Antony's College). She has written extensively on transitional justice and also has solo authored

Transcription

1 Overcoming Barriers to Justice in the Age of Human Rights Accountability About the authors Leigh A. Payne is professor of Sociology and Latin America at the University of Oxford (St Antony's College). She has written extensively on transitional justice and also has solo authored a number of books, including Unsettling Accounts: Neither Truth nor Reconciliation in Confessions of State Violence (Duke University Press, 2008). Francesca Lessa is postdoctoral researcher at the Latin American Centre and St. Anne s College (University of Oxford). She is author of Memory and Transitional Justice in Argentina and Uruguay: Against Impunity (Palgrave Macmillan 2013) and co-editor of numerous edited volumes, including Amnesty in the Age of Human Rights Accountability (with Leigh A. Payne, CUP 2012). Gabriel Pereira is Doctoral Candidate in Politics and Research Assistant at the Department of Sociology at University of Oxford. He is a human rights practitioner who has worked for local and national organizations in Argentina and regional human rights organizations in Latin America. Acknowledgments This research received support from the National Science Foundation (Grant No ) and the Arts and Humanities Research Council (Grant No AH/I500030/1) for the project entitled The Impact of Transitional Justice on Human Rights and Democracy ; from the Oak Foundation (Grant No OCAY ) for the project Overcoming Amnesty in the Age of Accountability ; and from the John Fell OUP Research Fund (Grant No 101/552) for the project Accounting for Amnesty: Justice for Past Atrocity. Any opinions, findings, conclusions or recommendations expressed in this study are those of the authors and do not necessarily reflect the views of the National Science Foundation, Arts and Humanities Research Council, 1

2 Oak Foundation or John Fell OUP Research Fund. The authors would like to acknowledge support from their fellow team members Alec Albright, Emily Braid, Brooke Coe, Geoff Dancy, Holly Dunn, Katherine Franzel, Marie-Christine Ghreichi, Daniel Johnson, Hun Joon Kim, Pierre-Louis Le Goff, Moira Lynch, Cameron Mailhot, Bridget Marchesi, Veronica Michel, Florencia Montal, Tricia Olsen, Andrew Reiter, Kathryn Sikkink, Farrah Tek, and Marcela Villarrazo. 2

3 Abstract Amnesty laws are viewed as a main barrier to justice for past human rights violations. Scholars and practitioners expected the age of human rights accountability to reduce the number or coverage of amnesty laws that block human rights trials. Based on analysis of an original database of amnesty laws and trials, this article challenges that outcome. Few discernible patterns regarding amnesty laws and accountability emerge; human rights trials are nearly as likely in the absence of amnesty laws or where partial laws in compliance with international standards and non-compliant blanket amnesty laws exist. Most of the countries that have overcome the amnesty law barriers to justice are in Latin America. This article thus uses the region to identify the factors that provide pathways to justice for past human rights violations. It considers policy recommendations to strengthen those four factors: civil society demand, international pressure, judicial leadership, and the absence of veto players. 1

4 Introduction Recent attention to the age of human rights accountability suggests that throughout the world courts have put perpetrators on trial for past abuses bringing justice to victims of atrocity. These efforts have chipped away at the impunity that once protected perpetrators of human rights violations. While our research concurs that human rights trials have increased, we remain cautious regarding the impact of the age of accountability on overcoming barriers to justice. This article sets out our argument and the evidence supporting it. Understanding the barriers and the limits to the age of accountability, we nevertheless use our findings to support specific policies and practices that we have found to increase accountability. The age of human rights accountability is usually identified with global institutional transformations of 1998, specifically the creation of the International Criminal Court. Since the end of World War II and the Nuremberg and Tokyo Trials, the international human rights system has attempted to replace the traditional practice of amnesty with a new norm of accountability for human rights violations. 1 International conventions adopted in the second half of the twentieth century now obligate state parties to provide redress for victims of torture and genocide. The UN international criminal tribunals for the former Yugoslavia and Rwanda, set up in the early 1990s, underscored the international duty to hold perpetrators accountable. The notion of universal jurisdiction, and its use in the effort to extradite former Chilean dictator General Augusto Pinochet from the United Kingdom to stand trial in Spain in the late 1990s, claims that courts in one country can hold foreign perpetrators accountable for crimes against humanity committed in another country. An accountability norm has spread throughout the world, producing dramatic and unprecedented results. 2 Although General Pinochet did not in the end stand trial in Spain, he did face charges in his own country before he died. Other heads of state responsible for human rights abuses have also faced prosecutions, convictions, and prison sentences, including: former Peruvian and Uruguayan presidents Alberto Fujimori and Juan María Bordaberry in domestic 2

