Assessing Gang Risks in Post-War Environments: The Case of Colombia

Countries emerging from armed conflict often experience heightened violence and youth gang activity. Following the signing of peace accords with the Revolutionary Armed Forces of Colombia — People’s Army (FARC-EP), what are Colombia’s risks in terms of youth gangs? To assess these risks, this article draws from gang research and literature on post-war violence to identify six factors that recur in post-war environments and are likely to fuel a rise in gangs: illicit economies and criminal networks, exposure to violence, marginalization, social disorganization, security gaps and state responses, and former combatants. After analyzing Colombia’s risks with reference to each of these, the article concludes that the strengthening ties between youth gangs and Colombia’s illicit economies, mediated by adult-run criminal networks, increase gang numbers and violence. Moreover, some disadvantaged, urban neighborhoods are vulnerable to gang escalation due to the effects on local youths of protracted violent exposure, marginalization, and social disorganization. Finally, while ex-combatant recidivism and security gaps are prominent concerns in Colombia, they are not expected to contribute significantly to youth gang dynamics in urban areas.


Introduction
In November 2016, the Colombian govern ment signed a peace accord with the Revolutionary Armed Forces of Colombia -People's Army (Fuerzas Armadas Revolu cionarias de Colombia -Ejército del Pueblo -FARCEP), the country's largest guerrilla group. Over three years later, mul tiple armed conflicts are still under way in the country, involving FARCEP ' dissident' or successor elements, the National Liberation Army (Ejército Nacional de Liberación -ELN), the Clan del Golfo (also referred to as That postaccord Colombia does not fit neatly with any commonsense concept of postwar is not unusual. Countries emerging from armed conflict often see an escalation in violent crime, which can topple peace pro cesses. Even when a negotiated settlement holds up, protracted insecurity precludes real recovery and erodes support for peace (Darby 2006). In an effort to reconcile such tendencies with a 'postconflict situation, ' Brown, Langer and Stewart (2011: 4-5) sug gest that postconflict is best understood as a process continuum that can suffer regress. Colombia's peace accord, ceasefire, and demobilization of most FARCEP members are milestones on that continuum, represent ing a significant reduction in political vio lence and creating space for peacebuilding.
I worked in Colombia for the International Organization for Migration (IOM) 1 from 2000 to 2002, during the peak of paramilitary vio lence, and again from 2011 to 2017, during the FARCEP negotiations and the start of peace implementation. My work involved excombatant reintegration, rehabilitation of child soldiers, and prevention of child recruitment into illegal armed groups. After the FARCEP accord was signed, the ques tion of Colombia's susceptibility to youth violence was a frequent subject of debate within the peacebuilding community. Since the rise of maras in postwar Central America was both the closestathand and most noto rious example of gang intensification in the aftermath of civil conflicts, these discussions often turned to how Colombia was -or, more often, was not -perceived as similar to El Salvador and Guatemala in the mid1990s.
While these accounts often describe the impacts of armed conflict and peace on local gang dynamics, what is missing is a framework for identifying those features of postwar situations that are most likely to fuel a rise in gang numbers and violence. To develop such a framework, I first reviewed ethnographic, case study, and quantitative research analyzing risk for gang recruitment and intensification, including compara tive research on other forms of organized youth violence to supplement gangspecific research in developing country contexts. I then reviewed the growing body of litera ture on postwar violence to identify recur ring features of postwar environments that intersect with major risk factors for gangs. This exercise generated a short list of risks that were likely to emerge in the aftermath of armed conflict and to contribute to gang intensification. Finally, each of these factors was applied to the Colombian context, taking account of its experience of armed conflict, the overall security context, and youth gang dynamics drawn from academic and institu tional literature, media reports, and personal communications with Colombian experts.
This paper is presented in three main parts. The first one provides an overview of Colombia's current security environment and gangs. The second part mines studies of gangs for major risk factors and the newer literature on postwar violence to posit -where these two bodies of scholarship intersect -prob able risk factors for postwar gang escalation. The third part analyzes Colombia's risks with reference to each of these factors. The article concludes that some disadvantaged, urban neighborhoods in Colombia are vulnerable to increased gang activity in light of their prolonged exposure to violence, marginaliza tion, social disorganization, and the strength of illicit economies. Excombatant recidivism and security gaps, though acute in Colombia, do not represent a firstorder threat of escalat ing urban street gangs. Moreover, Colombian authorities may attenuate these risks by expanding preventive measures grounded in citizen security and avoiding overly repres sive, militaristic approaches that have exacer bated gang violence elsewhere.

