ICC’s Effectiveness and the Explanatory Black Box: Deterrence or Cultural Prevention?

Joachim J. Savelsberg is a Professor of Sociology and Law and Arsham and Charlotte Ohanessian Chair at the University of Minnesota. He has authored several books on the representation and collective memory of atrocity, bridging the gap between criminology and genocide studies. His most recent book is Representing Mass Violence: Conflicting Responses to Human Rights Violations in Darfur (University of California Press).

“In the past rebel leaders would have directed me to nearby villages where they had left piles of corpses behind. They would have shown off their child soldiers. Now they know they may be held accountable. They have become more cautious.” These words, paraphrased from an interview I conducted with an Africa correspondent of a prominent European newspaper in 2011, resonate with empirical patterns revealed in Hyeran Jo and Beth A. Simmons’ paper (“Can the International Criminal Court Deter Atrocity”)? But they constitute anecdotal evidence, while Jo and Simmons fill a void of knowledge through systematic empirical investigation. The need for such data is urgent because the institutions of international criminal justice, specifically the ICC, are so new, and because we strongly desire effective intervention against those crimes over which the ICC has jurisdiction: war crimes, crimes against humanity, and genocide.

Jo and Simmons argue that the ICC and its prosecutions have a deterrent effect on both state actors and rebel leaders; that deterrence works along two lines: “prosecutorial deterrence” and “social deterrence;” and that these effects are contingent. These theses, and the evidence the authors provide, must be taken seriously. They are thorough and well presented. Also, their work is courageous: it steps beyond the traditional division of academic labor between scholars of criminal justice versus foreign policy and international relations. Doing so is an appropriate response to changing institutional realities, as Kathryn Sikkink (2011) made clear when, for good reasons, she subtitled her book on The Justice Cascade with How Human Rights Prosecutions are Changing World Politics. Appropriately, Jo and Simmons are less concerned with law on the books than with law in action, following Roscoe Pound’s early 20th century imperative and a century of law and society scholarship. Approaching the issue from the other side, law and society scholars and sociological criminologists have begun to address international relations (e.g., Hagan 2003; Savelsberg 2015).

Jo and Simmon’s paper yields crucial insights, but limits to their argument demand a reformulation of their central thesis. The authors show that the time following the introduction of the Rome Statute and the ICC and the onset of prosecution witnesses a reduction in killings by state actors, especially those who have supported the ICC and who depend on the world community. Also, rebel leaders kill less, especially those who lead secessionist movements that strive for recognition by the world community. These are important findings.

My critical commentary focuses on one issue: the causal interpretation of the statistical patterns Jo and Simmons identify. Even if the patterns observed can be attributed to ICC intervention, there is little evidence that they can be attributed to a deterrence mechanism. The authors fill the black box between court intervention and trends in violent deaths with the magic formula of “deterrence,” but they do not prove that this is indeed the causal mechanism at work.

The notion of deterrence is, of course, the first modern utilitarian justification of punishment, beyond the Kantian notion of retribution. Jeremy Bentham proposed the idea in his rational actor model, suggesting that criminal offenders weigh crime’s benefits against the costs of punishment (multiplied by the risk of detection). We know today that deterrence does work for certain offenders and under specific circumstances. More specifically, Jo and Simmons correctly cite criminological literature that attributes its effectiveness more to the certainty than to the severity of punishment. And indeed, they provide suggestive evidence that at least part of the effect they measure may result from deterrence experienced by rational actors (e.g., more reduction of killings in aid-dependent countries).

Yet, the logic of the argument is insufficient as the black box Jo and Simmons fill with “deterrence” contains at least one other, possibly more powerful, mechanism: a cultural effect of criminal justice intervention, the radical delegitimization of atrocious violence. Sikkink acknowledges this mechanism in The Justice Cascade (pp. 173f). While she takes exception to a too limited focus on collective memory, one aspect of culture, her finding that human rights records improve especially when trials are combined with truth commissions provides support for the cultural effect. Importantly, cultural transmission is – while potentially effective per se – also a precondition for deterrence to work: rational actors must know about past and current punishment before they can consider them. Prosecutions must have become part of the collective conscience, possibly of the memory, of political and military elites. Then, strategies involving mass atrocities may have disappeared from the decision tree of political and military leaders.

The cultural effectiveness of criminal justice intervention is suggested by a powerful line of neo-Durkheimian thought, and recent empirical work has demonstrated it for the ICC (Savelsberg 2015; Savelsberg and Nyseth Brehm 2015). Analyses of a comprehensive data set on Darfur resulting from content analysis of 3,400 media reports, and of interviews with experts in foreign ministries, human rights and humanitarian NGOs and Africa correspondents from eight Western countries, show that criminal justice intervention – from the International Commission of Inquiry on Darfur via UNSC decisions to ICC prosecution – produce a powerful public representation of those as criminal perpetrators who bear responsibility for the violence. Interventions keep the violence in public view, and they enhance its framing as human rights crimes and genocide. Qualitative analyses illustrate and multi-level, multivariate analyses confirm this effect for each of the eight countries, even if levels of receptivity for the criminal justice frame vary among them. Following ICC intervention (and the cumulative cultural effect of the ad hoc courts of the 1990s), powerful state leaders with responsibility for mass atrocity certainly no longer appear as heroes (as they did for much of human history), and their deeds may be less likely to be denied.

Does it matter if reduction of violence results from deterrence or cultural mechanisms? It certainly does in scholarly terms as we seek to avoid misinterpretations of empirical reality. It also matters for practice as the concern with consequences should inform how international justice institutions communicate their decisions and reasoning to the world public.

In conclusion, Jo and Simmons’ contribution is important and timely. The authors, however, do not prove deterrence to be the central causal mechanism. It would serve future scholarship (and practice) well to take seriously the cultural effect of new interventions in grave human rights crimes by criminal law and justice.


Hagan, John. 2003. Justice in the Balkans. Chicago: University of Chicago Press.

Savelsberg, Joachim J. 2015. Representing Mass Violence: Conflicting Responses to Human Rights Violations in Darfur. Oakland: University of California Press (open access online at: <http://www.luminosoa.org/site/books/detail/3/representing-mass-violence/>.

Savelsberg, Joachim J. and Hollie Nyseth Brehm. “Representing Human Rights Violations in Darfur: Global Responses, National Distinctions.” American Journal of Sociology 121(2):564-603.

Sikkink, Kathryn. 2011. The Justice Cascade: How Human Rights Prosecutions are Changing World Politics. New York: W.W. Norton.

The ICC Fails to Deter When Deterrence is Needed Most

Alette Smeulers is Professor in International Criminology at the Universities of Tilburg and Groningen. She has co-authored: International Crimes and Other Gross Human Rights Violations: a Multi- and Interdisciplinary Approach. Her other publications can be found via her website here.

The article of Jo and Simmons deserves praise: they not only address one of the most important issues in the field of international criminal justice – the alleged violence-reducing role of pursuing justice in international affairs – but more importantly they do so by conducting empirical research. On the basis of their empirical results they conclude that the ICC can potentially deter “governments and those rebel groups that seek legitimacy.” This is an important finding. In this comment I will explain what I believe the exact meaning of their finding is and whether this really means that the ICC can deter mass atrocities

Can the ICC really deter mass atrocities?

The authors suggest that the ratification of the ICC statute and the actions of the ICC have a potential deterrent effect on actors such as heads of states and to a lesser extent heads of rebel groups. Crucial is that within their analysis the authors distinguish between actors who ‘seek legitimacy’ and ‘are sensitive to social pressure’ and those who are not. The first type of actor can potentially be deterred the latter is much harder. Their conclusion seems fair and I fully agree with their analysis. I however do not believe that ICC ratification by itself can deter mass atrocities. The point is that such atrocities are the resultant of a complex combination of factors which interrelate and interact and make mass violence escalate into mass atrocities. The interrelationship between these factors is extremely complex and although it is possible to identify factors which play a role within this dynamic it is impossible to pinpoint one single deterministic cause of mass atrocities or -for that matter- pinpoint one factor which could prevent mass atrocities from being committed.

