Abstract
Over the past decade, significant global attention has been paid to the issue of ‘widespread and systematic’ sexual and gender-based violence (SGBV). To contribute to the prevention of SGBV, researchers have examined the relationship between the presence of armed conflict and the causes of SGBV. Much of this causal literature has focused on the individual and group perpetrator dynamics that fuel SGBV. However, we argue that research needs to lay bare the roots of SGBV in normalized and systemic gender discrimination. This article brings back structural gender inequality as a causal explanation for SGBV. In order to better understand and prevent SGBV, we propose a critical knowledge base that identifies causal patterns of gendered violence by building on existing indicators of gender discrimination.
Introduction
Ending the use of ‘rape as a weapon of war’ has become a new global mantra. Indeed, the last decade has seen historic recognition that egregious acts of sexual violence merit international political and legal attention. The 1998 Rome Statute was the first international legal document to recognize widespread and systematic acts of sexual and gender-based violence (SGBV) 1 as an act of genocide, a war crime and a crime against humanity. There are now no fewer than seven United Nations (UN) Security Council resolutions on the cross-cutting theme of Women, Peace and Security. 2 One of these, UN Security Council Resolution 1820, prompted the creation of the Office for the UN Special Representative of the UN Secretary-General on Sexual Violence in Armed Conflict in 2008. This resolution mandated the Secretary-General to annually report to the Security Council on country-specific situations of sexual violence in armed conflict. 3 This significant international attention to sexual violence in armed conflict culminated in the launch of the UK’s Prevention of Sexual Violence Initiative in May 2012 by Foreign Secretary William Hague. Thus far, the Initiative has generated the UN General Assembly (2013) Declaration of Commitment to End Sexual Violence in Conflict, which is supported by over 150 states, and a new International Protocol on the Documentation and Investigation of Sexual Violence in Conflict (UK Foreign and Commonwealth Office, 2014). These developments illustrate the high-level political actions being taken to address sexual violence in conflict around the world.
To date, much of the causal analysis of SGBV argues that because unequal gender relations are inherent in all situations of widespread and systematic SGBV, they cannot explain why it occurs in some situations but not in others (Cohen and Nordås, 2014). This scholarship has stressed understanding sexual violence as the product of individual and/or group perpetrator choices – particularly the behaviour and social cohesion of rebels and soldiers in existing conflicts (see, inter alia, Wood, 2006; Butler et al., 2007; Cohen, 2013). While we acknowledge that this research has been pivotal in identifying acts of sexual violence as mass atrocity crimes and elevating sexual violence as an act of political violence (Wood, 2015), in this article we reinstate the centrality of gender analysis in explaining and preventing SGBV (Cockburn, 2010). The UN Secretary-General’s 2015 annual report to the Security Council on situations of sexual violence in armed conflict states that
across the varied contexts examined a common point is that waves of conflict-related sexual violence take place against a backdrop of structural gender-based discrimination…. These dynamics affirm that efforts to prevent conflict, foster equality and build gender-responsive institutions are central to eradicating the scourge of sexual violence. (Ban, 2015: 4)
Gender analysis examines this backdrop of structural gender-based discrimination and informs the prevention framework for SGBV in high-risk situations. It raises three critical questions for the causal research on conflict-related sexual violence. First, how and when do unequal gendered structural conditions within a society lead to the intentional use of sexual and gender-based violence to achieve political ends? Second, how do situations of endemic gender discrimination affect our knowledge of where SGBV is taking place, and who are the perpetrators and the targets of it? Third, what is the relevance of conflict to the use of sexual and gender-based violence, and in what ways does it exacerbate gendered political violence? A gender analysis of empirical cases of SGBV is by definition reflexive. It reveals the gendered, racialized and sexualized contexts of violence (Agathangelou and Ling, 2004; Enloe, 2000; Ni Aolain et al., 2011). In addition, gender analysis uncovers the social and political nature of SGBV data, sources and knowledge (Ackerly et al., 2006; Ackerly and True, 2008; Kirby, 2012). When we critically reflect on the power dynamics of the concepts, the methods of data collection and analysis that we use, as well as our own situatedness in the research field, ‘power is the researcher’s subject and not the researcher power’s agent’ (Ackerly and True, 2008: 699).
