Bloomberg Tax
Free Newsletter Sign Up
Bloomberg Tax
Advanced Search Go
Free Newsletter Sign Up

Determining the Arm’s Length Interest Rate—Recent Guidance in France

Sept. 9, 2022, 7:00 AM

The deductibility of interest on intragroup financial transactions has become a major financial issue for corporate groups over the years, including in France. The context of the Covid-19 health crisis only reinforced this, since the financial difficulties faced by many French companies encouraged the use of intragroup financing to meet the financial needs of operating subsidiaries.

The French tax authority’s position, and the case law in this matter, have not always been clear, creating a risk of legal insecurity for taxpayers.

However, the landscape has become much clearer in recent years with a series of high-profile decisions by the Conseil d’État (Supreme Court), the publication in February 2020 of the Organization for Economic Co-operation and Development report, “Transfer Pricing Guidance on Financial Transactions” (which has been integrated into Chapter X of the new 2022 edition of the Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations), and the publication by the Public Finances Directorate General (DGFiP), the French tax authority, of eight “practical sheets” on the subject in January 2021.

As a reminder, in accordance with Article 212-I of the CGI—the general tax code—interest paid between related companies is deductible up to the limit provided for in section 1, subsection 3 of Article 39 of the CGI (1.16% between Jan. 31 and Feb. 27, 2022); or, if it exceeds this reference rate, based on the rate that the borrowing company could have obtained from independent financial establishments or organizations under “similar conditions” (the burden of proving the normality of the rate lies with the borrowing company).

To justify the arm’s-length nature of the interest rates applied to intragroup loans, the taxpayer most often uses the comparable uncontrolled price (CUP) method based on external data, in the absence of internal comparable transactions. This method consists first in determining the credit risk (or “scoring”) of the borrower, i.e., its ability to pay interest and repay the debt, and then in searching a reference financial database (such as Bloomberg, Moody’s, Standard & Poor’s) and for the loan period or the market rates applied for companies with a comparable scoring to loans with similar characteristics. Finally, an arm’s-length interest rate range is constructed from the rates observed for the selected comparable transactions.

This methodology, long criticized by the tax authorities as well as by certain courts (e.g., TA Paris, Jan. 18, 2018, no. 1707553/1-2, SAS Studialis), has now been expressly endorsed both by case law (CE, July 10, 2019, no. 429426, SAS Wheelabrator Group; CE, Dec. 10, 2020, no. 428522, WB Ambassador) as well as by the tax authorities within the framework of its published practical sheets (Sheet 2).

The clarifications provided by these different sources now make it possible to define an initial legal framework within which any economic study must be carried out to be able to survive in a litigious context, even if numerous practical difficulties remain.

Practical Difficulties in Determining a Borrowing Company’s Score

In general, a credit rating is used to estimate the risk of default for any borrower company. These ratings are obtained following a thorough analysis of the borrower’s financial statements, financial condition, and market position. However, the determination of a credit rating by a rating agency is part of a long and costly process.

In practice, it is difficult to require that such an exercise be carried out each time a company engages in intragroup financing. Taxpayers generally use scoring software to assess the creditworthiness of a borrowing company. This software is mainly based on the borrower’s historical quantitative data (accounting and financial data) and incorporates certain qualitative data—geographical area, business sector, macroeconomic risk, or the company’s competitiveness, etc.

  • Use of rating software

The tax authority has long questioned the use of such software, considering it unreliable insofar as it was based on historical quantitative data provided by the company itself (position followed by CAA Paris, June 23, 2020, 20PA00585, SAS Willink).

In its BSA decision (CE, Dec. 11, 2020, no. 433723, Société BSA), the Conseil d’État put an end to this debate by validating once and for all the use of publicly available rating software to evaluate the credit rating of a borrowing company. The rapporteur public highlighted in her conclusions that even if the use of scoring software does not have the same level of finesse as an official analysis conducted by a rating agency, and these scoring tools are supplied by the company itself, the data are sourced from the company’s balance sheets and income statements and the company cannot modify the parameters used.

