{"id":2114534,"date":"2026-09-01T01:16:01","date_gmt":"2026-08-31T23:16:01","guid":{"rendered":"https:\/\/kohenavocats.fr\/2026\/09\/01\/french-research-tax-credit-cir-foreign-owned-subsidiary-eligible-costs-proof\/"},"modified":"2026-09-01T01:16:01","modified_gmt":"2026-08-31T23:16:01","slug":"french-research-tax-credit-cir-foreign-owned-subsidiary-eligible-costs-proof","status":"publish","type":"post","link":"https:\/\/kohenavocats.fr\/en\/2026\/09\/01\/french-research-tax-credit-cir-foreign-owned-subsidiary-eligible-costs-proof\/","title":{"rendered":"French Research Tax Credit (CIR) for a Foreign-Owned French Subsidiary: Eligible Costs, Subcontracting and Proof"},"content":{"rendered":"

A foreign-owned French subsidiary can, in principle, claim the French Research Tax Credit (CIR, cr\u00e9dit d’imp\u00f4t recherche<\/em>) in the same way as another French company. The nationality of its shareholders is not the test. The decisive questions are whether the French company is within the statutory tax regime, whether it carries out qualifying research and development (R&D) work, whether the claimed costs are connected to that work, and whether the file can prove each step when the French tax administration or the Ministry of Higher Education and Research reviews it.<\/p>\n

This distinction matters in an international group. A parent may pay researchers abroad, recharge a technology team, or appoint a university or engineering company. None of those facts automatically makes the resulting invoice an eligible CIR expense for the French subsidiary. The subsidiary needs a clear map between the technical project, the entity that performs each task, the location of the work, the legal status of any research provider, the accounting treatment and the annual tax declaration. A foreign parent\u2019s invoice is evidence of a payment, not evidence that the French statutory conditions have been met.<\/p>\n

The current rules also require attention to timing. The credit is calculated by reference to the calendar year in which qualifying expenses are incurred, even where the company closes its accounts on another date. This article explains how a foreign founder or group finance team can qualify a French subsidiary\u2019s own research, assess foreign or related-party subcontracting, prepare the technical and financial evidence, and use a rescript procedure when the eligibility question is genuinely uncertain.<\/p>\n

I. Can a foreign-owned French subsidiary claim the CIR and which costs qualify?<\/h2>\n

A. Does foreign ownership prevent a French subsidiary from claiming the CIR?<\/h3>\n

Foreign ownership is not, by itself, an exclusion. Article 244 quater B of the French General Tax Code (CGI, Code g\u00e9n\u00e9ral des imp\u00f4ts<\/em>) identifies the companies that may benefit by reference to their activity, taxation and qualifying research expenses. The current wording covers industrial, commercial and agricultural businesses taxed on their actual profit, together with certain exempt businesses. It does not make French ownership of the share capital a condition. A French soci\u00e9t\u00e9 par actions simplifi\u00e9e<\/em> (SAS, a flexible French joint-stock company) owned by a United States, British, Canadian, Swiss or other foreign parent can therefore enter the analysis if its own situation meets the rules.<\/p>\n

The statutory starting point is the current version of Article 244 quater B of the CGI<\/a>. For ordinary research expenditure, the rate is 30% on the portion up to \u20ac100 million and 5% above that threshold. The same article contains other, separate categories and rates, including a distinct category for certain small and medium-sized enterprise prototype work. A calculation should not apply the 20% rate for that separate category to ordinary research without first checking the exact statutory conditions and the relevant accounting year. The figure that appears in a group budget is not a substitute for classifying each project.<\/p>\n

The company must then show that the underlying work is research rather than ordinary commercial development. The legal definitions in Article 49 septies F of Annex III to the CGI<\/a> distinguish fundamental research, applied research and experimental development. In practical terms, the project should address a genuine scientific or technical obstacle. A product roadmap, routine configuration, standard integration, ordinary maintenance, cosmetic improvement or a project that simply applies known methods is not converted into qualifying R&D by calling it \u201cinnovation\u201d. The company should be able to identify the state of knowledge at the start, the uncertainty that remained, the hypotheses considered, the technical method used and the results, including unsuccessful tests.<\/p>\n

For a foreign-owned subsidiary, the place where the work is performed is important. Article 244 quater B requires the expenses, apart from the particular intellectual-property exceptions in the text, to correspond to operations located in the European Union or in another European Economic Area state that has an administrative-assistance convention with France. The same provision also requires the expense to be retained for the ordinary determination of taxable profit. A French company cannot simply claim every research invoice paid by the group around the world. Work performed in the French subsidiary\u2019s laboratory, office or technical environment is easier to map; work directed and performed by the foreign parent needs its own analysis of the contracting entity, the provider, the location and the statutory category.<\/p>\n