5 courts; 3 former Rwandan Prime Minister Jean Kambanda in the ICTR; former Liberian president Charles Taylor before the Special Court for Sierra Leone; and Khieu Samphan, the former Head of State of Democratic Kampuchea at the Extraordinary Chambers in the Courts of Cambodia. 4 This tremendous and unprecedented global progress toward accountability would lead to the expectation that governments and international institutions would hold perpetrators of atrocities legally responsible for their acts. To do so, they would have to overcome barriers to justice such as amnesty laws. For some scholars and practitioners, amnesty laws are tantamount to impunity and therefore a barrier to the promotion of democratic right to redress, rule of law, and the deterrence of human rights violations. 5 They consider such laws out of compliance with international human rights law; particularly, since the late 1990s, the UN has taken the position that amnesties that prevent the prosecution of individuals charged with war crimes, genocide, crimes against humanity and other gross violations of human rights are inconsistent with states obligations under various widely ratified treaties as well as United Nations policy and may also be incompatible with emerging principles of customary law. 6 These scholars and practitioners seemed to expect that prosecutorial innovations such as the creation of the International Criminal Court with the 1998 Rome Statute and the application of the notion of universal jurisdiction would stymie the adoption of amnesty laws. And that new and old amnesty laws would become partial and consistent with international human rights standards, particularly excluding from the benefits of amnesty perpetrators of genocide, crimes against humanity, and war crimes. Some anecdotal evidence suggests that in the age of accountability amnesty laws around the world have faced challenges from domestic, regional, and international courts, as well as from mobilized local and international victims, survivors, and human rights organizations. The Inter-American Court of Human Rights has ruled against amnesty laws in Peru and elsewhere that block human rights trials. 7 While Chile has not annulled its amnesty law, lawyers representing victims have found ways to circumvent it to hold perpetrators accountable. Some 3

6 amnesty laws, such as those in Croatia and Guatemala, are partial, complying with international standards. Scholars have interpreted this anecdotal evidence differently presenting a range of expectations regarding the impact on justice of the age of human rights accountability. Kathryn Sikkink s notion of a justice cascade would appear to suggest that the global spread of individual criminal accountability for past human rights abuses could wash away earlier patterns of impunity. Sikkink (2012) has agreed with Louise Mallinder (2012), however, that cascading human rights trials might paradoxically lead to the increasing adoption of amnesty laws to protect perpetrators from the higher threat of prosecution. 8 Both scholars expect the age of human rights accountability to influence the type of amnesty laws that countries will adopt, specifically partial and compliant with international human rights standards in contrast to earlier blanket amnesty laws. Mark Freeman (2009) appears to lament the weakening of amnesty laws anticipated by the age of human rights accountability. 9 He sees amnesty laws as necessary evils that enable governments to bring warring factions to the negotiating table and to encourage former repressive leaders to withdraw from power. They offer a means to safeguard democratic and peaceful transitions, and establish human rights protections. They may also offer a more humane way to cope with victim-perpetrators, such as child soldiers. Moreover, Freeman and Max Pensky (2012) contend, contra Méndez, that amnesty laws are not inconsistent with international law, particularly article 6(5) of the 1977 Protocol II to the Geneva Conventions. 10 These scholars further argue that the duty to prosecute violations is distinct from a requirement to prosecute every perpetrator of past abuse or an amnesty ban. In their seminal article on amnesty laws, Jack Snyder and Leslie Vinjamuri (2003/2004) 11 consider amnesties an important tool for fragile governments emerging from authoritarian rule and civil conflict. Their appeal and utility would suggest that amnesty laws would remain in use 4

7 despite the age of human rights accountability and in contrast to Freeman s worry about the difficulty of adopting them. Olsen et al. (2010) seemed to confirm Snyder and Vinjamuri s expectation. 12 While Sikkink s justice cascade seemed to hold in terms of the increase in trials, these trials had very little impact on increases or decreases in amnesty laws. They could not confirm either an increase in amnesty laws in response to trials as Sikkink and Mallinder anticipated. Nor did they find a decrease in those amnesty laws as Freeman expected. Amnesty laws remained steadily in use, consistent with Snyder and Vinjamuri expectation. Despite the proliferation of assumptions about amnesty laws and justice in the age of human rights accountability what they do, what they block, and how they have changed over time very little empirical analysis has verified the presumed trends and outcomes. Answering these questions has been the focus of our project. We use our findings to then engage in a discussion of policy and practices that overcome the barriers to justice in the age of human rights accountability. The Research The data we use is drawn from a study carried out by twenty researchers based at the University of Oxford and the University of Minnesota from 2010 to The researchers used a variety of sources to track data on amnesty laws and human rights trials. 13 Using the criteria set out in Appendix I, the research team identified 63 amnesty laws in 34 transitional countries for analysis. We found that amnesty laws were unevenly distributed around the world, with the majority in Latin America (46%), as Table 1 below shows. We further investigated the specific provisions and beneficiaries of the existing amnesty laws, particularly whether the laws complied with international human rights standards. By compliance, the researchers considered those laws that excluded from amnesty benefits perpetrators of war crimes, genocide, and crimes against 5

