The Inadequacy of the US State Department Report on Arms Exports Assurances: Part I

[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]

After a short delay, the US State Department delivered its first report to Congress under National Security Memorandum 20 (“NSM-20”) providing its analysis of assurances that exported arms will not be misused. This post in two parts will review the implementation of NSM-20 by the State Department and its approach in its Report, and it will determine whether the US Department of State correctly assessed these assurances under any relevant international standards. In short, this post concludes that the State Department did not complete a thorough or convincing analysis of whether the assurances were credible and reliable for a number of reasons, and its failure suggests that, as a result, the US may not be in compliance with its international obligations to prevent the misuse of exported arms.

The underlying reason for NSM-20 and assurances is to operationalize multiple obligations in international and domestic law, such as the Geneva Conventions and Genocide Convention, to prevent serious violations  of international law by other states using arms exported from the US. For a more detailed analysis of the origins of these obligations, the reader can consult my earlier post on export assurances and those of many others. NSM-20 provides a means by which the US may export arms to states where there is a discernible risk of misuse by asking those states to give assurances that they will not misuse those arms. Essentially, the state can make a promise. If the assurances are credible and reliable, then, in the view of the US Government, any risk of misuse could be reduced in light of the promise, and those international obligations prohibiting export might not be implicated.

Following the issuance of NSM-20, the US State Department identified those states receiving arms exports that might be misusing them and called for assurances. It then reported to Congress on its assessment of those assurances. Much of the excellent commentary since the report was filed has focused on whether the reporting on past practice of international humanitarian law compliance and use of US arms was accurate. However, even if the reporting on past practice was accurate, there is another aspect of the analysis. The past does not necessarily predict the future. The question for this first part of the post is whether the US State Department performed an appropriate assessment of the assurances it received by applying an analytical framework of the foreseeable and relevant factors that international practice suggests would lead to a reliable prediction of risk in the future. The second part of this post critiques the evidence that was used in making this prediction. Ultimately, if the US failed to undertake a good faith assessment of future risk, then it might mean that the US was not complying with its obligations to ensure respect for the Geneva Conventions and Genocide Convention.

The Proposed Factors to be Used in Making a Prediction of Risk 

As the Court of Appeal in the Netherlands has recognized, the assessment of arms export under international law requires a prediction of the likelihood of misuse. There are disagreements over whether the standard is knowledge or merely awareness of a serious danger that the arms would or might be misused. In any event, both standards entail an identification of what may happen in the future with some degree of confidence, though this post will not explore the precise degree of risk. In his post, Blaha noted that the US focuses on past practice in its application of domestic arms export regulations, not future risk. But this approach is correct in regards to the Leahy law which specifically punishes states for poor past practice. International law on export, however, focuses not on punishment for bad behavior, but prevention of future risk. It may be that in assessing assurances under NSM-20, the US State Department has incorrectly borrowed its analytical framework from its Leahy vetting.

The question for this post is what particular factors states must assess in making this prediction of risk and whether the US applied the correct approach.

The fact that assurances were sought under NSM-20 does not, in itself, necessarily lead to any definitive conclusion that there is a risk of misuse. For comparison, in situations of expulsion assurances, some authorities have asserted that the very fact that assurances were requested is itself evidence that there were concerns about risk. But, other authorities have refuted that assumption, because assurances may be requested “out of an abundance of caution.” (See e.g. the Canadian Supreme Court in Badesha.) That being said, in some cases, the US had already demanded assurances of use even before NSM-20. After NSM-20, the US had to return again to the state to seek additional assurances, because the first round were not sufficient. If the first round was insufficient, one must wonder whether a second round would be sufficient, unless the assurances were considerably enhanced. As of now, the fact that assurances were requested demonstrates that the US had credible indications that violations of international law might predictably occur in the future, so this concern, along with other factors, might lead to a conclusion that there is risk, but the request alone does not establish risk.

The State Department Report correctly notes that “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 result, the State Department determined that assurances must be provided in writing. This appears to be a threshold requirement in the Report and is in alignment with the requirement under NSM-20. It also stated that assurances would be assessed for credibility and reliability based on:

• 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.” Whether the person has such authority will be assessed by their “position, responsibilities, and authority … in relation to the subject matter of the assurances”. In the view of State, “Minister-level officials from the relevant ministry or above would be appropriate in most circumstances.”

• “Whether the individual providing the assurances is understood to be credible in doing so”

• “The likelihood that the partner government will comply with both assurances”. This factor would be measured “based on past practice.”

While the requirement that the assurances be given in writing appears to be a threshold requirement, the other items appear to be factors, collectively weighed.

The State Department also noted that there may be other criteria for assessment, but did not necessarily list them in the Report (“Assessment of the credibility and reliability of these assurances is based on consideration of the following factors, among others”). It would have been helpful to know these factors for the sake of transparency and to provide the international community with confidence that the US was in compliance with its due diligence obligations.

In my prior post, I suggested a slightly different framework for assessing assurances. Specifically:

• “Assurances will be more persuasive and instill greater confidence if they are publicly available …, 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.” 

The basis for this framework came from two sources: (1) international law and its interpretation by courts and other authorities, and (2) state practice making predictions of behavior from assurances generally. The reader is invited to consider the prior post for a more detailed explanation of how these two sources result in the list of factors above. 