Colombia's Security Landscape
Traditional FARCEP strongholds with sig nificant criminal assets, like coca crops and illegal mines, have been hard hit by postwar violence as different criminal actors have scrambled to fill the vacuum. PARES (2019) reports that 107 municipalities of the 242 vacated by the FARC had suffered a significant incursion by other criminal actors by August 2019. In these municipalities, the homi cide rate increased enough in 2018 to offset reductions elsewhere and raise the national murder rate for the first time since the peace talks started in 2012 (Castillo, Lleras, and Suarez 2018).
Underlying these developments is a crimi nal underworld that feeds off the cocaine trade. Coca production increased throughout the peace talks and the start of peace imple mentation, breaking historic records in 2016 and again in 2017 (Asmann 2018). According to the Ideas for Peace Foundation (Fundación Ideas para la Paz -FIP), a Colombian think tank, the rise is partly attributable to increased output by smallscale farmers in anticipation of volunteer eradication pro gramming that many believed would com pensate farmers based on their production (Llorente and Garzón 2018). Although the number of cocagrowing crops levelled off in 2018, more intensive cultivation meant that production continued to rise (Colombian Observatory of Organized Crime 2019).
The ICRC (2019) identified five ongoing armed conflicts in Colombia. Four involve the Colombian army against illegal armed groups: the ELN, the EPL, the Clan de Golfo, and successor FARCEP groups. The fifth refers to a conflict between the EPL and the ELN. The five conflicts are in addition to the ' other situations of violence' 2 identi fied in numerous cities, such as Medellín, Buenaventura, Cali, and Tumaco (ibid.).
The ELN has grown in numbers, territory, and capabilities since the signing of the accord, rising from 1,100 to 3,000 combat ants and establishing a presence in 35 new municipalities (101 to 136 out of a total of 1,122 municipalities in Colombia) (PARES 2019). The Clan del Golfo, the strongest of the socalled bacrim, or criminal bands, has also expanded into 26 new municipalities, up from 276 when the accord was signed, and has 2,500 members (ibid.). The FARC ' dissident' or successor groups have also strengthened, with estimates in August 2019 of over 2,200 members, including at least 400 new recruits, spread across 23 groups and active in 85 municipalities (ibid.). Recent highlevel defections from the peace process and calls to rearm by demobilized command ers may herald greater military and ideologi cal consolidation among these groups (The Economist 2019b).
Colombian politics shifted to the right in 2018 with the election of a president who opposed the peace process throughout his campaign. President Iván Duque's adminis tration launched a new national security pol icy in 2019 that favors military solutions over negotiated settlements with armed groups like the ELN and Clan del Golfo (Alonso and Robbins 2019). The policy is largely focused on rural strategies and is unlikely to impact security conditions in cities (The Economist 2019a).

Colombia's gangs
For the purposes of this discussion, youth or street gangs are groups that display some continuity over time, engage in illegal behav ior, and consist mostly of members under the age of 25 (Rodgers and Hazen 2014: 8). The gangs that first appeared in Colombia in the 1960s were hierarchical, with large member ships and territories, and wellestablished rituals (Perea 2007). Starting in the 1980s, however, these gangs have undergone con tinuous transformations, and vary from one city to another, depending largely on the city's relationship with armed conflict and illicit economies. The number of gangs has increased over time, but they generally have fewer members and control smaller ter ritories (Semana 2018). Their structures are looser and more horizontal, and the old rules and rituals have largely disappeared (Perea 2007;Semana 2018). They are also more violent, more criminal, and better armed, especially in those cities where links to the adult criminal world are strongest (Resource Center for Conflict Analysis [CERAC] 2014).
In its recent report on gangs in seven cit ies, Colombia's Ministry of Justice (MOJ) observed that contradictory estimates, incomplete data, and divergent methods made it impossible to estimate the number of gangs and their members in Colombia (MOJ 2017). The ministry hazarded to say that in the seven cities in its study, there may be some 1,200 gangs with 21,000 members (ibid.). Credible research from other sources suggests that gangs are viewed as problem atic in at least 12 Colombian cities, with Cali and Medellín at the top of the list (El Tiempo 2015; Semana 2018). Gang activity is concen trated in crowded neighborhoods set apart from the city, with lowquality education and jobs, where young people are effectively excluded from the practices of citizenship (MOJ 2017). Gang members tend to come from dysfunctional families with high rates of domestic violence; they typically neither work nor go to school (Semana 2018).
With some exceptions, Colombia's gangs are not as persistent or rooted as many gangs in Central America or the US, but they may be on a development path that is headed in that direction. In his historical analysis of street gangs, Howell (2015) mapped five phases that describe the development of gangs in much of the US. The path starts with social disorganization in phase 1 and low neighbor hood and family control in phase 2, followed by the formation of gangs in phase 3 and the interplay with 'gang violence facilitators' like drug trafficking and weapons availability in phase 4, and culminates with widespread, transformed gangs in phase 5 (ibid.: 76-94). Many Colombian gangs would be classified in phase 3 or 4, but almost all gangs in Medellín and some in Cali meet the description for phase 5. Nationally, the major gang violence facilitators have been adult armed groups that bring drugs and weapons, and also gen erate the cultural conditions for violence.
While cities like Medellín and Cali have made sustained efforts to address youth gangs, national government attention to gang issues is relatively new, in part because the ' complex ity of the armed conflict did not leave space to focus on gangs' (Perea 2007: 14). When a Colombian senator brought the issue up for parliamentary debate in 2017, he remarked that the gang threat grew without notice while the country worried about the armed conflict (HSB Noticias 2017).

