Book Symposium ‘Proxy Warfare’: Arms Supply from Sponsor State to Proxy Armed Group and Applicable Regimes under International Law

About the author(s):

Stuart Casey-Maslen is a Special Advisor to the IHL in Focus Project at the Geneva Academy. He is also a Visiting Professor of International Law at the University of Johannesburg in South Africa.

[French version below]

It appears to be settled at law that when a State has overall control over a proxy armed group that is party to a non-international armed conflict (NIAC), an armed conflict between Sponsor State and Territorial State of an international character (IAC) comes into existence. Despite the misgivings of some international lawyers, including the present author, the position taken by the Appeals Chamber of the International Criminal Tribunal for the former Yugoslavia (ICTY) in 1999 in Tadic (para. 145) was effectively endorsed by a Trial Chamber of the International Criminal Court in its 2012 Lubanga judgment (at para. 541). 

What is less certain is whether, in the coming into being of the IAC, the existing NIAC is extinguished, and whether the Sponsor State is now responsible in international law for all of the internationally wrongful acts of its proxy. It was certainly the view of the ICTY – and particularly Antonio Cassese – that the existing NIAC is fully transformed and that the Sponsor State bears international responsibility for violations of international humanitarian law, a stance that has been adopted by the International Committee of the Red Cross (ICRC). 

In its 2024 Opinion Paper, ‘How is the Term “Armed Conflict” Defined in International Humanitarian Law?’, the ICRC writes (at p. 12) that a non-State armed group (NSAG) that is ‘under the overall control of a state is subordinated to that state as its de facto organ, and members of the group become the equivalent of agents of the state’. This means, of course, that, if this understanding were correct, the acts of the NSAG are acts attributable under international law to the Sponsor State. As the International Law Commission (ILC) stipulated in Article 4(1) of the 2001 Draft articles on Responsibility of States for Internationally Wrongful Acts, ‘The conduct of any State organ shall be considered an act of that State under international law.’

The International Court of Justice, however, has long taken a different view on State responsibility for the acts of proxy armed groups. In its 1986 judgment in the Nicaragua case (at para. 115), the Court declared: ‘United States participation, even if preponderant or decisive, in the financing, organizing, training, supplying and equipping of the contras, the selection of its military or paramilitary targets, and the planning of the whole of its operation, is still insufficient in itself, on the basis of the evidence in the possession of the Court, for the purpose of attributing to the United States the acts committed by the contras in the course of their military or paramilitary operations in Nicaragua.’ Referring directly to – and rejecting – the ICTY’s position in Tadic, in its 2007 judgment in the Bosnian Genocide case (at paras. 399–406), the Court reaffirmed that the higher threshold needed for State responsibility was indeed that of ‘effective control’ of the specific military or paramilitary operations in the course of which violations of international humanitarian law were committed. The Court further held (at para. 420): ‘This is the state of customary international law, as reflected in the ILC Articles on State Responsibility.’

According to Eugénie Duss, the proxy group under overall control must be regarded as belonging to its sponsor within the meaning of Article 4(a)(2) of Chapter III of the Geneva Conventions, and consequently, as forming part of that State’s irregular armed forces. More specifically, Eugénie considers that, even though the criterion of overall control is distinct from that of affiliation, a simple tacit agreement suffices for the purposes of affiliation, and that this agreement exists as long as the State does not refuse to allow the group to fight on its behalf (pp. 92–98 and 137–40). Admittedly, neither the proxy nor the sponsor has openly acknowledged the relationship between them. Nevertheless, she argues that since a party to the conflict cannot deprive individuals of the protections to which they are entitled by mere declarations, a superficial refusal should not be sufficient to preclude the establishment of affiliation, so long as the sponsoring State continues to exercise overall control over the proxy and, in so doing, participates in the planning of the latter’s military operations (p. 100). 

Such reasoning has the advantage of ensuring the coherence of the legal regime, of conforming to the principles underpinning IHL, and also of promoting compliance with it. Nevertheless, in the current author’s view, it should be considered that, in accordance with the general regime of international State responsibility, even if an armed conflict is of an international nature, where the State in question exercises only overall control and not effective control over its proxy, that non-State armed group does not act as an agent of the State and, consequently, its personnel are not composed of members of a militia ‘belonging’ to the State, let alone members of the armed forces. 

