Civilian confidence in domestic institutions, particularly in the security sector, is important for stability and state consolidation in post-conflict countries, where third-party peacekeepers have helped maintain peace and security after a conflict. Long-term peace and stability depend on a domestic security sector that is capable of maintaining its own security, as peacekeeping missions cannot provide stability indefinitely.Footnote 1 Citizens must trust the security sector to hold up its side of the bargain after the war, and after the peacekeeping mission leaves. Consequently, the security sector is often the first target for state-building reforms. Yet little is known about the types of reforms that are most successful in increasing citizens’ confidence in the security sector.
In a post-conflict context, reforms that meet certain criteria might be crucial for restoring confidence in the security sector. Yet Toft argues that strong security sectors are important for attenuating the credible commitment problem.Footnote 2 In addition to reforms that make the security sector more forceful, reforms that restrain the security sector and that are inclusive of a broader range of security needs, such as rape and domestic violence, may have an important effect in restoring the ‘trust gap’. The objective of this article is not to assess the public perceptions of reforms that increase strength versus ones that increase restraint or inclusiveness, but rather to demonstrate that reforms that meet alternative criteria – restraint and/or inclusiveness – are important for restoring public confidence in the security sector. The article does this by assessing how one reform, female ratio balancingFootnote 3 – or the increase in the ratio of women to men in various post-conflict institutions, including areas where women have been traditionally excluded such as politics and the security sector – affects trust in the state security forces.
Female ratio balancing reforms represent an observable shift in the composition of the forces from a pre-wartime and wartime security sector to a post-conflict one, and meet the criteria of restraint and inclusiveness because individuals often hold gender stereotypes about women, perceiving them as less abusive and more inclusive in addressing security needs. If women start to replace men in a security sector that is otherwise feared by the local population, this may help restore confidence precisely because women are perceived in these stereotypical ways. As a result, implementing female ratio balancing policies as a domestic security sector reform may help create the right conditions for a peacekeeping transition by increasing the demand for domestic security forces.
Post-conflict Liberia represents a ‘most-likely’ case to test these arguments. If female ratio balancing is to have any effect on civilian perceptions, it should in Liberia, because the country has prioritized female ratio balancing reforms in the security sector. Nevertheless, the research design tests the arguments in the communities that are least likely to have high levels of trust in the security sector: ex-combatant communities. In this way, the arguments are tested in a country where gender reforms are ubiquitous, but on a sample that is most likely to renege on peace. Survey evidence from ex-combatant communities finds that female ratio balancing reforms enhance demand for the domestic security sector, and that increased demand in the domestic security sector helps pave the way for peacekeeping transition.
The article makes two major contributions. First, the study contributes to the literature on post-conflict peace building – especially the literature on security sector reform (SSR), which finds that SSR helps prevent conflict recurrence, but fails to specify how and which types of reforms help reduce conflict, especially after peacekeeping missions leave. Until now there has been little theorizing about the types of reforms that might be better at restoring trust. This is one of the first articles to develop a conceptual framework to evaluate different forms of SSR and empirically test the theory using original survey data.
Second, the study makes an important contribution by demonstrating the importance of gender in the broader literature on state building and state consolidation. A vast literature has focused on women and peace building,Footnote 4 but these works have not touched on the effects of female ratio balancing on enhancing trust in security institutions. Additionally, anecdotal evidence has suggested that integrating women into the security forces improves the perceptions of security institutions, but there has been no systematic theory or empirical analysis of this potential relationship. Moreover, despite the fact that there have been several international mandates to include gender in SSR, little attention has been paid to how (or why) such reforms may be useful.Footnote 5 The study fills these gaps by testing whether female ratio balancing policies contribute to enhancing trust in domestic security sectors. It demonstrates that policies related to gender may contribute to improving confidence in domestic institutions, thereby exhibiting the importance of considering gender in theories related to post-conflict peace and international relations more broadly.
THE IMPORTANCE OF CIVILIAN CONFIDENCE IN DOMESTIC INSTITUTIONS
Civilian confidence in domestic institutions is necessary for long-term stability and state consolidation. Confidence is important for building the credibility of state institutions, since it allows a return to normal political life, in which citizens willingly concede to the state a monopoly over the use of force. As Weber and Tilly note,Footnote 6 state formation and consolidation are only possible when the state has a legitimate monopoly over the use of force, and when civilians consent to this monopoly. The second part of the statement – citizen consent – is important for legitimacy, and is perhaps only established when civilians have confidence in the domestic security sector. When citizens find the formal security sector to be credible, they are more likely to ‘defer to the institution’s decisions and rules and follow them voluntarily out of obligation rather than out of fear of punishment or anticipation of reward’;Footnote 7 they are also more likely to use state services over informal means,Footnote 8 and they may even provide intelligence or information to the security sector that makes it more effective in its duties.Footnote 9
Civilian confidence in the domestic security sector is also necessary for the successful withdrawal of third parties. Many post-conflict states receive third-party intervention, usually in the form of UN peacekeeping missions, which helps maintain peace and security.Footnote 10 A key part of the post-conflict bargain is that both sides trust each other to abide by the post-war agreement. The literature on the credible commitment problem has suggested that third-party enforcement, through peacekeepers, is one way to resolve the problem of mistrust.Footnote 11 Yet peacekeeping missions are unlikely to stay in a country indefinitely, which means that there must be some form of institutional design that enables both sides to trust each other after the third party leaves.