Risk factor analysis in gang studies
Risk factors here refer to experiences and characteristics that increase the probability that a young person will join a gang. Gang related risk is often discussed on two levels: at the micro level, research looks at risk across five domains of the individual, family, school, peer group, and community; the macro level takes on the larger forces at work in society -the ecology that surrounds and shapes the smaller domains (Howell and Griffiths 2016).
Much of the evidencerich analysis on risk has emerged from the US, which benefits from the 'longest and densest' gang scholar ship in the world (Rodgers and Hazen 2014: 9). While there is substantial research on gangs from Central America, and, to a lesser extent, from South Africa and Brazil, much of this literature lacks the statistical rigor to pinpoint and assess predictors of gang mem bership. Comparative research on youth engagement in a broader sample of armed groups, like Dowdney's (2005) research on youth involved in organized armed violence, can help fill the gap.
While findings from global literature must be applied cautiously to the Colombian case, risk profiles across countries and group types are surprisingly similar (Dowdney 2005;Hagedorn 2008). It is possible to identify core risk factors supported by solid evidence from the US and emerging support from other countries. Drawing from Howell and Griffiths (2016) and Thornberry et al. (2003) in the US, as well as the Small Arms Survey (2010), Higginson et al. (2018), Dowdney (2005) and Covey (2010) for international perspectives, the following factors may be relevant to gangs in Colombia.
In the individual domain, prior involvement with violence and drugs; antisocial attitudes, including the belief that aggression is neces sary to resolve conflict; traumatic life experi ences, especially victimization or exposure to substantial violence; as well as being young, marginalized, and male, increase the risk of gang involvement. Within families, economic disadvantage, domestic violence, family dys function, and gang membership within the home are risk factors. In the school domain, low educational attainment and attachment to school, together with dropping out, are associated with higher rates of gang involve ment. Regarding peers, having friends in a gang is one of the strongest factors in the US and elsewhere. At the community level, high risk neighborhoods tend to be urban and marginalized, with few opportunities, scarce services, and inadequate security provision. Firearms and drugs are often on hand, and the bonds among neighbors are weak.
More research on the risk factors for gang involvement for girls and young women is needed (Howell and Griffiths 2016;

Organization of American States [OAS] 2007).
Research to date suggests that risk profiles overlap, but some factors weigh more heav ily in girls' decisions to join a gang, including domestic violence and sexual abuse (Kinnear 2009;Taylor 2008), early sexualized behav iors, and difficulties in school (Howell and Griffiths 2016; Thornberry et al. 2003).
At the macro level, Thrasher's (1927) views on social disorganization resonate with the international literature. Social disorganization describes a situation in which conventional institutions like families, schools, and neigh borhood structures lose the ability to control delinquency. Such disorganization results from rapid and unpredictable change that dismantles old institutions without giving new ones a chance to consolidate (Hagedorn 2008). The impact on gangs is twofold: com munities lose the power to mediate juvenile misconduct and gangs emerge to provide the missing order (Papachristos and Kirk 2006).
Most of these risks are heightened in post war environments: families are fragmented, communities are uprooted, schools are destroyed, firearms are cheap, and social bonds are tenuous. Three areas of major risk identified in gang literature that feature heav ily in postwar settings are discussed below.

Exposure to violence
One explanation for the noted rise in homi cide during and after wars is that wars change societal norms and attitudes toward non combat violence (Archer and Gartner 1976). The rationale is that people whose daily lives are imbued with violence will cease to see violent acts as exceptional, while the tradi tional prohibitions on violence lose cred ibility when they are breached repeatedly without consequence. The effects deepen over time: as people become inured to vio lence, they are quicker to commit -and slower to punish -violent acts (Bos 2013). Increased violence reaffirms its normality and erodes communal capacities to resolve conflicts pacifically, further entrenching violence into daytoday affairs (Steenkamp 2009). Violence imprints on value systems as well as emotional and moral lives, thereby leaving a 'legacy' that makes violence more likely in the aftermath of war (ibid.: 8).
Persistent violence has serious consequences for young people whose development, sociali zation, and transitions to adulthood are threatened. Kurtenbach (2012: 8) explains that for a young person 'whose identity has been shaped by the experience of violence and a lack of even rudimentary forms of security, the development of a stable personality will be much more difficult.' She further describes how conflict alters or destroys the sources of socialization -families, schools, peers, media, institutions -and creates obstacles for transi tions to adulthood (Kurtenbach 2012).
The consequences of youth exposure to, victimization by, and participation in vio lence on future behavior are echoed in the gang literature. In their statistical analysis of children aged nine to fifteen years in dis advantaged neighborhoods in a midwestern city, Patchin et al. (2006), find that those who were more exposed through personal experience to community violence self reported higher rates of weapon possession and aggressive behavior. Youths who witness firearm violence are about twice as likely to engage in serious violence within the follow ing two years (Bingenheimer, Brennan, and Earls 2005). Writing about gang dynamics in Central America, Cruz (2005) views the ubiquity of violence in the home and school lives of many poor, urban youths as a key to understanding the extreme violence dis played by the maras.

Marginalization
The concept of marginalization refers to pushing certain people and groups out of mainstream society, away from opportuni ties, and relegating them to the economic, social, and geographic 'fringes of society' (Vigil 2002: 7). It can facilitate gang par ticipation when marginalized youths look to the gang for a sense of belonging, social recognition, and success. Marginalization is rife in countries coming out of conflict. The structural conditions that underlie marginalization, such as economic inequal ity, political exclusion, social divisiveness, and discrimination, are at the root of many civil conflicts. These conditions are rarely resolved -and are frequently exacerbatedby wars that act as ' development in reverse' (Collier 2007: 27). Internal wars often reduce incomes, exports, investment, and food pro duction and they weaken government and social institutions; their costs are rarely dis tributed equitably, and internal wars can increase individual and horizontal inequality (FakudaParr 2011).
Furthermore, forced migration produced by war often exacerbates marginalization. Migrants can face discrimination and stigma upon arrival in new communities (IOM 2015), compounding the difficulties of social and economic integration. Immigrant youth are especially vulnerable to marginaliza tion (Vigil 2002). Differences in language, culture, religion, ethnicity (van Gemert and Decker 2013), rural backgrounds, and per ceived political loyalties are used as grounds to exclude young newcomers, especially in communities under strain. The UN has noted that refugees and internally displaced persons are at increased risk of recruitment by armed groups due to the discrimination and deprivation many confront (UN 2015;IOM 2017). For forced migrants, the element of negative change may increase the risk of gang involvement beyond the usual cor relates of marginalization. Cruz (2014) cites research in El Salvador that suggests that young people who are forced to move into poor neighborhoods or drop out of school are more susceptible to joining a gang than those who have always been poor and never gone to school.

