Over the last ten years, Mexico and other countries in Latin America have gone through a deep crisis of insecurity (Bisogno et al. 2014). To cope with this crisis, some individuals have reached out to their neighbors in attempts to form stable or ephemeral organizations to address their concerns about crime.1 Although not all of these attempts have been successful, the ones that have pose a threat to the rights, lives, and livelihoods of alleged criminals, innocent bystanders, and vigilantes themselves (Bates, Greif, and Singh 2002; Hine 1998). Just months before this piece was written, for example, two surveyors working for a polling firm were confused for kidnappers and burned alive by citizens of a small town in Puebla (central Mexico).
Given that anti-crime organization attempts can result in the emergence of stable vigilante organizations, they can represent a serious challenge to the state’s capacity to monopolize the coercive use of violence (Silke 2001; Weber 1919). By 2014, for instance, a significant area spanning the states of Guerrero, Oaxaca, and Michoacán (southwest Mexico) was controlled by self-defense organizations. The dominance of these organizations led journalists, scholars, and policy makers to consider the possibility of Mexico becoming a failed state (Grant 2014). What forces propel individuals to reach out to their neighbors in an attempt to defend themselves from crime?
In this article, I propose that when citizens distrust the effective intervention of law enforcement, their perception of the trustworthiness of their neighbors translates into an increased probability of them reaching out to other community members to defend themselves from crime (whether through prevention or control). To test this claim, I analyze the determinants of anti-crime organization attempts over several years in Mexico, using the National Survey of Victimization and Perception of Public Safety (Encuesta Nacional de Victimización y Percepción sobre Seguridad Pública, ENVIPE). The findings presented here are important because, in recent years, citizens reaching out to their neighbors in attempts to fight crime have challenged the capacity of the Mexican state to maintain social and political order.2 These results contribute to our understanding of the state’s moderating role in society and suggest that, when the connection between citizens and the state erodes significantly, citizens turn to other actors (in this case their community) as potential substitutes for the state.
Neighborhood watches, communitarian police forces, paramilitary self-defense movements, and harmless anti-criminal committees are some of the types of organizations that can evolve from citizens’ attempts to address their concerns about crime. In Mexico and other countries, these organizations have introduced social uncertainty and have seriously challenged the power of the state.
In urban centers, citizens have reached out to community members to illegally close streets, set up threatening signs, and start neighborhood watches (Salgado 2014). In some neighborhoods of Mexico City, for instance, anonymous groups of citizens have posted signs promising severe punishments against thievery. In recent years, victims of crime have also turned to their neighbors to violently attack alleged criminals and police officers (Alarcón 2015).
In rural areas, citizens have turned to their communities to set up night watches, citizen-run community police forces (policías comunitarias), and other self-defense organizations (grupos de autodefensa) in an attempt to directly confront criminals.3 In one of the most extreme examples, citizens concerned about deforestation near the town of Cherán (in the state of Michoacán, western Mexico) mobilized their community to violently confront the illegal loggers that crossed the town every day. Soon after, they formed a policía comunitaria and expelled the elected authorities from the municipality. Today, citizens of Cherán manage their own security, elect their own authorities outside the established electoral system, and abstain from participating in state and federal elections (Cendejas, Arroyo, and Sánchez 2015).
Similarly, tired of being extorted by drug cartels, agricultural producers from Michoacán turned to their community to start an armed paramilitary movement popularly referred to as the Autodefensas de Tierra Caliente (the self-defense movement of Tierra Caliente). This group occupied several municipalities in the state of Michoacán. It collected donations, recruited men, distributed guns, and assassinated alleged drug cartel members in the region. At its peak (January 2014), this organization controlled an area larger than the state of Delaware and seriously challenged the capacity of the Mexican state (Malkin and Villegas 2014).4 Although these cases are extreme examples (as most anti-crime organizations only have a very local impact), they illustrate how anti-crime organization has the potential to carry with it important social and political consequences.
Organizations of this type vary in their structures, challenges, methods, and specific objectives. However, most anti-crime organizations (even the most violent and sophisticated) start from small, and seemingly harmless, individual attempts by citizens to fight crime with the help of their community.5 Thus, without dismissing the need for nuance in the treatment of anti-criminal groups, from a dynamic perspective, it is important to understand under what conditions individual citizens are more likely to turn to their community to attempt to defend themselves from crime. Only on this basis will it be possible to understand why some of these anti-criminal attempts consolidate and diverge in different types of organizations.
Although the individual origins of anti-crime organization have rarely been investigated, this is not the first time that their manifestations have been studied. Secondary (anti-criminal groups) and tertiary (violent events perpetrated by anti-criminal groups) manifestations of anti-crime organization attempts have captured the attention of social scientists for a long time. Contributions on the subject have come from a diversity of fields and can be organized into three general groups.
First, some propose understanding anti-criminal action as a manifestation of group conflict. They suggest that ethnic, political, economic, and demographic change increases citizens’ likelihood to associate with their in-group to confront threatening out-group members (see Hepworth and West 1988; Levine and Campbell 1972; Tadjoeddin and Murshed 2007).6 Others propose that, in hierarchical communities, anti-criminal action is more likely to target crimes committed against those with high social status. This type of behavior therefore serves to restore social order (Black 1976; Senechal de la Roche 2001; Tolnay and Beck 1995).7
A second group of scholars (mostly focused on vigilante groups) has argued that this phenomenon can be understood as a manifestation of institutional competition, that is, a conflict between the norms, values, and structures that preexist in specific societies and the legal framework of the state.8 Some authors have found that vigilante organizations are more prevalent in communities with indigenous legal frameworks (Owumi and Ajayi 2013; Vilas 2009), cultures of honor and violence (Clarke 1998; Hayes and Lee 2005; Tyson 2013; Wolfgang and Ferracuti 1967), and historically inherited security structures (Bateson 2013; Osorio, Weintraub, and Schubiger 2016).9
Finally, a third strand of scholarship has seen vigilantism as a direct reaction to crime and insecurity. Seligson (2003) argues that fear of crime erodes citizens’ support for democracy and can increase their support for vigilantism. For their part, Ungar (2007) and Smulovitz (2003) propose that vigilantism emerges as an attempt by citizens to reestablish order within their community. Finally, Huggins analyzes vigilantism as a way by which citizens defend themselves from crime in the absence of alternative resources (Huggins 1991, 2000). Consistent with this view, Malone (2012), Bateson (2012), Hinton et al. (2014), and others have found that citizens exposed to crime directly or indirectly are more likely to support vigilante justice and other drastic control measures.
Although these studies provide a strong theoretical and empirical platform from which to investigate citizens’ willingness to engage in anti-crime organization attempts, they are not without their limitations. First, most studies have focused on in-depth investigation of specific cases. While this strategy is useful for exploring the contexts in which extralegal violence occurs, it is of limited use for investigating why some individuals are more likely than others to engage in anti-crime organization attempts. To do so, it is necessary to compare across cases.
