In 1977 the socialist Eritrean People's Liberation Front (EPLF), while in the midst of a brutal conflict with the Ethiopian government, declared the overhaul of marriage customs in the territories they controlled a “major priority.”Footnote 1 The insurgents made sweeping changes, allowing women to marry without parental consent, banning polygyny, dowries, and bride prices, improving divorce rights, and giving women representation in civilian governance.Footnote 2
Forty years later, Ansar al-Sunnah, an Islamist insurgency loyal to the Islamic State, emerged in Mozambique. The group challenged customs in ways similar to the EPLF: encouraging women to marry without parental consent,Footnote 3 giving men loans to pay bride price,Footnote 4 and conducting their own marriage ceremonies.Footnote 5
Across every continent and ideology, armed groups interfere in the intimate lives of civilians, using precious resources to regulate marriage, divorce, and gendered behavioral expectations. These contentious and consequential strategies were described as “more of a threat than guns” by one Nepali commentator.Footnote 6 Yet scholars have yet to systematically analyze the risks and benefits of challenging civilian gender customs from a cross-national perspective.
Moreover, intervention in civilian conduct often provokes resistance from the local population.Footnote 7 Given the risk of backlash, it seems misguided for rebel groups to prioritize regulating personal conduct in this way. So why do so many insurgencies challenge civilian gender customs during conflict?
Using an original data set of 137 African armed groups fighting between 1950 and 2019 and two qualitative case studies, I argue that insurgencies challenge gender customs to reshape local power relations, undermining local elites by exploiting pre-existing gender grievances and empowering targeted subsections of the population. This strategy tightens social control in the territories where they operate, reducing the likelihood of resistance to rebel rule. Alternatively, rebels who wish to forge alliances with local elites will reinforce the gender status quo, using insurgent gender governance as a “bargaining chip” to mollify potential rivals and lessen the chance of elite-led resistance.
Responding to the call to “take love and care seriously” in security studies,Footnote 8 this paper makes a range of contributions to the study of conflict. First, it contributes to the literature on insurgent legitimacy, deepening our understanding of how armed groups interact with civilians.Footnote 9 Second, it improves our understanding of how gender inequality contributes to state fragilityFootnote 10 while contributing to scholarship that looks beyond violence to examine the importance of gender in conflict.Footnote 11 Third, it emphasizes the role of both gender and local elites in conflict governance, contributing to a growing scholarship on multilayered governance and state formation during and after conflict.Footnote 12 The article proceeds as follows. In the first section I examine the literature on gender and rebel governance before outlining my own theory on insurgent gender governance. Next, I introduce my original data set and conduct quantitative tests of my hypotheses. I further illustrate my findings using case studies of Katiba Macina and al-Qaeda in the Islamic Maghreb, two Islamist groups operating in Mali.
Rebel Governance As Social Contract
Insurgent governance, defined as “the set of actions insurgents engage in to regulate the social, political, and economic life of non-combatants during war,”Footnote 13 is costly, and is said to be motivated by the desire to cultivate legitimacy.Footnote 14 Rebels offer civilians public goods in exchange for their cooperationFootnote 15 while highlighting the failure of the state to fulfill the social contract,Footnote 16 boosting domestic and international support as well as military strength.Footnote 17
However, the advantages such a group derives from gender governance become uncertain when compared to other public goods. Even conventional rebel governance does not directly correlate with increased rebel strength or support, as interventions in civilian conduct often elicit resistance from the local population.Footnote 18 Surveys also find that civilians have a more favorable view of rebel groups that respect local traditions.Footnote 19 Challenging gender norms is bound to be particularly contentious, striking at the heart of intimate family dynamics. The gender interventions of armed groups can and do lead to civilian backlash. Ansar Dine, for example, faced civilian resistance after implementing dress codes in Mali.Footnote 20 The Democratic Union Party (PYD) in Rojava faced backlash against its attempts to outlaw polygamy.Footnote 21 The frequency with which insurgents’ interventions anger the local population shows that social contract theories do not tell the whole story when it comes to gender governance.
Ideological Explanations for Rebel Governance
Another explanation is that gender governance is driven by ideology,Footnote 22 with rebels sacrificing civilian compliance in a “power-ideology trade-off.”Footnote 23 Marxist, Islamist, and Secessionist groups engage in more state building during conflict,Footnote 24 with Stewart demonstrating that groups classified as “revolutionary” engage in more costly governance initiatives.Footnote 25 Ideology also shapes sexual violenceFootnote 26 and the role of women in armed rebellion, with leftist groups endorsing gender egalitarianism and incorporating more female leaders,Footnote 27 suggesting they might use gender governance initiatives that improve the lives of female civilians. Radical Islamic and other fundamentalist religious groups, in contrast, frequently strive to preserve patriarchal gender roles.Footnote 28 This implies they might use gender governance to reinforce male dominance.
Yet ideological categories cannot always predict how insurgents use gender governance. Some Islamist armed groups, like Algeria's Groupe Islamique Armé, adopt a repressive gender governance strategy, targeting women who participate in public life and carving out “enclaves” where women were barred from consulting male physicians.Footnote 29 Some, like al-Shabaab, combine repressive policies toward women, such as strict limitations on freedom of movement, with greater divorce and inheritance rights.Footnote 30 Others, like the Signed-in-Blood Battalion in Mali, leave civilian gender norms untouched.
Notably, gender governance is also used by armed groups that lack a clear ideology, including Sierra Leone's Revolutionary United Front and the National Patriotic Front of Liberia. There are also frequent discrepancies between groups’ stated goals and their behavior on the ground, with some left-wing organizations, like the Movimento Popular de Libertação de Angola, largely maintaining the gender status quo and failing to live up to their emancipatory rhetoric.Footnote 31 Overall, evidence suggests that the use of gender governance is not straightforward. The frequency with which gender governance provokes backlash suggests this strategy is not designed solely to boost civilian support. However, the unpredictable ways in which groups use these strategies suggest that ideological explanations are incomplete.
This study contributes to a small but growing literature on gender governance. Giri's pathbreaking study of the Maoist insurgency in Nepal describes how the group used the nonviolent governance of marriage and sexuality to accumulate control.Footnote 32 Recent scholarship has also examined gender and rebel governance at the group level, arguing that practices are reflective of a group's broader political projects, organizational cohesion, and resources,Footnote 33 while gendered participation dynamics also shape rebel attempts to govern civilian populations.Footnote 34 However, less attention is paid to rebels’ attempts to change gender norms. Responding to this noted gap in the literature,Footnote 35 I introduce an original theory on insurgent gender governance as a facet of rebel–civilian relations.
Theoretical Overview
I argue that gender governance is used as a tool to reshape power relations, driven by a group's approach to local elites. My framework draws on the state-in-society approach,Footnote 36 with state building a struggle for social control between state and nonstate orders. Defining social control as “access to people and their resources,”Footnote 37 I see this endeavor as crucial for any political leader.
Scholarship considers rebels’ provision of governance as both a challenge to the stateFootnote 38 and an attempt to win over the civilian population,Footnote 39 assuming that the distinction between civilian and state is clear. However, insurgencies tend to emerge in contexts where the state coexists with alternative modes of authority.Footnote 40 Nonstate actors—from tribal, traditional, ethnic, and religious institutions to private businesses, NGOs, and criminal groups—frequently take on state-like roles.Footnote 41 Local elites, and particularly traditional leaders,Footnote 42 who I focus on in this manuscript, exert authority over facets of governance including land allocation,Footnote 43 dispute resolution, and security.Footnote 44 Around 35 percent of the world's population is “traditionally organized,” but this rises to 83 percent in Sub-Saharan Africa.Footnote 45 The following section will consider why control over gender norms is an essential element of local elite power before turning to my theory of insurgent gender governance.
Why Gender and Local Elites?