Social disorganization
Postwar research in countries involved in World War II identified a crossnational rise in juvenile delinquency and an intensifica tion of gang activity (Fyvel 1961). Cavan and Cavan (1968) found that juvenile delin quency increased during and after World War II even in countries that were not exposed to combat or occupation. War destroyed social institutions and 'postwar conditions brought new problems ' (ibid.: 195). Through the lens of gang studies, the picture that forms is social disorganization on a massive scale, with the corollary impairment of traditional mechanisms to control youth delinquency.
Although the battles of many internal wars take place in rural areas, civil conflict is linked to urban social disorder (IOM 2015). Conflict chaos enters cities when armed actors use cities to recruit, restock, and recu perate; warfueled criminal networks set up shop in urban areas; economic losses impact on urban livelihoods; and depleted state resources mean fewer or worse services in cities, including security services.
Internal displacement also interacts with social disorganization. According to IOM's World Migration Report for 2015, forced migrants often move into cities for safety and opportunities. Cities that are already reeling from the urban impacts of conflict are ill equipped to meet the basic needs of new residents. These migrants may be required to start new lives in informal, poorly planned, and spatially segregated neighborhoods with numerous risks, including violence (ibid.). IOM also suggests that urban mobility is more likely to expose newcomers to violence when migration is forced, likely due to the ' disintegration of established community dynamics that provide a degree of control over crimes ' (ibid.: 78).
Human mobility patterns around war may also connect with gangs by facilitating the transmission of gang cultures and meth ods (Maxson 1998). For example, during El Salvador's conflict, hundreds of thousands of Salvadoran families became refugees in the US. As a response to intense, layered margin alization, and their need for protection, some Salvadoran youth joined gangs, including the MS13 and 18th Street gangs (Vigil 2002). Starting in 1996, the US increased deporta tion of Salvadorans, including an unknown -but significant -number of gang mem bers (ibid.). On arriving in El Salvador, these youths again found themselves culturally dislocated and marginalized, with the added stigma of gang membership (Covey 2010). They transmitted to El Salvador's street gangs the organizational culture, brutal violence, and brands of their USbased gangs (Cruz 2014;Rodgers and Baird 2015).

Contributions from literature on violence after wars
The frequency with which countries emerg ing from armed conflict see security condi tions worsen during peacetime lends support to Boyle's hypothesis that the experience of war changes the landscape for violent actors and 'primes' postconflict states for high lev els of violence (Boyle 2014: 52). Research across countries converges around common drivers of violence in postconflict situations, including the presence of 'spoilers' who use violence to prevent peace (Stedman 2000), inadequate social services and infrastruc ture, high rates of unemployment, property related disputes, changes in power relations and the emergence of power vacuums, illegal war economies (Geneva Declaration 2008), weak governance, gaps in public security, mass trauma (Boyle 2014), former combat ants, and widespread availability of arms (Kreutz, Marsh, and Torre 2012). Although each of these could contribute to collective youth violence, the following discussion is focused on three postconflict features that intersect in obvious and recurring ways with youth gangs.

Security gaps and state responses
States struggle to establish adequate secu rity in the aftermath of conflict for at least five reasons. First, some warending accords call for security sector reform, which can debilitate enforcement capabilities in the short term (Stanley 2000). Second, where law enforcement efforts have been focused narrowly on winning a war, the state may be short on the skills required for peacetime policing (Malan 2000). Third, a state may be unable to move into zones previously domi nated by nonstate actors fast enough to fill a security vacuum (Boyle 2014). Fourth, if police played a combat role during conflict, communities may be resentful or untrusting (Malan 2000). Fifth, conflicts often take a toll on national revenues, which has impacts on service delivery across sectors, including law enforcement.
The resulting security gap leaves some parts of a country or groups feeling vulner able to violence or crime, which incentiv izes gangs. Young people frequently cite the search for protection as motivation for their entry into gangs (Howell and Griffiths 2016). Further, residents who feel that the state cannot provide for their safety may seek out nonstate security providers, including street gangs (Geneva Declaration 2008).
Faced with a security gap and a perceived gang problem, many states have added fuel to the fire by adopting repressive, militaris tic tactics under the label of zero tolerance. It is easy to see how a country emerging from armed conflict could follow a path of least resistance from a war against, say, insurgents, to a war against gangs. Hagedorn (2008) observed how one war replaced the other in El Salvador. Wartime leaders who retain overt or behindthescenes power may well be more comfortable with militaristic approaches or see these as one way to main tain status in a shifting political landscape (Mani 2000;Stanley 2000).
If postwar settings make zero tolerance a more likely choice by lawmakers, they also render it a less promising one. Dowdney (2005) found that such policies routinely fail in countries lacking social investment and economic opportunities, which are recurring features in postwar contexts. Cruz (2014: 124) described how El Salvador's harsh anti gang laws -interwoven with the 'politics of violence' or the institutions, actors, and policies that made extreme violence nor mal among youth gangs -played a decisive part in the evolution of El Salvador's maras.
In postwar settings where violence has become engrained in institutional and com munity culture, zero tolerance pits the gangs against almost everyone. As Thrasher (1927) observed, conflict forges the gangs, deepens the bond, and incentivizes better organiza tion to survive the onslaught (Howell and Griffiths 2016).

