(practical answers to questions raised in the field)
We've become accustomed to other announcements. The OECD's Centre for Tax Policy and Administration (CTP), which until recently was waxing lyrical about the "new tax law" associated with the implementation of pillars 1 and 2-1, is now concentrating its efforts on ways of relieving taxpayers in the current unprecedented health crisis linked to the Covid-19 2 pandemic.
For the urgency is both pressing and real. At a press conference held on March 2, the Organization's Chief Economist, Laurence Boone, used measured formulas to temper the alarming statistics observed by analysts 3. It's a fact: beyond its sad health consequences, the crisis is also having a brutal impact on the global economy. In the space of just a few weeks, it brought commercial activity to a virtual standstill on every continent. And while some economic sectors have weathered the storm better than others, most have seen their budget projections shattered, and losses and shortfalls have been recorded in all or part of the value chain.
In our modern economy, these value chains are highly concentrated within groups of companies. The effects of this crisis will therefore first and foremost affect transfer pricing, which governs contractual and financial relations between entities under common control. The question then arises as to how to allocate these effects between the various players, with the underlying components of identification (which players must necessarily bear these effects) and quantification (what proportion of negative effects to attribute).
Of course, the concrete effects of this unprecedented situation will have to be assessed over the long term and with the necessary hindsight, in the light of the regulatory changes that are sure to follow; the paradigm shifts of the tax authorities; and the operational strategies of economic players. Any current analysis must therefore be both cautious and humble.
However, without claiming to be exhaustive, we shall endeavour in the following developments to provide answers to the questions that are on the minds of groups of companies, some of which have already been brought to our attention by those in the field. To do so, we will follow a logical, even chronological, intellectual path, addressing the attribution of losses and proposing a few recommendations for modifying existing transfer pricing policies.
Can all the companies in a group make losses?
The question of loss management in a transfer pricing environment highlights a binary and limiting approach, which in reality has often led over the years to the belief that losses can only be borne by one party. However, this belief cannot stand up to objective legal and economic analysis.
Let's refresh our memories. The arm's length principle developed by the OECD and adopted by almost all countries 4 requires the parties to a group of companies to remunerate the transactions that bind them in the light of practices observable in third parties, independent of each other and placed in comparable situations. Because it tends to focus on each linked transaction taken in isolation, this concept of arm's length competition has nevertheless gradually led to linked companies being placed in a blindly binary relationship, reduced to a creditor and a debtor, for whom practice has substituted the terms "entrepreneur" and "routine operator". But to quote the illustrious professor Nicolas Rontchevsky: "all practitioners copy and repeat each other, the problem is that the first one was an idiot".
The truth is that these terminologies are meaningless, and indeed are absent from the OECD guidelines, which form the basis of modern transfer pricing taxation.
This has not prevented practitioners, led by the tax authorities, from developing a Manichean approach to the modern economy, and from considering that the party to the related transaction who takes on the profile of the routine operator should receive a fixed remuneration, as opposed to the entrepreneur, who would be entitled to any residual profits or losses. Year in, year out, this binary vision has in many cases led to the conclusion that the routine operator cannot suffer losses.
One limiting belief generating another, we have seen the emergence of an appropriate concept, that of the "limited-risk" routine distributor or manufacturer, which is scarcely described any further. In the faith of a tax lawyer, we have read in numerous proposals for rectification that this "limited-risk" distributor or manufacturer, because he takes on a routine profile, should receive "a low remuneration, in line with his functions and limited risks, but necessarily positive".
But at what point in the history of the contemporary economy was it considered that an economic player would be protected against the risk of loss? It's a question that tax specialists have for too long avoided answering, no doubt hiding behind the principle of the autonomy of tax law (another misguided concept). In reality, we have to recognize that any company, regardless of its sector of activity, the functions it performs or the risks it carries (what practitioners call its "functional profile"), can at some point in its history sustain losses, whether chronic or sporadic.