This conclusion, if correct, has direct implications for the characterization under international law of any arms supply from Sponsor State to its proxy. Under Article 2(3) of the Arms Trade Treaty (ATT), ‘This Treaty shall not apply to the international transfer by or on behalf of a State Party of conventional arms intended for its own use, provided that such arms remain its property. ’ This is a treaty rule – although it is not yet, in all likelihood, a rule of customary law – and would therefore apply only to each of the 118 States Parties to the ATT. This all means that wherever the NSAG is not a State agent, since the ATT exception could not apply, any transfer of weapons to the group would be a transfer to a non-State actor and covered by all of the treaty’s obligations on export, as set forth in Articles 6 and 7. 

Under Article 6(1), breaching a UN arms embargo to supply an NSAG would also violate the ATT. Under Article 6(3), authorizing a transfer of conventional arms or ammunition in the knowledge they would be used to commit genocide, crimes against humanity, or (certain) war crimes would be a ATT violation. Under Article 7(1), if the transfer to the group would undermine peace and security (how could it not?), it would be unlawful. Otherwise, if none of the above applied, where a clear risk existed that the arms or ammunition would be used to commit or facilitate a serious violation of international humanitarian law, the ATT would again be violated.

The ATT rules operate side by side with rules of general international law on aid or assistance in the commission of an internationally wrongful act. James Crawford’s claim as ILC Special Rapporteur that shared intent to violate international law on the part of the supplier was necessary (his commentary para. 3 on Draft Article 16) was always miscast. It is enough, as in the draft article itself, that the arms supplier State ‘does so with knowledge of the circumstances of the internationally wrongful act’; and that ‘the act would be internationally wrongful if committed by that State’. The export by the Sponsor State to the NSAG would also be covered by the fundamental jus ad bellum rule in Article 2(4) of the UN Charter and customary law that prohibits the threat or use of force. As the International Court of Justice observed in its Nicaragua judgment (at para 195), ‘assistance to rebels in the form of the provision of weapons’ could breach that rule, or it might ‘amount to intervention in the internal or external affairs of other States’. Arguably, the export is better understood as amounting to a threat of force rather than a use of force. It would also –not in the alternative – be an unlawful intervention.

If, on the other hand, the NSAG is under the effective control of the sponsoring State and therefore acts as an agent of that State (see the commentary on paragraph 8 of Article 8 of the Draft Articles on State Responsibility), the supply of arms by that State to the group would not constitute an export, at least where the State is a party to the ATT. This is a conclusion also reached by Eugénie Duss in her work, albeit while extending it to groups that are merely under overall control (pp. 158–59) as ‘lex specialis (pp. 105–06). In any event, this does not mean, however, that such a supply is not governed by international law. Indeed, as noted by the ICRC and also by Eugénie (p. 159), the obligations to respect and ensure respect for Article 1 common to the 1949 Geneva Conventions by the armed forces ‘and other persons or groups whose conduct is attributable to the High Contracting Parties’ are relevant. Beyond the treaty strictures, it is also a customary rule whereby each party to an armed conflict ‘must respect and ensure respect for international humanitarian law by its armed forces and other persons or groups acting in fact on its instructions, or under its direction or control’.

In its 2025 commentary on Geneva Convention IV (at para. 225), the ICRC recalls that contributions to violations of the Conventions may take a number of forms, including ‘by providing aid or assistance, such as by providing weapons’. Thus, for example, any further supply of weapons by Rwanda to M23 in the east of the Democratic Republic of the Congo would breach this general duty to ensure respect for international humanitarian law. In its report submitted to the UN Security Council in July 2025, the Group of Experts on the Democratic Republic of the Congo found (at para. 44) that M23 was an armed group over which Rwanda exercised effective control, and which had engaged in widespread, serious violations of international humanitarian law (see e.g. paras. 93–106, 108, and 171). Rwanda had armed the group, in violation of a UN arms embargo – acts which the Group of Experts called upon Rwanda to cease (at para. 182 (l)). 

