Assessing Assurances that Exported Arms Will Not Be Misused

[William Worster has taught public international law, the law of international organizations and international migration and refugee law at The Hague University of Applied Sciences for more than fifteen years]

By now, many states have concluded that there is a risk that any arms they export to Israel may be misused to violate international humanitarian or human rights law. This predicted risk then triggers the obligation to block export. Because the prohibition on export involves the prediction of a risk of misuse, the US has requested assurances from Israel that the arms will not be misused and is currently assessing those assurances. Other states may follow. This post anticipates this expanding practice, reviews the law on assessing exports, and draws on the practice of expulsion assurances, to recommend an approach to assessing export assurances when there is a risk of misuse.

Several international obligations require states to assess risk of arms misuse. On Feb. 12, 2024, the Court of Appeals of The Hague, in the Oxfam Novib v Netherlands case, determined that The Netherlands cannot export F-35 parts from the storage facility in the country to Israel. While the case was based on Dutch export regulations, those regulations implemented Dutch obligations under international law. The Court observed that The Netherlands is bound by three relevant international legal obligations: the Arms Trade Treaty (“ATT”), the EU Common Position, and the Geneva Conventions, and did not address the Genocide Convention. The case is currently on appeal. But of course, the Dutch case is not an outlier. The Office of the UN High Commissioner for Human Rights posted a press release, signed by a number of human rights Special Rapporteurs and Independent Experts cautioning countries against the “transfer of [any] weapons or ammunition to Israel” as this could violate international humanitarian law if these weapons are used contrary to the Geneva Conventions, referring to Israel’s alleged failure to adhere to international law in Gaza. Exports of arms have been blocked by Belgium and Italy, and, just a few days ago, the Canadian Government decided against further arms exports to Israel. In addition, Nicaragua brought a case against Germany to the ICJ for its arms exports to Israel, among other claims, arguing that the exports potentially violate the obligation in the Genocide Convention to prevent genocide. The need to assess export compliance will most likely only grow more acute.

However, there are techniques for states to reduce the risk of misuse and, in so doing, potentially avoid arms exports restrictions. One method of reducing the risk that the weapons will be used for violations of international law is to request assurances of use. This approach is already underway. On Feb. 8, 2024, the Biden Administration issued National Security Memorandum 20 (“NSM-20”) that provides that the recipient state may be asked to provide “credible and reliable written assurances” that the state will not use the weapons in any manner contrary to international law. These assurances will not necessarily suffice to convince the US to permit the export. NSM-20 also adds that, should the US doubt the assurances, the US will consider “appropriate next steps to be taken to assess and remediate the situation.” Barak Ravid at Axois reported on March 14 that Israel provided those assurances. That letter, however, does not appear to be publicly available. On Mar. 25, the US State Department commented that, while it believed that the assurances were effective, it had not yet made a firm decision on that point. Secretary Blinken will need to report a definitive decision to Congress on the assurances and the possibility that any US arms transfers might be misused by May 8. Depending on the US’ experience with assurances, other states may wish to adopt this practice as well. In the pending Dutch Oxfam F-35 case, for example, it is unclear whether Israel has given The Netherlands any assurances.

The question is then, does international law provide any standards for assessing assurances that arms will not be used to commit or facilitate serious violations of international humanitarian or human rights law? This assessment is not a completely free exercise of political discretion, despite some governments appearing to act as if it were. The ATT, EU Common Position, Geneva Conventions and Genocide Convention provide some guidance and context for states to assess assurances, to the degree to which individual states are bound by these instruments. In addition, international practice assessing assurances in general also provides guidance on relevant factors that must be considered.

The ATT permits states to undertake mitigating measures that reduce risk. The ATT also encourages states to consider any information that the importing state might provide. When they do, the assessment is whether the risk is still overriding in light of the mitigating measures. A state party to the ATT could consider factors such as “the nature, type, and quantity of weapons to be exported; their normal and reasonably foreseeable uses; the general situation in the state of final destination and its surrounding region; the intended end user, including its record of compliance with international humanitarian law and international human rights law; actors involved in the export; and the intended route of the export.”

There is some guidance as well in the EU Common Position. Before exporting to a situation of risk of violations of international humanitarian and human rights law, states must “assess [] the recipient country’s attitude towards relevant principles established by international human rights instruments,” consider “the nature of the technology or equipment,” and consider “the record of the buyer country with regard to … its compliance … with international humanitarian law.” 

The Geneva Conventions do not expressly provide for any factors for assessing assurances, though the ICRC’s commentary does provide some advice. In the most recent commentary on Common Article 1 of the Gevena Conventions, the ICRC does not see any obligation to take any specific measures in particular to ensure respect for the convention. But in crafting its case-by-case approach, the state must consider “the gravity of the breach, the means reasonably available to the State, and the degree of influence it exercises over those responsible for the breach.” 

The Genocide Convention is also silent as to factors for assessing assurances, but, in the Bosnia v Serbia Genocide case, the ICJ has provided some insight into the factors for assessing compliance with the obligation to prevent. Mettraux has distilled the factors from that judgment as “(i) the geographical distance of the State concerned from the scene of the events; (ii) the strength of the political links with the relevant entity; (iii) links of other kinds between the authorities of that State and the main actors in the events; (iv) the particular legal position of that state vis-à-vis the situations and persons facing the danger of genocide; (v) the military, intelligence, and logistical support provided by the state to that other actors; (vi) the alignment and overlap of their political and military agendas; (vii) the strength of the political, military, and financial links between the two states; and (viii) other factors that might reflect a capacity to restrain those intent on committing acts of genocide.”