Former combatants
Former fighters sometimes challenge post war security with harmful behaviors that range from thuggery to derailing a peace process (Gamba 2006;Wessells and Jonah 2006). Based on his research in Congo and Sierra Leone, Themnér (2011) concluded that former fighters from groups that lacked economic opportunities, felt threatened, failed to reintegrate politically, retained access to weapons, and lived in insecure set tings were more likely to rearm. Kaplan and Nussio (2016) were able to identify risk fac tors in the individual domain by tracking recidivism among a sample of almost 1,500 Colombian former fighters. They found that having strong personal reasons for joining the group, the duration of a combatant's stay with the armed group, and resettlement in areas with criminal bands correlate with recidivism, while educational attainment, family bonds (including having children), and effective policing reduced risks (ibid.).
These same risk factors for recidivism can spur young former fighters to form gangs (Marcy 2014). Rodgers (2012) described how demobilized youth in Nicaragua, Guatemala, and El Salvador responded to insecurity and exclusion by starting territorial gangs or pan dillas to protect themselves and their neigh borhoods. He observed that while gangs predated civil conflict in these countries, the excombatants brought weapons and mili tary skills, resulting in better organization and more violence (ibid.). In Sierra Leone, street gangs that formed following independ ence were recruited into combat roles dur ing the civil war in 1991 (Hagedorn 2008). According to Utas (2014), in the aftermath of war, demobilized child and young soldiers recreated gang structures and infused them with more serious violence learned during war.

Illicit economies and criminal networks
The criminal underworld is a major factor in the prolongation of armed conflict and the persistence of violence after hostilities end. Boyle (2014) described how criminal networks can proliferate during wars where weak law enforcement is coupled with high societal tolerance for violence. Parties to a conflict often engage heavily in illicit com mercial activity, first to fulfill military goals and later, often, as an end in itself.
Illegal economies are dominated by adult structures designed primarily to perpetrate crime. Youth gangs are not the same thing. Generally, their members are younger and they sustain a mix of criminal and non criminal purposes (Covey 2010). With some important exceptions, youth gangs rarely play leadership roles within illegal econo mies, in part because gangs have typically lacked the discipline, organizational skill, and diligence to compete with more sophis ticated criminal networks (Hagedorn 2008; Kinnear 2009). However, the effects of illegal economies on gangs are powerful. Hagedorn (2008: 37) described how organized crime 'bends the form of area youth gangs to the needs of older armed groups.' Drugs, in particular, have altered gangs in ways that make them more problematic all over the world. Drug use is rampant in many gangs and dealing is increasingly common in many contexts (Covey 2010;Kinnear 2009;Rodgers and Baird 2015). Escalating violence by gangs is often attributed to their involve ment in the drug trade (Jütersonke, Muggah and Rodgers 2009). Looking at gangs together with other organized armed groups that involve children, Dowdney (2005) noted how the demands of drug dealing make groups more territorial, combative, and violent.

The Impact of Identified Risks on Gang Development in Colombia
Six major risk factors for postwar gang intensification were identified in the discus sion above: illicit economies and criminal networks, exposure to violence, marginaliza tion, social disorganization, security gaps and state responses, and former combatants. The following discussion situates these factors in the Colombian context to gauge their rele vance and assess the risks of gang escalation.

Illicit economies and criminal networks
Among the postwar ills that can contribute to gang growth, Colombia's gravest con cern comes from organized criminal activ ity around the production and distribution of cocaine and other illegal markets. These include illegal gold extraction, extortion, arms trafficking, and contraband (InSight Crime 2018). To assess this factor in the Colombian context, it is necessary to under stand how gangs relate to these markets and the criminal networks that dominate them. Their interactions vary enormously between cities, but the three defining char acteristics explored below are generally true in Colombia. First, youth gangs can be dis tinguished from the adult actors that domi nate the criminal underworld. Second, gang members sometimes participate, but rarely play leadership roles, in organized crime.
Third, youth gangs' relationships with organ ized crime have transformed -and continue to transform -youth gangs, with harmful consequences for youths, their communities, and national security.
In general terms, Colombian street gangs are younger, bounded to a smaller territory, and combine criminal activity with some social, identity, and noncriminal interests that set them apart from criminal bands and drugtrafficking organizations (CERAC 2014;Giraldo et al. 2015). These gangs are also more visible than their adult counterparts; they call attention to themselves and seek the recognition that comes from public mis behaving (Covey 2010;Perea 2007). Even in Medellín, where distinctions between organ ized crime and youth gangs are at their murki est (Semana 2018), the youth gangs or combos are distinguishable from older, more power ful criminal groups. Blattman et al. (2019) explain the relationships and points of differ ence between the city's street gangs or com bos and its more powerful, mafialike groups or razones. Based on interviews with combo and razón leaders, administrative crime data, and surveys of city residents, they observe that combo members are typically youths under the age of 25 who come from and reside in the neighborhood where the combo is active (ibid.). Razones manage broader, often non contiguous territories, and are primarily made up of men aged between 30 and 50 years who tend to live in Medellín's more affluent neigh borhoods (ibid.).
In many Colombian cities, criminal bands, drugtrafficking organizations, mafialike structures, or intermediaries hire gang mem bers for a variety of criminal tasks. Gang members provide labor for local drug sales, and act as messengers, lookouts, transport ers, dealers, debt collectors, and contract killers, but generally do not play leadership roles in domestic or international markets (CERAC 2014;Ordóñez Valverde 2017). Within these adult groups, gang members are generally treated as ' expendable' (Bargent 2015), and their access to upper manage ment is restricted (Semana 2018). In cities like Medellín and Cali, where recruitment by adult groups is highly structured, many gang members communicate only with interme diaries without learning the identity of their employer (DuránMartínez 2018). Youths in gangs are attractive recruits because of their knowledge of urban geography and warfare (RiañoAlcala 2006), as well as their inno vation, creativity, and technological skills (Semana 2018). Such recruitment is not new in Colombia, but the practice has increased since the 1990s as criminal structures have fragmented (Bargent 2015) and sought to exploit growing domestic markets in cocaine and marijuana (ibid.).
There are two common patterns of recruit ment in Colombia. In one, gang members are recruited by an adult organization, based on their reputation for violence and capabilities navigating and securing territory, to perform services connected to local drug sales, includ ing streetlevel dealing and transportation (Ordóñez Valverde 2017). As these gang mem bers prove themselves, they may be assigned a gun and enlisted to perpetrate additional criminal tasks, including murderforhire (MOJ 2017). In this pattern, individual gang mem bers, as opposed to the gang as a collective, work for the adult group (Castillo V 2015;Ordóñez Valverde 2017). They may be com pensated with cash, motorcycles, or weapons, and motivated by a boost in status from asso ciating with an adult group (Semana 2018).
The second pattern, which is wellestab lished in Medellín and increasingly employed by the bacrim elsewhere (Ávila 2019), involves coopting the gang as a whole. Durán Martínez (2018: 15) describes this practice as ' criminal outsourcing' or ' a situation where business activities are allocated to an exter nal source, and where lines of authority are fuzzy, but there is a clean alignment between organizations and contractor and some level of control.'

