Staying Power: Strategies for Weathering Criminal Violence in Marginal Neighborhoods of Medellín and Monterrey

Abstract

Criminal violence in Latin American cities is increasing. Meanwhile, with urbanization, greater numbers of people are moving to cities and into the crossfire. What self-protection strategies do residents adopt to keep safe in violent cities? Drawing on qualitative data from Medellín, Colombia, and Monterrey, Mexico, we document the strategies residents use to stay safe. We synthesize insights from studies of civil war, criminal governance, and urban violence to construct an analytical framework to systematically catalog and name these strategies. We posit that the type of violence residents face—indiscriminate or targeted—influences the strategies they pursue. Responding to either the indiscriminate or targeted form, residents employ survival strategies to avoid, withstand, or confront violence. Our research underscores the centrality of agency for residents’ “staying power” amid urban violence.

Resumen

La violencia criminal en América Latina va en aumento. Mientras tanto, la urbanización conlleva a que un mayor número de personas se mude a las ciudades y al fuego cruzado. ¿Qué estrategias de autoprotección adoptan los ciudadanos para mantenerse a salvo en las ciudades violentas? A partir de evidencia cualitativa de Medellín, Colombia y Monterrey, México, documentamos las estrategias que los ciudadanos utilizan para mantener a salvo. Sintetizamos los hallazgos de estudios sobre guerras civiles, gobernanza criminal, y violencia urbana para construir un marco analítico que cataloga y nombra sistemáticamente dichas estrategias. Proponemos que el tipo de violencia que enfrentan los ciudadanos —indiscriminada o selectiva— influye en las estrategias que emplean. Respondiendo ya sea a formas indiscriminadas o selectivas, los ciudadanos recurren a estrategias para eludir, soportar, o confrontar la violencia. Nuestra investigación enfatiza la centralidad de la agencia en el “poder de permanencia” ciudadana en medio de la violencia urbana.

You do not need to go to the rural areas to study violence here. Here, the war is fought in the city.

—Monterrey resident, December 2016

Organized criminal violence is increasingly prevalent across Latin American cities. Such violence frequently produces homicide rates rivaling or exceeding warzone casualties (United Nations Office on Drugs and Crime 2019, 12). Organized criminal violence refers to large-scale, sustained violence carried out by criminal rather than political actors, often with the purpose of maintaining economic dominance over a given territory. In Latin American cities, crime groups combat one another and state forces in what are sometimes deemed “low-intensity conflicts,” “gang wars,” or “urban warfare.” In Mexico, violence between organized crime groups and the state has intensified since the late 1990s, frequently seizing major metropolitan areas. Meanwhile, residents of Honduras’s San Pedro Sula—recently deemed the “murder capital” of the world—have been plagued by organized crime violence since the early 2000s (Bacon 2019). Similar conditions are found in Rio de Janeiro, Medellín, and Caracas, among other Latin American cities. This pattern reflects trends in violent conflict more generally, which is increasingly urban (Staniland 2010; Lucchi 2010).

The prevalence of urban criminal violence is noteworthy because more people live in cities today than ever before. In 1950, 751 million of the world’s inhabitants resided in cities. As of 2018, that number was 4.2 billion. The United Nations predicts that 68 percent of the world will live in cities by the year 2050 (United Nations Department of Economic and Social Affairs 2019, xix, 9). These trends are undeniable in Latin America, “the most urbanized region of the developing world,” where nearly 80 percent of the population lives in cities (Economic Commission for Latin America and the Caribbean 2012).

These intersecting trends paint a clear picture. While rural violence persists and is significant, residents of many Latin American cities live in conditions of sustained criminal violence. ¹ The causes, consequences, and dynamics of criminal violence in Latin America are well studied, but we know surprisingly little about how urban residents manage everyday life amid violence. What strategies and practices do they adopt to survive in a violent city?

Drawing on field research in marginal neighborhoods of two violent cities—Medellín, Colombia, and Monterrey, Mexico—we begin to answer this question by documenting the self-protection strategies we observed residents employ to keep safe. By focusing on individuals, we uncover the quotidian self-protection strategies that dictate daily life in violent contexts but are overlooked by the literature’s emphasis on collective responses. From this, we then develop a theoretical framework to empirically catalog these strategies. We posit that the strategies residents utilize are shaped by the type of violence they face, either indiscriminate or targeted. The former is dangerous only so far as it is proximate, while criminal groups aim the latter at particular residents. In response to either type, residents draw on one of two self-protection repertoires to avoid, withstand, or confront the violence. To develop this framework, we synthesize insights from scholarship on criminal governance, urban violence, and civil war. Our framework helps us understand urban residents’ reactions to pervasive violence, their strategies to persevere, and the agency with which they act.