Heads of states who believe in the importance of international law and aim to become a fully accepted member of the international community will seek legitimacy by showing their adherence to the international legal framework and it can be expected that they will take several measures: ratify human rights treaties, become a party to the ICC and in the meantime they will try to refrain from violating international norms and values. The conclusion would then be that both the reduction in civilian deaths and the ratification of the ICC statute is initiated by a head of state who (starts to) take abidance by international norms seriously. In other words: I believe that the ratification of the ICC statute as well as the implementation of these norms in their own penal system are outcomes of this stance rather than the cause thereof. These actors are much more deterred by the fact that committing mass atrocities is prohibited in international law than by their ratification of the ICC statute. I would suggest that ratification of the ICC statute is a sign of their adherence to international law. I do agree however that the deterrence following from the ratification adds to the deterrence which results from the existence of these norms and values within international law but would suggest that the role of the ICC is secondary rather than a prime deterrent.

This is an important finding but none the less we should not be too enthusiastic about this as the research of Jo and Simmons also seems to show that deterrence does not work in those cases in which it is needed most as I will explain in the following.

When the ICC fails to deter…

The ICC aims to focus on the most serious crimes of concern to the international community and it is precisely in these most extreme cases that it is most likely to fail to deter the perpetrators. The most extreme crimes are committed by ruthless dictators who do not care about the international legal order or their own legitimacy and they are much less likely to be deterred by the ICC (or any other international institution for that matter). Their prime focus is to gain or maintain power by whatever means including – if necessary – violent or genocidal policies. They are power hungry and ruthless in their struggle to stay in power and their survival instincts will make them focus merely on the (alleged) danger to their lives rather than the danger to their reputation or the possibility of at some point being prosecuted for their crimes. Besides, quite a few power hungry and ruthless dictators start to suffer from megalomania once in power. They feel they have superior divine-like powers and believe that they have been chosen to lead the people in their country. They often tend to believe their authority is superior to any man-made laws and they often seem to feel that the norms and values of the international community do not apply to them. They consider themselves to be above the law and above state like institutions. They thus do not care about the international legal framework nor do they believe that they have to abide by it. This type of dictator cannot be deterred by some kind of institution such as the ICC which in their eyes is inferior to them anyway. The ICC thus unfortunately fails to deter actors in those situations we would need deterrence most.

Another group of potential perpetrators who will not be deterred by the ICC are the middle and low ranking perpetrators who commit so-called crimes of obedience. They will not be deterred by the ICC for the very simple reason that the social context in which they operate is too overwhelming. There is usually a tremendous pressure to obey orders, to conform to the group and to do as they are told in order to protect their people and country. For these low and middle ranking perpetrators the ICC and whether or not their state has ratified the Rome Statue will not affect them in any useful way simply because the ICC as an international institution is too far outside of their reach to exert any influence on their behaviour. The danger of being punished for not obeying an order is more direct and specific than the far-fetched possibility of being prosecuted by the ICC at some time in the far future.

The deterrent effect of the ICC is thus very limited and in fact only deters actors who already aim to abide by international legal norms and values. This is however still a positive outcome as together with the legal framework it enforces, the ICC does seem to play an important (supporting) role in making sure heads of state abide by the law.


Deterrence of and Through Other Actors

As I mentioned in my earlier introduction to their groundbreaking piece, I believe Professors Jo and Simmon’s article (available here) is exceptionally important. In particular, I very much appreciate their addition of new theoretical nuance and empirical insight to the question of deterrence in international criminal law. My reactions are less a critique and more a set of pointers about other avenues through which empiricists (these or others) might think about measuring the role of international criminal justice in deterring atrocities in the future work they call for.  As with other aspects of my research, I am interested in the role of business in this regard, which is not a topic that is directly broached in this excellent paper. I briefly demonstrate the advantages of reorienting our thinking about deterrence towards economic actors as well as one particular danger this shift could entail.

Professors Jo and Simmons are rightly sensitive to the differences “type of actor” might generate for an assessment of the ICC’s deterrence. Astutely, they disaggregate states from rebel groups, then rebel groups with secessionist aspirations from those without. Likewise, in recognizing that deterrence might not operate uniformly across all international crimes, they wisely limit their project to a single international crime: intentional killing of civilians. In light of these limitations, they “encourage further research into a range of heinous crimes – from sexual violence to trafficking in children to widespread pillaging – that the ICC was meant to address.”

If this further research comes to pass, I would recommend: (a) further disaggregating the types of actors it focuses on beyond just states and armed groups, and (b) moving beyond the single crime model to assess the extent to which deterring some international crimes can ratchet up the deterrence of others.

Before I get to these arguments, I pause to reiterate a fact I hope is widely accepted, namely, that State actors and rebel groups are not the only agents implicated in atrocities—businesspeople and the corporations they represent are often instigators, masterminds and accomplices, too. I insist on these various forms of participation in deliberate opposition to a widespread but I think unfortunate perception that business invariably plays a role that is peripheral or auxiliary to mass violence. As others have shown (see infra), even the Nuremberg Judgment recognized that the most powerful German “industrialists” signed a petition calling on President Hindenburg to appoint Adolf Hitler Chancellor. In fact, Jonathan Bush has argued that a member of the company IG Farben paid a substantial bribe to facilitate that end. More recently, several modern cases in Africa also involve businesses at the helm of terrible bloodshed, not complicit in it.

Given this reality, it is curious that much of the literature on deterrence of atrocity to date has left business out, arguing that any rational incentive generated by criminal law is unlikely to restrain the fierce passion required to perpetrate offenses of this barbarity, particularly when the probability of prosecution is so low. Yet, as I have argued elsewhere, corporations and their representatives dispassionately pursuing profit rather than historical grievances, inter-ethnic rivalries or military control over capital cities also satisfy the formal elements of international crimes in certain circumstances. And importantly, the transnational corporations that sustain bloodshed are more exposed to foreign law enforcement, more prone to rational deliberation through their commitment to profit maximization, and likely to perceive conviction for a war crime as nothing short of a commercial catastrophe. Thus, they may be more easily deterred than the armed groups Jo and Simmons focus on.

Corporate offending should itself be deterred, but focusing on businesses may also have important trickle-down effects for the deterrence of armed groups Jo and Simmons address. In a recent debate about impunity staged by the International Center on Transitional Justice, I argued that prosecuting the arms vendors who provide weapons to notoriously brutal armed groups as accomplices may, in appropriate cases, be a way of incentivizing greater compliance with ICL norms by warring factions themselves. Prosecuting weapons vendors for complicity would say to states and rebel groups alike, “If your men don’t stop these intentional killings of civilians, you won’t get weapons because your suppliers will fear becoming implicated in these crimes, and without weapons, you’ll lose the war.” Tying military objectives to the need to observe law of war precepts may assist in deterring atrocity. Obviously, this basic model is very simplistic, but I wonder if it reveals possibilities that should feature in the future work Professors Jo and Simmons call for.

This brings us to “widespread pillaging.” Uncomfortably, in virtually every situation the ICC is presently addressing, commercial pillage of natural resources has provided a means and motivation for atrocity (I do not claim that it is necessarily the only or even the dominant motivation). Prosecuting commercial actors for pillaging conflict commodities, therefore, reveals another aspect of the new promise for deterring atrocity—the war crime of pillage is a gateway to many other international crimes. On the upside, focusing on commercial pillage of natural resources may deter actors who collectively make counterfactually dependent contributions to intentional killings of civilians in most modern conflicts – without the trade in pillaged diamonds, tin or oil, the perpetrators of mass violence will be less motivated to go to war and less able to bankroll atrocity once conflict erupts. On the downside, there is also a risk of over-deterrence, where the threat of sharp judicial redress deters legitimate commercial actors from operating in volatile political climates, thereby elevating worst actors into positions of authority and penalizing civilians who are dependent on illicit mining for basic sustenance in survival economies. Optimizing deterrence is thus another key question for the future.

In all, I view Professors Jo and Simmons’ article as a wonderful opening contribution to an emerging field. I hope this symposium will foster new scholarship on these critically important issues, and that this new work will also extend to and perhaps center on, the commercial sides of atrocity.

Grounds for Continued Skepticism about the ICC’s Deterrence

Julian Ku is the Maurice A. Deane Distinguished Professor of Constitutional Law at Hofstra Law School. Jide Nzelibe is the Harry R. Horrow Professor in International Law at Northwestern Law School. They are the authors of Do International Courts Deter or Exacerbate Humanitarian Atrocities?