Whether motivated by ideology (Farr, 2009), social hierarchy (Cohen, 2013) or instrumental purpose (Wood, 2015), the decision to order or condone SGBV demands examination of the gendered structural conditions – that is, gender roles in society and power inequalities in institutions of the family, the state and the market – that make SGBV possible in the first place (Enloe, 2000; Carpenter, 2006; Baaz and Stern, 2013). Sexual violence is an act of political violence. Recognized in the Rome Statute as an act of genocide, a war crime and a crime against humanity, it does not require the presence of conflict, nor does its recognition require a large number of cases of sexual violence to have been committed (Office of the International Criminal Court [ICC] Prosecutor, 2014). The inclusion of gender-based persecution in the Rome Statute was an acknowledgement of the reality observed to date: that sexual violence is the product of gendered power relations (Office of the ICC Prosecutor, 2014: 13–14). Here, we define political violence as ‘any form of organized violence carried out by political actors, which include governments, rebel groups, insurgents, or terrorist organizations’ (Valentino, 2014: 91, emphasis added). If appreciated as an act of political violence, sexual violence is not a random opportunistic crime. The crime always seeks to produce and reinforce gender stereotypes (Gagnon, 2004; Brass, 2006; Cockburn, 2010). Understanding why institutions would order or permit sexual violence crimes to occur, moreover, requires an understanding of the gendered power relations in a particular society. Accordingly, we contend that causal analysis of SGBV and critical feminist analysis of gendered violence are complementary projects. Both intend to generate knowledge that can contribute to the prevention of SGBV through policy and operational interventions on the ground. What has been lacking is a dynamic and contextualized indicator-based approach to deepen our understanding of the relationship between structural gender discrimination and mass SGBV. 4
We develop this argument in two main parts. First, we examine the existing causal explanations for SGBV in armed conflict. We argue that these studies unintentionally compartmentalize gender, often treating it as a single, isolable variable, and neglect the volume of feminist research that reveals that gender cannot be disaggregated from instrumentality and political violence. The individual and group motivations for sexual violence must also explore why individuals and groups carry out these acts of violence. This focus, in turn, requires engagement with the gender roles that shape that behaviour and observation of significant gendered inequality within and between groups in each case. Second, to demonstrate our argument that structural gender inequality is a cause of SGBV, we investigate the relationship between the cases of sexual violence in high-risk situations listed in the UN Secretary-General’s annual reports and indicators of structural gender discrimination in those situations. Gender inequality varies significantly across countries and high-risk situations and, as such, it can and needs to inform strategies of prevention.
Gender and causal explanations for sexual violence in armed conflict
Between 2000 and 2014, ten studies of four or more armed conflict-affected cases have sought to explain the occurrence of systematic sexual and gender-based violence (SGBV), as illustrated in the online appendix. 5 These studies offer two significant explanations for the onset and prevalence of sexual violence in conflict situations. One explanation is that SGBV is caused by the presence of armed conflict. Sexual violence – in most studies primarily defined as rape – occurs when there are perpetrator dynamics fuelling these primarily opportunistic crimes. The other major explanation is that sexual violence is an instrumental strategy deployed against civilians for the purpose of war gain or plunder, enabling the acquisition of resources, land and power. These studies have been vital in identifying patterns of sexual violence and documenting cases of sexual violence as atrocity crimes. However, these explanations do not convincingly answer why SGBV as opposed to another form of strategic or opportunistic violence is used to achieve these ends. Whether or not sexual violence is enabled by ‘individual’, ‘group’, systematic or opportunistic dynamics, we must not underplay the social, political and economic relations that lead an individual or group to target specific victims for this particular form of violence (True, 2012). Our concern is that gender inequality and discrimination are not by definition ruled out as causes of SGBV because gender inequalities exist in all cases of armed conflict, and thus appear not to be able to discriminate between those cases where SGBV has been documented and those where it has not. What remains relatively unexplored is how escalation or change in patterns of gender oppression, abuse and humiliation may fuel the tactical use of sexual violence for political gains.
Salient causal studies
Ten major studies of armed groups seek to understand the strategic use of SGBV crimes in distinction from their opportunistic perpetration in war and conflict. Elizabeth Wood’s (2006) study of eight cases ranging from World War II (WWII) to the El Salvador civil conflict and the US-led war in Vietnam asks if sexual violence is a strategic weapon in waging war, why do all armed groups not use it? She draws on existing studies to compare ‘non-event’ cases where allegations of the perpetration of sexual violence by at least one armed group involved in the conflict were noticeably ‘absent’. Wood defines ‘absence’ as the rare or minimal reporting of sexual violence. She tests a number of potential causes and argues that there was little support for the type of conflict, the instrumentality of SGBV to the group, the militarization of masculinity and pre-existing uneven gender relations as causes of wartime sexual violence. Wood uses the method of falsification, pointing to an anomalous case that she claims disproved these classic explanations. Wood (2006: 325–326) questions the relationship between gender and sexual violence in conflict, asking why would gender inequality produce sexual violence in one setting but in another lead to the participation of women in rebel militias? Of course, in many cases, notably Rwanda (Brown, 2014), we know that the presence of women in armed roles can co-occur with mass SGBV, and moreover sometimes be a factor that propels female recruitment into armed groups. Causal explanations based on gendered social relations, Wood argues, cannot account for the targeting of particular people and groups for sexual violence. But that assumption, of course, implies that unequal gendered social relations is a unitary concept rather than taking many varying forms with respect to individual behaviour within and across ethnic, racial and substate groups and with respect to institutions and collective norms (Beckwith, 2010; Chappell, 2010). Moreover, the ways in which women and men are advantaged and disadvantaged as women and men and with respect to race, class, sexuality and other axes of disadvantage are defined differently in different subnational and country contexts and can therefore be comparatively analysed (Weldon, 2006: 235).