The article proceeds as follows. We first review existing research on individual self-protection strategies and identify our contribution to that body of knowledge. We then outline our framework for categorizing the survival strategies we observed in Medellín and Monterrey. Next, we describe the research context in those cities and detail our methodological approach. The remainder of the article presents the qualitative evidence used to develop our framework. We conclude by identifying questions for future research and outlining policy recommendations. ²

Existing research

How do individuals and communities cope with and respond to criminal violence in Latin America? Existing scholarship interrogates the causes, consequences, and governance dynamics of organized crime violence in the region (Arias 2017; Barnes 2017; Durán-Martínez 2018; Lessing 2018; Yashar 2018; Trejo and Ley 2020). The literature also considers citizen responses to that violence, including fleeing (Ríos 2014; Atuesta and Paredes 2016; Cantor and Rodríguez 2016; Hiskey et al. 2018; Bada and Feldmann 2019), electoral participation (Trelles and Carreras 2012; Ley 2018; Berens and Dallendörfer 2019; Córdova 2019), antiviolence civic engagement (Bell-Martin 2019), victims’ legal advocacy groups (Gallagher 2017), and the search for and memorialization of the disappeared (Karl 2014; Robledo 2018; Paley 2020). Other scholars focus on contesting criminal domination, including “everyday” acts of resistance (Moncada 2020, 710) and the conditions under which different forms of resistance emerge (Moncada 2021; Ley, Mattiace, and Trejo 2022). A subset of these centers on collective efforts to shield communities from violence, including vigilante self-defense groups (Phillips 2017; Zizumbo-Colunga 2019), public shaming, protest, and collective collaboration or confrontation (Auyero and Kilanski 2015b; Arias 2019).

Comparatively less is known about the self-protection strategies that individuals living amid criminal violence adopt to maneuver daily life. Several authors make important inroads in this regard. Villarreal (2015) observes middle- and working-class residents of Monterrey, Mexico, “armoring” (e.g., securitizing the home), “camouflaging” (e.g., concealing signs of wealth), “caravanning” (e.g., traveling in groups), and “regrouping” (e.g., reestablishing community ownership of public space). Zubillaga, Llorens, and Souto (2019) shed light on how mothers protect their families in peripheral neighborhoods of Caracas, Venezuela. Their strategies range from hiding families indoors to collaborating with criminal groups or confronting them with the social power of gossip, among others. Third, Auyero and Kilanski (2015a, 2015b) apply an “ethics of care” lens to a neighborhood in Buenos Aires, Argentina. They show that older family members protect themselves and younger generations by creating everyday routines and, counterintuitively, by teaching and using violence against loved ones to deter even greater harm. These family practices occur alongside many of the same strategies Villarreal (2015) observes.

Our knowledge of individual self-protection strategies during civil war is more developed. There are several typologies classifying these strategies, which may range from fleeing violence or hiding in the home to taking a neutral stance in the conflict to collaborating with an armed group or relying on an external humanitarian actor or peacekeeping force for safety (Barter 2012; Baines and Paddon 2012; Justino 2012; Jose and Medie 2015; Suarez 2017). In their review of such typologies, Suarez and Black (2014) identify four common categories of civilian self-protection during civil war: fleeing violence or hiding from it, submission to or cooperation with armed groups, contesting violence by rousing public opinion and awareness, and confronting armed groups through collective action. Suarez (2017) and Justino (2012) further these typologies by emphasizing that civilian strategies emerge in response to different wartime threats rather than at random. In the eastern Democratic Republic of the Congo, Suarez documents how civilians first assess different types of threats, then negotiate with and sometimes deceive armed groups based on those assessments. Likewise, Justino (2012) finds that variation in violent actions—recruiting or looting, for instance—informs the survival strategies civilians are likely to employ in civil wars. Collective efforts to protect communities from violence also exist. In response to armed groups in Colombia’s civil war, scholars observe civilians fleeing en masse (Steele 2017), as well as using collective bargaining (Arjona 2016; Kaplan 2017) and strategic non-cooperation (Masullo 2021).

Scholars have observed that many approaches to civilian protection during war emphasize external, noncivilian actors, including the state, humanitarian agencies, and international peacekeeping forces, while neglecting the agency and knowledge of those living amid violence (Bonwick 2006; Metcalfe-Hough 2019, 1; Autesserre 2021). As Bonwick (2006) argues, such approaches assume that civilian protection is “done to” civilians who are conceived as “passive recipients” of external help rather than progenitors of their own protection and knowledge about wartime survival (270).

Our contribution: Self-protection strategies to endure urban violence

Our research builds on and advances existing knowledge in four ways. First, we draw attention to a significant but understudied phenomenon, individual- and household-level self-protection mechanisms to survive urban criminal violence. While self-protection is documented in civil war, criminal violence differs in important ways that we would expect to shape the form self-protection takes, including the motive for violence, how conflict is resolved, the organization of violence, and territorial governance arrangements (Kalyvas 2015). Second, we advance the few existing works on individual self-protection amid criminal violence by providing additional evidence from the urban periphery and a new theoretical framework through which to better understand self-protection. Our framework improves upon existing accounts by considering the conditions under which specific self-protection strategies develop. Rather than treat criminal violence as uniform, we identify two types of violence—indiscriminate and targeted—and consider which particular repertoires of self-protection mechanisms are likely to emerge in response. Third, we direct scholarly focus toward individual agency in the face of violence. Although much has been written on organized crime in Latin American cities, the agency of those affected by this violence has received far less attention. In line with scholars of civil war, we find that residents exercise choice and creativity over whether and how they respond to violence. Urban residents are not passive bystanders but engaged participants in their own protection in irrefutably difficult conditions. Finally, a common strategy in the face of violence is migration. ³ Our research, in contrast, centers on the survival strategies of those who remain.

