1. Introduction
This article identifies a framework for assessing any assurances on the use of exported arms and critiques the assessment by the United States in May 2024 of Israeli assurances in respect of arms used in the Gaza conflict. International humanitarian law (IHL) obliges states, at a minimum, not to aid or assist violations of the Geneva Conventions (GCs) and, arguably, also to ensure respect for them by other state parties. International instruments such as the Arms Trade Treaty (ATT) and the EU Common Position on the Control of Exports of Military Technology and Equipment of 8 December 2008 similarly prohibit the export of arms where there is a risk of serious misuse. Because it is a question of risk, the exporting state must make a prediction before approving export as to the likelihood that the arms would be used to violate international law. In balancing complex political and geopolitical factors, and the legal requirements for export, states may seek assurances from the recipient state as to the use of those exported arms. If those assurances are reliable and credible, any risk might not reach the necessary threshold that would require the state to refuse export. However, it remains unclear what standards international law provides for assessing assurances. This article identifies and clarifies this assessment framework and applies that framework to the situation of Israeli assurances given to the United States regarding arms used in Gaza.
Guidance for assessing assurances under international law is fragmented and partial, but can be synthesised into a coherent framework. The GCs do not provide expressly for any factors for assessing assurances, though the Commentary of the International Committee of the Red Cross (ICRC) does provide some guidance. Other international instruments are also relevant, including the ATT and the EU Common Position, either by binding states directly or by constituting evidence of how states articulate and operationalise what they must do to comply with their negative obligations under the GCs. This approach does not suggest that somehow non-state parties are bound by the ATT or Common Position by a juristic sleight of hand; instead, the practice under both instruments aids in interpreting the obligations under binding law. Both the ATT and EU Common Position also benefit from helpful secondary interpretative materials that provide supplementary opinion on how states are expected to assess the risk of the misuse of exported arms. Certainly, some states are not party to the ATT or bound by the Common Position per se but, in attempting to determine what the GCs prescribe about negative obligations under the GCs and risk of violations in the context of relying on assurances, those instruments constitute persuasive practice.
In addition to these sources, this article also argues that there is already an existing methodology for assessing the impact of assurances on situations of risk, which has been well developed in the context of non-refoulement and its related assessment of risk to the person. The leading case that best articulates the methodology for assessing expulsion assurances is Othman v United Kingdom before the European Court of Human Rights (ECtHR).Footnote 1 The approach in Othman has been largely adopted by other international courtsFootnote 2 and treaty bodies,Footnote 3 as well as states,Footnote 4 including states not party to the European Convention on Human Rights (ECHR),Footnote 5 and can safely be considered the widely adopted benchmark for assessing assurances generally in predictive-risk settings. Despite being applied in the field of expulsion, the underlying assessment methodology for the impact of assurances on risk can be transferred to the situation of arms exports. Both situations concern forming predictions of risk based on diplomatic statements as to the likelihood of how a state will behave in the future. The High Court of England and Wales has confirmed that this approach, again best articulated in Othman, is the appropriate methodology for assessing assurances even outside the expulsion context. In Al Maktoum, the Family Division of the High Court of Justice of England and Wales heard a child custody dispute that involved the Emir of Dubai and received assurances from the Dubai government that the state would comply with the custody decision.Footnote 6 The High Court noted that the approach to assessing assurances was already clear in the law as applied in the Othman case.Footnote 7 The fact that the contexts of child custody differed from expulsion was not deemed material because the underlying methodology applied to any assessment of assurances where the state had to make a prediction. In prior litigation on blocking arms transfers, the UK High Court, in CAAT v SSIT, assessed Saudi assurances with a forward-looking, due diligence evaluation that identified several criteria, which included institutional reforms, trend data, investigations and responsiveness that suggest whether assurances are reliable and credible.Footnote 8 Without citing Othman explicitly, the High Court nonetheless applied criteria similar to those in Othman, demonstrating that there is indeed a methodology for assessing assurances when making a prediction of risk with arms exports. This article takes one step further, makes the application of the Othman criteria explicit and draws on the jurisprudence underlying the Othman criteria to flesh out the methodology for assessing assurances in situations of arms exports.
Having developed the assessment framework, this article will then consider whether the assessment by the US of Israeli assurances for the use of exported arms in Gaza was satisfactory. This article focuses on this specific situation because it is both prominent and, at least, somewhat transparent, providing an opportunity to engage with substantive practice, whereas other uses of assurances are largely opaque. On 8 February 2024, the Biden administration issued National Security Memorandum 20 (NSM-20), which 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.Footnote 9 This policy built on the prior National Security Memorandum 18 (NSM-18), which forbids arms transfers where the US assessed that it was more likely than not that the arms would be used to commit genocide, crimes against humanity or other serious violations of IHL or human rights.Footnote 10 Alongside NSM-20, US law, as well as the law of many countries, has required assurances in situations where the lawful use of exported arms was in doubt. However, NSM-20 required the production of a publicly available report to Congress on the administration’s assessment of the assurances. In connection with NSM-20, the United States demanded and received assurances from several states, including Israel, and subsequently reported on its assessment. It concluded that the assurances were credible and permitted the US to continue exports without – in its view – violating the GCs. With this report, we gain some insight into the methodology the administration used to assess the assurances and have an opportunity to tentatively test whether the assessment truly complies with international law. This article will conclude, however, that the assessment by the US was unsatisfactory, on the ground that the US failed to apply a reasonable assessment by omitting critical lines of inquiry that were clearly required by international law. Having failed to conduct a due diligence assessment, the US may now be exposed to international responsibility for failing to respect its obligations under international law concerning arms exports.
2. The obligation to assess the risk of arms misuse
States are obligated, under a variety of instruments, to assess the risk of misuse of any arms they export. On 12 February 2024, the Court of Appeals of The Hague ordered the Dutch state to stop the export of any F-35 parts to Israel.Footnote 11 This case is currently on appeal before the Dutch Supreme Court, though the Office of the Dutch Procureur-Generaal has already issued a preliminary conclusion that the Court is likely to adopt, which would affirm the decision of the Court of Appeals to prohibit export.Footnote 12 The case was founded on Dutch export law, as informed by international and EU law. The Netherlands is party to the ATTFootnote 13 and the GCs,Footnote 14 and, being a member of the European Union, it must comply with the EU Common Position defining common rules governing the control of exports of military technology and equipment.Footnote 15
The Dutch case is not isolated.Footnote 16 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, arguing that the ‘transfer of [any] weapons or ammunition to Israel’Footnote 17 might violate international humanitarian law, if those arms were foreseeably used in violation of the GCs. Several states – including Belgium,Footnote 18 Canada,Footnote 19 ItalyFootnote 20 and the United KingdomFootnote 21 – have partly or entirely prohibited arms exports to Israel. In parallel with this issue, Nicaragua brought a case against Germany before the International Court of Justice (ICJ) in respect of its arms exports to Israel, claiming that those exports potentially violate Germany’s obligations under the Genocide Convention.Footnote 22 During the oral argument in the case, Germany noted that it was already drawing down its arms exports to Israel.Footnote 23
Aside from completely ending exports, states have at least one other option for reducing the risk of arms misuse and, thus, potentially avoiding state responsibility for violating international law. This option is for the state receiving the exported arms to give assurances about their use. Under NSM-20, the Biden administration decided to request assurances of use from several states, including Israel, in order for them to continue to receive exportsFootnote 24 (the Trump administration rescinded NSM-20 in February 2025Footnote 25). NSM-20 expressly contemplated international humanitarian law as well as ‘other international law’.Footnote 26 However, simply giving assurances is not in itself sufficient; they also need to be convincing. NSM-20 adds that, should the US doubt the assurances, it will consider ‘appropriate next steps to be taken to assess and remediate the situation’.Footnote 27 Axios reported that Israel provided assurances on 14 March 2024.Footnote 28 Those assurances, however, do not appear to be publicly available. On 25 March, the US State Department noted that the assurances appeared to be effective, but a final decision was still pending. On 10 May, the US State Department finally issued its definitive conclusion that the United States was satisfied by the assurances. Whether other states also issued assurances was not revealed until the State Department’s May Report. With this precedent, other states may now wish to adopt a similar practice.