The US State Department’s Report proposed an approach that partially deviates from these factors. It misses key considerations and glosses over important distinctions in making a reliable prediction of future behavior from assurances. None of the four inquiries for each state in the Report address the underlying issues of predicting risk of misuse other than to document past practice. They consider past practice of arms misuse but not compliance with past assurances (with the sole exception of the delivery of humanitarian aid in the case of Colombia). The fact that the US omitted clearly relevant factors in its analysis already suggests non-compliance with its international obligations.

The Factors Actually Used in Making a Prediction of Risk 

Having identified those criteria, albeit vague and incomplete, the State Department proceeded to largely ignore them in its written analysis. For each state’s assurances under scrutiny, the Report instead only considered four questions, rather than all of the factors it proposed. The first question in each state’s section was an “[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…” The second question covered the same content, but concerned “best practices for mitigating civilian harm” and implementation efforts. The third provided a description of deliveries that reached unintended recipient states to be used for unintended purposes. The final question for each state focused on impediments to the delivery of foreign aid, which this post will not address.

The Report states that the position of the person issuing the assurances must be such that they have authority to do so. However, the Report says nothing about whether the assurances will be assessed on whether they are binding under international and/or under domestic law. The Report notes that the State Department will consider “[w]hether the individual providing the assurances is understood to be credible in doing so”. Perhaps this is a diplomatic way of asking whether the state official giving the assurances is lying. The Department of State does not say what this assessment entails. 

As for content, the State Department did not evaluate the quality of the content of the assurances. Although the position of the person giving the assurances must be such that they have the authority to do so, the Report does not demand that they actually make any specific, substantive commitments. This distinction is important because assurances might merely contain statements of fact (e.g. confirmation of which treaties the state has adhered to) or it might contain a promise of future behavior (e.g. a pledge that the state will comply with those treaties). 

The Report states that the Department of State will consider the “likelihood that the partner government will comply with … assurances based on past practice.” What is unclear with this point is the level of specificity. Is the State Department looking at any documented record of the state refusing to abide by its assurances or other promises in the past? (Recall that the US required assurances long before NSM-20 and makes no mention of this in the report.) Or is it considering whether the state has ever previously used exported arms to violate international law generally? Or is it examining whether the state is known to have a poor attitude to (or eccentric interpretation of) international law generally? 

Furthermore, the State Department does not explain why it can rely on past practice as the primary consideration in predicting future misuse. In R (Al-Haq) v. Secretary of State for Business and Trade, the UK Government correctly submitted that “the mere fact of prior breaches does not equate to a risk of future breaches”. This is critical, especially if there has been a change of government or long periods since the cited practice. Thus, relying heavily on evidence of past breaches cannot constitute a full evaluation of likely future behavior. Past practice must be taken in context, and it must be analyzed for whether it has any predictive value. For example, the Report omits any mention of whether there are any reasons to expect that the various states’ governments and/or militaries would fail to comply with the assurances despite having given them.

The Report largely avoids explicitly mentioning 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. As we have seen, numerous authorities have opined that the use of large bombs has no place in attacking civilian areas because of the inherent likelihood of indiscriminate harm. It appears that the nature and type of weapons has been considered in internal State Department evaluations, because the Biden Administration specifically identified 2,000-pound bombs in its limited pause on those exports. In implementing this export hold, the US signaled that it was aware of the connection between the nature and type of certain weapons and their high potential for being used to violate international law. However, the Report says nothing about such considerations. It only vaguely refers to certain previously exported weapons that it presumes were used in violations of international law. It draws little connection between the types of arms that may be exported and their likely use in the future. All of these factors would be necessary in order to determine whether the specific commitments, if any, in the assurances actually addressed the specific risks of misusing the kinds of weapons that will be exported.

The Report also fails to mention any existence of independent verification or investigation mechanisms and whether the specific requirements of international humanitarian law and human rights law are unlawful under domestic law. We know that investigations do sometimes happen, yet the Report fails to evaluate the reliability and quality of such investigations. Given that in some cases, media has not been tolerated, one can presume that independent fact finding is limited.

Lastly, the Report does not mention the potential for legal, political, diplomatic or other influence over the state receiving exported arms that might persuade the state not to misuse them. As noted above, Blaha’s post observes that State Department’s Leahy law vetting specifically excludes political considerations from its analysis of past practice. Yet, for compliance with its international law obligations which focuses on future risk, political influence would be critical. This factor will differ from state to state considerably. The US obviously has differing relationships with all of the states whose assurances were under consideration. One would imagine that states that are significantly dependent on military aid, or have extensive contacts, might be more likely to comply with their assurances. The Report does not mention or attempt any discussion over the possibility of influence, or steps that the US will take, to compel the states to comply with their assurances.

Thus, the list of foreseeable factors that the State Department could and should use in making a prediction of risk from the assurances misses multiple critical lines of inquiry. It concentrates on just a few superficial criteria, and overlooks several key points and the overall context. Certainly, the State Department may have taken these factors into consideration in its internal, confidential evaluations and not mentioned them in its report to Congress, but the report lacks clarity and transparency in this regard. The fact that the State Department chose to only ask four brief, vaguely-phrased, and backwards-looking questions before making predictions on the likelihood of misuse in the future already suggests that it did not structure the Report’s analysis to reach a determination in good faith.

In the next part, I will examine the evidence that the US State Department considered in making its assessment of the assurances, and its failure to complete a good faith analysis of available evidence that arms were likely to be misused.