8 humanity. The researchers located, again, the majority of non-compliant amnesty laws in Latin America (52%). The researchers also tracked human rights accountability practices through trials. To determine the accountability of these trials, in this article we have focused on only those trials that reached a conclusion in a verdict and excluded ongoing trials or those that were officially suspended or unofficially shelved. Using a variety of sources, the researchers investigated 1812 trials and found 330 that reached a verdict, or 18 per cent. Of those, 267 reached final guilty verdicts, 15 per cent of the total number of trials investigated and 81 per cent of those with verdicts. As Table 1 below shows, the majority of these guilty verdicts were also delivered by Latin American courts (42%). Further information on the data and methods used in the project and the country and regional information concerning amnesties and guilty verdicts can be found in Appendix I-III. In the interest of providing systematic coverage across all countries, the research team used a consistent set of sources recognizing that this method undercounts trials and amnesties. Finding accurate information on specific countries, however, would bias the comparison since certain countries have better systems of counting and tracking trials and amnesties than others. For example, information contained in the database of the Centre for Legal and Social Studies (CELS) lists a total of 416 individuals found guilty in Argentina by 2013, while that of the Human Rights Observatory of the Universidad Diego Portales lists 266 individuals found guilty in Chile until August Our database seems to be undercounting in those two countries, showing instead 86 individual guilty verdicts for Argentina and 71 for Chile. 6

9 Table 1: Amnesties and Guilty Verdict by region Research Findings The research allowed us to address the following questions: Has the rate of amnesty law adoption changed over time and with the age of accountability? Has the type of amnesty law changed allowing for a higher level of accountability? Where amnesty laws have persisted, have courts found ways around amnesty laws to allow for prosecutions? Has justice increased with the age of human rights accountability or do barriers continue to block justice? Tracking amnesty laws At first appearance, the research seems to confirm the assumption that amnesty laws began to taper off with the age of human rights accountability. As Figure 1 shows, the number of countries adopting trials each year have increased over time while the number of countries adopting new amnesty laws each year have fallen, suggesting a trend away from amnesty law adoption. 7

10 Figure 1: Trends in Amnesty and Accountability, On closer inspection, however, our research questions that assumption. Amnesties adopted at an earlier time continue to have legal standing, as shown in Figure 2. Although there was a small dip in amnesties a few years after the 1998 Rome Statute marking the age of human rights accountability, over all amnesties that had been adopted in previous years continued to have the power of law. Figure 2: Total Amnesties with Legal Standing,

11 Given the persistence of legally standing amnesty laws, and following on the assumptions in Snyder and Vinjamuri and Olsen et al., we probed whether the rate of transitions, rather than the age of accountability, explained the apparent drop off in the adoption of new amnesty laws. As Figure 3 shows, amnesties did not steadily rise or decline but peaked at three key moments corresponding to political transitions. The first peak in the mid and late 1970s witnessed the transitions in Southern Europe, including in Portugal, Greece and Spain. The second, and higher peak, occurred with transitions from authoritarian rule in Latin America and transitions from communism in the former Soviet Bloc countries in the 1980s and 1990s. A third peak around relates to transitions in several African countries, including Algeria, Democratic Republic of Congo (DRC), and Liberia. The decline in the number of transitions since that period may explain the low level of new amnesty law adoption. Amnesty laws seem at least somewhat independent of the global human rights norm; consistent with some theoretical arguments their enactment appears to remain in use as a function of transitional governments needs. Figure 3: Trends in Amnesty and Transition,

12 To further examine amnesty trends and patterns, the team investigated which of the legally standing amnesty laws complied with international human rights law and which did not. Our analysis of amnesty laws, summarized in Table 2 below, could suggest that compliant amnesty laws have tended to emerge after the 1998 Rome Statute creating the International Criminal Court and are thereby consistent with the argument made by Sikkink and Malinder that more partial amnesty laws accompany the age of human rights accountability. Although the compliant amnesty laws are split evenly between those adopted before and after 1998, the period after 1998 is much shorter (13 years) than the years before (24 years). Detracting from that argument, however, are the data on non-compliant amnesty laws. Nearly the same number of non-compliant laws (4 in five countries) as compliant laws (5 in three countries) was adopted after The vast majority of legally standing amnesty laws, moreover, is not in compliance with international standards. Although ten of those laws have been overturned (two before 1998 and eight after), this involved only three countries (Argentina, Bangladesh, and Uruguay). These data provide fairly weak evidence of the impact of the age of human rights accountability with regard to the compliance of amnesty laws to international standards. 10

13 Table 2: Response to Accountability Norm Table 2 suggests additional trends that warrant further exploration. On the one hand, most of the compliant amnesty laws have been adopted in Africa (7 out of 10). On the other hand, the region that has moved the furthest in overturning non-compliant amnesty laws is Latin America (2 out of 3 countries; 9 out of 10 laws). The researchers began to consider regional patterns that might explain where the age of human rights accountability had penetrated the deepest and why. Regional patterns reveal a complicated picture of the diffusion of global human rights norms. Africa looks like the leader in compliant amnesty laws, with 7 out of the 10 compliant laws in the world. As Table 1 shows, this is 70 per cent of the world s amnesty compliant laws with two other regions (Latin America and Europe) sharing the remaining 30 per cent. Despite those impressive figures, this still amounts to only seven laws in the world. Moreover, the region 11