Following Htun and Weldon,Footnote 46 I define gender not as an aspect of individual identity but as a collection of institutions: a set of “rules, norms, and practices” that construct what it means to belong to a particular sex. Local elites, especially traditional leaders, derive their authority from the maintenance of gender rules, norms, and practices.Footnote 47 The origins of this dynamic are argued to lie in security provision for clans, tribes, and lineages.Footnote 48 Facing threats of violence from other men, males forge alliances with other males, often based on kinship ties.Footnote 49 Because the clan can survive only by controlling reproduction, gender relations and marriage practices become subject to patrilineal authority, leading to hierarchy among males within the group and the strict control of women.Footnote 50
The shift to agriculture is posited to have strengthened local elite control over gender relations. Long-term collaborative investment in the land increases focus on group survival and continuity,Footnote 51 making kinship and reproduction a key communal concern. Reproductive capacity in patrilineal societies depends on the capacity of leaders to negotiate women into the group.Footnote 52 Thus, to ensure that “the girls of their community are available for exchange,” elders must exert tighter control over women's freedom and rights.Footnote 53
Others have noted that colonial and postcolonial states reinforced these systems,Footnote 54 enhancing local elite control over women and young men through legal pluralism. Colonial states often retained and standardized customary or religious law to govern disputes within the “private realm”Footnote 55 while local elite power was eroded in other domains. This codification process allowed male elders to enshrine their interpretation of these laws in new legal systems,Footnote 56 even as their authority was weakened in other arenas. In the present day, resistance to improving women's legal status still frequently comes from religious and traditional leaders who benefit politically and economically from maintaining the gender status quo.Footnote 57
At the same time, local elite authority generates specific gender grievances, particularly among women and young men. Customary courts, typically dominated by powerful male elites, wield significant power,Footnote 58 often excluding women from decision-making councils and requiring male representation.Footnote 59 Gender inequality is also pervasive in many customary law systems, where women may need their husband's permission to open a bank account, enter into a contract, apply for a loan, or work outside the home.Footnote 60 Adultery is often a worse crime for women than men; divorce is easier for men to obtain; and child custody may default to fathers.Footnote 61 Many customary systems condone forced marriage, marital rape, and domestic violence,Footnote 62 permitting men to “discipline” their wivesFootnote 63 or “solving” rape cases by requiring rapists to marry their victims.Footnote 64
Marriage and reproduction heavily influence status in customary structures. Unmarried men may be considered boys and be ineligible for certain political positions.Footnote 65 Bride price and polygyny incentivize early and forced marriage, increase rates of domestic violence, and erode reproductive freedom for women.Footnote 66 These customs also block access to marriage for young men.Footnote 67 Moreover, young unmarried men and women are also often excluded from land ownership under customary law.Footnote 68 Local elite control over labor and land allocation further exacerbate these gender grievances. For example, young men in Sierra Leone often work off bride-price debts through unpaid “bride service.”Footnote 69
Overall, the authority of local elites over their local communities generates gender grievances, particularly among women and young men who are excluded from social, political, and economic power. Armed groups, I suggest, can use gender governance to reshape the power balance between local elites and these marginalized communities. I now turn to the factors determining rebels’ approach to local elites, before considering how this impacts their use of gender governance.
Determinants of Rebels’ Approach to Local Elites
Rebel groups vary in their approach to local elites.Footnote 70 Some choose a cooperate approach, consulting with local elites, relying on them for civilian governance, or incorporating them into their own institutions. Others take a pure challenge approach, undermining local elites using violence and replacing them with their own institutions.Footnote 71 Alternatively, a partial challenge approach combines a largely coercive approach with elements of co-option.
While rebel groups and states differ, insights can be drawn from state–customary relations. Ideology might theoretically influence the approach to local elites, with leftist leaders more inclined to challenge traditional structures and conservative leaders more inclined to preserve the status quo. But variation rarely falls neatly along ideological lines. Leftist governments in South Africa, Zimbabwe, Uganda, and Angola introduced legislation which actively bolstered the powers of local elites.Footnote 72 Variation is also seen among rebel groups. The National Liberation Front (FLN) in Algeria, for example, forged alliances with these leaders,Footnote 73 while the Marxist EPLF in Eritrea sought to undermine their authority.Footnote 74
Considering why states diverge from their stated approach to local elites, Boone contends that ideology is “too blunt an instrument for explaining institutional choice,” with regimes choosing strategies that maximize their advantage within their context.Footnote 75 My framework concurs with this assessment, arguing that the primary interest of armed groups is organizational survival,Footnote 76 with the two most crucial aspects being security and access to resources. The path an insurgency takes will depend on their assessment of these factors, as demonstrated in Table 1.
TABLE 1. Drivers of approach to local elites
The Local–Elite Dilemma
Rebel groups in the initial stages of rebellion are often vulnerable and dependent on civilian support.Footnote 77 They need recruits, personnel, food, weapons, finance, shelter, and information.Footnote 78 Access to these resources may hinge on cooperation with local elites. Just as politicians leverage local elite networks,Footnote 79 alliances with elites may enable armed groups to make use of new networks. This may help with recruitment, as individuals may be incentivized to join armed groups due to traditional social obligations.Footnote 80 More generally, local elites may foster cooperation from the civilian population,Footnote 81 making them useful allies.Footnote 82 Making use of existing structures may also be more efficient than building institutions from scratch.Footnote 83
On the other hand, cooperating with local elites comes with risks. Communities with strong leaders and pre-existing organizations are more able to resist rebel rule.Footnote 84 Local elites may forbid residents from collaborating with rebels or use the threat of resistance to extract concessions.Footnote 85 Cooperating with local elites therefore may strengthen their negotiating power vis-à-vis the rebels.Footnote 86
Rebel groups therefore encounter a dilemma. Strong local elites are more helpful for resource accumulation, but they are also more able to mobilize civilian resistance. Weak local elites pose less of a threat but may also be less effective as partners. When considering whether the benefits of cooperating with local elites outweigh the risks, a group will consider two questions.
First, are local elites likely to be loyal? Civil war tends to be marked by a lack of front lines.Footnote 87 With spies and agents of either side blending into the civilian population, betrayal often becomes a “pervasive obsession” for insurgents.Footnote 88 Local elites often cooperate with the state on matters of governance,Footnote 89 leading the armed group to view them as “threats behind the frontlines”Footnote 90 or functionaries of the state.Footnote 91 Local elites may also have ties to rival armed groups, raising the risk of collaboration. On the other hand, if that risk is low, these elites may be seen as valuable allies.
Second, are local elites essential for access to resources? One element of this calculation is the armed group's existing resource base. Alternative resources such as foreign government sponsorship, access to natural resources, criminal activity, or direct civilian taxation may reduce reliance on local elites. Another element is the resources of the local elites themselves, as local elites with easy access to social and material resources will be more useful and harder to bypass. Finally, rebels will consider the local context. Situations of intense government repression, for example, may allow armed groups to easily access recruits and other resources without the use of local elites as intermediaries.
The cooperate approach
A group that believes that local elites are unlikely to collaborate with rivals and relies on them for resources will be likely to adopt a cooperative approach. UNITA in Angola is an example. Despite its origins as a Maoist organization, the movement depended on traditional authorities (sobas) for social and material resources, using these leaders to build legitimacy among diverse communities in the territory it controlled.Footnote 92 With the ruling Movimento Popular de Libertação de Angola having little influence over the sobas in UNITA-held territory,Footnote 93 the movement could be relatively confident these leaders were not compromised by ties to their rivals or to the central state.Footnote 94 As one former UNITA soldier recalled, “the relationships [between UNITA and the sobas]) were all-embracing, revolutionary, tangible. There were never any problems between UNITA and the sobas … There was never any ill-feeling.”Footnote 95
The pure challenge approach
In contrast, a group with an alternative resource accumulation strategy that believes local elites are tied to the central state will adopt a pure challenge approach. The Revolutionary United Front in Sierra Leone is one example. This group killed chiefs or forced them into exileFootnote 96 before replacing them with its own “town commanders.”Footnote 97 As one person recalled, “No one wanted to be identified as ‘the chief’ as that meant you were going to die at any time there was an attack.”Footnote 98 This strategy can be explained by the fact that the group did not rely on local elites for material or social resource accumulation,Footnote 99 while the faltering legitimacy of these elites among some people made a pure challenge approach less risky.Footnote 100 Moreover, chiefs were closely tied to the state and therefore an inherent security risk.Footnote 101
The partial challenge approach
Other groups face a more complicated situation, relying on local elites for access to resources while seeing them as security risks. Alternatively, a group may have other means of accessing resources and see elites as unlikely to collaborate with the state. Local elites may also pose a threat because of their legitimacy with the general population. Thus the group might be incentivized to weaken these leaders to lessen the chance of elite-led resistance while at the same time avoiding the risks of indiscriminate violence. This would be a partial challenge approach.
