Forthcoming: Journal of Genocide Research

Transcription

1 THE EFFECTS OF DIPLOMATIC SANCTIONS AND ENGAGEMENT ON THE SEVERITY OF ONGOING GENOCIDES OR POLITICIDES Matthew Krain Department of Political Science The College of Wooster Wooster, Ohio USA [Phone] [Fax] Forthcoming: Journal of Genocide Research

2 BIOGRAPHICAL NOTES Matthew Krain is Professor of Political Science at the College of Wooster. His research examines the causes and consequences of repression and large-scale human rights violations, and the role of the state and other actors in the international community in causing, preventing or mitigating the severity of conflict, violence and genocide. ABSTRACT This study tests the effects of diplomatic sanctions or engagement on reducing the severity of ongoing instances of genocide or politicide. I argue that neither diplomatic measure will be effective in slowing or stopping the killing. I argue that diplomatic sanctions merely reduce the flow of information without credibly signaling intent or commitment, while diplomatic engagement does not challenge perpetrators. Neither policy raises the costs of perpetrating genocide or politicide. Therefore neither is expected to be useful in mitigating ongoing atrocities. Ordered logit analyses of ongoing genocides and politicides from 1976 to 2008 confirms these assumptions, and demonstrate that changing levels of diplomatic representation with atrocity perpetrators may make policymakers feel like they are doing something, but does little to reduce the lethality of ongoing mass killing. Under one set of circumstances, increased engagement even exacerbates the atrocities. 1

3 THE EFFECTS OF DIPLOMATIC SANCTIONS AND ENGAGEMENT ON THE SEVERITY OF ONGOING GENOCIDES OR POLITICIDES Foreign policy leaders and international institutions see diplomatic action as a useful tool to counter instances of ongoing atrocities. For instance, in November 2011, US Ambassador to the United Nations Dr. Susan Rice outlined a range of options available to the international organization and its state members to protect civilians during ongoing atrocities, including naming and shaming, diplomacy, sanctions and intervention. 1 This was echoed later in President Obama s acknowledgement that we possess many tools diplomatic and political, and economic and financial and our moral suasion that could be effective in the face of genocide or politicide. 2 These statements of US policy themselves echo the role of diplomatic (dis)engagement in sections of the UN charter, and in the emerging norm of a Responsibility to Protect (R2P). Both suggest that an imperative to first diplomatically engage, and later, if needed, to diplomatically disengage and isolate targets. Before authorizing force the UN charter requires that the security council attempt non-military strategies such as complete or partial interruption of economic relations and of rail, sea, air, postal, telegraphic, radio, and other means of communication, and the severance of diplomatic relations. 3 More recently, in a report on the Responsibility to Protect, UN Secretary-General Ban Ki-Moon wrote that when perpetrator States do not respond to diplomatic engagement collective measures may be authorized by the Security Council, under Articles 41 and 42 of the Charter [including] limiting diplomatic contact of States with a target entity. 4 2

4 Scholars and NGOs have also called for diplomatic action in the face of atrocities, though their recommendations often differ. In the face of continued targeting of civilians by the Assad regime in Syria, noted sanctions and human rights scholar George Lopez wrote: To maximize the isolation that Assad s actions warrant, all embassies in Damascus should close their doors and formally withdraw their ambassadors. 5 Conversely, the Executive Director of the Global Centre for the Responsibility to Protect called for a diplomatic surge rather than disengagement, as the only way out of the crisis. 6 Yet it is unclear as to whether either diplomatic sanctions or increased engagement have any mitigating effect in the face of ongoing mass killing. Tara Maller argues that diplomatic sanctions against Sudan did little to avert the slaughter there, and that while subsequent diplomatic engagement helped facilitate negotiations to end the Sudanese Civil War forward, it did little to address the killing in Darfur. 7 And continued slaughter in Syria, despite the withdrawal of US, EU and Gulf Cooperation Council States ambassadors in protest, seems to underscore Maller s point about the seeming futility of using changes in diplomatic representation as a way to mitigate mass murder. 8 Are these cases anomalies? Does this foreign policy tool hold some potential to ameliorate the worst atrocities? While we now know that perpetrators of mass killing may be sensitive to economic pressure in the form of sanctions, military pressure in the form of intervention, and political pressure in the form of naming and shaming, we do not yet understand the effects of diplomatic sanctions or engagement on the severity of ongoing mass killing. 9 This study aims to fill this gap. Diplomatic sanctions, diplomatic engagement and ongoing geno/politicides 3

