enough. 

   Tax refund for 2001/2 to 2004/5 ? You say “ABF actually agreed with us that using the ‘cash tax’ measure is appropriate: they said “we agree that cash tax paid is a good, clear and understandable measure of one element of a company’s contribution to the local economy.”

    …. Seriously?!?! ….the rest of that quote reads “…..however for the purposes of their report Action Aid fail to be totally clear.Corporate tax of R50m was paid in respect of the period 2006/7 – 2011/12. The receipt of the tax overpayment of R41 was for a prior period 2001/01 – 2004/5. These payments should not be netted as Action Aid have done.” You read that as agreement?!

   I understand you want to highlight the issue of of the retrospective tax refund and the court case but surely lumping the sum into 2006-2012 makes things like the quoted 0.5% tax rate and the comparison with  Irish Aid over the period meaningless,  as they are ratios based on different periods of time ( and discount any tax paid in 01-06).

   Have you got a chart of how you think the gap between 15% and whatever CIT ABF actually paid is accounted for and how they say it is, and similarly on WHT, what you think they avoided, where ABF agree with you and where they disagree. One barchart would simplify 1000 words!

6. Thanks Maya.

   1. Ireland management fees

   I’m looking at my spreadsheets here: the tax loss for this part is calculated based on reduction of taxable profits and thus CIT. The sentence – which I agree is definitely a bit confusing after editing – is simply meant to note that the payments also avoid WHT.

   However, if we’d calculated the tax loss in ‘Step 1’ based on the WHT avoidance (i.e. assuming that they were all *real* payments at cost, no profits shifted etc. but still avoiding WHT by routing via Ireland), the figure would have been almost the same. This is because the WHT for all these years is 15%, as is Zambia Sugar’s CIT rate for all years except 2006/7 (when it’s around 22%). So: calculation based on CIT loss is $7.4m; calculation based on WHT loss is $7.1m, the only variance being in 2006/7.

   By the way, I’d be more than happy to send you the spreadsheets showing all our calculations if useful.

   2. Capital allowances

   We don’t make a calculation of the tax loss to capital allowances because
   (i) we can’t tell from the publicly available information how much capital allowances are being claimed, and how much in which year; and
   (ii) we are confining our criticism to tax avoidance and special tax incentives actively and aggressively pursued by the company. We didn’t think it was fair to blame ABF for claiming general capital allowances available to all companies (although many Zambians might disagree). We don’t have a blanket position against all tax incentives.

   3. Factors causing low CIT rate

   As I said, (i) some will be due to capital allowances, (ii) some to the interaction of capital allowances + reduction of taxable profits from e.g. management fees, and (iii) some to the rate change rebate. I can tell how much is due to the rate change rebate ( -3% in 2008/9 and -32% in 2009/10). So the rest is presumably (ii) and (iii), but we can’t tell for sure: there may of course be other reasons too that aren’t discernible from the accounts alone.

   4. Interaction of tax rate reduction and capital allowances

   If you’re more than wiping out taxable profits with a capital allowance in one year, you generate a tax loss that you can generally carry into future years, so I think the argument still stands.

   And as Martin said, claiming capital allowances has absolutely nothing to do with the WHT avoidance, which is present in all three ‘tax haven transaction’ steps; nor to the rate change rebate.

   The ‘disagreement’ that ABF has foregrounded seems to have been largely invented by using the existence of capital allowances as an argument against avoidance; yet ignoring all of the avoidance we point out that has nothing to do with capital allowances.

   5. Borrowing rates

   You’re right that we didn’t make an assumption. Borrowing rates for these kind of loans are the result of a negotiation between the bank and the company, whose dynamics are very difficult to model. It’s certainly far from clear to me that in any negotiation between ABF and the banks, ABF would agree to shoulder the whole cost of the WHT through increased interest payments – on what is already a very high-interest loan (around 18%).