Staying safe in the city: Analytical framework

How can we make sense of patterns of self-protection during urban violence? The literature points to context-level factors like local political economies (Moncada 2021), state negligence (Auyero and Kilanski 2015b, 401), and level of gang dominance (Córdova 2019) as constraining or creating opportunities for residents to respond to violence (see also Mac Ginty 2014). In this same vein, we document self-protection strategies that emerge in contexts of marginalization, state absence or capture, and where individuals hold little clout with gangs. First, these strategies develop among the marginalized, meaning persons “relegated to the peripheries of society because they lack social, political, or economic power” (Bell-Martin and Marston 2021, 160). Marginalized individuals rely on the survival strategies we document because they do not have the same resources to manage violence as other urban residents, such as money to hire private security, political clout to solicit state assistance, or the finances and external social connections to migrate safely. ⁴ In other cases, individuals want to stay despite the violence due to social ties or other connections to the neighborhood. Second, these self-protection mechanisms develop when collaboration with the state is not a viable strategy because it is absent or corrupt. Although Arias (2019) shows that entire communities can collaborate with state entities to resist or control criminal groups, we find that this option is not possible for individuals because they often face reprisals from gangs for working with the state.

Finally, power dynamics between residents and criminal groups shape the context in which self-protection mechanisms emerge. Arias (2019) shows that communities can broker peace with gangs, particularly when control is divided between gangs and the state, thereby granting the community some power relative to the gang. Marston (2020) similarly shows that individuals can leverage social ties to gang members to protect at-risk community members from forced displacement. Building on these, we find that the range of survival strategies available to residents widens where they hold some clout over or an existing relationship with crime groups. In these cases, the gang may look out for particular residents, or residents may be able to negotiate a reprieve from violence. In contrast, where residents have little sway over criminal groups, the range of survival strategies is limited to what an individual can achieve without gang or state intervention.

In this context, what explains the choice to use one self-protection strategy over another? Scholars posit that the type of violence wielded against a population shapes civilian responses (Kalyvas 2006; Justino 2012; Jose and Medie 2016; Steele 2017; Berens and Dallendörfer 2019). We likewise argue that the type of violence urban residents face informs the self-protection strategy they pursue. Drawing on insights from Steele (2017), we define two categories of violence, indiscriminate and targeted. ⁵ Indiscriminate violence occurs when armed groups or state forces haphazardly produce violence, for instance, when civilians are inadvertently caught in the middle of clashes between rivals (Steele 2017, 24). Individual civilians are generally not the object of indiscriminate violence. Thus, their risk of falling victim is defined by their physical proximity to violence rather than any personal attribute. Targeted violence occurs when armed groups or state forces aim violence at specific residents, often over (perceived) grievances, such as collaborating with a rival (Steele 2017, 23). In contexts controlled by organized crime, residents may also be targeted for their wealth or for engaging in behaviors the criminal group prohibits, like domestic abuse or theft. In sum, whereas the risk of experiencing indiscriminate violence is defined by where residents are, the risk of targeted violence is defined by who they are or their actions (Steele 2017, 27).

We identify three categories of self-protection strategies: avoiding, withstanding, and confronting. Avoiding refers to strategies individuals wield to keep away from violence or preclude violence from emerging; these strategies are preventive and take place before violence occurs. Withstanding refers to individuals shielding themselves from active violence; they are utilized when prevention has failed and violence is underway. Confronting involves individuals violently or nonviolently engaging criminal groups to stop violence. Violent confrontations include threatening the perpetrators with physical harm. Nonviolent confrontations include advocating on behalf of an individual or negotiating a compromise.

Residents avoid, withstand, or confront both indiscriminate and targeted violence. However, the specific practices they use to do so vary depending on the type of violence they face (Table 1). Because the risk of indiscriminate violence is defined by location rather than personal identity or actions, the associated self-protection strategies typically involve avoiding violence by keeping away from high-risk spaces and, when violence is unavoidable, withstanding it by sheltering in place. Individuals confront indiscriminate violence when they negotiate or advocate with the criminal group to cease shootouts and not engage rivals or the police in the neighborhood. Since the risk of targeted violence is defined by personal identity or actions, the associated self-protection strategies involve avoiding targeting by concealing identity and modifying behavior and, once threatened, withstanding the attack by going into hiding or conceding to criminal demands. Individuals confront targeted violence when they negotiate with a criminal group or a prominent ally advocates with the group on behalf of the targeted individual, either violently or nonviolently.

Table 1. Self-protection strategies to endure urban violence

Compared to the other strategies, individual confrontation is rare because the risks are high, and any single resident has little power over criminal groups. Residents more often confront criminals collectively because they typically have more power over criminals as a group. However, confrontation can take the form of individual action, as we show here and as documented in the literature (e.g., Zubillaga, Llorens, and Souto 2019, 435–436; Marston 2020). When individuals confront gangs, they sometimes draw on the power of collective identities, such as motherhood, or social ties with gang members; nonetheless, their actions are overwhelmingly carried out individually, even as they draw on collective identities to give their individual acts meaning. Our focus is on individual and household-level self-protection strategies. Collective action is addressed elsewhere in the literature.

We ground our framework in the distinction between indiscriminate and targeted violence because these categories reflect the characteristics of violence our interlocutors described when discussing self-protection mechanisms. Alternative categorizations of violence exist, such as violent versus nonviolent crime (Berens and Dallendörfer 2019) or the frequency and visibility of violence (Durán-Martínez 2018). However, the self-protection mechanisms we document respond to violence specifically, so an emphasis on nonviolent crime is not useful. Furthermore, residents’ self-protection strategies do not appear to respond to variation in the frequency or visibility of violence, making that categorization unsuitable. Instead, our research participants describe their choices as shaped by whether they were inadvertently caught up in violence happening to others or were individually targeted. Consider the following evidence from Monterrey, which is typical of broader trends in our data and illustrates the relevance of the indiscriminate and targeted categories:

Author: Did you ever think of moving away … ?

Interviewee: No … Look, I wasn’t the target of the violence. It was the criminal groups fighting with each other and with the police.

Author: So if you weren’t involved in organized crime, you were safe?

Interviewee: Yes. Now, sometimes you would get impacted by the violence out of bad luck. For example, if you got caught in the crossfire of a shootout. Or, during those times, you could be eating in a restaurant, and they would enter and shoot someone there. But it was only to shoot that person—it was not to spray bullets across the restaurant. ⁶

While the type of violence a resident faces influences the repertoire of self-protection mechanisms they draw on, residents are not bereft of agency. In fact, marginalized residents’ capacity to exercise self-protection evinces agency in highly constrained settings. Agency is embedded in social, political, and economic relationships that structure opportunities and constraints on individual choice. The agency of those with the least access to traditional sites of power is typically the most limited (Davies and Msengana-Ndlela 2015; Banks, Lombard, and Mitlin 2020; Hume and Wilding 2020). Individual residents’ social, political, and economic clout further shape the agency with which they can act (Mac Ginty 2014; Banks, Lombard, and Mitlin 2020, 230–231; Hume and Wilding 2020). This lends some residents a greater range of action in the face of violence than others, even when they share similarly marginalized conditions and similar contexts of violence. For example, while our framework maintains that residents seek to fortify their homes to guard against indiscriminate violence, economic inequalities shape different households’ range of options for doing so. Gender, sexual orientation, age, religious affiliation, social connections, and other characteristics may further intersect with structural power dynamics (Mac Ginty 2014). In sum, agency varies based on contextual limitations, identity, and the particular crisis at hand such that some individuals have greater or lesser agency at a given moment (Davies and Msengana-Ndlela 2015, 132; Hume and Wilding 2020). Yet even the most constrained individuals typically have multiple self-protection strategies from which to select. When absent, our informants rely on innovation and improvisation to generate new strategies. Their actions align with conceptions of agency that emphasize “goal formulation, creativity and planning, especially … when confronted with novel challenges” (Davies and Msengana-Ndlela 2015, 132).

Self-protection strategies are not overtly political, yet their political import should not be dismissed. Others argue that the resolve to continue daily life amid violence constitutes “everyday resistance” against subjugation (Lees et al. 2018, 351; Arampatzi 2017; Hume and Wilding 2020). Pointedly, Lees, Annunziata, and Rivas-Alonso’s (2018) notion of survivability suggests that remaining in a space where social, political, and economic forces aim to dispel or dominate lays the minimum grounds for resistance. Likewise, the behaviors we document have political implications by facilitating nonconformity to forces of violence (Mac Ginty 2014; Lees, Annunziata, and Rivas-Alonso 2018, 351; Jenss 2021).

Research context

Medellín, Colombia, and Monterrey, Mexico, are contemporary cases of urban criminal violence. Both are densely populated, major metropolitan zones with high levels of violence attributed to local gangs and organized crime. Nestled in the Aburrá Valley, Medellín is an industrial and entrepreneurial hub. The city of 2.5 million is also skirted by marginal neighborhoods (barrios populares), which sit high up the surrounding hillsides. Despite differences in ethnic makeup of residents and access to public services, these neighborhoods share commonalities such as disenfranchisement from the city center, close-knit communities, high rates of poverty, and susceptibility to criminal control. Criminal organizations wield violence to regulate communities, including using targeted violence for social control, forced displacement, drug sales, and extortion. On the other hand, criminal gangs sometimes provide social services, such as dispute resolution, foodstuffs, and school supplies (Abello-Colak and Guarneros-Meza 2014; Gordon 2020). Residents interact with gangs (combos) whose members are generally from the neighborhood but are affiliated with higher-level criminal organizations, several of which control the city. Although largely absent for many years, the police began to make incursions and built small stations in the city’s marginal neighborhoods in the late 2000s and early 2010s. Police killed or harmed residents during their incursions, and shootouts between the police and gangs produced indiscriminate violence. After a peak in the 1990s, when the city was considered the most violent in the world, Medellín experienced a stark drop in homicides in the early 2000s. Rates of violence spiked again between 2009 and 2013 before declining in subsequent years (den Held and Robbins 2019).

Like Medellín, Monterrey is a locus of industry and innovation that simultaneously suffers from high rates of inequality. Socioeconomically vulnerable neighborhoods often flank wealthy ones, sometimes separated by nothing more than a freeway. Despite being geographically close to seats of city and state power, these spaces are typically excluded from political influence, experience high rates of poverty, and suffer from criminal predation. Neglect and mistreatment by state institutions, paired with decades of unrequited promises from politicians, produce general mistrust of the state, particularly the police.

Street-level gangs have a long history in the vulnerable Monterrey neighborhoods studied. ⁷ The nature of criminal threat escalated around 2008 with increasing tensions and later a turf war between two major organized crime syndicates, Los Zetas and the Gulf Cartel. This included street-level gangs subsumed and supported by them. One interviewee described the importance of this shift, observing, “Before, the gangs occupied the streets, but they were kids from the neighborhood … and they fought with rocks and knives … not AK-47s.” ⁸ As in Medellín, violence intensified further when the military intervened. The ensuing conflict generated indiscriminate violence, most frequently in the form of shootouts between the crime groups and between the crime groups and the state. It also included targeted violence, including reprisals, kidnappings, and forced recruitment. The conflict diminished by 2015. While there are still gangs in Monterrey neighborhoods with links to larger criminal organizations, violence has declined markedly.

