A New Moral Hazard? Military Intervention, Peacekeeping and Ratification of the International Criminal Court

The newly established International Criminal Court (ICC) promises justice to the victims of genocide, war crimes and crimes against humanity. Past offenders can be punished, while future potential offenders may be deterred by the prospect of punishment. Yet, justice is no substitute for intervention for the benefit of people at acute risk of being victimized. The Court may create a new moral hazard problem if the promise of ex post justice makes it easier for states to shy away from incurring the costs of intervention. This article indirectly tests for the relevance of this potential problem by estimating the determinants of ratification delay to the Rome Statute of the ICC. If the Court represents an excuse for inaction, then countries that are unwilling or unable to intervene in foreign conflicts should be among its prime supporters. Results show instead that countries that in the past have been more willing to intervene in foreign civil wars and more willing to contribute troops to multinational peacekeeping missions are more likely to have ratified the Statute (early on). This suggests that the Court is a complement to, not a substitute for intervention.


Introduction
The creation of an International Criminal Court (ICC) has been hailed by many as a major breakthrough in international law. Weller (2002: 693), for example, has called it 'the culmination of international law-making of the twentieth century'. Former United Nations (UN) General Secretary Kofi Annan regards it as 'a vital part of an emerging system of international human rights protection' (Annan, 1997/98: 365).
Yet, the ICC is no substitute for humanitarian intervention and multinational peacekeeping since at best it provides ex post justice together with the hope of deterrence of future crimes, but no immediate relief and assistance to people at risk of becoming the victims of grave offences. Critics have argued that the ICC might just be a cheap way for states, which are unwilling or unable to intervene militarily abroad and to contribute to peacekeeping, to demonstrate to their domestic publics seeming action against genocide, war crimes, aggression and crimes against humanity (Smith, 2002). In other words, is the ICC a form of organized hypocrisy behind which states hide who cannot or do not want to take real action?
This article attempts to answer this question indirectly by examining ratification of the Rome Statute of the ICC. Specifically, I test whether countries that have been more willing in the past to intervene militarily in conflicts abroad and who have contributed more often troops to multinational peacekeeping are more or less likely to have ratified the Rome Statute (early on). An answer to this question not only sheds light on the likely motivations of state parties, it also informs a judgement of the likely future effectiveness of the Court. The ICC is only as strong as its enforcement capacity and it is dependent on states for crucial assistance during all stages of its proceedings, particularly for the capture and extradition of indicted individuals. If those states that are willing to commit troops for military intervention and 4 peacekeeping abroad were less likely to support the ICC, then this would bode badly for the Court's enforcement capacity.
In brief, the results from a stratified proportional hazard model over the period 1998 to 2004 suggest that, despite the notable opposition to the Court from the United States, Russia, India, China and others, states that have been more willing to intervene abroad and more willing to contribute to multinational peacekeeping in the past have been more, not less, likely to ratify the Rome Statute (early on). Taking ratification as an indication of revealed preference for ICC support, this suggests that the Court is unlikely to function as a sorry excuse for inactive governments.