With respect to the causes of SGBV, Wood argues that the most compelling explanation, because it could not be falsified across any of the cases, is the individual opportunity to commit such violence facilitated by the lack of a hierarchical command structure and explicit norms against sexual violence within armed groups. Her study has spurred a debate within political science on the primary drivers of sexual violence in armed conflict, between scholars who argue that SGBV is largely opportunity-driven (Butler et al., 2007; Cohen, 2013) and those who argue that it is a deliberate strategy or weapon of war (Farr, 2009).
Debating the scale of the sexual violence being reported (isolated versus numerous reports) has also framed research on the causes of conflict-related sexual violence (Butler et al., 2007; Cohen, 2013). For example, Christopher Butler, Tali Gluch and Neil Mitchell’s (2007) investigation of why agents are motivated to perpetrate acts of SGBV refers to Wood’s (2006) earlier research. The authors restrict their study to SGBV perpetrated by members of state security forces. Data were collected for 163 countries for one year (2003), using a five-point scale that measured the level of rape and sexual violence from 0 to 4. Reports of ‘rape’, ‘sexual assault’ or ‘sexual abuse’ committed by security forces were sourced (using the US Department of State’s Annual Country Reports) to provide an SGBV scale from 0 (no reports) to 4 (reported as a ‘tool of war’ or ‘systematic weapon of war’) for each country. In this cross-sectional analysis, Butler et al. (2007) observed a high correlation between the presence of conflict and SGBV, particularly in situations where it was reported that the government had lost control of its army or commanding officers had lost control of their subordinates. In these contexts, sexual violence was again depicted as a form of ‘opportunistic’ violence that occurs with the breakdown of control and discipline. Level 4-type sexual violence occurred when the environment was permissive, rather than necessarily involving orchestrated and deliberate actions.
Kathryn Farr (2009) challenges Wood (2006) and Butler et al.’s (2007) studies, suspecting that there may be differences in the use of SGBV across types of conflict and perpetrator. She compared war rape in 27 armed conflicts reported by Project Ploughshares in 2007. Like others, Farr did not refer to the 1998 Rome Statute’s definition of sexual and gender-based violence, but she noted that there are few studies of ‘extreme war rape’, defined as ‘regularized, war-normative acts of sexual violence accompanied by intentional serious harm, including physical injury, physical and psychological torture, and sometimes murder’ (Farr, 2009: 6). Farr’s study is one of the few references to sexual violence as being part of a broader pattern of political violence, if not gender-based violence, against a particular group. Similar to Bastick et al. (2007), Farr was less concerned with explaining the scale of sexual violence in war and more concerned with understanding ‘differences in the prevalence and prominence of particular rape sites, perpetrator groups, and victim targets’ (Farr, 2009: 10). From 27 cases, Farr (2009) identified four war-rape patterns: field-centred/opportunistic; field-centred/woman-targeted; state-led/ethnic-targeted; and state-led/enemy-targeted.
Contesting previous findings, her analysis revealed varied reports of conflict-related sexual violence and different motivations and targets of violence, depending on the context of the conflict and what combatants were fighting over. In the conflicts she analysed, such as Liberia and the Democratic Republic of the Congo, where strong rebel groups were fighting to gain control over valuable resources in a country, SGBV tended to be highly prevalent and perpetrated by both rebels and state agents in the field area where the fighting was occurring (field-centred/opportunistic). Thus, rape may have a strong opportunistic component as well as a ‘strategic’ purpose. Field-centred/woman-targeted rape occurred more often in conflicts such as those in Afghanistan and Iraq, where both state and rebel groups were perpetrators, with women being specifically targeted as a way of controlling their movement and autonomy, including their dress, appearance in public and political activism. State-led/ethnic-targeted rape was perpetrated in contexts like Chechnya and Myanmar, where a powerful, controlling state attempted to quash a smaller secessionist or minority-representing group. In this category, rape victims were often targeted on the basis of ethnicity. Of note, Farr’s study reveals that gender norms within these societies appear to further inform the selection of victims for this particular form of violence. Finally, in Farr’s findings, state-led/enemy-targeted rape tended to target victims on the basis of ideological affiliations with rebel groups and activism critical of the state. In this situation, large and powerful rebel groups appear to have committed little SGBV, while state security forces tended to employ SGBV as a means of repression (e.g. Colombia, Sri Lanka). However, as Farr (2009) herself acknowledges, these findings are limited to situations already embroiled in violent conflict. What is not clear in this study is whether the motivations for different groups to deploy sexual and gender-based violence against different victims shifts depending on the phase of conflict, before, after and during armed fighting.
