WHICH PUNISHMENT? IMPRISONMENT, LUSTRATION, AND PUNITIVE SANCTIONS
Assume, then, that some individual W is liable to punishment for war-related crimes. But what kind of punishment, though? It is a common trope that the punishment should both fit, and be proportionate to, the crime. Fittingness and proportionality are not the same. Thus, one might think that imprisonment is a fitting punishment for assault, but that condemning someone to life imprisonment is a disproportionate punishment for a single (non-lethal) assault. Conversely, one might think (justifiably) that asking someone to pay a fine for killing another person is not on its own a fitting punishment (though punitive damages might be warranted, alongside a jail sentence), but that, if it were a fitting punishment, asking someone to give their victims n per cent of their financial resources might be proportionated 
What counts as fitting and proportionate punishment is one of the thorniest difficulties which any full-blown account of punishment must tackle. I cannot do so here. Instead, I shall take for granted that the more grievous the rights violation, the greater the punishment—subject to upper limits set by W’s right (which he may never be deemed to have lost) to humane treatment. Recall, however, that even if W is liable to be punished in particular ways, it does not follow that we are justified in punishing him all things considered. In particular, the requirement not to undermine goods such as third parties’ interests in being protected from W himself or from putative wrongdoers delineates conditions for justifiably punishing him (see s.7.2.1). Suppose that W belongs to a powerful gang of criminals at war against the state, and that if we punish him, his associates will launch a campaign of murderous violence against business owners who refuse to pay protection money. Punishing W might not be justified in this case, on the grounds that it would lead to the deaths of many innocent people.
Not only, then, must punishment be proportionate to the crime itself; it must also be proportionate in the sense that the goods it brings about must not be outweighed by the harm it will cause. To put the point differently, punishment, like defensive force, is subject to both narrow and wide proportionality. Narrow proportionality refers to the fit between his wrongdoing and the amount and kind of punishment to which he is subjected, whereas wide proportionality refers to the impact on third parties of punishing him. Thus, W is liable to punishmentp only ifp is narrowly proportionate to his wrongdoing; but inflicting P on W is justified all things considered only ifp is widely proportionate.
The distinction between narrow and wide proportionality is notably and regrettably absent from contemporary debates about appropriate sentencing.28 In the present context, it is absolutely crucial. To see this, consider punishment for significant contributions to violations of rights to life, bodily integrity, and property. On the plausible assumption that they generally warrant imprisonment in domestic contexts, one might think that they also warrant imprisonment when classified as war-related crimes. However, there are at least three very good reasons not to impose it other than in the clearest, most egregious cases such as massacres of civilians, or the instigating and planning of unjust aggression and concomitant unjust killings—for two reasons. First, punishing combatants merely for fighting might provide them with incentives to continue to fight so as to avoid falling into the hands of the enemy—at the cost of securing prospects for peace. The point is familiar, though it is not made in domestic settings, at least for violent crimes. For example, we do not hear it said that human trafficking should not be turned into a criminal offence lest it should provide traffickers with incentives to commit further rights violations in order to avoid punishment (e.g., kill victims who show signs that they might go to the police, corrupt police officers, which in turn weakens
On proportionality, see A. Ashworth, Principles of Criminal Law, 5th edn. (Oxford: Oxford University
Press, 2006), 19-20.
28 In domestic contexts, wide proportionality refers, for example, to harms accruing to a criminal’s dependent relatives as a result of his imprisonment. For a brief discussion of wide proportionality and (domestic) state punishment, see Tadros, The Ends of Harm, 358-9.
the legal enforcement of rights, etc.) The only reason why we might want so to distinguish between war and non-war criminals is the sheer numbers of the former, relative to the latter: we might be able to bear the costs of further rights violations for the sake of punishing non-war criminals to the degree to which they are liable; but bearing the costs of the vastly greater rights violations committed by combatants desperate to avoid punishment seems morally undesirable.29
Second, as we saw in s.7.4.1, many participants contribute to unjust wars or unjust phases therein in ways which do not warrant as severe a punishment as imprisonment—such as the citizen who knows that parts of her taxes fund an unjust war of aggression, the journalist who writes in favour of the war, the engineer who works in dual-use facilities, and so on. In so far as imprisonment would be disproportionate in the narrow sense, those wrongdoers are not liable to it.
Third, even if imprisonment is narrowly proportionate in some cases, it is likely to be widely disproportionate anyway. Remember that it is only if one can be under no reasonable doubt that the defendant committed the crime and either intended so to act or could reasonably be expected to know what he was doing (and the moral status of his act) that one may harm him to that extent. Given the sheer numbers of agents involved, establishing judicial institutions which would be able to reach a verdict on both actus reus and mens rea would be extraordinarily costly. Assuming that financial resources could justifiably be made available to that end, imprisoning all those found guilty would also be extraordinarily costly. And assuming that this could be done, the human costs of so doing would themselves be enormous, since thousands of individuals would be sent to jail who could take part in the rebuilding of their community. To put the point in the language of feasibility, even if it were practically feasible to put on trial, judge fairly, and imprison every single agent who is strongly suspected of having participated in grievous war-related crimes, the financial and human costs attendant on doing so would detract from the pursuit of other just ends, whether post bellum ends such as the reconstruction of war-torn communities or extra bellum ends such as the provision of non-war related essential goods and services. Partly for those reasons, this would also jeopardize prospects for a justifiedATC peace, and it is in that sense that it would be widely disproportionate. The wholesale imprisonment of suspected war criminals, in other words, is morally unfeasible given scarcity of material resources. The case of Rwanda illustrates the point well: seven years after the genocide, over 100,000 people, each suspected of having taking part in genocidal killings, were still in jail awaiting trial. 
The most that we can hope for, thus, is the imprisonment of high-ranking officials and civilian participants such as directors of implicated companies (e.g., Krupp, Flick, and IG Farben at Nuremberg) for large-scale war-related crimes, and of those who directly killed or raped, or directly gave orders for micro-threats such as massacres or torture. In fact, sometimes we ought not to punish those agents at all, particularly in the context of a civil conflict. Even if we may punish them, this leaves us with the question of what to do with other agents—many of whom (let us not forget) did commit wrongdoings which are justifiably punished with imprisonment within domestic contexts. In that vein, the Rwandese authorities released most of those suspects and had them face non-adversarial community courts modelled on traditional modes of conflict resolution. I shall discuss this case in some detail in s.9.4. In the remainder of this section, I examine two punitive options— lustration on the one hand, and financial penalties on the other hand. Both have been available to international criminal courts, as stipulated both in the Rome Statute for the International Criminal Court, and in the Statutes establishing the ICTY and the ITR. Both may provide a solution to the difficulty raised by the case of those who are liable to be imprisoned but whom we should not put in jail all things considered, and those who are not liable to be imprisoned but who are nevertheless liable to some form of punishment. In all those cases, the claim that wrongdoers are so liable, conjoined with the claim that punishment of some kind is in order (whether that to which they are liable or a lesser form thereof ), yields the conclusion that their war-related wrongdoings ought to be regarded as criminal offences—that there is a sense in which, thus, that the laws of war should track the morality of war.
-  On the connections between compensation and punishment, see A. Ashworth, ‘Punishmentand Compensation: Victims, Offenders and the State’, OxfordJournal of Legal Studies 6 (1986): 86—122.
-  Note that even if this point is correct, it does not undermine my earlier claim that war punishment is not sui generis in relation to domestic punishment. In principle, we must also take wide proportionality into account in both contexts. The point thus merely holds that the proportionalitycalculation is likely to differ across those contexts, such that war punishment is likely to be widelydisproportionate while domestic punishment for roughly similar wrongdoings is not. By implication,then, if, in a given pairwise comparison, the harms attendant on punishing combatants for seriousviolations of human rights would be widely proportionate while the harms attendant on punishingdomestic wrongdoers such as, e.g., gang members would be widely disproportionate, punishmentwould be warranted in the former case while not in the latter.