Introduction

According to reports by the National Center for Historical Memory, Medellín is one of the 20 municipalities most affected by the Colombian armed conflict in the last 30 years, and the most important of those 20 in terms of population, economics, and politics. For most of the period between 1985 and 2014, Medellín was the most violent city in the country, and one of the most violent in the world according to its homicide rate. The number of annual homicides oscillated between two thousand and six thousand from 1985 to 2003. It has been almost impossible to ascertain the influence of the city’s various wars on the number of violent deaths, but it can be assumed that they produced the majority of homicides and facilitated the disorder and obstruction of public authority that contributed to an increase in the diffusion of violence and opportunistic crime.

In 2013, various factors coincided to result in the city’s outstanding achievements in security, and its transformation into an international paradigm for relevant best practices. Medellín also became well-known for the demobilization of ex-combatants and social urbanism, to the point that US political scientists Francis Fukuyama and Seth Colby (2011) coined the expression ‘half a miracle’ to refer to Medellín’s experience.

This article examines the milestones that mark Medellín’s trajectory from being Colombia’s main urban theater for multiple national and regional armed struggles, to its current state as a model of public and private innovation. This model will, in 2015, help the city to leave the list of the 50 most violent cities in the world1 after having been in first place for fifteen years.

The paper will fulfill this purpose in three sections which will analyze: a) the accumulation of armed conflicts and the superposition of post-conflict phases in the city since 1985, while sustaining that central government interventions had the most dramatic results; b) how state strengthening and protagonism at the local level have been decisive in guaranteeing the sustainability of these security successes over the years; and c) main changes in the phenomenology of crime in the city, leading to conclusions about the challenges they pose for authorities and society.

Wars and National Interventions

This section presents an analysis of the different armed struggles that have affected Medellín in the last three decades, and the way they have impacted relevant security statistics during their peaks and declines. Throughout the section, reference will be made to the way the local state and central government have intervened to neutralize the intensity of the violence. Our main thesis is that conflict escalations that increased homicide rates exceeded local capacity for state control of the territory, and merited interventions from the central level of government.

In Medellín, conflicts and conflict neutralizations have been juxtaposed with diverse armed, insurgent, and criminal organizations. Conflicts and post-conflict phases have succeeded and replaced each other as much in neutralization (first of the cartel, then the militias, then the paramilitaries) as in the rise of new violent competitors who either disputed or claimed control over the criminal space left behind by the deactivated group. Disputes were due to activity by disobedient or rebel members who did not submit to deactivation, demobilization, or other reconfiguration processes of the criminal organizations. The dispute could also be due to limitations (if not failures) of previous deactivations.

According to statistics from the National Department of Statistics (DANE) of Colombia, 1,749 homicides were committed in Medellín in 1985, representing a rate of 113.87 per 100,000 inhabitants. Six years later (1991), there were 6,809 homicides, representing a rate of 395.74 per 100,000 residents. In numbers, the increase is dramatic (5,060 more homicides), but it does not describe the wake of violence left in Colombia’s second most important city. Those were the days of Pablo Escobar’s criminal empire and the Medellín Cartel with its direct war against the central state.

Beginning in the mid-1980s, and until recently, Medellín was caught between conflict and post-conflict phases, with partial peace and partial post-conflict agendas – a type of violent tempering of security and coexistence. We can therefore interpret the achievements reached until now as the result of a mixture of national crisis interventions and local stabilization policies. These processes have clearly not been lineal, for reasons we will examine later.

The behavior of the homicide rate in Medellín in the mid-1980s is related to cycles of armed fighting, both national and local, ranging from the Medellín Cartel against the State to more recent disputes between mid-level commanders of criminal organizations (see Graph 1). The ascending curves in the homicide rate therefore coincide with the escalation of the offensive against narco-terrorism and the militia phenomenon in the city in the 1980s, as well as the urban war strategy used by the guerrilla and paramilitary organizations in the early 2000s. The most pronounced decreases in the homicide rate resulted from specific events including: 1990 peace agreements; the dismantling of the Medellín Cartel and the death of Pablo Escobar (December 1993); Operation Orion in Comuna 13 (October 2002); and the demobilization of the Cacique Nutibara Bloc (December 2003). All of these were consequences of decisions made by the respective national governments (Giraldo-Ramírez 2008: 101).