The Court of Appeal (CAA Versailles, Dec. 29, 2021, no. 20VE03249, Sté BSA) recently confirmed this position by noting, on the one hand, that a rating agency does not have to be used, given its cost, in an intragroup transaction and that, on the other hand, the rating provided by the scoring software used in this case can be considered as sufficiently reliable, in the absence of any detailed criticism, to justify the risk profile of the borrowing company.

The OECD report published in February 2020 accepts the results obtained via such software provided that the financial information used was checked and that both quantitative and qualitative information were used. It also indicates that the use of such tools may be appropriate if it is possible to “demonstrate consistency of ratings using such tools with those provided by independent credit rating agencies” (OECD Guidelines, section 10.74).

  • Consideration of group membership

Beyond the use of software, the question of how to consider the effect of group membership on the scoring also arises.

In this respect, the Siblu decision (CE, March 18, 2019, no. 411189, SNC Siblu; see also CE, June 19, 2017, no. 392543, Sté General Electric France) of the Conseil d’État requires that the market rate be assessed with respect to the specific characteristics of the borrowing company, and not of the group to which it belongs—the “standalone” approach.

However, in its practical sheets (Practical Sheet 3), the tax authority has clarified the consideration of the implicit support of the group for the rating of a subsidiary, and has indicated that when a subsidiary is at least moderately strategic for a group, the rating agencies consider the implicit support of the group to establish the final rating of the subsidiary. This position is in line with that of the OECD (OECD Guidelines, sections 10.76–10.80).

  • Consideration of consolidated financial data

However, the standalone approach adopted by the Conseil d’État is not incompatible with the use of the consolidated accounts of the borrowing company and its subsidiaries. The Apex Tool decision (CE, Dec. 29, 2021, no. 441357, Apex Tool) confirms that “the risk profile must [...] in principle be assessed with regard to the consolidated economic and financial situation of the company and its subsidiaries.”

The rapporteur public justifies this decision by the fact that (i) the economic and financial situation of the subsidiaries affects the risk profile of their parent company, as their strength may contribute to improving it and their weakness to worsening it; (ii) an independent banking institution naturally takes this into account when developing its credit offer; and (iii) the consolidated financial statements make it possible to take into account the actual amount of assets and liabilities on the balance sheets of the consolidated companies.

This position is shared by the French government, which recommends that in assessing the credit risk of a borrowing company, “the assets that it controls directly or indirectly, and consequently the prospects of any of its subsidiaries, should be taken into account” (Practical Sheet 3).

  • Impact of the loan characteristics

The characteristics of a loan—repayment priorities, collateral, etc—can also have a substantial impact on the borrower’s credit risk.

In practice, it is common to apply adjustments to reflect the characteristics of the transaction. Some credit risk agencies propose adjustments applicable to credit risk depending on the characteristics of the issue tested (guaranteed, unguaranteed, subordinated, etc.). In particular, they specify that a subordination rank—for example, junior debt compared to senior debt—justifies the application of a downgrade of one to two notches of the initial rating. Conversely, if an issue has a guarantee, they justify raising the issuer’s initial rating by one notch.

Other questions also arise in practice, particularly regarding the inclusion of forecast data, exceptional transactions, or isolated purchase-resale transactions.

  • Practical difficulties of finding comparable data

On a practical level, taxpayers regularly use bond markets, which offer public transparent data on transactions concluded between independent parties and usable by all actors. The possibility of using bonds as comparables has been validated by the case law (SAS Wheelabrator Group; WB Ambassador), and then by the tax authority, provided that the comparison is made with companies in economic conditions comparable to those of the borrower and that the bonds constitute, in the hypothesis under consideration, a realistic alternative to an intragroup loan.