In this way, civilian confidence in the domestic security sector is particularly important for communities that may have trouble trusting that the state will abide by the agreement, such as those with high levels of ex-combatants. Without trust in the security sector, the negotiated settlement may come under threat. Those that do not trust the state and its security forces are less likely to find the state credibly committed to peace and may incite violence preventatively. Thus it is precisely those who are most likely to exacerbate the credible commitment problem to whom institutional reforms should be targeted.
CRITERIA FOR RESTORING CONFIDENCE IN THE DOMESTIC SECURITY SECTOR
Scholars and practitioners alike have suggested that SSR is important for restoring confidence in domestic institutions and preventing future conflicts,Footnote 12 but there may be several mechanisms for how confidence is achieved. For the purposes of this study, the security sector is defined as ‘organizations and entities that have the authority, capacity and/or orders to use force or the threat of force to protect the state and civilians’,Footnote 13 and a reform is defined broadly as an institutional change that occurs to the security sector that may ‘restore order and neutralize non-legal, non-statutory insurgents; rebuild the security forces such that these forces can take responsibility for the maintenance of public order; and build security related institutions that monitor and support the security forces’.Footnote 14
If the goal is to design security sector reforms in order to mitigate the credible commitment problem, one approach has been to implement reforms that increase the strength of the security sector. Toft has theorized on the relationship between security sector reform and peace duration, suggesting that without a strong security sector, resolving the commitment problem is not possible.Footnote 15 When the security sector is weak, the state lacks the capacity to punish ex-combatants for reneging on settlements. Thus a security sector with increased capability, which can use force credibly, may be necessary for peace.
While a strong security sector may be important for ensuring enforcement of the agreement by the other party, it does little to demonstrate to the other party that the state is respecting the agreement. Citizens, particularly ex-combatants, are unlikely to trust the state to abide by the rules, particularly those that relate to ceasing to abuse and exploit citizens. A deficit of trust in the police and military is common in post-conflict countries.Footnote 16 Scholars have found that in many cases, state security forces are culpable for violence against civilians;Footnote 17 state forces commit mass atrocities,Footnote 18 and government or government-sponsored security forces often perpetrate high levels of sexual violence.Footnote 19 Moreover, such a setting is often characterized by political and ethnic competition, corruption of the security services, excessive military spending, lack of professionalism and poor oversight,Footnote 20 which only contribute to more wariness of the domestic security sector. In this context, it is unlikely that reforms that strengthen the security sector will ameliorate fears; on the contrary, they may even exacerbate fears of the state. Comments from survivors demonstrate this sentiment: ‘how can we be expected to believe that men in the uniform of our new government will protect us when so many of them were themselves our brutalizers?’Footnote 21
Alternative – or at least complementary – security sector reforms should address the ‘trust gap’ by demonstrating to wary citizens that the state is no longer using its monopoly over the use of force to renege on the bargain or using force excessively against citizens for its own gain. Such reforms either (1) restrain or decrease state security forces’ ability to engage in unlawful behavior toward citizens and/or (2) are inclusive or increase the security force’s ability to address a wider variety of crimes that were previously ignored.
First, similar to the literature on government institutions, which holds that states implement institutions to limit governments’ political powers,Footnote 22 states may also implement reforms to restrain the security sector from using power to forcibly extract resources from the population. Such institutional restraints on the security sector may assure the public that the security forces are less likely to exploit and abuse the population in the post-conflict period. These reforms demonstrate that the state has minimized its power and no longer presents a credible violent threat to those that might oppose the state.
Secondly, reforms that are more inclusive of a wider range of protection needs may also help restore confidence because they address the needs of citizens who were harmed by the state during the war. Such reforms allow the state to redress its previous actions. Inclusive reforms enable the state to make amends and counteract the violence it inflicted during the war by providing protection to neglected populations. They also allow the state to cover a broader set of security needs. Intrastate conflicts require more than traditional civilian protection from armed violence; they also require protection that fulfills ‘human security’Footnote 23 needs or protection from chronic threats and sudden and hurtful disruptions in the patterns of daily life such as rape or domestic violence.Footnote 24 Much of the peace-building agenda has focused on addressing human security,Footnote 25 and has called for reforms to address human security needs.Footnote 26 The broader peace-building mandates include addressing rape, domestic violence and conflict resolution, among other issues, which requires a broader set of skills than just the provision of traditional security.Footnote 27 Thus reforms that are able to reach a broader range of people and that address a wider range of security needs may be particularly important for restoring trust.
The policy-making world has focused much of its attention on reforms that meet these requirements. In the early 1990s, SSR focused on transforming the structure of security bodies by making them transparent, professional and representative.Footnote 28 Mobekk states that, ‘the end goal of SSR is to effect shifts in mindsets and the transformation of a political system where the security sector is abusive, corrupt and politicized to one that is accountable, legitimate and transparent in the provision of internal and external security and the rule of law’.Footnote 29 Thus while reforms that make the security sector stronger may be correlated with longer peace duration,Footnote 30 they may not be the right reforms to restore trust in the state. Instead, reforms that meet the restraint and/or inclusiveness criterion may be more helpful in restoring confidence in the security sector, and help pave the way for a peacekeeping transition.Footnote 31
While restraining and inclusive reforms might help elicit trust among distrustful communities, their implementation must be visible to the public. Distrustful ex-combatants are unlikely to believe governments that merely declare they have implemented reforms, as governments may have an incentive to lie, especially if the reforms weaken their power. The only way states can demonstrate that they are not engaging in cheap talk is to ensure that reform implementation is observable. This means that citizens must interact with the newly reformed security sector. Karim finds that experiencing a reform leads to perceptional change as well as increased demand for the security sector, and that merely receiving information about reforms does little to change perceptions.Footnote 32 In this way, ‘contact’ with the reform or experiences with reform are crucial for the reforms to have any effect on improving perceptions of the security sector.
One of the most effective ways to demonstrate reform or ensure contact with reforms is if the reform focuses on changing the demographic characteristics of the security forces. Particularly in ethnically divided societies, or where violence between the state and civilians has been especially severe, the visible composition of the security forces ensures that minority groups feel safe to use the state’s security services.Footnote 33 If demographic reforms are implemented, then contact or experiences with new demographic groups may change perceptions, particularly if the reforms also fulfill the criteria of restraint and inclusiveness.Footnote 34
The goal here is to develop alternative and preliminary criteria for assessing the effect of reforms on trust. It is possible that when reforms meet one or both of the criteria, they mitigate the ‘trust gap’. Moreover, the task of this article is not to assess different reforms based on these criteria, but rather to examine the extent to which one reform that meets both criteria increases trust as a sort of proof of concept. Female ratio balancing meets the criteria and, like ethnic balancing, is a demographic reform, which means that when citizens interact with women, they are exposed to the reform. As such, it is an expedient reform through which to test whether either of the criteria described above helps increase confidence in the security sector. The rest of the article is devoted to assessing how well female ratio balancing reforms perform when they are implemented in ex-combatant communities.
THE RISE OF FEMALE RATIO BALANCING REFORMS GLOBALLY
The Convention on the Elimination of All Forms of Discrimination against Women established gender equality as an international principle, and scholars have established that reforms based on global principles such as human rights or gender equality can affect domestic policies and beliefs.Footnote 35 One such popular reform to promote gender equality globally is female ratio balancing, particularly in politics.Footnote 36 Female ratio balancing policies are enacted in order to increase the number of women relative to men in traditionally masculine institutions, so that women and men are equally represented in them; these policies are implemented through quotas or other measures that encourage women’s participation in the institution.Footnote 37 Since the passage of UN Security Council (UNSC) Resolution 1325 in 2000, the seminal resolution that institutionalized the international Women, Peace and Security (WPS) agenda, they have become increasingly popular as a post-conflict domestic security sector reform.
Emboldened by the WPS agenda, the UN has embarked on an ambitious agenda to female ratio balance the security sector. Nowhere has female ratio balancing been more salient than in UN policy toward post-conflict reconstruction.Footnote 38 Like with other international trends in peace building (such as, disarmament, demobilization and reintegration (DDR)), female ratio balancing is an increasingly popular reform advocated by third parties. Recent research finds that the presence of a peacekeeping mission strongly predicts the implementation of female ratio balancing policies,Footnote 39 and with the presence of a peacekeeping mission, female ratio balancing policies are more likely to be implemented than other policies such as DDR and civilian oversight.Footnote 40
The UNSC has endorsed female ratio balancing in a number of resolutions related to conflict, notably Resolution 1325 (2000), as a way to promote women’s participation and rights in post-conflict countries. As a result of UNSC 1325, the UN has sought to increase the number of women in peacekeeping missions.Footnote 41 Moreover, the international community has invested resources in female ratio balancing of domestic security forces as a form of SSR, and many states have institutionalized female ratio balancing through the adoption of UNSC 1325 National Action Plans or other policy changes. Figure 1 provides a map of female ratio balancing policies adopted by post-conflict countries during the period 1989–2012.Footnote 42 For example, the Sierra Leone police developed family support units and recruited women to fill these posts. The Policia Nacíonal de Timor-Leste enacted a 20 per cent quota, the Liberian National Police introdced a 30 per cent quota, and the Kosovo Police Service targeted women and ethnic minorities in recruitment.Footnote 43 These international and domestic efforts to female ratio balance represent the international community’s and states’ increased commitment to gender equality globally.
Fig. 1 Map of female ratio balancing policies in post-conflict states (1989–2012) Note: Figure 1 is taken from Huber and Karim Reference Huber and Karim2017.
MEETING THE CRITERIA: FEMALE RATIO BALANCING
Female ratio balancing demonstrates a visible demographic shift from the pre-war or wartime security forces, which most likely had very few women. With the exception of a few countries such as Israel and Eritrea, high numbers of women usually do not participate in state militaries and police forces.Footnote 44 Thus the mere presence of women in the domestic security sector is a visual representation of a change in the composition of security personnel from one comprised of men implicated in repressing the population to something new and different. Like with ethnic balancing, the presence of women signals reform implementation. When civilians have contact with female officers, they are exposed to the reform, which could elicit perceptional change.Footnote 45
Nevertheless, just because female ratio balancing is a visible reform does not mean that it will elicit positive perceptions. For example, it is possible that in countries where gender norms are rigid such as in Afghanistan, the visibility of women goes against society’s norms. Regrettably, a number of female police officers have been murdered in Afghanistan after the state started integrating women into the police force.Footnote 46 Thus female ratio balancing reforms may only have a positive effect in states that have moderate levels of norms of gender equality, particularly ones related to women’s visibility in public life. However, the introduction of women into traditionally male-dominated positions may start to encourage gender norms within society to be more ‘equal’.Footnote 47 Regardless of whether interactions with women elicit positive or negative reactions, the visibility of women in the security sector could somehow change perceptions of the police. It is possible that because female ratio balancing embodies restraint and inclusiveness, that these perceptions will enhance confidence in the security sector, not detract from it. Each criterion, as it relates to female ratio balancing, is explained and tested separately below.
Criterion 1: Female Ratio Balancing as a Restraining Reform
Female ratio balancing is a restraining reform due to gender stereotypes about women as less likely than men to perpetrate violence.Footnote 48 In general, though there are some exceptions,Footnote 49 female security personnel are not associated with prior repression of the population. Research indicates that female officers use less physical force, are less likely to use it to excess, and are better at defusing and de-escalating potentially violent confrontations with citizens.Footnote 50 Perceptions that women are less likely to use forceFootnote 51 may make them more trustworthy than men in the security sector.Footnote 52 Studies find that policemen are more likely to make arrests and use force excessively than policewomen.Footnote 53 Female police officers may not be perceived to use force when needed,Footnote 54 and others have found evidence that the inclusion of women in all levels of policing leads to the police force becoming less violent in general.Footnote 55 While most of these studies occurred in an American context, stereotypes about women as more restrained may be universal,Footnote 56 and they may be particularly salient for victims of state abuse. In many post-conflict countries, the justification for implementing female ratio balancing policies is that women are perceived as more restrained than their male counterparts.Footnote 57
In order to test this criterion directly, we can assess citizens’ views about security sector demand in a situation where excessive violence might be used. If women are perceived as more restrained than men when it comes to perpetrating violence, then individuals may prefer the domestic security sector to handle security and security threats involving the state-sanctioned use of violence – such as in an armed confrontation or riot – only when women are visible in the security sector. It is precisely in these circumstances that distrustful citizens might expect the security forces to renege on the peace deal and use excessive force. Put differently, ex-combatants are most likely to mistrust the government in emergency situations, because the state might use the opportunity to instigate unnecessary violence against them. However, a security sector that visibly employs women may be perceived as less prone to use excessive violence in such circumstances. Consequently, if distrustful individuals have had prior interactions with women in the security forces, they may be more likely to prefer state security services during security threats/emergencies because they perceive the security sector as less likely to use disproportionate force in these circumstances.
Hypothesis 1: When individuals have contact with women in the security sector, they will be more likely to prefer the domestic security sector to respond to armed violence and/or riots.
Criterion 2: Female Ratio Balancing as an Inclusive Reform
Inclusive reforms also have the potential to increase trust in the security sector. Again, gender stereotypes about women may lead citizens to perceive the security sector as more inclusive in addressing the needs of a wider range of individuals and security issues. Female ratio balancing may ensure that different skills, ideas and assets permeate the institution and give the perception that it is more inclusive in efforts to address human security.Footnote 58 Historically, the focus of many police institutions has not been representative of the needs of society, particularly women.Footnote 59 For example, women’s security needs – and more broadly, human security – was not a priority issue in the Liberian National Police before or during the war.Footnote 60 This means that there is a security gap in many post-conflict countries, which needs to be filled.
Female ratio balancing helps achieve this gap, because women are often perceived to be better at addressing previously ignored gendered issues such as rape and domestic violence.Footnote 61 Female police officers are more likely to prefer responding to domestic violence disputes than their male counterparts,Footnote 62 and female officers may be better at conflict resolution.Footnote 63 Women may also be perceived to be better at communicating with populations to address their needs.Footnote 64 These skills could also be important in a post-conflict context, where women may have experienced different forms of violence. The UN secretary-general’s report of 2008, which defines principles for UN SSR, asserted that incorporating a gender approach was important to ensure that security sector institutions are capable of effectively responding to the specific security needs of diverse groups. In this way, ‘women are assumed to bring a more gender-specific, value-added to broader security tasks’, which may help ‘foster civilian trust towards the security sector’.Footnote 65
In order to test this criterion directly, we can assess citizens’ views about security sector demand in the context of neglected security needs. If women are thought to be better than men at handling security issues that previously went unaddressed, such as rape and domestic violence, then individuals may prefer the domestic security sector to handle these issues when women are integrated into the sector. That is, with women in the security forces, citizens may perceive the security sector as better able and capable of handling rape and domestic violence, and this might increase demand for these services. Consequently, more citizens will trust the security sector.
Hypothesis 2: When individuals have contact with women in the security sector, they will be more likely to prefer the domestic security sector to respond to rape and/or domestic violence.
If female ratio balancing only meets the criteria because of gender stereotypes about women, then these gender stereotypes should be inelastic and universal in order to apply to different contexts. Though there is evidence that gender stereotypes are rigid,Footnote 66 they may not apply to all women in the same way because cultural sensitivity is not achieved solely through sex similarities. Interaction with just any woman may not lead to perceptions of restraint or inclusiveness. The category of ‘woman’ is not singular, and encompasses many different axes or dimensions that intersect with sex.Footnote 67 For example, a woman’s ethnicity, age, class and education level may help determine whether interactions with particular individuals facilitate trust. It is also possible that gender stereotypes do not apply, because women are sometimes perpetrators of violence.Footnote 68
Additionally, gender stereotypes may work in the opposite direction. If women are thought to be more restrained – and thus unwilling (or unable) to use force – individuals may perceive the security sector as less effective at providing protection. Protection from violence is largely considered a predominantly male task, and there is still the widely held view that women may not be able to protect civilians from violence.Footnote 69 Individuals may not perceive women to be as competent as their male counterparts, perhaps due to the perception of lowered standards for women’s integration. If individuals perceive the security sector negatively after having contact with women, this may support Toft’s argument,Footnote 70 as individuals may instead prefer reforms that are perceived to increase the capacity and strength of the security sector.
If gender stereotypes do not work in the assumed way, we should again expect to see null results for the hypotheses. But if gender stereotypes function in the expected direction, female ratio balancing meet the criteria of restraint and inclusiveness, then female ratio balancing reforms have the potential to restore some level of confidence in the state security forces.
CONFIDENCE IN THE SECURITY SECTOR AND PEACEKEEPING TRANSITIONS
One of the main concerns discussed above is that security sector reforms are needed for peacekeeping transitions. That is, to facilitate the successful withdrawal of peacekeeping missions, citizens must have confidence in the domestic security sector. Only then will they trust that the state will not renege on deals. Surprisingly, the literature on third parties does not address what happens after third parties leave, nor does it address the timing of peacekeeping transitions. Some have suggested that building effective and credible domestic institutions, which the public trust, is an important prerequisite for peacekeeping exit.Footnote 71 It is possible that third parties become obsolete when trust holds between the state and ex-combatants via institutions, but this relationship has not been empirically tested. Thus we should expect that as confidence in the security sector grows, demand for the peacekeeping mission declines.
Hypothesis 3: When individuals have confidence in the domestic security sector, they will be less likely to prefer an extension of the peacekeeping mission.
CASE STUDY: LIBERIA
Case selection is based on a ‘most likely’ design. If female ratio balancing is to affect perceptions, it is likely to have one in Liberia, because it is a post-conflict country with a large deficit of trust due to high levels of government-sponsored violence, and female ratio balancing in the security sector has been a priority. Liberia experienced intermittent civil war from 1989 to 2003 that claimed the lives of hundreds of thousands – mostly civilians – and led to a complete breakdown of law and order. By the end of the war, Liberia had fifteen different security agencies with overlapping functions that committed violence against civilians; many police stations had been abandoned, destroyed or taken over by rebel forces; the state lacked basic equipment, vehicles, fuel and communications systems; and many police officers and other government officials had fled the country.Footnote 72 Since 2003, it has had the presence of a multidimensional peacekeeping mission – the UN Mission in Liberia (UNMIL).