The Rome Statute of the International Criminal Court
The idea for an ICC equivalent was briefly floated in the post-Second World War era, but soon dumped in the wake of Cold War antagonism. It was reinvigorated by Trinidad and Tobago in the late 1980s with the support of other Caribbean countries, which wanted to see international drug trafficking and money laundering being prosecuted by an international court. The International Law Commission, a UN expert body of legal scholars, was charged with formulating a draft statute, a final version of which was presented to the UN General Assembly in 1994 .
Encouraged by the establishment of the two ad hoc tribunals for Yugoslavia and Rwanda, the General Assembly created a preparatory committee whose objective was to formulate a draft treaty with much wider coverage and competence for a permanent court to be established than initially envisioned by Trinidad and Tobago or the International Law Commission.
At a conference in the Italian capital, the Rome Statute of the ICC was opened for signature in July 1998 after five weeks of negotiation. International civil society, 5 in the form of the NGO Coalition for an International Criminal Court (CICC), exerted an important influence on the negotiations of the Rome Statute, but only after the decision to establish an ICC in principle had already been taken by states (Fehl, 2004: 374). The Rome Statute entered into force on 1 July 2002 after having been ratified by 60 state parties. As of February 2008, the treaty had been ratified by 105 states. The four most prominent non-parties are the United States, Russia, India and China.
After long negotiations, the state delegates present in Rome decided to include four categories of crimes under the remit of the ICC (Art. 5:1): genocide, crimes against humanity, war crimes and, at least in principle, aggression. These crimes are considered so heinous that their punishment should be a matter of international concern and jurisdiction. The list of crimes can be added to at a review conference, to be held seven years after the coming into force of the Rome Statute, i.e. in 2009.
Currently, it seems that state parties have little appetite for amending the list of crimes. In fact, it is even doubtful whether they will give full effect to the aggression category. This crime was formally included in the Statute, but essentially to no effect since the negotiators at the Rome Conference could not agree on a definition of what constitutes aggression and how it is to be prosecuted, leaving these essential details to future amendments instead (Art. 5:2).
According to Art. 6, genocide covers 'acts committed with intent to destroy, in whole or in part, a national, ethnical, racial or religious group'. Of note, much like the 1948 Convention on the Prevention and Punishment of the Crime of Genocide, the Rome Statute does not cover political and social groups within the scope of genocide, which according to Schabas (2004: 40) has been frequently criticized. Art. 7 of the Statute lists 11 acts of crimes against humanity, from the somewhat anachronistic crime of apartheid to the widely recognized crimes of murder, torture and enforced disappearance as well as 'rape, sexual slavery, enforced prostitution, forced pregnancy, enforced sterilization', only very recently recognized as crimes against humanity. Prosecution of such acts is always subject to the requirement that they are 'committed as part of a widespread or systematic attack directed against any civilian population, with knowledge of the attack' (Art. 7:1). Art. 8 of the Statute represents perhaps one of the greatest steps forward. Very explicitly, it extends the scope of war crimes from international conflicts to civil wars (Schabas, 2004: 54). It covers 'grave breaches of the Geneva Conventions' together with 'other serious violations of the laws and customs applicable in international armed conflict', but also 'serious violations of article 3 common to the four Geneva Conventions' as well as 'other serious violations of the laws and customs' for civil wars. However, state parties can opt out of the provisions for war crimes for a one-off transitory period of seven years (Art. 124).
The Court can exercise jurisdiction over individuals whether or not their nationstate is party to the ICC as long as the crime has been committed within the territory of state parties. However, the Court requires the consent of either the state of the accused national or the state, in which the crime took place, unless the case is referred to the Court by the UN Security Council. Also, the Court is only meant to take action under the principle of complementarity, i.e. only when a state 'is unwilling or unable genuinely to carry out the investigation or prosecution' (Art. 17). Commanders can be prosecuted for crimes committed by forces under his or her command if the commander knew about the crimes and could have prevented them (Art. 28).
Conversely, offenders cannot take recourse to the argument that they committed crimes upon superior orders (Art. 33). No offence can be prosecuted that took place before the Rome State came into effect in July 2002 (Art. 11).