  • Main comparison criteria

The comparison of financial market transactions is based primarily on these characteristics: issue date, currency, maturity, risk profile of the borrower, degree of subordination of the loan. In practice, however, it is extremely rare to find “perfect” comparables. It is therefore common to conduct multiple searches based on initially restrictive criteria that are gradually expanded to obtain an acceptable number of comparables.

For example, it is common practice to consider financial instruments issued by companies operating in a different industry from the borrower. Moreover, the borrower’s activity sector is already considered once at the stage of determining its credit risk. Logically, considering the risk profile of the borrower when searching for comparable financial transactions therefore already corrects any industry-related rate differences. This reasoning was explicitly validated by the Conseil d’État in the Apex Tool case.

This position is also in line with that of the OECD (OECD Guidelines, section 10.63), which considers that credit rating requires the consideration of qualitative factors, such as the sector of activity in which the company or group of multinational companies operates. The tax authority itself admits of broadening the panel to include companies that do not pursue the same activity as the borrowing company or that belong to different sectors than the borrowing company, in particular in situations where a more restrictive search criterion would lead to a very limited panel (Practical Sheet 8).

The same applies to the amount of the transaction and the size of the issuing company, the tax authorities admitting that the fact (i) that an issue does not concern an amount comparable to that of intragroup financing, and (ii) that an issuer is not of the same size as the borrowing company, does not lead to its exclusion from the panel for this reason alone (Practical Sheet 8).

In addition, it is also common to search for market rates applied to companies with a rating that is not strictly comparable to that of the borrower. In its decision in Apex Tool, the Conseil d’État approved the method of searching for comparables used by the taxpayer, which consisted of searching for financial instruments issued by companies with a credit rating one notch above and below the borrower’s rating.

Nevertheless, it should be borne in mind that broadening the selection criteria for comparables presents a risk that the tax authority and the administrative courts will challenge the comparability of these transactions. For example, in SAS Willink the court rejected the validity of a study because the comparables selected resulted in a panel that was too broad.

  • Possibility of using comparability adjustments

In some situations, the comparables obtained from public databases may not be sufficiently reliable as they stand, requiring some adjustments.

In this regard, the OECD recognizes that comparability adjustments may be necessary to eliminate the material effects of differences in liquidity, maturity, collateral or currency between the controlled intragroup loans and the alternatives (OECD Guidelines, section 10.93).

The tax authority accepts that the borrower company may make adjustments if they are documented and if their cumulative application leads to a result that is reasonably reliable (Practical Sheets 4 and 5). Adjustments that are not sufficiently reliable because they are undocumented or imprecise are not accepted by the authority. In practice, adjustments relating to country risk, currency, or maturity are those most commonly applied.

  • Whether a bond is a realistic alternative to an intragroup loan

Neither the tax authority nor the Conseil d’État have ruled on what is meant by a “realistic alternative to an intragroup loan.” Following the Wheelabrator opinion of the Conseil d’État of July 10, 2019, the Administrative Court of Versailles (TA Versailles, Dec. 6, 2019, no. 1607393 and 1806803, SAS Wheelabrator Group) however considered that the condition of realistic alternative was met in this case, since, in the context of a financial crisis, a bank would have agreed to grant a loan to a company with a high credit risk only on the condition of applying a significant risk premium.

The condition of a realistic alternative should nevertheless be able to be met in situations other than the particular one linked to a financial crisis. This idea is echoed by the rapporteur public in the Wheelabrator opinion, who noted that “economic actors who have access to bond financing constantly choose between this method of financing and bank borrowing and often resort to both at the same time.”

Thus, what is most important to understand is that the realistic alternative test ultimately amounts to asking whether the borrowing company would itself have been able to issue bonds. The contours of this notion nevertheless deserve to be further clarified by case law.

This article does not necessarily reflect the opinion of The Bureau of National Affairs, Inc., the publisher of Bloomberg Law and Bloomberg Tax, or its owners.

Author Information

Mohamed Haj Taieb is a Transfer Pricing Tax Lawyer with CMS France.

The author may be contacted at: