The ICC Would Increase Its Prevention Ability If the Prosecutor’s Discretion Were More Visibly Limited
The ability of the institution to prevent criminality must eventually, as with criminal law systems generally, rest upon a generally internalized set of norms, rather than the “naked” threat of possible enforcement.
The prosecutorial discretion granted institutionally to the ICC has the upside of allowing a resource-constrained prosecutor’s office to maximize its effect by being able to pick and choose where to act or threaten to act. The downside of this discretion, in what is otherwise a strongly rule-governed legal system established by treaty, is that if overplayed, it can undermine the legitimacy of the court and the norms that it attempts to uphold. The suggestion made here is that the long-term legitimacy of the court would be enhanced if prosecutorial discretion were cabined within greater rule-governed and accountable constraints. That legitimacy is crucial to deterrence over the long term, because the ability of the institution to prevent criminality must eventually, as with criminal law systems generally, rest upon a generally internalized set of norms, rather than the “naked” threat of possible enforcement.
A Politics, Not a Society
My own general views of the International Criminal Court are skeptical, principally on grounds that its mandate and aspirations do not correspond to an “international” body, because the legitimacy of a court of this kind requires, as Max Weber noted, that it be embedded in a genuinely social structure, and not merely in the institutions of international politics. The legitimacy of courts in enforcing the law in domestic societies depends fundamentally not upon naked deterrent calculations by would-be law-breakers, but upon the general casting of norms within a structure in which they are generally accepted and internalized. The internalization takes two directions, according to classic social theory, acceptance and adherence by nearly all and, if not obedience, then at least acceptance and internalization sufficient to regard violation as “deviance” from a legitimate social order. In order for there to be either of those forms of internalization leading to legitimacy, there must be a social order. Since I do not regard what passes for the international community as constituting a social order—a society, in Weber’s sense—it seems to me mere metaphor and analogy to consider that the ICC can play a role globally that criminal courts play domestically.
This is of course a distinctly minority view within international law as a field, and it is not my intention to pursue this level of argument here. Nonetheless, deep skeptics can sometimes play a salutary role in forcing enthusiasts to at least pause in the march forward to ask if it goes quite where they think it goes. In this case, the highest level question is, does the mission, goals, reach, ambitions and—notably—ability to deter of the ICC depend ultimately upon the “embedded” community being a society and not merely a politics? And if so, in what sense is the ICC embedded in a “society”?
The problem in raising this kind of question is that the ICC always has an answer to this kind of skepticism, even if it is granted as a category even in principle: give us time. We need more time to solidify and consolidate our institution, and the institutions among which it nestles; as international society, as you say, consolidates, it will too. If it becomes obvious that the ICC is achieving prevention of international crimes, then the question goes away. If there are questions about that, then at some point someone has to say, well, how much time before we are permitted to judge whether the institution is a success or failure or something in-between.
The Metrics of Prevention, or, the Dangers of Datasets
In practical terms, what are the metrics on which to ask whether the actions of the court or the prosecutor are having a preventative impact on international crimes? As I scan the scholarly literature, particularly that coming from political science, I am impressed with the efforts being made to establish measurable and quantifiable methods of establishing to the satisfaction of social science that an action, or threatened action, by the prosecutor or by the court had a deterrent effect, either in that direct case or indirectly elsewhere. And that the accumulation of such actions was having, or not having, such effects on a more global basis.
Methods such as examining statements by the subjects of such actions, whether reported in the press, or once in custody, can help give some sense of possible impacts. More indirect measures, looking to such things as child soldier recruiting practices, or rates of attacks in violation of the laws of armed conflict that were expressly raised by the prosecutor, can also help establish whether the institution has a preventative impact. I hope that scholars who deal in empirical methods elaborate more and better methods of evaluation. It is an improvement on the traditional methods of international law, which are not well-informed by social science.
Having said that, I want to express caution about the application of quantitative methods in an area so dependent upon long causal chains with vast arrays of potentially confounding variables. In an area in which data collection of the kind one would ideally like in experimental design is so difficult to come by, the temptation to assume that one’s dataset must say something important just because one was able to collect it is very powerful. Much of the argument necessarily takes the form of counterfactuals—akin to the ferocious argument today in the United States over “jobs created or saved,” which, after all, does have vast quantities of real data to draw on and yet remains contentious. Its equivalent in dealing with the effects of the court and the prosecutor, particularly over time, and beyond limited circumstances—which is to say, “but for the action of the court and prosecutor”—is just as contestable and, in my estimation, should not be the defining criterion.
Rather, the qualitative evaluation of such things as interviews with people directly involved, qualitative rather than narrowly quantitative inquiries, seem to me more useful, at least until there is something of the dense information networks that inform advanced societies in crime statistics. This creates its own set of difficulties; qualitative research carries, of course, many temptations to selection bias, preference for certain interpretations, what in retrospect looks to be over-emphasis on certain informants and kinds of information. And the “but for” problem inherent in the whole question of deterrence and prevention is just as real as with quantitative methods. One should be willing to offer different methods, and be willing to be skeptical and accept that they might reach quite different conclusions. And finally, it goes without saying, if one asks for the assessment, one cannot do so with an expectation of where the conclusion must come out.
The Model of Rational Deterrence
Classical criminal deterrence theory says that, assuming risk averse law-breakers who fear punishment, we balance off the resources put into monitoring and enforcement of the criminal law against the punishment attached to it. The greater the punishment, the less we need monitor for law-breaking; and vice-versa, the less the punishment, the more resources needed to monitor. The classical model assumes a further social assumption, the one noted in the opening section—viz., a general assumption of the legitimacy of the laws and legal order at issue.
We think that the international community has settled on a short list of the most heinous mass atrocity crimes that no rational human being could deny and committed them over to an international tribunal; but from out in the field, no one needs telling, it does not seem that way.
This might sound like pure theory, but it is the most practical deterrence problem of the court. None of these assumptions holds except at best weakly. Let us walk through them. Is the legal order at issue legitimate in the sense that its fundamental mandates are seen as legitimate and departures from them as “deviance”? For those of us from the human rights community, nothing could seem more obvious—but seen from the ground up, this is at best a weak assumption. We think that the international community has settled on a short list of the most heinous mass atrocity crimes that no rational human being could deny and committed them over to an international tribunal; but from out in the field, no one needs telling, it does not seem that way. This is for many reasons—yes, these are fundamental human rights, but …. Everyone knows this, but criminal law systems that do not enjoy this broad legitimacy among those it proposes to police will have to invest vastly more resources in monitoring and enforcement—and even then, traditional policing theory questions whether it can work.
Are the subjects of the ICC actions actually risk-averse? The question might seem surprising, but there are reasons to think it is only weakly true, at most. Yes, it is clear that in some circumstances, those named in court or prosecutor actions fear being arrested while traveling abroad. There are some indicators that some parties have at least discussed what the threat of ICC action might mean for them down the road. I don’t deny any of that. I want to suggest, rather, something that appears only sporadically in the literature—in part because it is not really part of the purely “rational” actor school of thought about deterrence and criminal punishment.
One of the peculiarities of the ICC as an institution is that it is addressed to individuals. The reasons why that is so are many and quite persuasive, but this fact nonetheless means that there is an odd gap between institutional responsibility for mass atrocity crimes and those of individuals. From the standpoint of rational deterrence theory, this criminalization of the individual agent’s activity should be a very good idea, because a risk averse individual agent will have good reasons to not want to internalize on an individual basis the crimes, so to speak, of the whole institution. Hence the risk aversion of the individual agent will run back to behavior by the institution as principal. This is all perfectly true in rational choice theory and agent-principal relations. It also fits well with the model of the criminal leader of a government or a society responding rationally to incentives and fear of punishment.
The models assume that criminals will cut and run to save themselves; the models underestimate, in my view, the extent to which many of the most important possible targets of the prosecutor are regarded by their followers and by themselves as genuine and faithful fiduciaries.
This account leaves out, however, a feature of even criminal leadership that is not rational in this way. Leaders are leaders, and though we correctly say that they are criminals and guilty of mass atrocity crimes, much of the time these acts are carried out—in their own self-perception and in that of their followers—as fiduciaries. However criminal their actions, they are invested by their followers with leadership, trust, obligation, and expectations of leading—what a society, a state, a government, a tribe, a clan, a family, whatever the criminality involved. The theories of prevention that underlie much of the advice to the prosecutor and the court depend fundamentally, it seems to me, on overly strong identities between “criminal” and “illegitimate”. The models assume that criminals will cut and run to save themselves; the models underestimate, in my view, the extent to which many of the most important possible targets of the prosecutor are regarded by their followers and by themselves as genuine and faithful fiduciaries.
I mean this only weakly, and such relationships eventually can be broken down—though a NATO campaign helps quite a lot—but this makes the proposition that a risk averse individual agent will convey such risk aversion to and through the institution only a weak proposition. In that case, the assumption of a risk averse target of criminal sanction is only weakly true.
On the classic deterrence model, prosecutorial discretion acts like a sort of “force multiplier” in military affairs—the threat that the prosecutor might intervene here or there gives either party a reason to be careful. In practical terms, however, the world is too big a place. The court asserts its legitimacy on the basis of a carefully negotiated and crafted rules-based system, a complex and highly structured treaty that establishes the terms of its own criminal law. Judges set about interpreting those structures to create an internally cognizable body of predictable rules. But then the system is run at the front end through a highly, if not purely, discretionary prosecutorial choice. And that prosecutorial discretion—though it fundamentally contravenes the legal tenets and indeed legitimacy of the rest of the system—is then announced as a virtue, not a necessary vice, because it enables this multiplier effect of enforcement.
While understanding all the reasons why the system works this way, let me say that on balance it is a very bad idea. In most domestic legal systems—leaving aside the altogether notorious American way of justice in this matter—discretion exists, but is cabined by many different things. Sometimes it is by formal mechanisms; in some places the law of bankruptcy gives a judge very wide discretion in dealing with creditors, for example, but locks the discretion within a stable set of otherwise predictable rules. Sometimes it is by informal mechanisms that are built into the cultural and extra-legal expectations of how the system is supposed to work. I suppose that many people, deeply involved in the elaboration of the ICC and its practices over many years, believe that this is the case here and now for the ICC prosecutor and discretion. Seen from the outside, however, the prosecutor’s office appears something of a wild card.
The desire, first, to obtain the force multiplier effect of discretionary prosecutor actions and, second, to assert the legitimacy that the court genuinely considers situations worldwide (and not just as a court for Africa, in the critical phrase)—yes, the reasons why the exercise of this discretion makes sense is clear. But the accompanying effect is also to de-legitimize both the prosecutor and the court, because the system is so much a rules based system otherwise. The prosecutor’s discretion on the front end undermines the legitimacy that accrues as Weber (to return to social theory) analyzed as a reason for bureaucratic rationalization of the rules and their enforcement.
A system for going after the world’s worst crimes and worst international criminals that has a feeling of simple misfortune to the participants will not fulfill very adequately either legitimacy or rational deterrence.
Moreover, the lack of predictability in a system in which the resources of the prosecutor are so small in relation to the whole world that intervention looks like a lightning strike, turns belief in the system as something no longer about legitimacy, or even about rational deterrence. It looks like just plain bad luck. A system for going after the world’s worst crimes and worst international criminals that has a feeling of simple misfortune to the participants will not fulfill very adequately either legitimacy or rational deterrence.
In a practical way, how to constrain this discretion and allow it sufficient play, while cabining it to make it more predictable? Well, this is not popular advice, but the first thought would be to announce in advance that the geographical reach of the court, not as a statutory matter, but as a resource matter, will be constrained—for example, to countries that might be listed in advance by a panel of the court, not the prosecutor, as being seen as at risk for the failure of the rule of law. Or simply explicitly limit, for now, the prosecutor’s reach to situations of actual civil war. These are tradeoffs, and any of these suggestions might have fatal downsides. My point, rather, it to observe that there is a tradeoff between the seeming invitation to evil-doers who will not be covered at that point by the prosecutor’s attentions (because they are in Asia or not in civil war, for example), and the sense that they are not really covered in a meaningful, deterrent sense anyway (because the world is too large, the OTP too small, and if discretion strikes, it was simply vagaries of fate).
Moreover, the relationship of the prosecutor to the Security Council in the exercise of discretion raises important questions about neutrality and the rule of law. I do not mean this in the way usually raised as a concern. I myself would vastly prefer to see the prosecutor constrained by the need of an actual referral in every case from the Security Council; I understand very well that the current system exists as it does in large part because outside actors, including NGOs and many countries, wanted a system that allowed them to “contract around” the dubious authority, and more importantly deadlock, of the Security Council. Point accepted for this discussion. The prosecutor, in that case, should either address situations that are already the subject of Security Council action, such as the authorization of humanitarian intervention, very strictly according to the terms of referral in a transparent and rule governed way—eschewing, in other words, the suggestion that the prosecutor is acting according to its pure discretion. Or else pursue its prosecutorial agenda purely according to its own criteria, without any particular attention to actions of the Security Council (outside of the formal referral process). One might question whether such a distinction can be maintained in fact, but appearances here matter.
The least attractive alternative is to be seen to be scurrying after the Security Council in a discretionary way. Why? Because it gives the unfortunate appearance that the sacred discretion of the prosecutor is something to be used in the agendas of other actors—even agendas that are, on their own terms, perfectly defensible. It is too strong and unfair to say that the prosecutor, in the case of Libya, has looked to the rest of the world as though the prosecutor were a mere camp-follower of the Security Council, or more precisely, NATO; the Security Council referral both helps the Prosecutor and hurts it in this regard. But a certain indecorous chasing-after has not gone unnoticed, as I read world opinion. Even if this is substantively quite unfair to the prosecutor’s office, a certain appearance is there.
The prosecutor ought to recognize that its discretion is useful, but only in very limited ways and rapidly becomes counter-productive to deterrence.
The broader practical point is that the prosecutor ought to recognize that its discretion is useful, but only in very limited ways and rapidly becomes counter-productive to deterrence. In any case, it runs strongly against much of the rest of a system of justice that is strongly rule-governed.
I have tried, I hope not too unsuccessfully, to make points that are within the structure of the court and prosecutor offices as they exist. I apologize for the excessive abstraction. But I do think there are certain practical points here concerning deterrence—viz., that the discretionary system as it exists now does not seem to me to promote deterrence. Finding ways for the prosecutor’s office to cabin that discretion voluntarily within a much more rule governed set of constraints would, paradoxically, increase legitimacy and the sense that action by the prosecutor or the court is more than just a misfortune.