UNMIL has been instrumental in rebuilding the Liberian National Police (LNP), and has helped institute numerous reforms, especially related to gender.Footnote 73 In 2005, one of the first tasks of UNMIL was to develop a Gender Policy for the LNP, which helped UNMIL and the LNP introduce a 15 per cent quota for women.Footnote 74 In 2012 the quota increased to 30 per cent women. By 2015, the LNP had 19 per cent female officers, up from 2 per cent in 2005.Footnote 75 While Liberia is not the only country with quotas for female security personnel, it has received international attention for these reforms.Footnote 76
Although Liberia fits the ‘most likely’ case selection, other countries do as well – Liberia is not an outlier case. Other post-conflict countries such as Haiti, DRC, Mali and others also have high levels of civilian mistrust in the state, have received large-scale, multidimensional peacekeeping operations, and have instituted gender reforms domestically. Most post-conflict countries’ security sectors also did not include women before and during wars. Moreover, it may seem that Liberia is an outlier case given women’s organizations roles in bringing peace to the country or due to the all-female police unit from India in the UNMIL mission, but women’s organizing is common in almost all civil wars,Footnote 77 and all-female units have been present in other missions. Understanding the Liberian case may thus yield important insights into these other post-conflict countries.
RESEARCH DESIGN
In order to test whether female ratio balancing affects perceptions about domestic security sector institutions, an original representative survey was conducted in Monrovia’s two main ex-combatant communities, West Point and Peace Island. The scope conditions for the article require that the hypotheses be tested in ex-combatant communities, among those who are least likely to trust the state. Nevertheless, to check the internal and external validity of the arguments, the observable implications were tested through experimental work by Karim and a general survey conducted by the Norwegian Refugee Council in four Liberian rural counties: Nimba, Grand Gedeh, Maryland and River Gee in 2014.Footnote 78
Sampling
The representative survey includes 1,280 randomly chosen respondents. The survey was conducted in 2015 among adults aged 18 and over. None refused to participate in the survey.Footnote 79 In total, 638 from Peace Island and 642 from West Point were interviewed.
Variables and Models
To operationalize female ratio balancing, the independent variable is contact with a female LNP officer. The survey questions asked about contact with the security sector, including with the LNP, and asked whether this contact was with women. Three variables were constructed using the survey questions: contact with females (41 per cent), contact with males (27 per cent) and no contact with police (31 per cent). In the models, contact with males is the reference category, and the model controls for no contact with police.
Contact with female officers means that the individual had a conversation with the officer in the past year after taking a complaint to the police, and does not necessarily mean that the officer took the case.Footnote 80 Based on multiple days and hours of observation at police stations in these two communities, contact usually occurs at a police station when citizens report a crime, and interactions with a police officer usually last 10–20 minutes when reporting a case to the police. Contact is best operationalized as a conversation with a policewoman than women taking the case (which may signify competency or performance), because it captures the idea that any short interaction with them should lead to changed perceptions and is a more rigorous test of the argument.
There are several reasons why contact with female LNP officers may be random. First, both West Point and Peace Island are close to police stations, which means that individuals are likely to seek attention from the most accessible officer.Footnote 81 It is highly unlikely that individuals will seek out policewomen,Footnote 82 as respondents from informal interviews in the two communities said they do not seek out female police officers but rather approach the first person they encounter when they experience a dispute.Footnote 83 Furthermore, the effort and resources needed to travel to a police station overshadow concerns about finding the ‘right’ officer. Secondly, almost every unit in the LNP includes both sexes, and the units serving the depots and police stations always include officers of both sexes. At each depot for the communities in the survey, the sex composition was 22–30 per cent women. Contact with police may also occur because police respond to disputes. The sex of officers that respond may be considered random, because it depends on the availability of officers at the depot. In Liberia, it is possible that only the most dedicated officers will respond, but there is no reason to believe that women are more dedicated than men.
Nevertheless, individuals who interact with female officers may be fundamentally different from those who interact with male officers. To alleviate these concerns, a balance test was conducted of individual characteristics associated with those who had contact with female LNP officers, and those who had contact with male officers (see Appendix). The two groups match on many characteristics. However, there are several areas where the two groups differ. Those who have contact with female officers are more likely to be male,Footnote 84 are slightly older, tend to live in West Point,Footnote 85 tend to be Muslim, are members of a Poro and Sande secret society,Footnote 86 are more likely to pay a bribe to the policeFootnote 87 and are more likely have contact with female peacekeepers.Footnote 88 To account for the imbalance between the two groups, the models include sex, age, a dummy for West Point, a dummy for paying a bribe to the police, a dummy for being Muslim, a dummy for being a member of a secret and a dummy for having contact with female peacekeepers.
The dependent variable measures confidence in the security sector. While there are many ways to measure confidence, such as asking about perceptions of corruption or discrimination, the study uses survey questions that asked people which security institution they preferred for responding to armed violence and riots, rape and domestic abuse.Footnote 89 This measures people’s willingness to turn to the formal security sector when they have a problem, instead of informal community institutions such as traditional leaders and instead of the peacekeeping mission,Footnote 90 and is a better measure than asking about perceptions of corruption or discrimination, because it better captures the idea of trust and confidence in institutions: individuals would not state that they want the formal security sector to respond to problems if they did not have trust or confidence in it. That is, increased demand for the state’s security sector signals confidence in the institution. Additionally, the questions used in the survey mitigate the likelihood of social desirability bias.Footnote 91
Finally, when individuals have confidence in their domestic security sectors, they may be less likely to prefer the peacekeeping mission to provide security (Hypothesis 3). Thus the survey included questions about whether individuals preferred UNMIL to extend the mission. If they did want UNMIL to extend, it indicates that they still feel insecure about the domestic security sector handling the security situation in Liberia.