Second, quantitative studies investigating vigilantism have largely analyzed media reports of spectacular displays of collective violence conducted with the manifest purpose of deterring or punishing crime (e.g., Hepworth and West 1988; Mendoza 2006). Although some of these events may involve anti-crime organization attempts, they are only imperfect proxies. On the one hand, some of these actions may reflect socially or politically motivated mob violence, only interpreted or rationalized as anti-criminal after the fact. On the other hand, even when vigilantes face incentives to engage in lethal violence (see Hine 1998), not all anti-crime organizations act violently. Thus, it is hard to know the extent to which analyzing media reports of lynchings (and other spectacular displays of collective violence) can help us understand anti-crime organization more generally.
More recently, some studies have turned to media reports of vigilante organizations (e.g., Osorio, Weintraub, and Schubiger 2016; Phillips 2015). However, not all anti-crime organization attempts crystalize into formal and stable organizations. Many of these attempts find little resonance within the larger community or result only in informal or ad hoc anti-criminal groups. Of the few that succeed, not all engage in crime-control actions or become violent, and even fewer engage in an act of sufficient notoriety to be reported by the media.10
Finally, to date, there is little consensus about the criteria for classifying an organization as “vigilante.” It is unclear whether we should look at intentions, actions, or effects as indicators of vigilantism, or how extensively or consistently these indicators should be present before a successful classification can be made. To mitigate the risks of this multilevel selection process, it is necessary to look at firsthand accounts of citizens’ behavior.
Although anti-crime organizations can vary in their structures, methods, and specific objectives, they frequently emerge from earlier (harmless and unorganized) attempts by individuals to fight crime as a collective. Thus, it is critical to understand the incentives (i.e., risks and opportunities) for citizens attempting to fight crime with their neighbors.
The first set of incentives is noncontingent, or independent of citizens’ success in fighting crime. Simply by broadcasting their intention to confront criminals, for example, citizens may raise wrongdoers’ perceived likelihood of being captured and, thus, deter some crime.11 Further, by engaging in anti-criminal action and standing up to perceived wrongdoers, citizens may experience positive emotions (de Quervain et al. 2004; Maitner, Mackie, and Smith 2006). Finally, just by broadcasting their intention to fight crime directly, citizens may feel gratified because they have expressed their nonconformity with the insecurity in their communities (Gollwitzer and Denzler 2009).
Conversely, citizens may experience noncontingent costs of anti-criminal action. First, citizens may have to invest valuable economic resources to mobilize their community. Second, they may need to spend time, effort, and resources that they could invest in more productive activities (Bates, Greif, and Singh 2002). Third, although some may experience a thrill from confronting wrongdoers, others may experience a diversity of negative emotions. Just days after the start of the anti-criminal movement in Cherán, for example, one of the leaders explained, “We are experiencing ire, anger, fear, dread, and rage. But above all, we are experiencing a sentiment of humiliation for the inaction of the authorities.”12
Although it is important to consider these consequences, it is critical to note that the biggest risks and opportunities of anti-criminal action are contingent. That is, they depend on whether citizens are successful in fighting crime with impunity. Citizens who successfully fight crime stand to benefit from their actions. Citizens who rally their community to fight crime effectively might experience a sense of satisfaction from punishing wrongdoers (de Quervain et al. 2004). Further, in contexts of high impunity, citizens may experience gratification from restoring what they see as justice (Lerner 1980). Further, citizens stand to recover lost income and property and achieve an even larger deterrent effect if they can confront criminals successfully. That is, they stand to gain from a reduction in the material and immaterial costs of crime.13
In contrast, those who unsuccessfully engage in anti-criminal actions may incur significant costs. First, citizens who do not succeed in fighting crime directly may unintentionally increase the future costs of crime. Kidnappers often kill their hostages after a failed rescue attempt, and many extortion rackets in Mexico have increased their rates after failed self-defense attempts (Scandizzo and Ventura 2015). Second, those engaging in anti-criminal action may suffer physical, emotional, and psychological harm. The Michoacán autodefensas, for example, were not always successful, and whenever they lost territory to the drug cartels they had to endure humiliation and the loss of members and leaders (Gómez et al. 2014). Finally, since many forms of anti-criminal action are considered illegal, even if successful in confronting crime, citizens risk duplicating official law enforcement efforts or incurring legal sanctions. The Mexican authorities have overtly condemned vigilante justice, and many vigilante leaders have been jailed or killed by federal forces (Cruz 2015).14
All in all, a significant portion of the risks and opportunities associated with anti-criminal action are contingent on citizens effectively fighting crime with impunity. Since this depends on the actions of other agents, it is important to consider citizens’ perceptions of others as a potential determinant of their willingness to participate in anti-criminal action. In the following section I focus on two actors: the community and law enforcement authorities.
The first factor that may influence an individual’s expectation of benefiting from organizing with others to fight crime is her perception that her community is united and trustworthy. First, well-connected and united communities provide citizens with a social scaffolding in which resources and information can be mobilized (Nahapiet and Ghoshal 1998; Narayan and Pritchett 1999). Citizens who perceive their community as strong and united may expect that news of their attempts at organizing others to fight crime will spread rapidly and resonate with their neighbors. Conversely, those who perceive their community as weak and divided may have to face criminals with the help of only a handful of neighbors. Thus, there is a greater probability that they will suffer the physical, emotional, and psychological harm discussed previously.
Second, trustworthy neighbors are more likely to interact with individuals pro-socially. Therefore, when citizens perceive their neighbors as trustworthy, they may anticipate experiencing positive emotions from engaging in effective collaboration with them (Jung et al. 2002). On the contrary, if citizens perceive their community as untrustworthy, they may anticipate frustration and other negative emotions from failing to mobilize their neighbors. Considering citizens’ perceptions of the trustworthiness of their neighbors may be important to identify the individuals who are most likely to reach out to their community to fight crime.
Many studies have considered community trust as an important determinant of collective action. Citizens who perceive others to be more trustworthy may be more likely to take a collective approach to problem solving (Nahapiet and Ghoshal 1998), be less fearful of free riders, be more likely to perceive that they will benefit from collective action (Ostrom, Walker, and Gardner 1992), and be more collectively efficacious (Welch et al. 2001; Putnam 1995).15 However, most of the literature has made strong normative assumptions about the nature of social trust. I join a new group of scholars challenging this assumption by considering social trust as a motor that propels collective action but that, depending on the context, can have mixed implications for the state and society.16 In what contexts are citizens more likely to invest their social capital outside the established institutions of the state?
Since the first works in modern political thought, the state has been seen as more than a repository of symbolic representation. For a long time, authors from very different backgrounds have conceded that, beyond its role as a representative, the state plays an important role as a social moderator (e.g., Hobbes 1651; Locke 1689), that is, as a force that structures the contexts in which citizens can exercise their liberty to pursue their individual interests at the expense of the common good. Building on this notion of the state, a number of scholars have proposed (and identified empirically) citizens’ distrust in state law enforcement to be an important predictor of state reliance (or a lack thereof).17
Undoubtedly, there are reasons to expect citizens to delegate to the authorities when they see them as trustworthy. When the likelihood of state authorities intervening effectively is high, citizens can expect to obtain a positive outcome from reporting crime and to incur legal costs if they take the law into their own hands. Since appealing to law enforcement authorities is less costly and risky than confronting criminals directly, ceteris paribus, as the probability of effective police intervention increases, the relative utility of citizens taking the law into their own hands decreases.