Like Farr, Jennifer Green (2006) examines ‘collective’ rape as a gendered form of political violence, but unlike Farr she explores the phenomena irrespective of the presence of conflict and draws on multiple English-language sources for reports of sexual violence from 1980 to 2003. Green (2006: ii) defines collective rape as ‘a pattern of sexual violence perpetrated on civilians by agents of a state, political group and/or politicized ethnic group’. She tests various causal factors, including gender inequality, militarization, armed conflict, ethnic cleavages, economic development, world culture and state structure, against cross-national data for situations where collective rape was reported. Her finding is that the reporting of collective rape increased dramatically from 1998, and that reports were not evenly distributed globally but tended to concentrate in Africa. Governments are found to be the most responsible, in 84% of episodes, for sexual assault committed against civilians. In nearly half (46%) of all episodes, members of a government group (or multiple government groups) were the sole perpetrators of collective rape (Green, 2006: 124). Green confirms the hypothesis that conflict intensity drives collective rape, as well as showing that militarization drives collective rape. An ambiguous causal relationship is observed between gender inequality and SGBV, possibly due to the presence of conflict affecting the gender variables. Similarly, Green finds that the presence of ethnic cleavages is not strongly correlated with the presence of SGBV. 6
More recently, Dara Kay Cohen (2013) has examined the occurrence of sexual violence in civil conflicts (defined by the Fearon and Laitin [2003] Civil War dataset). In this study, the operationalization of sexual violence (0–4 scale of intensity) and reporting source – US State Department annual reports – are similar to those of Butler et al. (2007). Cohen finds the socialization of armed groups to be the most significant explanation for the perpetration of rape. Insurgencies that forcibly conscript combatants, fight over lootable resources and are ‘aimed at the centre’ (i.e. not secessionist) are more likely to perpetrate rape in civil conflicts. State armed forces, however, are more likely to commit rape when they pressgang their fighters or when there is state failure (thesis provided by Butler et al., 2007). Cohen observes no evidence of a relationship between sexual violence and ethnic hatred, ethnic cleansing or genocide, nor any relation to gender inequality or discrimination. To test the effects of gender inequality on SGBV, she uses the (now discontinued) Cingranelli and Richards (2010) Human Rights Data Project on women’s economic, social and political rights, and fertility rates. According to Cohen (2013: 468), who cites a study by Caprioli (2005), the latter variable captures ‘cultural factors – such as personal choice and the need for children – and structural inequalities, such as lower levels of education and employment’.
Focused on perpetrator socialization, Cohen suggests that early-warning analysis should consider reports of abductions by armed groups as a sign of the escalating threat of wartime rape, which is especially acute in fragile states where there are resources available for plunder. Cohen’s finding that a permissive environment for rape perpetrators, more than the presence of gender inequality, ethnic grievances or the genocidal intent of the regime, causes SGBV has been corroborated by Cohen and Nordås (2014) using a larger conflict database (Uppsala Conflict Data Program) and expanded reporting sources (UN and non-governmental organization [NGO] reports in addition to US State Department annual reports). Quijano and Kelly (2012: 489) argue that Cohen’s study is an advance on the ‘[feminist] continuity [of violence against women] argument, which calls on gender subordination and culturally condoned violence against women to explain why wartime rape happens’ but ‘cannot adequately account for why some forms of [sexual violence in armed conflict] occur that are taboo in most cultures’. Cohen discounts feminist arguments about pre-conflict gendered conditions, noting that ‘scholars who study wartime killing are rarely asked to calculate rates of pre-war murder to determine if the culture was especially murderous before the outbreak of conflict’ (Quijano and Kelly, 2012: 489n154), yet sexual violence in conflict is subjected to continuity and ‘simplistic’ culture arguments (Quijano and Kelly, 2012: 488–489).
Variation in norms regarding rape and gender within and across groups, and variation of different types of sexual violence deployed in the phases and intensity of political violence and conflict, including situations prior to armed conflict, are neglected as causal factors (Cohen, 2013; Farr, 2009; Wood, 2006). Notably, conflict and open-source reporting are the master frames that determine the selection of the cases of SGBV studied (Peterman et al., 2011). None of the studies refers to or acknowledges the presence of gender-based violence or the ICC’s definition of sexual and gender-based violence. Even if the prohibition norms and strengthened punishment and sanctions regimes are vital for addressing SGBV, perpetrators’ attitudes, values, beliefs and interests must themselves be explained, as they do not arise ‘naturally’. Moreover, the lack of prohibition norms and punishment or sanctions regimes cannot explain why SGBV (and not other forms of violence) are overwhelmingly perpetrated against women and children. Such violence is constructed and legitimized within a social context. The culture of impunity for SGBV is itself a product of highly unequal and discriminatory gendered contexts. And we contend it is those gendered contexts that need to be understood as both part of the problem and the solution to systematic SGBV.
‘Missing variable’? Gendered analysis of sexual violence and conflict
A first step required in future studies of sexual violence in conflict-affected situations is to bring gender analysis in as a multifaceted form of explanation focused on structures, institutions and identities rather than single factors or individual-level variables (Harding, 1986; Connell, 1990; True, 2009; Davies et al., 2015). Discounting the relationship between gender inequality and SGBV on the basis of annual fertility rates, for example, does not unravel the degree of gendered discrimination in a society and has limited utility in revealing short- to medium-term changes in gender relations (Myrskyla et al., 2009). To date, gendered inequalities and discrimination have not been successfully integrated in causal studies.
Moreover, the current explanatory variables tested in studies of sexual violence in armed conflict – for instance, the recruitment strategies or moral codes of armed groups – are not ‘gender-neutral’, as they may be affected by socially constructed and culturally specific gender norms, and apply to groups differently depending on their gender makeup. 7 For instance, the targeting of civilian populations disproportionately made up of women and children should lead us to want to interrogate whether structural gender inequalities – disaggregated to provincial level – affect the patterns and location of SGBV. Thus, there is a need to gender-disaggregate data not only on suspected perpetrators and victims of SGBV crimes, but on the conflict itself more broadly.
