Commitment to the “National” in Post-Conflict Countries: Public and Private Security Provision in Lebanon

Abstract

A core function of contemporary states is to ensure the security of their citizens. Yet in many post-conflict settings, non-state actors provide security alongside the state, typically prioritizing their own ascriptive groups and potentially undercutting a sense of national political community. When do citizens prefer group-specific versus national security? While most studies focus on individual psychological factors, we argue that group-level characteristics also shape political preferences. Based on a conjoint experiment in Lebanon, we explore the relative appeal of group-specific versus national pledges to assure protection. We find that respondents view national security provision quite positively, while members of communities with stronger group-specific security simultaneously favor private provision. Individuals with closer ties to credible group security providers are also more likely to prefer those services. Citizens therefore do not see a clear trade-off between private and public protection, while group-specific legacies mediate heterogeneity in support for pluralist security provision.

Keywords

ethnic politics political violence security political parties voter behavior conjoint analysis public opinion Middle East Lebanon

Introduction

A first order responsibility of the state—perhaps the defining characteristic of the state as the monopolizer of violence within its territorial borders (Weber 1965, see also Acemoglu, Robinson, and Santos 2013) —is assuring the protection of its citizens. Yet during civil conflict or state collapse, the state may cease or curtail security provision, while non-state actors fill this void. Militias and rebel organizations, among others, are key suppliers of security during and after conflict. They often provide these services in an exclusionary manner along the lines of territorial control, co-ethnicity, political support, or some combination therein. In some states, non-state security provision extends well into the post-conflict era, even as state capacity re-emerges. Non-state security provision may also reflect and perpetuate citizen mistrust of the state as a source of protection. Under these conditions, what do citizens demand from states and other actors with respect to security?

In this paper, we develop and test a theory of citizen demand for national security in post-conflict settings. We contend that micro-level security preferences are conditioned by the structure of group representation and the nature of group-specific security provision.¹ In some post-conflict settings, national security institutions remain weak, allowing non-state actors to capitalize on this vacuum by offering protection as a way to consolidate control and further their interests. Yet not all non-state actors are equally effective at providing security to their constituents. Some may have highly organized militias or other protection rackets, while others feature weak or nonexistent security forces.

In line with existing research, varied individual-level experiences and attributes shape the demand for security, but we contend that these are mediated by group-specific variation in the nature of non-state governance. In short, both meso- and micro-level factors shape preferences for national security, calling for a synthetic framework to explain public preferences for distinct types of security provision.

We investigate preferences for security provision—national and group-specific—in a hybrid security environment where both forms exist. Using a conjoint experiment, we randomly expose respondents to statements by hypothetical Lebanese parliamentary candidates, who pledge to ensure either the security of the nation as whole or of their own communities. The results of the experiment indicate that preferences for group-specific and national protection are not mutually exclusive, but rather depend on the credibility of different security providers. In particular, we present two key findings. First, people favor politicians who frame security as a national good when national security provision is credible, even if a group-specific alternative exists. Second, people also favor politicians who frame security as a group-specific good when the salient group provider can credibly promise such services. Our framework therefore links variation in contextual, group-level variation to support for both national and group-specific security.

Lebanon is an appropriate site for this research because it features well-respected national security institutions, a history of civil conflict, and variation in the strength of group-specific security providers. Our analysis compares preferences for politicians between and within the country’s Sunni Muslim and Shia Muslim populations, two dominant identity groups in the country. We compare security preferences across these two groups because they are characterized by political representatives with different track records of group-specific security provision. Whereas dominant Lebanese Shia party representation has engaged in group-specific security provision since the country’s civil war (1975–1990), the dominant Sunni party only coalesced after the war and never developed comparable capacity for communal protection.

We preview our findings here. First, we find that both Sunni and Shia respondents prefer candidates who frame security as a national good. Second, we show that Shia—but not Sunni—respondents also favor group-specific security, in keeping with the stronger track record of group-specific provision by Shia political parties. Finally, we demonstrate that Shia respondents with stronger party ties express greater support for such group-specific provision. Descriptive evidence highlighting variation in party strength vis-à-vis security provision between these two respondent pools corroborates this interpretation.

Our work contributes to a growing literature on preferences for security provision in post-conflict societies (Hazbun 2016). It also speaks to larger debates about how state capacity is rebuilt (or not) after civil war (Hartzell and Hoddie 2003; Barma 2017). Specifically, the paper complicates assertions that preferences for national and group-specific services are mutually exclusive. Existing literature discusses how group-specific security may undercut state capacity by impeding its monopoly of force over territory (Meagher 2012). Yet our findings show that from the demand side, constituents do not perceive a trade-off between national and group security provision, and that the presence of a group-specific option does not bias constituents against that service being provided by the state. This has important implications for scholarship on security provision and post-conflict policy-making.

Demand for Security in Post-Conflict Settings

In post-conflict settings where group identities have been politicized and intergroup relations are strained, is it possible to (re)constitute support for national security institutions and providers? When do citizens favor national security in contexts where other actors offer group-specific security?

Answering these questions, we argue, requires attention to both micro- and meso-level factors. The micro-level, where much existing research focuses, relates to factors like exposure to violence, partisanship, ideological orientation, access to information, and even individual psychological characteristics. On the meso-level, the terrain of national and group-specific security provision sets the menu of realistic options. In this section, we elaborate on these points to provide a theory of post-conflict preferences for security provision.

Micro-Level: The Impact of Threats on Individual Political Attitudes and Behavior

Perceptions of threat and elite activation therein have clear political ramifications. Based largely on research in Western countries, a broad body of literature traces the impact of racial or ethnic appeals, including those cueing insecurity in the face of out-group threats. While most studies show that identity-based threats create or exacerbate inter-group tensions to the benefit of fear-mongering politicians, emerging research in developing countries finds that such appeals can backfire.

Social science research highlights the political power of emotions (Petersen 2002; Pearlman 2013), which are often experienced collectively (Tajfel and Turner 1979; Brewer 1999; Bar-Tal et al. 2007, inter alia). Increased risk perceptions arising from fear and insecurity (Lerner and Keltner 2001; Huddy et al. 2005; Nacos et al. 2011; Young 2019) have been shown to increase attachment to group identity and in-group solidarity, and drive a corresponding desire for in-group protection (Huddy et al. 2005). When people feel targeted as members of a group, they are more likely to develop negative affect towards out-groups (Brewer 1999), resulting in exclusionary attitudes and support for politicians who position themselves as defenders of the community (Huddy et al. 2005; Canetti-Nisim et al. 2009). Exposure to violence (Fearon and Laitin 2000; Das 2006; Bulutgil 2015; Lupu and Peisakhin 2017) can boost demand for group-specific security provision, heighten discriminatory attitudes towards out-group members, and increase levels of political participation.

Because of this, during elections, in-group threats can benefit candidates who highlight and offer protection from them (Cho, Gimpel, and Wu 2006), just as authoritarian rulers or “strongmen” who promise protection and promote intolerance towards out-group members gain support by boosting threat perceptions (Davis and Silver 2004; Nacos, Bloch-Elkon, and Shapiro 2011, inter alia). These findings are echoed in diverse contexts, including the U.S. (Mendelberg 2001; Valentino, Hutchings, and White 2002; Davis and Silver 2004), where white conservative voters recognize explicit racial appeals and often approve of them (Valentino, Neuner, and Vandenbroek 2017, 758). In non-Western settings, studies show that “cultural entrepreneurs” cue caste, ethnic, or religious identities in an effort to drive voter turnout and participation in demonstrations or riots (Young 1979; Brass 1997; Fearon and Laitin 2000).

