In recent years, article 39-1-3 of the French General Tax Code has become one of the best-known tax provisions. Like a kind of code, a digital password reserved for a community apart. We no longer speak of the deductible interest rate. We say "the 39-1-3 rate". Take the test and you'll be able to spot the tax experts in the audience.
As a reminder, this text, as incongruous and anti-economic as it may be, specifies that "interest paid to partners on sums they leave or make available to the company, in addition to their share of the capital, whatever the form of the company, [is deductible] up to the limit of that calculated at a rate equal to the annual average of the average effective rates charged by credit institutions and finance companies for variable-rate loans to companies, with an initial term of more than two years".
Incongruous, because it is a French speciality, a secret of Bercy's, and adds to an already long list of measures designed to restrict companies' financial burdens.
Anti-economic, as it forces companies to limit the tax deductibility of the interest they pay to minority (or even very minority) shareholders, whereas an exception is made for shareholders holding more than a majority of the debtor's share capital, thanks to the alternative provided for in Article 212-I. In practice, this can lead to companies having taken out a loan, or having issued bonds to third-party banking institutions, and which, as is often the case, acquire a stake in the said company in order to follow its development. Consider the paradox: the BPI, which contributes to the growth of our economy, very often holds minority shares in the capital of the companies it supports. These same companies will therefore not be able to deduct all the financial interest they pay it, since the interest rates charged to them when they are starting up, or for the bonds they issue, are statistically well above the rate of the famous article 39-1-3.
But if these minority shareholders are located in a country other than France, can the provisions of Article 39-1-3 of the CGI withstand the arm's length principle referred to in Article 9 of the OECD Model Tax Convention? The arm's length principle may in fact offer a radically different rate, since it reflects market conditions and what independent companies in a similar situation would have negotiated with each other.
If we consider for a moment, the limitation provided for in this article 39-1-3 contradicts article 9 of the model, which is found in all the treaties signed by France, and which allows an alternative rate to be applied. The principles of subsidiarity and the primacy of treaties would then have to play their role, dear to the minds of tax law theorists, and thus give precedence to the possibility for related parties to demonstrate the correctness of the rate they have actually applied.
In many conventions, Article 9§1 refers to the case of "an enterprise of a Contracting State [which] participates directly or indirectly in the management, control or capital of an enterprise of the other Contracting State, or the same persons participate directly or indirectly in the management, control or capital of an enterprise of a Contracting State and an enterprise of the other Contracting State". However, direct or indirect participation in the management or capital does not require a majority shareholding. If we leave it at that, the arm's length principle could therefore be applied to any partner who, by nature, "participates in the capital". But since every good tax expert is a potential paranoid, let's go a step further and take a moment to consider the notion of "control", which seems more ambiguous. It has to be said that the OECD's definition of this notion is as grey as the concrete of the château de la Muette that houses its services.
The OECD comments state that "Two enterprises are associated if one of them fulfils the conditions set out in Article 9 paragraphs 1a) or 1b) of the OECD Model Tax Convention in relation to the other enterprise".
The comments on Article 9 of the Model Convention state that "the Committee has devoted considerable time and effort (and continues to do so) to studying the conditions of application of this article, the consequences of such application, and the methodologies applicable for the adjustment of profits when transactions have been concluded under conditions other than at arm's length. The conclusions of this study are described in the report entitled "Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations", which is periodically updated to take account of developments in the Committee's work on this issue. This report represents internationally accepted principles and provides guidelines for applying the arm's length principle, of which Article 9 is the authoritative statement".
So let's continue the treasure hunt and refer to the OECD guidelines, which reproduce the same definition as the comments under the model convention.
It therefore appears that the notion of control, which is key to the applicability of the arm's length principle, is not explicitly or formally defined by international law standards. As nature abhors a vacuum, it is best to refer to the domestic law of each State.
In our case, Article 39-12 of the French General Tax Code stipulates that "two companies are deemed to be dependent on each other when: (a) one of them directly or through an intermediary holds the majority of the share capital of the other, or in fact exercises decision-making power over it; (b) both of them are placed, under the conditions defined in (a), under the control of the same third party". From the point of view of French positive law, therefore, control seems to be assessed from the angle of majority (of capital or of decision-making power).
Would this have led to a dead end, leaving the taxpayer to face the cold injustice of article 39-1-3? Well, perhaps not for long. On September 12, 2023, the European Commission unveiled two draft Directives, on the "BEFIT" initiative and on the harmonization of transfer pricing rules within the EU. The second draft directive aims to simplify the applicable rules and reduce the risk of double taxation, by incorporating the arm's length principle into EU law. One of the ways in which tax certainty could be strengthened is by harmonizing the main transfer pricing rules, creating the possibility for the Commission to establish common rules on specific subjects within the Union.
In order to define an associated undertaking, Article 5 of the draft Directive proposes a minimum threshold of 25% of an entity's voting rights, capital or profits, in order to establish a relationship of dependence. Transposition of the Directive as it stands would therefore automatically and necessarily lead to minority shareholders being able to rely on the arm's length principle, and thus defeat the provisions of Article 39-1-3 of the CGI, which, by freezing an interest rate, contradicts the alternative offered by international law, and more specifically Article 9 of the Model Tax Convention. The constraints currently imposed on minority shareholders by the provisions of article 39-1-3 of the CGI would therefore be lifted for certain shareholders only, who on the one hand hold at least 25% of the share capital, either directly or indirectly, or who are placed under a common entity exceeding this threshold; and on the other hand, who are resident in a country other than France.
However, our positive law would create a double discrimination, by keeping under the yoke of Article 39-1-3 ultra-minority partners, holding less than 25% of the debtor company's capital, and partners established in France. For the latter, reverse discrimination has already accustomed us to treating purely domestic operations less favourably.
For foreign partners holding less than 25%, the taxman will tell you: you always need someone smaller than you.