Trust in state authorities has been found to be a significant determinant of citizens’ willingness to support harsh anti-criminal punishment (Soss, Langbein, and Metelko 2003), report crime (Messing et al. 2015; Malone 2012), and uphold and collaborate with law enforcement authorities to fight and prevent crime (Hough et al. 2010; Tyler and Fagan 2008).18 In the midst of police reform in Latin America, citizens who trust the police have also been observed to be more likely to participate in programs in which the state authorities invite local citizens to help them prevent, monitor, denounce, or persecute crime (Arias and Ungar 2009; Sabet 2014).
Distrust in law enforcement seems to set the conditions in which citizens are more likely to seek alternative sources of criminal justice. As Diamond (1999, 91) famously noted, “In the context of weak states and inefficient, poorly disciplined police, crime may inspire drastic, illegal, unconstitutional, and grotesquely sadistic responses to try to control it.” Thus, I propose that trust informs citizens’ willingness to engage in an anti-crime organization attempt on two levels. First, distrust in law enforcement sets the stage for citizens to consider alternative sources of criminal justice.19 Second, a citizen’s perception of the trustworthiness of her community informs her likelihood of turning to her neighbors to confront criminals directly.
Even though this interactive hypothesis has not been explicitly tested, previous scholarship has hinted at its general validity. Gamson (1968) and then Seligson (1980) proposed that trust in the state influences the degree to which collective efficacy translates into alienated or democratic activism. Similarly, in his in-depth study of lynchings in Guatemala, Mendoza (2006) suggests that solidarity links among co-ethnics can trump barriers to collective action, especially in areas in which the state has traditionally failed to provide citizens with basic security services.
Along the same lines, in her work in Central America, Malone (2012, 117) notes, “in order for the perceived failure of the justice system to translate into collective action, citizens would need to have some sense of solidarity with other members of their community and view citizen action as a viable means for achieving their goals.” In his in-depth investigation of the Mano Amiga program in Honduras, Ungar (2009) found that, faced with untrustworthy law enforcement authorities, a number of citizens invested their newly acquired social resources into the constitution of vigilante organizations. This led to an unintended increase in vigilantism which, in turn, prevented the program from resulting in a marked reduction in violence.20
In the following section, I test the moderating effect of distrust in law enforcement over the effect of social trust. I do so using a rarely utilized data source that has collected information about citizens’ anti-crime organization attempts over the last ten years: ENVIPE.
Since 2006, Mexico has experienced one of its most serious security crises in modern history. In December of that year, the then newly elected president Felipe Calderón sent some 6,500 Mexican soldiers to the state of Michoacán to start the so-called war on drugs. Since then, Mexicans have faced an increase in the militarization of security (Anaya 2014) and the aggressiveness of drug cartels (Calderón et al. 2015; Phillips 2015). Consequently, there has been a rapid rise in homicides, kidnappings, and other acts of violence (Atuesta, Siordia, and Madrazo Lajous 2016; Schedler 2015; Shirk and Wallman 2015).
Kidnappings increased from 2,687 (2001–2006) to more than 6,500 (2006–2012), and summary executions climbed from fewer than 1,000 per year prior to 2006 to nearly 13,000 in 2011 (Atuesta 2016). In total, drug-related homicides jumped from fewer than ten thousand between 2000 and 2006 to more than fifty thousand during the first phase of the war on drugs (2006–2012)—more than a fivefold increase (Atuesta, Siordia, and Madrazo Lajous 2016; Rios 2013). As a consequence of this increase in violence, since 2010 Mexico’s National Institute of Geography and Statistics (INEGI) has conducted annual nationally representative surveys (n ≈ 65,000 each) that recover information about citizens’ perceptions of insecurity and crime victimization histories.21
Crucially, in addition to questions on these topics, the 460,329 respondents that participated in the six waves of ENVIPE were asked about their trust in neighbors, trust in the police, and their participation in anti-crime organization attempts.22 On this topic, INEGI asked respondents: “In the previous year, did this household take joint action with its neighbors to defend itself from crime [Yes or No]?”Although this question does not allow us to identify actions involving extralegal violence, it gives us a unique opportunity to study Mexicans’ willingness to turn to their community to attempt to prevent or control crime at the individual level. Importantly, it is possible to use this information to compare how frequently individuals with different characteristics (including trust in neighbors and the government) attempt anti-crime organization, while holding other factors constant.
Overall, amid the context of the ongoing war on drugs, 42.10 percent of Mexicans perceive the neighborhood in which they live as unsafe, and 32.9 percent of Mexican households report being victims of crime in an average year. This is particularly concerning if one considers that this number is double the average for the Latin American and Caribbean region (between 17 percent and 19.2 percent, according to Hinton et al. 2014). In this context, Mexicans tend to be much more trusting of their neighbors than of the police. On a scale of 0 to 100, Mexicans’ trust in their neighbors is about 64.21. In contrast, on this same scale, citizens’ trust in the police is only about 40.5. Not surprisingly, about 10.49 percent of citizens take joint action with their neighbors to defend themselves from crime every year, and this number increased from about 7.9 percent in 2011 to about 14.4 percent in 2015. This represents an 80 percent increase.23
As suggestive as these statistics may be, to evaluate whether citizens’ trust in their neighbors influences their participation in anti-crime organization attempts and whether this link is moderated by their distrust in the police, it is necessary to take a closer look at individual-level data. To do so, I specify a logistic regression model in which the probability of a citizen engaging in anti-crime organization is a function of the following equation:
In this equation, the term Trust N represents a variable that captures citizens’ trust in their community, Distrust P represents a variable that captures citizens’ distrust in the police, and Trust N × Distrust P represents the product of these two variables. This last term is included to estimate the degree to which distrust in the police modifies the effect of trust in neighbors. If the interactive trust hypothesis is consistent with this data, we should expect a positive and significant coefficient associated with this term.