Because assurances may allow export to a state that is reasonably suspected of using the arms to violate international law, we need to know how to assess assurances, remain in compliance with export controls, and not attract state responsibility for violating the underlying legal obligations. Surely, compliance with international law should not be a political issue.Footnote 29 This article will attempt to identify a framework for assessing these assurances. It will first identify standards for assessing assurances contained in the ATT, the EU Common Position, and the GCs. These standards will also be informed by comparative experience with assurances in situations other than arms control. The article will then turn to the specific case study of the US assessment of the Israeli assurances of 14 March.
3. Defining ‘assurances’
Before considering how assurances are assessed for arms export, it is helpful to briefly consider what assurances are. There is no single definition of assurances comparable with, for example, the definition of a treaty in the Vienna Convention on the Law of Treaties.Footnote 30 Instead, assurances are a heterogeneous collection of communications with similar purposes. States describe their assurances with many differing expressions: sometimes they use the term ‘assurances’, but at other times, they refer to the communications as ‘guarantees’, ‘commitments’, ‘undertakings’, and so on. Assurances are sometimes referred to as ‘diplomatic assurances’, though they can be issued by a variety of actors within a state,Footnote 31 and not only by diplomats, but are usually transmitted via diplomatic channels. This article will not use the expression ‘diplomatic’ for this reason.Footnote 32 While often included in a diplomatic note or memo, assurances may take a wide variety of forms – both oral and writtenFootnote 33 – and are contained in various types of legal instrument, including both unilateral and bilateral instruments;Footnote 34 some have even been considered treaties.Footnote 35 The content of assurances ranges from highly specific pledges of future behaviour to vague declarations of facts.Footnote 36 States have given assurances of compliance with treaty obligations;Footnote 37 use of nuclear weapons;Footnote 38 security assurances;Footnote 39 using military forceFootnote 40 or withdrawing troops;Footnote 41 granting safe conduct;Footnote 42 protecting minority rights,Footnote 43 waiving claims;Footnote 44 treatment of investments;Footnote 45 application of air transport services;Footnote 46 arresting,Footnote 47 detainingFootnote 48 or prosecuting individuals;Footnote 49 or even affirming the existence of facts, such as domestic laws.Footnote 50 They have given assurances to states as well as to a wide variety of international organisations and dispute settlement bodies,Footnote 51 including the ICJ,Footnote 52 the World Trade OrganizationFootnote 53 and the ECtHR.Footnote 54 In the export context, states have routinely given assurances on the use of imported nuclear materialFootnote 55 as a condition of the export, or assurances they would not transfer certain arms onwards to third countries.Footnote 56 Thus, assurances arise out of the long-standing diplomatic practice of making statements or agreements that give confidence to other states and are commonly assessed in a similar manner.Footnote 57 As such, the assurances under NSM-20 will fall within this pattern of practice.
4. A proposed framework for assessing arms export assurances
This section will review the GCs, ATT and EU Common Position to identify whether there are any standards or criteria for the assessment of assurances. For the reasons discussed in the Introduction, this section will also examine assurances given in the deportation/expulsion context that articulate international standards for considering the impact that assurances may have on the prediction of risk generally.
4.1. The Geneva Conventions
The first potential source of law is the GCs. This article will not investigate in detail whether there is a positive obligation to ensure respect for the GCs and will focus primarily on the negative obligation.Footnote 58
There is, however, some support for a positive obligation. Note that the Dutch Procureur-Generaal in the Oxfam case has already agreed with the conclusions of the ICRC in its Commentary on the positive obligation,Footnote 59 as did Germany in the Alleged Breaches in respect of the OPT case before the ICJ.Footnote 60 The EU also appears to accept the ICRC interpretation of a positive obligationFootnote 61 and the ATT Secretariat has observed that, based on the information gathered from ATT state parties, there was widespread agreement that the GCs imposed a positive obligation to prevent misuse of arms, confirming specifically the interpretation in the ICRC Commentary.Footnote 62 Thus, there is support for such a position.Footnote 63
Nonetheless, it is sufficient to note that there is at least a negative obligation to refrain from encouraging, aiding or assisting states to violate the Conventions.Footnote 64 The ICRC specifically observes arms exports as an example of an area where a state shall take measures to avoid encouraging, aiding or assisting states to violate the Conventions.Footnote 65 The Arms Trade Treaty Secretariat and Working Group on Effective Treaty Implementation have produced a ‘Voluntary Guide to Implementing Articles 6 & 7 of the Arms Trade Treaty’, in which the ATT Secretariat reports that state parties certainly agree that the Geneva Convention imposes a negative obligation not to aid and assist.Footnote 66 The ATT Secretariat reports that state parties appear to agree with the ICRC view that arms exports to situations of international law violations would constitute prohibited aid or assistance.Footnote 67 As such, states must exercise due diligence in assessing the risk that arms will be misused for IHL violations.Footnote 68 Whether one adopts a due diligence ‘ensure respect’ reading or a narrower ‘do not aid or assist’ view, assurances are relevant only in so far as they demonstrably reduce the risk that exported items will be used in breach; they cannot justify transfer once that risk meets the refusal thresholds in the applicable regime. In terms of a state discharging its duties under the GCs, the ICRC adds only that it ‘depends on the specific circumstances, including 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’.Footnote 69
Despite this obligation, at a minimum the GCs provide little to no assistance in providing for assurances assessment criteria under the obligation to ensure respect for the Conventions.Footnote 70 There is some guidance in the ICRC Commentary where it indicates that the exporting 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’,Footnote 71 when there is a risk of misuse, and, noting the support for the Commentary above, there is room to argue that those factors are relevant. These factors will be included in the list of factors developed in this article.
4.2. The Arms Trade Treaty
More detailed guidance on assessing assurances of arms use comes from the ATT. Obviously, this regime is applicable only for states party to the ATT, which includes 116 states: inter alia, all the Member States of the EU, Australia, Brazil, Canada, China, Japan, South Africa, the UK, but not India, Indonesia, Iran, Russia, Pakistan, the US and Israel. Arms exports, and related assurances, from any of the above-mentioned states parties would absolutely need to be assessed in line with the ATT’s requirements. In addition, while the obligations of the ATT regarding assurances strictly apply only to state parties, any practice under the ATT provides persuasive guidance as to how assurances regarding arms must be assessed to remain in compliance with the GCs and human rights law. Non-state parties would, of course, not be required to participate in the ATT reporting and other obligations, but they are bound by the GCs and human rights law, and the ATT prohibitions on export to situations of IHL violations is designed specifically to operationalise that prohibition. In short, the ATT merely requires an assessment of the likelihood of violating other norms of international law, including IHL and human rights law. At a minimum, a state that is on notice of the risk of exported arms being misused and seeking to discharge its due diligence obligation would be aware of the ATT assessment. This section identifies the factors that are relevant for predicting risk and will include those factors in the proposed list of factors for assessing assurances.