There is a second distinction between ownership and economic control. A parent may own 100% of the French subsidiary and still provide a real service. But common ownership creates a related-party context in which the price, allocation and benefit require careful support. The transfer-pricing principle in Article 57 of the CGI<\/a> can become relevant when a French company accepts terms that differ from those that independent businesses would have negotiated. The subsidiary\u2019s CIR file and its transfer-pricing file should tell the same story: who performed the work, who bore the technical risk, what the French company received and how the amount was calculated.<\/p>\n

That is why a French subsidiary should separate three situations in its project register:<\/p>\n

    \n
  1. research carried out by the French company\u2019s own employees or people placed at its disposal and working with its means;<\/li>\n
  2. research commissioned from a qualifying and approved external research body; and<\/li>\n
  3. ordinary group support, licence payments, management services or shareholder activities that may be deductible or taxable under other rules but do not automatically form part of the CIR base.<\/li>\n<\/ol>\n

    The third category is often where an international group makes its first mistake. A French subsidiary may receive a monthly \u201cR&D management fee\u201d from its parent. That label does not prove that the parent carried out a qualifying research operation on behalf of the subsidiary, that the parent is an approved research provider, or that the invoice falls within the statutory subcontracting limits. The invoice may need to be analysed as an ordinary intra-group service, a royalty, a cost contribution or a research subcontract. Each classification has different consequences.<\/p>\n

    The French company\u2019s legal form and registration remain relevant to the wider file even though they do not create the CIR right by themselves. Keep the current Kbis<\/em>, meaning the official extract showing the company\u2019s registration and legal details, and the information filed through the INPI Guichet unique<\/a>. INPI is the French National Institute of Industrial Property. The company\u2019s registered office, representative, activity and ownership chain should be consistent in the research contract, invoices, payroll records, board documents and tax returns. An inconsistent legal identity does not automatically destroy a claim, but it creates avoidable questions about who actually incurred and controlled the expense.<\/p>\n

    Foreign tax residence also needs to be kept separate from French eligibility. The parent may be resident in a country with its own R&D incentive, and the group may need to examine that country\u2019s rules on the same project. A French CIR analysis does not guarantee that the parent can claim a second benefit, and a foreign incentive does not enlarge the French base. The group should record which entity owns the technical results, which entity bears the expenditure, and whether a separate national rule restricts double incentives or requires an adjustment.<\/p>\n

    The most useful decision test is therefore not \u201cIs the company foreign-owned?\u201d It is \u201cCan the French company prove a qualifying operation and a qualifying expense under the rule applicable to the relevant year?\u201d If the answer is yes, foreign ownership is normally a background fact. If the answer is no, changing the invoice description or moving the charge through another group entity will not cure the weakness.<\/p>\n

    B. Which internal people, equipment and operating costs are counted?<\/h3>\n

    Article 244 quater B lists several categories. For internal work, the central category is the personnel cost of researchers and research technicians who are directly and exclusively assigned to qualifying operations. The French company should identify the people, their technical role, the projects on which they worked, the periods concerned and the proportion of their time that can be supported. A job title such as \u201cengineer\u201d or \u201cdata scientist\u201d is not enough. The file should explain the technical work and show how the payroll amount is allocated.<\/p>\n

    Time records do not need to be artificial if the company has a reliable project system, but they must be sufficiently detailed to reconstruct the allocation. A monthly record can link a person to a project code, task, experiment, test, review or deliverable. When an employee works on both research and commercial implementation, the company should split the time on a defensible basis. Claiming 100% for every technical employee in a young subsidiary is especially risky when the same people also sell, install, support customers or manage ordinary production.<\/p>\n

    The rule also treats other operating costs by a statutory flat-rate method. The current Article 244 quater B fixes those amounts at 75% of the qualifying depreciation allocations and 40% of qualifying personnel expenses. The company should not replace those percentages with a second estimate of actual overheads for the same category. It should first determine which personnel and depreciation amounts are eligible, then apply the legal mechanism and retain the calculation that connects the result to the project list.<\/p>\n

    Depreciation can qualify when it concerns new fixed assets created or acquired in new condition and directly assigned to scientific or technical research operations, including prototypes or pilot installations. A machine bought for general production does not become a CIR asset merely because a researcher uses it occasionally. Keep the purchase record, commissioning date, asset register, depreciation schedule, technical allocation and evidence of actual use. Where the asset is used for several purposes, document the allocation rather than treating the entire annual depreciation as research expenditure.<\/p>\n