5 Thanks to a series of disastrous misadventures as de facto bystanders the mass murder in the 1990s, recent successes in acting on behalf of targets of ongoing atrocities, and an emerging body of research on the effectiveness of such policies, in the face of genocide or political mass murder it is more difficult for policymakers to plead that we have no effective means of response. Given the recent commitment by the international community to the R2P doctrine, it is incumbent upon scholars and policymakers to continue to work on determining how best to respond in situations of ongoing systematic state-sponsored mass murder. One set of foreign policy options available to states in such instances is the use of diplomatic engagement and/or diplomatic sanctions. 10 Engagement is a foreign policy strategy which depends to a significant degree on positive incentives to achieve its objectives. 11 Types of diplomatic engagement that hold out the greatest allure to potential target regimes include normalizing diplomatic relations with a state, upgrading the diplomatic presence in that state, or regularizing interaction between high and low level diplomats. These actions legitimize the actor and signal a willingness to pursue a diplomatic resolution to disputes. They also insure more communication of information between actors, and may facilitate more effective negotiation and mediation. 12 Conversely, diplomatic sanctions include severing formal diplomatic ties with a country or significantly downgrading ties from the normal level of diplomatic activity for foreign policy purposes. 13 This is usually done to signal disapproval of the target regime s behavior, and to also signal the possibility of subsequent punitive measures should that behavior not change. Neither diplomatic engagement nor diplomatic sanctions rule out other options. Indeed, in practice, there is often considerable overlap of strategies. 14 Unfortunately, we have little information about the effect of either set of diplomatic actions in situations of ongoing mass killing. 15 4

6 On the other hand, we know a lot about other types of policies that members of the international community can adopt. One approach that has been shown to slow or stops the killing once it has begun, at least within a year after the act, is overt military intervention against perpetrators. 16 In the long run, however, neutral interventions may prove more effective in ameliorating or terminating atrocities. 17 Economic sanctions, though often ineffective in making headway on human rights issues, 18 have been found to force perpetrators of mass killing to reduce the duration of these ongoing atrocities, 19 though not their severity. 20 And naming and shaming by the media, and international governmental and non-governmental organizations may ameliorate human rights violations more generally, though its effectiveness is still debated. 21 Encouragingly, recent research suggests that naming and shaming can play a role in reducing the severity of campaigns of mass killing, 22 though not its duration. 23 These studies share some important assumptions that have guided much of the recent work on slowing or stopping mass killing. Such work has assumed that the choice to slaughter one s own citizens is a rational, if horrific, choice. The choice to engage in atrocities is a policy choice, one with clear goals. As such, the violence is instrumental. The implications of these assumptions are that perpetrators of atrocities (usually the state) will slow or stop the killings when any or all of the following conditions apply: (1) they have achieved their goals; (2) they are divided amongst themselves, and thus cannot effectively implement the policy; (3) the resources to carry out killings or otherwise enforce the policy are depleted; (4) the costs of killing increase, or other alternatives that might reach the same goals become relatively less costly; or (5) when perpetrators are defeated. Expected effects of diplomatic sanctions and engagement on atrocity severity 5

7 While diplomatic sanctions may be an attractive policy option for countries interested in doing something or signaling displeasure, and certainly adds to the sense of isolation of the perpetrator regime, it does not affect any of the above conditions. Policymakers see diplomatic sanctions or disengagement as largely symbolic, and as a low cost means of isolating and delegitimizing regimes. 24 Because they are lower in cost as compared to alternative actions such as economic sanctions or intervention, states sometimes use diplomatic sanctions for political or strategic reasons as cheap talk so as to be able to claim that they acted without actually having to do anything, or for other domestic political reasons. 25 Yet, regardless of why they are employed diplomatic sanctions neither impede nor divide perpetrators, and may actually increase cohesion given the perception of hostility from abroad. Even advocates argue that diplomatic sanctions are not likely to be successful on their own but they create severe inconveniences for the country s politicians and bureaucrats and cannot be ignored. 26 While other policies designed to signal displeasure and publicly highlight atrocities such as economic sanctions or naming and shaming carry with them economic or political costs, and thus have the potential to deplete much needed resources and raise the costs associated with continuing the killing, diplomatic sanctions do not. However lower in cost diplomatic sanctions are relative to other policies, it is not cost free. Diplomatic representation itself is costly, but occurs because it is beneficial to both the sender and the target. Thus, severing ties can be a costly message. 27 Those considering downgrading diplomatic relations, withdrawing ambassadors, or closing embassies understand that there are advantages in having diplomatic staff on the ground to keep track of events. 28 These advantages include the ability to gather accurate information, communicate threats or promises clearly, and engage in public diplomacy. 29 Indeed, it is possible that rather than 6

8 reducing the killing, diplomatic sanctions may even make things worse, particularly if the target of is unclear as to the message being sent by changes in representation, or what it needs to do to reverse these diplomatic sanctions. The lack of clarity is likely to be exacerbated with fewer (or no) skilled diplomats on the ground to communicate the message to the target. Ultimately, diplomatic sanctions might prove disadvantageous if they diminish contact with the country and thus potential influence. 30 Diplomatic sanctions may also increase the target regime s sense of isolation, as well as their perception that they are no longer as intensely under the spotlight. The lack of high-level attention may suggest to perpetrators that their actions are more hidden from scrutiny, 31 or at least more easily deniable. Perpetrators are likely to continue or even escalate the killing if they feel that the spotlight has shifted, 32 or if states have more difficulty in verifying reports of atrocities. On the other hand, increased diplomatic engagement might put regime actions increasingly under the spotlight, and might enable other international actors that disapprove to more clearly communicate with the perpetrators. As such, engagement seems to solve some of the issues noted above when discussing diplomatic sanctions. However, it might not be reasonable to presume that information, communication an diplomatic ties can thwart an ongoing genocide. 33 Engagement does nothing to affect the factors that contribute to the killings. It does not help the regime achieve its goals, nor does it do anything to divide the decision-makers. In fact, international recognition may actually enhance the legitimacy of the regime with important domestic constituents, or even reinforce their image of themselves as actors behaving legitimately, 34 strengthening rather than undercutting the elites. Perpetrators faced with increased engagement may begin to tally their diplomatic partners as potential friends (or at least not foes) 7

9 as they calculate the future costs of continuing their actions. Thus increased diplomatic engagement with a perpetrator might signal the international community s willingness to tolerate the atrocities without further international action to stop them. In the presence of bystanders, mass killing is likely to continue unabated. 35 Finally, engagement does not deplete the resources needed to carry out killings, nor does it raise the cost of the policy. Much of the research to date suggests that attempts by international actors to slow or stop the killing in ongoing genocides or politicides should focus on raising the costs of such murderous policies for the perpetrators. 36 To do so effectively, however, actors must signal a shift in the global context from permissive to prohibitive, make visible the actions occurring on the ground, challenge perpetrators, and make any current or future threat of action against perpetrators credible. 37 Neither diplomatic engagement nor diplomatic sanctions fulfill these criteria. As such, these policies may be ineffective when whole populations are threatened with extermination by their own governments. 38 I expect neither to reduce the severity of ongoing mass killings. Hypothesis 1: (Increased) diplomatic sanctions from perpetrators of an ongoing genocide or politicide will not reduce the severity of genocide or politicide. Hypothesis 2: (Increased) diplomatic engagement with perpetrators of an ongoing genocide or politicide will not reduce the severity of genocide or politicide. Alternative arguments and expectations 8

10 On the other hand, some believe that the international community is much more likely to induce positive change in problematic regimes through diplomatic engagement rather than through condemnation and coercive measures; coercion and isolation, it is felt, only heighten the regime s intransigence. 39 Indeed, this might be true under a variety of circumstances, particularly if the optimal approach to ameliorating atrocities does not include challenging perpetrators. One such option is to negotiate or intervene or otherwise deal with any underlying conflict via peace talks, diplomatic pressures or incentives, or via the use of neutral or impartial interventions. 40 Such a strategy seems to require at minimum no disengagement diplomatically, and ideally increased diplomatic engagement with the perpetrator regime (as well as other members of society, including guerrillas and victim groups). Alternatively, it might be more effective to try to resolve the state failure or civil conflict to remove the underlying threat and instability that drives the killing. 41 Such an approach might focus on diplomatic intervention that specifically tries to resolve the concurrent ongoing civil war, so as to also mitigate mass murder. It, too, seems to require at minimum no disengagement diplomatically, and ideally increased diplomatic engagement. Either way, increased diplomatic engagement may be effective in facilitating clearer signaling on the part of potential interveners, making the consequences of not changing behavior clear to perpetrators, and allowing actors to more clearly signal the credibility of their commitment to their positions. It may also facilitate bargaining that increases the relative benefits of alternative solutions for perpetrators, or helps them find other ways to achieve their goals. 42 As Maller argues, diplomatic engagement should by no means be expected to be a panacea for thwarting genocide... but it should be recognized as having utility compared to isolation-based strategies and simply imposing costs. 43 9

11 Alternatively, diplomatic sanctions might be useful early signals of interest and awareness on the part of actors in the international community, and clear and costly signals of commitment on their part. It may help them communicate clear intention to escalate beyond diplomatic sanctions to other types of action, thus changing the projected costs of continuing the slaughter. In short, there reasons to expect that ether diplomatic engagement or diplomatic sanctions might play some role in reducing mass killing. Alternative Hypothesis 1: (Increased) diplomatic sanctions from perpetrators of an ongoing genocide or politicide reduces the severity of genocide or politicide. Alternative Hypothesis 2: (Increased) diplomatic engagement with perpetrators of an ongoing genocide or politicide reduces the severity of genocide or politicide. Conditional effects of diplomatic sanctions and engagement on atrocity severity Another way in which diplomatic sanctions or engagement might have effects is in combination with other strategies. Even if ineffective on its own, changes in levels of diplomatic (dis)engagement may work as part of a set of policies. 44 These actions might either reinforce or limit the effectiveness of other policies available to states or non-sate actors. In trying to mitigate mass murder, interveners need to be ready to employ economic as well as diplomatic incentives and sanctions. 45 The civil war literature suggests that diplomatic action is most effective when employed in conjunction with other policies. 46 The optimal policy mix in the face of ongoing mass killing is unclear, however. Maller s research suggests that to insure a greater chance of foreign policy success it is optimal to remain 10

12 diplomatically engaged when employing other punitive measures such as economic sanctions or military intervention. Her study does not, however, suggest whether there are benefits to mixes of policies in which there are increases in overall engagement with the target state. 47 This might mean that further engagement combined with alternatives mitigates mass murder, or it might suggest only that disengagement will make other options less effective. That same study found that diplomatic sanctions decreased the effectiveness of economic sanctions by making it harder for the sender to gather intelligence about target actions or sanction effectiveness, communicate their desired outcomes, and maintain influence overall in the target state. If a target is unclear as to what it has to do in order to comply, it makes it unlikely that targets will successfully change their behavior in reaction to sanctions. 48 It also makes it harder to assess target state vulnerabilities, tailor appropriate sanctions policies to the situation, and gauge the target s reaction to sanctions and assess their impact. 49 Diplomats often rely on economic sanctions as an element of leverage in negotiations. 50 In part this is what makes sanctions effective not their effects directly on the regime, but their presence as a stick, but also a carrot (their removal) during diplomatic efforts at mediation or negotiation. When diplomatic representation is downgraded or eliminated, negotiation is made more difficult and sanctions may no longer serve that function. 51 One might expect similar problems in gathering intelligence and insuring effective information flows when disengaging diplomatically while simultaneously engaging in any type of overt military intervention. It is unclear whether these problems that come with diplomatic sanctions would be enough to decrease the effectiveness of the other policies that they accompany. Yet it is hard to see how diplomatic sanctions would facilitate more effective economic sanctions or intervention policies. Additionally, diplomatic sanctions do not signal 11

13 high levels of commitment on the part of states to a given issue, nor do they raise the costs of genocide or politicide. Therefore, I expect that diplomatic sanctions will limit, rather than enhance, the effects of economic sanctions or overt military intervention on the severity of ongoing mass killing. It is less clear as to the effect that increased diplomatic sanctions should have in the context of increased pariah status as a result of naming and shaming. 52 On the one hand, it is possible that isolation by states (diplomatically) further reinforces the isolation and labeling by non-state actors (naming and shaming), enhancing any effects of the latter policy. On the other hand, when states diplomatically sanction perpetrators, they make it harder for non-state actors to be able to leverage interconnections between perpetrators and other members of the international community. When human rights organizations, the media, or international organizations name and shame perpetrators, high level diplomats can play a problem-solving role by helping perpetrators avoid the spotlight by reducing violations. 53 They can enable naming and shaming to work by providing perpetrators a way out, reducing the carnage in the process. But diplomatic sanctions make that more difficult, which then makes it harder for perpetrators to find facesaving ways to change their behavior in the face of naming and shaming. The effects of increased engagement are equally unclear. As previously noted, diplomatic engagement by states in an environment in which regimes are identified by transnational advocacy networks as perpetrators and norm violators might enable transnational advocacy networks (TANs) to pressure states to use those connections to affect change in the perpetrator. However, engagement might also undercut the framing by TANs, yielding competing frames as to whether the perpetrators can be trusted. It may also signal an unwillingness to act on the part of sender states, despite pressure by TANs to do so. 12

14 Increased engagement could allow for more accurate assessment of sanctions effects, and clearer communication between sender and target, thus making sanctions more effective. 54 However, these seemingly conflicting policies (diplomatic engagement, economic sanctions) may also send mixed signals as to the importance of the issue in contention to the sender, emboldening the perpetrator to continue to slaughter. Similar dynamics could be envisioned regarding the interplay between diplomatic engagement and intervention. Engagement facilitates more accurate signaling of intent and commitment on the part of interveners, but sends mixed signals. Such dynamics often appeared to play out in the Darfur case, where repeated attempts at engagement in the context of a proscribed sanctions regime and a neutral intervention seemed to signal that the regime could act with impunity. In sum, though there are compelling arguments as to what we might expect regarding the effects of the interaction between these diplomatic options and other types of interventions by outside actors, there is no clear set of expectations regarding their likely effects. Therefore I do not develop clear hypotheses, and instead propose what amounts to a brush-clearing exercise. Methodology Unit of Analysis In this study I examine country-years in which a genocide or politicide is ongoing. 55 Data for the independent variables must be lagged to avoid attributing changes in the dependent variable to things that have occurred after it has been measured. I also include the year immediately following the end of that instance of state-sponsored mass murder. This allows me to examine the effects of these policy options and their various combinations in the final year of the genocide or politicide. 13

15 Dependent variable: genocide/politicide severity The Political Instability Task Force (PITF; formerly the State Failure Task Force) has developed a list of all genocides and politicides from This list is an updated version of the data previously compiled by Barbara Harff, which has been used as the basis for a number of comparative empirical studies examining genocide and politicide. 57 This study will examine cases from 1976 to 2008, as the earliest data available for some of the main independent variables is Since these independent variables must be lagged one year prior to the caseyear observation of the dependent variable, I can examine the effects of naming and shaming in 1975 on genocide/politicide severity in The list of cases examined here is reproduced here in Table 1, including location and start and end years of the instance of state-sponsored mass murder. << Tables 1 and 2 about here >> The data set also includes information about the magnitude of severity of the genocide or politicide. All data on the severity of state-sponsored mass murder have an inherent flaw the more successful the policy, the less we may know about it. In particular, body counts, the usual measures of severity, are problematic. 58 There will be underreporting of fatalities by the aggressors in most cases in order to hide the extent of the atrocity, and over-reporting of fatalities by the victims in order to highlight to degree of the atrocity. 59 The particular subset of the Political Instability Task Force data employed here consists not of exact body counts, but a more approximate scaled magnitude index representing the likely range for the estimated number of 14

16 deaths per year. 60 I adopt the PITF severity scale (Table 2) as the dependent variable in this study. The eleven-point scale is recoded to range from zero to ten, rather than zero to five, eliminating half-point changes in magnitudes to make interpretation more intuitive. Note that the dependent variable is an ordinal categorical variable, suggesting the necessity of an appropriate statistical method. Therefore, I employ ordered logit as my method of statistical analysis in this study. Independent variables: diplomatic engagement and diplomatic sanctions To operationalize diplomatic engagement and diplomatic sanctions against perpetrator states I employ the Correlates of War s (COW) Diplomatic Exchange data set, which tracks diplomatic representation at the level of chargé d'affaires, minister, and ambassador between states from The data set allows one to see whether diplomatic relations existed between states, and if so at what level. Comparing years also allows one to see changes in diplomatic exchanges over time, including diplomatic engagement (such as the US decision to send an ambassador to Myanmar to reward the government there for releasing political prisoners) or diplomatic isolation or sanctions (such as the October 2011 recall of the US ambassador to Syria to protest the ongoing carnage there, or the Bush administration s decision to diplomatically isolate rogue regimes). For the purposes of this study, the data of interest involves the yearly change in the level of diplomatic representation by other states in perpetrator states. Downgrading representation (or eliminating it entirely) suggests punitive diplomatic sanctions, while increases in representation suggest diplomatic rewards through higher-level interactions. Unfortunately, the very helpful diplomatic exchange data is only available for every five years from 1950 until 2005, limiting its 15

17 utility for this project. Using the COW approach, I coded the level of diplomatic representation that all states had with atrocity perpetrating states for each year for the country-years of interest to this study, from 1975 until I then created a series of variables: diplomatic engagement is a count of the number of sender countries that upgraded their diplomatic relations with the perpetrator regime in that year; diplomatic sanctions, is a count of the number of sender countries that downgraded their diplomatic relations with the perpetrator regime in that year; and overall diplomatic (dis)engagement a single measure capturing the yearly total change in levels of diplomatic (dis)engagement with a given perpetrator regime, measured as diplomatic engagement minus diplomatic sanctions. I also created three additional measures of these same variables where each act of engagement or disengagement was weighted based on by how much the diplomatic status of the sender s top diplomat in the target s country changed. Relying on the COW scale, representation level codings included: 0 = no representation; 1 = chargé d'affaires; 2 = minister; and 3 = ambassador. Changes from one level to the other were indicated by subtracting representation in year t from representation in year t These were then summed within each country year to produce: diplomatic engagement, weighted; diplomatic sanctions, weighted; and overall diplomatic (dis)engagement, weighted. Finally, I create interaction terms between the relevant diplomatic representation variable (engagement or sanctions) and another type of action previously found to affect the severity of ongoing mass killing. In all I create eight interaction terms, four each for engagement and disengagement. Each is coupled with naming and shaming, challenging interventions, neutral intervention, and economic sanctions. These interaction terms allow me to observe the conditional effects that diplomatic (dis)engagement has on the effectiveness of these other policy 16

18 tools that have previously been found to affect the severity of ongoing mass killing. I include constitutive parts of all interaction terms in the relevant statistical models. 64 Control variables: other policy options to slow or stop ongoing atrocities To capture the effects of overt military intervention by international actors, I employ measures from the International Military Interventions (IMI) data set, which codes all overt International Military Interventions (IMI) from 1946 to The IMI data set records in what direction the intervener acts. I compress a number of IMI categories into three main intervention types: I code all interventions that are explicitly anti-perpetrator, as well as those that are pro-target, as antiperpetrator interventions. I code as pro-perpetrator interventions all interventions that are explicitly pro-perpetrator, as well as those that are anti-target. Those that support neither side explicitly or are expressly impartial are coded as neutral interventions. I examined all other interventions (including those that support or oppose third party governments, or those that oppose rebel groups in sanctuary) and recoded them as to whether they indirectly were properpetrator or pro-target. I was thus able to code the number of interventions for and against the perpetrator, as well as those that were impartial. 66 To extend the data past 2005, in consultation with the IMI directors, I applied their coding scheme to the lone post-2005 case in my sample Sudan/Darfur. I operationalize economic sanctions by using the Threat and Imposition of Economic Sanctions (TIES) dataset. 67 For the purposes of this study, TIES is more useful than the data compiled by Gary Hufbauer and his colleagues. 68 Although the latter dataset is often used in research on the effects of sanctions, it does not code instances of threats of sanctions. The TIES data set defines economic sanctions as actions that one or more countries take to limit or end 17

19 their economic relations with a target country in an effort to persuade that country to change its policies. 69 The TIES dataset includes both multilateral and bilateral sanctions, with sanctions or threats against each individual target coded as separate events. TIES codes actions related to a wide variety of sender demands, 70 but given the nature of this study I follow Davenport and Appel s approach I include only threats or imposition of sanctions related to human rights violations or modifying target government coercive behavior. 71 TIES data is only available for sanction threats and actions from , so I employed their coding scheme to extend the data through 2008 (to the Angola and Sudan/Darfur cases). I checked my results through 2006 with the data produced by Gary Hufbauer and colleagues to confirm accuracy. 72 Another choice international actors have is to use information about abuses to frame perpetrators as norm violators, and to signal both disapproval and possible negative sanctions to follow. To capture such naming and shaming activities I employ a measure of the number of Amnesty International background reports produced per country year. This measure, originally developed by James Ron and his colleagues, has been used extensively in the literature to examine how naming and shaming affects human rights. 73 Recent research has also shown that Amnesty background reports have a negative effect on subsequent mass killing magnitude specifically. 74 Ron and his colleagues code all background reports and press releases found in the Amnesty International cumulative guide from 1986 to I employ this data, plus my previous extension of the data backwards to 1975 and forwards to 2008 for all cases of ongoing genocides and politicides. 75 Control variables: factors affecting atrocity severity or effects of policy options on atrocities 18

20 I also control for some key characteristics of the ongoing genocide or politicide itself that are likely to have an effect on its magnitude of severity the duration of the genocide or politicide and its previous magnitude of severity. Many of the policy options noted above affect the duration of civil conflicts more generally, 76 and genocides and politicides specifically. 77 Moreover, duration itself affects the severity of ongoing atrocities. 78 In order control for these effects, I code in which year of the genocide or politicide the observation occurs. For example, the first year of the genocide or politicide is coded as one, the second year as two, and so on. I also lag the dependent variable by one year to enable me to control for the effects of prior severity of genocide or politicide on current genocide or politicide severity. This helps to control for the effects of autocorrelation, and for the finding that previous levels affect current or future levels of atrocities. 79 Following what has become standard practice in the quantitative human rights literature I also control for the population of the country experiencing the atrocity by taking the natural log of the data to account for the skewed nature of such data. 80 I use population data from the World Bank s World development indicators, 81 and lag the data one year to insure that population affected this year s atrocity observation rather than that this year s killings depleted the population being observed. Accounting for total population controls for the possibility that the sheer number of people available to kill, and the pressures that large populations place on regimes and their resources, may affect severity of atrocities. Genocides and politicides are almost always a consequence of other types of state failure such as civil war. State failures promote domestic instability, and open windows of opportunity during which murderous policies become more likely. 82 Evidence suggests that recent state failure also increases the death toll in campaigns of mass killing. 83 I employ a dummy variable 19

21 measuring whether the state experiencing genocide or politicide is also experiencing another kind of state failure in the prior year. The Political Instability Task Force developed a list of all state failures from 1955 to State failures are defined as all revolutionary wars, ethnic civil wars, genocides or politicides, and disruptive regime transitions. 85 Since all cases examined here are ongoing genocides or politicides, this variable would be a constant if it included the ongoing state-sponsored mass murder. For the purposes of this study, I exclude the ongoing genocides or politicide from the state failure variable. Serious political and existential threats to leaders have been found to trigger genocide and politicide, and to increase the severity of ongoing mass killings. 86 As Wright and Escribà-Folch argue, an increased risk of irregular exit may therefore provide leaders with an incentive to increase repression in an effort to remain in power and forestall a particularly nasty post-exit fate. 87 In the context of such extensive threats to power, attempting to use sanctions to compel a regime to hand over power or territory to its mortal enemy can backfire catastrophically. 88 Given its predictive power regarding severity of ongoing atrocities, and its possible confounding effects on the utility of options meant to slow or stop the killing, such threats need to be accounted for explicitly. To do so, I employ Marshall and Marshall s data on the number of successful, attempted, plotted or alleged coups experienced in the prior year to capture direct internal threats to leadership and (the possibility of) extra-constitutional changes. 89 Regime type has been found by some researchers to play a direct role in affecting the overall severity of mass killing and an indirect role via conditioning how naming and shaming impacts regime choices to improve human rights compliance, even among non-democracies. 90 And Hafner-Burton and Ron note that most statistical work finds, when adequately distinguishing between the two, that regime type is a more important explanatory variable than 20

22 international human rights advocacy or policy. 91 Therefore, I control for regime type here. I employ the recommended Polity IV data set measure of regime type, the Polity IV democracy measure minus the autocracy measure. 92 The regime type score ranges from negative ten to positive ten, with lower scores denoting more autocratic states and higher scores more democratic states. I employ the recommended coding rules for transitional states (interpolation) and those in a state of interregnum (set at zero). 93 Evidence also suggests that states that are less open to trade, and thereby less connected to the global economic system, are more likely to experience instances of state-sponsored mass murder. 94 While economic marginalization is not a significant determinant of genocide or politicide severity, 95 it may affect the presence or intensity of international pressures faced by perpetrators or the degree to which regimes are sensitive to those pressures. 96 Therefore, I control for the level of international economic interconnectedness, measured as the degree of marginalization within the world economy. This is operationalized as a function of that country's percentage of world trade, computed by calculating the total imports plus total exports of a given nation divided by the total imports plus exports of the world economy. The inverse of that figure is then divided by 100, yielding the economic marginalization score. The score is thus the degree of marginalization low scores on this index indicate greater centrality within the world economy; high scores indicate greater marginalization. The data was collected from the International Monetary Fund s Direction of trade statistics yearbook. On the other hand, trade dependence has little effect on the success of sanctions once implemented, and there is no evidence that that foreign investment affects state repression. 97 This suggests that if we wish to control for economic factors that might condition the effect of interventions on severity of state-sponsored atrocities, alternative measures of economic 21

23 interconnectedness may be in order. One such candidate is foreign aid. Countries heavily dependent upon aid flows will be more vulnerable to human rights pressures than those not receiving such flows. 98 And DeMeritt finds that [a]s perpetrators grow increasingly dependent on foreign financial support, they become decreasingly willing to risk that support by executing noncombatants. 99 All of this suggests the need to control for foreign aid flows. I operationalize this variable by using the World development indicators measure of official development assistance (ODA). I use the natural log of ODA to account for skewness, and lag the data by a year to insure that I am not accidentally picking up changes to ODA resulting from reaction to prior geno/politicidal activity. Finally, the end of the Cold War changed the geopolitical realities in the international system, and placed international organizations and non-governmental organizations in a position of greater influence than they had during the prior era. 100 Moreover, this momentous change saw greater emphasis on global human rights norms, greater need to conform to liberal standards of rights, and the removal of Cold War strategic or ideological limitations on condemning in words or in policy changes the human rights behavior of other states. 101 Not surprisingly, it also led to an escalation in the use of economic sanctions and overt military intervention through international institutions to achieve a variety of ends, not the least of which was affecting states respect for human rights. 102 In order to account for the potential temporal differences caused by changes in the structure of the international system and the resulting effect on actor behavior, I created a dummy variable for the Cold War, with the years coded as one and the years coded as zero. Results 22

24 Below, I present the results of ordered logit models of factors affecting the severity of ongoing instances of genocide or politicide. 103 Models are estimated using STATA, version 11.2, and using the White estimator of robust standard errors to correct for heteroskedasticity. White estimators of variance are particularly useful when estimating ordered logit models using unbalanced panel data (each panel has a different number of observations because each genocide/politicide lasts a different number of years). 104 White robust standard errors help to produce estimates that account for the fact that observations are likely to be independent across countries but not within them. 105 I used test a variety of expectations regarding the effects of diplomatic engagement or diplomatic sanctions on the magnitude of severity (Tables 3-6). All models use the same set of control variables, but each includes different versions of the diplomatic action variables. The first set of models, depicted in Table 3, examine the effects of the number of times states engage or disengage diplomatically on the severity of the atrocities. In Table 4 each change in diplomatic representation is weighted based on by how much the diplomatic status of the sender s top diplomat in the target s country changed. Tables 5 and 6 include interactions in each model between the relevant diplomatic representation variable (engagement in Table 5, and sanctions in Table 6, respectively) and another type of action previously found to affect the severity of ongoing mass killing. The constitutive parts of all interaction terms are included in each of the statistical models in Tables 5 and The results reported in Tables 3 and 4 appear to lend support to my hypotheses. Neither diplomatic engagement nor diplomatic sanctions have significant effects on genocide or politicide severity, regardless of whether they are examined separately, together or as part of a continuum of an overall level of diplomatic representation, and regardless of whether they are 23

25 counted or weighted. Even when controlling for changes in diplomatic representation, challenging interventions and naming and shaming continue to have ameliorative effects on mass killing severity. Similarly unaffected is the significance of the escalatory effects of the lagged dependent variable. Duration of mass killing decreases severity in the models using either diplomatic engagement variable or its weighted version (the first model in Table 3 and Table 4). No other variables in the models are significant. In Tables 5 and 6 examine the conditional effects of diplomatic engagement or sanctions on other types of action previously found to affect ongoing mass killing naming and shaming, economic sanctions and challenging interventions. I also include an interaction between neutral interventions and levels of diplomatic representation given significance of neutral interventions found in other studies, 107 as well as the arguments put forward by advocates of neutral or nonpartisan diplomatic and military action discussed above. Most of the results presented in these two tables seem to also support the hypothesis that changes in diplomatic representation are not costly enough to affect the calculus of mass killing, nor will they affect other ongoing efforts to counter atrocities. Recall that even Maller who argues that diplomatic sanctions can make economic sanctions less effective, does not expect such action to have the same effect in instances of genocide or political mass killing. 108 Indeed, changes in diplomatic representation have no statistically significant conditional effects on economic sanctions, nor do they condition challenging interventions or naming and shaming efforts. Diplomatic sanctions also have no conditional effects on neutral interventions. The lone significant interaction is between diplomatic engagement and neutral interventions one that escalates the killing. Why might this be the case? On their own, neutral interventions have either been found to have no effect or a dampening effect on state-sponsored 24

26 mass murder. 109 They may not be effective, but they do not make things worse. Efforts by states to engage perpetrators diplomatically may be signals that perpetrators will not face consequences for their actions. Alone such signals have little effect on perpetrator decisionmaking. However, when coupled with similar signals by military forces on the ground in a neutral intervention they may create an environment in which perpetrators see a window of opportunity to escalate the killing with impunity. All signs seem to point to a preference on the part of both states and international organizations to settle the matter diplomatically, rather than on the battlefield. Even forces on the ground, typically as peacekeepers or monitors in a neutral intervention, are there to avoid rather than engage in confrontation. Faced with multiple signals of a desire to avoid confrontation on the part of the international community, it is not surprising that perpetrators might draw the conclusion that they can not only proceed with targeted killings of noncombatant civilians, but can even escalate them. It should be noted that, while these results seem to suggest that neither diplomatic sanctions nor diplomatic engagement affect genocide or politicide severity either directly or indirectly (save for one interaction), these results are at best suggestive rather than causally determinative. This analysis included a range of control variables to account for other policy options, environmental and regime characteristics, and other possible confounds, yet not all potential confounding factors could be completely controlled. It is possible that other omitted variables have led to bias in these estimates. In particular, while this study examines changes in diplomatic representation and assumes that this might affect other political processes and calculations, there are numerous other factors that also impact such processes, and perhaps with greater weight. Nevertheless, these results are certainly strongly suggestive of a relationship, and seem to warrant much further investigation. 25

27 Conclusion As expected, this study has found preliminary evidence to suggest that neither diplomatic engagement nor diplomatic sanctions have significant effects on genocide or politicide severity. Nor do changes in diplomatic representation appear to condition the effects of most other tools that international actors have to ameliorate mass killing. However, increased engagement coupled with neutral interventions does seem to send a signal to perpetrators that they will continue to get away with the atrocities, and appears to encourage an escalation of the killing. As such, while changing levels of diplomatic representation does not appear to be effective at mitigating mass murder, under particular circumstances encouraging new or deeper diplomatic engagement may actually make things worse. This suggests the need for the international community to rethink the role that greater diplomatic engagement or the use of diplomatic sanctions have in slowing or stopping the worst atrocities. While such measures may be necessary politically to garner support for or lay the groundwork for other actions, they do little on their own to make things better, and under particular conditions may even make things worse. Of course, this study is but a small step forward in our understanding of the toolbox available to those interested in mitigating ongoing mass killing. Future work needs to revisit this relationship, and expand the examination of diplomatic tools beyond existence or levels of diplomatic representation, to include mediation attempts, diplomatic initiatives, and other types of diplomatic sanctions. Furthermore, this study has demonstrated that examining the interactions between policy tools may also be a fruitful avenue of research. For policymakers to have a fully functioning toolbox, they need to know how tools work in combination to fix (or perhaps exacerbate) a particular problem. Future work should emphasize the interactions 26