In the first place, and unless the intra-group contract provides otherwise, the notion of "limited risk" should in no way imply total immunity from all risk. It is a common feature of any business to face up to the volatility of its market, and to bear the potentially adverse consequences of its decisions. Whether civil or commercial in nature, a company may incur expenses that exceed its profits, thereby generating a deficit. This remains true even for structures whose purpose is not to make a profit, such as sociétés de moyens or groupements d'intérêt général. In the case of the latter, case law has had occasion to point out that, even when placed in a relationship of dependence, their very purpose justifies the absence of profit 5. However, there is no reservation as to the impossibility of recording losses.
Generating deficits can even be part of an economic strategy, in line with a company's social interest. As the public rapporteur pointed out in his conclusions in the aforementioned case: "no provision of the General Tax Code obliges a company to make profits 6". This leads to a second argument. It has to be said that nothing in positive law prohibits the making of losses, as long as they do not reflect an abnormal act of management. This is the case for market penetration strategies; long-term strategic investments; and what the OECD calls the "portfolio approach", which can be found in the automotive or consumer goods sectors, to name but a few7.
The use of tax losses is the subject of several articles in the French General Tax Code. Just the third paragraph of Article 209-I describes the calculation of the loss that can be carried forward with almost automatic precision. Given that the use of losses is clearly and expressly described by the law, without reserving the regime to a specific category of taxpayers, how could we consider that some, on the grounds that they are part of a group, would respond to a different logic? This would violate the very essence of the arm's length principle, as well as breaking the adage learned religiously in law school that "where the law does not distinguish, there is no reason to distinguish".
However, there is another law that may apply in this case, namely "the law between the parties". This concept, which stems directly from the principle of the binding force of contracts, derives from article 1103 of our Civil Code. It assumes that, if a contract is balanced and based on free and informed consent (which is, in fact, a presumption between companies), the parties to the agreement undertake to respect its terms. Regardless of the specific nature of tax matters or the economic environment, this principle is a cardinal point in our legal system, which is based on written law. It is therefore imperative, before any adjustment or modification of the transfer pricing policy, to review the contracts in order to determine the allocation of the obligations of each party to the transaction.
This aside, the question could legitimately be asked about commission agents making losses. Although commercial in essence (the status of commissionaire derives from article L132-1 of the French Commercial Code), commissionaires fall automatically into the chapter of the Civil Code dedicated to agents. Once again, this status is clearly defined and framed by the Civil Code, in particular articles 1984 and following, which commercial judges have helped to enrich over the years. The French Supreme Court (Cour de cassation) initiated a praetorian trend in favor of gas station attendants operating a brand concession. The supreme judge ruled that the contracts binding them to oil groups must necessarily cover losses incurred during management 8. Lastly, it specified that article 2000 of the French Civil Code precluded the parties from contractually charging the agent for losses caused by an event attributable to the principal 9.
The syllogism then becomes tempting: considering that an agent cannot legally or contractually incur losses in the exercise of his mission; and that a commission agent is by nature invested with a sales mandate; should the latter be obligatorily protected against any risk of closing with a loss?
In our view, this is a rather short-sighted legal approach. While the commissionaire must be remunerated in such a way as to cover his costs, the law does not grant him the ability to spend lavishly. Expenses that are the result of his own decisions must be his responsibility. Here again, it is essential to review the contract with the client and ensure that it includes certain safeguards. Finally, we believe that the mandate given to the commission agent is not valid for all the activities commonly carried out during the year, but for each sales transaction. As soon as a sale is made, the commission agent must receive adequate remuneration. If demand collapses and no sales are made, then the commission agent could be considered to be unjustified in claiming his commission. In the event of a significant drop in sales affecting the entire value chain, this could justify losses on the part of the commission agent.
How do you allocate losses between the parties to an intra-group transaction?
The most logical approach would be to allocate these in proportion to each party's contribution to the total value chain. In practice, this amounts to implementing the "profit-sharing" method (which also applies correlatively to losses) described by the OECD10.