Research on the impact of threat-based rhetoric on individual attitudes and behavior in developing countries is more nascent. Some studies indicate that ethnic appeals, inflammatory rhetoric, and even electoral violence can motivate citizens to back co-ethnic politicians. Based on a case study of a district in Sri Lanka that experienced electoral violence, H⊙glund and Piyarathne (2009) note that politicians capitalize on an atmosphere of insecurity to deter defections to other parties by promising protection to supporters. Findings from an experimental study in Kenya suggest that candidates who position themselves as defending against out-group attacks are more acceptable to those who have experienced violence (Gutiérrez-Romero and LeBas 2020). Other research in developing countries suggests that the use of group-specific appeals can generate a backlash effect. Horowitz and Klaus (2020) show that political rhetoric cueing ethnic grievances around land rights had no effect on the majority of respondents, and might even work against candidates employing divisive rhetoric. Rosenzweig (2019), too, finds that hostile ethnic rhetoric, which can lead to violence, causes voters to withdraw their support from politicians, including from co-ethnic politicians. His results further suggest that politicians overestimate the electoral payoffs of threat-based appeals among voters, in part because “voter backlash [...] may be large enough to offset whatever electoral advantages violence may provide” (Rosenzweig 2019, 2)

A parallel line of research in political psychology also suggests that threat-based strategies can fail to resonate or even undercut political support (Chong and Druckman 2007; Bullock 2011; Huddy 2015). Exposure to information can interact with psychological characteristics to shape the consumption of rhetoric cueing racial threats (Bullock 2011). Ideology matters as well: In a meta-analysis of studies conducted in diverse national contexts, Jost et al. (2017) find that conservative constituents are more responsive to rhetoric centered on fear. Alternatively, threat-based appeals may undermine trust in politics and politicians, thereby reducing the perceived credibility of political rhetoric (Lau, Sigelman, and Rovner 2007).

Meso-Level: Group-Based Mediators of Threat-Based Appeals

Security is one of the most basic and vital functions of contemporary states (Tilly and Besteman 1985; Krasner 2007; North, Wallis, and Weingast 2009; Risse 2011), yet due to civil conflict or state collapse, private actors may provide it in lieu of the state. As a result, experiences of the state its security provision often vary across different territories and communities within conflict-affected countries, producing variation in perceptions of threat and insecurity along group lines. As Kalyvas notes, civil wars often involve a “convergence of local motives and supralocal incentives” (487), making multi-level analysis of preferences and beliefs in such contexts particularly appropriate. Beyond individual-level factors, therefore, subnational variation in the supply of security may shape citizen preferences for security provision.

Governance, including security provision, by militias and rebel groups has been extensively studied (Weinstein 2006; Arjona 2014; Staniland 2017). In some states, private security provision and low-intensity violence extend well into the post-conflict era, even as state capacity re-emerges (Daly 2016; Manning and Smith 2016; Balcells et al. 2015; Aliyev 2019). Private and public security provision are often not mutually exclusive in such contexts. Security provision may be jointly coordinated, during or after civil war, by the state and militias (Mampilly 2011; Arias 2020). This coheres with a large literature that stresses the diverse relationships that often emerge between formal and informal or non-state forms of governance (Helmke and Levitsky 2004). This literature often focuses on the production of basic services by public and private actors, in which private actors substitute for, collude with, or compete against the state (Cammett and MacLean 2014).

Seeking protection from violence, citizens may seek protection from a “hybrid state” comprising both the state police and private actors, such as gangs, in turn producing a hybridized form of citizenship in which both are seem as legitimate purveyors of force (Jaffe 2013; Bagayoko, Hutchful, and Luckham 2016). Such co-production can shape public opinion of what is considered acceptable action for different types of actors to provide in the security realm (Cooper-Knock 2014). This is true not only in post-conflict settings, but also more broadly in spaces where more traditional forms of authority are absent (Skarbek 2014).

Patterns of such hybrid security often spatially vary, with private and public actors competing or cooperating based on prior relationships and extant capacities (Moncada 2020). Arrangements between the state and non-state actors offer tacit permission for local militias to operate more extensively in some jurisdictions over others (LeBas 2013; Staniland 2015; Turnbull 2021). Hybrid security also varies according to the organizational type of the private actors involved—Lessing (2020), for example, distinguishes between private criminal governance structures according to their degrees of bureaucratization, and similarly Magaloni, Franco-Vivanco, and Melo 2020 argue that such groups vary by their level of collusion with the state and tendencies toward predation. In summary, this body of research suggests that both individual- and meso-level factors condition the provision of security and public perceptions therein.

Hypotheses

In this section, we develop several testable implications regarding the demand for different types of security provision—including national and group-specific security—in conflict-affected settings where both options are viable. We outline expectations for macro-level, meso-level, and micro-level variation in such contexts.

At the macro-level, insights from the literature on security provision in post-conflict states inform our expectation that demands for national and group-specific security provision are not mutually exclusive. In other words, there is no trade-off between preferences for candidates who promise strong national security and those who frame security around one’s communal group. When non-state providers continue to operate alongside the state, a “pluralist” or “assemblage” system of security provision ensues (Belhadj et al. 2015; Hazbun 2016). The everyday lived experience of citizens may therefore involve both types of security provision such that many individuals do not favor one type at the expense of the other. These points give rise to our first set of hypotheses:

H1: Within state contexts where national and group-specific security are both provided, demands for each type of security are not mutually exclusive.

H1a: Even individuals associated with credible group-specific security providers will not disfavor national security.

At the meso-level, we expect preferences for security provision to vary based on the credibility of the relevant provider for a given group. In many post-conflict contexts, including Lebanon, some identity-based groups are associated with stronger organizational structures that pledge (and deliver) group-specific security. For historical reasons related to the nature of prior conflict and relationships with the state, however, other communal groups may lack a competent, group-specific security provider. We anticipate that citizens will prefer group-specific security when their ascriptive identities associate them with groups with stronger legacies of security provision, relative to other individuals. These observations generate our second hypothesis:

H2: Demand for group-specific security is higher for groups associated with credible providers, and lower for groups without credible group-specific options.

At the micro- (or individual) level, we expect intra-group variation in preferences for security provision within communities featuring credible private providers. Specifically, we expect individuals to prefer group-specific security when they have stronger ties to group-specific providers, relative to other individuals within their community. Again, the credibility of provision is a key consideration. However, individuals with stronger ties to group-specific providers, such as more active partisans, are more likely to have benefited from group-specific security provision in the past, and are therefore more likely to prefer it in the future.

H3: Within groups with a credible supplier of group-specific security, individuals with stronger ties to the supplier will be more likely to favor that type of security.

H3a: Individuals from groups without credible in-group security providers will not express varied preferences for group-specific security depending on their connections to the provider.

Case Selection

Lebanon is an appropriate setting to explore these hypotheses on multiple grounds. First, the Lebanese civil war (1975–1990) saw the collapse of the state and rise of several militias that controlled and governed different regions of the country (Picard 2000; Rizkallah 2016). During this time, the state security apparatus effectively dissolved, and force was largely wielded by non-state political actors, each associated with one of the country’s ethnic groups.² Armed militias were formally disbanded by the war-ending Ta’if Agreement (1989), yet some armed groups continued to operate in the post-war period. Most prominently, Hezbollah, the Shia militia turned post-war political party, did not disarm and maintained its involvement in domestic and regional armed conflict. Other private, group-specific security providers also continue to operate alongside the state in the post-war era (Belhadj et al. 2015; Stedem 2020).