Graph 1 

Total number of homicides and armed fights 1985 – 2014. Source: (1980–2008) DANE life statistics, (2009–2014) INML, Police SIJIN System, Attorney General’s Office CTI System. Constructed by the authors.

The peace agreements with the April 19 Movement (M-19) and the Popular Liberation Army (EPL) in 1992 overlapped with the informal agreement for Escobar’s surrender to the justice system in 1991, his escape from jail in 1992, and his death in December 1993, all of which had a notable impact in Medellín. The demobilization of the EPL included a majority of guerrilla concentration zones in Antioquia and neighboring departments, as well as the main public act of reintegration to civilian life, which occurred in Medellín. The dismantling of the Medellín Cartel, the capture and judicial processing of some of its surviving members, and the agreements between other cartels and the Cali Cartel also had significant repercussions. All of these events coincided to result in the first post-conflict stage in the city.

However, the guerrilla demobilizations and dismantling of the Medellín Cartel did not facilitate the State’s monopoly on force in the short term. Rather, it generated an ‘oligopoly of coercion’ (Duncan 2014). In 1994, the city was full of activity conducted by various violent actors: different groups of popular militias,2 the National Liberation Army (ELN) urban militia, various armed groups employed by the Medellín Cartel, and an armed group called ‘Those Persecuted by Pablo Escobar’ (Pepes) which started to transform into another group at the same time. Conditions for operations in Antioquia and Medellín by the armed forces, another of the coercive actors, were not the best. The Antioquia department presented a shameful combination of the highest homicide rate and the lowest police presence. From 1990 to 1995, among the ten most important cities in Colombia, only Córdoba and Magdalena had less police per 10,000 inhabitants than Antioquia, and only Cúcuta and Barranquilla had less manpower than Medellín, which was well below the national average (Llorente 1999: 441), notwithstanding the fact that the city continued to be the most violent in the country.

Between 1994 and 1997, the atypical ‘reintegration without demobilization’ of the popular militia (Giraldo-Ramírez & Mesa 2013), coincided with the demobilizations of other militia groups, the repositioning of violent drug trafficking actors, and the national ascent of protagonist groups in the internal armed conflict – the ELN, the Revolutionary Armed Forces of Colombia (FARC), and the United Self-Defense Forces of Colombia (AUC). This scenario comprised the second conflict and post-conflict cycle in the city.

The phenomenon of popular militias in Medellín was renewed in the late 1980s in Comunas 1, 2, 3, and 4 (northeast) and Comuna 6 (northwest).3 The agreement between popular militias and the national government was signed on May 26, 1994 and benefited 843 members of the militia group, which, according to official estimates, represented 85 per cent of the militia forces. This was a paradigmatic case because it was the first exclusively urban peace process in Colombia. In addition, it was as big as the demobilization of the AUC Cacique Nutibara Bloc, saw more weapons surrendered than the Socialist Renewal Current (CRS), had a greater man-weapon ratio than the EPL (see Table 1), and occurred at a moment when violent homicide rates were particularly high (323 per 100,000 inhabitants, and a total of 5,793 homicides in 1993).

Year Group Demobilized people Deaths Deaths/ demobilized people Man-weapon ratio

1989 M-19 1211 108 8.9 0.74
1990 EPL 3790 316 8.3 0.53
1994 CRS 1041 75 7.2 0.42
1994 MP 843 187 22.2 0.70
1997 MIR-COAR 185 12 6.5 0.92
2003 BCN 868 105 12.1 0.57
1990–1994 Collective 4715 736 15.6 0.38
1990–1997 All 10530 839 8
2003–2006 Collective 31671 1734 5.5 0.57

Table 1

Demobilized people, deaths, and man-weapon ratio in demobilizations relevant to Medellín. Source: (Giraldo-Ramírez and Mesa 2013: 226)

Information about demobilized people and deaths, by group and individually, between 1989 and 1997, comes from the Ministry of Justice’s Transitional Justice Office.