The ATT divides the types of risk into two categories and treats them differently. First, under Article 6 of the ATT, states may not permit the export of arms that would violate that state’s international agreementsFootnote 72 or obligations under United Nations Security Council Chapter VII embargoes.Footnote 73 In addition, the state may not permit export if the state ‘has knowledge at the time of authorization that the arms or items would be used in the commission of genocide, crimes against humanity, [or war crimes]’.Footnote 74 Second, under Article 7, states also may not permit export when the arms ‘would contribute to or undermine peace and security’Footnote 75 or ‘could be used to … commit or facilitate a serious violation of international humanitarian law'Footnote 76 or human rights law,Footnote 77 or commit an act of terrorismFootnote 78 or transnational organised crime.Footnote 79 The ATT permits states to undertake mitigating measures that reduce risk,Footnote 80 but only for the risks under Article 7.Footnote 81 Mitigation is not an option when the state has knowledge that the arms would be used for genocide, crimes against humanity or war crimes.Footnote 82
4.2.1. Exports subject to Article 6: Genocide, crimes against humanity and war crimes
As noted above, Article 6 imposes an absolute export ban where there is a heightened risk, even the certainty of knowledge, of, inter alia, the commission of genocide, crimes against humanity or war crimes,Footnote 83 and, by excluding mitigation, it appears that any assurances from the export receiving state in attempting to diminish that certainty would be disregarded.
Normally, it would stand to reason that the export receiving state could communicate information in assurances, attempting to give the exporting state greater confidence that there was no risk of genocide, crimes against humanity and war crimes. Certainly, states are free to argue that they are not committing (or about to commit) genocide, crimes against humanity or war crimes. Indeed, some states report that they seem to blur any distinction between the assessment of risk and the assessment of risk mitigation into one assessment.Footnote 84 The precise form and degree of knowledge of risk under the ATT remains unclear. It may be possible for a state to submit assurances at the stage of forming the knowledge of risk rather than at the mitigation of risk stage. After all, the prohibition on export under this article will be implicated only when the arms ‘would’ be misused, not when they ‘could’, as provided in other parts of the ATT.Footnote 85
However, Article 6 is distinct from Article 7 in that Article 6 seems implicitly to exclude communication from the export receiving state. Under Article 7, the ATT specifically contemplates communications from the export receiving state.Footnote 86 Specifically, the ATT provides that the exporting state ‘shall, in an objective and non-discriminatory manner, taking into account relevant factors, including information provided by the importing State in accordance with Article 8(1), assess the potential [for misuse]’.Footnote 87 Yet, Article 6 omits this option, suggesting that it is not permitted.
This implicit exclusion is also supported by state practice. The Voluntary Guide reports on state practice under the ATT in implementing the treaty’s export obligations. It is not entirely clear whether this Voluntary Guide is intended to synthesise subsequent practice under the ATT or constitute an authentic interpretation of the ATT, in the sense of the interpretation methodology in the Vienna Convention on the Law of Treaties.Footnote 88 No claims to such weighty authority are made in the Voluntary Guide, which expressly disclaims any prescriptive authority.Footnote 89 However, the Voluntary Guide does collect state practice which constitutes evidence of how states interpret, understand and fulfil their obligations under the ATT.
According to the Voluntary Guide, states have adopted somewhat consistent practices for assessing arms exports. The assessment process generally involves personnel with a diversity of professional profiles with backgrounds in law, military technology, foreign policy, defence analysis, human rights and IHL.Footnote 90 Assessments typically include information from governmental entities, intelligence services, diplomatic missions, governmental research institutions, human rights institutions, cooperation arrangements, international organisations, commissions of inquiry, fact-finding missions, multilateral institutions, think tanks, media, and other non-governmental organisations.Footnote 91 These sources clearly do not include the export receiving state.
As for the topics that states consider in an assessment, the Voluntary Guide reports that they consider the recipient country’s attitude to human rights and principles of IHL, the specific type of material that is exported, the specific recipient and end user, and the anticipated use of the material.Footnote 92 One could conclude that this provision includes the export receiving state’s assurances; yet the only views that the exporting state can consider are views on human rights and IHL, which are Article 7 concerns, and not genocide, crimes against humanity or war crimes. Assurances on those topics appear to be excluded.
The Voluntary Guide also notes that exporting states exercise special caution in approving exports when there is a ‘conflict situation’ in the receiving country, noting that some states prohibit outright any exports to countries in conflict.Footnote 93 This statement demonstrates that exporting states may shift the burden of proof and not approve exports at all, despite any other factors. This supports the reasoning that when there is a risk of genocide, crimes against humanity or war crimes, states could simply prohibit exports under the ATT without regard to any communication from the export receiving state.
This reporting on practice is confirmed in another soft law document, the United Nations Office of Disarmament Affairs (UNODA) ‘ATT Implementation Toolkit’. The Toolkit states that ‘[i]n carrying out the assessment, the authorization agency should seek information and inputs from other relevant government entities as well as other sources, including open sources’.Footnote 94 However, implicitly, the Toolkit is referring to states other than the export receiving state.
Thus, it would appear that – in respect of Article 6 concerns of genocide, crimes against humanity or war crimes – there is no option for a state to provide assurances or any other form of mitigation.Footnote 95 Essentially, should a state submit unsolicited assurances regarding those topics, the state receiving such assurances would be obliged to exclude them from its analysis. As has been argued in cases of assurances in expulsion cases, if a state has to give assurances that it is doing nothing wrong, then there is necessarily a serious risk that it is.
4.2.2. Exports subject to Article 7: International humanitarian law and human rights law
Under Article 7, states also may not permit export when the arms ‘would contribute to or undermine peace and security’Footnote 96 or ‘could be used to … commit or facilitate a serious violation of international humanitarian law’Footnote 97 or human rights law,Footnote 98 or commit an act of terrorismFootnote 99 or transnational organised crime.Footnote 100 Compared to Article 6, where the standard is ‘would’ be misused, this article mixes usage of ‘would’ and ‘could’. Focusing on violations of IHL and human rights, which are the relevant topics for this article, the standard is ‘could’, a lower threshold than ‘would’.Footnote 101
The Toolkit states that the standard of inquiry into risk, specifically for potential human rights violations, should incorporate ‘the due diligence standard in human rights law’,Footnote 102 which ‘requires that exporting States engage in an effective inquiry in order to make a reasoned determination as to whether the proposed export carries a substantial risk of facilitating serious violations’.Footnote 103 The Toolkit suggests some contextual factors for assessing risk that the use ‘could’ violate IHL,Footnote 104 which are non-exhaustive.Footnote 105 Many of these factors closely track factors for assessing assurances from other sources or contexts.Footnote 106 Those factors that would be relevant include:Footnote 107
[w]hether the importer or end‐user has committed [or aided and abetting] serious violations of international humanitarian law; … [w]hether or not the importing State has taken all feasible measures to prevent violations of international humanitarian law [and has legal and judicial measures in place for punishment and repression of violations] … [w]hether or not the importing State disseminates international humanitarian law .. and has integrated international humanitarian law into its military doctrine, manuals and instructions.
