Can We Tell If the ICC Can Deter Atrocity?

Kate Cronin-Furman is a postdoctoral fellow in Law & International Security at Stanford University’s Center for International Security and Cooperation. She is the author of Managing Expectations: International Criminal Trials and the Prospects for Deterrence of Mass Atrocities.


The International Criminal Court opened its first investigations in 2004. In its first 12 years of operation, the court convicted two individuals of war crimes and crimes against humanity, and issued arrest warrants or summonses for 37 others. Today, the ICC is pursuing prosecutions of atrocities on the territory of eight countries and conducting preliminary examinations in seven more.

Most immediately, the goal of these efforts is to punish those responsible for egregious breaches of international law. But the bigger purpose of the court’s existence is to contribute to international peace and security. Underpinning this aim is the hope that by prosecuting the perpetrators of serious international crimes, the ICC can make mass atrocities rarer. In short, that it can deter this type of violence.

“Can the International Criminal Court Deter Atrocity?” by Hyeran Jo and Beth Simmons is one of the first rigorous empirical examinations of this claim. For skeptics (like me) of the ICC’s ability to produce deterrence, the article’s conclusions may come as something of a surprise. The authors find that state actors commit fewer intentional killings of civilians in conflict in the presence of ICC ratification, implementing statutes in domestic criminal law, and action by the court. They find that rebel violence also lessens in the face of ICC action, but not following ratification or legal change.

Even a limited and contingent reduction in violence against civilians would be great news for the ICC’s effectiveness. But the difference in findings across state vs. rebel perpetrators leads me to hesitate before interpreting these results as cause for optimism. The reason for this is that rebel violence seems like the best context in which to test ICC’s impact. The weak results on rebels therefore leave open the possibility that the strong findings on state actors are an artifact of selection effects.

Here is my logic: Analysis of the ICC’s effect on state actors is complicated by the fact that states choose to accept the court’s jurisdiction through ratification of the Rome Statute. For those who have elected to join up, it’s hard to know how the court has affected their behavior. Any reduction in a state’s use of illegal violence that follows accession to the ICC might be caused by the same factors that led the state to ratify—a democratic transition, a commitment to peace and justice, etc.

Jo and Simmons take heroic measures to address this issue, employing controls for dynamics that might explain both ratification and violence reduction in the main regressions and conducting a matching analysis as a robustness check. But the fact remains that selection effects are impossible to rule out. This is a general headache for researchers looking for effects of treaties, because the actors whose behavior we wish to study are the same actors who have chosen to join international legal regimes. The case of the ICC is something of an exception to this rule, though, because there is a set of actors who did not choose to join the treaty, but upon whom its effects might be expected to operate: rebel groups.

It is rebels who experience something closest to “random assignment” of ICC jurisdiction, because they do not participate in decisions to ratify or pass implementing statutes. (Although their behavior may be a causal factor in states’ decisions to join the court.) They therefore most closely approximate a scenario in which we could observe the effect, all else equal, of the ICC’s existence. Jo and Simmons’s results suggest that this effect exists, but is limited to a slight reduction in violence against civilians when the ICC has been more active (conducting preliminary examinations and investigations and issuing arrest warrants).

Could the comparatively weaker effect on rebels vs. state actors be explained by something other than selection effects? Absolutely. Jo and Simmons attribute it to the fact that rebels are “likely to be the most difficult case for ICC deterrence”. They argue that this is because they are difficult to apprehend and because they (particularly non-secessionist rebels) have fewer incentives to conform to international norms.

This may be true, and certainly there may be other characteristics of rebels that make them difficult to deter. They may be particularly likely to face the sort of “overriding interests” in committing atrocities that I identify in my 2013 IJTJ article as obstacles to deterrence. A rebel leader’s set of potential futures may look quite different from, and much shorter than, a high-ranking regime commander’s. The prospect of prosecution and imprisonment in The Hague, while unappealing, may simply be one more entry in a long list of equally unpleasant possible outcomes: death in battle, a domestic treason trial, overthrow and execution by your subordinates. ICC deterrence might therefore require a comparatively higher risk of prosecution. And as recent microfoundational work by Michael Broache on the effects of ICC action on rebel groups in Congo shows, their incentive structures are extremely complex—under certain circumstances, prosecution can even provoke them to escalate conflict.

However, there are countervailing dynamics: Evidence about rebel crimes is easier to gather because investigations proceed with the consent and assistance of the territorial state. Rebels can also be arrested more easily than state actors, who benefit from ingrained diplomatic practices of honoring immunities. This suggests that rebels, once charged, are more likely to be successfully prosecuted than state actors. Furthermore, for the first few years of the court’s existence, most of its targets were rebels, indicating that their risk of facing charges was also higher than state actors’.

It is therefore not clear that rebels are categorically less likely to be deterred by the threat of prosecution than state actors. But Jo and Simmons are theorizing broader deterrent effects of the ICC’s existence than those simply attributable to the increased risk of prosecution. They argue as well for a “social deterrence” mechanism, whereby social pressures impose costs on rule violators. This is a critical insight. The ICC (like many international institutions) has limited formal enforcement capabilities, but exists within a web of interconnected norms, institutions, and actors concerned with fighting impunity and protecting civilians. Treating it as a stand-alone institution and focusing on its independent effects on violence against civilians therefore misses most of the story of its impact on international relations.

Jo and Simmons’s account lends itself to the interpretation that prosecutorial and social deterrence are not separate mechanisms, but mutually reinforcing. Several of their findings demonstrate that extralegal mechanisms benefit from the role of the formalized legal institution as a focal point — evidence that both civil society and foreign aid reliance have stronger effects on reducing violence in the presence of ICC ratification than its absence is suggestive. The reverse should also be true, with prosecutorial deterrence operating more weakly in the absence of social deterrence, and kicking in only at higher levels of risk of prosecution. Empirically, that translates into an expectation that actors who are less integrated into the international community, and less beholden to domestic constituencies, would be less responsive to the presence of the ICC. This is likely to be more often true of (non-secessionist) rebels than state actors, perhaps explaining the varying findings across perpetrator types.

My own view is that the findings on the independent impacts of the International Criminal Court should be interpreted with caution, but that this hardly matters. The ICC is not independent of the broader normative context, and its position highlights what Jo and Simmons describe as “the central importance of extralegal deterrents to law violation”. Debates about the ICC’s impact have been too narrowly focused on the question of prosecutorial deterrence. The account of social deterrence offered here is a more nuanced approach to the question of how international institutions can affect behavior in the absence of robust formal enforcement capabilities, and lays the ground for a more productive research agenda on the ICC going forward.