A potential moral hazard problem: the ICC as an excuse for inaction?
Most extant scholarship has focused on trying to make sense of US resistance toward the ICC and has discussed at length the credibility and merits of the two main US objections toward the Court -see, for example, Forsythe (2002), Mayerfeld (2003), Fehl (2004), , Weller (2004), Ralph (2005) and Johansen (2006). The two main explicitly stated objections are the potential for abuse of the Court by enemy states and the risk to US military personnel on active duty abroad (Bolton, 2002).
Clearly, the Rome Statute could have been constructed in a way that would have made it easy for the US to ratify. In particular, it could have allowed permanent members of the UN Security Council to veto any investigation and prosecution, thus effectively enabling them to protect any of their nationals or those of their allies.
Doing so would have meant creating large "loopholes" in exchange for the benefit of support by the US.
This proved unacceptable to the vast majority of negotiating states, however.
Instead, the parties chose an option that risked US withdrawal from the entire project.
The UN Security Council can defer investigations and prosecutions repeatedly for a period of twelve months each (Art. 16), but none of its permanent members has the power to veto the initiation of such proceedings in the first place. As a result, the Bush administration withdrew Clinton's signature under the Rome Statute and it began to actively undermine the ICC by, first, refusing to support UN peacekeeping missions unless US citizens are explicitly exempted from ICC proceedings and, second, pressurizing state parties via the threat of economic and political sanctions into concluding immunity agreements guaranteeing that no US citizens will be held or extradited for the purpose of ICC proceedings (Johansen, 2006). However, the Bush 8 administration has recently wavered a bit in its resolution to refuse co-operation with the ICC. In particular, it has abstained from vetoing a UN Security Council vote referring the Darfurian crisis in Sudan to investigation by the ICC in March 2005 (Grono, 2006: 629).
The decision by the state parties to solve the crucial issue of the role and power of permanent UN Security Council members in the particular way described above has prompted critics to question the enforcement capacity of the ICC. Fehl (2004: 358), for example, wonders '…why was the new institution designed in such a way that it failed to gain the support of the one country which is, by all accounts, most important for enforcing its future decisions?'. This is a relevant question to ask, after all the US is in a unique position due to its quasi-universal deployment of troops and infrastructure and its logistical and military capacity to intervene virtually anywhere in the world.
Yet, this critique only scratches the surface of a potential problem that reaches much further than the enforcement of future Court decisions. The very creation of the ICC might present a fundamental problem, namely via the negative effect it potentially has on humanitarian intervention and multinational peacekeeping. There are two reasons why the ICC might have such a negative effect. One argues with a direct deterrence effect, the other one with a potential moral hazard problem. The former is a favourite argument of the US administration used in its opposition to the Court. Let us briefly discuss each argument in turn.
The ICC may directly deter humanitarian intervention and peacekeeping if those who intervene and contribute to peacekeeping must fear that their military personnel may face future indictment for their actions by the ICC. As David Scheffer, then US Ambassador-at-Large for War Crimes Issues has put it in addressing the UN General Assembly in 1998: 'The Rome treaty will become the single most effective brake on international and regional peacekeeping in the 21st century.' (cited in Lietzau, 2001: 125). Similarly, US Major Michael L. Smidt (2001: 240) contends that 'the court will likely make the world a much more dangerous place because it will likely deter the forces of good, which will allow the forces of evil to act with impunity'. India, another prominent critic of the ICC and, contrary to the US, a troop contributor to many multinational peacekeeping missions, has similarly raised the issue of a negative effect of the ICC on peacekeeping operations and concern for its army personnel involved in such operations as part of its case against ICC ratification (Koshy, 2004(Koshy, : 2439McDoldrick, 2004: 440).
Such arguments are rather implausible, however. It is possible, of course, that members of an intervening or peacekeeping force will commit crimes that fall under the auspices of the ICC and might be indicted for them. But judged from past experience, committing such crimes during interventions or peacekeeping missions is the rare exception, not the rule, and would constitute cases that should be prosecuted by the state to which these soldiers belong to itself, such that the ICC would never become active in the first place. Neither is it likely that the ICC would become

10
The moral hazard problem is more pertinent and a priori more plausible. As Smith (2002: 177) has formulated it succinctly: 'If international actors feel confident that human rights criminals will eventually be brought to justice, either in their own countries or before the ICC, they may be less inclined to intervene to stop human rights crimes while they are happening, something international actors have been reluctant to do in any case.' As a consequence, 'the ICC may become a virtuous excuse for states to turn a blind eye to atrocities, a moral free ride on the coattails of humanitarian law' (Smith, 2002: 178).
Such suspicion and criticism is fuelled by the reaction of the international community to the crimes committed during the Yugoslav wars and the Rwandan genocide. In both cases, states arguably failed to intervene or intervened far too late.
In both cases, states tried to make up for their failure by installing special ad hoc tribunals to prosecute crimes ex post. Aryeh Neier (1998: 112), co-founder and then director of Human Rights Watch, argues that the Yugoslav tribunal acted as a substitute for action: 'Facing domestic criticism for allowing the slaughter to continue unchecked, some governments seemed to feel obliged to show that they were doing something', and without much actual political cost. Similarly, Forsythe (1994: 403) has commented that European states 'felt the need to give the appearance of doing something about violations of humanitarian law in the former Yugoslavia', but 'lacking the political will to act decisively to curtail abuses of prisoners and civilians, they endorsed or went along with the creation of the Tribunal'. Is something similar happening with the ICC?