Al-Shabaab is an example. The group assassinated some local chiefs, and some decided to flee the area. Those who stayed had their responsibilities curtailed.Footnote 102 At the same time, the organization did utilize these local elites for governance efforts, ultimately deciding to keep the overall system of elders in place rather than get rid of it all together. This partial challenge approach can be explained by the fact that the organization was reliant on clan elders for both material and social resource accumulation, with these local elites perceived as legitimate by the general population. For example, clan elders were ordered to provide a certain number of recruits for enrollment into al-Shabaab forces. As one former commander described, they decided to keep the system of elders “because clan elders are important for the community. Without them, it would be difficult to rule the population.” However, the organization was also wary of the potential for collaboration with their rivals. Al-Shabaab saw many elders in the area as supporters of rival Islamist group Mu'askar Ras Kamboni. This suspicion led the organization to target certain clan leaders with violence, with one former commander saying that two elders were killed because “al-Shabaab saw them as secret supporters of Madobe, a Mu'askar Ras Kamboni commander.”Footnote 103
Approach to Local Elites and Insurgent Gender Governance
Overall, a rebel group's approach to local elites is driven by concerns for organizational survival. Every approach, however, continues to come with risks. Groups adopting a challenge and particularly a partial challenge approach are incentivized to undermine local elites to lessen the chance of elite-led resistance. Targeting these leaders with indiscriminate violence, however, may jeopardize resource access or trigger the backlash they wish to avoid, particularly when local elites are largely legitimate.
On the other hand, a group that believes these leaders are unlikely to supply intelligence to rivals and relies on them for resource access has less incentive to weaken them. It is likely to adopt a cooperative approach. Nevertheless, it continues to face the risk that elites will mobilize civilian resistance or otherwise undermine its rule. Rather than weakening local elites, however, such a group will focus on building loyalty, lessening the chance of elite-led mobilization by forging strong alliances. The next section considers how insurgent gender governance allows insurgents to achieve these goals.
The Challenge Approach and Insurgent Gender Governance
As we have seen, control over gender relations is the basis of local elites’ social, political, and economic power, but this authority frequently generates gender grievances, particularly among women and young men. I argue that rebels wishing to challenge local elites exploit these gender grievances to empower marginalized subsections of the population and undermine local elites, using this strategy as a means to establish social control (see Figure 1). I then illustrate this using the example of bride-price customs.
FIGURE 1. Mechanisms driving a challenge approach
Exploit gender grievances. Rebels will implement gender governance initiatives that explicitly address preexisting gender grievances. Concerns over rising bride-price costs, for example, may be addressed by banning or minimizing these payments, providing specific loans, or banning expensive wedding ceremonies.
Appeal to marginalized subsections of the population. These changes appeal to women and young men who are disadvantaged by the customary system. Changing local rules so that a woman receives her own bride-price payment, for example, gives her better economic leverage. And offering young unmarried men access to affordable marriages also gives them access to societal power and respect.Footnote 104
Erode the power of local elites. These changes also undermine local elites. Bride prices are a means through which wealth, land, and resources are concentrated,Footnote 105 and eliminating these customs cuts off a means through which local elites consolidate power. More generally, challenging gender customs erodes the authority of local elites, which is exerted through the regulation of gender norms.Footnote 106
Strengthen social control. As mentioned, groups forced to rely on local elites for resources face a dilemma because strong elites are more effective in resource accumulation but also pose risks in terms of civilian resistance. Slowly weakening untrustworthy local elites limits opportunities for elite-led mobilization. By exacerbating existing low-level tensions and altering power dynamics within communities and even families, insurgent gender governance undermines community cohesion.
This divide-and-conquer strategy weakens ethnic, religious, and familial solidarities which might otherwise facilitate civilian resistance.Footnote 107 It also boosts civilian loyalty among subsections of the population, who can be used to further strengthen the organization's intelligence and strength. Establishing new marriage practices, for example, can work to create dependence and loyalty to the rebel organization in ways which transcend existing kinship lineages.Footnote 108 Finally, asserting control over gendered institutions of justice, dispute resolution, and marriage also strengthens the group's ability to surveil the local population, identify informants, and extract fines or fees from the population.
The Cooperate Approach to Local Elites and Insurgent Gender Governance
On the other hand, a group that relies on local elites for resources and believes them unlikely to collaborate with rivals will adopt a cooperative approach. Nevertheless, these groups continue to face the risk that these elites will mobilize civilian resistance. As highlighted in the previous section, colonial and postcolonial states have used gender as a “bargaining chip” in relations with traditional elites, shoring up customary gender relations in an effort to maintain local elite control over rural areas.Footnote 109 I argue that rebel groups may act in a similar way, maintaining or reinforcing customary gender relations in an effort to stave off local elite resistance (Figure 2). These dynamics are then illustrated through the example of bride-price customs.
FIGURE 2. Mechanisms driving a cooperate approach
Reinforce the gender status quo. Much like the rebel groups just discussed, insurgencies using a cooperative approach will encounter widespread gender grievances. However, rather than challenging existing norms to appeal to marginalized subsections of the population, these groups will maintain or reinforce the gender status quo. Group members may, for example, marry into local elite families to consolidate alliances, pay high bride prices, or introduce laws that formally strengthen local elite control over kinship.
Appeal to local elites. These changes work to appeal to local elites. Inflated bride prices work to personally enrich these leaders while forming alliances with powerful local families. Strengthening elite control over gender relations also allows local elites to consolidate their power over key social institutions.
Strengthen social control. As mentioned, local elites pose a threat due to their ability to mobilize civilian resistance. Maintaining or reinforcing the gender status quo allows groups to placate these local elites, lessening the chance that they will mobilize the community against the rebels. Moreover, by strengthening the institutions that maintain local elite control over the civilian population, the group may strengthen the ability of these leaders to monitor the population and accumulate social and material resources. Since these groups believe these elites are unlikely to collaborate with the state or rival groups, the benefits of this strategy are likely to outweigh the costs.
Is Gender Unique?
I argue that gender offers specific opportunities for armed groups looking to undermine local elites. This is due to the role of gender norms as a source of local elite authority and to the widespread nature of associated gender grievances. This strategy may be particularly appealing in contexts where class consciousness is low,Footnote 110 with marginalized women and young men acting as a “surrogate proletariat” for the armed group.Footnote 111
This rationale was articulated by the EPLF. Lamenting the difficulty of mobilizing poor peasants, who often “suppress the deep hatred they harbored against their oppressors and chose not to stand or argue against the feudal chiefs,”Footnote 112 they decided to enact policies that appealed to young men and women, who they noted “constitute more than half of Eritrean society.”Footnote 113 This strategy, they argued, allowed the group to “identify and exploit” the differences between “feudal chiefs” and other classesFootnote 114 and limit the potential for an “alliance between poor peasants, nomads and serfs and the feudal chiefs.”Footnote 115
However, gender is not the only institution that is a source of both local elite authority and widespread grievance. In fact, ethnicity or caste distinctions can be conceptualized in a similar way. Like gender, ethnicity determines access to crucial resources. In much of Africa, for example, access to land is determined by ethnic “insider” status and subject to the authority of local elites, causing widespread grievances among ethnic “outsiders” at a local level.Footnote 116 One might expect groups who wish to challenge local elites to mobilize around the local grievances of ethnic “outsiders,” using governance to appeal specifically to these marginalized communities.