Despite this wartime history, the post-war Lebanese military is a relatively robust institution with high credibility among Lebanese of all sects. (See the Discussion and Appendix for descriptive statistics in support of this claim). Like its pre-war predecessor, the post-war military is structured by consociational power-sharing,³ in which power within and between branches is allotted on a proportional basis to different ethnoreligious groups. The Lebanese security sector is well-resourced, with an estimated 16% of the national budget devoted to the security sector, in addition to considerable foreign aid from the United States, France, and elsewhere (Carapico 2002; Berthier 2018). The military, like the Lebanese government as a whole, maintains an official stance of neutrality toward regional conflicts in Syria and elsewhere. Its highest-profile operations over the last decade have instead been in the service of containing domestic ethnic conflict, often associated with spillover of regional strife. As the following section describes, these conflicts are often ostensibly centered on the Sunni–Shia split in contemporary Lebanese politics.

Sunni–Shia Politics and Group-Specific Security Provision in Post-War Lebanon

The current macro-level political division between Sunni and Shia in Lebanon is a function of both regional and domestic factors. On the one hand, dominant Shia and Sunni ethnic parties in Lebanon have been ideologically opposed since the civil war, with tensions deepening following the 2005 assassination of former Prime Minister Rafiq Hariri and clashes between Sunni and Shia militants in Beirut in 2008. Since 2011, civil war in neighboring Syria further contributed to episodes of violence between Sunni and Shia parties and their security apparatuses in the urban centers of Tripoli and Sidon (Knudsen 2017). Though both Sunni and Shia parties have served alongside one another in subsequent national unity governments, their popular bases remain distinct and often opposed. (See the Methods section for a discussion of elite rhetoric in the two communities). Finally, regional tensions between the Gulf states and Iran have further politicized the contemporary divide between Lebanese Sunni and Shia political parties (Abdo 2013).

We focus our analysis on Lebanon’s Sunni and Shia populations because of their divergent post-war patterns of in-group security provision. The country’s most popular and well-represented Shia political party, Hezbollah, dates its origins back to the Lebanese civil war, when it was founded in 1982 with Iranian material support and ideological backing (Daher 2014). After operating as a militia during the latter part of the war, it evolved into a political party and has competed in post-war parliamentary and municipal elections. At the same time, it maintains a militia wing that conducts operations against Israel and, after 2011, mobilized in defense of regime forces in the Syrian civil war. Since 2004, Hezbollah has been closely allied with the Amal Movement, another Shia political party founded in the 1960s that also operated as a militia during the civil war. Unlike Hezbollah, however, Amal formally dismantled its militia wing after the war. Hezbollah and Amal have allied in electoral races since 2004, and both continue to provide private security within Shia-majority areas of Lebanon, often in cooperation with one another (Mroue 2020). In other words, in-group security provision by the Hezbollah–Amal alliance is quite credible to their Shia constituencies.

Meanwhile, Lebanon’s dominant Sunni political faction, the Future Movement, has little wartime or contemporary legacy of group-specific security provision. During the Lebanese civil war, various Sunni militant factions never coalesced into a single militia organization that controlled territory or outlasted the civil war. The reasons for the lack of a strong Sunni militia presence in the war and thereafter are beyond the scope of this paper, but may be traced to differing group-specific relationships to the Lebanese state and party structures prior to the war (Johnson 1986), as well as to the influence of Palestinian militant groups as a potential substitute for locally specific organizations (Dekmejian 1978; Hudson 1978).

The Sunni Future Movement derives its origins from the early post-war era and the leadership of Prime Minister Rafiq Hariri. Hariri maintained a large Parliamentary bloc throughout the 1990s and early 2000s and, in the aftermath of his 2005 assassination, his support base was formalized into the Future Movement under the leadership of his son, Saad. Like Hezbollah and Amal Movement, the Future Movement is an ethnic party that also directs considerable group-specific resources to supporters, both through and outside the state, including healthcare and education (Cammett 2014). Unlike major Lebanese Shia parties, however, the Future Movement maintains an explicit commitment to the disarmament of non-state militias and the routing of security assistance through the state. Though the party has allegedly supported Sunni militants in several instances of domestic conflict with Hezbollah, including the 2008 Beirut conflict and the 2011–2014 conflict in Tripoli, it largely refrains from providing private, group-specific security to its constituents (Stedem 2020). Therefore, in-group security provision by the Future Movement to its Sunni base is less credible.

How should we understand the dominant Lebanese Sunni faction’s comparatively weaker ability to commit to in-group security provision, relative to the dominant Lebanese Shia party alliance? We argue that this divergence is rooted in civil war-era dynamics that have subsequently persisted into contemporary (post-war) politics. During the Lebanese civil war, militant groups comprising both the Sunni and Shia dominant factions emerged and gained considerable popularity. That said, Sunni factions were disarmed as part of the war-ending Taif Agreement (1990), and no war-era Sunni militiant organizations were reconfigured as post-war parties (Nagle and Clancy 2019). On the other hand, Hezbollah was allowed to retain its armed capacity under the justification that this reassured against ongoing Israeli military action, much of it in Shia-majority areas of Lebanon. This exemption of Hezbollah from disarmament was undertaken with the endorsement of Syria, which continued to occupy the country until 2005 (Daher 2014). In other words, during the war and early post-war eras, dominant Lebanese factions developed divergent relative capacities for group-specific security provision that have persisted over time.

Finally, we note and account for the fact that support for respective dominant parties within Lebanon’s Sunni and Shia communities is not monolithic. Some members of Lebanon’s Shia population maintain opposition to Hezbollah, specifically for its continued presence as a militia within the country.⁴ Conversely, some Lebanese Sunnis support Hezbollah’s militia activities for various reasons linked to the regional environment.⁵ Therefore, as per our prior hypotheses, we investigate both meso-level and individual-level (i.e., within-group) variation in preferences for different types of security provision for Sunni and Shia Lebanese.

Methods

For empirical evidence, we rely on data from an original, nationally representative survey of voting-age adults (N = 2, 455) in Lebanon conducted in the fall of 2017. The survey consisted of a battery of demographic and sociopolitical questions followed by a conjoint experiment, in which respondents were asked to read two candidate biographies side-by-side and choose the candidate for whom they would be more likely to vote.⁶

Conjoint Experiment

In the conjoint experiment, respondents completed four tasks choosing between profiles of two hypothetical candidates running for a seat in the national legislature. Within each candidate’s biography, we randomly varied attributes to identify the causal effect of each attribute level on the likelihood of being elected. The order of attributes in a candidate’s biography was randomly assigned to avoid ordering effects, but remained consistent throughout all four rounds to ease the cognitive burden on respondents.⁷

Each pair of profiles, as depicted in Figure 1, came with two outcome questions: vote choice between the two candidates, and an assessment of how likely they were to attend a rally for each candidate.⁸ The ordering of these questions was randomized.

Figure 1.

Example of candidate profiles in conjoint task.