We want to thank Professor Stewart for inviting us to participate in this symposium on the new and very important paper by Professors Jo and Simmons.  The reason that the paper is important is that it fills a very large gap in the literature on international criminal tribunals (ICTs).  This gap is the lack of empirical data that would inform debates over the effectiveness of ICTs. In our 2006 survey of the academic literature on ICTs, we found that many proponents of the International Criminal Court claimed that ratification of the ICC would deter future crimes.  Yet there was, until now, almost no serious empirical study of this question in the academic or policy literature.

So Professors Jo and Simmons have done everyone a great service by establishing that there is some evidence for the claim that joining the ICC will deter the commission of war crimes against civilians.  Their claim is both careful and measured. It does not claim that the effect of joining the ICC is dramatic, but (when controlled for other factors), it makes a more modest claim that the effect is nonetheless non-trivial.

While we welcome the arrival of real empirical data on the deterrent effects of the ICC, we remain skeptical of some of the theoretical assumptions underlying their findings.  We also doubt that the small deterrent effect found by Professors Jo and Simmons will continue in future years.

First, we believe that the Jo-Simmons findings fail to establish the correct, or at least the most plausible, baseline for comparing the deterrent effect of joining the ICC.

As we argued in our 2006 paper, no claim of deterrence effects for criminal punishment can ignore the baseline sanctions that the potential perpetrator likely faces without the ICC.  We drew upon the substantial academic literature that has questioned the deterrent effect of capital punishment on potential murderers since an actual execution is rare given the number of appeals permitted by the U.S. legal system.  Given that the alternative to capital punishment is usually a life sentence, the additional (statistically unlikely) chance of being executed adds very little additional deterrent effect to potential killers.

So how does that apply to the Jo-Simmons analysis?  When one develops framework for deterrence, one has to have a plausible empirical baseline against which to make the comparison.  Jo-Simmons concede that prosecution before the ICC is a fairly distant possibility for any individual perpetrator (like the death penalty above), but they maintain it should make a difference because they also assume that the default baseline is impunity.   See Jo and Simmons, at 16 (“The absolute risk of punishment by the ICC remains small, but it is not negligible and is much higher than was the case when impunity was the default.”) & 16  (“ICC investigations, indictments and convictions or those triggered by complementarity are likely to encourage actual or potential perpetrators to reassess the risks of punishment – relative to the status quo, which is often impunity – and to moderate their behavior.”)

But there are good reasons to doubt that complete impunity is often the default.   It simply is not the case that individuals who engage in genocidal violence or ICC level atrocities and then get caught/captured tend to roam freely in the absence of ICC prosecution.   For instance, if Joseph Kony of Uganda’s Lord Resistance’s Army gets captured by the Ugandan army anytime soon, his realistic options will not be prosecution by the ICC or roaming free; in all likelihood, his choices will be to face long imprisonment or a firing squad by the Ugandan army or getting prosecuted by the ICC.  And notice that the ICC indictment does not change the likelihood of his capture because that is pretty much the responsibility of the Ugandan government.

Another more recent illustration is the case of the former President of Ivory Coast, Laurent Gbabgo. He is currently facing charges before the ICC for election-related atrocities; his wife (Simone Gbagbo) is not so fortunate because although she is under indictment by the ICC, she already has been convicted by an Ivorian court for “undermining state security” and is currently serving out her 20-year-sentence in a local prison. If Gbagbo were not before the ICC, he would pretty much being facing the same fate as his wife. As an interesting side note, Amnesty International actually implored the Ivorian government to transfer her case to the ICC rather than prosecute her locally. One can imagine that Amnesty International was more motivated by her welfare than the risk that she will be left to roam free.

One slight nuance that was not captured by much of the Jo-Simmons analysis is that the realistic relevant subsample of individuals who will ever face prosecution by the ICC are individuals who have been forced from power or captured.  If leaders who commit atrocities have not been forced from power, for instance, their chances of getting successfully prosecuted before the ICC are negligible or next to zero.  And when they are forced from powers or captured, the choices they face outside of ICC prosecution are not pretty.

In addition, the reasons why leaders are forced from power or captured has almost nothing to with the ICC.  Gbagbo was not forced from power in Ivory Coast because his political enemies wanted to have him arrested before the ICC; on the contrary, they forced him from power because they thought they were entitled to hold power instead of him.   Paradoxically, the real role of the ICC in such cases might not be to increase the penalty of perpetrators who do get caught, but to decrease and bring it more in line with “civilized” standards.

Thus, we remain doubtful that the Jo-Simmons study establishes the correct baseline from which to measure the ICC’s deterrence effect.

Second, Professors Jo and Simmons do not isolate the deterrent effects of actual ICC prosecutions from the deterrent effects of joining the ICC and adjusting domestic law to conform to the Rome Statute.  Given the comparatively few ICC prosecutions that have actually been brought as well as its very spotty record on obtaining  timely convictions, it is likely that much of the deterrence effect is a result of the changes in domestic law.  As Professors Jo and Simmons point out, the change in domestic law to conform to the Rome Statute, and the principle of complementarity, places the primary burden for prosecuting or investigating Rome Statute violations on state parties to the ICC.   Indeed, as Professors Jo and Simmons suggest in their theoretical framework for analyzing deterrence, it may be that the effect of joining the Rome Statute is to mobilize local groups in favor of criminal punishment for war crimes and that these political forces also operate to deter state actors.

If it is true that conforming domestic law to the Rome Statute is doing much of the deterrence work, it seems possible, or even likely, that the deterrence effect will substantially fade as time passes. 123 states are members of the Rome Statute, but (not surprisingly) the majority of them joined the treaty between 1999 and 2002.    While numerous other states have joined, only two states have joined since 2013, and only 10 since 2011.

Joining the Rome Statute, and adjusting domestic laws, probably provides the most deterrent bang for the buck.  But as time passes, the awareness and significance of the Rome Statute could fade and it is less likely that local groups can mobilize political activity in favor of it.

Finally, we also wonder if the momentum and euphoria generated by joining the ICC can withstand movements to withdraw from it.   In February of 2016, the African Union endorsed a resolution encouraging its members to withdraw from the ICC to protest perceived bias in the ICC.   Indeed, the ICC’s actual record of prosecutions have been both slow and narrow in geographic focus.  Almost every serious investigation has involved Africa.   While not all AU members will necessarily withdraw, the denigration of the ICC’s political reputation domestically will naturally weaken groups that had mobilized around ICC ratification.

In conclusion, while we believe the Jo-Simmons analysis is a much needed and important contribution to the academic literature on ICTs, we remain skeptical as to the significance of its conclusions.  The baseline for many key ICC perpetrators is not impunity, and so the claimed deterrence effect might be too strong.  Morevoer, given the passage of time and the rise of political forces lining up against the ICC in key countries, we think skepticism about the ICC’s ability to maintain whatever small deterrence effect it has remains warranted.

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.

Hey Look at Me: Deterreo, Ergo Sum

Mark A. Drumbl, Class of 1975 Alumni Professor of Law & Director, Transnational Law Institute, Washington & Lee School of Law.

“Everything must have a purpose?” asked God.
“Certainly,” said man.
“Then I leave it to you to think of one for all this,” said God.
And He went away.

— Kurt Vonnegut, Cat’s Cradle (1963)

General deterrence is routinely invoked to justify ordinary national criminal law. Intuitively, it seems plausible that would-be perpetrators of common crimes refrain from offending if they fear getting caught and then being sent before a judge. In addition, the existence of an institution that prosecutes and punishes may also deflate crime by inflating the public’s respect for the rule of law.

As a court of law, the ICC investigates, prosecutes, and punishes core international crimes. Unsurprisingly, then, the ICC’s capacity to deter such crimes emerges as a barometer of its effectiveness and legitimacy, as well as a basis upon which to rationalize its existence.

Professors Hyeran Jo and Beth Simmons investigate whether the ICC deters core international crimes. Their methodology begins with a data-set of countries with civil wars between 1989 and 2011. Noting that the Rome Statute entered into force in July 2002 (near the mid-point of the data-set’s lifespan), they attempt to disaggregate the effect of the ICC’s existence upon the number of civilians killed intentionally by government forces or rebel groups in a direct military confrontation. Jo and Simmons control and test for many interceding factors. Their research is comprehensive, meticulous, transparent, elegantly delivered, and expertly presented.