   And even if they did agree to shoulder *all* of the costs of the WHT through perfectly matched interest payment increases, bear in mind that any CIT lost due to the interest payments being larger would be much smaller than the WHT gained, because CIT would be charged at 15% of the 15% interest payment increase.

   Scenario 1 (straightforward loan, suffering 15% WHT)
   Interest payment: $100
   WHT at 15%: $15
   Reduction of CIT at 15%: $15

   Scenario 2 (doglegged loan, no WHT)
   Interest payment: $85
   WHT at 0%: $0
   Reduction of CIT at 15%: $12.75

   So in Scenario 2 (what we actually have) Zambia loses $15 of withholding tax, and only gets back $2.25 of CIT.

   6. Tax refund for 2001/2-2004/5

   Fair enough, agreement is probably too strong a word! They didn’t really ‘agree’ with us about much (although we agree with them about quite a lot…)

   But the point is that you can’t have it both ways: you can’t say “we think it’s fair to judge us on our cash tax, but we want you to calculate our cash tax differently to how we calculate it ourselves in our own accounts”.

   To be honest, I’d be quite happy to disaggregate as ABF suggests: i.e. say that the company paid the equivalent of just 4.5% of their pre-tax profits in CIT 2007-12, and then made the ZRA pay them back the equivalent of 4% of their profits in those years as retrospective refunds (for a tax break they won by presenting Africa’s largest industrial sugar factory as nothing but a farmer). And we’re clear when we present the refunds – including in ‘the bill’ table – that these are refunds received during these years due to a court case in these years, but for earlier years’ tax.

   I also disagree entirely that the aggregation makes things like the 0.5% figure and the comparison with Irish aid meaningless. What we are looking at with the 0.5% figure is the actual amount of money that the Zambian exchequer gained from Zambia Sugar during this period. This is what cash tax in company accounts shows – you can see it Zambia Sugar’s cash flow statement year on year from 2007-12, and we’ve made no adjustments to these figures at all. I think it’s pretty fair to compare this actual hard revenue in these yearswith aid receipts etc. in these same years.

   I suspect you’ll still disagree, although I hope you’ll agree that our disagreement is about the timing of the tax loss, not its size!

   7. Chart

   I’m not sure how we would produce such a chart, because the size of the capital allowances written down in each year aren’t available. In terms of our $27m figure for overall tax loss, I’d say:

   $9.3m farming rate change & refund: It seems ABF agree about their tax effect, but think it’s completely legitimate, and that the rebate should be compared with 2001-6 tax.
   $7.4m from Irish management fees: ABF disagrees this is avoided
   $3m from dog-legged interest payments: ABF accepts this arrangement cancels WHT, but says this made their borrowing cheaper so it’s a good thing because their investment can be larger.
   $7.4m avoided WHT on dividends: ABF has said, mystifyingly, that this doesn’t shift profits to the Netherlands. We never said it did. They accept that this arrangement leads to a 5% rather than 15% Zambian WHT, and have said nothing about how the Cooperatief enables them to take dividends out of the Netherlands without WHT too.

7. An official response from ActionAid to ABF’s statements is available here:
   http://www.actionaid.org.uk/103501/five_questions_for_associated_british_foods_to_answer_in_relation_to_actionaids_sweet_nothings_report.html

   It’s obviously not as detailed as Mike’s explanation of the various methodolgies, but gives an overview of the key arguments – plus a challenge to ABF.

8. Thanks for all the answers to the questions (I still have a couple more…) but the biggest question still is why put out a report and associated communications that put the headline attention on Zambia Sugar’s low CIT bill if that is not what you want them to respond about.

   e.g. this infographic distills the message for supporters

   http://www.actionaid.org/what-we-do/zambia-sugar/stop-abf-group-dodging-their-tax-zambia

   ‘Stop the ABF dodging their tax in Zambia’
   ‘ABF have paid “virtually no corporation tax” in Zambia since 2007 despite making 123m in profits’
   …
   ‘How much has Zambia lost: an estimated 27m since 2007’
   ….
   ‘How do they get away with it’
   -Shift profits into tax havens
   – Dodgy banking
   -The big shuffle
   -Get your own tax haven

   You and i know, having had this long and nerdy discussion, that in statements 1, 3 and 4 Action Aid is mainly talking about Witholding tax which as you say has nothing to do with corporation tax or profits. So statement 2 is a red herring, and the implied link between statement 2 and 3 and 4 is pure misdirection.