On the other hand, a gendered strategy offers some advantages over an ethnic strategy. Mobilizing around gender grievancesFootnote 117 may allow an armed group to access a greater proportion of the population. In contrast, mobilizing around ethnic outsiders may risk alienating ethnic “insiders,”Footnote 118 who may make up more of the population. This strategy may also encourage the “ethnic insiders” to rally around local elites, thus strengthening their authority rather than undermining it. While it is beyond the scope of this article to test these theoretical implications, this suggests my theory could be used to strengthen our understanding of how rebel groups govern in other domains and the conditions under which they mobilize around societal cleavages like caste, race, class, or ethnicity.
Hypotheses
This framework implies a series of hypotheses. First, I argue that armed groups exploit societal gender grievances to appeal directly to marginalized young men and women. However, while gender grievances exist in every known societal context, only armed groups that wish to challenge local elites will utilize this strategy. Groups that wish to align themselves with local leaders will maintain or reinforce the gender status quo.
H1 Rebel groups that challenge local elites are more likely to undermine local gender customs.
H2 Rebel groups that cooperate with local elites are more likely to reinforce or maintain local gender customs.
As we have seen, insurgent gender governance allows rebel groups to erode the powers of local elites without the destabilizing impact of indiscriminate violence. This strategy is likely to be particularly appealing to groups adopting a partial challenge rather than a pure challenge approach, as they have more to lose by adopting a more violent strategy.
H3 Rebel groups that adopt a partial challenge approach to local elites are more likely than any other groups to undermine local gender customs.
Empirical Approach
Data collection. There are currently no data sets providing cross-national information on armed groups and gender governance. I assembled an original data set of 137 African rebel groups fighting between 1950 and 2019. I collected data on nineteen separate strategies of insurgent gender governance (Figure 3). Every category has a binary indicator of presence or absence. I also collected data on each armed group's approach to local elites, assessing whether the group collaborates with, challenges, or partially challenges traditional and customary leaders in the territories it operates in. Details of the variables used in my models are provided later, while my data collection process is discussed in more detail in the online supplement.
FIGURE 3. Incidence of gender governance strategies
Scope conditions. My unit of analysis is the rebel group, and I limit the scope of my project to Africa due to the continued importance of traditional authority on this continent. However, I believe my theory would apply wherever local elites (including traditional, religious, or tribal authorities) exert authority in matters of local governance. Following Thomas and Bond,Footnote 119 my sample covers groups in nineteen countries, covering each major African subregion and including majority-Islamic as well as majority-Christian countries.Footnote 120 The sample is stratified to include countries with many conflicts and insurgencies, like Ethiopia, and those with fewer, such as Gambia. The sample used by Thomas and Bond was gathered from the Actor Dataset of the Uppsala Conflict Data Program,Footnote 121 the Global Terrorism Database,Footnote 122 and other sources and covered 1950 to 2011. Using the same sources, I extend this sample to 2019.
Research design. The models test the association between the approach to local elites and the use of gender governance, incorporating several insurgency-level variables as controls. All models are estimated using ordinary least squares (OLS) regression and include country-level and decade fixed effects. OLS was chosen rather than logit since linear regression allows direct interpretation of the coefficients as probabilities.Footnote 123 The tests I conduct do not aim to predict the probability of occurrence for the binary outcomes I examine. Instead, I estimate the effect of treatments on these outcomes. Given this, OLS is a suitable choice, allowing me to generate easily interpretable results with minimal assumptions.Footnote 124 I also ensure my results remain robust when using alternative modeling strategies, as demonstrated in the robustness checks and the online supplement.
Dependent variables: insurgent gender governance. I use various dependent variables from my own data set for this study. Change gender status quo is a binary indicator coded 1 if the group implements any governance to challenge existing gender norms, rules, or regulations.Footnote 125 Gender governance extent is an additive index using the nineteen separate strategies of insurgent gender governance—that is, the number of such strategies used by each group. See Figures 3 and 4 for details.
FIGURE 4. Gender governance extent versus frequency
Independent variables: approach to local elites. As mentioned, I test the relationship between the use of insurgent gender governance and the approach to local elites. The binaries pure challenge, partial challenge, cooperate, no contact, and challenge combinedFootnote 126 are the independent variables, with cooperate as the reference variable. This is because the no contact category (groups without documented interaction with local elites or where there are none to interact with) is difficult to interpret. On the one hand, my theoretical expectations are that groups that do not come into contact with local elites have no need to undermine them, and thus will change the gender status quo less than other groups. On the other hand, this category is more prone to a lack of information than other categories. By including cooperate as the reference category, I am therefore controlling for any potential bias introduced by the relative lack of information in the no contact category. The distribution of these variables is shown in Table 2.
TABLE 2. Approach to local elites
Alternative Explanations for Gender Governance Provision
Scholarship suggests that variation in gender governance will be driven by ideology. I draw on the Women in Armed Rebellion, Foundations of Rebel Group Emergence,Footnote 127 and Women's Activities in Armed Rebellion (WAAR) data sets for a leftist variable. I also use these data sets to record whether groups are motivated by fundamentalist Islamic ideals or theocracies based on sharia law (Islamist), as well as a variable capturing whether a group's main aim is self-determination or independence (secessionist).
Rebel strength is correlated with both rebel governance and group gender dynamics. I therefore include a series of variables that capture organizational strength and access to resources. In the Non-State Actors in Armed Conflict Dataset of Cunningham and colleagues, each insurgent group is coded as “much weaker,” “weaker,” “parity,” “stronger,” or “much stronger” in comparison to the incumbent government they are fighting (operationalized as an ordinal variable ranging from 0 to 4).Footnote 128 My binary variable rebel strength is set to 1 for “parity,” “stronger,” and “much stronger,” and 0 otherwise. As insurgent gender governance may be used to attract women or be shaped by women within the organization, I include a binary measure of female participation (female participants), using data from Thomas and Bond's data setFootnote 129 and the WAAR data set.Footnote 130
Territorial control and foreign state support also impact the provision of rebel governance.Footnote 131 I therefore include a territorial control variable, set to 1 if the group controlled any territory at all and 0 if it did not, and a foreign support variable, set to 1 if the group was allegedly or explicitly supported by a foreign state actor, and 0 otherwise. These variables are all drawn from the Non-State Actors data set.Footnote 132 As these data extend only to the year 2012, I hand-coded control variables for entries up to 2019 using a combination of UCDP data, the Big, Allied and DangerousFootnote 133 data set, and other secondary sources.
Results
The models support H1, H2, and H3, with a strong positive relationship between approach to local elites and change gender status quo (Figure 5). In other words, groups that undermine local elites in the territories they operate in are also more likely to challenge existing gender customs. This is demonstrated in model 1, an OLS regression which shows a strong, significant, and positive relationship between the provision of gender governance and a challenge combined approach to local elites, a relationship which remains statistically significant when accounting for a series of ideological control variables as well as female participation, group strength, territorial control, and foreign state sponsorship.
FIGURE 5. Determinants of change gender status quo
Model 2 demonstrates that this positive relationship is driven by groups using a partial challenge approach, with the coefficient for pure challenge positive but insignificant. This suggests, in line with H3, that changing the gender status quo strategy allows rebel groups to erode the powers of local elites without the destabilizing impact of indiscriminate violence. In contrast, groups adopting a pure challenge approach may prefer to overthrow local elites directly and be less incentivized to change the gender status quo. Among the model 1 and 2 covariates, no contact is negatively associated with changing the gender status quo. On the one hand, this is theoretically expected, as groups that do not interact with these leaders would have no incentive to use strategies designed to weaken them. On the other hand, as mentioned, this category is also disproportionately prone to bias created by an absence of information. This confirms my decision to use cooperate as the reference category to control for this potential source of bias.
Looking further at model 1 and 2, leftist, secessionist, and Islamist groups are no more likely to change the gender status quo overall. This suggests that broad-based ideological explanations are insufficient as an explanation for gender governance, although ideology does have more of an impact on gender governance extent, as I will discuss later. In contrast, female participants has a strong, positive relationship with gender governance, suggesting that gender governance is used partially as a tool to attract female recruits or be influenced by the presence of female participants. On the other hand, the mechanisms driving female recruitment may be similar to those shaping gender governance,Footnote 134 making the causal relationship difficult to assess. Likewise, foreign support has no impact, which suggests that insurgent gender governance is not driven by the desire to appeal to international audiences.