    The Conseil d\u2019\u00c9tat addressed the prototype point in a recent decision. In Conseil d\u2019\u00c9tat, 17 June 2026, no. 507371<\/a>, concerning Green Big, the court held that research depreciation could qualify \u201csans qu’ait d’incidence la circonstance que ces immobilisations rev\u00eatiraient le caract\u00e8re de prototypes.\u201d The exact wording means that the fact an asset is a prototype does not, by itself, defeat the depreciation claim. It does not remove the other requirements: the asset must be new, directly assigned to qualifying research, properly depreciated and connected to an operation that meets the scientific or technical test. The case is useful because it prevents one over-simple rejection, not because it turns every prototype into eligible expenditure.<\/p>\n

    A person temporarily made available to the French company can also require a precise analysis. In Conseil d\u2019\u00c9tat, 25 January 2017, no. 390652<\/a>, Intuigo, the court explained that eligible personnel expenses were not limited to remuneration and social charges for people directly employed by the company. The court stated that the rule could extend to \u201cCes dispositions ne limitent pas les d\u00e9penses de personnel susceptibles d’ouvrir droit au cr\u00e9dit d’imp\u00f4t aux seules r\u00e9mun\u00e9rations et charges sociales vers\u00e9es pour des personnes employ\u00e9es par l’entreprise et affect\u00e9es \u00e0 des op\u00e9rations de recherche susceptibles d’ouvrir droit \u00e0 ce cr\u00e9dit, mais s’\u00e9tendent aux r\u00e9mun\u00e9rations et aux charges sociales prises en charge par l’entreprise au titre de la mise \u00e0 sa disposition par un tiers de personnes afin d’y effectuer dans ses locaux et avec ses moyens des op\u00e9rations de recherche.\u201d This is an exact extract from the decision. The operational safeguards are the person\u2019s actual work in the French company\u2019s operations, use of the company\u2019s premises and means, an identifiable cost borne by the French company, and proof that the work is qualifying.<\/p>\n

    Do not apply older personnel incentives without checking the period. The conditions applicable to a historical double-counting rule for employees holding a doctorate have changed over time. For example, Conseil d\u2019\u00c9tat, 31 May 2024, no. 476354<\/a>, Awalee Consulting, held in the circumstances of that case that the double amount applied only from the date the doctorate was obtained when that date followed recruitment. The exact temporal version of Article 244 quater B and the declaration year must be checked before any such adjustment is entered. A foreign group should never copy a calculation from an earlier French tax year into a current return without reviewing the amended text.<\/p>\n

    As an illustration, assume a French subsidiary has \u20ac200,000 of supported qualifying research payroll and \u20ac80,000 of qualifying depreciation during a calendar year. The statutory operating-cost amounts would be \u20ac80,000 for the personnel component and \u20ac60,000 for the depreciation component, before other adjustments. The illustrative base would therefore be \u20ac420,000 and the ordinary 30% rate would produce a theoretical \u20ac126,000 credit. This is not a filing result: the company would still need to exclude non-qualifying work, verify the common-law deductibility and geographic conditions, account for any grant or other adjustment, and apply the correct rule for subcontracting. The example shows why a transparent bridge from payroll and asset records to the return matters.<\/p>\n

    Research work can also create intellectual-property and standardisation costs, but those categories have specific statutory conditions. They should be mapped separately instead of added to a general \u201cinnovation\u201d line. A project register should identify the relevant provision, the technical operation, the accounting account and the supporting evidence for each category. This approach is especially useful when a foreign parent reports the project in a different accounting vocabulary from the French subsidiary.<\/p>\n

    II. How should subcontracting, group charges and proof be handled?<\/h2>\n

    A. Can a French subsidiary subcontract research to its foreign parent or another group company?<\/h3>\n

    Subcontracting is possible, but an international group must distinguish an approved research provider from an ordinary related company. Article 244 quater B, II, d bis covers expenses for operations of the same nature entrusted to organisations approved by the French minister responsible for research, or to scientific or technical experts approved under the relevant rules. For providers established in a European Union state or in another European Economic Area state with the required administrative-assistance convention, the article describes routes involving French approval or an equivalent approval by the competent authority of the provider\u2019s country when a similar system exists.<\/p>\n

    The approval framework is set out in Article 49 septies H of Annex III to the CGI<\/a>. The provider should be able to show the research activity and the documents required for approval. The official Service-Public CIROCO information<\/a> explains the practical purpose of the approval: a provider carrying out research for clients needs the relevant approval for the client to treat the outsourced operation as a CIR expense. Service-Public is the French public-administration information service; CIROCO is the official tool used for information about approved research providers. Check the provider\u2019s status for the period of the work and keep the approval evidence with the contract.<\/p>\n