   But statement 2 is of course is the one that gets most of the headline mileage.

   I don’t think it helps the case of a grown up conversation about international tax systems to make complex subject more confusing with plays like this.

9. Hi Maya – I really don’t mean to prolong this discussion, but we are talking about both CIT and WHT. (Both, incidentally, are taxes on corporate income, just levied in different ways). I hope we have been clear about this.

   As our report and our new statement says, of the $27m tax loss we calculate, $10.4m is WHT loss, $16.6m is CIT loss thanks to a combination of avoidance and special tax breaks. I don’t think we’re ‘playing’, or misdirecting.

   1. Mike –

      Can I go back to the Irish management fees?

      I didn’t realise this was CIT. If you are calculating foregone Zambian CIT on the whole of the 47.6m, that means you are saying that it should all be declared as profit and none of it as a business expense?
      I.e. $47.6m goes out but none of it is used to pay for expat workers and contractors etc… Wouldnt that be fraud/ tax evasion?

10. We are not alleging illegality or fraud. We are saying that the result of offshoring management services – real or not- to an overseas, low-tax jurisdiction (in this case, Ireland) is tax loss to Zambia. (This is the same methodology, incidentally, as that used in our SABMiller report). It’s notable that in at least the last two financial years (2011/12 and 2010/11) the entirety of the payments from Zambia Sugar to Illovo Sugar Ireland appear to be management fees booked as income in Ireland, in the Irish company itself – a company, remember, that has no presence in Ireland and is somewhat confused about whether it has any employees or not. (The nature of the payments in prior years is unclear, since Zambia Sugar recharacterised past payments in its 2011/12 accounts wholly as management fees, but we can only see how much of this recharacterisation took place in 2010/11, the prior year shown in the 2011/12 accounts.)

    We’ve chosen to calculate it as a CIT loss, because we have been unable to establish the nature, substance or expenses of these services, in comparison to what is provided directly from the parent company in South Africa. But in addition, there’s the ‘tax on management fees’ perspective, which would lead to a similar figure (I know you’re not a fan of the WHT argument, so I hesitate to raise it here). In Zambia management fees paid to non-residents (including to non-related-parties) are charged 15% tax *in gross*. So if the counterfactual is that these fees are paid to expats resident in other countries and to Illovo Sugar Ltd in South Africa, they will bear a 15% tax *in gross* unless routed via Ireland. This is irrespective of the substance of the payments.

11. Incidentally, in the spirit of non-partisanship, I hope you’ve put similar questions to ABF about their response?! We’ve got five of them here:

    http://www.actionaid.org.uk/103501/five_questions_for_associated_british_foods_to_answer_in_relation_to_actionaids_sweet_nothings_report.html

    1. How does ABF justify booking 26% profits in low-tax Ireland on management fees paid out to a company that appears to have no presence or staff in Ireland?


    2. How have its Irish accounts been approved by its board – and cleared by its auditors – for seven years in a row, when these accounts, by ABF’s own admission, appear to have forgotten that the company had any employees?


    3. How can ABF claim it needed to move Zambia Sugar’s export sales operations to Mauritius in 2010/11 because “Zambia lacked the capability to manage [them]”, when the Mauritian company ostensibly doing the export sales has no employees in Mauritius at all, while Zambia Sugar has 3000-6000 staff?

    
4. How does ABF justify taking the Zambian government to court to get a tax break by presenting Africa’s largest industrial sugar factory as nothing but a domestic farmer?


    5. Why is ABF keeping confidential the details of the additional special ‘investment certificate’ tax incentive granted by the government to Zambia Sugar – especially when Zambian law requires the investment certificates the government awards to companies (the key qualifying document for a tax break), and details of the conditions and incentives granted to the company, to be on public record?