An Indirect Test for the Moral Hazard Problem
The potential moral hazard problem created by the ICC cannot be directly observed or tested for. As an indirect test, I estimate the determinants of its ratification. The idea is that if the ICC creates an excuse for states which are reluctant to act, then these states should be the prime supporters of the Court, which should reveal itself via (early) ratification. In other words, if the ICC represented a real moral hazard problem, then states which have been relatively inactive in the past as interveners in foreign humanitarian crises should be among the prime supporters of the ICC. If these states were the prime supporters of the ICC then this would suggest that the court functions as a substitute for action (the moral hazard problem is very relevant). If the opposite were the case -that is, if states which have been more active in the past were its prime supporters -then this would suggest that the ICC functions as a complement to action, which would imply that the moral hazard problem is rather irrelevant.
For this argument to be valid, it must not be the case that countries which have been active in the past ratify (early on) precisely because they are anxious to get the ICC established such that they have an excuse for being less active in the future. I contend that this is a safe presumption to make for either one of two reasons. First, if preferences for intervention in foreign humanitarian crises are allowed to differ across countries, then those countries which have been more active in the past may have a stronger taste for intervention than less active countries and will keep having this stronger taste since preferences have not changed. Second, if preferences for intervention are assumed to be equal across countries, then the side constraints they face will still differ. If domestic public pressure has managed to induce certain countries to be more active in the past than others even in the absence of an ICC, then it is likely that this same domestic public will manage to impose enough pressure on its government to remain active in the future as well even after the ICC has been established since the domestic pressure groups certainly regard the ICC as a complement to action, not as a substitute. 1 It is the countries, which have managed to withstand pressure for action in the past that are likely to be tempted by the potential moral hazard created by the ICC. For these countries, the establishment of an ICC could potentially represent a means to deflect criticism toward its inaction, thus enabling them to remain inactive in the future. This is not to say that the establishment of the ICC could not represent a moral hazard in a specific case even for a country, which has been a rather active intervener in the past. Even a generally active country may use the existence of the ICC as a means to deflect criticism toward its inaction due to the particular contextual circumstances of the specific case at hand (for example, because the country is on friendly terms with the offending government). However, one would expect that other more active countries would be willing to step in, unless intervention similarly goes against their interests. What this means, in sum, is the following: If states which have been relatively inactive in the past are the prime supporters of the ICC, then this would suggest that the moral hazard is very relevant, whereas the moral hazard is less relevant if it is mainly supported by countries that have been rather active in the past.
But even then, the moral hazard still exists and may, depending on the circumstances of the specific case, represent more or less of a problem.
A final critical hurdle for the indirect test of the moral hazard problem to work is the presumption that countries are generally under pressure to act in foreign humanitarian crises. If there is no such pressure, then there is no need to deflect 13 criticism for inaction and the ICC cannot exert any moral hazard for countries unwilling to act. The liberal democracies of developed countries are more likely to experience such pressure domestically than developing countries due to their stronger civil society and due to their commitment to human rights protection (Andersson, 2002;Lebovic, 2004). However, even developing countries can be under pressure from the outside. Several observers have attributed the increasing role of developing countries in post-Cold War peacekeeping missions to pressure from developed countries (see, for example, Byman & Waxman, 2002;Pugh, 2004;Neethling, 2004).
The main result of this article's analysis upholds if the sample is restricted to developed countries only. It is thus robust to excluding developing countries from the sample, for which the moral hazard problem may be less of an issue in the first place.

Research Design
As mentioned already, the indirect test of the moral hazard problem works via estimating the determinants of ratification of the Rome Statute. To be precise, I estimate the determinants of ratification delay with an event type model. Such a model uses more information than a pure cross-sectional probit or logit model with a simple ratification dummy as dependent variable. The extra information used is the time delay between the opening of the Rome Statute for signature and ratification until eventual ratification, for those countries that have ratified. The event type model thus not only distinguishes between ratifiers and non-ratifiers, but it further distinguishes among the ratifying countries between early and late ratifiers, assuming that early ratification reveals stronger support for the ICC than late ratification. This can be justified on two accounts: First, early ratification increases the probability that an international treaty will come into force as typically a minimum threshold of ratifications needs to be exceeded (60 in the case of the Rome Statute). Second, the more countries ratify early on, the more credibility the treaty gains and the higher the pressure on remaining states to ratify the treaty as well. Strong supporters of a treaty should thus not merely ratify a treaty, but will be keen to ratify early on.
Data on the year of ratification, acceptance, approval of or accession to the Rome Statute is taken from United Nations (2007). The country year rather than country day forms the basis of our sample since the control variables are only available on an annual basis. Smith (2004) argues that there are distinct regional norms of state participation in the ICC following from preference heterogeneity across regions. To account for this possibility, I use a stratified Cox proportional hazard model, where I stratify by regions following largely World Bank country classification. 2 This allows each region to have its own baseline hazard. Formally, the dependent variable is the conditional probability ρ(t) that ratification occurs at time t given that the country has not ratified before t; this is the hazard of ratification: where i stands for the ith stratum (region), ρ 0i (t) is the "baseline hazard", differing between regions (but uniform within a region), and β β β β T is the vector of parameters to be estimated. The time-variant underlying baseline hazard of ratification may depend on unobserved variables, possibly in complex ways. An advantage of the Cox model is that the baseline hazard does not need to be estimated.
The likelihood function is constructed using the observation that the probability that country i ratifies at time t i equals 2 The difference is that Canada and the US form one region together with Central America and the Caribbean and South American countries constitute their own region.