I would like to do two admittedly incongruent things in my comments. The first is to talk more about the ICC’s ability to deter. The second is to suggest that we talk a lot less about the ICC and deterrence. Whether the ICC deters atrocity is difficult to answer, but has become too comfortable and too exigent a question to ask.

Jo and Simmons locate “strong evidence of a reduction in intentional civilian killing by government actors when states implement ICC-consistent statutes in domestic criminal law,” which they indirectly attribute to the ICC’s influence. They conclude that the ICC has stronger positive effects on governments than rebel groups, although they discern an effect on rebels concerned with legitimacy. In the final sentence of their article, they posit that the ICC “has potential to save at least a few lives in some of the most violent settings in recent decades”. After fourteen years of being in operation and four years of becoming fit to begin operating, after the euphoria of Rome, and after well over $1 billion spent – there’s the bottom line.

Might the deterrent effect be even thinner than what the research – however accurate – suggests? The data-set (if I understand it accurately) extends to 2011. The data-set thereby ends early in the game. It ends before the ICC actually convicted anyone. On the one hand, the fact that the ICC has since issued a few convictions could embolden the deterrent effect by revealing the ICC’s punitive muscle to potential human rights abusers. On the other hand, a close look at the ICC’s activities since 2011 reveals the atrophied nature of whatever muscle the ICC may actually have.

The ICC’s two convictions (Lubanga and Katanga) led to prison terms of fourteen and twelve years respectively. Katanga is already free insofar as judges granted his application for sentence reduction last November. By any metric, these are lenient sentences. The ICC has also issued an acquittal. Actually taking cases to trial has exposed the frailty of much of the evidence the OTP relies on to convict and the precariousness of the accessorial liability theories it often submits. High-profile charges have been withdrawn in the Kenyan situation, where the ICC failed to guarantee witness security or testimonial accuracy and was bedeviled by obstruction on the part of the Kenyan government. OTP hibernated the stagnant Sudanese investigations late in 2014: the Security Council lacked assertiveness, Bashir is still in power (perhaps now even more firmly so), and government forces commit ongoing abuses. Happily, two high profile rebel fugitives – Bosco Ntaganda and Dominic Ongwen – are now in ICC custody. But these fugitives turned themselves in willingly. They did so because their other option – staying at large – meant they would probably be killed. They self-arrived at the ICC because it was likely a decent bet – a refuge even – for them.

The ICC may be more influential when it engages in preliminary examinations or initial investigations (i.e. when it threatens to prosecute) than when it actually brings a case to trial and sentences an offender. Colombia might be a case-in-point. This uncorks the fascinating question of the kinds of political conversations and interactivities that the specter of ICC intervention triggers. What are the domestic effects of complementarity? Dancy and Montal link ICC investigations to general domestic human rights prosecutions in Africa in what they call unintended positive complementarity. Nirej Sekhon emphasizes how complementarity reflects Foucault’s governmentality, in which international elites technocratically commune with domestic elites in a process that occludes the marginal (who mostly happen to be the victims of atrocity). Perhaps positive complementarity negates modalities of justice other than criminal trials. Selectivity, too, is a concern: in some situations in which it exercises jurisdiction, for example Uganda, the ICC achieves some justice (for LRA violence) only because it sanitizes an injustice by closing an eye to the violence of the Ugandan government.

Let’s return to the deterrence question. Jo and Simmons focus on the ICC’s ability to deter intentional killing of civilians. This category is capacious. It covers everything from group-based genocide against a defined population (Rwanda, the Yazidis today) to the targeted killing of a few individuals by a small band of specialized armed forces. Meg De Guzman is right to point out that the ICC has a gravity deficit. Although the ICC was created to deliver justice following massive jus cogens ruptures, in practice it has tended – to the frustration of many victims – to prosecute lesser crimes involving less rampant violence, including child soldiering, pillage, and destruction of historic/religious buildings. The ICTY, ICTR, ECCC, and SCSL have to date done heavier lifting when it comes to gravity. I have argued elsewhere that, as violence metastasizes into a collective project, the nature of the criminality shifts. The involvement of the rank-and-file as killers drifts from a deterrable act of deviance towards conformist behavior that normalizes violence in order to eliminate the ‘other’ for the sake of a perceived communal good. It remains unclear to me that criminal law can deter these kinds of killers or that these kinds of atrocity crimes resemble ordinary common crimes upon which deterrence theorists rest their claims. In sum, a refinement to ICC deterrence research might cleave discrimination-based mass violence (genocide and certain crimes against humanity) from other kinds of war crimes. It may be that the ICC’s deterrence capacity grafts better onto the latter category which, in turn, reflects violence that is less widespread.

Why should we talk less about whether the ICC deters? For one, repeatedly asking that very question reinforces a woefully inadequate status quo. This is the status quo in which the ICC is accepted as the best thing we can come up with internationally to deliver justice and prevent violence. There are nevertheless other ways to deter atrocity. These include actuating responsibility to protect and humanitarian intervention, reformatting a more nimble Security Council, and revisiting state sovereignty. These ways are less comforting because they require more skin in the game. Ironically, an ICC with (potential) jurisdiction over a crime of aggression might discourage humanitarian armed intervention because leaders may fear criminal liability when they intervene to help people from tyrants. Non-penal justice mechanisms, moreover, also may excel at building up the social deterrence Jo and Simmons value.

Activists invest greatly in the ICC. Activists do so, however, in a world where it’s unclear that actual atrocity survivors are single-mindedly keen on criminal trials as post-conflict priorities. Victims may wish for things that trials cannot guarantee, such as reparations, apologies, truths, reconciliation, and memorialization.

It is no answer to say: “Just because we have an ICC doesn’t mean we can’t do more!” It is no answer to posit that we live in a world of endless conjunctive permutations of “and,” “ands”, and even more “ands”. As Sarah Nouwen and Wouter Werner have argued, global policy decisions to manage conflict are not made in contexts of endless resources. These decisions entail disjunctive choices and clashes. They sit upon a topography of “ors”. We settle on the ICC as the icon of preventative justice at the expense of other options. The more we foreground the ICC, the more we settle. Yet saddling the ICC with altitudinous expectations – deterreo, ergo sum – just makes it seem weaker and lonelier.

New Symposium: Can the International Criminal Court Deter Atrocity?

Professors Hyeran Jo and Beth Simmons have authored what is arguably one of the most important articles in the field of international criminal justice in recent times. Their piece, entitled Can the International Criminal Court Deter Atrocity? (available here), defies almost all previous thinking about deterrence of atrocity by answering the question in the affirmative. Moreover, in rehearsing the previous literature on the topic, Professors Jo and Simmons lament how “[p]ractically no systematic evidence has been produced to date to support such concerns.” To correct for this deficit, they embark on a multi-faceted empirical analysis to measure the deterrent effect of the International Criminal Court, controlling for a whole range of extraneous factors that might explain the trends they observe.

I will not attempt to summarize their reasoning or findings in any great detail, but several aspects are worth emphasizing briefly. The authors begin by rehearsing much of the earlier (skeptical) literature on deterrence in international criminal justice, then offer a theory of the ICC’s “conditional impact.” The theory they develop assesses (a) prosecutorial deterrence (court-ordered punishment) as well as (b) social deterrence (extra-legal social costs associated with law violation). Having set out their theoretical expectations, Professors Jo and Simmons adopt a sampling strategy that involves 101 states and 264 rebel groups. From their ensuing analysis, they conclude that:

  • “If hypothetically, 100 civilians were killed by a non-ratifying government [of the ICC Statute], our estimates suggest about 53 civilians are likely to be killed, assuming ratification with all other control variables held constant.” (Page 28)


  • “a hypothetical well-organized secessionist movement that would have used tactics intentionally leading to the deaths of 100 civilians in the years prior to the ICC’s entry into force might have ‘only’ killed 82 civilians after entry into force, holding all other variables at their mean.” (Page 35)

Quite apart from the significance of offering a new empiricism that leads to a diametrically opposite conclusion to previous literature on this topic, the paper also adds new layers to debates about the value and legitimacy of international criminal justice as a field (I catch myself here since I am more interested in national trials for international crimes than international and critical of tendencies to lump all institutions capable of enforcing these crimes into a monolithic whole in assessing legitimacy). Even limiting our gaze to just international institutions, however, if the ICC deters meaningfully, perhaps its selectivity is less of a problem than previously imagined? Alternatively, perhaps the institution’s politics are worth tolerating?