Similarly, indicators for the risk of serious human rights violations include:Footnote 108
[w]hether or not the recipient State is a party to … human rights instruments; … [w]hether or not there is evidence that the type of arms … has been used for serious violations of international human rights law … [w]hether or not the importer … has aided or abetted a serious violation …; [w]hether there is a record of impunity … [w]hether or not the recipient State has adopted national legislation … to implement [those] instruments; … [w]hether or not the recipient State has … effective procedures for … investigation [and] accountable structures … competent, independent, impartial and functioning judiciary … independent monitoring bodies … [w]hether or not the recipient State disseminates international human rights law.
According to the UNODA, one factor that is particularly important for assurances regarding IHL is ‘[w]hether or not the importing State has made a formal commitment to apply rules of international humanitarian law and taken appropriate measures for their implementation’.Footnote 109 UNODA does not suggest a similar factor of human rights violations, though, again, the list is not intended to be exhaustive.
In reaching this determination on risk under the ATT, scholars have opined that a state party could also consider factors such as:Footnote 110
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.
Under Article 7, the ATT permits states to undertake mitigating measures that reduce risk.Footnote 111 Whether the risk arises under Article 6 or 7, the exporting state cannot export, but what distinguishes Article 7 from Article 6 is that acknowledged risks under Article 7 can potentially be overcome with mitigation measures.Footnote 112 According to the ATT, mitigation can occur only when a state undertakes ‘measures’.Footnote 113 The ATT gives two examples of qualifying measures: ‘confidence-building measures or jointly developed and agreed programmes’.Footnote 114 The Voluntary Guide adds post-shipment controls,Footnote 115 changes in legislation (including effective disciplinary and judicial bodies), and ‘national action plans’.Footnote 116 This suggests that the only qualifying measures would be concrete, physical behaviour or formal changes in policy, and not mere statements.
However, there does seem to be some space for diplomatic efforts, including assurances, to offer mitigation. The Voluntary Guide notes that measures ‘could also include formal commitments of the recipient State to relevant international instruments and their implementation on the national level’.Footnote 117 It also mentions ‘commitments to international instruments relevant to the use of arms, such as disarmament or export control agreements’ and ‘broader formal commitments, for example to respect IHL in the case of armed conflict’.Footnote 118 It adds that some states include ‘a description of the specific intended end use of the arms in the end-user certificate, as well a non re-transfer clause’.Footnote 119 This reporting on practice is affirmed in the UNODA Toolkit, in which UNODA confirms examples of statements that could qualify as mitigation measures, such as ‘a declaration … of intended use of the transferred weapons or items, accompanied by the undertaking/assurance/guarantee not to use them for other purposes’;Footnote 120 ‘disclosure by the importing State of its records regarding observation of relevant international human rights law, [and] international humanitarian law’;Footnote 121 ‘[u]ndertaking by the importing State not to re‐export or re‐transfer in a manner that would run counter to the provisions of the ATT’;Footnote 122 ‘[p]rovision of information on weapons or items stolen, lost or otherwise unaccounted for’;Footnote 123 ‘[e]nhancement of transparency on military matters’.Footnote 124 However, the Voluntary Guide reports that state parties would positively evaluate such statements only when ‘confirmed by actual State practice’Footnote 125 and in the light of the export receiving state’s compliance with international law generally.Footnote 126 Obviously, should states not comply with mitigation measures, the exporting state would need to reassess the export licence.Footnote 127
4.2.3. Conclusions on the ATT approach
Following from the above analysis of the ATT approach to the assessment of assurances, we can reach several conclusions. Firstly, when there is knowledge, and thus a high degree of certainty of risk, of the commission of genocide, crimes against humanity or war crimes, no assurances can overcome the risk. There is a narrow possibility for states to submit assurances questioning the certainty of risk, though states appear to be precluded from considering the assurances when there is sufficient information from other sources that there is knowledge that violations would occur. If the state is currently in an armed conflict, that assessment could involve a shift in the burden of proof.
Secondly, when the risk is a possibility (rather than a certainty) of IHL or human rights violations, states may attempt to mitigate the risk with assurances. The ATT does not preclude consideration of the state’s assurances. The standard is one of due diligence.
Following from the practice of states described in the Voluntary Guide, the ATT should be interpreted to require assessment by a team with broadly diverse backgrounds and expertise, and should be supplemented with information from a range of governmental and non-governmental sources. Assessment of the assurances should include:
• whether the state’s assurances constitute a formal commitment/undertaking to apply IHL and human rights;
• whether the assurances include an undertaking/guarantee not to misuse the arms or re-export them elsewhere;
• whether the assurances include disclosure of the state’s record of IHL and human rights compliance and any arms unaccounted for;
• the nature, type, and quantity of weapons to be exported and their normal and reasonably foreseeable uses;
• the state’s attitude to human rights and IHL generally;
• whether the state has committed or aided and abetted serious IHL or human rights violations, or otherwise violates international law generally;
• whether the state has complied with previous similar assurances;
• whether the state has taken all feasible measures to prevent IHL violations (including disseminating information and integrating IHL into its military doctrine, manuals and instructions);
• whether the state has mechanisms in place for investigation, punishment and repression of IHL violations or whether there is a record of impunity;
• whether the state is party to human rights instruments;
• whether the state has adopted national legislation to incorporate its IHL and human rights obligations; and
• whether there is evidence that the type of arms has been used for serious violations of human rights law.
4.3. The EU Common Position
Having considered the ATT, this article will now turn to the EU Common Position, the last of the sources referred to by the Dutch Court of Appeals.Footnote 128 Of course, the Common Position applies only to EU Member States; however, it is expressly intended to discharge its obligations specifically under the ATT and, generally, under IHL and human rights law. Thus, while it does not bind non-EU states, it does provide an interpretation of the content of binding law concerning exports, risks and, potentially, assurances. Just as in the prior two sections, this section will collect the factors for predicting risk under the Common Position and integrate them into the list of factors this article is developing for assessing assurances.
The EU Common Position provides some guidance that is relevant for assurances, including that an exporting state must ‘assess[] the recipient country’s attitude towards relevant principles established by instruments of international humanitarian law and respect for international humanitarian law’.Footnote 129 The Common Position notes that this assessment takes place in the context of ‘the recipient country’s attitude towards relevant principles established by international human rights instruments’, ‘the nature of the technology or equipment’, and ‘the record of the buyer or recipient country with regard to … its compliance … with international humanitarian law’.Footnote 130
The ‘User’s Guide for the EU Common Position’ adds several additional considerations for assessing risk. Although not binding, this soft law instrument explains and further articulates the methods for compliance with the Common Position and, as such, provides an interpretation of the content of the Common Position. In the earlier 2009 version of the User’s Guide, it mentioned assurances very briefly. It provided that the export licensing certificate should include not only an ‘[i]ndication of the end-use of the goods’ but also ‘[a]n undertaking, where appropriate, that the goods being exported will not be used for purposes other than the declared use; [and] [a]n undertaking, where appropriate, that the goods will not be used in the development of [weapons of mass destruction]’.Footnote 131 It also noted that the recipient country’s attitude to human rights and humanitarian law was relevant,Footnote 132 but did not mention how an exporting state should learn of the receiving state’s attitude, if not by assurances.