Attributes

Our conjoint experiment reflects the multidimensional choice voters face at the ballot box (Meyer and Rosenzweig 2016). Therefore, we use characteristics typically included in Lebanese candidate biographies and campaign materials. Table 1 lists all attribute levels. In this paper, we focus on the security attribute as it pertains to Sunni and Shia respondents, our main sub-groups of interest. We focus on three security attribute levels: framing of security as a Shia-centric good; framing of security as a Sunni-centric good; and framing of security as a national (public) good.

Table 1.

Conjoint attributes and levels.

Relationship with supporters	Security	Religion	Piety	Party	Platform	Education	Experience	Gender
Promises to work hard for the people in the electoral district	In a campaign speech, the candidate says: “I will ensure that the national security of Lebanon is the highest priority.”	Christian	Rarely attends religious services	Runs on the party list of a party that you neither support nor oppose	Discusses the problem of waste management and sanitation in the country	High school degree	None	Male
Distributes food baskets and cash during elections	In a campaign speech, the candidate says: “Wahhabi influences are threatening us. We must defend ourselves and our honor.”	Shia	Frequently attends religious services	Runs on the party list that you prefer	Discusses the problem of unemployment in the country	University degree	5 years	Female
Helps constituents to get parking tickets canceled	In a campaign speech, the candidate says: “Iranian influences are threatening us. We must defend ourselves and our honor.”	Sunni		Runs on the party list of a party you oppose	Has a detailed plan to address waste management and sanitation problems in the country	Post-graduate degree	15 years
Helps constituents get jobs for family members	In a campaign speech, the candidate says: “The extremists are anti-Christian. We will stand up to defend our values, liberty and justice.”				Has a detailed plan to boost job creation in the country
Arranges medical treatment for constituents

We focus on these attribute levels and sub-populations within the survey because they capture a core cleavage in contemporary Lebanese politics. References to an “Iranian threat” clearly cue Lebanese to think of threats from Shia armed groups, which have been targeted disproportionately at Lebanon’s Sunni population.⁹ Conversely, references to a “Wahhabi threat” clearly cue Lebanese to consider threats from Sunni armed groups, which have disproportionately affected Lebanon’s Shia population in an analogous way.¹⁰ Table 2 provides several examples of this rhetoric as used by Lebanese political figures in recent public addresses (emphasis added in italics).

Table 2.

Examples of group-specific security rhetoric from Lebanese political elites.

Sunni-specific security framing
“The assassination of Prime Minister Rafik Hariri paved the way for the Iranian project that has sought to dominate [Lebanon] and the region.”- Nohad Machnouk (Sunni; MP and former Interior Minister), 2018
“No person associated with the Iranian–Syrian project is allowed to represent the people of Tripoli.”- Ashraf Rifi (Sunni; former Justice Minister and MP candidate), 2018
“Hezbollah and its allies have many flaws, most notably that they are tools of the Iranian project in Lebanon.”- Ashraf Rifi, 2018
Shia-specific security framing
“Not just Muslims, but the whole world is suffering from the terrorist Wahhabi and takfiri ideology.”- Hashem Safieddine (Shia; head of Executive Council, Hezbollah), 2016
“ISIS and its ideology are nothing but Wahhabism, which is created in Saudi Arabia and exported with Saudi money to all parts of the world, at America’s request and support.” - Hassan Nasrallah (Shia; Secretary General, Hezbollah), 2017

Note: These statements were gathered for a larger study by the Lebanese Center for Policy Studies carried out between 2016 and 2019. Quotes were collected in Arabic from various Lebanese television, radio, newspaper, and social media sources.

We exclude Lebanese Christians from our core analysis because Christian-centric security pledges associated with Lebanese Christian elites often cue a wider variety of domestic and regional threats, such that a realistic attribute framing does not correspond to those associated with the Sunni and Shia sub-populations.

Sample and Fielding

We worked with Statistics Lebanon, a local survey firm, to recruit a nationally representative sample of 2,455 voting-age respondents. Interviews were conducted face-to-face in respondents’ homes by trained enumerators using the CAPI method and tablets between September 29 and November 18, 2017. Prior to the conjoint experiment, respondents answered a series of questions about their demographic characteristics, civic engagement, and political attitudes and behavior.¹¹ Enumerators read the questions and recorded respondents’ answers for most of the survey.¹² On average, respondents took about half an hour to complete the entire survey.

Statistics Lebanon maintains a list of populated areas covering the Lebanese national territory as well as the estimated population size in each location. The survey sample was drawn based on probability proportional to size, done in four stages: (1) selecting 240 clusters, each of which contains about 100–150 households; (2) randomly selecting 10 households from each cluster; (3) listing all adults of voting-age who were available at the time of the interview using a Kish table matrix; and (4) selecting an adult at random using the Kish table. The Appendix provides descriptive statistics of the sample.

Limits and Value of Survey Experiments

Conjoints are clearly shown to accurately represent real-world electoral behavior (Hainmueller et al. 2015; Bansak et al. 2020). Nonetheless, there is a clear discrepancy between our results and actual voter behavior observed in contemporary Lebanese politics. Since the end of the civil war, voters have consistently elected representatives from the country’s six main ethnic parties, all of which espouse promises of in-group protection. If a voter says they want public security provision but actually votes for candidates who promise otherwise, are our survey results less valid?

We argue that this seeming contradiction is a product of candidate supply rather than voter demand. Because a small set of ethnic parties have dominated contemporary Lebanese politics, they have also virtually monopolized control over candidate selection. In the 2018 elections, only one reformist candidate won a seat in Parliament, and secular-reformist lists received under 5% of the vote in almost every electoral district (Tavana and Parreira 2019). In prior post-war elections, no such reformist candidates ran or won seats. In other words, if credible alternatives to candidates promising in-group protection existed, the findings we will present suggest that at least some voters would be inclined to support them.

Results

The quantity of interest for our main analysis is the unadjusted marginal mean—the mean outcome across all appearances of a particular conjoint feature level, averaging across all other features.¹³ Marginal means, unlike the more frequently used average marginal component effect (AMCE), do not restrict the baseline categories to 0. Marginal means therefore provide a measure of favorability toward a given candidate characteristic (attribute level) for all attribute levels. In other words, results do not depend on the choice of the reference category (Leeper, Hobolt, and Tilley 2020).¹⁴

Preferences for National and Group-Specific Security

Figure 2 depicts the marginal means for different security framings, comparing across the Sunni and Shia sub-populations. First, we find that both Sunni and Shia favor candidates who frame security as a national good (no significant differences therein). Second, we find that Sunni respondents neither favor nor disfavor candidates who frame security as a Sunni-specific good, while Shia candidates disfavor such candidates. Third, we find that Shia respondents favor candidates who frame security as a Shia-specific good. In summary, while both sub-groups favor candidates who pursue a national security framing, only the Shia also favor candidates associated with an in-group security framing. As the Appendix shows, these findings are similar when the AMCE is measured.

Figure 2.

Marginal Means for Candidate Framing of Security Provision as National versus Group-Specific. Note: Point estimates represent marginal means. Horizontal lines represent confidence intervals. Because of the forced choice design, the grand mean is by definition 0.5. Marginal means above 0.5 indicate a positive preference for a candidate with a particular attribute level (e.g., a candidate prioritizing national security); marginal means below 0.5 indicate a negative preference. Standard errors are clustered by respondent.