    A foreign parent is not automatically an approved research organisation. It may have excellent laboratories, employ qualified researchers and own valuable patents while still failing the specific approval condition for the French subsidiary\u2019s outsourced CIR expense. The parent may invoice the French company for a genuine intra-group research service under ordinary tax and transfer-pricing rules, but the French company should not place that invoice in the CIR base unless the subcontracting rule and provider-status requirements have been checked. The analysis is more difficult where the parent is outside the European Union or European Economic Area because the location and approval wording in Article 244 quater B must be applied to the precise arrangement.<\/p>\n

    The statutory caps also matter. Article 244 quater B limits approved-provider expenditure to three times the total of the other qualifying research expenses before the additional limits are applied. The expenses enter the base within a global limit of \u20ac2 million per year. That limit is increased to \u20ac10 million where the outsourced research is entrusted to approved bodies and there is no relationship of dependence, within the meaning of the tax code, between the beneficiary company and those bodies. A group company should not assume that the \u20ac10 million ceiling applies simply because the provider is incorporated in another country. Related-party status must be analysed, and the lower limit may be relevant.<\/p>\n

    The operations must also be performed directly by the approved body to which they were entrusted, subject to the statutory exception allowing that body to use another approved body for certain necessary work. The contract should identify the operation, the scientific or technical question, the deliverables, the period, the location, the personnel or facilities involved and any permitted onward subcontracting. An invoice that only says \u201cR&D services\u201d will not explain whether the provider performed the work, passed it on, or delivered ordinary implementation.<\/p>\n

    The Conseil d\u2019\u00c9tat confirmed the anti-double-counting logic in Conseil d\u2019\u00c9tat, 18 June 2021, no. 433319<\/a>, Sopra Steria. The court held that sums received by approved private research organisations for qualifying operations entrusted by beneficiary companies \u201cconstituent, pour ces entreprises donneuses d’ordre, des d\u00e9penses \u00e9ligibles \u00e0 ce cr\u00e9dit.\u201d The same decision explains that a subcontracting research organisation cannot include the expenses it incurred to carry out those operations in its own CIR base. In other words, the structure is not designed to create two credits for the same research operation. The French subsidiary should obtain enough information to identify the provider\u2019s role and avoid counting the same payroll, equipment or external invoice in two places.<\/p>\n

    A parent-company cost pool needs an additional layer of analysis. If the French subsidiary participates in a group research programme, the agreement should state whether the subsidiary is a buyer of a defined research operation, a contributor to a cost-sharing arrangement, a licensee of results, or simply receiving central management. These are not interchangeable descriptions. The file should identify the research question that benefits the French business, the work assigned to each entity, the allocation key and the ownership or use of results. A group-wide statement that \u201call entities support innovation\u201d is too general for a French tax file.<\/p>\n

    Transfer pricing and CIR eligibility should be reconciled, but they are not the same test. A charge can be arm\u2019s length and still not be an eligible research expense. Conversely, an operation can be technically eligible in principle while the price or allocation is challenged as excessive. Keep the intercompany agreement, functional analysis, cost pool, allocation calculations, deliverables and payment trail together. If the parent receives a cost-plus remuneration, identify the underlying qualifying expense and the mark-up separately; the statutory CIR base does not automatically follow the total invoice amount.<\/p>\n

    Grants, public funding and other support can also affect the amount retained. The group should record the source, beneficiary, project and period of every subsidy or funding stream. The French tax computation should then follow the rule applicable to that funding rather than assuming that a gross invoice is the eligible base. The official BOFiP guidance on research expenditure<\/a> is a useful administrative reference alongside the statute, but the legal classification must still be made on the facts of the project.<\/p>\n

    B. What documents and tax filings prove the claim?<\/h3>\n

    The proof file should be assembled while the project is running. The French tax administration can review the tax calculation, while research authorities can assess the reality and scientific or technical nature of the operations. Article L45 B of the French Tax Procedure Code (LPF, Livre des proc\u00e9dures fiscales<\/em>)<\/a> permits agents of the research ministry to verify the reality of the research allocation, while the tax administration remains responsible for the tax reassessment. Article R45 B-1 of the LPF<\/a> sets out the role of agents appointed by the research ministry and the regional academic delegation for research and innovation (DRARI). The ministry\u2019s technical-control process and review period should be anticipated when the file is assembled.<\/p>\n

    A serious file should allow a reviewer who was not present at the start of the project to understand what was unknown, what was attempted and what was learned. For every project, retain a short technical narrative covering:<\/p>\n