Gang transformations
These interactions with organized crime have a powerful impact on Colombian gangs. There is a sense that youth gangs are los ing space, morphing into adult structures or being squeezed out by them. The MOJ (2017: 213) suggests that the gangs' relative autonomy and fights for territory may soon be ' a thing of the past.' Gang transformations are occurring at different rates and degrees among Colombia's cities. The process is most advanced in Medellín (CERAC 2014, MOJ 2017, where the vast majority of combos are dedicated to criminal purposes and subordi nate to older, more powerful actors in organ ized crime (Blattman et al. 2019).
In Cali, where gangs and adult groups have likewise coexisted for several decades, the police believe that 31 of an estimated 104 gangs citywide have been coopted by criminal organizations, while others retain some autonomy (MOJ 2017). The Cali Mayor's Office finds support for these esti mates in its implementation of an exem plary program to reduce gang involvement. Rocío Gutiérrez Cely, Cali's Peace and Citizen Culture Secretary (2016-2019) explains that the program team has been able to inter vene in about 70 per cent of the city's gangs to persuade some members to exit. 3 While efforts are continuing among the remaining 30 per cent, Gutiérrez recognizes that some of these are too attached to organized crime to allow for members to explore exit strate gies with municipal authorities.
Gangs in other cities are less attached to adult criminal networks. For instance, Cartagena lies at the far end of the spec trum with gangs that are minimally related to organized crime (Álvarez 2016). Cartagena confronts similar conditions of marginaliza tion and social disorganization as Medellín and Cali but was not impacted to the same degree by the armed conflict or the ' criminal sagas' described in the next section.
The engagement of gangs by organized crime is deeply concerning. In their work on Central American gangs, Jütersonke, Muggah, and Rodgers (2009: 385) conclude that 'professionalization is ominous, insofar as the corrosive roles that organized crime can play in developing contexts is extremely wellknown.' Colombia's gangs become better armed, exposed to more violence, and increasingly combative over territory because of their links to organized crime (Castillo V 2015). The situation in Quibdó, the departmental capital of one of Colombia's poorest states (Chocó), illustrates a pattern that is repeated in wartorn cities around the country. Gangs there started spontaneously, spurred by con ditions of marginalization and social disor ganization with roots in the armed conflict. According to local officials, youths probably joined these gangs for social reasons, but the situation spiraled when adult groups started offering members ' an immediate future' with guns, phones, and salaries in exchange for criminal labor (Semana 2018: 55).
This association also means worse out comes for the young people who join gangs. In their research on US gangs, Howell and Griffiths (2016) explain how gang member ship, which is a step in escalating delinquent behavior, sets up an individual for increased risks into adulthood, with higher probabili ties of prison, drug use, poor health, and unemployment compared to former delin quents who did not join gangs. In Colombia, there is a worrying trend in which young people join gangs, start using drugs, start dealing, take on escalating criminal tasks for adult groups (CERAC 2014;Giraldo et al. 2015), and -in exceptional cases -become fullfledged members of adult armed struc tures (DuránMartínez 2018).
Another negative consequence of criminal outsourcing is that it appears to generate more violence than what criminal groups produce when they rely on internal human resources (DuránMartínez 2018). Based on research in five cities in Colombia and Mexico, DuránMartínez (2018: 15) explains that 'where criminals outsource to youth gangs, they…lose the ability or willingness to control their "soldiers."' Looking at Medellín, the author observes that the city's most vio lent periods are characterized by outsourcing to gangs (DuránMartínez 2018). Colombian analysts are concerned that youth gangs pro vide organized crime a ready stream of urban labor and enable more sophisticated crimi nals to evade law enforcement by reducing their visibility (Ávila 2019; Semana 2018).
Given the strength and numbers of adult criminal groups in many cities, it is remark able that many gangs still stand as distinct entities with some of the territorial, social, and identity concerns that typically charac terize youth gangs (CERAC 2014;Semana 2018). Their continued existence in cities dominated by adult criminal actors sug gests that gangs fulfill a different purpose for young people than adult criminal struc tures. The availability of criminal jobs does not eliminate the organic demand for youth gangs. Borrowing from Hagedorn (2008: 49), a criminal job may solve a gang member's problem of 'survival,' but leaves unaddressed his problem of 'meaning.'