Preferences by Candidate and Respondent Party Affiliation

Next, we further divide Sunni and Shia respondents based on the party affiliation of the candidates they were asked to evaluate. We argue that the dominant Shia party is associated with robust group-specific security institutions, while the dominant Sunni party has no similar legacy of security provision. Both respondent pools are grouped based on whether the candidate being evaluated was from the party the respondent supported, versus a neutral or opposed party. The exact party was not specified, and therefore contingent on whatever favored or opposed party the respondent had in mind.

Figure 3 shows marginal mean results for the four main sub-groups of this analysis. We find that both Sunni and Shia largely favor (or do not disfavor) a national security framing, regardless of whether the candidate is from their preferred party. Shia favor a national security framing even more strongly when it comes from a candidate from their preferred party. Second, Sunni respondents do not favor in-group security pledges when they come from a candidate from their preferred party. Finally, by contrast, Shia respondents do favor in-group security provision when it is pledged by a candidate from their preferred party. This provides corroborating evidence that party-level behavior and, especially, citizen views of the security-related activities of parties are key drivers of the differences in in-group protection perceptions between Shia and Sunni respondents.

Figure 3.

Marginal Means for Candidate Security Framing by Candidate Party. Note: Point estimates represent marginal means. Horizontal lines represent confidence intervals. Because of the forced choice design, the grand mean is by definition 0.5. Marginal means above 0.5 indicate a positive preference for a candidate with a particular attribute level (e.g., a candidate prioritizing national security); marginal means below 0.5 indicate a negative preference. Standard errors are clustered by respondent.

Figure 4 depicts an analogous set of results, with Sunni and Shia respondents once again grouped by party support. This time, however, Sunni and Shia respondents are divided based on self-reported support for their ethnic group’s dominant party. We separate Sunni supporters of the dominant Sunni party (the Future Movement) from other Sunni respondents, and Shia supporters of the dominant Shia party alliance (Hezbollah–Amal Movement) from other Shia respondents. While Figure 3 allows both sub-groups to envision pledges of in-group security by any party of their choice, this sub-group analysis constrains support to the parties most prominently associated with each group.

Figure 4.

Marginal Means for Candidate Security Framing by Respondent Party. Note: Point estimates represent marginal means. Horizontal lines represent confidence intervals. Because of the forced choice design, the grand mean is by definition 0.5. Marginal means above 0.5 indicate a positive preference for a candidate with a particular attribute level (e.g., a candidate prioritizing national security); marginal means below 0.5 indicate a negative preference. Standard errors are clustered by respondent. Respondent party support is coded based on who the respondent anticipated voting for in the next (2018) Parliamentary elections. The dominant party for the Sunni sub-group is the Future Movement; the dominant party for the Shia sub-group is the Hezbollah–Amal alliance (either party named).

Figure 4 shows, similarly, that both dominant party supporters and other respondents for both Sunni and Shia largely favor (or are neutral toward) a national security framing, with no relevant differences across sectarian groups. It also shows that Sunni supporters of the Future Movement do not significantly favor in-group security framings relative to non-supporters. However, Shia supporters of the Hezbollah–Amal alliance do favor in-group security framings relative to other Shia respondents. This provides further evidence that perceptions of party-level differences, specifically between these particular dominant parties, are key drivers of Sunni and Shia security preferences.

Robustness Checks

We conduct a variety of robustness checks to account for important confounders. The Appendix provides all of these results. First, we conduct all analyses presented in the Results section estimating the AMCE, rather than the marginal mean. We also estimate security preferences while excluding unrealistic profiles, notably Sunni candidates discussing Shia-centric security or Shia candidates discussing Sunni-centric security. We also estimate security preferences when the outcome variable is a 1–7 scale indicating how likely the respondent would be to attend a rally for the candidate they evaluated. Additional sub-group analyses in Supplemental Appendix A.7–A.14 reassure against a variety of alternative hypotheses to explain disparate preferences between Sunni and Shia respondents.

Discussion: Trust in National and Group-Specific Security Providers in Lebanon

Why do Lebanese Sunni and Shia respondents have divergent preferences for candidates who use different security framings? We argue that the credibility of providers can explain these differences. While national security is a valid promise from candidates of all ethnic groups and party persuasions, some parties can more convincingly provide group-specific security than others. Specifically, while Shia parties have a strong track record of group-specific security provision in contemporary Lebanese politics, the dominant Sunni party does not maintain the same credibility in this domain. Here, we present corresponding descriptive evidence that these differences have resulted in divergent levels of trust in and support for Sunni versus Shia parties.

Trust in Parties and the Credibility of Group-Specific Security Framing

Table 3 displays descriptive differences in political attitudes between our Sunni and Shia respondent pools. These findings are corroborated with data from Wave IV of the Arab Barometer, as presented in the Appendix.

Table 3.

Differences in political attitudes between Sunni and Shia respondents.

	All	Shia	Sunni	p-Value
Demographic characteristics
Female	0.502	0.502	0.502	0.997
Female	(0.000)	(0.001)	(0.001)
Household socioeconomic security	2.707	2.585	2.824	0.00
Household socioeconomic security	(0.001)	(0.001)	(0.001)
Level of education	4.66	5.216	4.13	0.00
Level of education	(0.002)	(0.003)	(0.003)
Political attitudes
Party supporter	0.285	0.36	0.214	0.00
Party supporter	(0.000)	(0.001)	(0.001)
Hezbollah–Amal supporter	0.187	0.343	0.038	0.00
Hezbollah–Amal supporter	(0.000)	(0.001)	(0.000)
Future movement supporter	0.052	0.002	0.101	0.00
Future movement supporter	(0.000)	(0.000)	(0.000)
Feels part of religious community	0.831	0.832	0.831	0.934
Feels part of religious community	(0.000)	(0.001)	(0.001)
Feels unsafe	0.654	0.665	0.644	0.446
Feels unsafe	(0.000)	(0.001)	(0.001)
Trust in institutions
Trusts Lebanese political parties	0.169	0.26	0.082	0.00
Trusts Lebanese political parties	(0.000)	(0.001)	(0.000)
Trusts Lebanese parliament	0.318	0.42	0.22	0.00
Trusts Lebanese parliament	(0.000)	(0.001)	(0.001)
Trusts Lebanese army	0.913	0.919	0.908	0.519
Trusts Lebanese army	(0.000)	(0.000)	(0.000)

Note: Data come from a nationally representative survey (N = 2455) conducted by the authors in collaboration with the Lebanese Center for Policy Studies in 2017. Household socioeconomic status is coded on a 1 to 4 scale (1 = very insecure with no savings; 4 = very secure with savings). Education is coded on a scale of 1–9. Attachment to religious community, perceived safety, and trust in institutions are coded using a binary variable.

First, we find that Sunni and Shia respondents maintain very different levels of support for Lebanese political parties, and their dominant sectarian parties in particular. 36% of Shia respondents report supporting a political party, compared to only 21% of Sunni respondents. About 95% of Shia party supporters (34% of the whole Shia sample) back the Hezbollah–Amal alliance, compared to the 50% of Sunnis who support the only national Sunni party, the Future Movement (10% of the whole Sunni sample). With respect to other relevant factors, however, such as perceptions of safety and religious belonging, the Sunni and Shia populations exhibit no real variation (i.e., no significant differences). This provides some reassurance that disparate security threats against Sunni and Shia respondents are not driving security preferences in the conjoint setup.