This method, which has long been reserved for transactions in which the parties make unique, high-value contributions (single intangible assets, for example), is being revitalized under the impetus of the BEPS project. Indeed, the Organization is advocating greater use of this approach, given the known weaknesses of traditional methods, or the "transactional net margin method", which remains the most widespread. It is also at the forefront of recent work on Pillar 1 and the unified approach that will eventually lead to new taxing rights for States.
However, it has to be said that the profit-sharing method is complex to implement and requires the use of accounting and economic aggregates that are difficult to recover in multinational groups. For the time being, experience shows that administrative authorities still lack maturity in the use of this method, and thus tend to set it aside in favor of more traditional formulas, such as the transactional net margin method.
No more profit-sharing. To test the arm's length nature of a transaction, it seems to us more appropriate and simpler to calculate a theoretical remuneration, such as would have been derived from the activity if it had been carried out at a normal period. In practice, this amounts to granting a profitability based on an appropriate basis (costs incurred or sales achieved) and theoretically defined on the basis of known, anticipated aggregates, disconnected from any extraordinary event such as the current crisis. Budgets, when drawn up jointly by the two parties to a transaction, can serve as an appropriate reference. Their reliability will be all the greater if, in the past, it can be demonstrated that the variations between budget projections and the actual data recorded at year-end were small. In some cases, this discrepancy may even serve as an adjustment variable in the implementation of our theoretical method.
In this respect, this approach is in line with that previously validated by the tax judge in the Unilever 11 case. In this case, the company manufactured margarine under the Astra brand and sold it to a Belgian company belonging to the same group. Following an audit of its accounts, the tax authorities questioned the selling prices on the grounds that the company's margins had become structurally unprofitable. Not only was the company's operating profit (often used as a benchmark by the authorities) negative, but its gross margin was also negative, as the company was unable to cover its manufacturing costs. The company argued that this situation was due to the obsolescence of its production units, which in another environment corresponds to a situation where the site is not operating under normal, stable circumstances.
The Court dismissed the administration's claims, agreeing with the recommendations of its Government Commissioner. With economic wisdom, he considers that "it is in a manufacturer's interest to sell at a loss in order to cover its variable costs and part of its fixed costs, in other words, to survive". In his conclusions, he also lists the cases in which invoicing at a loss is sometimes the only way for a company to sell at the market price. The last case corresponds to a company in difficulty, due to the economic situation. In this case, selling at a loss in order to sell at the market price may be a way of continuing to operate, covering part of the fixed costs, the time needed to invest, or even to retrain. In addition, and commercial considerations aside, it should be remembered that tax judges do not rule out on principle the advantage of abnormal invoicing between related companies, if this is the only way to maintain employment on a site12.
In order to take a fresh look at any abnormalities linked to intra-group transactions, the Court recommended reconstituting the margins if Astra's production line were operating to normal standards, i.e. free of the extraordinary circumstances that affected its margins. In such an environment, we believe it would be just as conceivable to recalculate the company's profit and loss accounts by placing it in a stable and sustainable situation, for example, by taking into account the budgets achieved before the crisis. Unanticipated costs resulting from the effects of the crisis will not be taken into account, even though they will have a negative impact on the company's total net margin.
What comparables should be used to test the arm's length nature of transactions?
As we have already seen, it should be not only permissible, but perfectly defensible, to book losses on intra-group transactions, even if the transfer pricing policy initially allowed for a margin to be granted to the company. There is, however, a limit to this approach: it must reflect market conditions. It must therefore be demonstrated that, in the same situation, independent and comparable companies would not do much better. This brings us back to the problem of finding reliable comparables on which to base our argument.
Clearly, the comparables currently available on the databases do not incorporate the economic effects of the Covid-19 pandemic. At the time of writing, the most recent data concern financial years ending March 31, 2019, i.e. one year ago. Back then, no one imagined what the world would go through, and yours truly didn't know what a pangolin was. It is therefore necessary to artificially adjust the comparables available to us. Such adjustments are, moreover, permitted and recommended by the OECD, precisely to neutralize observable differences in comparability factors, which include economic circumstances.
A first solution might then be to apply the percentage decline observable in a given sector or country, and derived from statistics kept by INSEE, the OECD or other professional bodies, to the margins of comparables.
An alternative, or corroborative, method might be to calculate the profitability of an activity over the months of the financial year not yet affected by the crisis. In this way, it would be possible to return a true picture, observed over a normal period. These data would then be compared with the most recent references from databases. In the event of a match, it would then be possible to conclude that, apart from exceptional cases such as the current crisis, the transfer pricing policy pursued until then reflected an arm's length situation. However, this approach only works for companies that do not close on December 31, and for which a sufficiently representative range of months is available.
Finally, a third avenue could be to investigate the behavior of the comparables used to previously test the transfer pricing policy, during periods of crisis. For example, once a panel of comparables has been used in 2019, it would be interesting to look at the variation in margins of these same comparables during the financial crises of 2008 and 2010. Of course, despite their violence at the time, it has to be admitted that these crises had nothing in common with the one we're experiencing today. But this approach will at least demonstrate that independent companies with comparable reputations can also see their results rocked during troubled times. Once again, the authorities will be able to argue that companies with a similar, albeit "limited", profile can suffer economically without being protected against losses.
Can intra-group contracts be suspended or modified?
Although transfer pricing is heavily influenced by economic theory, it is not immune to the most basic legal concepts, starting with the binding force of contracts. Before taking any action, it is therefore imperative to review the content of intra-group agreements with a critical eye, in order to ascertain the conditions under which a contract may be suspended, terminated or modified.
A first temptation could be to invoke force majeure. It should be remembered that force majeure constitutes an event beyond the debtor's control, irresistible and unforeseeable at the time of conclusion of the contract13. The advantage of force majeure is that it leads to temporary or definitive exemption from performance of obligations, or to termination of the contract, as the case may be, which in this instance could allow related parties to depart from transfer pricing policies.
In fact, force majeure has never been recognized in the case of pandemic events, since they have not, until now and fortunately, sufficiently affected the activities of economic players. The situation could be different in the case of Covid-19, as there have already been a number of court rulings in the specific cases of administrative detention or administrative measures to remove foreign nationals. In these specific cases, the judges considered that the impossibility of acting as a result of the pandemic constituted a case of force majeure, notably due to the closure of borders. In addition to these remote examples, we might also be tempted to add a reference to Bruno Lemaire's general position on public procurement, according to which the French government will consider the coronavirus as a case of force majeure for companies.
In any event, it is essential to check the terms of any force majeure clauses in contracts and general terms and conditions that may apply, as well as the procedure to be followed where applicable. Indeed, the rules laid down contractually may be more flexible than those of ordinary law, which are merely suppletive. The fact remains, however, that it will be more difficult to withhold payment on the grounds of force majeure, as judges will consider that payment is not fundamentally prevented.
In certain situations, therefore, it may be more appropriate to examine whether unforeseeability can be more usefully invoked than force majeure. Force majeure does not apply to difficult circumstances that have made performance of the obligation particularly onerous, but not impossible. In such cases, renegotiation of the contract can be envisaged, which is logically easier between bound parties. In this context, the parties could legitimately make some temporary adjustments to the transfer pricing policy, and in particular, transgress the remuneration methodology, while providing for an adequate and fair allocation of losses.
What impact do exceptional regulatory provisions have on transfer pricing policies?
To support their taxpayers in this turmoil, the vast majority of countries adopted emergency measures, some aimed at reducing the tax burden, others, like France, preferring to postpone or adjust the timetable. Some corporations have even obtained temporary aid in the form of subsidies, grace periods or grants (notably interest-free loans). However, once these measures have an impact on the company's earnings, the question arises as to how they should be treated when calculating the transfer pricing method. Most transfer pricing methods aim to attribute a fixed margin to the tested party. Should these subsidies be included in the calculation of this margin?