The current User’s Guide of 2019 appears to take a slightly different approach to assurances. It largely follows the language of the earlier Guide in advising that certain elements should always be demanded as part of the export licence, including:Footnote 133
[a]n undertaking, where appropriate, that the goods being exported will not be used for purposes other than the declared use; [and] [a]n undertaking, where appropriate, that the goods will not be used in the development, production or use of chemical, biological or nuclear weapons or for missiles capable of delivering such weapons.
It adds the possibility that states might also demand, as circumstances so require, ‘[a]n undertaking that: the goods will be used for civil end-use; the goods will be incorporated; [and] the goods will be re-exported to a specified country’.Footnote 134 Also, in terms of assessing the state’s export request, the 2019 User’s Guide recommends that states take into consideration ‘the recipient’s intentions as expressed through formal commitments’.Footnote 135
According to the updated User’s Guide, this assessment takes place in the context of:Footnote 136
inquiry into the recipient’s past and present record of respect for international humanitarian law, … and the recipient’s capacity to ensure that the equipment or technology transferred is used in a manner consistent with international humanitarian law and is not diverted or transferred to other destinations where it might be used for serious violations of this law.
This poses questions such as ‘[i]s there national legislation in place prohibiting and punishing violations of international humanitarian law? … [h]ave mechanisms been put in place to ensure accountability for violations of international humanitarian law committed by the armed forces and other arms bearers, including disciplinary and penal sanctions?’.Footnote 137 In addition, the User’s Guide advises states to consider ‘an inquiry into the recipient’s past and present record of respect for international humanitarian law’ and ‘the recipient’s capacity to ensure that the equipment or technology transferred is used in a manner consistent with international humanitarian law’.Footnote 138 The Guide also notes that only a ‘pattern of [IHL] violations’ should serve as the basis for an export refusal, not ‘[i]solated incidents’.Footnote 139 It mentions only that a state must consider its degree of influence over the receiving state, and therefore its ability to ensure compliance with assurances, in discharging its positive obligation to ensure respect for international law by its partner state.Footnote 140 We can deduce that the state must also consider its degree of influence in connection with its negative obligation to refuse export. The negative obligation is, after all, already set out in the Common Position itself, whereas the User’s Guide needed to discuss expressly the positive obligation that is implicated only after export.
Therefore, the Common Position accepts assurances with similar considerations as the ATT. Assurances should be examined for (a) an indication of the end use of the goods; (b) an undertaking that the goods being exported will not be misused; (c) an indication of whether the assurances constitute formal commitments; (d) an expression of attitude to human rights and humanitarian law; (e) the nature of the technology or equipment; (f) the record of state compliance with international humanitarian law, with particular attention to patterns of misuse; (g) the capacity to ensure that the arms are not misused or re-exported; and (h) the existence of national legislation prohibiting and punishing violations of IHL and accountability mechanisms.
4.4. Assurances in other contexts: Expulsion
In addition to these factors, states can also consider their experience of making predictions from assurances in other situations, especially their extensive experience with expulsion assurances. As mentioned above, assurances are routinely employed in a pattern of practice for a wide range of state-to-state communications on various topics. Ultimately, the issue for the assessment of assurances in any context is whether the state pledge is credible.Footnote 141 However, assurances made in some contexts (such as the transfer of nuclear material) might never be subjected to litigation and therefore the precise assessment framework might escape clarification. Expulsion assurances, on the other hand, have been intensively litigated and multiple courts, tribunals and international treaty monitoring bodies have opined on factors for their assessment. Thus, expulsion assurances provide a rich source for comparative purposes. This section will review the factors in the jurisprudence for assessing expulsion assurances and, where they can be generalised in making predictions for compliance with any assurances, will include them in the list of factors for assessing arms export assurances.
Beyond all the examples given previously where assurances have been given, states make extensive use of assurances in situations of expulsion, deportation and extradition, when there is a risk that the person will be executed, tortured, severely mistreated, or suffer some other grave violation of human rights.Footnote 142 This is done to avoid implicating the non-refoulement obligation. That being said, states do not have a perfect track record with expulsion assurances and states have mistreated expelled people despite their assurances.Footnote 143
Nevertheless, assurances form part of the attempt at an intelligent prediction of risk. States already know from their experience with expulsion assurances that mere pieces of paper do not necessarily eliminate risk and they have a methodology of taking assurances into account in the risk analysis. They need to be scrutinised carefully and in the light of the relevant context. As the question is one of making a good faith prediction, there may be cases where a state is unable to issue any assurance that would ever instil confidence, but this remains a matter to be considered on a case-by-case basis. Arms export is the same. Practice with expulsion assurances can help to identify important variables in making a prediction of whether a state will comply with its promises. This analogy with expulsion assurances is all the more relevant when we consider that the purpose of expulsion assurances is to avoid the implication of non-refoulement, whereas in the situation of arms assurances, the purpose is also to avoid the implication of another applicable legal rule: the obligation to prohibit export. If the question with arms-use assurances is essentially whether they are reliable and credible, and whether they reduce the risk of misuse, then, in conducting risk assessment in the arms export context, states must apply the tools they have from their experience with other risk assessments.
As mentioned above, this article will rely on the Othman case for the best statement of the various factors that a state must consider in conducting a risk assessment with assurances. In that case, the Court identified eleven factors for assessing assurances that have arisen from the Court’s jurisprudence over the years. Specifically, the Court concluded that states must consider:Footnote 144
(a) whether the terms of the assurances have been disclosed;Footnote 145
(b) whether the assurances are specific or are general and vague;Footnote 146
(c) who has given the assurances and whether that person can bind the receiving state;Footnote 147
(d) whether local authorities can be expected to abide by them;Footnote 148
(e) whether the assurances concern treatment that is legal or illegal in the receiving state;Footnote 149
(f) whether they have been given by a state party to the ECHR;Footnote 150
(g) the length and strength of bilateral relations between the sending and receiving states, including the receiving state’s record in abiding by similar assurances;Footnote 151
(h) whether compliance with the assurances can be objectively verified;Footnote 152
(i) 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;Footnote 153
(j) whether the applicant has previously been ill-treated in the receiving state;Footnote 154 and
(k) whether the reliability of the assurances has been examined by the domestic courts of the sending/contracting state.Footnote 155
Criterion (c) above, in particular, calls for additional comment. It refers not only to the competence of the individual giving the assurances, but also the legal effect of the assurances and whether it is binding under international law.Footnote 156 When the Court stated that this criterion was applicable, it did so in reaction to the argument that the assurances in the case were not reliable because they were not binding under international law, compared to treaties that prohibit torture.Footnote 157 In addition, the Court decided Othman on the same day as Harkins and Edwards v United Kingdom, in which it noted specifically that the assurances in that case were binding under international law because they had been provided by the US Embassy, which was competent to bind the United States under international law.Footnote 158 Therefore, that criterion must be read to include both the question of whether the assurances were binding and whether the individual who gave the assurances was competent to bind the state.