Information about the man-weapon ratio in the collective demobilizations comes from Vranckx et al 2009.

Information about demobilized people and deaths between 2003 and 2006 is from the OAS 2011.

Calculations about deaths / demobilized people and man-weapon ratios for the EPL, CRS, MP, and MIR-COAR are our own and are based on Ministry of Justice information.

Information about the BCN comes from the SIR (Reintegration Information System).

The descriptor ‘All’ indicates the total of collective and individual demobilizations in Colombia between 1989 and 1997, consolidated by the Ministry of Justice.

The negotiation process with the popular militias was innovative in that it involved the creation of a security entity formed by part of the demobilized population, called the Security and Community Service Cooperative – Coosercom. This new body had state-provided vehicles and weapons, and capabilities that ranged from arresting people and taking them to the authorities, to advising, conducting research, and carrying out crime prevention. Between June 1994 and November 1996, Medellín therefore had a legal para-police force that did not contribute to a decrease in homicide but rather allowed for the reproduction and diversification of private justice groups, prolonged the absence of the State in the urban periphery, and probably served as an immediate precedent for the extensive, massive, and complex paramilitary phenomenon in the region (Giraldo-Ramírez and Mesa 2013: 235). The failure of the demobilization management mechanism and general reintegration model Coosercom proposed, coupled with the adverse number of killings of demobilized people led to Coosercom’s termination in November 1996.4

Between 2000 and 2002, Comunas 6, 7, 8, 9, and 13 were the main centers of confrontation between the FARC and ELN guerrilla groups, the self-defense forces (Metro and Cacique Nutibara Blocs), and the armed forces. This escalation of conflict broke away from the decreasing trend in homicides, causing an increase in annual rates beginning in 1999 (see Graph 2).

Graph 2 

Homicide rate per 100,000 inhabitants 1998–2014. Source: Colombian National Police SIJIN. Constructed by authors, with data systematized from the SISC.

These illegal organizations (FARC and paramilitaries) constructed an order different from state order and, thanks to their territorial control, established enclaves where illegal structures, as opposed to legal ones, dominated. In 2001, the convergence of local citizen insecurity dynamics and armed conflict was clear, and coincided with the escalation of the war fought by national actors, and with the guerrillas’ and paramilitaries’ express decision to bring the conflict to urban contexts. This urbanization included an intense takeover process of urban illegal infrastructure represented by ‘combos’ and gangs who coordinated the criminal framework as per the logic of internal conflict.

Once again in 2002, as in 1993, a strategic state intervention fulfilled a fundamental role in citizen security in Medellín. ‘Operation Orion’ was an armed takeover of various areas of the city – especially Comuna 13 – by the State. Its objective was to terminate expansion of the armed conflict inside the urban perimeter. It was a joint operation by the Police, Army, Air Force, DAS, and Attorney General’s Office, conducted between the 16th and18th of October 2002, at which point the urbanization of the war was at a boiling point. The armed forces confronted the FARC and ELN guerrilla units and Armed People’s Commanderships (CAP). According to international parameters (designed to apply to countries), the situation in Medellín at that time fulfilled the conditions of an armed conflict given that confrontations between the State and armed groups generated more than 25 combat deaths per year.