Second, we find that trust in Lebanese political parties is significantly higher for Shia respondents than Sunni respondents. 26% of Shia respondents report high levels of trust in parties (with no particular party named), compared to only eight percent of Sunni respondents. This is echoed in levels of trust in the Lebanese Parliament in the two respondent pools report (42 vs 22% for Shia and Sunnis, respectively). Given that the Parliament is dominated by representatives of the country’s major sectarian parties, including the Hezbollah–Amal alliance and the Future Movement, cross-group variation in this dimension of political trust is not surprising.

Finally—and importantly for our analysis—we find that Sunni and Shia respondents have equal and very high levels of trust in the Lebanese military. 92% of Shia respondents and 91% of Sunni respondents report high levels of trust in the army, demonstrating the credibility of this institution as a provider of national security irrespective of ethnic affiliation. As previously discussed, we think high trust in the military is a foundation of our results: While preferences for group-specific security framing diverge between Sunni and Shia, both favor candidates who employ a national security frame.

Trust in the Military and the Credibility of National Security

We argue that positive perceptions of the role of the military in Lebanon, a common phenomenon in both the Arab World (Lotito 2018) and the OECD (Johnson 2018), shape our finding that all Lebanese respondents prefer candidates who use a national security framing. As noted above, trust in the Lebanese armed forces among respondents is extremely high (91% overall) and far greater than for any other political institution, a finding validated with data from the Arab Barometer, which can be found in the Appendix. Table 3 shows that, across sects, the military is viewed more favorably by several orders of magnitude than the Parliament or political parties. Certain characteristics of the Lebanese military that apply to some, but not all post-conflict contexts may play a role in these high levels of trust. Unlike in Iraq or Libya, for example, Lebanon’s armed forces have not been historically associated with exclusionary policies, such as autocratic regime maintenance and coup-proofing (Sayigh 2018).

Finally, high levels of trust in the capacity of Lebanese national security providers may stem in part from the nature of sectarian power-sharing, which also applies to military institutions. As in other polities with power-sharing arrangements, Lebanon’s consociational system both cements sectarian identity and at least partially encourages cross-cutting, multi-ethnic coalition building. With over 50 identified cases of institutionalized ethnic power-sharing (Hartzell and Hoddie 2020), the track record of such institutions for delivering effective governance is mixed (Horowitz 1993; Reilly 2012; Rothchild and Roeder 2005; Fakhoury 2019). We do not take a particular side in this debate, but instead note that Lebanon’s power-sharing system may have contributed to the maintenance of a diverse military that engendered buy-in from all major sectarian groups (Hartzell and Hoddie 2003).

Conclusion

In Lebanon and elsewhere, conflict and state breakdown have generated a diverse array of institutional providers of basic protection. How does such pluralist security provision persist, and why do some individuals or groups prefer one option versus the other? Based on a conjoint experiment in Lebanon, our findings show that respondents value pledges to focus on national security as much, if not more, than group-specific promises. At the same time, we argue, the credibility of in-group protection drives varying attitudes towards promises of in-group protection.

These results indicate that unifying appeals can have greater resonance than divisive ones, even in conflict-affected contexts where elites frequently deploy ethnocentric rhetoric and fear-mongering as political tactics. At the same time, parallel preferences for group-specific security may help to illuminate why non-state actors continue to operate alongside states, which face challenges in maintaining the monopoly on violence in conflict-affected countries. In short, our findings in Lebanon may reflect the nuanced realities of pluralist security provision as an equilibrium.

The conditions under which individuals favor certain forms of private protection over others deserve further investigation in conflict-affected or insecure settings, including Lebanon. Our survey in Lebanon was conducted during a time period (2017) just prior to the start of massive, unprecedented anti-government protests, precipitated by an economic crisis widely attributed to the parties discussed in this paper, among others. Future work should explore how Lebanese public opinion has evolved in light of these developments. Future research should also explore whether the provision of security on exclusionary terms—that is, for specific in-group communities—is driven by supply or demand-side factors. Although public goods provision by private actors is often assumed to be demand driven (Alesina, Baqir, and Easterly 1999; Keefer 2007), this may not always be the case, particularly in the case of security.

Supplemental Material

sj-pdf-1-jcr-10.1177_00220027221079401 – Supplemental Material for Commitment to the “National” in Post-Conflict Countries: Public and Private Security Provision in Lebanon

Supplemental Material, sj-pdf-1-jcr-10.1177_00220027221079401 for Commitment to the “National” in Post-Conflict Countries: Public and Private Security Provision in Lebanon by Melani Cammett, Christiana Parreira, Dominika Kruszewska-Eduardo and Sami Atallah in Journal of Conflict Resolution

Footnotes

Acknowledgments

For valuable feedback and suggestions, we thank Carol Abi-Ghanem, Daniel Corstange, Nadim El Kak, Anselm Hager, Adam Harris, Mike Hoffman, Ellen Lust, Mashail Malik, Juan Masullo, Gwyneth McClendon, Ahmed Mohamed, Jana Mourad, Lama Mourad, Laura Paler, David Romney, Leah Rosenzweig, Steven Rosenzweig, Bassel Salloukh, Aytug Sasmaz, Daniel Smith, Kelly Stedem, Anton Strezhnev, Daniel Tavana, Lily Tsai, Sami Zoughaib, and participants in seminars at the American Political Science Association, Boston University, Gothenburg University, Harvard University, Leiden University, Notre Dame University, and U.C. Berkeley, as well as researchers at the Lebanese Center for Policy Studies. All errors are our own.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The data collection for this research was supported by a grant from the European Union to the Lebanese Center for Policy Studies. The authors received no funding for the authorship and publication of this article.

ORCID iD

Christiana Parreira

Supplemental Material

Supplemental material for this article is available online.

Notes

References

Abdo

2013. The New Sectarianism: The Arab Uprisings and the Rebirth of the Shi‘a-Sunni Divide. Technical Report 29. Washington, DC: Brookings.

Acemoglu

Robinson

J. A.

Santos

R. J.

2013. “The Monopoly of Violence: Evidence from Colombia.” Journal of the European Economic Association 11 (S1): 5-44.

Alesina

Baqir

Easterly

1999. “Public Goods and Ethnic Divisions.” The Quarterly Journal of Economics 114 (4): 1243-1284.

Aliyev

2019. “When and How Do Militias Disband? Global Patterns of Pro-Government Militia Demobilization in Civil Wars.” Studies in Conflict & Terrorism 42 (8): 715-734.

Arias

E. D.

2020. “Understanding Criminal Networks, Political Order, and Politics in Latin America.” In Ungoverned Spaces, pp. 115-135. Stanford, CA: Stanford University Press.

Arjona

2014. “Wartime Institutions: A Research Agenda.” Journal of Conflict Resolution 58: 1-30.

Bagayoko

Hutchful

Luckham

2016. “Hybrid Security Governance in Africa: Rethinking the Foundations of Security, Justice and Legitimate Public Authority.” Conflict, Security & Development 16 (1): 1-32.

Balcells

Daniels

L.-A.

Escriba`-Folch

2015. “The Determinants of Low-intensity Intergroup Violence: The Case of Northern Ireland.” Journal of Peace Research 53 (1): 33-48.

Bansak

Hainmueller

Hopkins

D. J.

Yamamoto

(2020). Using Conjoint Experiments to Analyze Elections: The Essential Role of the Average Marginal Component Effect (AMCE).

10.

Bar-Tal

Halperin

De Rivera

2007. “Collective Emotions in Conflict Situations: Societal Implications.” Journal of Social Issues 63 (2): 441-460.