12. Mike,

    The question of whether the management services in Ireland are real or not seems pretty central.

    If they are not real (i.e. Zambia sugar paid $47m to a sister company in Ireland on the basis of false invoices for non existent contractors…) then that would be fraud, surely?

    If they are real, then why would you expect Zambia sugar to pay for contractors out of their profits, rather than as a business expense before tax?

    I don’t understand how the calculation that this was a CIT loss works out without the assumption of fraud.

    I agree you could have calculated a similar figure from the WHT loss, which would have been simpler (this is what I had assumed you had done), but you didn’t, and I am trying to understand the basis of the calculations and allegations.

13. Maya –

    As I’m sure you’ve seen, questions around the substance of the transactions to Ireland & Mauritius form 3 of the 5 questions we’ve challenged ABF to answer:

    1. How can ABF justify booking 26% profits in low-tax Ireland on management fees paid out to a company that appears to have no presence or staff in Ireland? If this isn’t for tax purposes, what is it for?


    2. How have its Irish accounts been approved by its board – and cleared by its auditors – for seven years in a row, when these accounts, by ABF’s own admission, appear to have forgotten that the company had any employees?


    3. How can ABF claim it needed to move Zambia Sugar’s export sales operations to Mauritius in 2010/11 because “Zambia lacked the capability to manage [them]”, when the Mauritian company ostensibly doing the export sales has no employees in Mauritius at all, while Zambia Sugar has 3000-6000 staff?

    Perhaps you could help get these points cleared by taking them to the company directly? They have not yet responded to ActionAid.

14. They are all good questions, and I hope that ABF answers them.

    But my question is how do you calculate that ABF has wrongfully deprived the Zambian exchequer of US$7.4 million by spending US$47.6m on purchases and management services, unless those services were in fact non-existent? And if that was the case wouldn’t that be fraud?

15. ActionAid’s research and campaigning focuses on tax avoidance by ABF, and the ineffective national & international tax rules that enable this to take place.

    Just yesterday, the OECD said on tax avoidance by multinationals, “These strategies, though technically legal, erode the tax base of many countries and threaten the stability of the international tax system,”

    http://www.oecd.org/tax/oecd-urges-stronger-international-co-operation-on-corporate-tax.htm

    We think ABF has an individual responsibility to shun aggressive and articfical tax practices, but that ultimately, governments need to make international tax rules fit for purpose.

16. Chris – that doesn’t answer the question. Here is the question again:

    Action Aid’s report says that Zambia Sugar avoided $7.4m of CIT by paying $46.7m to its sister company in Ireland. If that is the case then Zambia Sugar must have received 0 management services for this money. That sounds illegal to me. Please clarify.

17. ActionAid’s report says that Zambia Sugar legally avoided paying $7.4m since 2007 in *withholding taxes*. I suggest you read page 16-19 of the report:

    Click to access sweet_nothings.pdf

18. Chris.

    I have read pages 16 to 19..It is quite hard to tell whether it is based on WHT or CIT .I asked for clarification on it above and Mike said the calculation is definitely on Corporate Income Tax.

    February 13, 2013 at 9:44 am

    Mike: “I’m looking at my spreadsheets here: the tax loss for this part is calculated based on reduction of taxable profits and thus CIT. The sentence – which I agree is definitely a bit confusing after editing – is simply meant to note that the payments also avoid WHT”

    Then again

    February 13, 2013 at 7:26 pm

    Mike: “As our report and our new statement says, of the $27m tax loss we calculate, $10.4m is WHT loss, $16.6m is CIT loss thanks to a combination of avoidance and special tax breaks.”

    1. Wow! This is definitely my busiest post ever! But I think this discussion is going in circles now and you are not reading each other carefully enough.