Immediately afterwards, another national intervention presented itself, this time through diplomatic means: the demobilization of the many blocs that formed the AUC. The demobilization of the Cacique Nutibara Bloc (BCN) must be mentioned as the most recent example of how the development of conflict negotiations can determine the configuration of new conflicts or challenges in terms of citizen security. On December 25, 2003, 874 members of the AUC bloc reintegrated into society; this was the first of a series of demobilizations that would continue until 2006. In this case, there was a direct proportional relationship between demobilization and the decrease in homicides in the city, to the point that areas with a lower percentage of reintegrating people also had the lowest variation in the homicide rate. The southeast and southwest, which did not register reintegrating people, saw decreases in homicide of 9.6 per cent and 20 per cent respectively. Meanwhile, in the central-eastern area, which reintegrated 17.1 per cent of the population, there was a 43.5 per cent decrease. In the northeastern area which received 34.6 per cent of the reintegrating population, homicide dropped by 55.1 per cent. This was the beginning of a drastic structural reduction in homicide.

Generally speaking, the demobilized paramilitary assistance program has been accepted as a success, with the national government basing its reintegration policy on the model constructed and implemented in Medellín since Fajardo’s administration. Fundación Ideas para la Paz (FIP) found that as of February 2009, 86 per cent of the demobilized paramilitary members were reported as active in the program, despite the pressure on demobilized people to re-arm.

The upturn in the homicide rate between the end of 2008 and mid-2010 (see Graph 2) is the product of a confluence of various factors: a) the dispute between surviving commanders of the drug trafficking groups, residual paramilitary groups who continued after extradition, judicial processing, and internal revenge; b) the corruption crisis between the metropolitan police and attorney generals in Medellín, which led to the destitution and judicial processing of various representatives; and c) the breakdown in cooperative relationships between the government of President Álvaro Uribe and the municipal administration of Mayor Alonso Salazar.

As of 2003, the context of violent homicide in Medellín became more complex, especially after the crisis of 2009–2010, due to the absence of large organizations who could control the violence and create a clear and significant armed struggle. Since then, insecurity has been characterized by the confluence of organized crime and illegal economies. In addition, youth gangs, stirred up by petty interpersonal conflicts and who commit violent expressions of machismo and intolerance, control various territories. Finally, there is a high percentage of armed individuals predisposed to use lethal force in daily altercations.

Strengthening the Local State for Security Management

This section will explain the way the local state intervened and acted with respect to the conflicts mentioned. It is based on the thesis that there has been a gradual strengthening of the local state, which both increased social investment as a result of a positive financial situation, and progressed in the design of public policy and other local management activities related to security.

With respect to political treatment, the security problem – particularly that of violent homicide – has been on the Medellín public agenda since the 1990s. However, different mayors’ administrations have addressed it in a variety of ways.

In the militia phase and until the 1990s, the local institutional situation could be called a ‘borrowed state’ according to Leyva (2010), and is characterized by intervention by the central state, national civil organizations, and even international entities in response to the instability of local institutions. This instability resulted from municipal fiscal weakness, limitations on municipal competencies to act, minimal scope of institutional design, and the conceptualization of public policy. This led to a process that Fukuyama calls ‘from outside’ (2005) in which social groups with influence, international cooperation agencies, and the national government impact construction of the local state through social investment in infrastructure and grassroots community work. That drive, however, is inhibited by the absence of clear rules, lack of local state force, and little relative influence of the combination of actors. At this stage, there was only some governance (Leyva 2010: 278) because the local state, strictly speaking, had not been created, making the Mayor’s Office an actor, but not the protagonist.

In this context, public citizen security policy in the last decade of the twentieth century reinforced the local state’s tendency to be absent from territories under its purview, such as neighborhoods and comunas. The local state did not take charge of responsibilities related to local conflict management and regulation of community transactions in private illegal groups. This lack of action produced a disordered negotiation model that frequently saw mediators as protagonists (Alonso-Espinal, Giraldo-Ramírez & Sierra 2012). Between 1988 and 1994, local governments with low budgets, high levels of debt, and low investment capacity avoided local security management. They perceived this area to be the responsibility of the national government, while their own obligations centered on ‘coexistence’ matters such as citizen participation strategies, alternative conflict resolution mechanisms, and education. This is the political scenario in which the conflict phase stemming from the militia phenomenon occurred (Giraldo-Ramírez and Fortou 2014: 69).