Virtually all of these Othman factors for predicting risk in the light of assurances are also relevant for predicting the risk of misuse of arms in the light of assurances. None of them appear to be uniquely limited to human rights issues, but are instead generalisable standards for making predictions from assurances in any context. As such, they are the implicit standards for any risk assessment with assurances, whether it is expulsion or arms export. Some of the criteria already closely resemble similar factors mentioned above for assessing the risk of arms misuse, further demonstrating that these standards are common to risk assessment generally. Applying the Othman factors to the situation of arms misuse introduces additional justifiable factors for determining the likelihood of compliance with international commitments beyond the minimal requirements in the various treaties and other instruments mentioned above. At a minimum, states are aware that in order to assess assurances in the expulsion context, they must have regard to these factors to make an intelligent prediction, satisfy their due diligence obligation and avoid violating their international obligations.
4.5. The proposed assessment framework
Following from the above sources of assurance assessment, we can synthesise the following framework approach.
States that are parties to the ATT cannot accept assurances from states, or give any weight to them, when there is a risk of genocide, crimes against humanity or war crimes.Footnote 159 Thus, the assessment of assurances does not come into play.
Aside from this preliminary requirement under the ATT, when implementing their export obligations, states must make a prediction of whether there is a risk that the arms will be used to violate international law and may consider assurances in this analysis.Footnote 160 Any assurances must be scrutinised for reliability and credibility following from the requirements of the GCs as informed by practice under the GCs, such as that under the ATT and EU Common Position. Drawing on these sources, there are a variety of factors that states must consider in order to complete a sufficient due diligence assessment of assurances. Failure to address these factors would strongly suggest that the state making the assessment did not discharge its international obligations. Those factors are as follows.
4.5.1. Form of the assurances
Assurances will be more persuasive and instil greater confidence if they are publicly available (for example, reported to the ATT Secretariat),Footnote 161 articulate specific commitments,Footnote 162 and are binding (formal undertaking) under international law, in addition to being binding under the domestic law of the receiving state.Footnote 163
4.5.2. Content of the assurances
The assurances should specifically address the relevant IHL and/or human rights concerns, or concerns of re-export to situations of possible misuse, and pledge to comply with the applicable international norms.Footnote 164
4.5.3. Quality of the arms that the assurances address
The states must consider the nature, type and quantity of weapons technology or equipment and their normal or foreseeable uses,Footnote 165 including the general situation in the importing stateFootnote 166 and the gravity of the breach of international law that such use would entail.Footnote 167 In terms of the text of the assurances, they must address this particular foreseeable use.Footnote 168
4.5.4. Enforcement measures
An exporting authority must consider whether there is an independent mechanism for verification of the use of arms and/or whether there is any effective internal mechanism for investigating, prohibiting and punishing misuse of the arms, as well as mechanisms for investigation, punishment and repression of IHL violations with a record of preventing impunity for violations.Footnote 169
4.5.5. Complementary legal obligations
Contextually, these assurances will be more persuasive if the state giving the assurances is party to IHL and human rights legal instrumentsFootnote 170 and/or has incorporated/legislated those norms in the domestic law.Footnote 171 This includes taking all feasible measures to prevent violations (such as disseminating information and integrating international law into its military doctrine, manuals and instructions).Footnote 172
4.5.6. Context of the assurances
Furthermore, the exporting state must consider whether it has strong bilateral influence with the receiving stateFootnote 173 and whether the receiving state has a record of abiding by its assurances in the past (including export assurances and assurances of mistreatment),Footnote 174 and whether there are any other reasons to expect the military of the receiving state to comply with the assurances and not use the arms in a prohibited manner.Footnote 175
4.5.7. Context of state engagement with international law
Of course, the assurances will be less persuasive if the state is known to have a poor attitude to IHL and human rights law generally,Footnote 176 especially if the state has previously used exported arms to violate IHL or human rights law.Footnote 177 Assurances could include the disclosure of the record of compliance with IHL and human rights law generally (and noting any arms unaccounted for).Footnote 178
4.5.8. The assessment team
Finally, the assurances must be assessed by a team with broadly diverse backgrounds and expertise, and should be supplemented with contextual information from a range of governmental and non-governmental sources.Footnote 179
Noticeably, this methodology for assessing assurances does not permit a counterbalancing consideration of generalised foreign policy (or domestic policy) concerns that might persuade the state to grant the export despite there being a risk. The factor of the general bilateral relations between states is aimed solely at determining whether the political relationship between the states is such that the state can effectively prevent violations, and not whether risky exports should be permitted to maintain political favour. Thus, there is no space for a state to say that its foreign policy needs compel it to overlook risk.
Ultimately, assurances are based on a case-by-case analysis, but the methodology for assessing them remains consistent as a persuasive interpretation of the requirements under international law. States, domestic and international courts or tribunals, and even treaty compliance bodies such as the ATT Conference of States Parties, must take these factors into consideration in assessing any arms export assurances. Equally, other domestic oversight committees and the media may also scrutinise assurances in the 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 orderFootnote 180 means that the state is not in compliance with its international obligations in this situation generally. Of course, inability or unwillingness to give assurances at all would suggest that the receiving state refuses to promise not to misuse the arms. However, international law does have something to say about the assessment of assurances. The next section will examine whether the United States fulfilled its international due diligence obligation in assessing Israeli assurances when it issued its first Report under NSM-20.
5. Case study: The assessment of Israeli assurances by the US State Department in May 2024
After Biden issued NSM-20, the US State Department contacted those states in receipt of arms exports that might be misusing them, including Israel, and requested assurances. In May 2024, the US State Department delivered its first report to Congress under NSM-20 and provided its analysis of those assurances.Footnote 181 Initially, most of the attention focused on whether the reporting on pact compliance with IHL and use of transferred arms was accurate.Footnote 182 We could speculate that the State Department took this approach to focus on past practice because that is the approach for the assessment of compliance with US domestic law on arms transfers, though this speculation has not been confirmed. However, in addition to what may have been factual errors in documenting past practice, focusing almost exclusively on past practice is not in compliance with the predictive methodology under international law described in this article, and thus may have amounted to a violation of international law.
The US is obviously not a party to the ATT, nor is it an EU Member State, and is not bound by those obligations directly. As such, the US did not take the ATT or Common Position into consideration. Although NSM-20 includes violations of ‘other international law’,Footnote 183 presumably including genocide, the US is not required specifically to omit consideration of mitigation assurances, as the ATT requires. However, as this article has argued, the ATT and Common Position seek to interpret and operationalise the obligations under the GCs, and do provide some standards for assessing assurances, so the US should have informed itself of those international standards.
Requesting assurances alone is not particularly persuasive of any risk of misuse.Footnote 184 By way of comparison, some authoritiesFootnote 185 have argued that the mere fact that a state has requested assurances means that it must have perceived a risk. That said, the Canadian Supreme Court, in Badesha, concluded that states may ask for assurances ‘out of an abundance of caution’.Footnote 186 In the situation concerning Israel, however, the US had previously requested arms-use assurancesFootnote 187 even before issuing NSM-20. The fact that it had to return to Israel and request new additional assurances after NSM-20 suggests that the first round of assurances was already deemed insufficient. Such a conclusion then further suggests that a second round would not be sufficient, unless the assurances were considerably enhanced. Thus, this background should have formed part of the State Department’s analysis in the Report and shows that the US had credible indications that violations of international law might occur.