This paper is also exciting is that, together with Kathryn Sikkink’s recent work, it marks an important empirical shift in literature on international criminal justice that is supportive of the project. In recent years, much of the discourse around international criminal justice has been critical in orientation (for my own contribution in this spirit with Asad Kiyani, see here). Nevertheless, as this blog’s manifesto intimates, I am enthusiastic about the rise of a different, relatively new, and sympathetic discourse about international criminal justice, which I hope informs a broad intellectualism in the field. Thus, I’m grateful to authors and commentators for sharing their expertise on these immensely important issues over the next fortnight. The commentators’ names and posts are listed here.

Judicial Rejection of “Specific Direction” is Widespread

I hadn’t thought to use this blog to write individual posts on new judgments or decisions in international criminal law but at the instigation of some friends, I’m persuaded to offer some very short reactions to a number of interesting blog posts in the past week on “specific direction.” The posts include commentary by Marko Milanović, Kevin Heller, Dov Jacobs and Jens Ohlin. Despite my initial reticence to re-engage with this topic in the blogosphere, it struck me that offering some thoughts on these ideas would also be an appropriate topic to end the year on and a convenient pretext for me to wish readers happy holidays and a prosperous new year.

By way of background, the latest discussions of “specific direction” in complicity stem from a judgment by the ICTY Appeals Chamber in the Stanišić & Simatović case (hereafter “Stanišić), which again rejected “specific direction” as a relevant aspect of the actus reus for aiding and abetting (see paras 94 – 109). As most readers will know, a differently constituted Appeals Chamber at the ICTY had adopted the controversial “specific direction” standard two years ago in a case called Perišić (paras 17 –74), before the same body (differently constituted) reversed itself in a very detailed judgment called Šainović (paras 1617 – 1651). But as I point out below, the judicial treatment of this question since Perišić is actually a lot thicker than this brief history would suggest: rejection of “specific direction” is far wider.

To review, I was opposed to “specific direction” when it first emerged in Perišić. I have always thought that “specific direction” as announced by the ICTY in that case was a misreading of casual language in Tadić. I won’t rehearse everything I wrote opposing “specific direction” at the time, but I do want to reiterate my empirical findings that the concept had no grounding in customary international law, comparative criminal law or previous discussions of the theory of complicity by leading experts (see here). In addition, I also wrote a blog post on Opinio Juris responding to Kevin Heller’s thoughtful defense of the doctrine. To the extent that experts at the national level have considered this problem, this is the established orthodoxy on the topic.

To complete this (overly) long introduction, I should say that I’ve always insisted that specificity is one of the many difficult questions in the theory of complicity (see here). For this reason, I’m looking forward to reading Sasha Greenawalt’s new draft article on the topic in the new year, which discusses “specific direction” by drawing on much of the voluminous and difficult literature on the theory of complicity (sorry for the delay getting to this Sasha). I very much admire Sasha’s work and I’m very pleased to have colleagues engaging with this thorny literature, especially if they reach different conclusions to mine. As I’ve mentioned public recently, I believe that a major public debate on complicity will be a great benefit to the world, even if no consensus ever emerges about its contours.

With all these preliminaries said and done, I set out below a very short list of thoughts about the most recent emanation of the “specific direction” debate in the blogosphere. I hope some of my reactions are helpful:

  1. I would like to suggest that our discussion of “specific direction” would be far clearer if we dropped the word “direction” out of “specific direction,” calling this the specificity issue in complicity or some other clearer label depending on what we mean. As I say, “specific direction” was very casual language first employed in Tadić that didn’t really mean terribly much before Perišić (see here). The definition the Perišić court gave the concept really does the specificity problem a disservice – recall that according to the Perišić definition “specific direction” entails: (a) an assessment of whether the recipient of the aid is “an organisation whose sole and exclusive purpose was the commission of crimes” (Perišić Appeal Judgment, para. 52); (b) whether the aider “endorsed a policy of assisting” (Perišić Appeal Judgment, para. 52); and (c) a distinction based on whether the aider is present at the scene of the crime or not (Perišić Appeal Judgment, para. 39, 70). I maintain that these positions are clear misreadings of complicity and that they only detract from: (a) whatever there is that’s genuinely problematic about the specificity problem in complicity; and (b) attempts to account for that residual difficulty in the theory of blame attribution writ large. I also believe that, because advocates seldom define what they mean by the term “specific direction”, we frequently talk past one another.
  1. My main contribution to the discussions in the blogosphere, however, is to place a far larger number of cases on the table for discussion. In particular, I think it bears noting that many courts have now rejected “specific direction,” such that a differently constituted court in Stanišić could not really have taken us back to Perišić even if it had decided to readopt the controversial concept. The list of cases that have rejected “specific direction” now includes:
  • The ICTY Appeals Chamber in Šainović (paras 1617 – 1651);
  • The ICTY Appeals Chamber in Popović (para 1758)
  • The ICTY Appeals Chamber in Stanišić (paras 94 – 109)
  • The ICTR Appeals Chamber in Nyiramasuhuko et al (see para 44 of Judge Agius’s Separate Opinion)
  • The Charles Taylor Appeal Judgment (see here, paras 466 – 481).
  • The ECCC in Case No 002/01 (see paras 707 – 710)

I got the sense from some of the commentary in the last week or so that the rejection of “specific direction” here again in Stanišić was somehow a farce given the composition of the bench in this case or the lack of reasoning substantiating the position. To my mind, the first of these arguments plays down that a variety of courts, at both trial and appellate levels, have rejected the standard. Thus, the supposition that a differently constituted court would have just reinstated Perišić as a norm in ICL as a field is, I think, unconvincing.