Exposure to violence
Colombia's conflict encroached on the ev eryday life of civilians. In the government's database, the number of civilian victims since 1985 is close to nine million, primar ily victims of forced displacement, followed by homicide, threats and forced disappear ance (Unit for Assistance and Comprehensive Reparations to Victims 2020). The National Center for Historical Memory (2013: 32) reported that civilians were killed in much larger numbers than parties to the conflict: 82 per cent of the 220,000 conflict deaths were noncombatants.
Further, armed actors exerted a powerful and lasting influence in some Colombian cities. The MOJ (2017: 99-101) traced what it called ' criminal sagas,' in which different armed actors arrived in waves to Colombia's largest cities, each taking on a mythical status for the cities' youths. Cali is a case in point: in the 1980s, as part of peace negotiations, the M19 guerrilla group was authorized to set up urban camps in Cali, where members provided political training to youths; the 1990s brought in drugtrafficking organiza tions that employed brutal tactics to main tain territorial control; in the late 1990s, the ELN and FARCEP committed largescale kid nappings in Cali (MOJ 2017). This may have paved the way for the rightwing and ostensi bly antiguerrilla paramilitaries to exert more influence in Cali in the 2000s; after their demobilization in 2005, the socalled crimi nal bands or bacrim filled the vacuum (ibid.).
The incursion of armed actors into some Colombian cities exacerbated urban conflict, normalized violence, and supplied gangs with new criminal skills and brutal meth ods. Describing Medellín, Pécaut (1999: 142) characterizes urban conflict as an interplay of ' extremely heterogenous' violence, includ ing armed confrontations between guerrillas and state forces, extortion by urban militias, paramilitary operations, drug trafficking, political assassinations, delinquency, organ ized crime, and intergang warfare. Perea (2007: 2) describes youth gangs as being at the 'heart of the urban conflict,' rejected by communities, intermittently hunted down and recruited by organized crime and armed groups, and continuously battling for terri tory with other gangs.
The prolonged exposure of Colombia's cit ies to conflict has normalized violence for some urban youth. The MOJ (2017) concludes that one legacy of the criminal sagas was to position violence as the principal method of resolving disputes. Waldmann (2007: 66) sees evidence of cultural norms that pro mote violence in Colombia in its widespread incidence, high homicide rates, entrenched armed actors, and the 'glaring discrepancy between the brutality of means and mod esty of the ends pursued.' The consequences creep into everyday life. Santos (2015) notes, for instance, a high correlation between household exposure to armed conflict and domestic violence against women, leading him to argue that Colombians are 'sick with violence,' which makes them believe that violence is ' a valid mechanism to resolve dis putes.' Prolonged exposure to brutality may also deaden societal responses to loss of life. Looking at Colombia and Mexico, Lemaitre Ripoll (2014: 4) finds the absence of 'moral indignation' to homicides, especially of poor, young men of color, as a sign of widespread habituation to violence.
The Colombian case sheds light on an additional outcome of violent exposure on gangs: criminal learning. According to the MOJ (2017), in cities like Cali, Medellín, and Bogotá, youth groups learned how to com mit violent crime from a succession of armed actors, whose methods are mirrored in today's gang activity. Gangs learned about local net working and weapons from the guerrillas; workforce organization from drugtrafficking organizations; and recruitment and territorial control with 'invisible frontiers' from paramil itaries and criminal bands (ibid.).

Marginalization and social disorganization
Though costly, Colombia's conflict did not destroy the national economy as often hap pens in protracted internal wars. Poverty, extreme poverty, and, to a lesser extent, income inequality declined from the mid2000s until 2017 (Joumard andLondoño 2013, World Bank 2019). Nonetheless, economic marginalization is a significant risk factor in Colombia, which remains one of the most unequal countries globally and, contrary to expectations on post conflict economic growth, experienced a rise in both poverty and urban income inequality in 2018 (World Bank 2019).
Young people who were uprooted by armed conflict are especially prone to layered mar ginalization. A report on protracted displace ment by the UN (Kälin and Entwisle 2017) concludes that most internally displaced per sons (IDPs) in Colombia were forced to reset tle in precarious settlements or slums. Kälin and Entwisle (2017) observed that even years after displacement, these families remain at a stark economic disadvantage, and many displaced youths face stigma and discrimina tion in underserved, highcrime areas, with out access to formal education or livelihood opportunities.
Internal displacement and other conflict shocks can produce social disorganization. Based on a survey of more than 2,000 dis placed households throughout Colombia, Ibañez, Moya, and Velásquez (2006) find that forced displacement stripped victims of their customs, ways of life, and identi ties, and deprived them the support from social networks and decision makers they relied on in their places of origin. Displaced persons typically moved with their nuclear families, only rarely migrated with neigh bors or other relatives, and sharply reduced their participation in communal and civic spaces after displacement (Ibañez et al. 2006). The writers conclude that 'displace ment obstructs the formation of commu nity organizations and contributes to the destruction of social networks and social capital' (ibid.: par. 15).