Nevertheless, the violations continued. In its report of 30 December 2025, the newly mandated Group of Experts noted that UN human rights monitoring indicated that, between April and October 2025, summary executions were the most frequently documented violation in the Democratic Republic of the Congo, with M23 listed as the main perpetrator, responsible for approximately 45 per cent of all summary executions reported across the country. The number of victims of abduction, including for forced recruitment, and ill-treatment by AFC/M23 surpassed the number of victims killed by the same group (para. 80).

In sum, international law is all-encompassing of the supply of weapons to a proxy, even if there remain differences of opinion with the clear position taken by the International Court of Justice on issues of State responsibility. Sometimes the provision of weapons from Sponsor State to proxy will be a ‘transfer’ as that term is understood in disarmament law; on other occasions it will be supply. The movement of arms will not always be prohibited. But it will always be regulated.

Il semble établi en droit que lorsqu’un État exerce un contrôle global sur un groupe armé « proxy » partie à un conflit armé non international (CANI), un conflit armé de caractère international (CAI) entre l’État « sponsor » et l’État du territoire prend naissance. Malgré les réticences de certains juristes internationaux, dont l’auteur du présent article, la position adoptée par la Chambre d’appel du Tribunal pénal international pour l’ex-Yougoslavie (TPIY) en 1999 dans l’affaire Tadic (paragraphe 145) a été effectivement entérinée par une Chambre de première instance de la Cour pénale internationale dans son arrêt Lubanga de 2012 (paragraphe 541). 

Ce qui est moins certain, c’est de savoir si, avec la naissance du CAI, le conflit armé non international existant prend fin, et si l’État sponsor est désormais responsable, en droit international, de tous les actes internationalement illicites commis par son proxy. C’était certainement l’avis du TPIY – et en particulier d’Antonio Cassese – que le CANI existant est entièrement transformé et que l’État sponsor porte la responsabilité internationale des violations du droit international humanitaire, une position qui a été depuis adoptée par le Comité international de la Croix-Rouge (CICR). 

Dans son document d’opinion de 2024 intitulé « Comment le terme « conflit armé » est-il défini en droit international humanitaire ?», le CICR écrit (à la page 12) qu’un groupe armé non étatique (GANE) qui est « Un GANE placé sous le contrôle global d’un État est subordonné à cet État comme son organe de facto, et les membres du groupe peuvent être assimilés, en pareil cas, à des agents de cet État ». Cela signifie bien entendu que, si cette interprétation était correcte, les actes du GANE seraient, en droit international, imputables à l’État qui le soutient. Comme l’a stipulé la Commission du droit international (CDI) à l’article 4(1) du Projet d’articlesde 2001 sur la responsabilité des États pour fait internationalement illicite, « Le comportement de tout organe d’un État est considéré comme un fait de l’État d’après le droit international. »

La Cour internationale de Justice a toutefois longtemps adopté une position très différente sur la responsabilité de l’État pour les actes des groupes armés « proxy ». Dans son arrêt de 1986 dans l’affaire Nicaragua (paragraphe 115), la Cour a déclaré que : « même prépondérante ou décisive, la participation des Etats-Unis à l’organisation, à la formation, à l’équipement, au financement et à l’approvisionnement des contras, à la sélection de leurs objectifs militaires ou paramilitaires et à la planification de toutes leurs opérations demeure insuffisante en elle- même, d’après les informations dont la Cour dispose, pour que puissent être attribués aux Etats-Unis les actes commis par les contras au cours de leurs opérations militaires ou paramilitaires au Nicaragua. » Se référant directement à la position du TPIY dans l’affaire Tadic – et la rejetant –, la Cour a réaffirmé, dans son arrêt de 2007 rendu dans l’affaire du génocide en Bosnie (paragraphes 399 à 406), que le seuil plus élevé requis pour engager la responsabilité de l’État était bien celui du « contrôle effectif » des opérations militaires ou paramilitaires au cours desquelles les violations en question se seraient produites. La Cour a en outre estimé (au paragraphe 420) que sa position existe : « en un sens qui ne diffère pas de manière sensible [sic] de celui que possèdent ces notions dans le droit général de la responsabilité internationale. »