The mayoralties of Sergio Naranjo (1995–1997) and Juan Gómez Martínez (1998–2000) marked the end of local government’s exclusive focus on coexistence, which had guided its work since the late 1980s. The first of these mayors progressed in negotiations with urban armed groups (particularly militias) while the second, in conjunction with support from multi-lateral organizations, promoted a conceptual framework that linked violence to poverty and social inequality. For his part, Luis Pérez (2001–2003) broke the continuity of security and coexistence-related public policy and realigned financial efforts towards assistance, thereby reinforcing the power of private mediators over municipal administrators and neighborhood communities. This facilitated the paramilitary and guerrilla groups’ urbanization of war.

Since 1999, the local state has experienced a phase of ‘forced autonomy’ involving a relative increase in investment and the gradual independence of the local body with respect to other social actors and departments of the State. However, local conditions have worsened as a result of the influence of mid and long-term development and regional order plans, despite the increasingly positive fiscal situation. The armed conflict’s influence between 1998 and 2002 has two explanations related to state policy: on the one hand, the central government avoided conflict-related language applied to urban centers, and on the other, the local government persistently denied responsibility and capability for urban security. It should also be noted that the local government in Bogota was creating a standard for local governments’ operations by the variety of actions available to municipal authorities at the time.

National government interventions through Operation Orion and the bloc demobilization process, both of which occurred in the framework of the Democratic Security policy adopted by then-President Álvaro Uribe’s administration, defined the conditions in which the following two mayoral administrations (Sergio Fajardo, 2004–2007 and Alonso Salazar, 2008–2011) would act. According to Leyva, Medellín was in a phase of ‘relational autonomy’ in which the local state becomes integral to municipal management. This occurred thanks to greater fiscal and technical capacity, and cooperation with a broad range of social actors (especially public-private partnerships), which allowed for more efficiency in local policy.

Since 2002, the most important change in local management occurred when the municipal government expressly and directly assumed responsibility for local security governance. It also continued the central role it has had since the 1980s in managing coexistence issues related to crime prevention, culture, and shared responsibility. This led to an increase in local government security investments (which continued in the following years) and interventions based on the legitimization of state authorities in the region. This included increasing investment in the most vulnerable areas, administration of public space, assistance programs for the demobilized population, logistical support and strengthening, and technology updates for security and justice entities.

Mayor Alonso Salazar denounced corruption in security and justice entities which had contributed to the increase in homicides between 2008 and 2010. The crisis was evidence of a corruption network that involved the sectional director of the attorney generals’ offices, part of the national Attorney General’s Office, a police commander, a businessman, and a criminal boss. This generated tense relationships between the local government and the metropolitan police, thereby obstructing inter-institutional coordination which had been integral to the decrease in violence since 2003. In addition, Salazar had to confront both changes in criminal dynamics due to internal disputes in organized crime that had temporarily interrupted the decrease in homicide in 2003, and problems that distanced him from the national and departmental government. This last point was especially important considering the role of high-level interventions in the de-escalation of the conflict’s urbanization.

During the administration of Mayor Aníbal Gaviria, Medellín reached its lowest homicide rate in 35 years and made the greatest regional contribution to the decrease in national-level homicides. This meant a reaffirmation of the decreasing trend in violence – empirically demonstrated since 2003 – and the consolidation of a period of relative peacefulness that young people had never seen, and that the elderly had not experienced since the 1970s. This can be attributed to factors such as the national government’s focus on its new public policy addressing urban security, including a new presidential council for this topic. Other factors include the municipal administration’s constant high-priority attention to the security and coexistence sector, as shown in three ways: better coordination at the national level, increases in security investment, and progress in the formulation of public policy. One tangible result of these factors was an increase in police manpower to levels the city had not reached since the climax in the persecution of Escobar.

One of the most important lessons of this period comes from the analysis of homicides between 2004 and 2008 and the ‘natural experiment’ presented by the phased truce and demobilization process with the paramilitaries, in addition to the jailing of their main commanders (see Graphs 1 and 2). In this period, the homicide rate descended at a dizzying speed and managed to sustain itself near 30 deaths per 100,000 people for three consecutive years (2005–2007) and near 50 for another two years (2004 and 2008). These were the lowest homicide rates in the city in three decades.

In addition, the most victimized age group went from being the youngest (18 to 24 years) to an older group (25 to 30 years). Use of firearms in homicides decreased from around 95 per cent to slightly more than 70 per cent, and the regional distribution of crime went from unstable hillside areas to formal settlements beside the Aburrá River, which traverses the center of the city. The hypothetical conclusion is that in the absence of activity by a large criminal structure, physical violence derived from other factors produced a rate of 30 homicides per 100,000 inhabitants in Medellín – double that of Bogotá, and 50 per cent more than the Latin American average.

The variations in Medellín produced by national interventions in 2002 and 2003, and by the institutional and local political shifts at the time, consolidated themselves over the next ten years. As a result, the city went through various structural changes.

The first of those changes relates to a critical imbalance seen in the homicide rate trend, which goes from a moderate descent at the end of the last century until Escobar’s surrender in 1991, to rates that soared through the symbolic barrier of 30 in 2014, and probably in 2015. As can be seen in Table 2, 2003 marked the beginning of the end for the structural rate of homicide.

Chow Breakpoint Test: 1985 2003
Null Hypothesis: No breaks at specified breakpoints
Varying regressors: All equation variables
Equation Sample: 1975 2012
F-statistic 8.551299 Prob. F(6,29) 0.0000
Log likelihood ratio 38.70569 Prob. Chi-Square(6) 0.0000
Wald Statistic 51.30779 Prob. Chi-Square(6) 0.0000
Dependent Variable: TM
Method: Least Squares
Date: 03/10/14 Time: 12:29
Sample: 1975 2012
Included observations: 38
Coefficient Std. Error t-Statistic Prob.

C 48.30000 18.02772 2.679208 0.0112
D85 171.5333 22.48451 7.628955 0.0000
D2003 –158.9333 22.48451 –7.068569 0.0000

Table 2

Test of structural change for the homicide rate.

The second significant change was due to the dramatic decrease in the homicide rate, so that while in 2002 the number of violent deaths was almost 4,000, in 2013, 2014, and probably 2015, there were less than 1,000 violent deaths annually (the number is actually closer to 600 deaths). This has brought to light two new and important pieces of data for city public policy: inequality in the provision of security-related public goods, and territorial heterogeneity in the distribution of the same. With respect to inequality, in 2013 Medellín had comunas such as El Poblado, or Popular, with homicide rates lower than those of Boston or Miami (USA) respectively, while in the city center and Comuna 13, the rates were as high as the most violent states in Mexico. In terms of heterogeneity, we can identify eleven polygons where slightly less than a third of the homicides in the period under consideration are concentrated (see Map 1)5. These polygons are dispersed throughout nine of the 16 comunas and five settlements of Medellín and are located in 53 of the 249 neighborhoods (21.3 per cent of the total). They cover an area of 11.78km2 (twelve per cent of the urban area and three per cent of the total area of the city) and in total constituted 32.1 per cent of the homicides in the eleven years analyzed.

Map 1 

Homicide polygons 2003–2013. Source: SISC 2014 Concept by Jorge Giraldo.

Challenges to Normalizing Medellín

Having analyzed these conflicts, public interventions, deactivations, and local state capacities and the way they evolved in relation to security-related public policy, this paper will conclude with a review of the challenges that Medellín faces in stabilization and decreasing homicide violence.

Perhaps the greatest challenge in maintaining the decrease of homicide rates in Medellín will be that of locational distribution. Criminal activity is not endogenous and directly relates to a ‘regional ecosystem’ characterized by an informality of economic activity, the presence of criminal activity such as the cultivation, production, and exportation of cocaine, illegal gold exploitation, contraband, and a culture of illegality (Giraldo, Eslava et al 2013). This indicates that in addition to the known relationship between criminality in Medellín and that of Valle de Aburrá, the same relationship can be argued to exist between the city, its immediate surroundings, and the rest of the Antioquia department, with all of the challenges that implies. It is also important to remember that, as was stated earlier, Antioquia has been the region most affected by the internal armed conflict in the last three decades, and was also the department with the greatest presence of armed actors. It can also be assumed that Antioquia will be a priority region for demobilization, reintegration, and related interventions in the approaching post-conflict phase that could result from the Havana-based negotiations between the national government and the FARC.