Methods

The evidence we present is based on more than two years of combined field research in Medellín and Monterrey carried out between 2016 and 2018. We leverage a multimethod data collection strategy, drawing on participant observation with local communities beset by violence, in-depth interviews with residents of these communities, and an original survey. Interviewees included men and women, LGBT persons, persons of different ethnic backgrounds and ages (older than 18), and community leaders. Interviewees were asked to share their observations and experiences on several topics, including living amid criminal violence, how they and their communities respond to violence, and the logic behind those choices. The interviews focused on times of high violence when falling victim to both indiscriminate and targeted violence was a threat to most residents. This tended to be 2009–2013 in Medellín and 2009–2015 in Monterrey.

Our interviews were carried out alongside participant observation in marginal neighborhoods. One author lived in a peripheral neighborhood in Medellín for six months and spent additional time volunteering with a local organization there; the other lived in a peripheral neighborhood in Monterrey for three months and spent additional time volunteering with area neighborhood improvement groups. We also accompanied residents in their quotidian activities, like attending birthday parties, looking after sick neighbors, doing morning exercise, and attending neighborhood meetings to address local problems. Observing how residents managed and talked about the safety aspects of otherwise ordinary tasks—picking up their children from school, walking to the internet café, or going to the corner store for a Coke—led to important insights about the practices residents adopt to manage daily life in a violent city. Finally, Marston carried out an original survey in three marginal neighborhoods of Medellín as part of a larger project on forced displacement (n = 613). ⁹ For the present article, we analyze qualitative data from open-ended survey questions about whether and how the respondent leveraged different coping mechanisms to manage the threat of violence.

Self-protection in violent cities: Evidence from Medellín and Monterrey

Indiscriminate violence

Criminal groups operating in cities generate indiscriminate violence through confrontations with other groups or state security forces. The risks of falling victim to indiscriminate violence are defined by civilians’ physical proximity to it because indiscriminate violence is not aimed at particular civilians. The self-protection mechanisms most frequently associated with indiscriminate violence are thus those that maximize distance. Among our interlocutors in Medellín and Monterrey, being caught in the middle of a shootout between criminal groups or between these and the state—and the stray bullets that result—were the most frequently cited manifestations of indiscriminate violence.

Avoiding indiscriminate violence

Residents in the peripheral neighborhoods we studied leveraged an array of coping mechanisms to avoid indiscriminate violence. Common avoidance behaviors include staying indoors, especially inside the home, and modifying one’s schedule to avoid leaving during high-risk periods, like after nightfall. A Monterrey-area taxi driver described the indiscriminate violence he and other residents faced and the avoidance mechanisms they employed:

Just a few years ago, the whole city was racked by crime … It was because the organized crime groups were fighting each other and the police … there were shootouts that sometimes people would get caught up in. During those times, everyone was very afraid. Lots of businesses shut down because people wouldn’t go out at night—the streets were empty after 7 p.m. ¹⁰

Residents in Medellín’s marginal neighborhoods likewise avoided going outside to protect against indiscriminate violence. One mother recounted how “I kept them [her children] shut in, and we came in to go to sleep at six. It was horrible around here … there was no one around, the police wouldn’t come up here, and around here the people kept themselves shut in due to all of the shootouts.” Her sentiments were echoed by other Medellín residents. One 19-year-old unemployed woman relayed, “Well, we didn’t go out to the neighborhood at all because we were afraid and worried that something would happen to us.” A 44-year-old construction worker similarly reported: “I hid. I didn’t leave my house if it wasn’t strictly necessary.” According to a then legislator in Monterrey, the idea that violence could happen to anyone—that is, indiscriminately—facilitated the integration of avoidance mechanisms into daily life:

We entered a very dark period. The nightlife disappeared because people were scared to go out at night … There was a lot of changing habits at the time … People are now accustomed to checking out the place they’re going to ahead of time, and parents won’t let their children be in public spaces alone anymore. When I was a kid, we would walk the streets, go to the park—we had complete freedom. Not anymore. ¹¹

During particularly intense peaks of violence between crime groups and the state, residents often deemed it too dangerous to leave home and opted instead to miss work, ¹² close their businesses early, ¹³ and keep children home from school. ¹⁴ Yet for marginalized residents subsisting on daily wages, staying home is not always possible. In these instances, residents would modify their schedules to avoid high-risk times. This might require leaving extra early in the morning to be home before nightfall. ¹⁵ If residents found themselves away from home in the late afternoon, they would call family members to see if there were signs of imminent confrontations in the neighborhood (such as armed patrols or graffiti messages) and, if necessary, stay elsewhere in the city. ¹⁶

Residents of criminally controlled areas often receive advance notice of impending indiscriminate violence, such as confrontations between gangs or state security operations. In Medellín and Monterrey, residents described receiving WhatsApp or word-of-mouth messages advising them to remain indoors due to events that, though not targeting residents, could place them in harm’s way. The fact that armed groups and security forces warn residents of such clashes evinces the distinction between indiscriminate violence, of which residents are not the object, and targeted violence, which aims to harm specific residents.