Selon Eugénie Duss, le groupe proxy placé sous contrôle global doit être considéré comme appartenant à son sponsor au sens de l’art. 4 (a) (2) CG III et, partant, comme composant les forces armées irrégulières de cet État. Plus précisément, Eugénie estime que, quand bien même le critère du contrôle global est distinct de celui de l’appartenance, il convient de se ranger à l’avis de ceux pour lesquels un simple accord tacite suffit aux fins de l’appartenance, et que cet accord existe aussi longtemps que l’État ne refuse pas que le groupe combatte en son nom (p. 92-98 et 137-40). Certes, aucun proxy ni sponsor n’ont reconnu au grand jour la relation qui les lie. Néanmoins, puisqu’une partie au conflit ne saurait priver des individus des protections auxquelles ils ont droit par de simples déclarations, un refus de façade ne devrait pas être à même de déjouer l’établissement de l’appartenance, aussi longtemps que l’État sponsor continue de contrôler globalement le proxy et, ce faisant, de participer à l’élaboration des opérations militaires de ce dernier (p. 100). Un tel raisonnement présente l’avantage d’assurer la cohérence du régime juridique, de se conformer aux logiques sous-tendant le DIH mais aussi d’en favoriser le respect. Il reste que, de l’avis de l’auteur de cette publication, il convient toutefois de considérer que, conformément au régime général de la responsabilité international des États, même si un conflit armé revêt un caractère international, lorsque l’État de référence n’exerce qu’un contrôle global et non un contrôle effectif sur son proxy, ce groupe armé non étatique n’agit pas en tant qu’agent de l’État et que, par conséquent, son personnel n’est pas composé de membres d’une milice « appartenant » à l’État, et encore moins de membres des forces armées. Considérer que le groupe sous contrôle global équivaut ou non aux forces armées de l’État qui le contrôle a des implications directes sur la qualification, au regard du droit international, de tout transfert d’armes entre l’État de référence et son proxy. En vertu de l’article 2, paragraphe 3, du Traité sur le commerce des armes (TCA), « Le présent Traité ne s’applique pas au transport international par tout État Partie ou pour son compte d’armes classiques destinées à son usage, pour autant que ces armes restent sa propriété. » Il s’agit d’une règle conventionnelle – bien qu’elle ne soit pas encore, selon toute vraisemblance, une règle de droit coutumier – et elle ne s’appliquerait donc qu’à chacun des 118 États parties au TCA. 

Tout cela signifie que, dès lors que le groupe armé non étatique (GANE) n’est pas un agent de l’État, l’exception prévue par le TCA ne pouvant s’appliquer, tout transfert d’armes à ce groupe constituerait un transfert à un acteur non étatique et serait soumis à l’ensemble des obligations du traité en matière d’exportation, telles qu’énoncées aux articles 6 et 7. En vertu de l’article 6 (1), le non-respect d’un embargo sur les armes décrété par l’ONU pour approvisionner un GANE constituerait également une violation du TCA. En vertu de l’article 6 (3), autoriser un transfert d’armes conventionnelles ou de munitions en sachant qu’elles seraient utilisées pour commettre un génocide, des crimes contre l’humanité ou (certains) crimes de guerre constituerait une violation du TCA. En vertu de l’article 7 (1), si le transfert au groupe portait atteinte à la paix et à la sécurité (comment pourrait-il en être autrement ?), il serait illégal. Dans le cas contraire, si aucun des cas susmentionnés ne s’appliquait, mais qu’il existait un risque manifeste que les armes ou les munitions soient utilisées pour commettre ou faciliter une violation grave du droit international humanitaire, le TCA serait à nouveau violé.

Les règles du TCA s’appliquent parallèlement aux règles du droit international général relatives à l’aide ou à l’assistance dans la commission d’un fait internationalement illicite. L’affirmation de James Crawford, en sa qualité de rapporteur spécial de la CDI, selon laquelle il fallait une intention commune de violer le droit international de la part du fournisseur (son commentaire, paragraphe 3, sur le projet d’article 16) a toujours été erronée. Il suffit, comme le prévoit le projet d’article sur la responsabilité internationale des États lui-même, que l’État fournisseur d’armes « agisse en connaissance des circonstances de l’acte internationalement illicite » ; et que « l’acte serait internationalement illicite s’il était commis par cet État ». L’exportation par l’État sponsor vers le groupe armé non étatique serait également couverte par la règle fondamentale du jus ad bellum énoncée à l’article 2, paragraphe 4, de la Charte des Nations Unies et par le droit coutumier qui interdit la menace ou l’emploi de la force. Comme l’a observé la Cour internationale de Justice dans son arrêt Nicaragua (paragraphe 195), « l’assistance aux rebelles sous forme de fourniture d’armes » pourrait enfreindre cette règle, ou pourrait être « l’équivalent d’une intervention dans les affaires intérieures ou extérieures d’autres Etats. ». On peut soutenir que l’exportation doit être mieux comprise comme constituant une menace de recours à la force plutôt qu’un recours à la force. Elle constituerait également – et non à titre alternatif – une intervention illicite.