Some models include contact with men and no contact in order to assess how other non-gender-related reforms affect perceptions. LNP officers have undergone professional training through an academy, and they have created specialized units, among other reforms.Footnote 92 While it is not possible to test the effects of these individual reforms, assessing contact with male officers gives a sense of exposure to other reforms besides female ratio balancing.
The models include several control variables.Footnote 93 Controls were included based on the measure on which the groups did not balance. Additionally, even though the two groups were balanced, if an individual has been a victim of a crime it may affect their perception of the police and whether or not they seek an officer, thus previous victimization is included. In the models with a dependent variable about extending UNMIL, a variable about perceptions of UNMIL is also included.Footnote 94
A logit model is used to test the hypotheses. Different model specifications were tested by systematically removing each control variable. The model including all control variables is displayed below. The substantive results are presented using simulation to generate 95 per cent confidence intervals.
RESULTS
According to Hypothesis 1, we should observe that contact with female LNP officers leads to demand for the domestic security sector to respond to armed violence and/or riots (or other traditional security situations), and based on Hypothesis 2, we should observe that contact with female LNP leads to demand for the police to respond to rape and domestic violence (beating). There is support for both of these hypotheses. Contact with female police officers increases demand for the police to respond to armed violence, riots, rape and domestic violence. Table 1 shows the relationship between contact with women and the different security issues,Footnote 95 and Table 2 provides the predicted probabilities of preferring the police to respond to security concerns based on whether they had contact with a male or female police officer. Table 2 shows that there is a perceptional change depending on whether individuals interacted with a male or female officer, albeit in some cases the effect is small.
Table 1 Coefficient Table for Contact with Female Police and Preferences about Security Responses
Note: *p < 0.1; **p < 0.05; ***p < 0.01
Table 2 Predicted Probability of Contact with Female versus Male Police on Preferences for Police Response
Note: 95 per cent confidence intervals in brackets.
Taking a closer look at the data, many of the effects are driven by women in the sample. Particularly with regards to preferences for police response to rape and domestic violence (Hypothesis 2), women are much more likely to prefer the police if they have had contact with female police officers, whereas in some cases males were less likely to prefer the police to respond to rape if they had contact with female officers (see Appendix). Hypothesis 1 is driven by both men and women (coefficients are positive), though the results are statistically significant mostly for women. This suggests that restraint and inclusiveness are key components of successful SSR for women. Female ratio balancing thus increases overall confidence in the security sector by encouraging more women to trust the security forces.
The results are not only robust to different model specifications, but also hold in other analyses using experimental methods (demonstrating internal and external validity) and in other surveys (demonstrating external validity). Karim conducted a field experiment in rural Liberia in which household visits by the police were randomized based on officers’ sex and found that households that experienced visits by female police were more approachable and also had a higher demand for police services, particularly for addressing rape,Footnote 96 indicating support for Hypotheses 1 and 2 in this article. Moreover, in another survey conducted in four rural counties in Liberia, compared to contact with males, those living in rural countries who had contact with female officers were more likely to prefer the police to respond to a community dispute (hala hala), indicating support for Hypothesis 1.Footnote 97
Evidence in support of Hypotheses 1 and 2 is also strong when using alternative measures of restraint and inclusiveness.Footnote 98 When citizens had contact with female rather than male police officers, they were more likely to feel safer with the police in their community (restraint),Footnote 99 and to perceive the LNP as more respectful of the traditions and beliefs of the people who live in the community (inclusive). Given that multiple surveys and an experiment have found consistent evidence, there is robust internal and external validity for the causal claim that contact with policewomen enhances demand for police services in Liberia.
In total, 68 per cent of the population of West Point and Peace Island wanted UNMIL to extend their mission. Hypothesis 3 suggests that the higher the confidence in the domestic security sector, the less likely citizens are to prefer the peacekeeping mission to extend its presence. There is consistent support for this hypothesis. Table 3 shows the relationships. Substantively, holding all other variables constant, demand for the police to provide security and respond to armed violence, riots, rapes and domestic beatings leads to a 5 per cent, 4 per cent, 9 per cent, 5 per cent and 10 per cent decline, respectively, in preferring the peacekeeping missions to extend. While the substantive results are small, there is at least some indication that domestic legitimacy in security institutions could help obviate the need for peacekeeping missions. And, in order to achieve domestic legitimacy, states may do well to implement reforms that meet the criteria of restraint and inclusiveness.
