Based on a pilot study of seven African countries – Burundi, Democratic Republic of Congo (DRC), Liberia, Mozambique, Rwanda, Sierra Leone and South Africa – this Policy & Practice Brief presents insights on amnesty regimes and conflict-related sexual violence (CRSV)2 beyond conflict termination. It shows that CRSV often continues past conflict termination and that most of the conflicts under study negotiated amnesties. Yet, this finding is not enough to confer that impunity leads to continued CRSV. The combination of amnesties and continued CRSV raises important questions about how to address sexual violence in post-conflict situations and build durable peace.
Rape, gang-rape, sexual mutilation, sexual slavery and other forms of CRSV3 are accompanied with immediate mental and health-related risks. There are also the devastating long-term social and economic effects such as stigmatisation, unwanted pregnancies and the spread of sexually-transmitted diseases, including the Human Immunodeficiency Virus (HIV).4 These effects have implications for the overall health and well-being of individuals, and for repeated patterns of psychosocial trauma. They are rarely considered in efforts to rebuild societies or implement peace processes. Indeed, deliberations about peace focus on lethal armed violence, or rather its absence, and rarely on non-lethal violence such as sexual violence.5 CRSV is particularly relevant for Africa as the forthcoming Peace Research Institute of Oslo (PRIO) database on Sexual Violence and Armed Conflict has found that ‘sexual violence often persists after the battle deaths have ended.’6
Related to our understanding of peace is the highly popular and stylised debate on ‘peace versus justice’.7 In general, this debate contrasts the effects of retributive accountability with those of negotiated amnesties on peace, ranging from the end of lethal violence to reconciling societies.8 The debate has mainly been discussed within policy circles, but also received some academic attention. While most case studies contend that long-term peace requires retributive accountability for which amnesties represent an antidote,9 a recent statistical study finds support for long-term peace resulting from negotiated amnesties.10 Amnesties have also been shown to extract non-legalistic accountability when conditioned; for example, by granting amnesty in return for truth-telling.11 Although many policy actors tend to claim that amnesties and impunity for past crimes fuel CRSV, there is no empirical proof that this is the case. This brief questions this common assumption and generates three policy recommendations:
The findings in this brief are based on a pilot-study of CRSV in state-based12 armed conflicts in: Burundi, DRC, Liberia, Mozambique, Rwanda, Sierra Leone and South Africa13, during 1989–2008.14 To begin exploring the relationship between amnesty regimes and continued CRSV, a review is necessary of the existence and scope of negotiated and implemented amnesties at the time of, or closely following, the formal cessation of armed fighting – understood as fewer than 25 battle-related deaths in one calendar year between a rebel movement and a government. In addition, CRSV by the conflict parties15 is examined three years following the cessation, which may represent a victory, disintegration or negotiated solution.16
Sources of information for the research of this brief include academic writings, official documentation, publications of local and international organisations, databases17, as well as news reporting. By examining a wide range of sources, it is possible to address inherent biases such as underreporting of sexual violence.18,19 Domestic sexual violence, though sometimes attributed to experiences of armed conflict, is not included as CRSV.20
Table 1 presents an overview of the findings, including conflict parties – rebels and governments – with active years of fighting. The existence, year and scope of negotiated amnesties are specified, with specific references to eventual limitations of international law, where crimes such as genocide, crimes against humanity and war crimes are excluded from the amnesties and can theoretically be prosecuted. The table clarifies if there was international accountability for past crimes – through hybrid courts, tribunals, commissions of inquiries – and whether there was any domestic accountability for past crimes in the form of: trials, vetting procedures, accountability through truth-telling (TRC) or local measures. Finally, the table presents whether CRSV continued past these resolutions or victories and in what form. CRSV that is referred to as systematic includes widespread use of rape or other sexual violence, sometimes connected to armed violence. Slavery relates to continued sexual slavery, forced marriages and similar hostage situations. Revenge-based CRSV covers retributive rape, gang-rape and other sexual violence; and sporadic CRSV refers to situations of separate, indiscriminate incidents of sexual violence perpetrated by representatives of the conflict parties.
| Country | Conflict Party, Active | Negotiated Amnesty | International Accountability | Domestic Accountability | Continued CRSV |
| 1. Burundi | Frolina (1980s–1997) |
28 August 2000 Limited: int. law |
No | No | No reports |
| Palipehutu (1991–92) |
28 August 2000 Limited: int. law |
No | No | No reports | |
| CNDD (1994–98) |
28 August 2000 Limited: int. law |
No | No | No reports | |
| CNDD-FDD (1998–2003) |
02 November 2003 Limited: int. law |
No | No | Systematic Revenge-based Sporadic <100 |
|
| Palipehutu-FNL (1997–2008) |
No* | No | No | Few sporadic | |
| Government 1 (–2000) |
28 August 2000 Limited: int. law |
No | No | Systematic Some revenge-based Sporadic <100 |
|
| Government 2 (2001–03) |
02 November 2003 Limited: int. law |
No | No (2 trials for new CRSV in 2006) | Systematic Some revenge-based Sporadic <150 |
|
| Government 3 (2004–08) |
No | No | No | Few sporadic | |
| 2. Democratic Republic of the Congo | AFDL (1996–97) |
No | No | No | Systematic Revenge-based Sporadic >200 |
| MLC (1998–2001) |
02 April 2003 Limited: int. law, not war crimes |
No | Limited Trials in 2003 Ineffective TRC | Systematic Revenge-based Sporadic <250 |
|
| RCD (1998–2001) |
02 April 2003 Limited: int. law, not war crimes |
No | Limited Trials in 2003 Ineffective TRC | Systematic Slavery Revenge-based <150 |
|
| RCD-ML (1999–2001) |
02 April 2003 Limited: int. law, not war crimes |
No | No Ineffective TRC | Sporadic <25 | |
| Government 1 (1996–97) |
No | No | No | Sporadic | |
| Government 2 (1997–2001) | 02 April 2003 Limited: int. law,not war crimes |
No ICC invited to investigate violence in Ituri | Limited Trials in 2006 Ineffective TRC | Systematic Revenge-based Sporadic >300 |
|
| 3. Liberia | INPFL (1990) |
No | No | No | No reports |
| NPFL (1989–90) |
17 August 1996 Blanket |
No | No | Sporadic <50 | |
| LURD (2000–03) |
18 August 2003 Blanket, later limited by int. law |
No | Limited TRC | Systematic Slavery Revenge-based Sporadic <150 |
|
| MODEL (2003) |
18 August 2003 Blanket, later limited by int. law |
No | NoTRC | Systematic Revenge-based Sporadic <150 |
|
| Government 1 (1989–90) |
17 August 1996 Blanket |
No | No | Revenge-based Sporadic <50 |
|
| Government 2 (1996–2003) |
18 August 2003 Blanket, later limited by int. law |
|
No TRC | Sporadic <25 | |
| Government (1977–92) |
04 October 1992 Blanket |
No | No | Revenge-based Sporadic | |
| 4. Mozambique | Renamo (1977–92) |
04 October 1992 Blanket |
No | No | Slavery (>5000) |
| Government (1977–92) |
04 October 1992 Blanket |
No | No | Revenge-based Sporadic |
|
| 5. Rwanda | FPR (1990–94) |
No** | No (ICTR not apply) | Limited Arrests, trials | Revenge-based Sporadic |
| Government (1990–94) |
No** | 10 ICTR-indictments | Broad Arrests, trials | Slavery Revenge-based Sporadic |
|
| 6. Sierra Leone | AFRC (1997–99) |
10 November 2000 Blanket, later limited by int. law |
Decent SCSL trials | Limited TRC | No reports |
| RUF (1991–2000) |
10 November 2000 Blanket, later limited by int. law |
Decent SCSL trials, including Charles Taylor | Limited TRC | Systematic Slavery Revenge-based >100 |
|
| CDF/Kamajors (1997–98) |
10 November 2000 Blanket, later limited by int. law |
Decent SCSL trials | Limited TRC | No reports | |
| WSB (2000) |
No | No | No | No reports | |
| Government (1991–2000) |
10 November 2000 Blanket, later limited by int. law |
No | Limited TRCSome vetting | Sporadic <10 | |
| 7. South Africa | ANC (1981–88) |
18 November 1993 Limited by TRC |
No | Decent Commissions of Inquiry TRC truth-telling | Sporadic |
| Government (1981–88) |
18 November 1993 Limited by TRC |
No | Decent Commissions of Inquiry TRC truth-telling | Systematic Revenge-based Sporadic <100 |
An important insight emanating from this pilot study is that the end of lethal armed violence does not always imply the end of CRSV. This verifies the general findings of the Sexual Violence in Armed Conflict data on Africa and thus validates the pilot study. According to available data in the pilot study, 14 out of 20 rebel movements continued CRSV in different form three years after a final agreement or end of fighting. Furthermore, all governments were responsible for continued CRSV, which is quite remarkable and seldom discussed.
Some of the continued CRSV has been widely reported in the media. These include the systematic violence of Revolutionary United Front (RUF) in Sierra Leone, after the Abuja ceasefire (10 November 2000), and the continued turbulence in DRC post the Final Act of the Inter-Congolese Dialogue (02 April 2003), as well as in Liberia, immediately after the Accra agreement (18 August 2003). Some has received less international attention, like the continued sexual torture in South African prisons after 1993; the sexual slavery under the former Rwandan government in exile; and a similar situation in Mozambique under Resistência Nacional Moçambicana (Renamo). While there are only a few registered incidents by the African National Congress (ANC), and the Liberian and Sierra Leonean governments – after 2003 and 2000 respectively –these acts count, as they would have during active armed fighting.21
The study points to three more insights on CRSV. First, many conflict parties that committed intense and widespread CRSV during armed fighting were also the ones responsible for most of the continued CRSV, though often on a lower intensity level than before. During armed fighting, 11 out of 20 rebel groups, and 7 out of 11 governments, used CRSV on a high intensity level and out of these, 8 rebel movements and 4 governments continued similarly intense CRSV after conflict termination. This goes for rebel movements Renamo, RUF, Liberians United for Reconciliation and Democracy (LURD), Movement for Democracy in Liberia (MODEL), Rassemblement Congolais pour la Démocratie (RCD), Movement pour la Libération du Congo (MLC) and Conseil National pour la Défense de la Démocratie–Forces pour la Défense de la Démocratie (CNDD-FDD), as well as two of the three governments of Burundi, the second government in the DRC, and South Africa. The intensity of CRSV for these actors varied across time, either peaking during the first year after conflict ended, or escalating throughout the three-year period, in parallel to other armed conflicts.
A second insight is that dissatisfaction with an agreement or new parallel, armed challenges to governments22 appear to influence continued CRSV. A dissatisfied RUF reneged on the Abuja ceasefire and committed numerous rapes between November 2000 and May 2001. In Burundi, the Global Ceasefire Agreement (16 November 2003) was followed by intense and systematic CRSV by both the government and the CNDD-FDD during January–May 200423; in Liberia, MODEL and LURD sexually assaulted internally displaced persons and other civilians in the turmoil that ensued until UN peacekeepers were effectively deployed in February 2004; and in DRC, sections of RCD and MLC systematically looted, raped, and attacked civilians during mutinies fuelled by discontent over the peace process. Similarly, CRSV increased when the Liberian government – incorporating National Patriotic Front of Liberia (NPFL) since 1996 – was challenged by LURD in 2000; when the DRC government – taken over by the Alliance des Forces Démocratiques pour la Libération du Congo-Kinshasa (AFDL) in 1997 – faced new uprisings in the east in the late 1990s; and when the Front Patriotique Rwandais (FPR) government engaged in battle with Forces Democratiques de Liberation du Rwanda (FDLR) on Congolese territory in the late 1990s.
A third insight relates to the types of continued CRSV. Among the most common forms of CRSV are: situations of hostage, sexual slavery or forced marriages, rape motivated by revenge, and sporadic, independent incidents of rape or torture. Sexual slavery continued, for at least three years, under the control of RUF, Renamo, the former Rwandan government in exile, LURD, and RCD. CRSV described as being motivated by revenge was either committed in the context of government institutions, such as prisons or detention centres, or in relation to other armed violence, such as when a government was still engaged in fighting another, still active rebel group or when rebel movements fought each other. There were reported rapes in detention centres and prisons by representatives of the Frelimo government, the Rwandan FPR government, as well as the South African government. The second category of reprisal attacks in parallel to armed fighting occurred in, for example, Burundi and DRC. Finally, most common forms of CRSV past conflict termination are reports of independent incidents, often rape, but sometimes also gang-rape. This points to a mix of both institutionalised forms of CRSV and more sporadic, independent incidents.
Almost all conflicts ended with negotiated amnesties attached to agreements negotiated at the time of or following end of fighting. In the case of seven rebel groups and the Rwandan government, there were either no negotiations or revoked former amnesties. The amnesties in Burundi and DRC were restricted by international law, the South African amnesty in turn restricted by domestic criminal law, covering only political crimes, while blanket amnesties were negotiated in Mozambique, Liberia and Sierra Leone. The Liberian amnesty later became circumscribed by international law in the mandate of the TRC, as was the Sierra Leonean amnesty by the international law-mandate of the Special Court for Sierra Leone (SCSL) in late 2000. There is, in other words, a great variation in the amnesty formation and a strong prevalence of negotiated amnesties in these particular conflict countries.