Women are less likely to be politically represented in pre-conflict situations (Buvinic et al., 2013). In these environments, the policies and institutions that provide opportunities and incentives to report sexual violence are likely to be absent. The presence of institutionalized gender discrimination could be a major permissive condition for rape perpetrators. What distinguishes critical feminist analysis of SGBV, as Baaz and Stern (2009: 503) argue in their study of rape in the Democratic Republic of the Congo, is the understanding that all acts of violence exist on a continuum of violence facilitated by a (further) breakdown in law and order, which is intimately related to society’s hierarchy of gender, ethnicity, political and civil rights. In many post-conflict situations, there may be more opportunities to report sexual violence owing to the presence of humanitarian actors. Further, there are real incentives for reporting sexual violence as conflict-related because of the services and compensation now targeted at that form of violence (True, 2012). However, we also know that displacement and deaths by gender (including indirect deaths due to health and disability) continue to disproportionately affect women in post-conflict situations (see Ghobarah et al., 2004). 8 In situations where this is the product of ethnic and political discrimination, gender analysis can explain and contextualize the accuracy of SGBV reporting and identify whether gaps in reporting may reflect unequal access to reporting for one group versus another (Sooka, 2014; see also Cohen, 2013).
Finally, while causal analysis reveals important insights into the patterns of sexual violence, the concept of gender-based persecution identifies men and boys as victims, women as perpetrators, and a broad range of ICC–related sexual and gender-based crimes that may be deliberately deployed against members of minority groups, including lesbian, gay, bisexual and transgender (LGBT) persons (see International Gay and Lesbian Human Rights Commission [IGLHRC], 2014). Contrary to the view that feminist perspectives cannot appreciate men as victims of sexual violence (Quinjano and Kelly, 2012: 483–484), we argue quite the opposite. Though the implications of rape and other forms of sexual violence are different for women and men, they are equally gendered. The reason why rape is so traumatic and under-reported for men and boys is precisely because it undermines their achievement of masculine agency and political identity (Cockburn, 2010; Dolan, 2002; Zarkov, 2007). Though we may not agree with Osama bin Laden, it is significant that as the leader of Al-Qaeda he stated that there was nothing worse for a Muslim man than to be treated like a woman – to be raped, that is (Tickner, 2002).
A gendered analysis of society explains precisely why sexualized violence would be effectively used in societies where it is culturally taboo and associated with women as lesser subjects, not citizens (Engel, 2006; True, 2004). Indeed, reflexive feminist methodology leads us to ask why there are silences in the reporting of violence against men and boys (Lewis, 2014), as well as against women and girls. Accordingly, we question all the studies to date in which it has been claimed that SGBV did ‘not’ occur (Wood, 2009; Peterman et al., 2011; Human Security Report Project, 2012) because we have little understanding of the socially and culturally specific barriers to reporting for men and women, girls and boys.
Bringing gender back into the study of widespread and systematic SGBV enables a more dynamic and precise understanding of how gendered structural conditions affect the incidence and the reporting of SGBV crimes. We need greater sensitivity to understand the effects of sustained gender inequalities and sudden shifts in gender discrimination over a short period of time, particularly if these are accompanied by escalating discrimination against other groups at risk of exclusion. Gendered analysis is possible, for example, using the UN Convention on the Elimination of all Forms of Discrimination Against Women state party and shadow reports (Liebowitz and Zwingel, 2014), but it is not being further developed to seriously examine the relationship between gender inequality, discrimination and risk of SGBV. The next section argues that feminist perspectives on SGBV and conflict enrich causal analysis by mainstreaming gender within and across structural, institutional and individual-level variables. We illustrate the explanatory potential of gendered analysis of sexual violence in conflict through one approach using indicators of gender discrimination. This indicator-based approach reveals gender to be a multifaceted set of structures, institutions and identities in the widespread and systematic sexual violence prevention toolkit.
Using gendered indicators to analyse systematic sexual violence
The crucial intersection between causal and feminist analysis, we argue, is to identify when the use of SGBV becomes a logical, tactical form of political violence. To prevent these crimes, it is crucial to trace and intercept the moment where gender discrimination and gendered violence were normalized in a society to such an extent that, in situations of civil unrest and political violence, it becomes logical and instrumental to direct sexual and gender-based violence against political opponents (Von Joeden-Forgey, 2012; Valentino, 2014).