       There are two reasons why both Maya and Chris seem to have got confused. First, because the figure $7.4m appears twice in AA’s calculations: coincidentally both the WHT lost through the Netherlands and the CIT lost through management fees both come to $7.4m (see the end of Mike’s comment on February 13, 2013 at 9:44 am). I think Chris has got confused and is referring to the former.

       Second, because Mike says he chose between two possible assumptions: 1) the payments are all real, but routed through Ireland, in which case WHT at 15% is avoided but CIT at 15% is not, 2) the payments are false, in which case 15% CIT is avoided but 15% WHT is not. As you can see, either way, the tax that AA estimates to be avoided works out around the same (there are the odd few anomalies that mean it’s not exact). In other words, the point about fraud may well be an interesting one, and it may well affect the total tax effect on ABF, but because of the coincidence that the WHT and CIT rates are the same, it doesn’t actually affect AA’s estimate of the tax foregone by Zambia.

       In the comment to which I’m replying, Maya, you could have pulled out the following couple of quotes from Mike’s comments, from which I worked this out:

       Mike Lewis February 12, 2013 at 7:18 pm

       “And while Step 1 *is* calculated a CIT loss, by routing management fees (to related or unrelated parties) via Ireland the company *also* avoids the 15% withholding tax ordinarily levied on management fees. Were we to calculate the tax lost from ‘Step 1’ based on the WHT loss rather than CIT loss, the tax lost would actually be greater, since Zambia Sugar’s ‘farming’ CIT rate dropped to 10% in 2012, while WHT remains at 15% and is rising to 20% in 2013.”

       Mike Lewis February 13, 2013 at 9:44 am:

       “However, if we’d calculated the tax loss in ‘Step 1’ based on the WHT avoidance (i.e. assuming that they were all *real* payments at cost, no profits shifted etc. but still avoiding WHT by routing via Ireland), the figure would have been almost the same. This is because the WHT for all these years is 15%, as is Zambia Sugar’s CIT rate for all years except 2006/7 (when it’s around 22%). So: calculation based on CIT loss is $7.4m; calculation based on WHT loss is $7.1m, the only variance being in 2006/7. By the way, I’d be more than happy to send you the spreadsheets showing all our calculations if useful.”

19. Martin

    Chris refers to pages 16-19 which are about Mystery Management not tax free takeaway (i.e. the other $7.4m) so i’m not sure that is the source of the confusion.

    Yes I see that the calculation would have worked out roughly similar whether you chose the CIT or WHT route (give or take 0.3m). But the report takes the CIT route in its allegations, so that is the one I am asking about.

    My question is: if ABF have done what the report implies in its CIT calculation on the Irish management fees, would that be illegal?

    That doesn’t seem to be a trivial question to me!

    1. OK! That wasn’t your question on February 14, 2013 at 1:57 pm, when you asked:

       “But my question is how do you calculate that ABF has wrongfully deprived the Zambian exchequer of US$7.4 million by spending US$47.6m on purchases and management services, unless those services were in fact non-existent? And if that was the case wouldn’t that be fraud?”

       An answer to that question is “through WHT avoidance”.

       Would it be fraud to deduct management fee payments if there was zero substance underlying them? I don’t know the answer to that, as it’s not my area. But from a tax side it would certainly not be arm’s length. From what I have seen in countries that don’t have much transfer pricing capacity, up to a certain threshold it tends to be understood that a company can get away with a payment without needing to justify it. I don’t know if such tacit acceptance is a factor here.

       1. Martin –

          I know this thread has gone on a bit, so just for clarity

          Mike confirmed on February 13, 2013 at 9:44 am that the Mystery Management calculation is based on CIT

          I have asked the same question at

          February 13, 2013 at 10:10 pm
          February 14, 2013 at 10:02
          February 14, 2013 at 1:57 pm (oops I forgot to specify that i was still talking about CIT)
          February 15, 2013 at 8:01 am

          And it still hasn’t been answered (other than by you saying ‘i don’t know’).

          It does not seem like a particularly complex question to me. There are two parts to it.