It is also notable that stronger groups seem less likely to provide gender governance, a relationship which reaches significance in both models. The relationship between territorial control and changing gender status quo is not significant. These findings run contrary to the theoretical expectations of much of the rebel governance literature and may support the idea that changing the gender status quo is a strategy used by weaker groups that are wary of competition from rival civilian leaders or civilian backlash. Strong groups with established territorial control may be confident in their ability to quash civilian backlash or overpower civilian rivals with their coercive forces and may be less likely to invest time and resources in this strategy.
The models depicted in Figure 6 provide additional support for all three hypotheses, with a strong positive relationship between approach to local elites and gender governance extent. This is demonstrated in model 3, an OLS regression which shows a significant and positive relationship between gender governance extent and a challenge combined approach to local elites, a relationship which remains statistically significant when accounting for a series of ideological control variables as well as female participation, group strength, territorial control, and foreign state sponsorship. Model 4 demonstrates that this relationship is again stronger among groups using a partial challenge approach, although both pure challenge and partial challenge reach significance in this model.
FIGURE 6. Determinants of insurgent gender governance
However, these models also demonstrate that both leftist and Islamist groups engage in more extensive insurgent gender governance. This suggests, in line with existing theories,Footnote 135 that ideology can be important in driving transformative and ambitious projects of rebel governance.Footnote 136 In contrast, secessionism has no impact on gender governance. This may be because secessionist governance is driven primarily by the desire to win recognition as a sovereign entity from both the local population and the international community. These groups may therefore prioritize governance initiatives that win them support from the whole population or demonstrate their capacity to govern effectively to an international audience, for example, public goods provision. Female participants again has a strong positive relationship with gender governance extent, while territorial control, foreign support, and rebel strength have no impact.
Robustness Checks
I address certain concerns with a series of robustness checks. As mentioned, I choose OLS regression for its ease of interpretation. However, others might contend that logistic regression is appropriate for the binary variables used in models 1 and 2. To ensure these results are not a product of the estimator used, I also run these models using logistic regression, and my results remain significant (supplemental Table C).
I also examine the relationship between a cooperate approach to local elites and a subcategory of groups that use gender governance, for example, enforcing the payment of high bride prices, to actively reinforce the gender status quo—a binary variable I call reinforce gender status quo. This model shows a positive relationship, providing more evidence for H2. Armed groups that cooperate with local elites reinforce gender relations as a tool to strengthen their alliances, lessening the chances of elite-led resistance.
This result also allays concerns that the results are driven by some untested dimension of rebel capacity—for example, if groups that are strong enough to intervene in gender governance are more likely to challenge local elites. Stronger groups are statistically less likely to change the gender status quo (Figure 5), and more likely to specifically intervene to reinforce the gender status quo (supplemental Table D). This suggests that many groups that have the capacity to provide gender governance and/or overturn local elites in fact choose to use that capacity specifically to appeal to these leaders and prop up existing gender relations.
Finally, I also check that my results remain robust when using alternative specifications of the independent variable—that is, by excluding groups that have no interaction with traditional leaders at all. Both challenge combined and partial challenge remain positive and statistically significant in these models, while cooperate shows a negative and significant relationship, suggesting that these results are not driven by a lack of information on one category of groups (supplemental Table E).
To conclude, this analysis provides persuasive evidence that the use of gender governance is strongly associated with the approach to local elites. These findings, which are robust to the inclusion of a range of controls and checks, challenge some of the accepted wisdom regarding gender and insurgency. As expected, the presence of female participants is positively associated with the use of gender governance. While ideological categories are not associated with the use of gender governance overall, both leftist and Islamist ideologies are associated with more extensive gender governance, while the impact of secessionist ideologies is limited. The negative relationship between group strength and gender governance is also surprising, suggesting that gender governance is a “weapon of the weak” for armed groups rather than a display of strength.
Overall, given the nature of my data, I cannot make confident causal claims based on my quantitative results alone. I therefore use mixed methods which enable me to trace the causal process in a particular time and place while representing broader patterns through statistical analysis.Footnote 137 Triangulating my findings using qualitative case studies allows me to draw stronger conclusions regarding my mechanisms, and to test competing theories.
Case Studies
This section explores insurgent gender governance through case studies of two Islamist groups operating in Mali. The groups operate in similar geographical, temporal, and societal contexts, supporting reasonable comparison. They share an Islamist ideology, and both recognize al-Qaeda as a moral authority.Footnote 138 And in 2017 they merged, forming the Jama'a Nusrat al-Islam wa al-Muslimin’ (JNIM), suggesting a shared vision and goals. Moreover, both groups provide civilian governance, suggesting that any differences in gender governance are not driven by differences in capacity to supply public goods.
Gender Status Quo in Mali
Mali is ranked 158th out of 162 countries in the UNDP's 2019 Gender Inequality Index. Violence against women is common, with 63 percent of men and 41 percent of women agreeing with the statement “there are times when a woman deserves to be beaten.”Footnote 139
Women are also frequently marginalized within the customary system. The vast majority of Malian chiefs say they never ask a woman's opinion when making decisions about the village.Footnote 140 In surveys throughout the country, about 60 percent of women say that women are always or often treated unequally by traditional leaders.Footnote 141 As one woman in Ménaka, northeastern Mali, reported: “The grievances we have are mainly the lightness with which women's rights are treated by these traditional authorities … Women are the losers in the majority of court decisions.”Footnote 142
Marriages in Mali are traditionally planned and managed by elders, who frequently “compel young people into unions that might reinforce their own networks or be materially advantageous.”Footnote 143 Elders in one rural community describe marriage as “too serious a matter to be founded on the caprice of youth … who often do not truly know what they want or what is good for them.”Footnote 144 The practice of bride price is almost universal.Footnote 145 This puts marriage out of reach for many young men; many stay single into their forties as they “struggle to save money for their wedding day.”Footnote 146 In rural areas bride prices have inflated to an average of a million CFA francs (around USD 1,500).Footnote 147 Polygamous marriages are also extremely widespread, a dynamic that further distorts the marriage market for young men.Footnote 148
However, despite the similarities between the two organizations and the context of widespread gender grievance, Katiba Macina engaged in an extensive overhaul of existing gender customs, while al-Qaeda in the Islamic Maghreb (AQIM) did not. The following case studies show how their approach to local elites shaped each organization's use of gender governance.