The presence of structural gendered inequalities and pervasive gender-based discrimination in a group or society remains a promising line of inquiry for understanding the causes and the utility of widespread and systematic SGBV. Currently, we do not have dynamic gender-disaggregated indicators that are able to capture changes and escalations in inequality, discrimination and phases of conflict, which could help to explain and predict SGBV. In this section, we outline one methodological approach to address this gap. This approach is consistent with a critical feminist perspective that emphasizes gender as a social relation and form of explanation, and that sees specific class, racial/ethnic and sexual constructions of identity as intersecting with gender. To develop a more systematic analysis of the relationship between gender relations and the use of SGBV, it draws upon indicators of gender inequality and discrimination founded on qualitative research by gender and country experts, and on the UN Secretary-General’s annual report on conflict-related sexual violence. Such analysis has the potential to transform early warning and prevention analysis of SGBV (and indeed all genocide early-warning analysis) (Davies et al., 2015). We are fully aware of the reductionism involved in indicators that quantify social relationships. However, potential reductionism can be mitigated with the addition of contextual analysis of particular SGBV and conflict situations in cases that show both strong and weak relationships between gender discrimination and the presence or high risk of SGBV.
To examine the relationship between gendered inequalities and the use of SGBV, we use the 2012 Indices of the Organisation for Economic Co-operation and Development’s Social Institutions and Gender Index (SIGI), the largest dataset of its kind (OECD, 2012). In our view, the SIGI is the best available indicator for gender discrimination by country because it provides a ‘composite measure of social institutions which are mirrored by societal practices and legal norms that produce inequalities between women and men in non-OECD countries’ (Branisa et al., 2009: 1). The SIGI measures the underlying factors that lead to gender discrimination and the various forms in which it manifests are captured in five subindexes – restricted physical integrity, discriminatory family codes, son bias, restricted civil liberties, and restricted resources and entitlements. We compared average levels of gender inequality among all developing countries (SIGI list) with the levels of gender inequality reported in the UN Secretary-General’s list of countries with documented occurrence of widespread and systematic sexual violence (SG list). 9 Table 2 shows the results of this comparison of the two sets of countries.
Comparing 2012 SIGI average and average score for gender inequality in the 2012–15 Secretary-General list of sexual violence.
Data for countries in the OECD SIGI may be affected by the presence of armed conflict in the territory concerned, but what is presented in the SIGI score is not just a one-year snapshot of fertility rates or women’s access to labour and education, where the presence of armed conflict may dramatically affect the score (Ghobarrah et al., 2004). Rather, the SIGI measure of a country has been literally years in the making. Interestingly, most but not all of the bottom ten countries with the poorest gender discrimination scores on the SIGI in 2009 and 2012 are in conflict or in transition from conflict (Branisa et al., 2009: 10; OECD, 2012). Table 3 shows the conflict intensity and one-sided violence dataset from the Uppsala Conflict Data Program (UCDP, 2014) to provide a context for the situation on the ground in the countries reported by the UN Secretary-General as being at high risk for sexual violence. The UN Secretary-General’s reports on situations of high risk for SGBV crimes are informed by the presence of political violence rather than armed conflict per se. We argue that the higher-than-average rates of gender inequality in SG list countries combined with the intensity of political violence in these situations warrants further attention to the relationship between systemic structural gender inequality and the use of SGBV as a rational and instrumental act of political violence. By comparing SIGI developing countries with the subset of SG list countries, Tables 2 and 3 highlight three trends that call for further analysis of gendered inequalities as part of the complex causal pattern of SGBV.
2012–15 lists on sexual violence in armed conflict, gender inequality and mass atrocities.
Notes: *SIGI range starts from highest value (out of 0–1) and stops above median – Myanmar at 0.245. †These countries did not have aggregate scores across the five subindexes in the 2012 SIGI; however, Angola, Central African Republic and East Timor are included in Table 2’s correlational analysis as they reported data under four of the five subindexes. South Sudan did not exist when the SIGI data were collected. ‡As discussed in the third section of the article, the criteria for Colombia and Libya were slightly different from those of the other cases.
First, as Table 2 shows, gender discrimination in all its dimensions is much higher on average in the 29 countries on the SG list than it is on average across the 92 countries that are on the SIGI list but not on the SG list. 10 In particular, statistically significant relationships were found between the lack of a law against domestic violence in a particular country and being on the SG list; greater acceptance in public opinion of domestic violence and being on the SG list; poorer access to property, resources and entitlement for women and being on the SG list; and limited access to public space for women and being on the SG list. In the context of a global movement promoting the diffusion of laws against domestic violence (Htun and Weldon, 2012), countries that either do not have such laws or have inadequate legal frameworks are countries where the state has failed to address its responsibility to protect against and prevent SBGV wherever it occurs and are therefore likely complicit in the strategic use of sexual violence. The analysis of land, property and credit taken together suggests that women’s access to and control of key economic resources and entitlements may affect the propensity of state and non-state actors to engage in – or sanction – systematic sexual violence, especially in contexts where many conflicts, including those on the SG list, are related precisely to the control and exploitation of resources. 11
The strong association between countries with documented sexual violence and women’s limited access to public space suggests that SGBV is indeed a strategy of political domination and violence. In contexts where women’s and girls’ bodies and dress become symbolic of group identity and difference, particularly in ethnic, nationalist and extremist groups, controlling women’s mobility and public participation is part of the dynamic of intergroup conflict and itself a way of further carrying out conflict. In Bosnia and Herzegovina, Charli Carpenter (2010) refers to the pervasiveness of ‘ethnoreligious identity’ before the outbreak of war – the systematic construction and imagery of women as ‘symbols and bearers of national culture … their perceived role as cultural symbols representing national territory is precisely what makes women vulnerable to violence, including rape, during conflict’ (Carpenter, 2010: 140–141). Perpetrators of sexual violence targets women, particularly of mixed ethnicities, because they are seen to be reproducers of ‘impurities’ (Korac, 1996). Thus, SGBV is targeted at civilians whose gender or sexual status intersects with minority status, which may be based on ethnicity, religion or political affiliations.
Second, as Table 3 shows, between 2012 and 2015 the UN Secretary-General’s combined list of conflict, post-conflict and other situations of concern for SBGV included more non-conflict countries than countries currently experiencing armed conflict. 12 This indicates not only that SGBV occurs in non-conflict situations and must be studied independently of conflict, but also that it may be a specific risk factor for conflict as well as the escalation or return to conflict.
Third, among the 29 situations listed in the report, 15 of these experienced one-sided violence over the 2011–12 period (UCDP, 2014). This finding is important, as it correlates with recent research in political science on the deliberate targeting of civilians by military organizations rather than random acts of primordial violence based on ‘ancient’ hatreds (Valentino, 2014: 93, 99). Accordingly, early warning and prevention efforts should be directed to the structural political instability that gives rise to regimes that use violence, including opportunistic and strategic SGBV, to retain power.
With respect to the strong correlations observed in Table 2, there are only two countries that are above the SIGI world average on gender equality and discrimination where there is ongoing violence, as well as widespread and systematic situations of sexual violence: Colombia and Libya. Because these are the outlier cases that could potentially falsify our argument about the complex causal relationship between gendered inequality and discrimination, political violence and the prevalence of SGBV, it is worth analysing them in more depth.
Though gendered inequality and discrimination in Colombia are marginally above average for a developing country, gendered inequalities are nonetheless part of a complex pattern of causation for widespread and systematic SGBV in that country. Prolonged conflict and post-conflict land-redistribution agreements as a result of conflict have disproportionately affected Colombian women, especially Afro-Colombian and indigenous women and girls, who have also been disproportionately targeted for sexual violence in the Colombian state (Ban, 2012: 6, 2013, 2014, 2015). Colombia has a history of intense conflict despite being listed as an area of minor conflict with less than 1,000 deaths per annum (UCPD, 2014). Government troops, paramilitary formations supported by the government and armed militias have all perpetrated violence against displaced and indigenous women. These women have often been victims of one-sided attacks solely aimed at achieving territorial control and control over vulnerable local populations for rent-seeking purposes (Ban, 2012). The continued high rates of SGBV crimes documented in the UN report do not necessarily correspond to a war strategy; however, according to the UN, ‘they constitute a generalized practice that takes advantage of the conditions of subordination of women, their precarious economic conditions resulting from lack of protection by the State, and the acceptance of existing ideas in the local culture, such as that a woman’s body is an object that belongs to men’ (Ban, 2012: 6). Of particular importance, the gendered economic and social status of displaced and indigenous Colombian women sheds light on SGBV perpetrators’ motivation of achieving larger political and resource ends – for instance, halting the return of populations to paramilitary/rebel-controlled areas and controlling the income of those who do return. Rather than discounting the causal relationship between racialized, gendered inequalities and SGBV, the Colombian case illustrates the value of a gendered analysis of SGBV crimes. It demonstrates the need to critically disaggregate national-level indicators to overcome biases in static, aggregate data that are not adequately dynamic or detailed enough to be sensitive to gender and other minority groups at risk of being targeted for SGBV crimes.
The complexity around indicators of gender discrimination with respect to physical/bodily integrity and their causal relationship to sexual violence is best illustrated with the Libya case, first listed in the UN Secretary-General’s 2012 report on sexual violence in armed conflict (Ban, 2012). At the time of this report, this was an intense conflict, with armed combatants fighting for complete control and administration of the Libyan government and all its territory (UCPD, 2014). A high level of one-sided attacks on civilians by both government armed forces and militias was reported (UCPD, 2014). However, sexual violence was not widely reported during the Libyan conflict. This remains a puzzle until we learn that in Libya, ‘prior to the onset of the conflict there was an absence of state structures to address sexual violence which not only exacerbates the situation [but] strengthens the silence of victims and witnesses’ (Ban, 2012: 12). The few reports of rapes mostly came from external witnesses reporting rapes they documented in prisons, especially crimes of rape and sexual mutilation committed against women and men, girls and boys imprisoned by the forces of Muammar Gaddafi. Since the end of conflict in 2011, there have also been reports of sexual violence perpetrated by the new administration against prisoners alleged to be supporters of the Gaddafi regime (Ban, 2012: 11–12). Most testimonies indicate sexual violence was used as a form of punishment – a distinctly gendered form of punishment in a society where women and men would be reluctant to come forward with reports of sexual violence because it is a profoundly shaming and feminizing act in Libyan society. As the 2012 report of the UN Secretary-General noted, there is little doubt that many women who were raped feared to report it because of the social stigma associated with being a victim of rape. This does not mean that sexual violence did not take place. Rather, it indicates that at the national level gender measures may not reflect – and indeed may obscure – the actual gendered obstacles faced by women, particularly those of a minority group, when seeking to report violations of their physical integrity (Ban, 2015: 30).