14 does not appear to have used those compliant amnesty laws to become a world leader in prosecuting perpetrators of past human rights abuses; Africa follows Latin America (42%) and Europe (33%) with only 15% of the world s guilty verdicts. Examining the regional patterns suggests that compliant amnesty laws do not necessarily guarantee accountability for past human rights violations. Latin America further complicates the regional picture. It is the global leader in amnesty law adoption in general (46%), and amnesty laws that do not comply with international standards in particular (52%). These figures hint that the region has the lowest level of legal accountability for human rights violations in the world. Defying that profile, the region has become a global leader in convictions for human rights violations. Over 40 per cent of the global guilty verdicts are located in the Americas. The presence of amnesty laws, and even non-compliant amnesty laws, does not necessarily block accountability. To summarize our key research findings, the researchers drew the following conclusions about the pattern of amnesty law adoption in the age of human rights accountability. First, there is little evidence to fully support the claim that the lowered rate of amnesty law adoption corresponds to the age of human rights accountability. Moreover, the majority of legally standing amnesty laws are not in compliance with international human rights norm even after Despite this finding, non-compliant amnesty laws have not necessarily blocked accountability as the data from Latin America show. On the other hand, we have also seen how compliant amnesty laws have not guaranteed accountability as Africa shows. These findings suggest that something other than amnesty laws are blocking or promoting accountability. We explored possible factors to explain these phenomena. 12

15 Tracking Challenges to Amnesty Laws The evidence from Latin American suggests that courts have found ways around amnesty laws to convict perpetrators of past human rights violations. The researchers thus explored where amnesty laws had been legally challenged, opening up the pathway to accountability. (A discussion of the definition of and determination of challenges can be found in Appendix I.) As the researchers expected given the outcome of human rights trials, the vast majority of legal challenges are concentrated in Latin America: 90 per cent of the legal challenges to amnesty laws in the world. This finding illustrated in Table 3 suggests that where amnesty laws have been legally challenged either through international pressure or domestic pressure, amnesty laws will likely be weakened, circumvented through loopholes, or overturned in the rare case. Latin America had two out of three of those rare cases. Region Total No. (%) Dom Intl INGO Africa 3 (2%) Americas 145 (90%) Asia 5 (3%) Europe 4 (2.5%) MENA 4 (2.5%) TOTAL Table 3: Regional Distribution of Challenges The high number of challenges in Latin America with a corresponding high level of guilty verdicts suggests judicial leadership and capability. For example, Argentina represents the country with the greatest number of domestic legal challenges compared to the rest of the region 13

16 and the world. It is also the country with the highest number of guilty verdicts. These challenges eventually culminated in the annulment of the amnesty laws. (More country and regional details on challenges can be found in Appendix II.) The challenges, in other words, may reflect factors such as judicial leadership and capability to overcome non-compliant amnesty laws and promote accountability. The region also has a high level of international legal challenges, suggesting that the presence of an active Inter-American Human Rights system ruling against the region s amnesty laws helps explain the higher degree of accountability in the region. International pressure in the form of international non-governmental organizations is also much higher in the Americas compared to other regions. The researchers drew two main conclusions from the data on challenges: international pressure on countries with non-compliant amnesty laws tends to weaken those laws and promote accountability and responsive domestic courts are likely to find ways to bypass amnesty laws and promote accountability. To further develop these findings the researchers began to investigate the trials and guilty verdicts. Tracking Domestic Prosecutions and Amnesty Laws With the age of human rights accountability, the expectation is not only that there will be an increase in human rights trials, but also in guilty verdicts resulting from those trials. Evidence does not entirely support that assumption, however. Of the 1812 human rights trials in the database, the vast majority (82%) ended in no verdict at all, suggesting either a lack of information about trials once they are initiated or a high degree of cases in which trials are still ongoing or have been shelved. 15 Of the 330 trials that reached a verdict, 267 (81%) found perpetrators guilty of human rights abuses. The project s data on human rights trials also reveal that certain regions (Latin America) and certain countries within those regions are global leaders with regard to accountability (guilty 14

