The Netherlands Appeals Court Order on F-35 Parts Delivery to Israel


| Feb 27, 2024

F-35 Israel Netherlands

A domestic court judgment in the Netherlands recently captured global headlines for ordering the government to suspend participation in a multinational program supplying F-35 parts to Israel. The Court of Appeal decision reversed an earlier judgment rendered by the District Court refusing to grant a request to order the suspension (an unofficial translation is available here). The same day the Court of Appeal judgment was published, the government announced it will appeal the order to the Supreme Court.

This post addresses two fundamental errors that render the Court of Appeal decision deficient as a matter of law. First, the decision fails to persuasively determine whether there is a clear risk that Israel is committing serious violations of international law during the ongoing conflict in Gaza. Second, the judgment’s reasoning is drawn exclusively from subsidiary sources of customary international law, without addressing the primary sources of law that compel a contrary ruling.

The judgment does not address the recent provisional measures order published by the International Court of Justice (ICJ) in the dispute between South Africa and Israel, so this post likewise does not analyze the provisional measures order. The ICJ order merely maintains the status quo as it existed before the order was published. As such, the popular suggestion that States must take steps to ensure Israel complies with the order is mistaken since reinforcing the existing status quo does not create additional obligations.

The Court of Appeal decision raises several other important issues. These include controversy regarding an ostensive external dimension to the obligation “to respect and to ensure respect” for the 1949 Geneva Conventions reflected in Common Article 1 (CA 1) of the treaties as well as the propriety of the judiciary taking an active role in shaping foreign policy.

Though these issues warrant consideration, determining there is a clear risk that Israel is committing serious violations of the law of armed conflict (LOAC) is decisive for the Court of Appeal ruling. Fundamental errors committed by the court in support of that finding will need to be addressed on appeal to the Supreme Court, so this is the primary focus of the present inquiry.

Regulatory and Legal Framework Applied by the Government

The Ministry of Foreign Affairs has developed a general permit regulation to establish the procedure for the supply of military goods to parties, including Israel, that are affiliated with approved agreements in the context of the F-35 Lightning II program. Exports to Israel pursuant to this general license are effectively suspended as of February 19, which is seven days from the date of the Court of Appeal order.

The court ultimately determined that continuing to deliver F-35 parts to Israel pursuant to the export license violates three separate but related international legal obligations (para. 5.51). The first is an agreement known as the “Common Position” adopted by the European Council defining rules governing control of exports of military technology and equipment. This document is the European Union Global Strategy, referred to by the acronym EUGS throughout the decision. Second is the Arms Trade Treaty (ATT), which has been accepted by the Netherlands. Third is the ostensive external dimension of CA 1.

The court found that the ostensive external component of the CA 1 obligation to “ensure respect” must clearly be interpreted in such a manner that there is no conflict between CA 1 and relevant provisions of the EUGS and ATT, both of which regulate Dutch shipments of F-35 parts (para. 5.25).

In an article published in 2020, Professors Michael Schmitt and Sean Watts persuasively (and, I assess, correctly) contend that CA 1 includes no external component. Even so, States such as the Netherlands that are bound by the EUGS or have accepted the ATT have adopted an external obligation to “ensure respect” for LOAC in the context of arms exports as a matter of policy or of conventional law.

For States bound by the EUGS such as the Netherlands, both the ostensive external component of CA 1 and relevant provisions of the ATT can adequately be addressed by implementing the “clear risk” standard established in the EUGS. This assesses whether there is a clear risk that military technology or equipment might be used in the commission of serious LOAC violations.

As a news release published on the judiciary website announcing the decision summarizes, the Court of Appeal determined there is a “clear risk” (duidelijk risico) that Israel will commit “serious violations of the humanitarian law of war” (ernstige schendingen van het humanitaire oorlogsrecht) with F-35s in the Gaza Strip. The Court of Appeal decision merges the EUGS, ATT, and CA 1 legal/regulatory obligations in this manner and applies it as a consolidated legal standard to determine whether the court should direct the government to suspend the export license as it relates to Israel.

Assessing “Clear Risk” of Serious LOAC Violations

With that legal framework in place, the court’s first fundamental error is in analyzing whether there is a “clear risk” that Israel is committing serious LOAC violations using F-35s during the conflict in Gaza. The analysis is fundamentally flawed in each of the three primary stages of the assessment: describing the standard for evaluating LOAC compliance; selecting external evidence used to support the analysis; and reaching conclusions based on the evidence. At each stage, the court erroneously applies an effects-based approach to assessing compliance with LOAC. The most conspicuous fundamental errors are addressed in turn below.

Developing Effects-Based Legal Standards

In section 5.b.1. (paras. 5.6-5.9), the decision articulates relevant law of war obligations. It is reasonable to dispute whether Israel is involved in an international or a non-international armed conflict (NIAC) against Hamas and, if it is the latter, whether each LOAC rule addressed by the decision actually applies to the current conflict. While I agree with Professor Michael Schmitt’s conclusion that this is a NIAC, I will assume for the sake of a comprehensive analysis that all the LOAC rules described in the judgment actually apply to the conflict. Although Israel has not ratified the Rome Statute, the Netherlands has accepted the treaty. As such, the text of the Rome Statute is adopted as a standard against which to evaluate the rules articulated by the Court of Appeal decision.

Regarding the LOAC discrimination rule, the court correctly observes that attacks that fail to distinguish between military objectives and civilian objects are prohibited (para. 5.9). For support, the decision also refers to Article 51(4) of Additional Protocol I to the 1949 Geneva Conventions (AP I).

However, the assertion that civilian targets should not be attacked (aangevallen) is erroneous. The relevant text of AP I describes as “indiscriminate” those attacks “which are not directed at a specific military objective” (emphasis added). The rule described by the court is centered on the outcome of an attack, while the text of AP I correctly focuses on the process of an attack.

To operationalize this standard, the Rome Statute formulates this war crime as “[i]ntentionally directing” attacks against the civilian population or civilian objects (emphasis added). Evaluating this offense requires a determination that the attacker was aware the attack would be directed against civilian persons or objects and deliberately engaged in the attack. This is fundamentally at odds with the standard used by the court and constitutes an ex post determination that civilians were attacked rather than whether the attack was intentionally directed against civilians.

The same effects-based error is adopted by the court when articulating the LOAC proportionality rule. According to the judgment, if collateral damage is excessive in relation to the concrete and direct military advantage anticipated, the attack cannot take place (para. 5.9). Again, the decision cites a provision of AP I. And again, the court impermissibly converts the doctrinal rule into an effects-based standard.

The relevant text of AP I prohibits an attack that may be expected to cause incidental damage that is excessive in relation to the concrete and direct military advantage anticipated. The Rome Statute operationalizes the LOAC proportionality rule by proscribing “[i]ntentionally launching an attack in the knowledge that such attack will cause [collateral damage] which would be clearly excessive in relation to the concrete and direct overall military advantage anticipated” (emphasis added).

By professing that an attack must not proceed if the collateral damage is excessive in relation to the concrete and direct military advantage anticipated, the standard articulated by the court is centered on the outcome of an attack. This is inconsistent with the doctrinal rules established in AP I and the Rome Statute.

Finally, the Court of Appeal order also asserts that attacking “agricultural areas” and “drinking water installations” is forbidden (para. 5.9). This is only partially accurate. Here, the decision cites Article 54(2) of AP I. This provision does indeed prohibit attacks on objects (such as agricultural areas and drinking water installations) that are indispensable to the survival of the civilian population. However, the attacks must be conducted “for the specific purpose of denying them for their sustenance value to the civilian population” (emphasis added). The Rome Statute operationalizes this rule by prohibiting[i]ntentionally using starvation of civilians as a method of warfare by depriving them of objects indispensable to their survival” (emphasis added).

By omitting the specific intent components of the relevant AP I and Rome Statute provisions, the standard developed by the court allows for another effects-based assessment of LOAC compliance. Any attack(aanvallen) on agricultural areas or drinking water installations can be described as a serious LOAC violation even in the absence of evidence regarding the requisite mental element. This is inconsistent with existing doctrinal LOAC rules reflected in AP I and the Rome Statute.

Errors in Evidence

The Court of Appeal amplifies these errors when developing the factual foundation upon which to support the “clear risk” determination by selecting sources that also utilize flawed effects-based standards. The objective of the present inquiry is not to engage in a detailed analysis of each source upon which the court relies to support its conclusions. Instead, a sampling of these sources is assessed here for illustrative purposes.

Section a. of paragraph 5.11 cites two separate reports published by Amnesty International: a “Damning Evidence” report from October 2023; and a “US-made Munitions” report from December 2023. The judgment notes that both reports conclude from on-site investigations that civilian targets have been attacked or indiscriminate attacks have occurred and, as such, these incidents should be investigated as war crimes.

However, both reports from Amnesty International adopt the same flawed methodology to reach these conclusions. For example, the “Damning Evidence” report relates, “The Israeli army claims it only attacks military targets, but in a number of cases Amnesty International found no evidence of the presence of fighters or other military objectives in the vicinity at the time of the attacks.”

Reportedly failing to find evidence of the presence of fighters or other military objects at the time of an attack during a post-strike investigation provides little, if any, information regarding the knowledge and intent of those responsible for an airstrike before and during the attack. This post-strike evidence, then, is not an adequate basis upon which to conclude the attacks should be investigated as war crimes.

The “US-made Munitions” report adopts the same erroneous effects-based methodology. According to the report, “Amnesty International did not find any indication that there were any military objectives at the sites of the two strikes or that people in the buildings were legitimate military targets, raising concerns that these strikes were direct attacks on civilians.” This report also suggests “the use of explosive weapons with wide-area effects in such densely populated areas could make these indiscriminate attacks.” However, as I have previously indicated while analyzing this conclusion, the “use of explosive weapons in a densely populated area is not alone a conclusive factor” in determining whether an attack is “indiscriminate” and therefore should be investigated as a war crime.

The remainder of the evidence considered by the court in para. 5.11 exhibits similarly flawed methodology. For example, instead of conducting a doctrinal legal analysis, the court relies upon a press statement from a UN special rapporteur observing that the “healthcare infrastructure in the Gaza strip has been completely obliterated” and that an “‘immeasurable’ number of violations of the special protection afforded to civilians, children, and medical personnel under international humanitarian law” have been committed in Gaza (para. 5.11b).

Both of these assertions are based on observations regarding the effects of attacks, whereas evaluating whether a “clear risk” exists that Israel is committing LOAC violations with F-35s in Gaza requires an assessment regarding the “intent and knowledge” of the personnel responsible for each attack analyzed. It is true that UN Security Council Resolution 2712 (2023) expresses “deep concern” about the “disproportionate effect” on children of the conflict in Gaza and demands “that all parties comply with their obligations under international law” (emphasis added), but at no point does the resolution conclude or even suggest that Israel is failing to do so.

The decision also relies upon separate press releases from two different groups of UN special rapporteurs (here and here, paras. 5.11d and 5.11e). These engage in an effects-based analysis of Israel’s responsibility for the destruction in Gaza while articulating an offense, domicide, that does not exist in international law. Similar foundational deficiencies are discernible from each source upon which the decision relies to develop the factual basis for the court’s conclusions.

The judgment conveys that the findings of non-governmental organizations, such as Amnesty International, must be taken “extremely seriously” (uiterst serieus), especially when these involve potential LOAC violations (para. 5.12). Likewise, the Court of Appeal expresses in the judgment that it assumes special rapporteurs should be considered “very capable” (zeer wel in staat) of assessing whether civilian targets have been attacked in Gaza and whether the damage inflicted is proportionate (para. 5.13).

None of the sources relied upon by the Court of Appeal engage in a doctrinal analysis of LOAC compliance, for which direct information related to knowledge and intent is an absolute requirement. Instead, each source determines solely from observable effects that serious violations have been committed by Israel. Based on these extensive errors in both the standards articulated by the court and the evidence developed to apply to these standards, the conclusions reached by the Court of Appeal decision are predictably flawed.

Conclusions Reached in Support of Order to Suspend Export License

Several concerns can be raised by the findings in Section b.4. concluding a “clear risk” exists that Israel is committing LOAC violations using F-35s in Gaza. For present purposes, the one decisive passage (para. 5.17) is the focus.

The decision recalls evidence that Israel has caused a large number of civilian casualties during the conflict before describing specific examples at length. Then, the court finds it is “unlikely” (niet aannemelijk) the extent of the destruction in Gaza was inflicted only on military targets or represents legitimate “collateral damage.”

With this finding, the decision again applies an effects-based analysis to Israel’s entire military campaign in Gaza to conclude a “clear risk” of LOAC violations exists. In the first sentence of the subsequent paragraph (para. 5.18), the court finds support in the conclusions of Amnesty International and various experts affiliated with the United Nations.

This seems reasonable; a court implementing a non-doctrinal analysis to support effects-based conclusions may well feel supported by sources that do the same. However, neither the court nor the sources upon which it relies develop adequate evidence, doctrinal standards, or valid conclusions to adequately support the “clear risk” finding.

Earlier in the judgment (para. 5.16), the court acknowledges that a final judicial decision will require a careful factual investigation that examines what information a commander had available at the time of an attack to determine whether LOAC has indeed been violated. This is undoubtedly correct, but the assertion by the court that a “clear risk” determination does not require the same information is mistaken. Behavior of the recipient country (here, Israel) in the “recent past” (recente verleden) is certainly relevant for evaluating potential “future use” (toekomstig gebruik) of weapons systems such as the F-35.

However, concluding a “clear risk” of LOAC violations exists requires a doctrinal LOAC analysis even when evaluating conduct in the recent past. A doctrinal LOAC analysis requires a careful factual investigation that examines the information available to the decision-maker at the time of an attack, whether that analysis is utilized to support a final judicial decision or a “clear risk” determination.

Excluding Analysis of Primary Sources of Customary International Law

The final significant error reflected in the Court of Appeal decision to be considered here involves the sources of customary international law considered by the court. None of the significant evidence relied upon by the court, whether reports from Amnesty International or expressions from experts affiliated with the United Nations, or media reports and the like, constitutes primary sources of customary international law.

The persuasive report (with commentaries) published by the International Law Commission (ILC) regarding identification of customary international law presents detailed information regarding the distinction between primary and subsidiary sources of customary law. As the ILC report notes, the teachings of the “most highly qualified publicists . . . may serve as a subsidiary means for the determination of rules of customary international law.”

Assuming the Court of Appeal finds Amnesty International and various experts affiliated with the United Nations to qualify for designation as the “most highly qualified publicists,” then the “teachings” of these sources can, at best, qualify as a subsidiary source of customary law.

As the ILC report also conveys, it is primarily “the practice of States that contributes to the formation, or expression, of rules of customary international law.” The report also describes what qualifies as the practice of States, and the commentaries explain the process for identifying State practice in greater detail.

It is certainly true that each State is responsible for making an independent determination regarding whether a clear risk exists that a recipient of security assistance will use the defense articles received to commit LOAC violations. However, when canvassing sources of customary international law to support that determination, the foundation for this inquiry should be the practice of other States.

In this case, if a few governments have withheld security assistance to Israel over concerns that serious LOAC violations will be committed with the defense articles received, this is certainly some evidence of the practice of States. If nearly every government that provides security assistance to Israel has halted the supply of weapons and expressed a legal obligation to do so, this would constitute persuasive evidence indicating that no country should be allowed to provide security assistance to Israel.

The Court of Appeal acknowledges that all other countries providing security assistance to Israel are responsible for making their own determinations regarding compliance with international law (para. 5.45). However, the decision does not survey the practice of other governments in this regard. Instead, the decision relies extensively on subsidiary sources of customary international law—to the complete exclusion of the practice of other States involved in providing security assistance to Israel as a primary source of customary law—to support the court’s conclusions. This is another fundamental flaw reflected in the analysis and conclusions rendered by the Court of Appeal.


As the conflict in Gaza rages on and debate continues regarding the advisability and permissibility of continuing to supply weapons to Israel, there are many important considerations to weigh. Separating policy preferences from legal requirements and fostering an accurate understanding of the doctrinal application of actual legal obligations is vitally important for encouraging a productive and balanced debate.

Due to extensive methodological flaws, the decision by the Court of Appeal suspending the export license for F-35 parts by the government of the Netherlands to Israel incorrectly analyzes and applies relevant aspects of international law. In this regard, the judicial order more closely resembles a preferred policy outcome rather than a doctrinal legal analysis.

The Supreme Court in the Netherlands must identify and account for the pervasive errors inherent in the methodology, legal analysis, and conclusions that contributed to the Court of Appeal decision ordering the government to suspend the export license for F-35 parts to Israel.

The factual record considered by the Court of Appeal unquestionably supports a determination that a clear risk of harm to civilians in Gaza exists for arms—including F-35 parts—that are exported to Israel. However, even extensive civilian harm is not a reliable proxy for evaluating compliance with relevant rules of international law since this data is derived solely from effects-based observations and sources.

There is no substitute for developing and implementing a balanced, doctrinal analysis of relevant aspects of international law. This is true even when merely applying the “clear risk” standard. In this regard, the analysis and conclusions of the Court of Appeal are irreparably flawed.


Brian L. Cox is a doctoral candidate lecturer and J.S.D. candidate at Cornell Law School, a visiting scholar at the University of Ottawa Faculty of Law (Common Law Section) in Ontario, and a retired U.S. Army judge advocate.



Photo credit: William Lewis