However, recent analyses also allow for new interpretations of homicide in Medellín that imply a deviation from the focus taken by past efforts to intervene.

In the last few years, particularly since 2009 when the decrease began, homicide in Medellín has been characterized very differently than in past decades. Currently, the origins of homicide violence stem from organized crime’s illegal incomes and economies, strong youth gangs with regional ties that become involved in small personal conflicts and violent expressions of machismo and intolerance, and a percentage of armed individuals predisposed to use lethal force in daily altercations.

Regional violence by ‘combos’ or youth gangs, seen in the polygons of homicide recurrence, is tied to community life. Inhabitants of neighborhoods who suffer this violence are sometimes displaced by it. However, others live with it, and even if they are not complicit, they benefit from the armed dominion exercised by family members, friends, or neighbors. This explains the recurring riots by inhabitants of these neighborhoods against the National Police as a reaction to capture or operations against gang bosses.

The challenge, then, is not a group of wayward youths seduced by crime, but is broader and relates to the entire social fabric around the ‘combo.’ This could be called ‘bad civil society,’ to use the terms of Chambers and Kopstein (2001) in that attention should be directed towards communities that resist, if not outright oppose, judicial order, coexistence norms, and morals.

Another element to take into account is the fact that 67.9 per cent of homicides from 2003 to 2013 were not linked to areas with higher levels of organized criminal activity. We could intuitively suggest that two of every three homicides corresponded to diffuse spontaneous violence and citizen intolerance, which assuredly constitutes one of the greatest challenges to the continued decrease in homicides. The fact that the greatest recurrence of homicide occurs between Friday night and Sunday morning reinforces this hypothesis.

However, a high guard must be kept with respect to organized crime, as although the technology of violence (massacre, high-powered explosives, and kidnapping) has been reduced to a minimum level,6 other activities that denote high levels of organization and relatively sophisticated value chains (such as car thieving) show that these organizations continue to gravitate towards the city (see Graph 3). Medellín currently experiences 20 per cent of the car thefts in the country.

Graph 3 

Car Thefts 2003–2014. Source: Colombian National Police SIJIN. Constructed by authors with data systematized from the SISC.

As the previous sections have demonstrated, Medellín is the Colombian city with the largest cumulative total of disarmament, demobilization, and reintegration processes. Its experience includes negotiations with militias, drug cartels, paramilitary groups, and guerrilla groups. The city has applied institutional designs from the central government and incorporated many of its own initiatives. The approaching post-conflict phase could imply some additional fragility to the elements mentioned, but there is no city in Colombia better prepared for this challenge.

Conclusion

Medellín has been highly praised by observers and international experts for its achievements in this area over the last eleven years. However, these accomplishments are partial and fragile. They are partial because the homicide rate is still eight points above the Latin American average and almost 20 points above the standard fixed by the World Health Organization to qualify as ‘epidemic rates.’ They are fragile achievements, as security is always subject to unpredictable risks and undetermined threats, but also because organized crime and a lack of adherence to norms by a large part of the population continue to be significant.

The topic of security is here to stay. It is indicative of the global context, and detaches itself from diagnostics about the Latin American situation. The increasingly probable post-conflict phase will maintain the topic of security in a spotlighted position among citizen demands and government priorities.

Given that the determining factors behind homicide in Medellín have been very dispersed, disorganized, and varied, lowering the number of points from the homicide rate will be more difficult from now on. It will demand more intelligence and less force; more efficiency than quantity in the application of resources; more focus and specificity than general measures; more culture, and less punishment.