In addition to these factors, states can also borrow from their experience generally with making predictions from assurances, especially their extensive experience with expulsion assurances. States give each other assurances on a wide variety of topics for a wide variety of purposes, but they all have the same function: trying to convince other states that they will act as they say they will. Unlike most areas where assurances are given, expulsion has been subject to extensive litigation, so international courts and tribunals have developed a framework for their assessment. States often give assurances when expelling, deporting, or extraditing a person to another state where there is a risk that the person will be executed, tortured, severely mistreated, or suffer some other grave violation of human rights. This is not to say that states have an excellent track record with expulsions assurances. Everyone knows that states have mistreated expelled people despite assurances. Assurances cannot be a perfect fix. They are, after all, only part of an attempt at an intelligent prediction of risk. But what states can learn from expulsion assurances is that mere pieces of paper do not necessarily eliminate risk. Practice with expulsion assurances can help to identify important variables when making a prediction whether a state will comply with its promises. 

In assessing whether assurances issued in expulsion cases were likely to reduce the risk to the person, the European Court of Human Rights, most famously in the Othman v UK case, considers eleven factors based on its jurisprudence: whether the terms of the assurances have been disclosed to the Court; whether the assurances are specific or are general and vague; who has given the assurances and whether that person can bind the receiving State; whether local authorities can be expected to abide by them; whether the assurances concerns treatment which is legal or illegal in the receiving State; whether they have been given by a [ECHR] Contracting State; the length and strength of bilateral relations between the sending and receiving States, including the receiving State’s record in abiding by similar assurances; whether compliance with the assurances can be objectively verified; whether there is an effective system of protection against torture in the receiving State… and whether it is willing to investigate allegations of torture and to punish those responsible; whether the applicant has previously been ill-treated in the receiving State; whether the reliability of the assurances has been examined by the domestic courts of the sending/Contracting State. These Othman factors have been largely adopted by other international courts, tribunals, and treaty bodies, as well as states, including states not party to the ECHR.

Many of these Othman factors for predicting risk are relevant for predicting the risk of misuse of arms. Some of them already closely resemble similar factors for assessing the risk of arms misuse. Borrowing the Othman factors introduces additional factors for determining the likelihood of compliance with international commitments beyond the minimal requirements in the various treaties mentioned above. 

Therefore, when implementing their international, European or domestic export arms obligations, states must make a prediction whether there is a risk that the arms will be used to violate international law. In making this assessment, they may request assurances that the state receiving the arms will not misuse them. But these assurances will need to scrutinized for reliability and credibility, and for whether they do indeed reduce the risk. We can derive a framework from international practice on the factors that states must consider in order to complete a sufficient due diligence assessment of the assurances. Failure to address these factors would strongly suggest that the state making the assessment did not discharge its international obligations under the ATT, EU Common Position, Geneva Conventions and/or Genocide Convention. Those factors are:

• Assurances will be more persuasive and instill greater confidence if they are publicly available (e.g. reported to the ATT Secretariat), articulate specific commitments, and are binding under international law, in addition to being binding under domestic law in the receiving state. 
• The states must consider the nature, type, and quantity of weapons technology or equipment and their normal or foreseeable uses, including the general situation in the importing state and the gravity of the breach of international law that such use would entail. In terms of the text of the assurances, they must address this particular foreseeable use. 
• An exporting authority must consider whether there is any independent mechanism for verification of the use of arms and/or whether there is any effective internal investigation mechanism for investigating, prohibiting and punishing misuse of the arms. 
• Contextually, these assurances will be more persuasive if violations of international humanitarian law and human rights law are unlawful under the domestic law of the state giving assurances. 
• Furthermore, the exporting state must consider whether it has strong bilateral influence with the receiving state and whether the receiving state has a record of abiding by its assurances in the past (including export assurances, assurances of mistreatment and other assurances), and whether there are any other reasons to expect the receiving state’s military to comply with the assurances and not use the arms in a prohibited manner. 
• And of course, the assurances will be less persuasive if the state is known to have a poor attitude to international humanitarian law and human rights law generally, especially if the state has ever previously used exported arms to violate international humanitarian law or human rights law. 

Certainly, this assessment is still case-by-case, but the above analysis provides a framework for assessing assurances. States, and any domestic or international courts or bodies (e.g. the ATT CSP) reviewing export, must take these factors into consideration in assessing any assurances. Legislative oversight committees and media may also wish to scrutinize any decision on assurances in light of these factors. In the specific case of exports to Israel, any authority monitoring export compliance will need to consider a variety of facts, including whether the need for humanitarian aid drops is relevant or whether Israel’s failure to file a report with the ICJ under the South Africa v Israel provisional measures order means that state is not in compliance with its international obligations in this situation. Of course, inability or unwillingness to give assurances to any exporting state would be a significant mark against export, as the receiving state would be refusing to promise not to misuse the arms. As for the Biden Administration, it appears to be keeping any assurances confidential for now, which constitutes a factor against the reliability of the assurances. In any event, the next phase of the export situation may very well include the widespread use of assurances, which will need to be assessed. Assurances might have a steep hill to climb, but there is a legal framework for integrating assurances into a prediction of risk, and deciding whether any risks were overcome with words alone.