Al-Qaeda in the Islamic Maghreb: A Hands-Off Approach to Gender Governance
AQIM was established in 2007. The Islamist group originated in 1998, during the Algerian Civil War, as the GSPC (Salafist Group for Preaching and Combat).Footnote 149 As the war dissipated, it shifted its focus to Northern Mali,Footnote 150 changing its name to AQIM in 2007 after aligning with al-Qaeda.Footnote 151
Cooperative Approach to Local Elites, 1998–2012
Local elites as a source of resources
AQIM adopted a cooperative approach to local elites in the region, motivated by three factors. First, it facilitated access to material resources. Moving into the Sahara allowed AQIM to overcome organizational weakness,Footnote 152 accumulating considerable wealth through kidnapping for ransom,Footnote 153 which by 2014 had brought in USD 91.5 million.Footnote 154
However, converting financial resources into arms required infiltrating Northern Mali's smuggling networks, which were controlled by powerful local elites.Footnote 155 With “social and family connections” the “backbone of everything in the Sahara,”Footnote 156 cooperating with these elites was necessary.Footnote 157 This strategy is articulated in a 2009 letter to Osama bin Laden from AQIM's leader, Abdelmalek Droukdel.Footnote 158
The group used its kidnapping profits to attract local elites as business partners in their smuggling ventures. They offered these leaders a share of profitsFootnote 159 and gifts like four-wheel-drive cars.Footnote 160 Mokhtar Belmokhtar, one of AQIM's leaders, also financed construction projects for local Arab elites who facilitated access to smuggling networks.Footnote 161 The successes of this strategy are confirmed by Droukdel in another letter to bin Laden in 2010, describing how AQIM relies on the Malian tribes to access arms and how tribal leaders have provided access to “weapons and ammunition.”Footnote 162
Collaboration with local elites also helped AQIM access new recruits. Tuareg elite Iyad ag Ghali said that between 2003 and 2009 he went through “the stage of getting to know the mujahidin and beginning to make ties with them.”Footnote 163 This process led to his becoming a broker in networks connecting the Tuareg elite to AQIM.Footnote 164 Hamada Ag Hama, Ghali's nephew and fellow member of the elite Ifoghas Tuareg clan, was appointed leader of one of three new AQIM brigadesFootnote 165 designed to attract and accommodate a surge in new Tuareg recruits.Footnote 166 By 2007 a report prepared for al-Qaeda's senior leadership said that local recruits made up 95 percent of AQIM's Saharan-based brigades.Footnote 167 By 2010 the influx of Sahelian recruits drove the group to create three new units at the katiba (brigade) level.Footnote 168
Local elites as a source of security
AQIM's cooperative relationship was motivated by access to material and social resources. At the same time, the risk that these elites would collaborate with the state was low, with many local elites, including Iyad ag Ghali, involved in long-running conflicts with the central stateFootnote 169 and known for their roles in previous uprisings.Footnote 170
With the risk of disloyalty seemingly low, AQIM's cooperative strategy was explicitly designed to ameliorate the risk of backlash to rebel rule. This approach was articulated in correspondence between AQIM and al-Qaeda's central leadership during this time.Footnote 171 Following on from the downfall of al-Qaeda in Iraq, in which tribal leaders turned against the organization following violence toward them,Footnote 172 Bin Laden emphasized to AQIM that tribal leaders were an important constituency to court,Footnote 173 stating that “if the mujahideen deal well with the tribes, the tribes will for the most part side with them” and warning that “killing one member of a tribe is sufficient to provoke blood vengeance.”Footnote 174
AQIM Insurgent Gender Governance, 1998–2012
An insurgency using a cooperative approach has less incentive to weaken local elites, as it would jeopardize resource access and alienate potential allies. These groups will instead reinforce or maintain the gender status quo, using this as a “bargaining chip” to appeal to local elites and stave off elite-led resistance. AQIM provided extensive governance during this period. The rebels offered medicine, treated the sick, distributed money, and provided cell phone access for the local population.Footnote 175 They also financed development projects such as house construction and well digging and organized medical clinics that provided free medicine and examinations for people and animals.Footnote 176 In many ways “AQIM was acting as an Islamic charity, with the exception that its members carried arms and did not hesitate to use them.”Footnote 177
However, despite the existence of widespread gender grievances,Footnote 178 the group made no effort to challenge local gender customs. For example, there is no evidence they implemented gendered dress codes or limited freedom of movement during this period, nor did they create or regulate dispute resolution systems.Footnote 179 Instead, the group made use of existing marriage norms to successfully infiltrate local networks and consolidate bonds with local elites. Mokhtar Belmokhtar, one of AQIM's leaders, strategically married four women from local Tuareg and Arab Berabiche families,Footnote 180 relying on these alliances for “logistical support and protection.”Footnote 181 Nabil Abu Alqama, another AQIM leader, also married into a prominent Kounta Arab family at this time.Footnote 182 AQIM also reinforced bride-price customs, setting high prices to marry local women.Footnote 183
Overall, reinforcing the gender status quo fostered the group's relationships with key tribal leaders. For example, in 2006 relationships between AQIM and Tuareg tribesmen were occasionally hostile.Footnote 184 By 2010, however, AQIM leaders said that the group had forged “brotherly relationships” with these tribes and that this had contributed to their success in the region.Footnote 185
Background to the 2012 Crisis
AQIM's “hands-off” approach to gender governance was put to the test from January 2012, when it allied with fellow jihadist groups Ansar Dine and MUJAO (Mouvement pour l'unicité et le jihad en Afrique de l'Ouest) in launching a revolt. In April 2012, following a military coup in March, a Tuareg independence group, the MNLA (Mouvement national de libération de l'Azawad), proclaimed an independent state of Azawad. In June, however, the group was expelled after the collaborative relationship collapsed. This left the jihadist groups in control of almost all of northern Mali, with AQIM sharing authority with Ansar Dine in Timbuktu.
AQIM's Approach to Local Elites and Insurgent Gender Governance, 2012 Onward
AQIM's leaders continued to prioritize strengthening relationships with local elites during this time. Letters written by Droukdel say that the group has a “golden opportunity” to “extend bridges to the various sectors and parts of Azawad society … particularly the big tribes and … the elite of Azawad society, its clerics, its groupings, its individuals and its noble forces.” The aim of building these bridges, he argues, is “to make it so that our Mujahedeen are no longer isolated in society, and to integrate with the different factions, including the big tribes … and tribal chiefs.” He emphasizes that this should be AQIM's primary goal: “If we are only able to accomplish this limited, positive measure … it would be enough for us.”Footnote 186
In line with the expectations of this paper, cooperation with local elites is described as a security issue: “You are walking in a minefield full of tribalism, conspiracy, and revenge, corruption and arrogance. So you have to be diligent.” Droukdel quotes advice he received from bin Laden, saying that “states are not created from one night to the next and there needs to be a lot of elements for it to succeed. One of these very important elements is to take the allegiances of the tribes.”Footnote 187 Droukdel also noted in 2014 that the power of local elites can be used to the organization's advantage, as “tribal animosities,” when properly directed, are “greater than we can imagine.”Footnote 188
Droukdel emphasizes the need to incorporate elites from every large tribe into the government of the emirate. Tribalism is consistently prioritized, with Droukdel arguing in 2014 that “competency is important, but tribalism is also important and we absolutely have to take that into account, so let us … ensure that we include some of the tribal chiefs and, if they can be found, competent people, from every large tribe.”Footnote 189 Surprisingly, he stipulates that “loyalty to Islam and the Sharia” should not be a requirement for these positions.
In these letters, Droukdel also articulates a hands-off approach to insurgent gender governance. Being “aware that salafi-jihadist ideology in itself had little traction among the tribal chiefs,” he believed the radical implementation of Sharia gender governance would “push people away.” Radical gender policies, he explains to other jihadist leaders, are “contradictory to the policy,” meaning “your officials need to control themselves.”Footnote 190 The crucial mistakes the coalition has already made in this regard include the “extreme speed with which you applied Sharia.” He particularly criticizes “the application of the ḥudūd [religious punishment] in the case of adultery … the lashing of people and the use of force to try to stop things that are haram,” and “the fact that you prevented women from going out, and prevented children from playing.” Instead, the priority should be to “make sure to win allies … and compromise on some rights to achieve greater interests … Not every concession to the enemy is forbidden.” Moreover, “you should limit the circle of confrontation and of your enemies to the maximum” and instead “make sure to integrate everybody … notables, people with qualifications in every town.”Footnote 191
Not challenging the gender status quo is a clearly articulated strategy designed to avoid pushback from powerful local elites. Howoever, Droukdel's instructions were “blatantly ignored”Footnote 192 by the other elements of the jihadist coalition, with Ansar Dine, MUJAO, and the rebellious southern commanders pressing ahead with the implementation of extremist policies. Yet some have argued that AQIM's leadership succeeded in modifying the more extreme elements of gender governance in Timbuktu. For example, Droukdel insisted that the Islamic Court should base its judgments on the local Maliki school of law, with which people were already acquainted, rather than Salafist principles. And indeed, Islamic Courts under AQIM's jurisdiction refrained from imposing harsh punishments for crimes like adultery.Footnote 193
To sum up, AQIM prioritized cooperation with local elites, using them for resource access while believing them to be more useful as allies than as enemies. Despite its extensive rebel governance, the organization maintained the gender status quo, using marriage to forge powerful alliances with these leaders. While the group's leadership lost control over this approach when it entered into a jihadist coalition, Droukdel's personal correspondence demonstrates that this approach stemmed from a desire to build alliances with powerful local chiefs and a fear of elite-led backlash.