  1. To expand on this observation, I plot here the number of judges across all courts and tribunals who have voted for and against “specific direction,” from Perisic onwards. By my rough count, at least 20 different judges have had opportunity to pronounce on “specific direction” if one includes the Perišić court and everyone since in the different cases I list in 2 above. Three judges endorsed the concept in Perišić, and now Judge Afande has on entirely different grounds, but that still leaves a full 16 judges who have voted to have it overturned, some multiple times. By the by, this includes Judge Khan in Nyiramasuhuko, which means that even in Stanišić, the Agius/Afande coalition would likely have been inadequate to reinstate the standard had the judicial changes many lament not taken place. Nevertheless, even if Khan had been on the case and helped reinstate “specific direction” in Stanišić, the resulting judgment would still be at sharp odds with the vast majority of judicial thinking on the topic. Four times more judges think it is incorrect than are willing to endorse it.
  1. Looking through these more recent cases post Perišić, I read one additional judgment (not in my list in 2 above) that I think warrants mention. The ICTR’s Ngirabatware Appeals Judgment was presided over by Judge Meron and included Judge Liu, but it also involved three other judges who were entirely new to the issue. Logically, counsel for the defense argued, drawing on Perišić, that “the Trial Chamber erred in failing to determine whether the ‘specific direction’ requirement of aiding and abetting had been satisfied in his case.” (see para 145). The Appeals Chamber unanimously rejected this argument, despite their finding that “the Interahamwe used at least some of the weapons Ngirabatware distributed […] during the attacks and killings” (see para. 148) (my emphasis). Again, I’m not sure what “specific direction” means in its best light, but if it operates to deny complicity where conduct has a dual use, then surely Ngirabatware’s conduct was not “specifically directed” and he should have been acquitted. I suspect that people may argue about this given Ngirabatware’s intentions (separate from “specific direction”), but I wanted to highlight the case to suggest that even the minority of judges who advocate for this controversial standard in complicity are less than clear about when it applies and how.
  1. On the issue of substantive reasoning, I don’t necessarily share the concern about the absence of deep reasoning in the Stanišić Judgment. I take this position because the prior decision in Šainović was surely amongst the most meticulously researched judgments in the history of this discipline, drawing on the criminal law of an enormous number of states (see Šainović Appeal Judgment, paras 1617 – 1651). I’d written a doctorate that addressed the comparative law and theory of accomplice liability in ICL, then four years of further research on the topic thereafter, but still there were many sources in this judgment I’d never even heard of before. I can’t imagine what it took to acquire and analyse all these legal materials in such a short period of time, but however one views “specific direction” as a normative concept, I think we have to acknowledge that this depth of research and justification goes far beyond what criminal courts normally offer. In fact, to the best of my knowledge, it is without parallel on any other issue in ICL before or since. Accordingly, I didn’t see terribly much reason for the Stanišić Appeal Judgment to reinvent the wheel on this; it had been fully addressed previously. To be clear, no court has offered a compelling theoretical explanation of this problem, but I believe that task falls to academics.
  1. This brings me to Judge Afande’s apparent endorsement of “specific direction” by way of dissent in Stanišić, wherein he is the first and only judge to defend the concept since Perišić (Judge Tuzmukhamedov questioned the need to address it on the facts in Šainović but he did not opine on the propriety of the doctrine itself and Judge Agius has included a paragraph in two judgments maintaining his earlier reasoning without further argument). Although I respect Judge Afande’s attempt at finding a third way through a difficult legal problem and competing dissenting positions, it strikes me that: (a) the account he offers is no longer speaking about “specific direction” as espoused by Perišić (see my point 1 above); (b) his methods for this novel third way are at times highly suspect, like the use of dictionary definitions to cut through all previous debates; and (c) the resulting position is hard to reconcile with any of the different schools of thinking about blame attribution in the theory of complicity (see here). Perhaps others will defend Afande’s reasoning, but personally, I have some difficulty imagining that it will prove adequately convincing to bring so many judges around to readopting “specific direction” given the widespread judicial disagreement with it now. My guess is that only the in-depth work of scholars will be able to do that at this point, to the extent that they are able to establish that the problem of specificity (or whatever more accurate label we can attribute to the problem) cannot be accommodated anywhere else in conventional theories of blame attribution.
  1. Finally, a brief word about the changes to the composition of the bench in Stanišić. As my colleagues have pointed out, two of the judges were replaced with others before the Stanišić appeal hearing. I know very little about this backstory and have consistently steered clear of the various political controversies surrounding “specific direction”, but I confess that I don’t fully understand the complaint that the composition made the decision predictable or arbitrary. Judge Meron, himself an advocate of “specific direction,” appointed one judge for and another against “specific direction.” I see no scandal. The orders doing so were perfectly hum-drum, speaking about “the appeal management and case distribution needs of the Tribunal.” Isn’t this entirely normal and very banal? As we now know, it’s unlikely to have changed anything in the concrete case (Khan appears to be against “specific direction”) or the field as a whole (the vast majority of ICL judges clearly oppose “specific direction.”) As for predictability, I’m not sure what the problem with that is either. Many would argue that predictability is highly desirable in a criminal trial. Moreover, judges the world over almost always come with a known set of legal commitments. Think of the US Supreme Court. The parallel isn’t entirely direct, but I wanted to raise these points because I’m not convinced that this situation at the ICTY is quite as bad, surprising or irregular as some of the previous commentators have suggested. To my mind, the only thing that would be objectionably arbitrary is if, by chance, the very few international judges prepared to endorse “specific direction” again found themselves on a single appellate bench.

In any event, I hope some of the foregoing is helpful. I’ve written this uncomfortably quickly, so I hope readers will correct my errors, if there are any. Once again, I encourage scholars of all stripes to engage with this issue and the very many other difficult problems in the theory of complicity. As I mention, I believe they are part of an important struggle to lead ethically decent lives in a world that is at once highly inter-connected and very dysfunctional.

Happy holidays to one and all.


The Argor Heraeus Decision on Corporate Pillage of Gold

Earlier this year, a Swiss federal prosecutor wrote a reasoned opinion declining to prosecute a company named Argor Heraeus for pillaging Congolese Gold. I am grateful to Bénédict De Moerloose at TRIAL in Geneva and Ken Hurwitz at the Open Society Justice Initiative for their blessing to post the prosecutor’s decision (hereafter “the Decision”) here. The original German version of the Decision and an English translation the Open Society commissioned are now available in the links in this sentence. To the best of my knowledge, these documents are not online elsewhere, although the prosecutor did make them public by sending them to journalists (see here) and to the parties to the complaint.

I thought to write a neutral and constructive legal assessment of the prosecutor’s decision given that, some years ago, I wrote a legal study of pillage as applied to natural resources (see the English version here and the French version here), as well as an academic article that used the fact of a formal investigation against Argor Heraeus as an illustration of an important new legal development (see here). For present purposes, my neutrality is ensured by the fact that I know nothing whatsoever about the veracity of the allegations in the complaint, and I certainly do not vouch for or endorse any statement of fact about this case in the Decision or elsewhere. Accordingly, I also cannot form an opinion about whether or not the conclusion is correct.

Instead, I want to offer a balanced legal appraisal of the prosecutor’s reasoning in this the first formal document to discuss corporate pillage of natural resources ever. As will become apparent, I consider that the bulk of the legal reasoning is excellent but it is occasionally slightly erroneous – to the company’s benefit as well as its detriment. I hope my reactions are useful to reflections about this and other pillage cases moving forward, especially for a separate new case against a Belgian businessman involving the alleged pillage of Sierra Leonean diamonds, which has just resulted in an arrest in Belgium.

I resist the temptation to rehearse the factual allegations against Argor-Heraeus since these are contained in the Decision itself. Instead, I focus my legal assessment on pillage alone, even though the back end of the Decision also considers money laundering. The paragraph numbers I use within my headings below correspond to the paragraph number used in the Swiss prosecutor’s Decision. I have also cross-referenced relevant portions of my work with the Open Society on pillage whenever possible in the hope of providing a resource for those who wish to pursue these questions in greater depth than I can offer here.

Para 5.1.2 – For Private or Personal Use

The Decision adopts the definition of pillage set out in the ICC Elements of Crimes, which include the requirement that “The perpetrator intended to deprive the owner of the property and to appropriate it for private or personal use.” In the ICC Elements of Crimes, this particular element is accompanied by an asterisked footnote, which reads: “As indicated by the use of the term “private or personal use,” appropriations justified by military necessity cannot constitute the crime of pillaging.” I certainly understand that adopting this definition of pillage from a source as apparently authoritative as the ICC is attractive, but I have argued that this element of the definition of pillage in the ICC Elements of Crimes is not part of the crime of pillage and courts have vindicated this position.

In the Pillage Manual (see paras. 16-20), I set out how this “private or personal use” element in the ICC Elements of Crimes is: (a) not binding even on the ICC; (b) inconsistent with the exceptions set out in the Hague Regulations of 1907; (c) at odds with the vast majority of pillage cases post WWII, which involved prodigious pillage of natural resources to further the Nazi war machine, not for personal or private profit; (d) inconsistent with the case law of other international courts and tribunals that define pillage without referencing “personal or private use”, and (e) at odds with the explicit finding of the SCSL that “the requirement of ‘private or personal use’ is unduly restrictive and ought not to be an element of the crime of pillage.”

It is not evident that the issue played any real role in the ultimate result in the Decision, but legally speaking, the question matters a great deal because this aspect of the definition in the ICC Elements of Crimes arguably implies that armed groups can expropriate resources for military purposes during war, which I think is not correct save under very specific circumstances I set out in the Pillage Manual (see paras. 78-100). Although not especially pertinent here, the error is somewhat unfortunate in that it disseminates an inaccurate message about the significance of pillage for resource wars.

Para 5.1.2  – The Element of Force

Within the portion of the reasoning dedicated to defining pillage, the Decision also stipulates that “[t]o some extent it is also required that an element of force must be present during pillaging.” I do not agree that this phrase should be included in this reasoning insofar as it misleadingly implies that companies or the armed groups they purchase from must use force to commit pillage. While their operations certainly must be “closely related” to an armed conflict (see Pillage Manual, paras. 32-39), force is not an element of the offense of pillage itself. I explain the various reasons why below.

As support for its statement about pillaging requiring force, the Decision refers to in the ICTY’s Mucić Trial Judgment, which reads:

“While it may be noted that the concept of pillage in the traditional sense implied an element of violence [footnote 604] not necessarily present in the offence of plunder, [footnote 605] it is for the present purposes not necessary to determine whether, under current international law, these terms are entirely synonymous.”