In the cases of both Colombia and Libya, sexual violence was referred to as a form of punishment, where the intention was to use sexual violence specifically to shame the particular groups attacked. The increased targeting of women, men and children for sexual violence may be an early indication of the intention to commit mass atrocities against specific political and ethnic groups. This possibility is noted in the reports of the UN Secretary-General on other country situations of sexual violence where sexual torture during imprisonment (Syria) and the targeting of single mothers (Mali and the Central African Republic) were, in hindsight, precursors to dehumanization and displacement practices used to single out minority political and ethnic groups (Ban, 2012; Human Rights Watch, 2013). Although Colombia and Libya appear to be outliers, not supporting the statistically significant relationship between high levels of gender discrimination and perpetration of SGBV, they nonetheless evidence the causal role of gendered norms influencing the deployment of sexualized forms of torture. Both countries had previously been known to use gendered forms of punishment and torture within state and military institutions. This record alone should lead researchers to focus on the ostensibly ‘hidden’ structural gendered conditions that promote the widespread perpetration of rape (Kirby, 2012).
Conclusion
Current research views sexual violence as a by-product of armed conflict, as opportunistic acts that occur in the theatre of war or as instrumental acts of war used to achieve other tangential aims, such as seizure of land or resources and political repression. By contrast, we conceptualize widespread and systematic SGBV as an act of political violence. This form of political violence is informed by discriminatory societal norms around gender, including discriminatory family codes that entrench men’s domination over women and children; severe restrictions on women’s civil liberties and access to resources and entitlements; institutionalized bias toward sons, boys and men; and normalized, everyday violations of women’s physical and bodily integrity. Such endemic gendered discrimination and oppression makes the use of SGBV an effective way of shaming, destabilizing and displacing whole groups in civil unrest, before, during and after conflict.
In this article, we have shown how the existence of gendered structural inequalities, institutions and identities are part of the complex pattern of causation that gives rise to SGBV in particular contexts. Much of the causal scholarship on sexual violence in conflict, however, has assumed that because gender inequality is everywhere it cannot logically explain SGBV anywhere. We question the absence of gender analysis from this scholarship and challenge the ‘single variable’ representation of gender inequality operationalized through a composite human rights score or fertility rate score. Gender inequality is a multifaceted concept rather than a single variable; thus structural gender discrimination cannot be ruled out as an explanation for widespread and systematic sexual violence. If the social environment of armed groups is a cause, that social environment is a gendered construction that shapes the perpetrators’ attitudes, values, beliefs and interests. These are intimately related to the ‘value’ of women and men, along with their political, economic and social relationships in a given society. In some cases where armed groups consciously and actively choose to use SGBV to achieve group cohesion, such a strategy is only effective because of uneven gender power relations in that group and society.
Equally concerning is that much of the research on the causes of sexual violence in conflict has not adequately appreciated the politicized context of sexual violence reporting and data collection on the ground, especially in humanitarian and conflict-affected situations. Where there is endemic gender oppression within a group or society and a culture of impunity for SGBV, the lack of any reports (especially prior to the onset of conflict) may actually be indicative of widespread and systematic SGBV rather than evidence that it has not occurred. Very real challenges with data collection, including gaining access to SGBV reports by local groups, has led researchers to analyse the causes of sexual violence globally on the basis of incomplete data, despite the potential biases and selectivity of such an approach. For instance, many studies focus exclusively on the explanatory power of a particular type of armed conflict, perpetrator and crime – usually rape (Lancet, 2014). Because of the reporting limitations and dangers, we have relied on existing quantitative datasets or on secondary qualitative data, but we need to contextualize the serious limitations of those data. At the very least, to enable us to assess knowledge claims about conflict-related sexual violence, scholarship should highlight and explore the ramifications of these limitations on its conclusions.
Bringing back gender analysis to explain widespread and systematic SGBV not only contributes to the methodological rigour of international peace and security research, but is crucial for prevention at the point of escalating violence and in peacebuilding. Policy recommendations on conflict-related sexual violence increasingly favour prevention focused on the prosecution and socialization of individual perpetrators and reforming justice systems to end impunity for SGBV crimes. These interventions are important, but so too is addressing the significance of gendered inequalities and discriminatory practices in societies, especially where these are intended to exclude and marginalize minority groups. Failing to mainstream analysis of the gendered structures, institutions and identities of both perpetrators and victims will ultimately undermine the effectiveness of efforts to eliminate conflict-related SGBV and prevent it in the long term.
Footnotes
Acknowledgements
We would like to acknowledge the research assistance of Sarah Hewitt, Sara Meager, Dyonne Pennings, Maria Tanyag, and Barbara Trojanowska. We would like to thank the University of Queensland, Monash University, the Oceanic Conference on International Studies and the Australian Political Studies Association for hearing earlier versions of this article and providing generous feedback. We sincerely thank the anonymous reviewers and editors of the journal for constructive comments and engagement with the piece.
Funding
This research was supported under the Australian Research Council’s Discovery Projects funding scheme (project no. DP140101129).