Katiba Macina: A Radical Approach to Gender Governance
Katiba Macina, also known as the Macina Liberation Front, emerged in central Mali in 2015 under preacher Hamadoun Kouffa. This Islamist group vowed to restore the theocratic Macina Empire, which once spanned central Mali.Footnote 194
Local elites as a security threat
As we have seen, AQIM believed cooperation with local elites would ameliorate the threat of backlash. Katiba Macina, meanwhile, saw local elites as a threat due to their potential collaboration with the state. Compared to those in the north, local elites in central Mali have greater representation in national politics,Footnote 195 with the jowros (chiefs) collaborating with the state at a local level.Footnote 196 Katiba Macina therefore adopted a “safer out than in” strategy, thinking that chiefs might “collaborate with state agents and give away sensitive information about the radical elements in their community.”Footnote 197
Violence toward these local elites was widespread during Katiba Macina's emergence (2015–2017); the group was responsible for killing more than thirty chiefs,Footnote 198 while many others fled after realizing the “jihadists were mainly attacking notables.”Footnote 199 The assassinations often followed accusations of collaboration with national authorities.Footnote 200
Local elites not required for resource accumulation
Katiba Macina also calculated that local elites were not needed for access to social and material resources. While information on Katiba Macina's financing is “rather opaque,”Footnote 201 it is reported that the group received funding and weapons from Ansar Dine in its early years,Footnote 202 suggesting that the group did not initially require local elites for access to material resources. As the organization grew, however, it increasingly relied on extraction from the civilian population.
Local elites in central Mali have historically held a crucial role in the management of crucial resources, with jowros demanding up to five million CFA (over USD 8,000) for access to pastures.Footnote 203 These fees were a source of local tension, with pastoralists entering the pastures illegally and jowros involving formal law enforcement.Footnote 204 Rather than co-opt local elites for resource access, as groups like al-Shabaab do,Footnote 205 Katiba Macina chose to challenge local elite control over access and establish its own direct taxes as a replacement. Kouffa criticized the jowros’ collusion with the state,Footnote 206 their privatizing of pastures, and their high access fees.Footnote 207 This system, Katiba Macina said, was unjust and irreligious, as “the bourgou [pasture] belongs to the good God, and so does the rain which makes it grow.”Footnote 208
Moreover, rather than relying on local elites for recruitment, Katiba Macina exploited pre-existing grievances with these elites among Fulani pastoralists. As one herder described, “The only feeling that animates us is that we can free ourselves from the yoke of the domination of our elites … This is why many of us are in the bush with weapons.”Footnote 209
However, local elites remained widely legitimate among the general population, with 63 percent of those surveyed by Afrobarometer in Mopti at the time of the group's emergence saying they trusted traditional leaders “a lot,” a figure in line with the national average at the time.Footnote 210 Notably, this level of trust is no lower when looking specifically at the Fulani in the Mopti region, where 61 percent again say they trust traditional leaders “a lot.”Footnote 211 With local elites unlikely to be loyal, this legitimacy among the general population was a risk for Katiba Macina. At the same time, it meant that indiscriminate violence came with a greater risk of backlash. The next section looks in detail at how Katiba Macina combined targeted violence with insurgent gender governance, exploiting pre-existing gender grievances to appeal to marginalized subsections of the population, erode local elite power, and establish social control among the civilian population.
Exploiting Gender Grievances to Appeal to the Marginalized
Marriage
In this region, as in the rest of Mali, tensions over marriage are widespread, with rising bride prices making the institution inaccessible.Footnote 212 In response, Katiba Macina challenged the authority of parents to arrange marriages and bolstered the right of young women to consent.Footnote 213 Other interventions included reducing bride prices to a “reasonable” sum,Footnote 214 banning expensive ceremonies, and providing resources to help young men marry and establish a home.Footnote 215 Kouffa also reversed prohibitions on inter-caste marriages.Footnote 216
By reducing the costs associated with marriage, the group appealed to women and to marginalized young men. Remarkably, in one survey 100 percent of women and 90 percent of men said the group has improved access to marriage.Footnote 217 Allowing inter-caste marriages also appealed to women from lower castes, who previously struggled to access marriage.Footnote 218
One man quoted by Guichaoua and Bouhel said, “The jihadists have helped to reduce the celibacy of women. Now everyone finds someone … They reduce all the expenses of the ceremonies that prevented young people from getting married, so they get married more easily.”Footnote 219
Others show similar findings, with focus groups in Mopti reporting that “high bride prices are now prohibited, they have to be reasonable”Footnote 220 and that “nowadays you're free to marry without money being demanded from you.”Footnote 221
Discrimination against women
Discrimination against women is prevalent throughout the region, with power described as “firmly in the hands of men.”Footnote 222 Women also face discrimination in terms of access to property and inheritance.Footnote 223
It is reported that Katiba Macina appointed their own judges and implemented mobile courts led by representatives on motorbikes.Footnote 224 These courts frequently resolved disputes related to divorce and other family law disagreementsFootnote 225 and developed a reputation for working more fairly than the customary system.Footnote 226 This alternative justice system likewise appealed to marginalized groups, with local women noting that “if a girl is forced into marriage, she may now appeal to the armed extremists to uphold her right to consent.”Footnote 227
Gender roles
Gendered labor divisions are strictly adhered to, reinforcing strict divisions between men and women as well as between castes. A man without work is described as “like a broken branch of a tree … not a real man, he's useless.”Footnote 228 Elite Fulani women, meanwhile, are meant to stay confined to the household,Footnote 229 while non-elite women perform labor dictated by their caste.Footnote 230 Jallube women (non-elite, pastoral herders), for example, engage in the processing and selling of milk. Ex-slave women (Riimaybe) sell goods in the market and pound millet.Footnote 231
Katiba Macina radically changed gender roles, forbidding women from traveling without a male family member, selling milk door to door,Footnote 232 or trading at the market.Footnote 233 They also enforced strict dress codes for both men and women and prohibited tattooing around the mouth of women after marriage.Footnote 234
Interestingly, many women welcomed these changes.Footnote 235 One said, “With Sharia law in force women won't have to work until they're exhausted. They will stay at home and have cherished children.” Another said, “It would be better for government to be based on religion. We would be protected and we'd stay quietly at home.” Overall, 100 percent of the women surveyed by International Alert agreed with the idea that “governance inspired by Islam would be better at meeting the needs of women.”Footnote 236
This can be understood within the context of Fulani society. Strict gender roles maintain hierarchical social divisions, with elite women restraining their own mobility.Footnote 237 By changing gender norms so that women are “protected” and “stay quietly at home” and banning specific forms of work associated with non-elite castes (such as selling milk and going to market), Katiba Macina may have appealed to non-elite women who wanted a high-status lifestyle.
Interviews conducted in Mopti also show many women approving of the new dress code.Footnote 238 This reaction can be interpreted in the same way, as wearing a full veil also serves as a marker of status, signaling the ability to avoid work.Footnote 239
Security
While it is likely that many women do not welcome the interventions, they may be seen as a worthwhile trade-off for security. One woman explained that jihadist groups were preferable to government soldiers and other armed groups because “the jihadists are responsible for less sexual abuse compared with the others … And any of their people who are found guilty of these kinds of acts are executed.”Footnote 240 In one survey, only 5 percent said that jihadist groups “commit acts of violence (rape or harassment) against women,” compared to 61 percent for ethnically based self-defense groups and 32 percent for armed state actors.Footnote 241 Another survey identified physical protection as a primary motivation for women to support jihadist groups.Footnote 242
Eroding the Power of Local Elites
Marriage
Katiba Macina's use of marriage governance reduced local elite power. Bride wealth is an “obligatory flat tax paid by young men to older men.”Footnote 243 By disrupting this system, Katiba Macina destabilized local elite authority and reallocated key resources. The group's support of inter-caste marriages also allowed the group to upend the customary hierarchies through which elites perpetuate dominance.Footnote 244
Discrimination against women
Allowing civilians to “forum shop” between customary and rebel institutions constitutes another blow to local elite authority. One local interviewee said, “When there is a dispute or disagreement, instead of going to the village chief's hall, the people appeal to the jihadists.”Footnote 245 Another said, “Whenever a local is unsatisfied with a decision made by a customary chief, he can call the local representative of the radical groups and inform him about it. In return, the customary chief would either receive a phone call and be asked to change the judgement, or would be kidnapped and killed in the bush.”Footnote 246
Gender roles
Katiba Macina also challenged local elites by overhauling gender roles. As mentioned, gender roles reinforce hierarchical social divisions. In banning non-elite women from engaging in the specific forms of labor associated with their castes, Katiba Macina's interventions can be seen as a challenge to the strict customary divisions of labor and the caste-based system which perpetuates local elite dominance.