In my view, there are several problems with the prosecutor’s reliance on this statement in the Decision:

  • The use of force is not an aspect of the ICC Elements that the Argor Decision draws on earlier to define pillage. It seems methodologically inconsistent to rely on the ICC Elements for “private and personal use,” then disagree with them later without good reason;
  • There is a great deal of case law saying pillage and plunder are synonyms, and they are translated between French and English as such. (see Pillage Manual, paras. 10-14) None of these definitions require force;
  • To the best of my knowledge, the words “implied an element of violence” in the Mucić Trial Judgment have not appeared in any of the many pillage cases since this passing reference. It is therefore somewhat surprising that the language resurfaces here; and
  • Footnotes 604 and 605 in the Mucić Trial Judgement provide very weak support for the proposition in any event. They are nowhere near evidencing virtually uniform state practice capable of supporting the proposition’s existence in customary international law, perhaps explaining why the passing comment in Mucić has not appeared elsewhere.

Section. 5.1.3 – Indirect Appropriation

A survey of case law governing pillage shows that the term “appropriate” in the ICC Elements of Crimes includes direct appropriation of property from the owner as well as indirect appropriation from an intermediary by purchasing stolen property (see Pillage Manual, paras. 40-49). The indirect appropriation limb of this interpretation is supported by the literal language of the ICC Elements of Crimes and at least twenty-six cases from post WWII trials, as well as conceptual first principles (see Pillage Manual, paras. 44-49). The Decision attempts to avoid these sources of authority by arguing that “[i]f even one accepts the view of STEWART, the accused could not be considered as perpetrators since they never acquired the gold in the legal sense.”

I address the argument about conversion rather than theft of the property that the second aspect of the sentence references further below. For now, I want to make a minor point, namely that the basis for indirect perpetration is not reducible to my view. As the Pillage Manual shows (see paras. 44-49), there are at least twenty-six incidents of indirect appropriation in the history of pillage cases, and generally, this extent of practice is more than adequate to ground an interpretation of customary international law in international criminal law. As I go on to mention, this interpretation of pillage is also conceptually coherent – although many legal systems disaggregate receiving stolen property from theft, others do not, and it is coherent to treat subsequent transfers of property as a new appropriation depriving the true owner of the property (see Pillage Manual, para. 48)

One recent development in the law of pillage that is not mentioned in my earlier work with the Open Society warrants mention here, too. If the prosecutor was concerned that the indirect appropriation principle could implicate consumers of products manufactured from pillaged conflict commodities, there were better ways of addressing that concern. The recent case law on pillage emanating from the ICC requires a “substantial” acquisition of property to constitute the crime, which helps address this concern. Although the term “substantial” is often difficult to define with perfect precision, there is no doubt that it should exclude consumers who purchase commodities constructed from pillaged natural resources. For reasons I mention below (see Section 5.2 below), the indirect appropriation principle is also important in addressing ex post facto causation in complicity.

Section. 5.1.3 – Conversion not Theft

As I mention above, the prosecutor avoids indirect appropriation as an aspect of pillage by arguing that, in any event, “the accused could not be considered as perpetrators since they never acquired the gold in the legal sense.”  Apparently, the gold remained the property of a third party intermediary while Argor was involved in refining it. I do not think this argument is quite as clear cut at the Decision suggests; regardless of whether a company actually acquires title, they may have converted the property and conversion may suffice for pillage. In Anglo-American criminal law, the difference between conversion and theft turns on whether the party misappropriating the property wants to take title in the property or not. Interestingly, if one looks at the table of cases at the back of the Pillage Manual (see Annex A), a number of WWII pillage cases involved conversion alone (note that some of the of the “coercion” references should read “conversion”). Conceptually, this makes some sense. The overarching legal test is whether the accused “appropriated” property without the consent of the owner; it makes no mention of whether the deprivation must be permanent or just temporary. In any event, this is a question that requires more careful research and thought. Moreover, in the abstract, an agreement to help an intermediary dispose of conflict gold could make a refinery complicit in the intermediary’s pillage if the agreement existed ahead of time, even if pillage does require an intention to acquire the property permanently. I address a closely related issue next.

Section. 5.2 – Causation in Complicity

The prosecutor makes an argument in the complicity section of the Decision that goes against the company, which I think is probably incorrect. The decision states that:

“The refining and hence increased value of the raw gold by [ARGOR-HERAEUS] is therefore causative of the pillaging by the FNI in the sense that, without the prospect of refining it to fine gold with a standardised gold context, pillaging, illegal trading and smuggling of raw gold would by no means have been a lucrative affair for the FNI.”

This argument is compelling on its face but it is actually an instance of ex post facto aiding and abetting, since the original pillage is complete by the time the company allegedly acquired the gold. There is an interesting discussion of this problem of ex post facto aiding and abetting in modern international criminal law. To make a longer debate short, you cannot make a causal contribution to a completed crime unless you had an agreement to do so ahead of time. Here, it is not clear that this is the case, meaning that the company could not be complicit in the original act of pillage by purchasing the resources subsequently absent some type of collaboration ab initio. 

Consequently, this type of allegation is probably better conceived as complicity after the fact, which was carved off from complicity proper in most jurisdictions over a century ago precisely because of these types of problems with retroactive causation. These problems are part of the reason why indirect appropriation (see Section 5.1.3 above) is such a significant aspect of the law governing pillage – it overcomes this difficulty with ex post facto aiding and abetting by implicating purchasers in a separate act of pillage. For these reasons, following the case law that mandates indirect appropriation was probably preferable to over-extending causation in the realm of complicity.

Section. 5.2 – The Mental Element for Complicity.

There is much discussion about the mental element(s) required for complicity, and I have written about these from a comparative perspective (see here) as well as at the ICC as part of an expert symposium held on this blog some months ago (see here). The Decision weighs in on these interesting discussions by, I believe correctly, insisting that “should have known” is too low for complicity. There are very few systems of criminal justice that consider negligence appropriate as a standard for accomplice liability and I believe it has no role in international criminal justice (although I believe it should be the central touchstone in business and human rights. See here). Nonetheless, there are still grounds for questioning the mental element for complicity the Decision adopts.

In particular, the Decision makes no mention of dolus eventualis. One of the leading Swiss textbooks on criminal law indicates that “Le complice doit avoir l’intention de favoriser la commission de l’infraction, mais le dol éventuel suffit.” (The accomplice must have the intention to favor the commission of the crime, but dolus eventalis suffices). See Michel Dupuis, Bernard Geller & Gilles Monnier, Code Pénal: Petit Commentaire (2012), p. 191. In other words, intention is required, but intention includes a cognitive appreciation of a risk plus a volitional “making peace” with that risk. It is somewhat strange that this standard does not feature in this analysis. I also understand that dolus eventualis was an issue in an earlier Swiss decision by a prosecutor when problems with “neutral acts” arose. Thus, it is unclear why no recognition of this broader mental element for complicity emerged in the Decision, and why so-called “neutral acts” did not reappear here either.

* * *

I hope the foregoing is somewhat helpful. Again, I am in no position to express an opinion about whether this Decision is rightly decided on issues of fact. My kind thanks again to Bénédict De Moerloose at TRIAL and Ken Hurwitz at the Open Society Justice Initiative for offering to post this material here.


The Expressive Value of Corporate Prosecutions

Wolfgang Kaleck and Miriam Saage-Maasz are Director and Vice Legal Director respectively of the European Center for Constitutional and Human Rights.

The birth of international criminal justice coincided with commercial responsibility for international crimes: the Nuremberg trials were based on a broad understanding of the political, social and economic causes of the Nazi Regime’s unspeakable atrocities. In this light, the prosecution concluded that corporations and businesspeople were relevant actors in the commission of international crimes. In the face of this legacy, the current practice of prosecuting corporations and their managers seems rather regressive. Apart from the ATS civil litigation in the USA, which hardly ever leads to any final judgements or admissions of legal responsibility, there is little modern practice. Neither the International Criminal Court nor any of the international or hybrid tribunals have investigated the responsibility of business actors in a meaningful way.