The term ϕ′ CONTROLS′ represents a vector of control variables. This vector is composed of demographics (sex, education, age, size of the locality, and occupation) and other variables the literature highlights as important determinants of anti-crime organization.24
First, I control for citizens’ perceived and actual exposure to crime and insecurity. Both variables have been shown to be negatively associated with social trust (Bateson 2012; Corbacho, Philipp, and Ruiz-Vega 2015) and trust in law enforcement authorities (Blanco 2013; Blanco and Ruiz 2013; Corbacho Philipp, and Ruiz-Vega 2015; Malone 2010), but positively associated with citizens’ attitudes toward vigilantism (Bateson 2012; Malone 2012; Seligson 2003). Thus, these variables could negatively bias our results. To be precise, if those who trust their neighbors are less likely to be exposed to crime and more likely to support vigilantism, our regression coefficient could fail to find a relation when one exists. To account for this potential source of bias, I control for crime victimization, for an index of perceived insecurity, and for an index of exposure to crime in the neighborhood.25
Second, I consider whether respondents are embedded in communities with a strong indigenous presence. Previous studies have hypothesized a link between these contexts and citizens’ willingness to participate in anti-crime organization. In areas where strong traditional justice systems exist, the argument goes, citizens may be more inclined to turn to their community to apply what they consider fairer and more legitimate criminal sanctions (Clunan and Trinkunas 2010; Vilas 2001).26 Given this argument and the stylized observation that violent anti-criminal events have been most frequently reported in traditionally indigenous regions (Hawkins, McDonald, and Adams 2013; Vilas 2009; Mendoza 2006), I control for the proportion of citizens self-identifying as indigenous in the interviewee’s municipality (from the 2015 inter-census survey). Although imperfect, this variable (municipal % indigenous) allows us to account for some of the variation associated with the diversity of indigenous cultural systems in Mexico.27
Finally, I control for the subnational context in which the respondents are embedded. This is important, since different states have experienced different levels of violence. While some states, such as Guerrero and Michoacán, are very conflictive, states such as Yucatán, Campeche, and Aguascalientes tend to be more peaceful. Further, while some years have been particularly bloody, others have gone by with relatively less violence. To achieve this level of control, I include a vector of 192 dummy variables (θ′ STATEYEAR′) that identify each state-year in which the study was conducted.28
Table 1 shows the results of estimating the model described earlier. First, I find that middle-aged and educated individuals are more likely to seek out their neighbors to fight crime, holding other variables relating to insecurity, violence, and trust in the authorities constant. Second, in line with the previous literature, I find that individuals who have been personally victimized by crime, who perceive a high level of insecurity, and who live in crime-ridden neighborhoods are more likely to engage in anti-crime organization attempts, holding other variables constant.29
|Trust in neighbors||0.652***|
|Distrust in the police||–0.445***|
|Trust in neighbors × distrust in the police||0.443***|
|Crime in the neighborhood||0.619***|
|Size of town (urban vs. rural)||–0.0794**|
|Size of town (semiurban vs. rural)||0.270***|
|Municipal % indigenous||–0.0295|
|Fixed effects||State year (192)|
In substantive terms, increasing an average citizen’s perception of insecurity from the minimum to the maximum is associated with a 2.5 percent increase in her probability of engaging in an anti-crime organization attempt. Moving this citizen from a household that has been victimized by crime to one that has not is associated with a 4.7 percent increase in the probability of her engaging in this type of behavior. And moving an average citizen from a household victimized by crime from an extremely safe to an extremely dangerous neighborhood is associated with a 6.78 percent increase in her probability of reaching out to her neighbors to fight crime. However, I do not find support for the sociocultural hypothesis. That is, I find that an average citizen living in a majority-indigenous municipality is not significantly more likely to engage in an anti-crime organization attempt than one living in a minority-indigenous municipality.
Although Table 1 shows significant coefficients associated with trust in neighbors and distrust in the police, these effects are not directly interpretable. Because of the inclusion of the interaction between both variables in the model, it is necessary to fix one variable to its mean before estimating the effect of the other. When doing this, I do not find the effect of citizens’ distrust in the police to be statistically distinguishable from zero. However, I find citizens’ trust in their neighbors to have a significant and strong effect on their likelihood to engage in an anti-crime organization attempt.
As Figure 1 shows, when distrust in the police is set to its mean, a minimum to maximum change in an average citizen’s trust in her neighbors is associated with an 8.3 percent change in her probability of engaging in an anti-crime organization attempt. Moreover, consistent with the idea that citizens’ distrust in the authorities activates the link between these two variables, I find a positive and significant interaction between police distrust and citizens’ trust in neighbors (in bold in Table 1). In other words, I find that as respondents’ distrust in the police increases, so does the effect of trust in neighbors. Figure 2 illustrates this moderating effect more closely. The vertical axis in the figure displays the estimated min-max effect of trust in neighbors, and the horizontal axis displays an average citizen’s distrust in the police. The confidence intervals in the figures (the shaded area) account for the complexity of the ENVIPE samples.
As Figure 2 shows, among citizens with a low level of distrust in the police, a min-max change in an average individual’s trust in neighbors is associated with a 6.5 percent change in her probability of engaging in an anti-crime organization attempt. Figure 2 shows that among those with a high level of distrust in the police, a maximum change in an average individual’s trust in neighbors is associated with a 9.3 percent change in her probability of engaging in an anti-crime organization attempt. This represents an increase of 43.4 percent in the effect of trust in neighbors.31
Taken together, these results are consistent with the common observation that anti-crime organization tends to emerge in crime-ridden and insecure neighborhoods. They show that individuals are more likely to engage in anti-crime organization attempts when they perceive more insecurity and live in a neighborhood where crime occurs more frequently. The latter factor carries the strongest effect. This is perhaps due to citizens considering it more likely that others will join in with their anti-crime organization attempts when they share the costs of crime.
Further, they show that, when these contextual variables are held constant, individuals who possess the most educational, physical, and social resources are most likely to turn to their neighbors to confront crime. Indeed, middle-aged individuals with high levels of education and trust in neighbors, in semirural contexts with high insecurity, seem to be the most likely to engage in an anti-crime organization attempt. The effect of increasing an average citizen’s trust in neighbors from the minimum to the maximum equates to moving her from a neighborhood in which gunshots, extortions, assaults, or kidnappings do not occur to one in which they do.32
Finally, although I do not find distrust in the police to have a significant effect for the average citizen, in line with the interactive hypothesis, I find this variable to have a significant moderating role. That is, I find distrust in the police to increase the positive effect of citizens’ trust in neighbors.33 While trust in neighbors has a positive effect among both sectors, it translates most strongly into an increased likelihood to engage in an anti-crime organization attempt among Mexicans who distrust the police.34
This article has sought to contribute to our understanding of anti-crime organization by empirically investigating the factors that influence individuals’ likelihood to reach out to their neighbors to defend themselves from crime. I have argued that, all other things being constant, citizens’ likelihood to engage in an anti-crime organization attempt is influenced by their evaluation of their neighbors. Further, I have argued that the effect of this variable is exacerbated in contexts in which citizens distrust law enforcement bodies. To test these ideas, I analyzed individual-level data that INEGI has collected over the last ten years.
Consistent with the framework presented earlier, I find evidence that citizens who trust their neighbors more are more likely to join them in an attempt to defend themselves from crime. Further, I find this effect to be stronger among citizens who distrust the police. More broadly, the theoretical approach presented in this article is supportive of Gamson’s (1968) view that trust in the government moderates the effect of citizens’ collective efficacy. However, it also pushes this framework forward by theorizing about the way citizens’ distrust in state authorities moderates their intra- and extra-state investments. All in all, the evidence presented here suggests that when citizens lose their confidence in law enforcement authorities, they are more likely to engage in what Hirschman (1978) once described as an “exit strategy” and invest their social resources in an attempt to resolve their most urgent problems by themselves.