Indeed, such warnings often prove vital to residents’ ability to avoid indiscriminate violence. However, as outlined in our theoretical framework, residents’ ability to employ such strategies can vary at the individual level. In Medellín, advance notice was sometimes reserved for those who supported the gang or were otherwise highly respected. ¹⁷ One 81-year-old resident explains: “Since I am old and very deaf, the guys [in the gang] look after me. They would tell me when they were going to shoot someone [iban a dar bala] and to shut myself with my wife into the house.” Another resident similarly recounts: “The guys in the gang respected us. They told us when there was going to be a shootout so that we could shut ourselves in with the children.”

These strategies are a logical response to the risk of indiscriminate violence specifically. Because such risk is defined by proximity to crossfire, residents reason that they can stay safer by keeping away from the spaces and times when violence is likely to erupt. At the height of violence, this was almost anywhere outside the home. We would not expect the same strategy from individuals facing targeted violence, in which one’s risk is defined by their identity or actions. Under this logic, staying home is a misstep because perpetrators can more easily locate their target.

Withstanding indiscriminate violence

Residents are not always able to avoid indiscriminate violence. In these instances, residents relied on several mechanisms to withstand the violence as it happened. These involve what we call a “shelter in place” strategy. If caught in the middle of a shootout while outside, residents retreated from the violence as quickly as possible, ¹⁸ took cover if unable to leave, or sought refuge in nearby shops or homes. In one Monterrey neighborhood, residents described getting caught in shootouts while walking their children home from school. They would knock on the doors of nearby houses. If nobody opened, they ran to their own home. ¹⁹

Yet being inside does not guarantee safety. In both research sites, a common concern was stray bullets entering homes, and many houses remain pockmarked. Once inside, residents took additional measures to protect their physical well-being, like hiding where wayward bullets are less likely to strike. ²⁰ This could include in the bathroom or interior rooms or under tables, desks, or the bed. ²¹ Reflecting the simultaneous refuge and weakness of the home as protection, residents may “bulletproof” (blindar) their homes to minimize the likelihood of bullets passing through exterior walls. This may include placing mattresses against weaker walls, doors, and windows, ²² reinforcing walls with additional layers of cinderblock, or, as a 52-year-old woman from Medellín explains, “We had to fill in the windows with bricks so that bullets … wouldn’t enter.”

While sheltering, residents relied on mental coping mechanisms to make withstanding indiscriminate violence psychologically bearable. For many, solace in their faith helped them endure. One Monterrey resident noted she would tell anxious friends and family: “Pray. Pray to God because this is hell we are living in, and only He can save us.” ²³ Meanwhile, adults used distraction and words of reassurance to help children remain calm. In Medellín, a teacher described instructing students to take cover under their desks and having them sing songs to divert their attention from shootouts outside. ²⁴ Another Medellín resident remembers hiding as a child in the bathroom, where her mother turned on the shower to conceal the sounds of a gang shootout: “My mom told me everything was going to be OK, I would play with my dolls, the shower in the bathroom made so much noise that you almost couldn’t hear the gunshots.” ²⁵

Confronting indiscriminate violence

Confronting strategies involve violently or nonviolently engaging criminal groups with the aim of stopping violence. In Medellín, community leaders and other prominent residents scolded gang members and shamed them for shootouts in the community or for bringing indiscriminate violence to the neighborhood. For some women, their ability to confront the gang stems from their sacred identity as a mother, something Zubillaga, Llorens, and Souto (2019) also observe in marginal neighborhoods of Caracas. The residents of one Monterrey neighborhood, meanwhile, credit the peacefulness of their neighborhood to one man who used threats to confront new criminals attempting to encroach on the neighborhood, which could lead to clashes with existing groups. His previous contributions to expelling criminals lend him power and influence that other residents lack. ²⁶

Although infrequent relative to strategies to avoid and withstand violence, our evidence demonstrates the existence of individual strategies to confront indiscriminate violence. This is in line with our framework, which posits that most individuals lack sufficient power over criminal groups to keep them from using indiscriminate violence. The rare instances we observed emerge when the individual can overcome the power imbalance—often by leveraging collective identities or preexisting social clout.

Targeted violence

Targeted violence is directed at particular individuals or families. The risks of targeted violence are defined by a resident’s identity or actions. The self-protection mechanisms that emerge in response aim to conceal vulnerable aspects of one’s identity or modify personal behaviors. Our interlocutors in Medellín and Monterrey cited targeted threats for suspected collaboration with a rival or the state (including reporting a crime), for disobeying neighborhood rules or norms, to appropriate their property, to kidnap for ransom, or to recruit or coerce them into work. In Monterrey, residents were forcibly recruited into working as drug dealers, ²⁷ foot soldiers, ²⁸ housekeepers and cooks, ²⁹ and in drug production warehouses. ³⁰ We also heard of residents in both cities employing criminal groups to carry out personal objectives, such as to collect unpaid debts or to appropriate property. ³¹

Avoiding targeted violence

Residents employ strategies preemptively to avoid being targeted. The phrase, “See, hear, and keep quiet” (ver, oir, y callar) captures a common strategy among Medellín and Monterrey residents to keep to themselves, to avoid getting involved in other people’s business, and to not interfere in criminal activities nor inform anyone about crimes witnessed. In the three neighborhoods surveyed in Medellín, 64 percent of respondents reported that keeping to oneself described them or their family’s behavior “very much.” ³² One resident credited survival itself to adhering to this “law of silence.”