          Action Aid calculates that ABF has deprived the Zambian exchequer of US$7.4 CIT million by spending US$47.6m on purchases and management services from Ireland (‘Mystery Management’).

          Q1) I think this implies that Action Aid is alleging that these payments had no substance. Is this correct?

          Q2) Is it illegal for companies to submit and pay false invoices for business expenses?

          I am not a tax person, but Q2 seems pretty straight forward to me, or am I missing something?

          1. The trouble with question 2 is that “yes” would be defamatory, which may be why you are not getting an answer.

             Mike says at 08.59 on the 14th: “We are not alleging illegality or fraud. We are saying that the result of offshoring management services – real or not- to an overseas, low-tax jurisdiction (in this case, Ireland) is tax loss to Zambia…We’ve chosen to calculate it as a CIT loss, because we have been unable to establish the nature, substance or expenses of these services, in comparison to what is provided directly from the parent company in South Africa.”

             In the absence of certain evidence, they have chosen to use the upper bound of the possible CIT loss in their working out, which is also the lower bound of WHT. That choice is open to question, but i hope it’s at least clearer now.

          2. I don’t see how the answer to question 2 can be defamatory – it is a basic question about the law.

             It is the answer to question 1 that might be defamatory.

             Is there a difference between “they have chosen to use the upper bound of the possible CIT loss in their working out” and “they have chosen to assume that Zambia sugar received no services for their payments to the Irish company”?

          3. It depends in what context you use the word “assume”

          4. As in what assumptions underlie the calculation.

20. Hi Maya, I’m afraid I too feel we’re going round in circles a bit, but am happy to respond.

    My answers to your two questions are substantially the same as Martin’s, I’m afraid. And as Martin says, when writing and publishing about tax I am continually aware of m’learned friend and the UK libel laws.

    Q1) Despite extensive research we were not able to establish or verify the nature or substance of services provided from Ireland, no. We are therefore puzzled about profits being booked in Ireland for those services.

    Error and contradiction in the accounts also makes it hard, probably impossible, for us to tell what is substantive in the payments being made from Zambia to Ireland. First there’s the confusion over the existence of employees in the Irish company. Then there is the puzzle about an additional ‘secondment fee’ shown in the Illovo Sugar Ireland accounts as coming from Zambia Sugar Plc (three-quarters again of the management fees duly declared in both Zambian and Irish accounts). This ‘secondment fee’ payment doesn’t appear to be disclosed at all as a related party transaction in the Zambia Sugar accounts, despite it being clearly declared as coming from Zambia Sugar in the Irish accounts.

    As you’ll see in our correspondence with the company, in our first letter to ABF we asked them to reconcile the figures given as received and paid in the Zambian and Irish accounts. They declined to do so. It would be great if they still wanted to do this.

    What we can see is the impact of these payments on Zambia Sugar’s Zambian tax liabilities; not least because these payments are clearly declared as being deducted from profits in the operating profit note of Zambia Sugar’s annual accounts.

    Q2) It is in many jurisdictions illegal for companies to submit and pay false invoices for business expenses. Whether that’s happening or not in any given case is for the relevant tax authorties to determine. For clarity with m’learned friend, we are not making an allegation of this kind against ABF or its subsidiaries.

21. I can see why you’d be worried about libel laws, and I am sure you have legal advisors ensuring you don’t step over the line.

    But what about the integrity of the message and responsibility to readers and supporters?

    An informed public debate on taxation relies on the integrity of the information that supports it. 99.9% of people reading the report and associated publicity are going to come away with impression that Action Aid’s one year study found that ABF deprived the Zambian exchequer of $7.4 m in income tax by making dodgy payments to its Irish sister company (because thats what it says)…and further more that this is legal.

    In fact what you are telling me is that you have no evidence either way (although you have some suspicions) and the figure could be anything between 0 and $7.4m (but if it was $7.4 it would most likely be illegal).

    That these payments are declared as being deducted from profits in the Zambia sugar’s annual accounts doesn’t tell us anything. If they were 100% Kosher, or 100% crooked they would still look the same.