It is true that many community efforts to replace the state can be preventive, innocuous, or go completely unnoticed. However, citizens’ attempts to defend themselves from crime can sometimes become active and violent. They can infringe on the alleged criminal’s individual rights, carry with them important opportunity costs, and in extreme cases shake the very foundations of the state. Even if this article is not directly connected to these outcomes, it tries to contribute to an empirical platform through which the emergence of established, violent crime-control organizations can eventually be understood by shedding light on the determinants of early individual manifestations of this phenomenon.
2The Mexican government’s monopoly on violence has also been eroded by citizens hiring private security firms. In complementary analyses (available upon request), I find support for the interactive trust hypothesis when predicting citizens’ likelihood to organize with neighbors to hire private security for the public space.
4This vigilante organization had a presence spanning 3,842 square miles (Martínez-Elorriaga 2014). In other Latin American countries, citizens have also attempted to mobilize their neighbors to confront crime directly. Indigenous communities have started community police forces (Cendejas, Arroyo, and Sánchez 2015; Handy 2004), and farming groups have gone so far as to initiate armed militias (Martínez-Elorriaga 2014; Mazzei 2009).
5Zizumbo-Colunga (2015) documents how the movement in Cherán emerged from a range of anti-crime organization attempts that preceded the 2011 uprising.
6Although there is evidence that group violence can increase in times of demographic, political and economic change, this approach has not gone unchallenged. Green, Glaser, and Rich (1998) find that the connection between negative economic perceptions and aggressive impulses against minorities decays quickly in the absence of political actors affixing economic blame on target groups. Further, in their analysis of lynching victims in Mexico, Rodriguez Guillén and Mora Heredia (2005) find that lynching victims tend to be young male co-ethnics living in the same place as their victimizers. Thus, even if insightful, this perspective has had limited success in explaining lynching in contexts in which group conflict is not politically or socially salient.
7It is important to note that, so far, group conflict and pure sociology have most often been applied to the study of lynching. While some lynchings may have their origins in anti-crime organizations, many others do not. Much theoretical and empirical work needs to be done before linking the two phenomena. Still, nothing about these theories limits their scope to just lynching.
8Note, however, that it is still uncertain to what extent this argument can be applied to urban vigilantism (Castillo-Claudett 2000) or vigilante actions in countries whose populations are not primarily indigenous (Martins 1991; Koonings and Kruijt 2004). Further, even if vigilantism correlates with preexisting informal institutions, it is necessary to study the more proximal variables driving this relation. Finally, although some quantitative studies validate this perspective (Mendoza 2006), others have challenged the mechanisms linking indigenous culture and violence (Fernandez 2004; Handy 2004; Vilas 2009).
9Bateson (2013) contends that the lynchings in Guatemala emerged from the rural patrols established during the country’s civil war. Osorio, Weintraub, and Schubiger (2016) use the location of the Cristero guerrillas as an instrument for armed self-defense groups in Michoacán.
10In Latin America, most anti-crime organizations are likely to be local groups that get together for a specific crime-prevention project. Groups that move forward into crime-control initiatives are scarce, and groups that engage in violent crime control are even scarcer. Using reports of violent crime-control organizations to understand the emergence of anti-crime organization attempts is likely to yield heavily biased results.
11As Hine (1998) points out, the violence of an anti-criminal action multiplies its deterrent effect. Thus, more violent anti-crime organization can be expected to be a stronger deterrent against crime. Having said this, unless criminals can expect to be completely unharmed, even the least violent anti-crime organization attempts will increase criminals’ expected risks.
12“El pueblo de Cherán, Michoacán recurre a la autodefensa para parar la violencia,” Regeneración Radio, May 11, 2011, http://www.regeneracionradio.org/index.php/represion/abuso/item/2903-el-pueblo-de-cher%C3%A1n-michoac%C3%A1n-recurre-a-la-autodefensa-para-parar-la-violencia. Translation is my own.
14It is important to note that not all types of anti-criminal action are persecuted. Authorities often tolerate anti-crime organizations that involve low levels of violence. Thus, citizens organizing in nonviolent ways face a low risk of being caught by the authorities. In contexts in which peaceful anti-crime organization is the dominant way vigilantes confront crime, the role of distrust of the police (which I highlight in the next section) may be reduced.
16A growing number of studies have found that, under certain circumstances, social capital can lead to outsider exclusion (Waldinger 1995), corruption (Graeff 2010), the emergence and consolidation of gangs and militias (Bourgois 1996; Tarrow 1998), participation in pro-authoritarian movements (Armony 2004; Satyanath, Voigtlaender, and Voth 2013), and other outcomes that can put democracy at risk.
17Note, however, that other variables have also been linked to state reliance. On wealth and crime reporting, see Goudriaan, Wittebrood and Nieuwbeerta (2006) and Malone (2012). On gender and crime reporting, see Tarling and Morris (2010) and Malone (2012). On age, migratory status and crime reporting, see Malone (2012), and Gutiérrez and Kirk (2017). Finally, on the relation between crime severity and crime reporting, see Tarling and Morris (2010).
18Malone (2012) finds an association between distrust in the judicial system and citizens’ likelihood of denouncing crime in Central American countries with a low level of equilibrium.
19Note, however, that there are also mechanisms by which anti-criminal actions could be directly and positively associated with an individual’s trust in the authorities. First, in line with Merolla and Zechmeister’s findings (2009), it is possible for the threat that arises from confronting crime directly to boost citizens’ perception of the strength and competence of the authorities. Second, authorities could have incentives to cultivate trust among anti-crime organizations to reestablish social order. Finally, citizens may understand trust as the likelihood of the authorities “looking the other way” after an attack on alleged criminals. Thus, it is hard to propose a unidirectional relation between an individual’s distrust in law enforcement and his/her willingness to engage in collective anti-criminal action.
20It is important to recognize that there is heterogeneity within the judicial system. In most communities, the state is represented by the local police; however, this actor is but a part of a larger criminal justice system. Other state officials (e.g., prosecutors, judges, etc.) can build trust independently of the police. Here, I analyze distrust in law enforcement as a single index because of the empirical correlation among the items (see the online appendix, section 2). Future studies may extend my contributions by analyzing the effect of different judicial actors separately.
22INEGI has measured trust in the police in different ways over the years. From 2010 to 2015, measurement was agency specific (i.e., was measured independently for municipal, federal, and state police agencies) and was asked only if the agency was recognized by the interviewee on a scale of 1 (a lot) to 3 (not at all). In 2010, INEGI included a direct agency-specific question in addition to the multistep agency-specific question. In 2011, in addition to these agency-specific questions, INEGI included a question on general distrust in the police for all citizens, scaled 1 (a lot) to 4 (none). I aggregated the questions in an index of general distrust in the police to minimize measurement error and missing data. The online appendix, section 2, shows the aggregation process.