∑ ∑
(2) The likelihood function to be maximized with respect to the vector β β β β T then The two central explanatory variables are military intervention abroad and contributions to peacekeeping. I use the Uppsala/PRIO dataset as source for the variable measuring intervention in foreign civil conflicts (Gleditsch et al., 2002). The specific measure used is the sum of country years in which a country intervened in a foreign civil conflict since 1990, weighted by conflict intensity. The decision to use 1990 as a cut-off point is to account for the fact that the world fundamentally changed after the end of the Cold War. The measure is weighted by conflict-intensity since it takes much more courage and commitment to intervene in more intense foreign conflicts. No attempt was made to distinguish "humanitarian" intervention from "selfinterested" interventions since there is no objective way of doing so. Data on contributions to peacekeeping are taken from a database of the Stockholm International Peace Research Institute (SIPRI), available at http://conflict.sipri.org/SIPRI_Internet/. The specific measure used is the sum of country years in which a country contributed troops to a multinational peacekeeping operation abroad since 1990, whether led by the UN, by a regional inter-governmental organization or by an ad hoc group of states.
There are six control variables. Democracy is measured by Polity data and accounts for the fact that democracies are much more likely to ratify international human rights treaties (Landman, 2005). Per capita income accounts for the possibility that demand for the ICC may be a normal good, using data from World Bank (2006).
The remaining control variables are meant to capture the expected future potential 16 costs of ratification to the ratifying state, its leaders and citizens. Note that since no offences prior to the entry into force of the Rome Statute can be prosecuted and the Court will exert its major impact only in years to come, the variables I use approximate the future costs of ratification by looking at potentially offensive behaviour in the past. This will be a good proxy for expected future costs of ratification only if one is willing to make the assumption that states (or rather their political leaders), which have engaged in potential offences in the past, have a taste for and are therefore likely to commit such potential offences in the future.
To start with, I include a variable measuring the sum of country years since 1990, in which a genocide or politicide has taken place in a country, weighted by the intensity of the crime, with data taken from Harff (2006 To our knowledge, there are no data measuring war crimes. In its absence, I use data on conflict involvement. Put simply, the argument is that nationals from a country that is involved in more and more intense armed conflicts are more likely to be indicted for war crimes. The link may be weak, but I see no other way of capturing this aspect. The specific variables I use are the sum of country years since 1990, in which a country was involved in an international or civil armed conflict, weighted by conflict intensity, using again data from the Uppsala/PRIO project. The last control variable is the number of times a state has initiated an international armed conflict since 1990, using and slightly extending data from Gleditsch (2004). Whilst, as discussed above, aggression is not a fully specified crime under the auspices of the Rome Statute yet, this may change in the future, which may deter states with a preference for initiating international conflicts from ratifying. Specifically, if the moral hazard argument applies, then those states which have been rather reluctant to intervene militarily and to contribute to multinational peacekeeping in the past should also be the prime supporters of the ICC and therefore be the first ones to ratify its Rome Statute. However, I find the exact opposite. Those states that are willing to play an active role in interventions and peacekeeping are also, on average, the prime supporters of the ICC. It would therefore appear that the ICC is not a sorry excuse for inactive states, but another strategy for active states to combat situations in which acts of genocide, war crimes as well as crimes against humanity can happen and can be committed with impunity.

Results
Yet, given the indirectness of our test for the moral hazard created by the ICC, only the future will tell whether the ICC will complement or substitute action. The 20 current situation in Darfur, Sudan, does not provide a clear picture. On the one hand, several hundred thousand people have already reportedly been killed and while there is the will from several, mainly Western and African states to intervene in a meaningful way, this has yet to happen. The case is under ICC investigation following a referral by the UN Security Council, but it appears that this move was part of a strategy to increase pressure on the Sudanese government to facilitate real protection of people at risk rather than a substitute for intervention.
Another caveat is that there remains the possibility that the existence of the ICC creates an incentive even for a state, which has been rather active in the past, not to intervene in a specific case if intervention goes against the state's interests. That is, even if in general the ICC does not represent a moral hazard, it can do so for specific countries in specific cases. However, on the whole the fact that the ICC is primarily supported by states, which have been active in the past, suggests that it will function as a complement to, not a substitute for action. If accompanied by action, the ICC can form an important step forward in the protection of victims. Our estimation results paint a cautiously optimistic picture in demonstrating that the ICC has been ratified early on by states, which have been willing to act in the past. This bodes well for the ICC and its future contributing role to the fight against heinous political crimes, despite continued opposition to the Court from the US in particular, but many other countries as well, including China and India. ***, **, and * indicate significance at the .01, .05, and .10 levels, respectively.