22. Hi Maya,

    thanks for your comment; but as before, that’s clearly not at all what I’ve said. So all I can do is say what I’ve said again.

    1) Offshoring high-value services outside of Zambia deprives Zambia of tax revenues, whether those services are ‘kosher’ or not.

    We recommend in our report: “Where paid-for services are being provided from tax havens, Zambia Sugar and ABF should aim ultimately to build the skills and expertise for those services in the countries where its main operations are located; and ensure that payments for them, at market prices, go directly to the company providing them.”

    2) In this case, we are also puzzled by the substance of those payments. Far from having “no evidence either way”, we looked at all the available accounting documentation, visited both Ireland and Zambia, and spoke with people involved with both Irish and Zambian companies, to try to establish the substance of services provided from Ireland. We couldn’t find such substance. This evidence is set out in our report. It’s now up to ABF to show what is being provided from Ireland to justify profits being booked there.

    3) Making an accusation of illegality is incredibly serious. That’s for a tax authority investigation to do, not us.

    Our responsibilities to readers and supporters are absolutely important. I’ve spent 5 days this week responding to the queries and concerns of just one reader, so I hope I’m not being lax in that department! (5 days when I am in fact on leave).

    One question: have you put any questions or queries to ABF? If not, why not?

23. Many apologies, I pressed ‘send’ too early.

    As we’ve said before, the possible Zambian tax losses from these transactions aren’t “between $0 and $7.4m”. The maximum is $7.4m (CIT loss on non-substantive payments). The minimum is $7.1m (entirely substantive payments, avoiding WHT by routing them through Ireland).

24. Mike – thanks for your time.

    I am still trying to understand how you calculate the 7.4m CIT loss.

    As i understand it, the CIT calculation is based on the idea that if Zambia Sugar did not pay any of the 47.6m to Illovo Ireland it would have had to declare it as profit in Zambia and pay tax on the whole amount at 15%. Is that right?

    So in this scenario if Zambia sugar were to have kept the money, cancelling any contractors it would have got through Illovo Ireland and not replacing them with people contracted elsewhere (because that would cost money and reduce taxable profits) it would have continued to operate just as well without them. That is what this calculation seems to suggest.

    A key point in the questions for ABF are whether it booked the 26% profits on management fees in Ireland for tax purposes. Lets assume it did. Lets assume that Zambia sugar could have contracted the same people itself at the same cost, keeping the 26% and thus booking am extra 4m profit in Zambia instead of Ireland. By my calculations this would result in an extra 0.6m in tax for Zambia.

    Why is this 0.6m particulatly important if you estimate that overall the Irish operation has cost Zambia 7.4m in CIT?

25. 1) Because we couldn’t find the substance of the management services provided for the management fees paid to Ireland, we’ve calculated that the whole as CIT loss to Zambia, yes. This is $7.4m.

    2) But even if you ignore all the evidence we gathered about lack of substance, and assume that in fact these are all ‘kosher’ management services provided from staff or contractors in South Africa and other countries, simply *routed* through Ireland, then this arrangement is still avoiding the usual Zambian WHT on management fees, which comes to $7.1m.

    Do let me know if that isn’t clear?

26. Hi Maya,

    See also my reply at February 14, 2013 at 8:59 am, which says the same thing.

    Best,

    Mike

    1. Mike – if the whole of the “management fees” quoted in the report were spurious transactions with no real underlying value the loss to the Zambian taxman would be $2.3m.

       The $7.4m figure implies that Zambia Sugar has spent the full sum of nearly $50m on phantom contractors over six years. That seems like an extraordinary claim, and would indeed be a scandal.

       Thanks for clarifying.

27. Pingback: Action Aid’s Sweet Nothing’s tax report: What’s the beef? « Hiya Maya

28. Thanks again for taking so much time clarifying and discussing with me. I have written up my takeaways on the report and communications around it.

29. Pingback: A campaigners’ code of conduct and the ABF story « Martin Hearson

Comments are closed.