25The index of perceived insecurity is composed of three variables: citizens’ perceptions of insecurity in their neighborhood, their municipality and their state. These three indicators of perceptions of insecurity are identically worded and are very close together in the questionnaire. Consequently, they load into a single factor with an eigenvalue of 1.29 and have an alpha reliability coefficient of 0.72. The index of crime in the neighborhood is composed of twelve indicators: citizens’ reports of public alcohol consumption, illegal alcohol sales, public drug consumption, shootings, gang activity, gun sales, Assaults, extortions, floor charges, drug stores, kidnappings, and/or piracy sales. Again, these twelve indicators of crime in the neighborhood are identically worded and are very close together in the questionnaire. Consequently, they load into a single factor with an eigenvalue of 2.37 and have an alpha reliability coefficient of 0.75. The wording of all these questions can be seen in the online appendix, section 1.
29Although perceptions of insecurity, crime in the neighborhood, and crime victimization are correlated with each other, I do not find that including them simultaneously introduces multicollinearity problems. First, I find the results to be robust to including the crime and insecurity variables independently. Second, I find the main results are robust to omitting all control variables.
30The statistical significance of the interactive term also suggests that trust in neighbors moderates the effect of distrust in the police. That is, I find that, at low levels of trust in neighbors, trust in the police increases a citizen’s chance of engaging in anti-criminal action. It is important, however, to take this finding with a grain of salt, as the maximum effect of trust in the police on anti-crime organization is never greater than 2.7 percent. A figure displaying this effect can be found in the online appendix, section 4.
31These findings resonate with what can be observed in the cases of Cherán and Tierra Caliente. The first was initiated (and led) by a group of public school teachers. The second was initiated and commanded by a medical doctor (Manuel Mireles).
32In complementary analyses (available upon request), I find these results to be robust to using citizens’ likelihood to organize with neighbors to hire private security for the public space as an alternative operationalization of the dependent variable.
33These analyses cannot disentangle the endogenous relation between citizens’ trust in neighbors and anti-crime organization. I find, however, that the results are robust to aggregating the independent variables at the municipal level and including them as lagged independent variables (analyses available upon request).
Daniel Zizumbo-Colunga is currently Assistant Professor at CIDE’s Drug Policy Program and a Research Assistant Professor at Vanderbilt University. He studies the political psychology of violence, discrimination, and corruption. In addition to his academic appointments he currently serves as director of CIDE’s Social Science Experimentation Unit (UECS). He is a principal investigator in the Program for Longitudinal Studies, Experiments and Surveys (PANEL) and is a frequent collaborator of the Latin American Public Opinion Project (LAPOP).
Anaya, Alejandro. 2014. “Violaciones a los derechos humanos en el marco de la estrategia militarizada de lucha contra el narcotráfico en México.” Cuadernos de Trabajo del Monitor del Programa de Política de Drogas, no. 4. Aguascalientes, Aguascalientes: Centro de Investigación y Docencia Económicas.
Arias, Enrique Desmond, and Mark Ungar. 2009. “Community Policing and Latin America’s Citizen Security Crisis.” Comparative Politics 41 (4): 409–429. DOI: https://doi.org/10.5129/001041509X12911362972476
Atuesta, Laura. 2016. “Un análisis de la evolución del crimen organizado en México a través de los narcomensajes.” Cuadernos de Trabajo del Monitor del Programa de Política de Drogas, no. 20. Aguascalientes, Aguascalientes: Centro de Investigación y Docencia Económicas.
Atuesta, Laura, Oscar S. Siordia, and Alejandro Madrazo Lajous. 2016. “La ‘Guerra contra las Drogas’ en México: Registros (oficiales) de eventos durante el periodo de diciembre de 2006 a noviembre de 2011.” Cuadernos de Trabajo del Monitor del Programa de Política de Drogas, no. 14. Aguascalientes, Aguascalientes: Centro de Investigación y Docencia Económicas.
Bates, Robert, Avner Greif, and Smita Singh. 2002. “Organizing Violence.” Journal of Conflict Resolution 46 (5): 599–628. DOI: https://doi.org/10.1177/002200202236166
Bateson, Regina Anne. 2012. “Crime Victimization and Political Participation.” American Political Science Review 106 (3): 570–587. DOI: https://doi.org/10.1017/S0003055412000299
Bisogno, Enrico, Jenna Dawson-Faber, Michael Jandl, Kristiina Kangaspunta, Labib Kazkaz, Lucia Motolinia-Carballo, Serena Oliva, and Felix Reiterer. 2014. Global Study on Homicide 2013. Vienna: United Nations Office on Drugs and Crime.
Blanco, Luisa R. 2013. “The Impact of Crime on Trust in Institutions in Mexico.” European Journal of Political Economy 32 (December): 38–55. DOI: https://doi.org/10.1016/j.ejpoleco.2013.06.004
Blanco, Luisa R., and Isabel Ruiz. 2013. “The Impact of Crime and Insecurity on Trust in Democracy and Institutions.” American Economic Review 103 (3): 284–288. DOI: https://doi.org/10.1257/aer.103.3.284
Calderón, Gabriela, Gustavo Robles, Alberto Díaz-Cayeros, and Beatriz Magaloni. 2015. “The Beheading of Criminal Organizations and the Dynamics of Violence in Mexico.” Journal of Conflict Resolution 59 (8): 1455–1485. DOI: https://doi.org/10.1177/0022002715587053
Cendejas, Josefina María, Omar Arroyo, and Angélica Sánchez. 2015. “Comunalidad y buen vivir como estrategias indígenas frente a la violencia en Michoacán: Los casos de Cherán y San Miguel de Aquila.” Revista Pueblos y Fronteras Digital 10 (19): 257–284. DOI: https://doi.org/10.22201/cimsur.18704115e.2015.19.53
Clarke, James W. 1998. “Without Fear or Shame: Lynching, Capital Punishment and the Subculture of Violence in the American South.” British Journal of Political Science 28 (2): 269–289. DOI: https://doi.org/10.1017/S0007123498000167
Cohen, Mark A. 2008. “The Effect of Crime on Life Satisfaction.” Journal of Legal Studies 37 (S2): S325–S353. DOI: https://doi.org/10.1086/588220
Corbacho, Ana, Julia Philipp, and Mauricio Ruiz-Vega. 2015. “Crime and Erosion of Trust: Evidence for Latin America.” World Development 70 (June): 400–415. DOI: https://doi.org/10.1016/j.worlddev.2014.04.013
De Quervain, Dominique J.-F., Urs Fischbacher, Valerie Treyer, Melanie Schellhammer, Ulrich Schnyder, Alfred Buck, and Ernst Fehr. 2004. “The Neural Basis of Altruistic Punishment.” Science 305 (5688): 1254–1258. DOI: https://doi.org/10.1126/science.1100735
Fong, Eric, and Ly-yun Chang. 2011. “Community under Stress: Trust, Reciprocity, and Community Collective Efficacy during SARS Outbreak.” Journal of Community Health 36 (5): 797–810. DOI: https://doi.org/10.1007/s10900-011-9378-2
Goldstein, Daniel M. 2004. The Spectacular City: Violence and Performance in Urban Bolivia. Durham, NC: Duke University Press. DOI: https://doi.org/10.1215/9780822386018
Goldstein, Daniel M. 2012. Outlawed: Between Security and Rights in a Bolivian City. Durham, NC: Duke University Press. DOI: https://doi.org/10.1215/9780822395607
Gollwitzer, Mario, and Markus Denzler. 2009. “What Makes Revenge Sweet: Seeing the Offender Suffer or Delivering a Message?” Journal of Experimental Social Psychology 45 (4): 840–844. DOI: https://doi.org/10.1016/j.jesp.2009.03.001
Goudriaan, Heike, Karin Wittebrood, and Paul Nieuwbeerta. 2006. “Neighborhood Characteristics and Reporting Crime Effects of Social Cohesion, Confidence in Police Effectiveness and Socio-economic Disadvantage.” British Journal of Criminology 46 (4): 719–742. DOI: https://doi.org/10.1093/bjc/azi096
Graeff, Peter. 2010. “Social Capital: The Dark Side.” In Handbook of Social Capital: The Troika of Sociology, Political Science and Economics, edited by Gert Tinggaard Svendsen and Gunnar Lind Haase Svendsen, 143–161. Northampton, MA: Edward Elgar.
Green, Donald P., Jack Glaser, and Andrew Rich. 1998. “From Lynching to Gay Bashing: The Elusive Connection between Economic Conditions and Hate Crime.” Journal of Personality and Social Psychology 75 (1): 82–92. DOI: https://doi.org/10.1037/0022-35126.96.36.199
Gutiérrez, Carmen M., and David S. Kirk. 2017. “Silence Speaks: The Relationship between Immigration and the Underreporting of Crime.” Crime and Delinquency 63 (8), 926–950. DOI: https://doi.org/10.1177/0011128715599993
Handy, Jim. 2004. “Chicken Thieves, Witches, and Judges: Vigilante Justice and Customary Law in Guatemala.” Journal of Latin American Studies 36 (3): 533–561. DOI: https://doi.org/10.1017/S0022216X04007783
Hayes, Timothy C., and Matthew R. Lee. 2005. “The Southern Culture of Honor and Violent Attitudes.” Sociological Spectrum 25 (5): 593–617. DOI: https://doi.org/10.1080/02732170500174877
Hepworth, Joseph T., and Stephen G. West. 1988. “Lynchings and the Economy: A Time-Series Reanalysis of Hoyland and Sears (1940).” Journal of Personality and Social Psychology 55 (2): 239–246. DOI: https://doi.org/10.1037/0022-35188.8.131.52
Hine, Kelly. 1998. “Vigilantism Revisited: An Economic Analysis of the Law of Extra-judicial Self-Help or Why Can’t Dick Shoot Henry for Stealing Jane’s Truck?” American University Law Review 47(5): 1221–1255.
Hinton, Nicole, Daniel Montalvo, Arturo Maldonado, Mason Moseley, and Daniel Zizumbo-Colunga. 2014. “Crime and Violence across the Americas.” In The Political Culture of Democracy in the Americas, 2014: Democratic Governance across 10 Years of the AmericasBarometer, edited by Elizabeth J. Zechmeister, 3–21. Nashville: Vanderbilt University.
Hirschman, Albert O. 1978. “Exit, Voice, and the State.” World Politics 31 (1): 90–107. DOI: https://doi.org/10.2307/2009968
Hough, Mike, Jonathan Jackson, Ben Bradford, Andy Myhill, and Paul Quinton. 2010. “Procedural Justice, Trust, and Institutional Legitimacy.” Policing 4 (3): 203–210. DOI: https://doi.org/10.1093/police/paq027
Jung, Insung, Seonghee Choi, Cheolil Lim, and Junghoon Leem. 2002. “Effects of Different Types of Interaction on Learning Achievement, Satisfaction and Participation in Web-Based Instruction.” Innovations in Education and Teaching International 39 (2): 153–162. DOI: https://doi.org/10.1080/14703290252934603
Koonings, Kees, and Dirk Kruijt. 2004. “Armed Actors, Organised Violence and State Failure in Latin America: A Survey of Issues and Arguments.” In Armed Actors: Organized Violence and State Failure in Latin America, edited by Kees Koonings and Dirk Kruijt. London: Zed Books.
Lerner, Melvin. 1980. The Belief in a Just World: A Fundamental Delusion. New York: Springer. DOI: https://doi.org/10.1007/978-1-4899-0448-5
Maitner, Angela T., Diane M. Mackie, and Eliot R. Smith. 2006. “Evidence for the Regulatory Function of Intergroup Emotion: Emotional Consequences of Implemented or Impeded Intergroup Action Tendencies.” Journal of Experimental Social Psychology 42 (6): 720–728. DOI: https://doi.org/10.1016/j.jesp.2005.08.001
Malkin, Elisabeth, and Paulina Villegas. 2014. “Enemies of Mexican Drug Gangs Pose a Security Challenge.” New York Times, January 13. http://www.nytimes.com/2014/01/14/world/americas/enemies-of-mexican-drug-gangs-pose-a-security-challenge.html.
Malone, Mary Fran T. 2010. “The Verdict Is In: The Impact of Crime on Public Trust in Central American Justice Systems.” Journal of Politics in Latin America 2 (3): 99–128. DOI: https://doi.org/10.1177/1866802X1000200304
Martínez-Elorriaga, Ernesto. 2014. “Autodefensas tienen presencia en la sexta parte del territorio de Michoacán: INEGI.” La Jornada, January 7. http://www.jornada.unam.mx/ultimas/2014/01/07/auto?defensas-tienen-presencia-en-la-sexta-parte-del-territorio-de-michoacan-inegi-9261.html.
Martins, José de Souza. 1991. “Lynchings—Life by a Thread: Street Justice in Brazil, 1979–1988.” In Vigilantism and the State in Modern Latin America, edited by Martha K. Huggins, 21–32. New York: Praeger.
Mazzei, Julie. 2009. Death Squads or Self-Defense Forces? How Paramilitary Groups Emerge and Challenge Democracy in Latin America. Chapel Hill: University of North Carolina Press. DOI: https://doi.org/10.5149/9780807898611_mazzei
McCollister, Kathryn E., Michael T. French, and Hai Fang. 2010. “The Cost of Crime to Society: New Crime-Specific Estimates for Policy and Program Evaluation.” Drug and Alcohol Dependence 108 (1–2): 98–109. DOI: https://doi.org/10.1016/j.drugalcdep.2009.12.002
Mendoza, Carlos A. 2006. “Structural Causes and Diffusion Processes of Collective Violence: Understanding Lynch Mobs in Post-conflict Guatemala.” Paper presented at the Latin American Studies Association meeting, San Juan, Puerto Rico, March 15–18.