Security gaps and state responses
As described above, postwar security gaps can increase gang activity both by moti vating youths to join gangs for protection and by triggering counterproductive state responses. Regarding the first effect, just as Howell and Griffiths (2016) find among US gangs, the search for protection is a common motive for a youth to join a gang in Colombia (CERAC 2014). The strength of this motive may depend on both objective and subjective factors. Objectively, Colombian youths are dis proportionately impacted by violence: almost half of all murder victims in 2016 and 2017 were aged between 15 and 29 years (Castillo 2018). Subjectively, Colombian city and town dwellers perceive themselves as highly inse cure. In a national survey from 2017, 78 per cent of respondents reported feeling fearful of being murdered, mostly by common crimi nals rather than by conflict actors (FIP 2017).
In terms of state responses, there is reason to be optimistic that Colombia can avoid the kind of zero tolerance approaches to youth gang reduction that backfired in Central America. While the Duque administration has adopted a more militaristic approach to organized criminal groups, with fewer options for negotiations (Alonso and Robbins 2019), street gangs are generally treated as issues of urban crime and youth development that fall within the purview of city mayors and police departments (International Institute for Strategic Studies 2018). Colombia was the first country in Latin America to launch citi zen security programming, with a heavy focus on rights, civic engagement, and government responsibility, and cities like Medellín, Cali, and Bogota are recognized for their innova tive preventive approaches (Cousins 2019; Muggah and Aguirre 2018;Muggah, cited in Gagne 2015).
In general terms, Colombian policy makers have demonstrated awareness of the soci etal factors at play in the creation of urban gangs and the importance of integrated carrotandstick solutions. For example, in its recent report on gangs, the MOJ's foremost recommendation to the government is to reduce the factors that make young people ' easy prey' to crime and violence by invest ing in education and economic opportuni ties in problem neighborhoods (MOJ 2017: 234). Finally, the same survey that found that Colombian city and town dwellers are fear ful of murder also found that a large majority of them (70 per cent) favor preventive meas ures, such as job and education opportunities over increasing punishment (28 per cent), as the preferred way for government to reduce homicides (FIP 2017).

Former combatants
As in other countries emerging from conflict, Colombia's former combatants are viewed as a source of insecurity. And indeed, the deci sion by some FARCEP members to rearm rep resents a serious threat to national security and the peace process (The Economist 2019b). However, demobilized combatants of the FARCEP are unlikely to join street gangs. First, while some former fighters are moving into cit ies, excombatants are mostly from rural areas and the majority have expressed an interest in working in agriculture (Kaplan 2017). Second, the demographics of the group do not align with the profiles of Colombian gang members. At the time of demobilization, the average age among the excombatants was 33, almost a quarter were women, and over half were par ents (ibid.).
While the FARCEP historically recruited children in large numbers, recruitment was sharply reduced during the peace talks, and only 123 were released under the accord, most of whom are now adults (UN 2019). The risk that children released from the FARCEP will join urban gangs is minimal. Former child combatants benefit from com prehensive rehabilitation support from the government until they reach adulthood, at which point they are entitled to adult social and economic services (Colombian Family Welfare Institute n.d.). Like their adult coun terparts, most children released by the FARC are from rural areas and many are expected to opt for rural reintegration.
More generally, Nussio (2018) argues that focusing on the relatively small number of former combatants may distract authorities from addressing potentially larger sources of insecurity. He notes that youths, especially young males, living in Colombia's marginal ized neighborhoods are at risk of becoming the 'human resources of postconflict violence' (ibid.: 143). Children and youths who have been recruited by criminal bands and drug trafficking organizations may be especially vul nerable. They rarely receive the kind of holistic and sustained support provided to children associated with paramilitaries and guerril las even though they face many of the same vulnerabilities and reintegration challenges.

Conclusion
In order to gauge Colombia's future risk of gang violence, this article drew from gang research and literature on postwar violence to identify six factors that recur in postwar settings and tend to drive gang escalation. Their application to the Colombian context suggests that Colombia's street gangs are likely to become more violent, harmful, and criminal if they continue to strengthen ties to illicit economies and criminal networks.
By looking at how Colombia's gangs have developed over time and the differences between cities, there is no doubt that asso ciating with organized crime underlies gang mutations that are injurious to gang mem bers and their communities.
Furthermore, absent concerted preventive efforts, gang membership is likely to grow among urban youth in cities where conflict exacerbated the conditions of marginaliza tion, social disorganization, and violent expo sure. Notably, Colombia's conflict has resulted in the marginalization of many young peo ple who were forcibly uprooted from their homes, saw their opportunities reduced, and put at sharp economic disadvantage. The armed conflict also damaged, and sometimes destroyed, the social institutions that enabled communities to mediate youth misconduct, resulting in widespread social disorganiza tion. Many young people have been exposed to extreme violence and socialized within neighborhoods where violence has become normal, tolerated, and even encouraged.
The risk that former combatants could create or join gangs was the least relevant of the six factors to the Colombian context. Recidivism by the FARCEP is a security threat, but members' demographic profiles and rural aspirations make it unlikely that they will engage with youth gangs in a significant way.
The final factor related to postwar inse curity and state responses. Urban insecurity and homicide rates remain high, especially for youths, and some youths will continue to seek out gangs for protection. State responses are difficult to predict, but the article posited that Colombia may well avoid socalled zerotolerance approaches to gang reduction that have proven counterproduc tive in other postwar environments.
The Colombian government can mitigate the risks identified in this article by investing in prevention, including by increasing oppor tunities for marginalized urban youths, devel oping law enforcement approaches to sever links between street gangs and organized crime, and supporting communities to design solutions locally. In doing so, Colombia can build on existing municipallevel programs to address youth violence and draw from international experience and the growing 'what works' literature -See, for example, Cunningham et al. (2008); O' Connor and Waddell (2015); Office of Juvenile Justice and Delinquency (n.d.); and WHO (2015). Colombia can also learn from the country's decadeslong efforts to prevent child recruit ment by actors to the conflict. As noted in the discussion on gang research, the risk profiles associated with vulnerability to join a violent group are surprisingly similar across contexts.