17 verdicts) for human rights violations. Within Latin America, four countries (Argentina, Chile, Guatemala, and Peru) 16 dominate the number of trials and guilty verdicts in the region; four out of the 14 countries with trials concentrate 65% of the guilty verdicts. Two out of 15 countries in Europe with trial verdicts (Croatia and Serbia) have nearly half (47%) of all of the guilty trial verdicts in the region. Nearly half (49%) of the trial verdicts in Africa are rendered by courts in a single country (DRC). All four trials and guilty verdicts were reached in a single MENA country (Turkey). In Asia, two countries (South Korea and Indonesia) reached 70 per cent of the guilty verdicts. Thus rather than a regional or global phenomenon of accountability, there are a few isolated countries that seem to be driving the assumptions regarding the age of human rights accountability for past human rights abuses. (See Appendix II for the details of trials by country.) Examining the countries that lead in trials further illustrates this concentration. Only seven countries in the world have more than ten trials with guilty verdicts. Yet those seven countries constitute half of all the trials with guilty verdicts in the world. The seven countries Argentina, Chile, Guatemala, Peru, Croatia, Serbia, and the DRC should be able to identify a clear pattern of relationship of accountability to amnesty laws. But they do not. We would expect these countries to have either no amnesty law or compliant ones. While that is the case for about half of the countries (Argentina, Guatemala, Croatia, Serbia, and DRC), the other half have non-compliant amnesty laws (Argentina, Chile, Guatemala, and Peru). 17 Examining the country leaders in guilty verdicts in the other world regions, the patterns are even less clear. Turkey, the only country among the MENA countries with guilty verdicts has no amnesty law, but the two guilty verdict leaders in Asia South Korea and Indonesia both have noncompliant amnesty laws. Moving beyond those countries in which guilty verdicts are concentrated, we cannot find clear sets of patterns. What is the relationship of these human rights trial leaders to amnesty laws? Twenty-eight per cent of the countries with guilty verdicts (13 out of 47 countries) 15

18 concentrated 70 per cent (174) of the world s guilty verdicts. These countries have the following break down in terms of their amnesty law: over half (8) have either amnesty laws that respect international human rights law (Argentina, Croatia, DRC, and Guatemala) or have no amnesty law at all (Indonesia, Ethiopia, Czech Republic, and Paraguay). Countries with non-compliant amnesty laws (Argentina, Chile, Panama, Peru, South Korea, and Former Yugoslavia) are also among the leaders in guilty verdicts. In other words, it does not appear that the kind of amnesty law (or its absence) shapes accountability behaviour. If one looks at the absence of guilty verdicts for human rights trials, more doubt is cast on whether removing an amnesty law or making it more partial and more compliant with international standards enhances accountability. Returning to the seven countries that have compliant amnesty laws, over half (5) have not had a single guilty verdict: Bangladesh, Burundi, Liberia, Uganda, and Uruguay, and a sixth country (Albania) had only one guilty verdict. These data further confirm that an amnesty law that complies with international human rights standards does not guarantee accountability. Similarly, the absence of an amnesty law does not mean that accountability is guaranteed. Countries without amnesty laws that bar human rights prosecutions still contributed only 29 per cent (73) of the guilty verdicts in the world. Compliant amnesty laws that allowed for prosecutions of certain human rights violations did slightly better with 32 per cent (89) of the guilty verdicts. Together these countries with no amnesties or compliant amnesties included about a third of the countries in the database but nearly two-thirds of the guilty verdicts. Among the countries with blanket amnesty laws, twelve issued no guilty verdicts while eighteen rendered guilty verdicts. These data further emphasize that the type or absence of amnesty laws does not determine the success in accountability processes, leading us to several related observations. Principally, we found that the absence of amnesty laws does not guarantee accountability and, as we saw with the data from Africa, partial amnesty laws that reflect international human rights standards do not necessarily guarantee accountability. On the other hand, we have also seen that 16

19 non-compliant amnesty laws, such as those found throughout Latin America, do not always block accountability. In fact, by looking at certain accountability leaders we could reveal processes in some countries whereby amnesties were bypassed and a degree of justice could be achieved. Guilty Verdicts No Amnesty TOTAL: 28 Africa: 4 Asia: 5 Americas: 5 Europe: 13 MENA: 1 Compliant Amnesty TOTAL: 3 Africa: 1 Asia: 0 Americas: 1 Europe: 1 MENA: 0 Blanket Amnesty TOTAL: 17 Africa: 4 Asia: 3 Americas: 7 Europe: 3 MENA: 0 No Guilty Verdicts TOTAL: 31 Africa: 11 Asia: 4 Americas: 3 Europe: 11 MENA: 2 TOTAL: 3 Africa: 3 Asia: 0 Americas: 0 Europe: 0 MENA: 0 TOTAL: 6 Africa: 4 Asia: 0 Americas: 0 Europe: 0 MENA: 2 88 countries total TOTAL: 48 TOTAL: 40 Table 4: Patterns of Amnesty and Guilty Verdicts 17

20 The lack of clear patterns can be interpreted in a number of ways related to the assumptions behind justice and barriers to justice in the age of human rights accountability. On one hand, the table above shows that transitional justice countries are almost split in terms of allowing for justice (at least one trial with a guilty verdict) and those where impunity continues. A positive interpretation would claim that more countries are allowing for justice than denying it. The concentration of countries with guilty verdicts, and the fact that a third of all countries (16 of 48) with guilty verdicts have only one such trial, somewhat diminishes this positive interpretation. Nonetheless, as we have shown elsewhere, trials have increased over time. More surprising in these results is the relationship to amnesty laws. We would expect that where the barriers to justice do not exist either because of the absence of an amnesty law or with a compliant amnesty law guilty verdicts would be more likely reached. The summary table above questions that assumption. While there are more guilty verdicts in countries with no amnesty law, there are even more countries without amnesty laws that have not rendered a single guilty verdict. Unexpectedly, the existence of blanket amnesty laws has not prevented justice. We see that there are a higher number of countries with blanket amnesty laws that have rendered guilty verdicts than those with no guilty verdicts. Blanket amnesty laws, in other words, are not protecting perpetrators from prosecution. How countries have found their way around these barriers to justice is the focus of the next section of the article. Pathways to Accountability Concentrating on Latin America where amnesties and accountability have co-existed, we have found four different pathways to accountability. In a piece written together with two other researchers (Lessa et al. 2014), we considered these pathways in details. The four pathways are summarized in Figure 4 below. 18