11.

Barma

2017. The Peacekeeping Puzzle: Political Order in Post-Conflict States. New York City: Cambridge University Press.

12.

Bassam

2018. New Snags Hold up Formation of Lebanon Unity Government. Beirut: Reuters.

13.

Bassam

Karouny

2013. Over 50 Hurt as Car Bomb Hits Hezbollah Beirut Stronghold. Beirut: Reuters.

14.

Belhadj

van der Borgh

Jaffe

Price

Stel

Warren

M. J.

2015. Plural Security Provision in Beirut. Policy Brief. The Hague: Knowledge Platform Security & Rule of Law.

15.

Berthier

2018. The Lebanese Warfare State. Beirut: Synaps.

16.

Blanford

2008. Uncertainty Deepens in Lebanon as Hezbollah Seizes Control of West Beirut. Beirut: Christian Science Monitor.

17.

Brass

P. R.

1997. Theft of an Idol: Text and Context in the Representation of Collective Violence. Princeton, NJ: Princeton University Press.

18.

Brewer

M. B.

1999. “The Psychology of Prejudice: Ingroup Love or Outgroup Hate? ”Journal of Social Issues 55 (3): 429-444.

19.

Brubaker

2002. Ethnicity Without Groups. Cambridge, MA: Harvard University Press.

20.

Bullock

J. G.

2011. “Elite Influence on Public Opinion in an Informed Electorate.” American Political Science Review 105 (3): 496-515.

21.

Bulutgil

H. Z.

2015. “Social Cleavages, Wartime Experience, and Ethnic Cleansing in Europe.” Journal of Peace Research 52 (5): 577-590.

22.

Cammett

2014. Compassionate Communalism: Welfare and Sectarianism in Lebanon. Ithaca, NY: Cornell University Press.

23.

Cammett

MacLean

L. M.

2014. The Politics of Non-State Welfare. Ithaca, NY: Cornell University Press.

24.

Cammett

Malesky

2012. “Power Sharing in Postconflict Societies: Implications for Peace and Governance.” The Journal of Conflict Resolution 56 (6): 982-1016.

25.

Canetti-Nisim

Halperin

Sharvit

Hobfoll

S. E.

2009. “A New Stress-based Model of Political Extremism: Personal Exposure to Terrorism, Psychological Distress, and Exclusionist Political Attitudes.” Journal of Conflict Resolution 53 (3): 363-389.

26.

Carapico

2002. “Foreign Aid for Promoting Democracy in the Arab World.” The Middle East Journal 56, 379-395.

27.

Chandra

2006. “What is Ethnicity and Does it Matter?” Annual Review of Political Science 9: 397-424.

28.

Chang

Krosnick

J. A.

2010. “Comparing Oral Interviewing with Self-Administered Computerized Questionnaires: An Experiment.” Public Opinion Quarterly 74 (1): 154-167.

29.

Cho

W. K. T.

Gimpel

J. G.

2006. “Clarifying the Role of SES in Political Participation: Policy Threat and Arab American Mobilization.” Journal of Politics 68 (4): 977-991.

30.

Chong

Druckman

J. N.

2007. “Framing Public Opinion in Competitive Democracies.” American Political Science Review 101 (4): 637-655.

31.

Cooper-Knock

S.-J.

2014. “Policing in Intimate Crowds: Moving Beyond ‘the Mob’ in South Africa.” African Affairs 113 (453): 563-582.

32.

Daher

2014. Hezbollah: Mobilization and Power. Paris: PUF.

33.

Daly

S. Z.

2016. Organized Violence After Civil War: The Geography of Recruitment in Latin America. Cambridge, UK: Cambridge University Press.

34.

Das

2006. Life and Words: Violence and the Descent into the Ordinary. Berkeley: University of California Press.

35.

Davis

D. W.

Silver

B. D.

2004. “Civil Liberties vs. Security: Public Opinion in the Context of the Terrorist Attacks on America.” American Journal of Political Science 48 (1): 28-46.

36.

Dekmejian

R. H.

1978. “Consociational Democracy in Crisis: The Case of Lebanon.” Com- Parative Politics 10 (2): 251-265.

37.

Fakhoury

2019. “Power-sharing after the Arab Spring? Insights from Lebanon’s Political Transition.” Nationalism and Ethnic Politics 26 (1): 9-26.

38.

Fearon

Laitin

2000. “Violence and the Social Construction of Ethnic Identity.” International Organization 54 (4): 845-877.

39.

Guti´errez-Romero

LeBas

2020. “Does Electoral Violence Affect Vote Choice and Willingness to Vote? Conjoint Analysis of a Vignette Experiment.” Journal of Peace Research 57 (1): 77-92.

40.

Hainmueller

Hangartner

Yamamoto

2015. “Validating Vignette and Conjoin Survey Experiments against Real-world Behavior. Proceedings of the National Academy of Sciences 112 (8), 2395–2400.

41.

Hainmueller

Hopkins

D. J.

(2015). The Hidden American Immigration Consensus: A Conjoint Analysis of Attitudes toward Immigrants.” American Journal of Political Science 59 (3): 529-548.

42.

Hartzell

Hoddie

2003. “Institutionalizing Peace: Power Sharing and Post-Civil War Conflict Management.” American Journal of Political Science 47 (2): 318-332.

43.

Hartzell

C. A.

Hoddie

2020. Power Sharing and Democracy in Post-Civil War States: The Art of the Possible. Cambridge, UK: Cambridge University Press.

44.

Hazbun

2016. “Assembling Security in a ‘Weak State:’ the Contentious Politics of Plural Governance in Lebanon since 2005.” Third World Quarterly 6597: 1-18.

45.

Helmke

Levitsky

2004. “Informal Institutions and Comparative Politics: A Research Agenda.” Perspectives on Politics 2 (4): 725-740.

46.

H⊙glund

Piyarathne

2009. “Paying the Price for Patronage: Electoral Violence in Sri Lanka.” Commonwealth & Comparative Politics 47 (3): 287-307.

47.

Horowitz

1993. “The Challenge of Ethnic Conflict: Democracy in Divided Societies.” Journal of Democracy 4 (4): 18-38.

48.

Horowitz

Klaus

2020. “Can Politicians Exploit Ethnic Grievances? An Experi- mental Study of Land Appeals in Kenya.” Political Behavior 42: 35-58.

49.

Hubbard

Saad

2021. “Prominent Lebanese Critic of Hezbollah Is Killed.” New York Times.

50.

Huddy

2015. “Group Identity and Political Cohesion.” In Scott

Kosslyn

(Eds.), Emerging Trends in the Social and Behavioral Sciences. Hoboken, NJ: John Wiley & Sons.

51.

Huddy

Feldman

Taber

Lahav

2005. “Threat, Anxiety, and Support of Antiterrorism Policies.” American Journal of Political Science 49 (3): 593-608.

52.

Hudson

M. C.

1978. “The Palestinian Factor in the Lebanese Civil War.” Middle East Journal 32 (3): 261-278.

53.

Jaffe

2013. “The Hybrid State: Crime and Citizenship in Urban Jamaica.” American Ethnologist 40 (4): 734-748.

54.

Johnson

2018. Trust in the Military Exceeds Trust in Other Institutions in Western Europe and U.S.

55.

Johnson

1986. Class and Client in Beirut. London: Ithaca Press.

56.

Jost

J. T.

Stern

Rule

N. O.

Sterling

2017. “The Politics of Fear: Is There an Ideological Asymmetry in Existential Motivation?” Social Cognition 35 (4): 324-353.

57.

Keefer

2007. “Clientelism, Credibility, and the Policy Choices of Young Democracies.” American Journal of Political Science 51 (4): 804-821.

58.

Knudsen

A. J.

2017. “Patrolling a Proxy War: Citizens, Soldiers and Zu’ama in Syria Street, Tripoli.” In Civil-Military Relations in Lebanon, pp. 71-99. New York City: Springer.

59.

Krasner

S. D.

2007. “Sovereignty.” The Blackwell Encyclopedia of Sociology.

60.

Lau

R. R.

Sigelman

Rovner

I. B.

2007. “The Effects of Negative Political Cam- paigns: A Meta-Analytic Reassessment.” Journal of Politics 69 (4): 1176-1209.

61.

LeBas

2013. “Violence and Urban Order in Nairobi, Kenya and Lagos, Nigeria.” Studies in Comparative International Development 48 (3): 240-262.

62.

Leeper

T. J.

Hobolt

Tilley

2020. “Measuring Subgroup Preferences in Conjoint Experiments.” Political Analysis 28 (2): 207-221.

63.

Lerner

J. S.

Keltner

2001. “Fear, Anger, and Risk.” Journal of Personality and Social Psychology 81 (1): 146-159.

64.

Lessing

2020. “Conceptualizing Criminal Governance.” Perspectives on Politics 19, 1-20.

65.

Lotito

N. J.

2018. Public Trust in Arab Armies. Technical report, Carnegie Endowment for International Peace.

66.

Lupu

Peisakhin

2017. “The Legacy of Political Violence Across Generations.” American Journal of Political Science 61 (4): 836-851.

67.

Magaloni

Franco-Vivanco

Melo

2020. “Killing in the Slums: Social Order, Criminal Governance, and Police Violence in Rio de Janeiro.” American Political Science Review 114 (2): 552-572.

68.

Mampilly

2011. Rebel Rulers: Insurgent Governance and Civilian Life During War. Ithaca, NY: Cornell University Press.

69.

Manning

Smith

2016. “Political Party Formation by Former Armed Opposition Groups After Civil War.” Democratization 23 (6): 972-989.

70.

Meagher

2012. “The Strength of Weak States? Non-State Security Forces and Hybrid Governance in Africa.” Development and Change 43 (5): 1073-1101.

71.

Mendelberg

2001. The Race Card: Campaign Strategy, Implicit Messages, and the Norm of Equality. Princeton, NJ: Princeton University Press.

72.

Meyer

Rosenzweig

L. R.

2016. “Conjoint Analysis Tools for Developing Country Contexts.” The Political Methodologist 23 (2): 2-6.

73.

Moncada

2020. Cities, Business, and the Politics of Urban Violence in Latin America. Stanford, CA: Stanford University Press.

74.

Mroue

2020. Lebanon’s Amal and Hezbollah: The Past in the Present? Technical report. London: London School of Economics.

75.

Nacos

B. L.

Bloch-Elkon

Shapiro

R. Y.

2011. Selling Fear: Counterterrorism, the Media, and Public Opinion. Chicago: University of Chicago Press.

76.

Nagle

Clancy

M.-A.

2019. “Power-sharing after Civil War: Thirty Years since Lebanon’s Taif Agreement.” Nationalism and Ethnic Politics 25 (1): 1-8.

77.

North

D. C.

Wallis

J. J.

Weingast

B. R.

2009. Violence and Social Orders: A Conceptual Framework for Interpreting Recorded Human History. Cambridge: Cambridge University Press.

78.

Parkinson

S. E.

2021. “Practical Ideology in Militant Organizations.” World Politics 73: 1-30.

79.

Pearlman

2013. “Emotions and the Microfoundations of the Arab Uprisings.” Perspectives on Politics 11 (2): 387-409.

80.

Petersen

R. D.

2002. Understanding Ethnic Violence: Fear, Hatred, and Resentment in Twentieth-Century Eastern Europe. New York: Cambridge University Press.

81.

Picard

2000. “The Political Economy of Civil War in Lebanon.” In War, Institutions, and Social Change in the Middle East. Berkeley: University of California Press.

82.

Reilly

2012. “Centripetalism: Cooperation, Accommodation and Integration.” In Wolff

Yakinthou

(Eds.), Conflict Management in Divided Societies: Theories and Practice, pp. 57-65. New York: Routledge.

83.

Risse

2011. Governance Without a State?: Policies and Politics in Areas of Limited Statehood. New York City: Columbia University Press.

84.

Rizkallah

2016. The Electoral Legacies of War. Working Paper. Los Angeles: University of California.

85.

Rosenzweig

S. C.

2019. Dangerous Disconnect: Voter Backlash, Elite Misperception, and the Costs of Violence as an Electoral Tactic.

86.

Rothchild

Roeder

P. G.

2005. “Power Sharing as an Impediment to Peace and Democracy.” In Roeder

P. G.

Rothchild

(Eds.), Sustainable Peace: Power and Democracy after Civil Wars, pp. 29-50. Ithaca, NY: Cornell Univeristy Press.

87.

Sayigh

2018. Hybridizing Security: Armies, Militias and Constrained Sovereignty.

88.

Skarbek

2014. The Social Order of the Underworld: How Prison Gangs Govern the American Penal System. Oxford, UK: Oxford University Press.

89.

Staniland

2015. “Militias, Ideology, and the State.” Journal of Conflict Resolution 59 (5): 770-793.

90.

Staniland

2017. “Armed Politics and the Study of Intrastate Conflict.” Journal of Peace Research 54 (4): 459-467.

91.

Stedem

2020. Reconceptualizing Security: Political Parties as Policemen.

92.

Tajfel

Turner

1979. “An Integrative Theory of Intergroup Conflict.” In Austin

W. G.

Worchel

(Eds.), The Social Psychology of Intergroup Relations, pp. 7-24. Monterey, CA: Brooks/Cole.

93.

Tavana

Parreira

2019. Lebanon’s 2018 Election: New Measures and the Resilience of the Status Quo. Technical report. Beirut: Lebanon Support.

94.

Tilly

Besteman

1985. “War Making and State Making as Organized Crime.” Violence: A Reader, 35-60.

95.

Turnbull

2021. “When Governments Tolerate or Repress Armed Groups: Evidence from Nigeria.” Working Paper.

96.

Valentino

N. A.

Hutchings

V. L.

White

I. K.

2002. “Cues That Matter: How Political Ads Prime Racial Attitudes during Campaigns.” American Political Science Review 96 (1): 75-90.

97.

Valentino

N. A.

Neuner

F. G.

Vandenbroek

L. M.

2017. “The Changing Norms of Racial Political Rhetoric and the End of Racial Priming.” 80 (3).

98.

Weber

1965. Politics as a Vocation. Philadelphia: Fortress Press.

99.

Weinstein

J. M.

2006. Inside Rebellion: The Politics of Insurgent Violence. Cambridge, UK: Cambridge University Press.

100.

Young

1979. The Politics of Cultural Pluralism (Revised edition ed.). Madison, WI: University of Wisconsin Press.

101.

Young

L. E.

2019. “The Psychology of State Repression: Fear and Dissent Decisions in Zimbabwe.” American Political Science Review 113 (1): 140-155.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.77 MB