Table 3 Prefer UNMIL to Extend
Note: *p < 0.1; **p < 0.05; ***p < 0.01
DISCUSSION AND CONCLUSION
The results indicate a proof of concept: at least one post-conflict domestic security sector reform that meets the criteria of restraint and inclusiveness – female ratio balancing – helps to improve confidence in the domestic security sector. Testing the criteria in ex-combatant communities, female ratio balancing appears to increase demand for the security sector during a security threat, indicating support for the restraint criteria; it also appears to increase demand for the security sector to protect citizens from crimes that were previously neglected. While the implementation of female ratio balancing cannot restore trust in the state on its own, a combination of reforms that meet these criteria might go a long way towards mitigating the commitment problem.
When it comes to the criteria of restraint, the evidence shows that instead of gender stereotypes working in a negative way – women perceived as less effective in providing security – they might be working in favor of the security sector. The integration of women may represent a more restrained security sector, thereby allowing a population that feared the security sector to be more accepting of it. If the survey results had indicated that individuals do not prefer domestic institutions to respond to traditionally male-gendered tasks such as providing security or responding to armed violence when they have contact with women, this may have indicated that they prefer a ‘stronger’ security sector. In other words, this would have indicated support for Toft’s criteria of security sector strength.Footnote 100 If individuals perceived women to be ineffective, we would have observed a negative relationship between contact with women and security provision by the institution. Instead, contact with women led to preferences for the domestic security sector to respond to traditionally masculine tasks, and made community members feel safer, which provides some evidence that reforms designed to improve trust may not induce negative gender stereotypes about women and may enhance confidence in the security sector.
Female ratio balancing reforms also give the perception that a wider variety of security concerns will be met. Individuals, particularly women, preferred the police to respond to rape and domestic violence when they had contact with female officers, and they also perceived the police force as more respectful and fair.
Importantly, these changes in preferences for the police mostly came from women, which suggests that, like in politics, substantive representation is important for successful security sector reform.Footnote 101 In this case, the sex of the officer represents restraint, as well as the police’s ‘newfound’ ability to respond to gendered crimes that affect women disproportionately. Implementing the reform thus increases trust among women, who were neglected before and during the war. The finding could also have broader, indirect ramifications for trust in the state. Reforms that are restraining and inclusive might help change the perceptions of female ex-combatants, and help women who are related to male ex-combatants (sisters, wives and mothers) or live in communities with male ex-combatants feel more confident in the state, which means that the pool of potential recruits and support system for potential male combatants might decrease. Extensive research has pointed to the important role women play in insurgent groups.Footnote 102 If the state can ‘win over’ women, it raises the costs of initiating conflict in the first place, because insurgent groups have to work harder to convince women to support their cause.
Evidence from the survey also found that individual contact with men often led to negative perceptions about the institution, whereas contact with women did not. Interacting with male officers exposed individuals to a ‘new’ police force, as the Government of Liberia created a professionalized police force in 2005. However, exposure to other types of reforms, which perhaps did not meet the criteria, may not be enough to bolster confidence. Instead, governments might do well to prioritize female ratio balancing, or at least implement it in conjunction with other reforms.
There is also some evidence that confidence in the domestic security sector helps ensure that peacekeeping transitions are successful. When survey respondents expressed confidence in the domestic security sector, they were less likely to prefer an extension of the peacekeeping mission. Confidence in the domestic security sector may be an important pre-requisite for peacekeeping transition. Peacekeepers do not, however, take public opinion into consideration when drawing down, but this research suggests that public opinion may be an important gauge of the progress of the transition as peacekeeping mandates come up for renewal.
The urban survey results were corroborated by a rural survey and experiment. That contact with female LNP officers positively affected perceptions of the domestic security sector in non-ex-combatant rural populations suggests that these results may also hold in communities that are more predisposed to trust the security services. That the results hold in an experimental context suggests that the causal mechanisms are internally valid.Footnote 103
Female ratio balancing is just one reform of many that may help establish the necessary conditions for restoring trust. The study has found evidence of the benefits of restraining and inclusive reforms, but future work could test the effects of other security sector reforms on civilian perceptions, and test which criteria are necessary and sufficient for enhanced perceptions. For example, while civilian oversight reforms affect restraint but not necessarily inclusion, they may still enhance confidence. Thus more work is needed to identify which of the criteria studied here are most important for generating confidence.
These findings may only generalize to post-conflict countries where violence has ceased. Context may be important for understanding which mechanisms are more important to citizens. If a conflict is ongoing, or if there are still high levels of violence from insurgents, Toft’s suggestions about security sector strength may be more important than restraint and inclusiveness.Footnote 104 Future studies could unpack the role of country context and SSR.
Finally, the study demonstrates the importance of considering gender in theories related to post-conflict peace building and international relations more broadly. Without considering how gender affects state- and peace-building activities, scholars and practitioners alike may miss out on potential new avenues to help create long-term stability in post-conflict states.