The group also targeted other gendered markers of hierarchy and caste identity. Fulani tattooing practices, for example, function as “embodied indicators of clan and geographic affiliations,” to communicate information about clan affiliation and social status.Footnote 247 Styles of dress are used to identify particular Fulani communities, with practices like veiling used specifically to distinguish between high- and low-caste women.Footnote 248 By contesting the gender norms that communicate social hierarchy, Katiba Macina eliminated markers of caste distinction, undermining a key principle of local elite authority.
Security
Finally, traditional authorities are responsible for security provision, yet frequently fail when it comes to gender-based violence.Footnote 249 The ability of Katiba Macina to provide better protection from sexual violence highlights the incompetence of local elites in this area.
Strengthening Social Control
Taking control of the gendered institutions previously controlled by local elites allows the armed group to accumulate information on the local population. As one local said, “They know everyone by name, they are infiltrated everywhere.” Others note that informants “keep tabs on all those who visit the court.”Footnote 250
Challenges to gender roles also facilitate social control. By appealing to women and marginalized youth, the group accesses a network of local informants and strengthens surveillance. Young men are given “motorcycles and phones” and in return “spy on the communities and relay the information.”Footnote 251 While Katiba Macina does not recruit female combatants, civilian women are also used as informants or “dormant cells,” operating a system of surveillance on behalf of the organization.Footnote 252
Interestingly, these women are also given opportunities to circumvent some of the more rigid elements of the gender governance regime. One woman in Mopti, for example, recalled that the group allowed her to work as a trader on the condition that she did favors for them.Footnote 253 This flexibility suggests Katiba Macina's gender governance policies are not motivated purely by ideology.
In line with my expectations, undermining local elites enabled Katiba Macina to access population-level resources without facing backlash. For example, the group limited and replaced jowro taxes with its own zakat (Islamic tax), collected from owners of cattle herds at one-fortieth the value of the herd. Reportedly, most view these taxes as less burdensome than those paid to the jowros.Footnote 254
The group's ability to enforce these changes and establish effective social control is facilitated by the changes it has made to local power relations. As one elder interviewee said, “I saw a case in a village that affected me deeply, where the son called his father a heathen because he had not paid the zakat. It was the arrival of the jihadists that led to this insult. Children no longer respect their parents.”Footnote 255
While this elder sees these shifts as “negative changes in the population that should not be encouraged,” this anecdote shows that Katiba Macina have effectively exploited cleavages between generations, appealing to marginalized subsections of the population and bolstering their support for the project while eroding the power of existing authority figures.
Alternative Explanations
The cases we have just seen suggest that insurgencies use gender governance to reshape local power relations, with variation driven by a group's approach to local elites. These cases also show us how to address a range of alternative explanations. The first is that gender governance is used to forge a social contract with civilians. Given the prevalence of gender grievances, changing the gender status quo could theoretically appeal to any armed group looking to boost civilian support. The approach of AQIM, however, shows this approach can be seen as risky. Believing that sharia law was unpopular with tribal chiefs and that radical gender policies would “push people away,”Footnote 256 leader Droukdel prioritized the “allegiance of the tribes” instead.Footnote 257 Given that civilian preferences are frequently heterogeneous, this case shows that social contract theories are insufficient.
Another proposed explanation is that gender governance is a manifestation of ideology. These cases suggest this too is unlikely. As mentioned earlier, two groups share an ideology and goals.Footnote 258 They also joined forces in 2017 to form the Jama'a Nusrat al-Islam wa al-Muslimin’ (JNIM), suggesting ideological compatibility. Moreover, AQIM's leaders clearly justify their approach as strategic rather than ideological, while Katiba Macina demonstrates flexibility over time. This suggests that ideological explanations are likewise insufficient.
Another possibility is that gender governance is shaped by differences between the groups or their contexts. For example, AQIM was more transnationally oriented, while Katiba Macina originated locally, which may have shaped their approach to local elites. On the other hand, other transnational Islamist groups, such as al-Qaeda in Iraq, chose to adopt a “challenge” approach to local elites in Iraq.Footnote 259
In terms of local context, certain differences between central and northern Mali do play a role, as the integration of central Mali's elites into national politics leads Katiba Macina to see them as a greater security risk.Footnote 260 However, the two regions also have many similarities: both are shaped by the legacies of slavery, a rigid caste system, conflicts between pastoral and sedentary communities, and contestations over resource allocation and local elite authority.Footnote 261 As we have seen, both the legitimacy of local elites and the prevalence of gender grievances were similar across the two regions and in line with the national average.Footnote 262
It is also notable that other northern Islamist armed groups, such as MUJAO, adopted both a more confrontational approach to local elitesFootnote 263 and a more radical approach to gender governanceFootnote 264 despite emerging in the same area as AQIM. This suggests that AQIM might have adopted a different strategy if resource and security calculations had been different.
Also, it looks like Katiba Macina shifted its strategy after joining the JNIM. Though it was still based in central Mali, once in the coalition it strategically softened its approach to local elites and took a more “hands-off” approach to gender governance.Footnote 265 This suggests that explanations based purely on local context are likewise insufficient.
Overall, these cases demonstrate that each group's approach to local elites is shaped by strategic calculations rather than by the need to build a social contract, the group's ideology, or the local context.
Conclusion
This study delves into the conditions under which insurgents choose to challenge or maintain local gender norms and the benefits they derive from such actions. Analyzing an original data set of 137 armed groups, I argue that gender governance is a tool to reshape local power relations, providing evidence that armed groups challenge existing gender customs when they wish to undermine local elites. This result, which is robust to the inclusion of a range of controls and robustness checks, is strongest among groups adopting a partial challenge approach, suggesting that gender governance is an alternative to more violent approaches.
Using the case of Katiba Macina, I show that armed groups use gender governance to undermine these local rivals. By exploiting pre-existing gender grievances, they build support among people dissatisfied with the status quo, slowly eroding the power base of traditional leaders and limiting opportunities for coordinated resistance. The case of AQIM shows why armed groups may choose a contrasting approach. Cognizant of the risks of alienating powerful civilian leaders, they take a “hands-off” approach to gender governance, prioritizing the forging of alliances with local elites and the avoidance of division.
The insights provided here make a substantial contribution to the literature on gender, civil war, and state building. First, they improve our understanding of how gender inequality contributes to the outbreak of conflict. Feminist scholarship has long emphasized the linkages between patriarchy, militarization, and conflict, with recent scholarship confirming this relationship empirically.Footnote 266 This article contributes to this body of work by specifically highlighting the roles of both armed actors and local elites in this dynamic.
Second, this project deepens our understanding of wartime governance. Scholars have called taxation and security the “two most basic realms of rule.”Footnote 267 Yet recent cross-national work shows that only around a third of armed groups collect taxes or improve security.Footnote 268 That 40 percent of the groups in my sample implement gender governance shows that the regulation of gender customs is a third “basic realm of rule.” These findings also have implications beyond the study of conflict, contributing to our understanding of gender as a central societal cleavage in the process of state formation.Footnote 269
Future research might consider how both insurgent gender governance and approach to local elites impact other conflict outcomes, including the use of violence, rebel fragmentation, or civil war. Others might assess whether my findings generalize beyond the African continent or to other types of conflict. Our understanding of the “demand side”Footnote 270 of rebel governance would also be strengthened by using individual-level data to test the specific mechanisms I outline here. Overall, I hope these findings strengthen our understanding of the conditions that enable violence, how rebels interact with civilians, and the legacies of insurgency in the postconflict environment.
Data Availability Statement
Replication files for this article may be found at <https://doi.org/10.7910/DVN/G65R6V>.
Supplementary Material
Supplementary material for this article is available at <https://doi.org/10.1017/S0020818324000419>.
Open access