Very few agreements, and not a single amnesty, contain explicit references to CRSV. The only two countries that mention CRSV in their negotiated agreements are Burundi and the DRC.24 This is disappointing, considering these countries’ track records of CRSV after peace agreements. The general silence on CRSV in peace processes is not surprising, since acknowledgement of CRSV as not just being a bi-product of war, but an integral part of it, is relatively new. Though documentation exists on CRSV since World War II, coordinated international action to combat CRSV only began at the beginning of the twenty-first century. Efforts to address CRSV early in peace processes were first formalised only two years ago in the UN Security Council Resolution 1888 (2009).25
Based on this pilot study, it is clear that amnesties and continued CRSV are relatively prevalent in the seven studied countries. Yet, this does not mean that one leads to the other. While certain initiatives – such as the Liberian TRC and the domestic trials against mutinied MLC, RCD and government soldiers extracted some accountability26 in the complex conflict dynamics in Burundi, Liberia and DRC – a culture of almost total impunity, for both past as well as new CRSV, is often described as contributing to continued CRSV. The one case of full impunity – Mozambique – was followed by continued CRSV by both parties, which some suggests muted international criticism against continued sexual slavery.27 Also, the explicit threat of accountability, specified to past CRSV, included in the 2003 indictments of the SCSL, appears to have had a dampening effect on continued CRSV, though past the studied period.28
However, this does not mean that a clear relationship can be established between amnesties and continued CRSV. There needs to be more systematic studies for that, which in turn depends on accessible data. One needs to test the assumption that because amnesties exist, conflict parties expect impunity, which in turn encourages them to continue CRSV. Researchers have not established whether explicit amnesties differ from de facto impunity and what comes first. In other words, it remains unknown whether amnesties represent already prevailing cultures of impunity that may already have affected the calculations of the conflict parties, or whether their existence in fact brings new signals on impunity for past crimes.
Amnesties alone fall short of explaining changes over time after conflict termination. Although they may be important to understanding impunity, amnesties do not, for example, tell us why CRSV exploded in Burundi in early 2004 (conflict ended in 2003) or in Sierra Leone in 2001 (conflict ended in 2000), or why there were no reports of continued CRSV for the rebel movements to the Burundian Arusha agreement (28 August 2000).
To understand this, one needs to turn to trigger factors. Some of the trigger factors for continued CRSV that surfaced in the pilot study, and have also been explored in other research, were: organised armed violence; intact command structures of rebel movements; as well as government security forces, and limited DDR processes inadequately equipped to deal with the more structural legacies of CRSV, such as sexual slavery. For example, more institutionalised forms of CRSV continued in areas under the effective control of elements of Renamo in Mozambique, RUF in Sierra Leone and LURD in Liberia. Similarly, where the structures and staff of the apartheid regime remained relatively intact within the South African police, such as in the Vanderbijlpark Murder, and Robbery Unit in Johannesburg, old practices of systematic sexual torture continued.29
A final general lesson pertains to the still underexplored potential of alternative accountability measures anchored with the people, particularly in rural areas, where access to the formal judiciary often remains limited.30 Though local support may change over time, the Mozambican blanket amnesty and the South African amnesty-for-truth-recipe were indeed initially widely supported among the population.31 Conversely, many individuals and civil society organisations have voiced criticism against effective impunity in Liberia, Sierra Leone, as well as Burundi.32
No blue-print accountability mechanism can work effectively in all situations. Where arduous mechanisms, such as truth commissions, special tribunals or special courts take time to establish and formalise, alternative, already existing fora may play an important role, such as the Bashingantahe in Burundi or the Gacaca court system in Rwanda. However, more often than not, traditional conflict resolution mechanisms in Africa are also infused with gender-inequalities that will not benefit addressing CRSV.33 Pragmatism, creativity, flexibility and perceptiveness are crucial to address continued CRSV in a relevant and timely manner.
One of the main insights emanating from this Policy & Practice Brief is that CRSV does not usually end because incompatibilities are resolved or arms have been put down. The seven African countries attest to a persistent, and sometimes even institutionalised, legacy of CRSV. This legacy appeared in the form of sexual slavery under the command of former rebels and sexual torture practices within governments’ security apparatus.
Another general lesson is that almost all seven cases ended with negotiated amnesties at the time of or following end of fighting, though their form and implementation differed mainly based on limitations of international law. This does not, however, mean that they explain the dynamics of continued CRSV. Though commonly referred to as a fertile ground for continued CRSV in policy spheres, it is still unknown whether amnesties, in fact, play this role in reality. One needs to study amnesties in relation to other, more direct trigger factors for continued CRSV that remain relatively unexplored.
Not just acknowledging, but also addressing the pink elephant of continued CRSV, is a huge and pressing task that rests upon all nations. In order to fulfil this mission, one needs to be flexible, creative and in tune with the desires of the people bearing the bulk of the costs of armed conflict, not least CRSV.