21 Obstinate Accountability Creative Democratic Amnesties Impasse Circumvention Displacement Full Impunity Full Accountability Figure 4: The Accountability-Impunity Continuum We found that the placement along the accountability-impunity continuum in Latin America, depended on the strength or weakness of four main factors: civil society demand, international pressure, judicial leadership, and the absence of veto players. Where all four factors were strong, the amnesty law was democratically displaced leading to the promotion of full accountability. This is the Argentine example. Although Uruguay also democratically displaced its amnesty law, it has had very little prosecutorial activity. The pathway to accountability in that case did not open up with the removal of the amnesty barrier; instead the process remains blocked due to the presence of strong veto players and the weakness of judicial leaders. Where all of the factors were weak, amnesty laws remain obstinately in place and impunity is the outcome. Brazil represents this case. In between those two extremes on the continuum, the factors vary in strength. The researchers found that where civil society strongly demanded accountability, and international pressure reinforced the legitimacy of that demand, judicial leadership often emerged in response, legally challenged amnesty laws and promoted accountability by finding legal innovations to creatively circumvent the amnesty law. Veto players therefore block efforts to remove the legal standing of the law, but they fail to block accountability processes that bypass these laws. Chile represents this pathway. The fourth pathway is where civil society and international pressure promote accountability and have some modest success. The strong presence of veto players 19

22 influences the judiciary and blocks most justice initiatives, however. An accountability impasse is reached, greatly constraining the possibility of justice as the Salvadoran case illustrates. Three key findings emerge from this analysis. First, civil society demand, international pressure, judicial leadership, and the absence of veto players are important. While each factor plays a key functional role in overcoming impunity, no single factor is sufficient to bring about pathways to accountability. We find that the combined strength of these four factors is the strongest predictor of overcoming impunity. Second, it is rare that all factors are strong, thus limiting progress toward overcoming impunity. Particularly in the early years following a repressive dictatorship, civil society demand tends to be weak, strong veto players prevail and block accountability, international pressure may not emerge and judiciaries have not yet experienced generational shifts or reform to challenge impunity. The weakness of these four factors, we contend, predicts persistent or obstinate amnesties. Third, variation in the strength of these factors explains countries placement along the continuum from impunity to accountability. Strong veto players and weak judicial leadership are most likely to block full accountability, even where civil society and international pressure is strong, leading to accountability impasse. Where judicial leadership emerges, civil society demand combined with international pressure finds ways around persistent blanket amnesties to promote justice. Creative circumvention thus depends on the judiciary, but the judiciary in turn relies on civil society demand, international pressure and waning support for the old authoritarian leaders Conclusions and Policy Implications The research team s careful research and analysis of the data on amnesty laws, challenges to amnesties and guilty verdicts, and the paths to accountability lead to a series of conclusions with relevant policy implications. The policy implications correspond to the findings on the four main questions set out at the beginning of this article. 20

23 We ask, first, if the rate of amnesty law adoption changed over time and with the age of accountability. New amnesty laws appear to have dropped off over time and even legally standing amnesty laws have peaked and appear to have begun to taper off. Nonetheless, the decline in new amnesty laws appears to correspond to the decline in the number of transitions rather than to a new set of global norms, international legal constraints, or a response to the rising threat of prosecution owing to the age of human rights accountability. The number of legally standing amnesty laws grew during the age of human rights accountability, only slightly dipping with the annulment of amnesty laws in two countries in the post-1998 period. While some might interpret those annulments as a sign of the impact of the age of human rights accountability on the rate of amnesty laws, a third country annulled its amnesty law prior to Moreover, of the three countries with annulled amnesty laws, only one has successfully prosecuted perpetrators of past human rights abuses. The accountability process in that country began well before the age of accountability, moreover. Our findings thus suggest that the accountability trend has not significantly shifted patterns of amnesty adoption. The failure of the age of human rights accountability to remove legal barriers to prosecution, however, has not blocked justice, as we show further below. The policy implications of our research thus inclines toward working within existing legal and institutional boundaries, whatever they may be, to overcome barriers to justice, rather than focusing on legal and formal barriers to justice per se. Our second question explores those legal and formal barriers; it probes the relationship between the type of amnesty law and the rate of accountability. We found that the type of amnesty law had not dramatically changed with the age of accountability. Partial amnesties in the post-1998 period only outpaced blanket amnesty laws by one law. Moreover, we found that the same number of partial amnesty laws had been enacted before the age of accountability as after Although we found that few discernable patterns emerged in the relationship between the type of amnesty law and the likelihood of successful prosecution, we note several significant relationships. The first implication is that impunity and accountability are not necessarily a 21

24 function of the type of amnesties but a function of the context in which amnesties are embedded. Take the case of Africa and Latin America. Even though Africa had the highest proportion of partial amnesties, Latin America was the region that had the highest number of guilty verdicts, despite having a large number of blanket amnesties. Thus, accountability outcomes seem to be driven by other factors, beyond the mere existence of amnesties and different typologies of amnesties. Indeed, Latin America had the highest concentration of blanket amnesties, but these did not block the accountability process. These regions thus suggest policy implications. On one hand, Africa suggests that policies should not merely advocate compliance with international human rights standards, but consider the barriers beyond amnesties that block accountability. Latin America provides insights into what factors or actors policies should target to replicate the region s success in overcoming impunity and promoting accountability. The analysis of the pathways to accountability in the region is particularly telling regarding the type of factors and how they interact to advance accountability. The policy orientation toward challenging the legality or legitimacy of amnesty laws that was expected to arrive with the age of human rights accountability has not had such an effect outside Latin America. Only one of the three countries that annulled blanket amnesty laws has had successful prosecutions (Argentina). In addition, the absence of amnesty law has not guaranteed successful prosecution. We find that guilty verdicts were no more likely (and actually even slightly less likely) than failing to prosecute in countries with such laws. On the positive side, however, we found that blanket amnesty laws have not blocked accountability. More countries with blanket amnesty laws rendered guilty verdicts than fail to do so. These findings suggest that the focus on removing the laws themselves may not be as successful a prosecutorial strategy than finding legal loopholes to get around them. In this way, our research differs from the work of other scholars such as Louise Mallinder who released in October 2013 the Belfast Guidelines on Amnesty and Accountability developed by an Expert Group of independent, interdisciplinary scholars and practitioners. 18 Those guidelines aim to assist those seeking to 22

25 make or evaluate decisions on amnesties and accountability in the midst or in the wake of conflict or repression, focusing especially on the scope, possible beneficiaries, and the adoption and implementation of amnesties. Our study suggests that the focus should be aimed less toward the amnesty laws themselves and instead around the capacity of civil society and state actors in finding ways around amnesty laws and other barriers to justice. In sum, amnesties do not necessarily block trials and the absence of amnesties does not, on its own, promote accountability; other barriers exist that block accountability. The policy implications of our study is thus to identify the barriers and the pathways around them. That finding relates to the third question regarding the capacity to overcome barriers to justice in the age of human rights accountability. Our study of challenges to amnesty laws shows that the region with the highest level of justice (Latin America) has also experienced the highest number of challenges to impunity. Indeed, four Latin American counties constitute 65% of the guilty verdicts in the region and 63% of the amnesty challenges in the region. Outside Latin America, however, there are fewer legal challenges to amnesty laws. The prosecutorial success of leaders in guilty verdicts in other geographic regions have not faced challenges to their amnesty laws. Thus, legal barriers may not be as significant a hurdle to justice as other social and political factors. The policy implications of our study is to focus efforts on strengthening civil society demand for accountability; heightening international pressure for accountability; enhancing judicial leadership behind accountability; and weakening veto players who promote impunity and block accountability. Civil society demand is fundamental to the promotion of accountability. Without victims of human rights claiming their right to redress, judiciaries are unlikely on their own accord challenge amnesty laws and promote accountability. The civil society groups that have achieved high visibility through the media, using innovative mobilizational techniques, and reaching beyond the generation of directly affected individuals to appeal to subsequent generations, have resonated at the domestic and international level, applying pressure on the judiciary to challenge 23

26 amnesty laws and impunity and promote accountability. Civil society demand is insufficient on its own, but challenges are unlikely to occur without it. Thus strengthening civil society demand for accountability through domestic and international media visibility, the development of innovative mobilization techniques and inter-generational appeal of accountability issues is only one policy implication of our study. The role of international pressure is also insufficient on its own. It can amplify civil society demand through international governmental and non-governmental organizations. Regional or foreign courts can also weaken amnesty laws by ruling against them and promoting prosecutions. By recognizing the contribution of judicial leaders to accountability, international forces enhance the leaders prestige domestically, providing some security and legitimacy for their work. International pressure thus plays a fundamental role in sustaining and enhancing domestic civil society mobilisation for accountability; this can be achieved through referring cases to international, regional, and foreign human rights courts. Ultimately the capacity to overcome impunity and promote accountability is in the hands of the judiciary. Judicial leaders have emerged in a few countries around the world, using innovative legal strategies to challenge amnesty laws, to bypass them, and in rare cases to annul them. It is not enough, as this research has shown, to weaken or remove amnesty barriers; judicial leaders also have to advance trials. This may often require innovative and creative interpretations of existing laws or the application of international human rights provisions to break the logic of impunity that often surrounds human rights violations as in the example of Argentina in the 1990s. 19 Our study demonstrates the importance of enhancing judicial leadership behind accountability, which can be achieved through the identification and recognition of judicial accountability leaders, the protection of their personal security, as well as the diffusion and adaptation of successful innovative legal strategies, such as with the interpretation of enforced disappearance as a permanent crime which has been used in several 24

27 countries such as Argentina, Brazil and Chile. In that effort, they may be blocked by veto players. Where veto players possess significant power they can counter civil society demand for, and judicial leadership behind, accountability. They sometimes threaten civil society and judicial leaders. They can use their power to remove judicial leaders from their positions as in the case of judge Mariana Mota in Uruguay in February or Guatemalan attorney general Claudia Paz y Paz in May Protecting civil society and judicial actors from these threats is thus crucial to promoting leadership behind accountability. Weakening veto players power would also remove a significant barrier to accountability. Our study shows that former veto players begin to distance themselves from their previous positions, and weaken the barrier to accountability, with the discrediting of the former regime. Thus our policy implications include active efforts within and outside borders to delegitimize the former authoritarian regime through the denunciation of atrocity and corruption, fragmenting political and economic elites, reducing the support and allegiance to the regime, and opening up the possibility of prosecution without veto. Finally, we examine whether justice has increased with the age of human rights accountability and what barriers to justice continue to prevail. Few accept that the progression toward justice in the age of human rights accountability is inexorable or linear (Sikkink 2012; Clark 2012). While we have seen an increase in trials beginning even before the age of human rights accountability, we cannot fail to acknowledge the barriers to justice that persist. Even in cases where amnesty laws never emerged, where they were removed, or where they are partial, justice barriers in the form of veto players, lack of civil society demand, or the absence of leadership in the judiciary block accountability. To end on a positive note, however, we have also found that even where blanket amnesty laws could and should block accountability, they usually fail to do so. Civil society actors have mobilized together with international pressure and domestic judicial leadership to find creative and innovative ways to bring justice and overcome those barriers. 25

28 Appendix I: Overview of Data and Methods The data used in this paper is drawn from two databases. A team of researchers based at the University of Oxford constructed a database of amnesty laws while a second team based at the University of Minnesota built a database of human rights trials. The researchers used the following sources to build these databases: Keesing s World News Archive; U.S. Department of State Country Reports on Human Rights; Human Rights Watch country reports; the Amnesty Law Database; the International Center for Transitional Justice country reports and studies; transitional justice laws and decrees; secondary literature on specific countries; and in-country media sources. Both databases and the archive of amnesty laws and human rights trials will soon become available at In building the amnesty database specifically, the researchers drew on two pre-existing data bases: Louise Mallinder s Amnesty Law Database and Tricia D. Olsen, Leigh A. Payne and Andrew G. Reiter s Transitional Justice Database Project ( The researchers also updated the data on amnesties from 2009 to 2011, reproducing the Olsen et al. methodology of extracting such information from Keesing s World News Archive. The researchers carried out additional research on each amnesty law to determine if it fit with the project s definition of an amnesty for human rights violations carried out during the authoritarian period. They defined a fit amnesty as any legislative, constitutional, or executive provision that includes crimes a) committed during the authoritarian period, b) involving human rights violations, and c) carried out by state authorities or individuals/groups working on behalf of the state. They determined that of the 433 amnesty laws that they analysed in the 88 transitional countries (see Appendix I for the full set of countries and amnesty laws), only 63 laws in 34 countries fit their definition. They created an archive of these amnesty laws. The researchers thus refined the amnesty law data in terms of definitional criteria. In addition, they began to examine particular types of amnesty laws to test the assumptions about 26

29 trends. They made distinctions between partial amnesty laws, or those that exclude from their beneficiaries perpetrators of war crimes, genocide, and crimes against humanity, to determine which laws complied with international human rights standards. They further examined the process of challenging amnesty laws so that they would adhere to international human rights standards. By domestic challenges, the project identifies attempts to modify the legal scope of amnesty laws via courtroom, parliament, ballot box (e.g., exclude certain types of crimes, particular types of perpetrators, time period of the crimes); cancel legal effects of amnesties retroactively, for the future (derogation), or both (nullification). International challenges include decisions regarding the legal standing of amnesty laws by international governmental organizations (e.g., Inter-American or UN System). The researchers tracked international non-governmental organizations reports that targeted amnesty laws. However, because these reports do not have a direct impact on the legal scope and standing of the amnesties, they were not included as challenges in this policy brief. The researchers include these reports as evidence of international pressure, however, here and in their other work. In constructing the human rights trials database, the researchers based at the University of Minnesota look at prosecutorial activity as any activity before, during, and resulting from criminal procedures brought against perpetrators of human rights violations. From the wide range of factors characterized as prosecutorial activity in the database, the current analysis draws only on guilty and not guilty verdicts for crimes carried out during the authoritarian period and adjudicated after the transition from authoritarian rule. Such trials were coded by the team when they found individuals and groups held accountable by a criminal court, defined as an official judicial body created, financed, and operated by state actors, international governmental organizations, or both. In constructing the sample, the researchers only included democratic transitions in the two databases. To determine these cases, the team began with the list of all countries in the world over one million in population and then followed an accepted method used by other 27