Residents of Monterrey’s peripheral neighborhoods followed a similar strategy. Residents rarely reported crimes due to fear of reprisals. Reprisals are targeted violence because they are typically aimed only at the offending individual or their household. Fear of reprisals against one murdered woman’s family kept them from reporting her death to law enforcement. “There was a lot of fear,” the woman’s sister explained, “because in the days after they found her [deceased], people were saying that there were men walking around here [the neighborhood] asking where her family lived. I don’t know if that was true, or not, but it made us very afraid.” ³³ This fear of targeted reprisals extended to residents who had not experienced crime directly:

We couldn’t talk to anyone because you never really knew who you were talking to. And it wasn’t safe to report crime because you put yourself at risk of a reprisal from the criminals … my dad told us to not go out at night and that if we ever saw anything going on, to just keep going and to never tell anyone about it. ³⁴

As this interviewee intimated, residents of Monterrey became wary of speaking or interacting with anyone, even neighbors, lest they unwittingly anger the criminal group by sharing information they should not or offend a neighbor who then enlists the group to retaliate. ³⁵ One Monterrey grandmother who described herself as “outspoken” and very involved in her community changed these behaviors to prevent targeted violence against her or her family. Afraid that his wife might offend the wrong people, her husband beseeched her to talk less, keep her opinions to herself, and stay out of other people’s business. ³⁶

Criminal groups may also target residents if they suspect them of collaborating with a rival group. Knowing this, residents heed the social and spatial boundaries between criminal groups to portray neutrality (see also Baines and Paddon 2012). Socially, residents may avoid speaking with or greeting members of any criminal group, lest one group perceive this as support for rivals. ³⁷ Spatially, residents avoid targeted violence by remaining within one group’s territory. Crossing these “invisible borders” (fronteras invisibles) could invite assault if a group perceives the border crossing to be evidence of collaboration with their rival, like conducting reconnaissance. One interviewee in Medellín described “being trapped in a cage”: church, work, and the children’s school were in other neighborhoods, but crossing into them was sometimes impossible. ³⁸

Residents with property or any sign of wealth are common targets of violence, even in peripheral neighborhoods. In Monterrey, we documented individuals killed or disappeared so that neighbors, family members, or gangs could appropriate their homes. ³⁹ Extortion by criminal groups and the police is also common. ⁴⁰ Accordingly, residents hide their belongings and conceal or “camouflage” (Villarreal 2015) their wealth. Some bury their valuables; ⁴¹ others store their money with family members outside their household. ⁴² Others conceal the value of their larger holdings. For example, some keep bank accounts and titles to houses or cars in others’ names; ⁴³ others refrain from making exterior home improvements. ⁴⁴ These practices prevent targeted violence for wealth appropriation or extortion by giving the appearance of modest homes and possessions.

We also observed residents of marginal neighborhoods in Medellín and Monterrey traveling together in groups to avoid being targeted, that is, “caravanning” (Villarreal 2015). Traveling in groups may deter attacks that are easier to carry out against an individual, such as extortion, robbery, and kidnapping. Following this logic, women in one Monterrey neighborhood never ventured far from home alone and recounted traveling to schools, markets, and church in groups. ⁴⁵ Similarly in Medellín, survey respondents report running errands and commuting to work together. ⁴⁶ In both research sites, few mothers allowed children to walk home from school alone: “My mom always went for us at school, and when she couldn’t, a neighbor took us and picked us up.” ⁴⁷ Furthermore, perpetrators may be less likely to carry out targeted attacks in front of witnesses. Residents of one Monterrey neighborhood remember their neighbor’s insistence on being accompanied in the days leading up to her death. They suspect she knew she was being targeted for her relationship with a police officer and hoped to prevent the attack by never being caught alone. She was later killed in her home, alone, save for two young children hiding under the bed. ⁴⁸

Finally, residents avoid becoming a target of violence by cultivating a good reputation in the neighborhood. This is based on the logic that, should a criminal group find reason to target them or ask around about them, neighbors will speak well, thereby decreasing the chances of targeted violence. ⁴⁹ For example, if the criminal group suspects a family reported them to the police, but neighbors say the family is loyal and keeps to themselves, the group is more likely to issue a stern warning rather than use violence. A 70-year-old Medellín resident summarizes this strategy for living amid the threat of targeted violence: “Make myself known to the people and show them that I am someone decent.” Like the other avoidance strategies described in this section, cultivating a good reputation is a logical response to the threat of targeted violence in particular because it responds to targeting’s emphasis on individual identity and actions.

Withstanding targeted violence

These mechanisms help civilians avoid becoming a target for violence in the first place. Once they have been targeted, how do civilians endure the imminent threat of harm? When we refer to imminent targeted violence, we mean instances in which a credible threat of violence has already been issued and the targeted individual must take action to address the threat. Just as civil war scholars suggest civilians may “accommodate” (Baines and Paddon 2012), “support” (Barter 2012), or “engage” (Jose and Medie 2015) armed groups, many residents survive targeted violence by conceding to criminal groups’ demands. ⁵⁰ In Monterrey and Medellín, this included paying extortion fees; ceding property; agreeing to work for the gang; storing drugs, arms, or money; or providing in-kind items like food, among others. ⁵¹