Si, en revanche, le GANE est sous le contrôle effectif de l’État sponsor et constitue donc un agent de l’État par rapport à ses actes (voir commentaire paragraphe 8 sur l’article 8 du Projet des articles relatives à la responsabilité de l’État), la fourniture d’armes par cet État au groupe ne constituerait pas une exportation, du moins lorsque l’État est partie au TCA. Une conclusion à laquelle arrive aussi Eugénie Duss dans son ouvrage, tout en l’étendant aux groupes qui ne seraient que contrôlés globalement (p. 158-59) à titre de lex specialis (p. 105-06). Dans tous les cas, cela ne signifie toutefois pas que cette fourniture n’est pas régie par le droit international. En effet, comme l’a dûment noté le CICR mais aussi Eugénie Duss (p. 159), les obligations de respecter et de faire respecter l’article 1 commun aux Conventions de Genève de 1949 par les forces armées « et les autres personnes ou groupes dont le comportement est imputable aux Hautes Parties contractantes » sont pertinentes. Au-delà des dispositions du traité, il s’agit également d’une règle coutumière selon laquelle chaque partie à un conflit armé « doit respecter et faire respecter le droit international humanitaire par ses forces armées ainsi que par les autres personnes ou groupes agissant en fait sur ses instructions ou ses directives ou sous son contrôle. ».

Dans son commentaire de 2025 sur la quatrième Convention de Genève (paragraphe 225), le CICR rappelle que la contribution à des violations des Conventions peut prendre diverses formes, notamment en fournissant une aide ou une assistance, par exemple en fournissant des armes. Ainsi, par exemple, tout nouvel approvisionnement en armes du Rwanda au M23 dans l’est de la République démocratique du Congo constituerait une violation de cette obligation générale de veiller au respect du droit international humanitaire. Dans son rapport soumis au Conseil de sécurité des Nations unies en juillet 2025, le Groupe d’experts sur la République démocratique du Congo a conclu (au paragraphe 44) que le M23 était un groupe armé sur lequel le Rwanda exerçait un contrôle effectif et qui s’était livré à des violations graves et généralisées du droit international humanitaire (voir par exemple les paragraphes 93 à 106, 108 et 171). Le Rwanda avait armé ce groupe, en violation d’un embargo sur les armes décrété par l’ONU – des actes que le Groupe d’experts a demandé au Rwanda de cesser (paragraphe 182 (l)). 

Néanmoins, les violations se sont poursuivies. Dans son rapport du 30 décembre 2025, le Groupe d’experts nouvellement mandaté a noté que la surveillance des droits de l’homme par l’ONU indiquait qu’entre avril et octobre 2025, les exécutions sommaires constituaient la violation la plus fréquemment documentée en République démocratique du Congo, le M23 étant désigné comme le principal auteur, responsable d’environ 45 % de toutes les exécutions sommaires signalées à travers le pays. Le nombre de victimes d’enlèvements, y compris à des fins de recrutement forcé, et de mauvais traitements infligés par l’AFC/M23 a dépassé le nombre de victimes tuées par ce même groupe (paragraphe 80).

En résumé, le droit international couvre l’ensemble des livraisons d’armes à un proxy, même si des divergences d’opinions sur la responsabilité des États subsistent. Parfois, la fourniture d’armes par un État sponsor à un proxy constituera un « transfert » au sens où ce terme est compris dans le droit du désarmement ; dans d’autres cas, il s’agira d’une livraison. Le mouvement d’armes ne sera pas toujours interdit. Mais il sera toujours réglementé.

(Visited 10 times, 10 visits today)

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: