Aggression and Criminal Responsibility (for Whom?)

Aggression and Criminal Responsibility (for Whom?)

[Nikola Hajdin is a doctoral candidate in international law at Stockholm University and a fellow at the Stockholm Centre for International Law and Justice]

On December the 14th, 2017, the Assembly of States Parties to the Rome Statute adopted a resolution that activates the International Criminal Court’s (ICC) jurisdiction over the crime of aggression (CoA). After the adoption of aggression amendments in 2010 in Kampala, controversy arose regarding the scope of its jurisdiction (for a brief overview on what has been happening in the last seven years, see here and here). In this post, however, I draw attention to yet another of many conundrums that the Office of the Prosecutor (OTP) will ultimately have to deal with—namely, the outer limits of criminal responsibility for the CoA.

The problem with the CoA’s leadership nature

Article 25 of the Rome Statue—labeled ‘Individual criminal responsibility’—stipulates that a person who commits a crime under the ICC purview is individually responsible and liable for punishment. Paragraph 3 of the Article echoes what has been now broadly accepted as a ‘differentiation model’ of participation, whereby perpetrators commit crimes (a) and accessories merely contribute to their commission (b-d). Or, in words of ICC Trial Chamber II, perpetrators’ conduct constitutes commission of the crime, whilst accessories’ is solely connected to the commission of someone else’s crime (see Katanga judgment, para.1384). Added in 2010, paragraph 3(bis) nevertheless reads:

In respect of the crime of aggression, the provisions of this article shall apply only to persons in a position effectively to exercise control over or to direct the political or military action of a State.

This so-called leadership clause (similarly mentioned in Article 8bis(1) as well) narrows responsibility to those in position to ‘control or direct’ state policy. This is nothing new for the CoA. Since the 1940s, aggression has been perceived as a leadership crime—a ‘supreme crime’ that may be committed only by the highest state officials. This is one of the main conceptual differences between the CoA and other crimes under the ICC’s jurisdiction. The latter are usually perpetrated by common foot soldiers and masterminded by leaders, whilst commission of the CoA is exclusively reserved to state leadership. This leadership nature adds an additional layer to the system of imputation of criminal responsibility: before the assessment of who committed and who assisted the crime in terms of Article 25, the ICC has to identify leaders that meet the requirement of ‘control or direct’ standard. This raises the question of what it means for an individual to qualify as the leadership of a state.

Different approaches to ‘control or direct’

To predict how the OTP and ICC will interpret ‘control or direct’ is challenging. Even worse, there’s no analogous jurisprudence: the leadership clause in the Nuremberg era was a lower standard ‘shape or influence’ (see High Command, p.488). An easy case for prosecution is the highest political and military state officials. A harder one is with non-immediate, civil society leaders who contribute to acts of aggression, yet do not have any formal a priori relationship with the state apparatus: e.g., business leaders whose companies sell weapons that are used in an aggressive war. (This issue was already raised by Heller in 2007.)

Here I will lay out three possible approaches to the ‘control or direct’ standard: ‘effective control’ from superior responsibility; ‘control’ from the ‘control over the crime theory’; and ‘control’ from the ‘overall control’ test.

One way to inform interpretation is to rely on the ‘effective control’ standard from the doctrine of superior responsibility (Article 28 of the Rome Statute). Effective control is defined as ‘the material ability to prevent offences or punish the offender’ (see Popović et al. Appeal Judgment, para. 1857). In a vast bureaucratic apparatus, once an aggressive policy has been devised, only a few people may prevent acts of aggression from happening. Even fewer are capable to use their authority to bring the offenders to justice. Accordingly, it is questionable whether anyone apart from the highest state leadership would be able to meet this requirement. Be that as it may, there’s nothing in the CoA definition suggesting this rather elusive standard. So there is room for a less stringent interpretation.

Another possibility is the ‘control over the crime’ theory and its definition of ‘control’. The contours of this theory may indeed give some guidance in interpreting the CoA’s leadership clause. In its variant of co-perpetration, control means the power to frustrate the commission of the crime by not performing essential tasks. Those who are assigned essential tasks have joint control over the crime (see Lubanga Confirmation Decision, para. 342).

The problem with this theory, however, is that its normative nature is different from ‘control or direct’. The ICC has developed the ‘control over the crime’ theory in order to distinguish between principals who commit the crime (by virtue of their control over it) and accessories that merely participate in the commission of someone else’s crime. In this vein, principal responsibility is ascribed to direct perpetrators — those who carry out material elements of the crime (‘with blood on their hands’) — and to masterminds who are often far behind the scenes — those who devise a criminal policy that sets in motion the chain of events leading towards commission of the crime (‘without blood on their hands’). By contrast, the CoA’s ‘control or direct’ standard purports to limit both perpetratorship and accessoryship to the policy level. Only when the policy level leaders are identified could the ‘control over the crime’ theory be applied to assign responsibility. But if the ‘control over the crime’ theory is used to determine leadership for the CoA, the same standard could not be re-applied to distinguish between perpetrators and accessories, because those in ‘control’ pursuant to the ‘control over the crime’ theory are regarded as perpetrators. Consequently, all participants in the crime of aggression would be labeled as perpetrators. This outcome might be normatively satisfactory to some (proponents of the unitary participation model), but it would be incoherent under the differentiation model of individual criminal responsibility. There needs to be room for accessoryship as well in the CoA, as Article 25(3bis) suggests, and therefore the standard of ‘control’ has to be different from the ‘control over the crime’ theory.

The third possibility might be to define ‘control’ as equivalent to ‘overall control’ test set forth in Tadić. In an article published earlier this year, I suggested this approach, whereby both the immediate and non-immediate state leadership are covered by the current definition of the CoA. This standard was examined in the context of responsibility of a state for the conduct of private individuals (or groups), rather than the responsibility of individuals for the action of a state. That notwithstanding, the analysis could be useful for the CoA as it identifies the position of (overall) control. However, the rationale behind the interpretation goes in the opposite direction, focusing on the individual and his degree of control over state policy. Based on jurisprudence and scholarship, I conclude that as long as a person has an ‘impact’ on the policy, he should satisfy the ‘control or direct’ requirement. I acknowledge, however, that ‘impact’ is not a very precise term and as such is not easy to define.

It bears stressing that this approach does not overly broaden criminal responsibility for aggression. Current doctrine requires the mens rea of dolus directus, which is already cumbersome for the prosecution. Specifically, the OTP has to prove that the perpetrator acted purposefully, i.e. he was aware of his position of ‘control’ and consequences of his deeds, and he consciously chose to engage in the conduct or cause a result. Only those who acted in this way can be prosecuted for CoA.

In short, once leaders are identified, the ‘control over the crime’ theory should be used to distinguish between perpetrators and accessories.


Once seen as a ‘supreme crime’, today the CoA has a different social value. Experience from the international criminal tribunals shows that other crimes have the capacity to stigmatize the highest state leaders for their misdeeds. Certainly, the ICC’s CoA will not stop violence between states. It is, moreover, an open question when we will see the first aggressors sitting in the dock in The Hague. Nonetheless, it should be clear that the CoA is not reserved for high political and military leaders. The ICC or any other court should bear this in mind when interpreting the leadership standard. Civil society leaders, such as industrialists, ought to be prosecuted as long as their impact on the governmental policy is substantial and they meet the actus reus and mens rea of the CoA. Every criminal law system in the world, as peculiar as it can be, recognizes the responsibility of those who furnish the murder weapon. The crime of aggression should be no exception.

Print Friendly, PDF & Email
Featured, International Criminal Law
No Comments

Sorry, the comment form is closed at this time.