Merolla, Jennifer L., and Elizabeth J. Zechmeister. 2009. Democracy at Risk: How Terrorist Threats Affect the Public. Chicago: University of Chicago Press. DOI: https://doi.org/10.7208/chicago/9780226520568.001.0001
Messing, Jill Theresa, David Becerra, Allison Ward-Lasher, and David K. Androff. 2015. “Latinas’ Perceptions of Law Enforcement Fear of Deportation, Crime Reporting, and Trust in the System.” Affilia 30 (3): 328–340. DOI: https://doi.org/10.1177/0886109915576520
Nahapiet, Janine, and Sumantra Ghoshal. 1998. “Social Capital, Intellectual Capital, and the Organizational Advantage.” Academy of Management Review 23 (2): 242–266. DOI: https://doi.org/10.5465/amr.1998.533225
Narayan, Deepa, and Lant Pritchett. 1999. “Cents and Sociability: Household Income and Social Capital in Rural Tanzania.” Economic Development and Cultural Change 47 (4): 871–897. DOI: https://doi.org/10.1086/452436
Osorio, Javier, Michael Lee Weintraub, and Livia Isabella Schubiger. 2016. “Vigilante Mobilization and Local Order: Evidence from Mexico.” Paper presented at the meeting of the Midwest Political Science Association Meeting, Chicago, IL.
Ostrom, Elinor, James Walker, and Roy Gardner. 1992. “Covenants with and without a Sword: Self-Governance Is Possible.” American Political Science Review 86 (2): 404–417. DOI: https://doi.org/10.2307/1964229
Owumi, Bernard, and Johnson Olusegun Ajayi. 2013. “Traditional Values, Beliefs and Reliance on Indigenous Resources for Crime Control in Modern Southwest Nigeria.” African Research Review 7 (1): 73–94. DOI: https://doi.org/10.4314/afrrev.v7i1.6
Phillips, Brian J. 2015. “How Does Leadership Decapitation Affect Violence? The Case of Drug Trafficking Organizations in Mexico.” Journal of Politics 77 (2): 324–336. DOI: https://doi.org/10.1086/680209
Putnam, Robert D. 1995. “Tuning In, Tuning Out: The Strange Disappearance of Social Capital in America.” PS Political Science and Politics 28 (4): 664–683. DOI: https://doi.org/10.1017/S1049096500058856
Rea, Daniela. 2013. “Ayutla de los libres, el lugar donde los civiles ejercen la justicia.” CNN México, February 1. Online edition. http://mexico.cnn.com/nacional/2013/02/01/ayutla-de-los-libres-el-lugar-donde-los-civiles-ejercen-la-justicia.
Rios, Viridiana. 2013. “Why Did Mexico Become so Violent? A Self-Reinforcing Violent Equilibrium Caused by Competition and Enforcement.” Trends in Organized Crime 16 (2): 138–155. DOI: https://doi.org/10.1007/s12117-012-9175-z
Sabet, Daniel. 2014. “Co-Production and Oversight: Citizens and Their Police.” Working Paper Series on Civic Engagement and Public Security in Mexico. Washington, DC: Woodrow Wilson International Center for Scholars.
Salgado, Agustín. 2014. “Vecinos de Xochimilco se organizan contra la delincuencia.” La Jornada, March 22. http://www.jornada.unam.mx/ultimas/2014/10/28/vecinos-de-xochimilco-se-organizan-contra-la-delincuencia-9993.html.
Satyanath, Shanker, Nico Voigtlaender, and Hans-Joachim Voth. 2013. “Bowling for Fascism: Social Capital and the Rise of the Nazi Party.” Working Paper 19201. National Bureau of Economic Research. DOI: https://doi.org/10.3386/w19201
Scandizzo, Pasquale L., and Marco Ventura. 2015. “Organized Crime, Extortion and Entrepreneurship under Uncertainty.” European Journal of Law and Economics 39 (1): 119–144. DOI: https://doi.org/10.1007/s10657-014-9479-3
Seligson, Mitchell A. 1980. “Trust, Efficacy and Modes of Political Participation: A Study of Costa Rican Peasants.” British Journal of Political Science 10 (1): 75–98. DOI: https://doi.org/10.1017/S0007123400002015
Senechal de la Roche, Roberta. 2001. “Why Is Collective Violence Collective?” Sociological Theory 19 (2): 126–144. DOI: https://doi.org/10.1111/0735-2751.00133
Shirk, David, and Joel Wallman. 2015. “Understanding Mexico’s Drug Violence.” Journal of Conflict Resolution 59 (8): 1348–1376. DOI: https://doi.org/10.1177/0022002715587049
Silke, Andrew. 2001. “Dealing with Vigilantism: Issues and Lessons for the Police.” Police Journal 74: 120. DOI: https://doi.org/10.1177/0032258X0107400204
Smulovitz, Catalina. 2003. “Citizen Insecurity and Fear: Public and Private Responses in Argentina.” In Crime and Violence in Latin America: Citizen Security, Democracy, and the State, edited by H. Hugo Frühling, Joseph S. Tulchin, and Heather Golding. Washington, DC Woodrow Wilson Center Press.
Soss, Joe, Laura Langbein, and Alan R. Metelko. 2003. “Why Do White Americans Support the Death Penalty?” Journal of Politics 65 (2): 397–421. DOI: https://doi.org/10.1111/1468-2508.t01-2-00006
Tadjoeddin, Mohammad Zulfan, and Syed Mansoob Murshed. 2007. “Socio-economic Determinants of Everyday Violence in Indonesia: An Empirical Investigation of Javanese Districts, 1994–2003.” Journal of Peace Research 44 (6): 689–709. DOI: https://doi.org/10.1177/0022343307082063
Tarling, Roger, and Katie Morris. 2010. “Reporting Crime to the Police.” British Journal of Criminology 50 (3): 474–490. DOI: https://doi.org/10.1093/bjc/azq011
Tarrow, Sidney. 1998. Power in Movement: Social Movements and Contentious Politics. New York: Cambridge University Press. DOI: https://doi.org/10.1017/CBO9780511813245
Tyson, Adam. 2013. “Vigilantism and Violence in Decentralized Indonesia: The Case of Lombok.” Critical Asian Studies 45 (2): 201–230. DOI: https://doi.org/10.1080/14672715.2013.792570
Ungar, Mark. 2009. “La Mano Dura: Current Dilemmas in Latin American Police Reform.” In Criminality, Public Security, and the Challenge to Democracy in Latin America, edited by Marcelo Bergman and Laurence Whitehead, 93–118. Notre Dame: University of Notre Dame Press.
Vilas, Carlos M. 2001. “(In)justicia por mano propia: Linchamientos en el México contemporáneo (Taking the Law into One’s Hands: Lynchings in Mexico).” Revista Mexicana de Sociología 63 (1): 131–160. DOI: https://doi.org/10.2307/3541204
Villalpando, Rubén. 2009. “Inaceptable, una policía comunitaria en Le Baron, dice diputado.” La Jornada, December 7. http://www.jornada.unam.mx/2009/07/12/estados/026n2est.
Waldinger, Roger. 1995. “The Other Side of Embeddedness: A Case Study of the Interplay between Economy and Ethnicity.” Ethnic and Racial Studies 18 (3): 555–580. DOI: https://doi.org/10.1080/01419870.1995.9993879
Welch, Michael, Robert Rivera, Brian Conway, Jennifer Yonkoski, Paul Lupton, and Russell Giancola. 2001. “Determinants and Consequences of Social Trust.” Sociological Inquiry 75 (4): 453–473. DOI: https://doi.org/10.1111/j.1475-682X.2005.00132.x