The case of a clothes vendor in Monterrey elucidates how residents withstanding targeted violence make concessions to ensure their survival. This interviewee was kidnapped and extorted for money. He suspects he was specifically targeted because a disgruntled employee, seeking revenge over a dispute, told gang members he was wealthy. His kidnappers insisted he pay a monthly fee he could not afford. He first refused, explaining he did not have the money. His appeal fell on deaf ears; he was held for days and beaten repeatedly until he agreed to pay the requested amount. Upon this concession, the kidnappers released him. Knowing he could not pay the fee and would be killed upon failing to do so, he fled Monterrey. ⁵²

While residents withstanding indiscriminate violence typically shelter in place, residents withstanding targeted violence are more likely to go into hiding elsewhere, at least temporarily. Survey respondents in Medellín report hiding at a friend’s house or in a neighboring town after receiving a targeted threat. For some, the objective of hiding is to buy more time while they try to comply with the gang’s demands (e.g., while they collect money to pay extortion fees). Others hide and wait for the problem that provoked the targeted violence to blow over (e.g., waiting for rumors about their collaboration with a rival to settle). As one respondent explained, “We hid in another house, further down [the hillside], a friend’s place, but they [the criminal group] even went there to get us.” While it appears to have failed in this case, this was a survival strategy nonetheless.

Confronting targeted violence

We documented cases of residents who confronted crime groups to protect themselves from targeted attacks and who confronted crime groups on behalf of others. In one Monterrey neighborhood, for example, a neighbor accompanied another to violently threaten gang-backed family members attempting to appropriate his house. ⁵³ However, navigating when and how to confront criminal groups presents many challenges and may require switching between several mechanisms. A 56-year-old woman in Medellín describes challenging a gang to keep her home:

I was very sincere and spoke head-on [hablaba de frente] when I had to … they threatened me many times, and also tried to displace me and steal my house from me, but I had to stop them and speak with them, because this is my house and I’m not going to let anyone take it, I had lived many years in the streets with my kids. A while later, when everything got worse, and they continued threatening my family, I had to see, hear and keep quiet.

Moreover, some residents are more likely to employ strategies to confront targeted violence based on their personal characteristics and preexisting clout with criminal groups, including older women and mothers, religious figures, or others with social ties to gang members (e.g., being childhood friends).

Conclusion

In this article, we ask how residents weather large-scale, organized crime violence in marginal neighborhoods of Latin American cities. Based on qualitative evidence from Medellín, Colombia, and Monterrey, Mexico, we shed light on the self-protection strategies residents utilize and inductively generate an analytical framework through which to catalog and make sense of these strategies. Our framework not only categorizes resident strategies to avoid, withstand, and confront violence. It also considers the type of violence residents face, either indiscriminate or targeted, as conditions under which we observe distinct repertoires of self-protection.

Our objective is not a comprehensive theory. Instead, we aim to advance knowledge by providing a theoretically grounded mapping of self-protection mechanisms that will support future research. Such conceptual and typological work lays important foundations for that research. To advance theory building and hypothesis testing, future research might investigate how additional variables like socioeconomic status, gender, or ascriptive group membership influence individuals’ choice of self-protection strategy and the degree to which targeted versus indiscriminate violence weighs in these decisions. Additionally, our framework facilitates documentation and cross-case analysis of self-protection mechanisms in other Latin American cities. Medellín and Monterrey share important commonalities with other metropolitan areas in the region, including high levels of inequality, segregation of vulnerable populations, legacies of criminal and political repression, and violence from confrontations between large-scale, localized crime groups, and state efforts to subdue them. Collecting additional evidence of individual confrontation strategies is particularly relevant. While uncommon compared to avoidance and withstanding strategies, our evidence from Monterrey and Medellín demonstrates the existence of individual confrontation strategies, consonant with evidence collected in other contexts (e.g., Zubillaga, Llorens, and Souto 2019, 435–436). Reflecting confrontation strategies in our theoretical framework, despite limited evidence from our specific cases, allows for their collection in contexts where they might be more prevalent. Finally, by illuminating residents’ survival strategies, we shed light on an open question in the forced migration literature: given mass flight, how do some individuals and households manage to remain amid violence (Ibáñez and Moya 2016)?

This article complements work by other scholars studying how individuals and communities manage the fear, vulnerability, and violence characteristic of an increasing number of Latin American cities. We depart from existing research by synthesizing the urban violence and criminal governance literatures with scholarship on civilian self-protection in civil wars, a literature that naturally offers many insights on enduring pervasive violence. We import the distinction between indiscriminate and targeted violence from the civil wars literature, advancing our understanding of the selection of individual self-protection strategies in the marginal neighborhoods of Latin American cities. Through this, we draw attention to a significant but understudied research topic while emphasizing the centrality of agency for residents’ “staying power” amid urban violence.

For policy makers and practitioners, our findings underscore that humanitarian aid and protection interventions in marginal neighborhoods facing criminal violence should reinforce the strategies residents already employ to keep safe. Rather than import strategies from elsewhere, practitioners should recognize residents’ agency and follow their lead to identify the most effective self-protection strategies based on the type of violence they face. For example, in Medellín, a humanitarian organization worked with teachers and children to develop programs that promote dialogue instead of violence and teach students how to keep safe if a shootout occurs outside their school (International Committee of the Red Cross 2015). Ultimately, residents are the experts in their context and, crucially, will live with the consequences of any intervention (Autesserre 2021, 171). As such, academics and practitioners must be attentive to residents’ know-how and preferences when it comes to staying safe in violent cities.