The State Department correctly noted in the Report thatFootnote 188
NSM-20 requires the Secretary of State to obtain credible and reliable written assurances from certain foreign governments that they will … use certain U.S. government (USG)-funded defense articles in accordance with international humanitarian law (IHL) and, as applicable, other international law.
As a threshold requirement, the US appears to demand that these assurances are given in writing. This article has not concluded that arms-use assurances must necessarily be given in writing, so this requirement goes above and beyond international law.
The State Department further noted that it had developed its own framework for assessing the credibility and reliability of assurances, based on NSM-20 (and not necessarily international law), including: (a) whether they were provided ‘by a senior official or officials in the partner government with authority to make commitments on behalf of their government related to the required assurances’,Footnote 189 measured by whether the person has such authority will be assessed by their ‘position, responsibilities, and authority … in relation to the subject matter of the assurances’Footnote 190 (in the view of the State Department, ‘Minister-level officials from the relevant ministry or above would be appropriate in most circumstances’Footnote 191); (b) ‘[w]hether the individual providing the assurances is understood to be credible in doing so’;Footnote 192 and (c) ‘[t]he likelihood that the partner government will comply with both assurances’,Footnote 193 measured ‘based on past practice’.
In addition to these factors, the State Department alluded to other criteria that it deemed important for its assessment, though it did not name them: ‘[a]ssessment of the credibility and reliability of these assurances is based on consideration of the following factors, among others’.Footnote 194 Thus, while the Report is somewhat transparent, it is not as open as it could be. This vague statement means that the State Department might very well have applied some or all of the criteria in this article and not reported it. However, given the State Department’s analysis, it seems doubtful that it did assess all of the criteria in this article, as will be explained below.
From the few factors that are listed, we can already deduce that the State Department is unlikely to have followed international practice in respect of assurances. The Report omitted explicit reliance on a number of critical factors, as identified above, and took an unfocused approach to those factors it did consider. The fundamental error in the Report, however, is overlooking the need to make a prediction of risk rather than merely document past practice. While the Report considered past practice of the misuse of arms, it did not consider past practice on compliance with assurances. Curiously, the Report did note past compliance with assurances but only in the case of Colombia, so the State Department was clearly aware that this should be part of the analysis.
5.1. The factors actually used in making a prediction of risk
What is remarkable is that, having identified certain factors in the Report, the State Department did not appear to actually apply those very factors that it claimed should apply. In practice, the Report instead considered the following topics: (1) ‘[a]ssessment of credible reports or allegations that certain defense articles … have been used in a manner not consistent with international law, including international humanitarian law … and if so, whether the recipient country has pursued appropriate accountability’;Footnote 195 (2) an assessment of ‘best practices for mitigating civilian harm’ and implementation efforts;Footnote 196 (3) a description of deliveries that reached unintended recipient states to be used for unintended purposes;Footnote 197 and (4) impediments to the delivery of foreign aid. This article will not assess the last topic as it does not concern arms exports.
The Report does note that the position of the person issuing the assurances is important,Footnote 198 but it omits to mention whether that authority is relevant for whether the assurances are binding under international and/or domestic law. Instead, the Report focuses on ‘[w]hether the individual providing the assurances is understood to be credible in doing so’Footnote 199 and provides no further information on how credibility is to be determined.
Turning to the content of the assurances, the Report does not engage at all with what the assurances undertake. In short, while the State Department considers the person giving the assurances to be important, it does not seem to concern itself with the specific substantive commitments that were given. Because the assurances were never released to the public, we have no way of knowing whether they simply reiterated well-known facts (such as IHL or human rights treaties to which the state has adhered) or whether they promised to take action (such as promising to comply with those treaties).
As mentioned above, the State Department noted that it will identify the ‘likelihood that the partner government will comply with … assurances based on past practice’.Footnote 200 Yet, the Report does not clarify which practice is relevant. Past practice could include past practice of arms use, but it could also include past practice with assurances of arms use,Footnote 201 or even assurances generally.Footnote 202 It might even consider past practice complying with IHL and international law generally. At one point, the Report notes that the Israeli government resisted cooperating with the US assessment of how exported weapons were being used and its application of IHL,Footnote 203 though this observation does not clearly fit within the Report’s own stated methodology.
Furthermore, the State Department does not explain its reasoning for the correct judgement that past practice is indicative of the risk of future misuse. While this conclusion might be intuitive, the Report does not state how it will infer future risk from past practice. This is important because the UK High Court in CAAT v SSIT concluded the opposite: that ‘the mere fact of prior breaches does not equate to a risk of future breaches’.Footnote 204 Past practice under a different government might not be indicative of what the present government might do. Thus, while past practice is relevant, it cannot serve as the central criterion for predicting future behaviour and must be taken in context. For example, the Report does not mention whether there are any reasons to expect that the actors mentioned in the Report are unlikely to comply with the assurances regardless of having given them.
The Report makes no mention of the nature, type and quantity of weapons technology and equipment subject to export, and the typical or foreseeable uses of that material; nor does it consider the overall situation and the possible gravity of IHL violations that using such weapons would produce:Footnote 205 for example, using large bombs in civilian areas would be more likely to result in serious indiscriminate harm than smaller bombs.Footnote 206 This omission is surprising considering that the previous Biden administration was well aware that different weapons have the potential for different violations when it paused the export of 2,000-pound bombs.Footnote 207 Yet, the Report makes no mention of such considerations. It does mention previously exported arms and presumes that they were used in what were likely to have been violations of international law but, again, fails to make a prediction of future use. Taking note of the specific arms involved would be necessary to determine whether the commitments in the assurances addressed the potential risk of misuse.
We see no evidence of any independent verification or investigation mechanisms in the Report, nor any observation of the adoption of IHL and human rights law into domestic law. Clearly, investigations happen,Footnote 208 which is why the Report’s failure to evaluate the reliability and quality of such investigations is surprising. The quality and nature of the investigations would be very informative because, in many cases, independent media cannot access the conflict zones.Footnote 209
Finally, the Report does not consider the legal, political, diplomatic or other influence that the US may have over the state in receipt of exported arms. Indeed, under the Leahy law, the State Department influence is not a factor; the only consideration is whether there is credible information that a unit has committed a gross violation of human rights. However, under the assessment of assurances in international law, influence can be important to know. Strong influence suggests that the state can persuade the other state to comply with its assurances and they are more reliable. The scope and scale of influence will clearly differ from state to state, depending on the nature of the unique relationship, but the Report completely omits any discussion of influence, or steps that the US could take, to compel states to comply with their assurances.
The conclusion on the factors used in the Report must be that they are inadequate for predicting the risk of future misuse. The State Department did not examine multiple lines of inquiry that are easily discernible from the jurisprudence on assurances and risk assessment. It has to be acknowledged that the State Department may have taken these factors into consideration internallyFootnote 210 and simply failed to mention them in the written Report, though such lack of transparency would harm the value of the Report for Congress and the public. In addition, the public assessment suggests that the four questions for each state were the primary motivating factors in its assessment, which, in turn, suggests that the assessment did not fulfil the US’s due diligence obligation under IHL and human rights law.