In the substratum, this question inevitably refers back to the Philips decision by the Conseil d'Etat 14. The question then was whether, for the purposes of calculating the net margin on the basis of total costs that Philips re-invoiced to its parent company, the latter was entitled to deduct the research tax credit from which it benefited from the cost base subject to re-invoicing. Quite logically, the French tax authorities replied in the negative, arguing that the subsidies received were in the nature of investment subsidies and were therefore unrelated to operating expenses, which form the basis of the transfer pricing method used by the group. However, the French Supreme Court overturned the department's claims, confirming the appeal decision, on the grounds that the tax authorities had not demonstrated that independent companies in comparable circumstances would not have deducted the amount of the CIR from their calculation basis.
While this ruling unfortunately does not settle the fundamental question (how to take subsidies into account when calculating the transfer pricing method), it does have the merit of referring to the "normal" behavior of parties in the free market. In this respect, it is unlikely that in times of crisis, when every player in the value chain is affected, one party to a transaction will make the other bear the brunt of a gift offered by another party, in this case the State. In other words, any exceptional benefits received should not be added to the cost base subject to the cost price method, nor, conversely, should they reduce the base retained under the resale price method. What's more, in order to prove that the non-inclusion of these subsidies is abnormal, the tax authorities would still have to produce reliable comparable references, which, given the current state of the databases, is an impossible task.
However, a more recent decision could cast doubt on our position. In the Laps France 15 ruling, the court held that the taxpayer had to re-invoice the CVAE to its related party on the same basis as its other operating costs, since the method adopted by the group required it to re-invoice all operating costs. As the CVAE was booked as an operating expense (and not as tax), it was automatically included in the base on which the net margin was calculated. In our view, however, this line of reasoning suffers from a number of flaws, in addition to referring to an issue that is quite different from the one analyzed here.
In fact, in this case, it was a question of re-invoicing an expense that had a negative impact on the company's profit, and not a subsidy received. If the benefits received were to be re-invoiced, the company would automatically benefit twice, in addition, as we have seen, to condemning its economic partners a little more. What's more, a comparability analysis based on data, adjusted as we suggested above, would undoubtedly produce greatly reduced margins, whereas our comparables' margins would necessarily be increased.
In short, benefits, subsidies and other gifts received should be considered as exceptional income, and therefore not included in the calculation of any transfer pricing method.
1 On March 31, 2019, the OECD-G20 Inclusive Framework on BEPS published its "Work Programme to develop a consensus solution addressing the tax challenges raised by the digitization of the economy". This document, adopted by the 129 members of the Inclusive Framework, was approved by the G20 Finance Ministers on June 8 and 9, 2019.
2 See on this subject "OECD, FORUM ON TAX ADMINISTRATION, Tax Administration Responses to COVID-19: Measures Taken to Support Taxpayers, 26 March 2020", available on the OECD website.
3 OECD, OECD Economic Outlook, Interim Report Coronavirus: The Global Economy at Risk, March 2, 2020, available at http://www.oecd.org/perspectives-economiques/
4 In France, Article 57, the cornerstone of our transfer pricing regulations, has long been recognized as compatible with Article 9 of the OECD Model Convention, which incorporates the arm's length principle. See, for example, CE, ruling of March 14 1984, no. 34430 and no. 36880.
5 See in this respect CE 25 nov. 2009, 3rd and 8th ss sect. réunies, no. 307227, Cie Rhénane de Raffinage.
6 Conclusions M. Geffray sous CE 25 nov. 2009, 3ème et 8ème ss sect. réunies, n°307227 in BDCF 2010 2/10 n° 106.
7 OECD Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations, July 2017, §3.10.
8 Cass com 17 déc. 1991, n° 89-20688 ; 90-11661.
9 Cass com 26 oct. 1999, n° 96-20063.
10 OECD Transfer Pricing Guidelines, §2.114 et seq.
11 CAA Versailles, 6ème ch. 5 déc. 2011, n° 10VE02491.
12 For example CAA Nancy 6 mars 1996 société Nord éclair n° 94-1326 n°1464.
13 Article 1218 of the French Civil Code.
14 CE 19 sept. 2018, n°405779, Sté Philips SAS.
15 TA Montreuil, 9ème ch. 14 février 2019, n°1801945.