Still, on the national level there is some change: Since the US Supreme Court restricted the ATS in the Kiobel decision, there are even more efforts to use (international) criminal law at the national level to hold corporations and their managers to account. In the Netherlands a Dutch business man has been convicted of aiding and abetting war crimes committed by Saddam Hussein against the Kurdish minority by supplying chemicals which were used to produce mustard gas. Also in the Netherlands there has been prosecution against a businessman for supplying weapons to Charles Taylor. Many other efforts have been initiated by victims, their representatives and civil society groups. In France and Germany several criminal complaints have been filed against corporations and their managers for allegedly supplying security agencies in repressive states like Syria, Bahrain and Libya with surveillance technology. Also in Germany there are ongoing investigations against a manager of a Timber trading company, which allegedly called police and military forces to raid a village in the DRC and which paid those forces after having raped several women and girls. Such prosecution in not only taking place in the home countries of the companies involved. In Argentina there are several criminal proceedings pending against large Argentine as well as EU and US companies focussing on the corporate complicity in the dictatorship crimes..

Many factual and legal obstacles remain, including the lack of corporate criminal liability in many legal systems, modes of liability which make it difficult to establish the culpability of corporate actors involved in international crimes, complex corporate structures, and factual difficulties in establishing mens rea in these cases. In light of this yet very imperfect and unsatisfying practice, can any human rights lawyer ever promise her clients that criminal prosecution of corporations will bring justice? Should we even try to pursue these sorts of prosecutions or do we simply need solutions within the economic sphere?

Social movements in the Global South and their lawyers have developed a very particular approach to these problems that transcends existing legal categories, without discarding the value of law and criminal prosecutions. A good example can be found in suits that were filed in the aftermath of unionist persecutions under military rule (Argentina) and in the context of corporate complicity in crimes committed during armed conflict (Colombia). Colombia is home to one of the highest levels of attacks on unionists in the world, accounting for half of unionist murders worldwide.[1] According to some sources, over the last 25 years, about 2,500 unionists have been murdered in Colombia by paramilitary and state security forces.[2] At the same time, the level of impunity is high. Unionists are subjected to threats, arbitrary detention, torture and killings, and in some instances, their participation in unions is criminalized.

The government of Colombia is either unable or unwilling to protect unionists adequately. Even after repeated interventions of the Inter-American Commission on Human Rights,[3] the government failed to take the steps the Commission had imposed to protect threatened unionists. An emblematic case is the murder of the unionist Luciano Romero who, on 10 September 2005, was murdered by paramilitaries in Colombia while working for a subsidiary of the Swiss company Nestlé.[4] Five lower-level paramilitary members have been convicted of this murder,[5] and criminal proceedings have been initiated against other paramilitaries, informants and members of the former Colombian secret service.[6] In one of the rulings in this case, justice Nirio Sánchez ordered the prosecution to investigate the company’s role.[7]

In March 2012, the European Center for Constitutional and Human Rights and the Colombian union Sinaltrainal filed a criminal complaint against Nestlé AG and several of its leading figures.[8] The case became necessary because investigations against employees of the Colombian company have not progressed within Colombia,[9] and since they did not extend to the potential criminal liability of the foreign parent corporation. Thus, we have formally raised this question with the Swiss criminal investigative authority in order to determine the possible liability of Nestlé and its managers.[10]

Although regulations on corporate liability have entered into force as part of Swiss criminal law, they have yet to be applied in practice at all, let alone in cases of extraterritorial human rights violations. Corporate liability according to article 102(1) of the Swiss Criminal Code is clearly distinct from common criminal concepts. It is an offense against the administration of justice and is relevant only if the crime was perpetrated from within a company in which there is no identifiable individual who can be held to account. Often, individuals cannot be identified due to a lack of organization, surveillance, or documentation. Thus, the company is punished for lack of organization rather than for the crime perpetrated.[11]

The Swiss Federal Court rejected the complaint on 21 July 2014, as it considered the offences in question time-barred.[12] It thereby deviated from the opinion of broad parts of the literature and the Swiss Federal Council, which consider violations on the basis of corporate liability as continuing offenses, so that a statute of limitations would not run before the underlying shortcomings in the company’s organization are remedied.

Regardless of this outcome, human rights organizations that supported the Swiss litigation hope that the analysis of corporate behavior according to criminal law standards will contribute to the human rights performance of businesses operating in armed conflict and weak states. The proceedings will provide multinational corporations active in these fragile political environments with guidance as to the necessary risk assessment they undertake before entering into these commercial ventures. At the same time, they will contribute to further defining the “corporate responsibility to respect human rights,” as postulated by the former United Nations Special Representative on Business and Human Rights in his Guiding Principles.[13]

Criminal and civil proceedings against corporations for complicity in human rights violations face both practical and legal difficulties. Political interventions in favor of the accused companies constitute another challenge. All these problems can be observed in the cases discussed in this post. They are complemented by insufficient legal regulations to deal with these questions in countries such as Germany and the United States, where related legal challenges have failed. However, it might have been possible to overcome the legal problems raised by this type of litigation—just as it would potentially be possible to overcome them now in the United States and Argentina, where cases are currently pending—if the economic power of the defendant companies had not stymied proceedings.

Still, the cases mentioned here demonstrate that the judicial outcome is not the only measure of success of this litigation. The investigation, documentation and reporting of these sorts of cases are creating a new historical narrative in parts of German, Argentine, and Latin American societies. Often, an important and necessary first step toward ending impunity is to investigate cases of direct involvement in human rights violations. When justice and society accept the occurrence of human rights violations as a fact, it is then reasonable to inquire into the political and legal responsibility of economic and political actors. Legal proceedings in Europe, and later in the transitional countries themselves, can serve as models for human rights organizations and prosecutors all over the world, stimulating public discussions, academic research, and artistic engagement with the topic. In the course of this process, people have the opportunity to interrogate the root causes of a regime’s ruthless past, which constitutes a fundamental element of any effort to come to terms with a history of violence.


[1] IACHR, Second Report on the Situation of Human Rights Defenders in the Americas, OEA/Ser.L/V/II, 31 December 2011, para.262; Escuela Nacional Sindical (ENS), 2.515 o esa siniestra facilidad para olvidar, Ediciones ENS, Medellin, 2007, p.11, http://www.ens.org.co/aa/img_upload/45bdec76fa6b8848acf029430d10bb5a/cuaderno_19.pdf

[2] ENS, n.1, p.11.

[3]The IACHR has granted precautionary measures in favour of several trade unionist, see: IACHR, Precautionary measures of Marta Cecilia DíazSuárez and María MancillaGamboa-ASTEMP, 22September 2006, http://www.cidh.org/medidas/2006.eng.htm ; IACHR, Precautionary measures in favour of Francisco Eladio Ramírez Cuellar, 29 October 2004, http://www.cidh.org/medidas/2004.eng.htm ; IACHR, Precautionary measures in favour of Álvaro Vélez Carriazo et al, 19 May 2004, http://www.cidh.org/medidas/2004.eng.htm.

[4] Amnesty International (AI), Colombia: Killings, arbitrary detentions, and death threats – the reality of trade unionism in Colombia, London, July 2007, pp.40-41.

[5] European Center for Constitutional and Human Rights (ECCHR), Special Newsletter on the Criminal Complaint Against Nestle in the Case of theMurdered Colombian Trade Unioninst Luciano Romero, p.6, http://www.ecchr.de/index.php/nestle-518.html

[6] Ibid.

[7] Criminal Court of the District of Bogota, Judgement in proceedings against Jose Ustariz Acuna and Jonathan David Contrera Puella, 26 November 2007, p.106 cont.

[8] ECCHR, Nestle precedent case: Charges filed in murder of Colombian trade Unionist, 6 March 2012, http://www.ecchr.de/index.php/nestle-518.html

[9] ECCHR, n.6, pp.6-7.

[10] Ibid.

[11] This solution is satisfactory only in part, for it continues to view individual criminal liability as the norm. Large companies often operate in a decentralized manner and with divided tasks. This allows companies to escape criminal liability by appointing nominal directors who do not actually attend to any managerial tasks of their division. They do, however, take responsibility in criminal proceedings in order to unburden the company as a whole. Those nominal directors are then financially compensated for any inconvenience occurring in the course of the proceedings.

[12] Amerika 21, Schweizer Bundesgericht lehnt Mordklage gegen Nestlé ab, 5 August 2014 [in German] https://amerika21.de/2014/08/103485/nestle-bundesgericht.

[13] U.N. Special Representative of the Secretary-General on the issue of human rights and transnational corporations and other business enterprises, J. Ruggie, “Guiding Principles on Business and Human Rights: Implementing the United Nations “Protect, Respect and Remedy” Framework”, A/HRC/17/31 (21 March 2011), Principle 11.