5.2. Limitations on the scope of the inquiry
As if the limited list of criteria was not alone problematic, the Report failed to engage meaningfully with those criteria it did consider. The Report fails to note which state official gave the assurances, whether that person was credible and had the proper authority to give assurances, despite mentioning that these were all important considerations. While we do not know who gave the assurances from the other states, we have media reports that the Israeli assurances were given by Yoav Gallant, who was serving as Minister of Defence at the time.Footnote 211 The Report does not confirm this. Of course, the Report was published several days before the arrest warrant for Gallant was issued by the International Criminal Court (ICC).Footnote 212 Presumably, an arrest warrant for an issue that overlaps with the assurances undermines their credibility, so the omission of identities from the Report undermines its conclusion.
A further problematic practice in the Report is the secrecy of the assurances. NSM-20 required assurances in writing, but did not order that they be made public, even with redactions, if necessary. Instead, the Report merely summarised whether recipient states had previously complied with international law, focusing on the use of arms, accountability, mitigation of civilian harm, whether arms reached their intended recipient, and delivery of humanitarian aid. Following from the factors mentioned above, and following the lead of the ECtHR in Othman, we can speculate that the lack of publicity suggests that the assurances that it received were not very persuasive.Footnote 213
5.3. Analysis of the evidence in the report
The article will critique in this section how the State Department analysed the evidence it had before it. As noted above, most of the evidence that could have been potentially helpful was not included in the Report and it concentrated only on past practice of complying with international law. As such, the limited evidence leads us to conclude that the State Department assumed that the ‘likelihood that the partner government will comply with … assurances based on past practice’Footnote 214 can be proved solely with evidence of past practice. However, even this evidence of past practice is itself limited to past practice of misuse of arms provided by the US and does not include any past practice of compliance generally with IHL or with assurances.
With this limited range of evidence, the Report lacks clarity in how the evidence was evaluated. The guiding question for the assessment of assurances must be whether there is a risk that the recipient state will use the arms to violate international law. The Report, however, does not explain why past practice was persuasive about there being no risk, even while acknowledging that there have been past examples of violating international law. The Report further takes note of past practice of investigating misuse, yet does not observe whether there are any present independent verification or investigation mechanisms.
The section on Israel, in particular, was notable for its slightly different analysis of evidence. While it used the same methodology and scope of inquiry, it included several pages of extra introductory material framing the conflict, which were not present for other states.Footnote 215 For example, this section of the Report notes that the conflict began on 7 October 2023, which shows the choice of omitting prior elements of the conflict in the area.Footnote 216 It stated that ‘Hamas does not follow any portion of and consistently violates IHL’,Footnote 217 which, while potentially correct, is not supported by any evidence. The Report also mentions the ‘stated objectives’ of the Israeli military operation ‘of destroying Hamas’s military capabilities and dismantling its infrastructure’.Footnote 218 It observed, without deciding, that there are conflicting numbers of civilian casualties from the ‘Hamas-controlled’ Gaza Ministry of Health and Government of Israel.Footnote 219 The Report offers no reason for this different structure and slightly different fact-finding approach, leading us to speculate that it may have been added by a different author or included because of the significant public interest in the Gaza conflict (though the Ukraine conflict presumably has similar levels of public interest and does not have additional contextual facts). Most importantly, the Report does not mention how the additional facts and conclusions have any bearing on whether Israel is likely to comply with its assurances in the future, which suggests that the State Department may have considered irrelevant facts in reaching its assessment on the risk of misuse.
6. Conclusion
This article has argued that not only does international law prohibit the export of arms when there is a risk of misuse, international law also offers a framework for assessing any assurances on the use of those exported arms. As a case study, the article then critiqued the United States’ first NSM-20 Report on assurances that it has received for exported arms, including assurances from Israel, among other states, and found that assessment to be lacking. This finding leads to the conclusion that the US may have violated the GCs by permitting exports to Israel for use in the Gaza conflict when there is a risk of misuse that was not effectively mitigated by assurances.
The article determined that there were a number of factors that states would need to consider in assessing assurances, drawn from various sources. The GCs offered some limited guidance for such assessment, but the EU Common Position and, especially, the ATT were more helpful. Of course, not all states are members of the EU or parties to the ATT, but those sources are still informative in so far as they give more content to enforcing the export requirements already existing under international law to which all states are bound. In addition to these sources, the article also drew on assurances in the context of expulsion for comparative purposes. Unlike assurances in most other situations, expulsion assurances have been subjected to extensive judicial scrutiny and there is considerable jurisprudence on how to assess them. This practice has resulted in a highly developed list of factors to consider in determining whether assurances change the assessment of predictable future risk to the person. Because they perform the same function of predicting risk, these factors are also informative for assessing assurances of arms use that are similarly designed to change an assessment of predictable risk. Following all of these sources, the article proposed a multi-factor due diligence analysis for assurances in respect of arms use.
Having concluded that international law also provides guidance on how to assess arms export assurances, the article turned to the case study of the recent US State Department NSM-20 Report and determined that the Report failed to perform a due diligence assessment for the assurances from Israel and the other states in the Report. In short, the Report was not focused on making a prediction of the likely use of arms in the light of the assurances. Instead, the Report focused almost exclusively on evidence of past practices violating international law to conclude that the partner government was likely to comply with assurances. At some point prior to issuing NSM-20, the US had sufficiently clear evidence to predict that there was a risk of misuse or else it would not have requested the assurances at all. To effectively address the concerns and comply with its obligations under the GCs, the State Department should have considered all foreseeable and relevant factors. It seems difficult to justify this Report as a good faith, due diligence examination of the risk of arms misuse. In fact, we know that internal State Department cables between the non-political departments had made a very different assessment of the assurances before the Report was modified by the political department and senior officials, and delivered to Congress.Footnote 220 At least one, non-political, State Department employee, who was a contributing author on the first draft of the Report, resigned because the Report was now incorrect and they could not support it.Footnote 221
As a result of not engaging with the necessary factors to conclude a good faith due diligence analysis, the narrowness of the Report suggests that the US may have violated the GCs, as well as domestic law, in exporting arms to Israel in reliance on its assurances. The White House stated expressly that the purpose of NSM-20 assurances was to comply with the requirements of international lawFootnote 222 and the State Department must have been well aware that it might not be completing a defensible assessment of assurances from other states. It remains to be seen whether the State Department is wilfully disregarding international rules or merely neglected to inform itself properly on the applicable international legal standards. Nonetheless, it seems quite clear that the US did not complete a thorough analysis. If the assurances were accepted without question and arms were exported to Israel for use in Gaza when there was a credible and predictable risk of misuse, then the US has incurred state responsibility for failing to ensure respect for international law.
Acknowledgements
The author thanks the organisers and participants of the 19th Annual Minerva Conference on International Humanitarian Law at the Hebrew University of Jerusalem for many thoughtful comments on an earlier draft. The article developed from a series of posts in Opinio Juris, and the author is grateful to the editors and readers there for helpful feedback. The author also thanks the Israel Law Review’s anonymous reviewers and Professor Yaël Ronen, Academic Editor, for insightful guidance and editorial assistance. Any remaining errors are the author’s own.
Funding statement
This research received no external funding.
Competing interests
The author declares none.
Open access
